Rule: Air Pollutants, Hazardous National Emission Standards: Hazardous Waste Combustors

Wednesday,
October 12, 2005
Part II
Environmental
Protection Agency
40 CFR Parts 9, 63, 260 et al.
National Emission Standards for
Hazardous Air Pollutants: Final Standards
for Hazardous Air Pollutants for
Hazardous Waste Combustors (Phase I
Final Replacement Standards and Phase
II); Final Rule
VerDate Aug<31>2005 20:20 Oct 11, 2005 Jkt 208001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\12OCR2.SGM 12OCR2
59402 Federal Register / Vol. 70, No. 196 / Wednesday, October 12, 2005 / Rules and Regulations
ENVIRONMENTAL PROTECTION chromium, dioxins and furans, DATES: The final rule is effective
AGENCY hydrogen chloride and chlorine gas, December 12, 2005. The incorporation
lead, manganese, and mercury. by reference of Method 0023A into
40 CFR Parts 9, 63, 260, 264, 265, 266, Exposure to these substances has been § 63.14 is approved by the Director of
270 and 271 demonstrated to cause adverse health the Federal Register as of December 12,
[FRL–7971–8] effects such as irritation to the lung, 2005.
skin, and mucus membranes, effects on ADDRESSES: The official public docket is
RIN 2050–AE01 the central nervous system, kidney the collection of materials that is
National Emission Standards for damage, and cancer. The adverse health available for public viewing at the
Hazardous Air Pollutants: Final effects associated with exposure to these Office of Air and Radiation Docket and
Standards for Hazardous Air Pollutants specific HAP are further described in Information Center (Air Docket) in the
for Hazardous Waste Combustors the preamble. For many HAP, these EPA Docket Center, Room B–102, 1301
(Phase I Final Replacement Standards findings have only been shown with Constitution Ave., NW., Washington,
and Phase II) concentrations higher than those DC.
typically in the ambient air. FOR FURTHER INFORMATION CONTACT: For
AGENCY: Environmental Protection
This action also presents our decision more information concerning
Agency (EPA).
regarding the February 28, 2002 petition applicability and rule determinations,
ACTION: Final rule.
for rulemaking submitted by the Cement contact your State or local
SUMMARY: This action finalizes national Kiln Recycling Coalition, relating to representative or appropriate EPA
emission standards (NESHAP) for EPA’s implementation of the so-called Regional Office representative. For
hazardous air pollutants for hazardous omnibus permitting authority under information concerning rule
waste combustors (HWCs): hazardous section 3005(c) of the Resource development, contact Michael
waste burning incinerators, cement Conservation and Recovery Act (RCRA). Galbraith, Waste Treatment Branch,
kilns, lightweight aggregate kilns, That section requires that each permit Hazardous Waste Minimization and
industrial/commercial/institutional issued under RCRA contain such terms Management Division, (5302W), U.S.
boilers and process heaters, and and conditions as permit writers EPA, 1200 Pennsylvania Avenue, NW.,
hydrochloric acid production furnaces. determine to be necessary to protect Washington DC 20460, telephone
EPA has identified HWCs as major human health and the environment. In number (703) 605–0567, fax number
sources of hazardous air pollutant that petition, the Cement Kiln Recycling (703) 308–8433, electronic mail address
(HAP) emissions. These standards Coalition requested that we repeal the galbraith.michael@epa.gov.
implement section 112(d) of the Clean existing site-specific risk assessment SUPPLEMENTARY INFORMATION:
Air Act (CAA) by requiring hazardous policy and technical guidance for
hazardous waste combustors and that Regulated Entities
waste combustors to meet HAP emission
standards reflecting the performance of we promulgate the policy and guidance The promulgation of the final rule
the maximum achievable control as rules in accordance with the would affect the following North
technology (MACT). Administrative Procedure Act if we American Industrial Classification
The HAP emitted by HWCs include continue to believe that site-specific risk System (NAICS) and Standard Industrial
arsenic, beryllium, cadmium, assessments may be necessary. Classification (SIC) codes:
SIC
Category NAICS code Examples of potentially regulated entities
code
Any industry that combusts hazardous waste as defined in

the final rule.
562211 4953 Incinerator, hazardous waste
327310 3241 Cement manufacturing, clinker production
327992 3295 Ground or treated mineral and earth manufac-
turing
325 28 Chemical Manufacturers
324 29 Petroleum Refiners
331 33 Primary Aluminum
333 38 Photographic equipment and supplies
488, 561, 562 49 Sanitary Services, N.E.C.
421 50 Scrap and waste materials
422 51 Chemical and Allied Products, N.E.C
512, 541, 561, 812 73 Business Services, N.E.C.
512, 514, 541, 711 89 Services, N.E.C.
924 95 Air, Water and Solid Waste Management
This table is not intended to be company, business, organization, etc., is Abbreviations and Acronyms Used in
exhaustive, but rather provides a guide regulated by this action, you should This Document
for readers regarding entities likely to be examine the applicability criteria in Part acfm actual cubic feet per minute
regulated by this action. This table lists II of this preamble. If you have any Btu British thermal units
examples of the types of entities EPA is questions regarding the applicability of CAA Clean Air Act
now aware could potentially be this action to a particular entity, consult CFR Code of Federal Regulations
regulated by this action. Other types of the person listed in the preceding FOR DRE destruction and removal efficiency
entities not listed could also be affected. FURTHER INFORMATION CONTACT section. dscf dry standard cubic foot
To determine whether your facility, dscm dry standard cubic meter
Federal Register / Vol. 70, No. 196 / Wednesday, October 12, 2005 / Rules and Regulations 59403
EPA Environmental Protection Agency F. Testing Requirements IV. Use of Surrogates

FR Federal Register G. Monitoring Requirements A. Particulate Matter as Surrogate for Metal
gr/dscf grains per dry standard cubic foot H. Relationship Among Emission Rates, HAP
HAP hazardous air pollutant(s) Emission Rate Limits, and Feedrate B. Carbon Monoxide/Hydrocarbons and
ICR Information Collection Request Limits DRE as Surrogates for Dioxin/Furan
kg/hr kilograms per hour I. Changes C. Use of Carbon Monoxide and Total
kW-hour kilo Watt hour X. Overview on Floor Methodologies Hydrocarbons as Surrogate for Non-
MACT Maximum Achievable Control Part Three: What Are the Major Changes Dioxin Organic HAP
Technology Since Proposal? V. Additional Issues Relating to Variability
mg/dscm milligrams per dry standard cubic I. Database and Statistics
meter A. Hazardous Burning Incinerators A. Data Sets Containing Nondetects
MMBtu million British thermal unit B. Hazardous Waste Cement Kilns B. Using Statistical Imputation to Address
ng/dscm nanograms per dry standard cubic C. Hazardous Waste Lightweight Aggregate Variability of Nondetect Values
meter Kilns C. Analysis of Variance Procedures to
NESHAP national emission standards for D. Liquid Fuel Boilers Assess Subcategorization
HAP E. HCl Production Furnaces VI. Emission Standards
ng nanograms F. Total Chlorine Emissions Data Below 20 A. Incinerators
POHC principal organic hazardous ppmv B. Cement Kilns
constituent II. Emission Limits C. Lightweight Aggregate Kilns
ppmv parts per million by volume A. Incinerators D. Liquid Fuel Boilers
ppmw parts per million by weight B. Hazardous Waste Burning Cement Kilns E. General
Pub. L. Public Law C. Hazardous Waste Burning Lightweight VII. Health-Based Compliance Alternative for
RCRA Resource Conservation and Recovery Aggregate Kilns Total Chlorine
Act D. Solid Fuel Boilers A. Authority for Health-Based Compliance
SRE system removal efficiency E. Liquid Fuel Boilers Alternatives
TEQ toxicity equivalence F. Hydrochloric Acid Production Furnaces B. Implementation of the Health-Based
µg/dscm micrograms per dry standard cubic G. Dioxin/Furan Testing for Sources Not Standards
meter Subject to a Numerical Standard C. National Health-Based Standards for
U.S.C. United States Code III. Statistics and Variability Cement Kilns.
A. Using Statistical Imputation to Address VIII. Implementation and Compliance
Table of Contents Variability of Nondetect Values A. Compliance Assurance Issues for both
Part One: Background and Summary B. Degrees of Freedom when Imputing a Fabric Filters and Electrostatic
I. What Is the Statutory Authority for this Standard Deviation Using the Universal Precipitators (and Ionizing Wet
Standard? Variability Factor for Particulate Matter Scrubbers)
II. What Is the Regulatory Development Controlled by a Fabric Filter B. Compliance Assurance Issues for Fabric
Background of the Source Categories in IV. Compliance Assurance for Fabric Filters, Filters
the Final Rule? Electrostatic Precipitators, and Ionizing C. Compliance Issues for Electrostatic
A. Phase I Source Categories Wet Scrubbers Precipitators and Ionizing Wet Scrubbers
B. Phase II Source Categories V. Health-Based Compliance Alternative for D. Fugitive Emissions
III. How Was the Final Rule Developed? Total Chlorine E. Notification of Intent to Comply and
IV. What Is the Relationship Between the Part Four: What Are the Responses to Major Compliance Progress Report
Final Rule and Other MACT Combustion Comments? F. Startup, Shutdown, and Malfunction
Rules? I. Database Plan
V. What Are the Health Effects Associated A. Revisions to the EPA’s Hazardous Waste G. Public Notice of Test Plans
with Pollutants Emitted by Hazardous Combustor Data Base H. Using Method 23 Instead of Method
Waste Combustors? B. Use of Data from Recently Upgraded 0023A
Part Two: Summary of the Final Rule Sources I. Extrapolating Feedrate Limits for
I. What Source Categories and Subcategories C. Correction of Total Chlorine Data to Compliance with the Liquid Fuel Boiler
Are Affected by the Final Rule? Address Potential Bias in Stack Mercury and Semivolatile Metal
II. What Are the Affected Sources and Measurement Method Standards
Emission Points? D. Mercury Data for Cement Kilns J. Temporary Compliance with Alternative,
III. What Pollutants Are Emitted and E. Mercury Data for Lightweight Aggregate Otherwise Applicable MACT Standards
Controlled? Kilns K. Periodic DRE Testing and Limits on
IV. Does the Final Rule Apply to Me? F. Incinerator Database Minimum Combustion Chamber
V. What Are the Emission Limitations? II. Affected Sources Temperature for Cement Kilns
VI. What Are the Testing and Initial A. Area Source Boilers and Hydrochloric L. One Time Dioxin and Furan Test for
Compliance Requirements? Acid Production Furnaces Sources Not Subject to a Numerical
A. Compliance Dates B. Boilers Eligible for the RCRA Low Risk Limit for Dioxin and Furan
B. Testing Requirements Waste Exemption M. Miscellaneous Compliance Issues
C. Initial Compliance Requirements C. Mobile Incinerators IX. Site-Specific Risk Assessment under
VII. What Are the Continuous Compliance III. Floor Approaches RCRA
Requirements? A. Variability A. What Is the Site-Specific Risk
VIII. What Are the Notification, B. SRE/Feed Methdology Assessment Policy?
Recordkeeping, and Reporting C. Air Pollution Control Technology B. Why Might SSRAs Continue To Be
Requirements? Methodologies for the Particulate Matter Necessary for Sources Complying With
IX. What Is the Health-Based Compliance Standard and for the Total Chlorine Phase 1 Replacement Standards and
Alternative for Total Chlorine, and How Standard for Hydrochloric Acid Phase 2 Standards?
Do I Demonstrate Eligibility? Production Furnaces C. What Changes Are EPA Finalizing With
A. Overview D. Format of Standards Respect To the Site-Specific Risk
B. HCl-Equivalent Emission Rates E. Standards Can Be No Less Stringent Assessment Policy?
C. Eligibility Demonstration Than the Interim Standards D. How Will the New SSRA Regulatory
D. Assurance that the 1-Hour HCl- F. How Can EPA’s Approach to Assessing Provisions Work?
Equivalent Emission Rate Will Not Be Variability and its Ranking E. What Were Commenters’ Reactions to
Exceeded Methodologies be Reasonable when they EPA’s Proposed Decision Not to Provide
E. Review and Approval of Eligibility Result in Standards Higher than the National Criteria for Determining When
Demonstrations Interim Standards? an SSRA Is or Is Not Necessary?
F. What Are EPA’s Responses to the Minority Populations and Low-Income regulatory structure overseeing the safe
Cement Kiln Recycling Coalition’s Populations treatment, storage, and disposal of
Comments on the Proposal and What is XI. Congressional Review hazardous waste. We issued RCRA rules
EPA’s Final Decision on CKRC’s to control air emissions from hazardous
Petition? Part One: Background and Summary
X. Permitting waste burning incinerators in 1981, 40
I. What Is the Statutory Authority for CFR Parts 264 and 265, Subpart O, and
A. What is the Statutory Authority for the This Standard?
RCRA Requirements Discussed in this from cement kilns and lightweight
Section? Section 112 of the Clean Air Act aggregate kilns that burn hazardous
B. Did Commenters Express any Concerns requires that the EPA promulgate waste in 1991, 40 CFR Part 266, Subpart
Regarding the Current Permitting regulations requiring the control of HAP H. These rules rely generally on risk-
Requirements? emissions from major and certain area based standards to assure control
C. Are There Any Changes to the Proposed sources. The control of HAP is achieved necessary to protect human health and
Class 1 Permit Modification Procedure? through promulgation of emission the environment, the applicable RCRA
D. What Permitting Approach Is EPA
standards under sections 112(d) and (in standard. See RCRA section 3004 (a)
Finalizing for New Units?
E. What Other Permitting Requirements a second round of standard setting) (f). and (q).
Were Discussed In the Proposal? EPA’s initial list of categories of major The Phase I source categories also are
Part Five: What Are the CAA Delegation and area sources of HAP selected for subject to standards under the Clean Air
Clarifications and RCRA State regulation in accordance with section Act. We promulgated standards for
Authorization Requirements? 112(c) of the Act was published in the Phase I sources on September 30, 1999
I. Authority for this Rule. Federal Register on July 16, 1992 (57 FR (64 FR 52828). This final rule is referred
II. CAA Delegation Authority. 31576). Hazardous waste incinerators, to in this preamble as the Phase I rule
III. Clarifications to CAA Delegation Portland cement plants, clay products or 1999 final rule. These emission
Provisions for Subpart EEE. manufacturing (including lightweight standards created a technology-based
A. Alternatives to Requirements. aggregate kilns), industrial/commercial/
B. Alternatives to Test Methods.
national cap for hazardous air pollutant
C. Alternatives to Monitoring.
institutional boilers and process heaters, emissions from the combustion of
D. Alternatives to Recordkeeping and and hydrochloric acid production hazardous waste in these devices. The
Reporting. furnaces are among the listed 174 rule regulates emissions of numerous
E. Other Delegation Provisions categories of sources. The listing was hazardous air pollutants: dioxin/furans,
IV. RCRA State Authorization and based on the Administrator’s other toxic organics (through
Amendments To the RCRA Regulations. determination that these sources may surrogates), mercury, other toxic metals
Part Six: Impacts of the Final Rule reasonably be anticipated to emit one or (both directly and through a surrogate),
I. What Are the Air Impacts? more of the 186 listed HAP in quantities and hydrogen chloride and chlorine gas.
II. What Are the Water and Solid Waste sufficient to designate them as major Where necessary, Section 3005(c)(3) of
Impacts? sources.
III. What Are the Energy Impacts?
RCRA provides the authority to impose
IV. What Are the Control Costs? II. What Is the Regulatory Development additional conditions on a source-by-
V. What Are the Economic Impacts? Background of the Source Categories in source basis in a RCRA permit if
A. Market Exit Estimates the Final Rule? necessary to protect human health and
B. Waste Reallocations the environment.
VI. What Are the Social Costs and Benefits Today’s notice finalizes standards for A number of parties, representing
of the Final Rule? controlling emissions of HAP from interests of both industrial sources and
A. Combustion Market Overview hazardous waste combustors: of the environmental community,
B. Baseline Specification incinerators, cement kilns, lightweight sought judicial review of the Phase I
C. Analytical Methodology and Findings— aggregate kilns, boilers, process rule. On July 24, 2001, the United States
Social Cost Analysis heaters 1, and hydrochloric acid Court of Appeals for the District of
D. Analytical Methodology and Findings— production furnaces that burn
Benefits Assessment Columbia Circuit granted portions of the
hazardous waste. We call incinerators, Sierra Club’s petition for review and
Part Seven: How Does the Final Rule Meet
the RCRA Protectiveness Mandate? cement kilns, and lightweight aggregate vacated the challenged portions of the
I. Background kilns Phase I sources because we have standards. Cement Kiln Recycling
II. Evaluation of Protectiveness already promulgated standards for those Coalition v. EPA, 255 F. 3d 855 (D.C.
Part Eight: Statutory and Executive Order source categories. We call boilers and Cir. 2001). The court held that EPA had
Reviews hydrochloric acid production furnaces not demonstrated that its calculation of
I. Executive Order 12866: Regulatory Phase II sources because we intended to MACT floors met the statutory
Planning and Review promulgate MACT standards for those requirement of being no less stringent
II. Paperwork Reduction Act source categories after promulgating than (1) the average emission limitation
III. Regulatory Flexibility Act
MACT standards for Phase I sources. achieved by the best performing 12
IV. Unfunded Mandates Reform Act of 1995
V. Executive Order 13132: Federalism The regulatory background of Phase I percent of existing sources and, for new
VI. Executive Order 13175: Consultation and and Phase II source categories is sources, (2) the emission control
Coordination with Indian Tribal discussed below. achieved in practice by the best
Governments
A. Phase I Source Categories controlled similar source for new
VII. Executive Order 13045: Protection of sources. 255 F.3d at 861, 865–66. As a
Children from Environmental Health Phase I combustor sources are remedy, the court, after declining to rule
Risks and Safety Risks regulated under the Resource on most of the issues presented in the
VIII. Executive Order 13211: Actions Conservation and Recovery Act (RCRA), industry petitions for review, vacated
Concerning Regulations that which establishes a ‘‘cradle-to-grave’’ the ‘‘challenged regulations,’’ stating
Significantly Affect Energy Supply,
Distribution, or Use that: ‘‘[W]e have chosen not to reach the
1 A process heater meets the RCRA definition of
IX. National Technology Transfer and bulk of industry petitioners’ claims, and
a boiler. Therefore, process heaters that burn
Advancement Act hazardous wastes are covered under subpart EEE as
leaving the regulations in place during
X. Executive Order 12898: Federal Actions to boilers, and are discussed as such in subsequent remand would ignore petitioners’
Address Environmental Justice in parts of the preamble. potentially meritorious challenges.’’ Id.
at 872. Examples of the specific III. How Was the Final Rule Developed? promulgated separate MACT standards
challenges the Court indicated might We proposed standards for HWCs on for sources that do not burn hazardous
have merit were provisions relating to April 20, 2004 (69 FR 21197). The waste within the following source
compliance during start up/shut down public comment period closed on July 6, categories: commercial and industrial
and malfunction events, including 2004. In addition, on February 4, 2005, solid waste incinerators (40 CFR Part 60,
emergency safety vent openings, the we requested certain key commenters to Subparts CCCC and DDDD); Portland
dioxin/furan standard for lightweight comment by email on a limited number cement manufacturing facilities (40 CFR
aggregate kilns, and the semivolatile of issues arising from public comments
Part 63, Subpart LLL); industrial/
metal standard for cement kilns. Id. commercial/institutional boilers and
on the proposed rule. The comment
However, the Court stated, ‘‘[b]ecause process heaters (40 CFR Part 63, Subpart
period for those issues closed on March
this decision leaves EPA without DDDDD); and hydrochloric acid
7, 2005.
standards regulating [hazardous waste production facilities (40 CFR Part 63,
We received approximately 100
combustor] emissions, EPA (or any of Subpart NNNNN). In addition, EPA
public comment letters on the proposed
the parties to this proceeding) may file considered whether to establish MACT
rule and the subsequent direct request
a motion to delay issuance of the standards for lightweight aggregate
for comments. Comments were
mandate to request either that the manufacturing facilities that do not burn
submitted by owner/operators of HWCs,
current standards remain in place or hazardous waste, and determined that
trade associations, state regulatory
that EPA be allowed reasonable time to they are not major sources of HAP
agencies and their representatives, and
develop interim standards.’’ Id. emissions. Thus, EPA has not
Acting on this invitation, all parties environmental groups. Today’s final
established MACT standards for
moved the Court jointly to stay the rule reflects our consideration of all of
lightweight aggregate manufacturing
issuance of its mandate for four months the comments and additional
facilities that do not burn hazardous
to allow EPA time to develop interim information we received. Major public
waste.
standards, which would replace the comments on the proposed rule along
Note that non-stack emissions points
vacated standards temporarily, until with our responses, are summarized in
are not regulated under Subpart EEE.2
final standards consistent with the this preamble.
Emissions attributable to storage and
Court’s mandate are promulgated. The IV. What Is the Relationship Between handling of hazardous waste prior to
interim standards were published on the Final Rule and Other MACT combustion (i.e., emissions from tanks,
February 13, 2002 (67 FR 6792). EPA Combustion Rules? containers, equipment, and process
did not justify or characterize these vents) would continue to be regulated
standards as conforming to MACT, but The amendments to the Subpart EEE,
Part 63, standards for hazardous waste pursuant to either RCRA Subpart AA,
rather as an interim measure to prevent BB, and CC and/or an applicable MACT
adverse consequences that would result combustors apply to the source
categories that are currently subject to that applies to the before-mentioned
from the regulatory gap resulting from material handling devices. Emissions
no standards being in place. Id. at 6793, that subpart—incinerators, cement
kilns, and lightweight aggregate kilns unrelated to the hazardous waste
6795–96; see also 69 FR at 21217 (April operations may be regulated pursuant to
20, 2004). EPA also entered into a that burn hazardous waste. Today’s final
rule, however, also amends Subpart EEE other MACT rulemakings. For example,
settlement agreement, enforceable by Portland cement manufacturing
the Court of Appeals, to issue final to establish MACT standards for the
Phase II source categories—those boilers facilities that combust hazardous waste
standard conforming to the Court’s are subject to both Subpart EEE and
mandate by June 14, 2005. That date has and hydrochloric acid production
furnaces that burn hazardous waste. Subpart LLL, and hydrochloric acid
since been extended to September 14, production facilities that combust
2005. Generally speaking, you are an
affected source pursuant to Subpart EEE hazardous waste may be subject to both
B. Phase II Source Categories if you combust, or have previously Subpart EEE and Subpart NNNNN.3 In
combusted, hazardous waste in an these instances Subpart EEE controls
Phase II combustors—boilers and
incinerator, cement kiln, lightweight HAP emissions from the cement kiln
hydrochloric acid production
aggregate kiln, boiler, or hydrochloric and hydrochloric acid production
furnaces—are also regulated under the
acid production furnace. You continue furnace stack, while Subparts LLL and
Resource Conservation and Recovery
Act (RCRA) pursuant to 40 CFR Part to be an affected source until you cease NNNNN would control HAP emissions
266, Subpart H, and (for reasons burning hazardous waste and initiate from other operations that are not
discussed below) are also subject to the closure requirements pursuant to RCRA. directly related to the combustion of
MACT standard setting process in Affected sources do not include: (1) hazardous waste (e.g., clinker cooler
section 112(d) of the CAA. We delayed Sources exempt from regulation under emissions for cement production
promulgating MACT standards for these 40 CFR part 266, subpart H, because the facilities, and hydrochloric acid product
source categories pending reevaluation only hazardous waste they burn is listed transportation and storage for
of the MACT standard-setting under 40 CFR 266.100(c); (2) research, hydrochloric acid production facilities).
Note that if you temporarily cease
methodology following the Court’s development, and demonstration
burning hazardous waste for any reason,
decision to vacate the standards for the sources exempt under § 63.1200(b); and
you remain an affected source and are
Phase I source categories. We also have (3) boilers exempt from regulation under
still subject to the applicable Subpart
entered into a judicially enforceable 40 CFR part 266, subpart H, because
consent decree with Sierra Club that they meet the definition of small 2 Note, however, that fugitive emissions
requires EPA to promulgate MACT quantity burner under 40 CFR 266.108. attributable to the combustion of hazardous waste
standards for the Phase II sources by See § 63.1200(b). from the combustion device are regulated pursuant
June 14, 2005, since extended to If you never previously combusted to Subpart EEE.
3 Hydrochloric acid production furnaces that
September 14, 2005—the same date that hazardous waste, or have ceased
combust hazardous waste are also affected sources
(for independent reasons) is required for burning hazardous waste and initiated subject to Subpart NNNNN if they produce a liquid
the replacement standards for Phase I RCRA closure requirements, you are not acid product that contains greater than 30%
sources. subject to Subpart EEE. Rather, EPA has hydrochloric acid.
EEE requirements. However, even as an higher rates of premature deliveries) inadequate to determine whether
affected source, the emission standards have been observed in female workers beryllium is carcinogenic when
or operating limits do not apply if: (1) exposed in an antimony processing ingested.
Hazardous waste is not in the facilities. Similar effects on the heart,
Cadmium
combustion chamber and you elect to lungs, and reproductive system have
comply with other MACT (or CAA been observed in laboratory animals. Chronic inhalation or oral exposure to
section 129) standards that otherwise EPA assessed the carcinogenicity of cadmium leads to a build-up of
would be applicable if you were not antimony and found the evidence for cadmium in the kidneys that can cause
burning hazardous waste, e.g., the carcinogenicity to be weak, with kidney disease. Cadmium has been
nonhazardous waste burning Portland conflicting evidence from inhalation shown to be a developmental toxicant in
Cement Kiln MACT (Subpart LLL); or studies with laboratory animals, animals, resulting in fetal malformations
(2) you are in a startup, shutdown, or equivocal data from the occupational and other effects, but no conclusive
malfunction mode of operation. studies, negative results from studies of evidence exists in humans. An
oral exposures in laboratory animals, association between cadmium exposure
V. What Are the Health Effects and little evidence of mutagenicity or and an increased risk of lung cancer has
Associated With Pollutants Emitted by genotoxicity.4 As a consequence, EPA been reported from human studies, but
Hazardous Waste Combustors? concluded that insufficient data are these studies are inconclusive due to
Today’s final rule protects air quality available to adequately characterize the confounding factors. Animal studies
and promotes the public health by carcinogenicity of antimony and, have demonstrated an increase in lung
reducing the emissions of some of the accordingly, the carcinogenicity of cancer from long-term inhalation
HAP listed in Section 112(b)(1) of the antimony cannot be determined based exposure to cadmium. EPA has
CAA. Emissions data collected in the on available information. However, the classified cadmium as a Group B1,
development of this final rule show that International Agency for Research on probable carcinogen.
metals, hydrogen chloride and chlorine Cancer in an earlier evaluation, Chlorine gas
gas, dioxins and furans, and other concluded that antimony trioxide is
organic compounds are emitted from ‘‘possibly carcinogenic to humans’’ Chlorine is an irritant to the eyes, the
hazardous waste combustors. The HAP (Group 2B). upper respiratory tract, and lungs.
that would be controlled with this rule Chronic exposure to chlorine gas in
are associated with a variety of adverse Arsenic workers has resulted in respiratory
health affects. These adverse health Chronic (long-term) inhalation effects including eye and throat
effects include chronic health disorders exposure to inorganic arsenic in humans irritation and airflow obstruction. No
(e.g., irritation of the lung, skin, and is associated with irritation of the skin information is available on the
mucus membranes and effects on the and mucous membranes. Human data carcinogenic effects of chlorine in
blood, digestive tract, kidneys, and suggest a relationship between humans from inhalation exposure. A
central nervous system), and acute inhalation exposure of women working National Toxicology Program (NTP)
health disorders (e.g., lung irritation and at or living near metal smelters and an study showed no evidence of
congestion, alimentary effects such as increased risk of reproductive effects, carcinogenic activity in male rats or
nausea and vomiting, and effects on the such as spontaneous abortions. male and female mice, and equivocal
central nervous system). Provided below Inorganic arsenic exposure in humans evidence in female rats, from ingestion
are brief descriptions of risks associated by the inhalation route has been shown of chlorinated water. The EPA has not
with HAP that are emitted from to be strongly associated with lung classified chlorine for potential
hazardous waste combustors. cancer, while ingestion or inorganic carcinogenicity. In the absence of
arsenic in humans has been linked to a specific scientific evidence to the
Antimony form of skin cancer and also to bladder, contrary, it is the Agency’s policy to
Antimony occurs at very low levels in liver, and lung cancer. EPA has classify noncarcinogenic effects as
the environment, both in the soils and classified inorganic arsenic as a Group threshold effects. RfC development is
foods. Higher concentrations, however, A, human carcinogen. the default approach for threshold (or
are found at antimony processing sites, nonlinear) effects.
and in their hazardous wastes. The most Beryllium
common industrial use of antimony is Chronic inhalation exposure of Chromium
as a fire retardant in the form of humans to high levels of beryllium has Chromium may be emitted in two
antimony trioxide. Chronic been reported to cause chronic forms, trivalent chromium (chromium
occupational exposure to antimony beryllium disease (berylliosis), in which III) or hexavalent chromium (chromium
(generally antimony trioxide) is most granulomatous (noncancerous) lesions VI). The respiratory tract is the major
commonly associated with ‘‘antimony develop in the lung. Inhalation exposure target organ for chromium VI toxicity for
pneumoconiosis,’’ a condition involving to high levels of beryllium has been inhalation exposures. Bronchitis,
fibrosis and scarring of the lung tissues. demonstrated to cause lung cancer in decreases pulmonary function,
Studies have shown that antimony rats and monkeys. Human studies are pneumonia, and other respiratory effects
accumulates in the lung and is retained limited, but suggest a causal have been noted from chronic high does
for long periods of time. Effects are not relationship between beryllium exposure in occupational settings due to
limited to the lungs, however, and exposure and an increased risk of lung chromium VI. Limited human studies
myocardial effects (effects on the heart cancer. We have classified beryllium as suggest that chromium VI inhalation
muscle) and related effects (e.g., a Group B1, probable human exposure may be associated with
increased blood pressure, altered EKG carcinogen, when inhaled; data are complications during pregnancy and
readings) are among the best- childbirth, while animal studies have
4 See ‘‘Evaluating THe Carcinogenicity of
characterized human health effects not reported reproductive effects from
Antimony,’’ Rish Assessment Issue Paper (98–030/
associated with antimony exposure. 07–26–99), Superfund Technical Support Center,
inhalation exposure to chromium VI.
Reproductive effects (increased National Center for Environmental Assessment, July Human and animal studies have clearly
incidence of spontaneous abortions and 26, 1999. established that inhaled chromium VI is
a carcinogen, resulting in an increased Dioxins and Furans reproductive effects, depending on dose
risk of lung cancer. EPA has classified Exposures to 2,3,7,8- received, which varied widely in the
chromium VI as a Group A, human tetrachlorodibenzo-p-dioxin (2,3,7,8– experiments.9
carcinogen. TCDD) and related compounds at levels Hydrogen chloride/hydrochloric acid
Chromium III is less toxic than 10 times or less above those modeled to
chromium VI. The respiratory tract is Hydrogen chloride, also called
approximate average background
also the major target organ for hydrochloric acid, is corrosive to the
exposure have resulted in adverse non-
chromium III toxicity, similar to eyes, skin, and mucous membranes.
cancer health effects in animals. This
chromium VI. Chromium III is an Chronic (long-term) occupational
statement is based on assumptions
essential element in humans, with a exposure to hydrochloric acid has been
about the toxic equivalent for these
daily intake of 50 to 200 micrograms per reported to cause gastritis, bronchitis,
compounds, for which there is
day recommended for an adult. The and dermatitis in workers. Prolonged
acknowledged uncertainty. These effects
body can detoxify some amount of exposure to low concentrations may
include changes in hormone systems,
chromium VI to chromium III. EPA has also cause dental discoloration and
alterations in fetal development,
not classified chromium III with respect erosion. No information is available on
reduced reproductive capacity, and
to carcinogenicity. the reproductive or developmental
immunosuppression. Effects that may be
effects of hydrochloric acid in humans.
Cobalt linked to dioxin and furan exposures at
In rats exposed to hydrochloric acid by
low dose in humans include changes in
Cobalt is a relatively rare metal that is inhalation, altered estrus cycles have
markers of early development and
produced primarily as a by-product been reported in females and increased
hormone levels. Dioxin and furan
during refining of other metals, fetal mortality and decreased fetal
exposures are associated with altered
especially copper. Cobalt has been weight have been reported in offspring.
liver function and lipid metabolism
widely reported to cause respiratory EPA has not classified hydrochloric acid
changes in activity of various liver
effects in humans exposed by for carcinogenicity. In the absence of
enzymes, depression of the immune
inhalation, including respiratory specific scientific evidence to the
system, and endocrine and nervous
irritation, wheezing, asthma, and contrary, it is the Agency’s policy to
system effects. EPA in its 1985 dioxin
pneumonia. Cardiomyopathy (damage classify noncarcinogenic effects as
assessment classified 2,3,7,8–TCDD as a
to the heart muscle) has also been threshold effects. RfC development is
probable human carcinogen. The
reported, although this effect is better the default approach for threshold (or
International Agency for Research on
known from oral exposure. Other effects nonlinear) effects.
Cancer (IARC) concluded in 1997 that
of oral exposure in humans are the overall weight of the evidence was Lead
polycythemia (an abnormally high sufficient to characterize 2,3,7,8–TCDD
number of red blood cells) and the Lead can cause a variety of effects at
as a known human carcinogen.6 In 2001 low dose levels. Chronic exposure to
blocking of uptake of iodine by the the U.S. Department of Health and
thyroid. In addition, cobalt is a high levels of lead in humans results in
Human Services National Toxicology effects on the blood, central nervous
sensitizer in humans by any route of Program in their 9th Report on
exposure. Sensitized individuals may system, blood pressure, and kidneys.
Carcinogens classified 2,3,7,8–TCDD as Children are particularly sensitive to the
react to inhalation of cobalt by a known human carcinogen.7
developing asthma or to ingestion or chronic effects of lead, with slowed
The chemical and environmental cognitive development, reduced growth
dermal contact with cobalt by stability of dioxins and their tendency
developing dermatitis. Cobalt is as a and other effects reported. Reproductive
to accumulate in fat have resulted in effects, such as decreased sperm count
vital component of vitamin B12, though their detection within many ecosystems.
there is no evidence that intake of cobalt in men and spontaneous abortions in
In the United States and elsewhere, women, have been associated with lead
is ever limiting in the human diet. accidental contamination of the exposure. The developing fetus is at
A number of epidemiological studies environment by 2,3,7,8–TCDD has particular risk from maternal lead
have found that exposures to cobalt are resulted in deaths in many species of exposure, with low birth weight and
associated with an increased incidence wildlife and domestic animals.8 High slowed postnatal neurobehavioral
of lung cancer in occupational settings. residues of this compound in fish have development noted. Human studies are
The International Agency for Research resulted in closing rivers to fishing. inconclusive regarding lead exposure
on Cancer (part of the World Health Laboratory studies with birds, and cancer, while animal studies have
Organization) classifies cobalt and mammals, aquatic organisms, and other reported an increase in kidney cancer
cobalt compounds as ‘‘possibly species have demonstrated that from lead exposure by the oral route.
carcinogenic to humans’’ (Group 2B). exposure to 2,3,7,8–TCDD can result in EPA has classified lead as a Group B2,
The American Conference of acute and delayed mortality as well as probable human carcinogen.
Governmental Industrial Hygienists has carcinogenic, teratogenic, mutagenic,
classified cobalt as a confirmed animal histopathologic, immunotoxic, and Manganese
carcinogen with unknown relevance to Health effects in humans have been
humans (category A3). An EPA been externally peer reviewed but has not yet been
associated with both deficiencies and
assessment concludes that under EPA’s incorporated in IRIS.
6 IARC (International Agency for Research on excess intakes of manganese. Chronic
cancer guidelines, cobalt would be exposure to low levels of manganese in
Cancer). (1997) IARC monographs on the evaluation
considered likely to be carcinogenic to of carcinogenic risks to humans. Vol. 69. the diet is considered to be nutritionally
humans.5 Polychlorinated dibenzo-para-dioxins and essential in humans, with a
polychlorinated dibenzofurans. Lyon, France.
5 See ‘‘Derivation of a Provisional Carcinogenicity 7 The U.S. Department of Health and Human recommended daily allowance of 2 to 5
Assessment for Cobalt and Compounds,’’ Risk Services, National Toxicology Program 9th Report milligrams per day (mg/d). Chronic
Assessment Issue Paper (00–122/1–15–02), on Carcinogens, Revised January 2001.
Superfund Technical Support Center, National 8 This does not necessarily apply in regard to 9 Eisler, R. 1986. Dioxin hazards to fish, wildlife,
Center for Environmental Assessment, January 15, laboratory testing, which tend to use 2,3,7,8 TCDD and invertebrates: a synoptic review. U.S. Fish and
2002. This is a provisional EPA assessment that has as the test compound. Wildlife Service Biological Report. 85(1.8).
exposure to high levels of manganese by studies have reported an increased risk include older adults and people with
inhalation in humans results primarily of lung and nasal cancers from exposure heart and lung disease.
in central nervous system effects. Visual to nickel refinery dusts and nickel This is only a partial summary of
reaction time, hand steadiness, and eye- subsulfide. Animal studies of soluble adverse health and environmental
hand coordination were affected in nickel compounds i.e., nickel carbonyl) effects associated with exposure to PM.
chronically-exposed workers. Impotence have reported lung tumors. The EPA has Further information is found in the 2004
and loss of libido have been noted in classified nickel refinery subsulfide as a Criteria Document for PM (‘‘Air Quality
male workers afflicted with manganism Group A, human carcinogen and nickel Criteria for Particulate Matter,’’ EPA/
attributed to inhalation exposures. EPA carbonyl as a Group B2, probable 600/P–99/002bF) and the 2005 Staff
has classified manganese in Group D, human carcinogen. Paper for PM (EPA, ‘‘Review of the
not classifiable as to carcinogenicity in National Ambient Air Quality Standards
humans. Organic HAP
for Particulate Matter, Policy
Mercury Organic HAPs include halogenated Assessment of Scientific and Technical
and nonhalogenated organic classes of Information: OAQPS Staff Paper,’’ (June
Mercury exists in three forms: compounds such as polycyclic aromatic 2005)).
elemental mercury, inorganic mercury hydrocarbons (PAHs) and
compounds (primarily mercuric Selenium
polychlorinated biphenyls (PCBs). Both
chloride), and organic mercury PAHs and PCBs are classified as Selenium is a naturally occurring
compounds (primarily methyl mercury). potential human carcinogens, and are substance that is toxic at high
Each form exhibits different health considered toxic, persistent and concentrations but is also a nutritionally
effects. Various sources may release bioaccumulative. Organic HAP also essential element. Studies of humans
elemental or inorganic mercury; include compounds such as benzene, chronically exposed to high levels of
environmental methyl mercury is methane, propane, chlorinated alkanes selenium in food and water have
typically formed by biological processes and alkenes, phenols and chlorinated reported discoloration of the skin,
after mercury has precipitated from the aromatics. Adverse health effects of pathological deformation and loss of
air. nails, loss of hair, excessive tooth decay
HAPs include damage to the immune
Chronic exposure to elemental and discoloration, lack of mental
system, as well as neurological,
mercury in humans also affects the alertness, and listlessness. The
central nervous system, with effects reproductive, developmental,
respiratory and other health problems. consumption of high levels of selenium
such as increased excitability, by pigs, sheep, and cattle has been
irritability, excessive shyness, and Particulate Matter shown to interfere with normal fetal
tremors. The EPA has not classified development and to produce birth
elemental mercury with respect to Atmospheric particulate matter (PM)
is composed of sulfate, nitrate, defects. Results of human and animal
cancer. studies suggest that supplementation
The major effect from chronic ammonium, and other ions, elemental
carbon, particle-bound water, a wide with some forms of selenium may result
exposure to inorganic mercury is kidney
variety of organic compounds, and a in a reduced incidence of several tumor
damage. Reproductive and
large number of elements contained in types. One selenium compound,
developmental animal studies have
various compounds, some of which selenium sulfide, is carcinogenic in
reported effects such as alterations in
originate from crustal materials and animals exposed orally. We have
testicular tissue, increased embryo
others from combustion sources. classified elemental selenium as a
resorption rates, and abnormalities of
Combustion sources are the primary Group D, not classifiable as to human
development. Mercuric chloride (an
origin of trace metals found in fine carcinogenicity, and selenium sulfide as
inorganic mercury compound) exposure
particles in the atmosphere. Ambient a Group B2, probable human
has been shown to result in
PM can be of primary or secondary carcinogen.
forestomach, thyroid, and renal tumors
in experimental animals. EPA has origin. Part Two: Summary of the Final Rule
classified mercuric chloride as a Group Exposure to particles can lead to a
variety of serious health effects. The I. What Source Categories and
C, possible human carcinogen. Subcategories Are Affected by the Final
largest particles do not get very far into
Nickel the lungs, so they tend to cause fewer Rule?
Nickel is an essential element in some harmful health effects. Fine particles Today’s rule promulgates standards
animal species, and it has been pose the greatest problems because they for controlling emissions of HAP from
suggested it may be essential for human can get deep into the lungs. Scientific hazardous waste combustors:
nutrition. Nickel dermatitis, consisting studies show links between these small incinerators, cement kilns, lightweight
of itching of the fingers, hand and particles and numerous adverse health aggregate kilns, boilers, and
forearms, is the most common effect in effects. Epidemiological studies have hydrochloric acid production furnaces
humans from chronic exposure to shown a significant correlation between that burn hazardous waste. A
nickel. Respiratory effects have also elevated PM levels and premature description of each source category can
been reported in humans from mortality. Other important effects be found in the proposed rule (see 69 FR
inhalation exposure to nickel. No associated with PM exposure include at 21207–08).
information is available regarding the aggravation of respiratory and Hazardous waste burning
reproductive of developmental effects of cardiovascular disease (as indicated by incinerators, cement kilns, and
nickel in humans, but animal studies increased hospital admissions, lightweight aggregate kilns are currently
have reported such effects, although a emergency room visits, absences from subject to 40 CFR part 63, subpart EEE,
consistent dose-response relationship school or work, and restricted activity National Emission Standards for
has not been seen. Nickel forms released days), lung disease, decreased lung Hazardous Air Pollutants (NESHAP).
from industrial boilers include soluble function, asthma attacks, and certain Today’s rule revises the emissions limits
nickel compounds, nickel subsulfide, cardiovascular problems. Individuals and certain compliance and monitoring
and nickel carbonyl. Human and animal particularly sensitive to PM exposure provisions of subpart EEE for these
source categories. The definitions of separate stack, such as an alkali bypass that burns hazardous waste. The final
hazardous waste incinerator, hazardous stack. In that case, the emission rule does not apply to a source that
waste cement kiln, and hazardous waste standards would apply separately to meets the applicability requirements of
lightweight aggregate kiln appear at 40 each of these stacks.12 § 63.1200(b) for reasons explained at 69
CFR 63.1201(a). FR at 21212–13.
Boilers that burn hazardous waste are III. What Pollutants Are Emitted and
also affected sources under today’s rule. Controlled? V. What Are the Emission Limitations?
The rule uses the RCRA definition of a Hazardous waste combustors emit You must meet the emission limits in
boiler under 40 CFR 260.10 and dioxin/furans, sometimes at high levels Tables 1 and 2 of this preamble for your
includes industrial, commercial, and depending on the design and operation applicable source category and
institutional boilers as well as thermal of the emission control equipment, and, subcategory. Standards are corrected to
units known as process heaters. for incinerators, depending on whether 7 percent oxygen. As noted at proposal,
Hazardous waste burning boilers will a waste heat recovery boiler is used. All we previously promulgated
continue to comply with the emission hazardous waste combustors can also requirements for carbon monoxide, total
standards found under 40 CFR part 266, emit high levels of other organic HAP if hydrocarbon, and destruction and
subpart H (i.e., the existing RCRA rules) they are not designed, operated, and removal efficiency standards under
until they demonstrate compliance with maintained to operate under good subpart EEE for incinerators, cement
the requirements of 40 CFR part 63, combustion conditions. kilns, and lightweight aggregate kilns.
subpart EEE, and, for permitted sources, Hazardous waste combustors can also We view these standards as unaffected
subsequently remove these emit high levels of metal HAP, by the Court’s vacature of the
requirements from their RCRA permit. depending on the level of metals in the challenged regulations in its decision of
Finally, hydrochloric acid production waste feed and the design and operation July 24, 2001. We are therefore not re-
furnaces that burn hazardous waste are of air emissions control equipment. promulgating and reopening
affected sources under today’s rule. Hazardous waste burning hydrochloric consideration of these standards in
These furnaces are a type of halogen acid production furnaces, however, today’s final rule, but are summarizing
acid furnace included in the definition generally feed and emit low levels of these standards in Tables 1 and 2 for
of ‘‘industrial furnace’’ defined at metal HAP. reader’s convenience.13 See 69 FR at
§ 260.10. Hydrochloric acid production All of these HAP metals (except for 21221, 21248, 21261 and 21274.
furnaces that burn hazardous waste will the volatile metal mercury) are emitted Liquid fuel boilers equipped with dry
continue to comply with the emission as a portion of the particulate matter air pollution control devices are subject
standards found under 40 CFR part 266, emitted by these sources. Hazardous to different dioxin/furan emission
subpart H, until they demonstrate waste combustors can also emit high standards than liquid fuel boilers that
compliance with 40 CFR part 63, levels of particulate matter, except that are not equipped with dry air pollution
subpart EEE, and, for permitted sources, hydrochloric acid production furnaces control devices.14 Liquid fuel boilers
subsequently remove these generally feed hazardous wastes with processing hazardous waste with a
requirements from their RCRA permit. low ash content and consequently emit heating value less than 10,000 BTU/lb
low levels of particulate matter. A must comply with the emission
II. What Are the Affected Sources and majority of particulate matter emissions concentration-based standards
Emission Points? from hazardous waste combustors are in (expressed as mass of total HAP
Today’s rule apply to each major and the form of fine particulate. Particulate emissions per volume of stack gas
area source incinerator, cement kiln, emissions from incinerators and liquid emitted) for mercury, semivolatile
lightweight aggregate kiln, boiler, and fuel-fired boilers depend on the ash metals, low volatile metals, and total
hydrochloric acid production furnace content of the hazardous waste feed and chlorine. Liquid fuel boilers processing
that burns hazardous waste.10 We note the design and operation of air emission hazardous waste with heating values
that only major source boilers and control equipment. Particulate greater than 10,000 BTU/lb must
hydrochloric acid production furnaces emissions from cement kilns and comply with thermal emissions-based
are subject to the full suite of subpart lightweight aggregate kilns are not standards (expressed as mass of HAP
EEE emission standards.11 The significantly affected by the ash content emissions attributable to the hazardous
emissions limits apply to each emission of the hazardous waste fuel because waste per million BTU input from the
point (e.g., stack) where gases from the uncontrolled particulate emissions are hazardous waste) for those same
combustion of hazardous waste are attributable primarily to fine raw pollutants. Low volatile metal standards
discharged or otherwise emitted into the material entrained in the combustion for liquid fuel boilers apply only to
atmosphere. For facilities that have gas. Thus, particulate emissions from emissions of chromium, whereas the
multiple combustion gas discharge kilns depends on operating conditions low volatile metal standard for the other
points, the emission limits generally that effect entrainment of raw material, source categories applies to the
apply to each emission point. A cement and the design and operation of the combined emissions of chromium,
kiln, for example, could be configured emission control equipment. arsenic, and beryllium. Semivolatile
to have dual stacks where the majority IV. Does the Final Rule Apply to Me? metal standards apply to the combined
of combustion gases are discharged emissions of lead and cadmium.
The final rule applies to you if you
though the main stack and other For any of the source categories
own or operate a hazardous waste
combustion gases emitted through a except hydrochloric acid production
combustor—an incinerator, cement kiln,
10 A major source emits or has the potential to
lightweight aggregate kiln, boiler, or 13 We are also republishing these standards, for
emit 10 tons per year of any single hazardous air hydrochloric acid production facility reader’s convenience only, in the new replacement
pollutant or 25 tons per year or greater of hazardous standard section for these source categories. See
air pollutants in the aggregate. An area source is a 12 We note that there is a provision that allows § 63.1219, § 63.1220 and § 673.1219.
source that is not a major source. cement kilns with dual stacks to average emissions 14 Liquid fuel boilers equipped with a wet air
11 See Part Four, Section II.A for a discussion of on a flow-weighted basis to demonstrate pollution control device followed by a dry air
the standards that are applicable to area source compliance with the metal and chlorine emission pollution control device do not meet the definition
boilers and hydrochloric acid production furnaces. standards. See §§ 63.1204(e) and 63.1220(3). of a dry air pollution device.
furnaces, you may elect to comply with Incinerators and liquid and solid fuel Tables 1 and 2. However, for
an alternative to the total chlorine boilers may elect to comply with an semivolatile metals, the alternative
standard under which you would alternative to the particulate matter standard applies to the combined
establish site-specific, health-based standard that would limit emissions of emissions of lead, cadmium, and
emission limits for hydrogen chloride all the semivolatile metal HAPs and low selenium; for low volatile metals, the
and chlorine based on national exposure volatile metal HAPs. Under this standard applies to the combined
standards. This alternative chlorine alternative, the numerical emission emissions of chromium, arsenic,
standard is discussed in part two, limits for semivolatile metal and low beryllium, antimony, cobalt, manganese,
section IX and part four, section VII. volatile metal emission HAP are and nickel. See § 63.1219(e).
identical to the limitations included in
TABLE 1.—SUMMARY OF EMISSION LIMITS FOR EXISTING SOURCES
Hydrochloric acid
Lightweight aggre- Solid fuel-fired Liquid fuel-fired boil-
Incinerators Cement kilns production fur-
gate kilns boilers 1 ers 1 naces 1
Dioxin/Furans (ng 0.20 or 0.40 and 0.20 or 0.40 and 0.20 or rapid CO or HC and 0.40 for dry APCD CO or HC and
TEQ/dscm). temperature temperature quench below DRE stand- sources; CO or HC DRE standard
control < 400°F control < 400°F 400°F at kiln ard as a and DRE standard as surrogate.
at APCD inlet 6. at APCD inlet. exit. surrogate. as surrogate for
others.
Mercury .................. 130 µg/dscm ....... Hazardous waste 120 hazardous 11 µg/dscm ... 4.2E-5lb/MMBtu 2, 5 Total chlorine
feed restriction waste MTEC 11 or 19 µg/dscm 2; standard as
of 3.0 ppmw feed restriction depending on BTU surrogate.
and 120 µg/ or 120 µg/dscm content of haz-
dscm MTEC 11; total emissions. ardous waste 13.
or 120 µg/dscm
total emissions.
Particulate Matter ... 0.013 gr/dscf 8 ..... 0.028 gr/dscf and 0.025 gr/dscf ....... 0.030 gr/dscf 8 0.035 gr/dscf 8 ........... Total chlorine
20% opacity 12. standard as
surrogate.
Semivolatile Metals 230 µg/dscm ....... 7.6 E-4 lbs/ 3.0E-4 lb/MMBtu 5 180 µg/dscm 8.2 E–5 lb/MMBtu 2, 5 Total chlorine
(lead + cadmium). MMBtu 5 and and 250 µg/ or 150 µg/dscm 2; standard as
330 µg/dscm 3. dscm 3. depending on BTU surrogate.
content of haz-
ardous waste 13.
Low Volatile Metals 92 µg/dscm ......... 2.1 E-5 lbs/ 9.5E-5 lb/MMBtu 5 380 µg/dscm 1.26E–4 lbMMBtu 4, 5 Total chlorine
(arsenic + beryl- MMBtu 5 and 56 and 110 µg/ or 370 µg/dscm 4; standard as
lium + chromium). µg/dscm 3. dscm 3. depending on BTU surrogate.
content of haz-
ardous waste 13.
Total Chlorine (hy- 32 ppmv 7 ............ 120 ppmv 7 .......... 600 ppmv 7 .......... 440 ppmv 7 .... 5.08E–2 lb/MMBtu 5, 7 150 ppmv or
drogen chloride + or 31 ppmv 7; de- 99.923% sys-
chlorine gas). pending on BTU tem removal ef-
content of haz- ficiency.
ardous waste 13.
Carbon Monoxide 100 ppmv CO 9 or See Note # 10 100 ppmv CO 9 or (2) 100 ppmv CO 9 or 10 ppmv HC
(CO) or Hydro- 10 ppmv HC. below. 20 ppmv HC.
carbons (HC).
Destruction and Re- 99.99% for each principal organic hazardous pollutant. For sources burning hazardous wastes F020, F021, F022, F023,
moval Efficiency. F026, or F027, however, 99.9999% for each principal organic hazardous pollutant.
Notes:
1 Particulate matter, semivolatile metal, low volatile metal, and total chlorine standards for solid and liquid fuel boilers apply only to major
sources. Particulate matter, semivolatile and low volatile metal standards for hydrochloric acid production furnaces apply only to major sources,
although area sources must still comply with the surrogate total chlorine standard to control mercury emissions.
2 Standard is based on normal emissions data, and is therefore expressed as an annual average emission limitation.
3 Sources must comply with both the thermal emissions and emission concentration standards.
4 Low volatile metal standard for liquid fuel-fired boilers is for chromium only.
5 Standards expressed as mass of pollutant contributed by hazardous waste per million BTU contributed by the hazardous waste.
6 APCD means ‘‘air pollution control device’’.
7 Sources may elect to comply with site-specific risk-based emission limits for hydrogen chloride and chlorine gas
8 Sources may elect to comply with an alternative to the particulate matter standard.
9 Sources that elect to comply with the CO standard must demonstrate compliance with the HC standard during the comprehensive perform-
ance test that demonstrates compliance with the destruction and removal efficiency requirement.
10 Kilns without a bypass: 20 ppmv HC or 100 ppmv CO 9. Kilns with a bypass/mid-kiln sampling system: 10 ppmv HC or 100 ppmv CO9 in the
bypass duct, mid-kiln sampling system or bypass stack.
11 MTEC means ‘‘maximum theoretical emission concentration’’, and is equivalent to the feed rate divided by gas flow rate
12 The opacity standard does not apply to a source equipped with a bag leak detection system under 63.1206(c)(8) or a particulate matter de-
tection system under 63.1206(c)(9).
13 Emission concentration-based standards apply to sources processing hazardous waste with energy content less than 10,000 BTU/lb; thermal
emission standards apply to sources processing hazardous waste with energy content greater than 10,000 btu/lb.
TABLE 2.—SUMMARY OF EMISSION LIMITS FOR NEW OR RECONSTRUCTED SOURCES

Hydrochloric acid
Lightweight aggre- Solid fuel boil- Liquid fuel boilers 1
Incinerators Cement kilns production fur-
gate kilns ers 1 naces 1
Dioxin/Furans (ng 0.11 for dry APCD 0.20 or 0.40 and 0.20 or rapid CO or HC and 0.40 for sources with CO or THC and
TEQ/dscm). and/or WHB 5 temperature quench DRE stand- dry APCD; CO or DRE standard
sources; 0.20 control <400 °F <400 °F at kiln ard as a HC and DRE as a surrogate.
for other at APCD inlet. exit. surrogate. standard as a sur-
sources. rogate for other
sources.
Mercury .................. 8.1 µg/dscm ........ Hazardous waste 120 hazardous 11 µg/dscm ... 1.2E–6 lb/MMBtu 2 4 TCl as surrogate.
feed restriction waste MTEC 10 or 6.8 µg/dscm 2;
of 1.9 ppmw feed restriction depending on BTU
and 120 µg/ or 120 µg/dscm content of haz-
dscm MTEC 10; total emissions. ardous waste 12.
or 120 µg/dscm
total emissions.
Particulate matter 0.0015 7 ............... 0.0023 and 20% 0.0098 ................. 0.015 7 ........... 0.0087 7 ..................... TCl as surrogate.
(gr/dscf). opacity 11.
Semivolatile Metals 10 µg/dscm ......... 6.2E–5 lb/ 3.7 E–5 lb/ 180 µg/dscm 6.2 E–6 lb/MMBtu 2 4 TCl as surrogate.
(lead + cadmium). MMBtu 4 and MMBtu 4 and 43 or 78 µg/dscm 2;
180 µg/dscm. µg/dscm. depending on BTU
content of haz-
ardous waste 12.
Low Volatile Metals 23 µg/dscm ......... 1.5E–5 lb/ 3..3E–5 lb/ 190 µg/dscm 1.41E–5lb/MMBtu 3 4 TCl as surrogate.
(arsenic + beryl- MMBtu 4 and 54 MMBtu 4 and or 12 µg/dscm 3;
lium + chromium). µg/dscm. 110 µg/dscm. depending on BTU
content of haz-
ardous waste 12.
Total Chlorine (Hy- 21 ppmv 6 ............ 86 ppmv 6 ............ 600 ppmv 6 .......... 73 ppmv 6 ...... 5.08E–2 lb/MMBtu 4 6 25 ppmv or
drogen chloride + or 31 ppmv 6; de- 99.987% SRE.
chlorine gas). pending on BTU
content of haz-
ardous waste 12.
Carbon monoxide 100 ppmv CO 8 or See note #9 100 ppmv CO 8 or 100 ppmv CO 8 or 10 ppmv HC
(CO) or Hydro- 10 ppmv HC. below. 20 ppmv HC.
carbons (HC).
Destruction and Re- 99.99% for each principal organic hazardous pollutant. For sources burning hazardous wastes F020, F021, F022, F023,
moval Efficiency. F026, or F027, however, 99.9999% for each principal organic hazardous pollutant.
Notes:
1 Particulate matter, semivolatile metal, low volatile metal, and total chlorine standards for solid and liquid fuel boilers apply only to major
sources. Particulate matter, semivolatile and low volatile metal standards for hydrochloric acid production furnaces apply only to major sources,
although area sources must still comply with the surrogate total chlorine standard to control mercury emissions.
2 Standard is based on normal emissions data, and is therefore expressed as an annual average emission limitation.
3 Low volatile metal standard for liquid fuel-fired boilers is for chromium only. Arsenic and beryllium are not included in the low volatile metal
total for liquid fuel-fired boilers.
4 Standards expressed as mass of pollutant contributed by hazardous waste per million BTU contributed by the hazardous waste.
5 APCD means ‘‘air pollution control device’’, WHB means ‘‘waste heat boiler’’.
6 Sources may elect to comply with risk-based emission limits for hydrogen chloride and chlorine gas.
7 Sources may elect to comply with an alternative to the particulate matter standard.
8 Sources that elect to comply with the CO standard must demonstrate compliance with the THC standard during the comprehensive perform-
ance test that demonstrates compliance with the destruction and removal efficiency requirement.
9 Greenfield kilns without a bypass: 20 ppmv HC or 100 ppmv CO 8 and 50 ppmv HC. Greenfield kilns with a bypass/mid kiln sampling system:
Main stack standard of 50 ppmv HC and 10 ppmv HC or 100 ppmv CO 8 in the bypass duct, mid-kiln sampling system or bypass stack. Green-
field kilns with a bypass/mid-kiln sampling system: 10 ppmv HC or 100 ppmv CO 8 in the bypass duct, mid-kiln sampling system or bypass stack;
Non-greenfield kilns without a bypass: 20 ppmv HC or 100 ppmv CO 8. A greenfield kiln is a kiln whose construction commenced after April 19,
1996 at a plant site where a cement kiln (whether burning hazardous waste or not) did not previously exist.
10 MTEC means ‘‘maximum theoretical emission concentration’’, and is equivalent to the feed rate divided by gas flow rate.
11 The opacity standard does not apply to a source equipped with a bag leak detection system under 63.1206(c)(8) or a particulate matter de-
tection system under 63.1206(c)(9).
12 Emission concentration-based standards apply to sources processing hazardous waste with energy content less than 10,000 BTU/lb; thermal
emission standards apply to sources processing hazardous waste with energy content greater than 10,000 btu/lb.
VI. What Are the Testing and Initial note, however, that today’s final rule kilns, and lightweight aggregate kilns as
Compliance Requirements? revises some of these requirements as well. Even though we are not
The testing and initial compliance they apply to all or specific HWCs (e.g., repromulgating the compliance and
requirements we promulgate today for one-time dioxin/furan test for sources testing requirements for those source
solid fuel boilers, liquid fuel boilers, not subject to a numerical dioxin/furan categories, those sources must
and hydrochloric acid production standard; dioxin/furan stack test demonstrate compliance with the
furnaces are identical to those that are method; hydrogen chloride and chlorine replacement emission standards
applicable to incinerators, cement kilns, stack test methods) promulgated today.
and lightweight aggregate kilns at We also discuss compliance and
§§ 63.1206, 63.1207, and 63.1208. We testing dates for incinerators, cement
A. Compliance Dates midway between comprehensive 3. You must install an automatic

The time-line for testing and initial performance tests). If your source is not waste feed cutoff system that links the
compliance requirements is as follows: subject to a numerical dioxin/furan operating parameter limits to the waste
1. The compliance date is October 14, emission standard (e.g., solid fuel feed cutoff system;
2008; 15 boilers, lightweight aggregate kilns that 4. You must control combustion
2. You must submit a comprehensive comply with the 400 °F temperature system leaks;
performance test plan to the permitting limit at the kiln exit, liquid fuel boilers 5. You must establish and comply
authority for review and approval 12 equipped with wet or no air pollution with an operator training and
months prior to commencing the test. control system, and hydrochloric acid certification program;
3.You must submit an eligibility production furnaces), you must conduct 6. You must establish and comply
demonstration for the health-based a one-time dioxin/furan test to enable with an operation and maintenance
compliance alternative to the total the Agency to evaluate the effectiveness plan;
chlorine emission standard 12 months of the carbon monoxide/hydrocarbon 7. If your source is equipped with a
before the compliance date if you elect standard and the destruction and baghouse, you must install either a bag
to comply with § 63.1215; removal efficiency standard in leak detection system or a particulate
4. You must place in the operating controlling dioxin/furan emissions for matter detection system; 18 and
record a Documentation of Compliance those sources. Previous dioxin/furan 8. If your source is equipped with an
by the compliance date identifying the emission tests may be used to meet this electrostatic precipitator or ionizing wet
operating parameter limits that, using requirement if the combustor operated scrubber, you must either establish site-
available information, you have under the conditions required by the specific control device operating
determined will ensure compliance rule and if design and operation of the parameter limits which limits are linked
with the emission standards; combustor has not changed since the to the automatic waste feed cutoff
5. For boilers and hydrochloric acid test in a manner that could increase system, or install a particulate matter
production furnaces, you must dioxin/furan emissions. The Agency detection system and take corrective
commence the initial comprehensive will use those emissions data when measures when the alarm level is
performance test within 6 months after reevaluating the MACT standards under exceeded.
the compliance date; CAA section 112(d)(6), when VII. What Are the Continuous
6. For incinerators, cement kilns, and determining whether to develop Compliance Requirements?
lightweight aggregate kilns, you must residual risk standards for these sources
commence the initial comprehensive The continuous compliance
pursuant to section 112(f)(2), and when requirements for solid fuel boilers,
performance test within 12 months after determining whether the source’s RCRA
the compliance date; liquid fuel boilers, and hydrochloric
Permit is protective of human health acid production furnaces are identical to
7. You must complete the initial and the environment.
comprehensive performance test within those applicable to incinerators, cement
You must use the following stack test kilns, and lightweight aggregate kilns.
60 days of commencing the test; and methods to document compliance with
8. You must submit a Notification of See § 63.1209. We note, however, that
the emission standards: (1) Method 29 today’s final rule revises some of these
Compliance within 90 days of for mercury, semivolatile metals, and
completing the test documenting requirements as they apply to all or
low volatile metals; and (2) Method 26/ specific HWCs (e.g., bag leak detection
compliance with emission standards 26A, Methods 320 or 321, or ASTM D
and continuous monitoring system system requirements; optional
6735–01 for hydrogen chloride and particulate matter detection system
requirements. chlorine; 16 (3) either Method 0023A or requirements; compliance assurance for
B. Testing Requirements Method 23 for dioxin/furans; and (4) thermal emissions-based standards).
All hazardous waste combustors must either Method 5 or 5i for particulate You must use carbon monoxide or
commence the initial comprehensive matter. hydrocarbon continuous emissions
performance test under the time lines C. Initial Compliance Requirements monitors (as well as an oxygen
discussed above. The purpose of the continuous emissions monitor to correct
comprehensive performance test is to The initial compliance requirements the carbon monoxide or hydrocarbon
document compliance with the for solid fuel boilers, liquid fuel boilers, values to 7% oxygen) to ensure
emission standards of the final rule and and hydrochloric acid production compliance with the carbon monoxide
establish operating parameter limits to furnaces include: 17 or hydrocarbon emission standards.
maintain compliance with those 1. You must place in the operating You must also establish limits (as
standards. You must also conduct record a Documentation of Compliance applicable) on the feedrate of metals,
periodic comprehensive performance by the compliance date identifying the chlorine, and ash, key combustor
testing every five years. operating parameter limits that, using operating parameters, and key operating
If your source is subject to a available information, you have
numerical dioxin/furan emission determined will ensure compliance 18 A major difference between a bag leak detection
with the emission standards; system and a particulate matter detection system is
standard (i.e., incinerators, cement the way the alarm level is established. The alarm
kilns, lightweight aggregate kilns that 2. You must develop and comply with level for a bag leak detection system is established
comply with the 0.2 ng TEQ/dscm a startup, shutdown, and malfunction using concepts in the Agency’s bag leak detection
standard, and liquid fuel boilers plan; system guidance document while the alarm level
for a particulate matter detection system is
equipped with a dry air pollution established based on the detector response during
control device), you must conduct a 16 Note that you may be required to use other test
the comprehensive performance test. The ash
dioxin/furan confirmatory performance methods to document emissions of hydrogen feedrate limit for incinerators and boilers is waived
chloride and chlorine if you elect to comply with if you use a particulate matter detection system but
test no later than 2.5 years after each the alternative, health-based emission limits for not if you use a bag leak detection system because
comprehensive performance test (i.e., total chlorine under § 63.1215. See § 63.1208(b)(5). the bag leak detection system alarm level may not
17 These same requirements currently apply to provide reasonable assurance of continuous
15 See 69 FR at 21313 for rationale. We received incinerators, cement kilns, and lightweight compliance with the particulate matter emission
no adverse comments at proposal. aggregate kilns. standard.
parameters of the air pollution control IX. What Is the Health-Based off-site receptor location which
device based on operations during the Compliance Alternative for Total concentrations are attributable to all on-
comprehensive performance test. You Chlorine, and How Do I Demonstrate site hazardous waste combustors) and
must continuously monitor these Eligibility? thus may exceed the 1-hour average
parameters with a continuous HCl-equivalent emission rate limit when
A. Overview
monitoring system. complying with the annual average HCl-
The rule allows you to establish and equivalent emission rate limit, absent an
VIII. What Are the Notification, comply with health-based compliance hourly rolling average limit on the
Recordkeeping, and Reporting alternatives for total chlorine for feedrate of total chlorine and chloride.
Requirements? hazardous waste combustors other than (6) Submit your eligibility
hydrochloric acid production furnaces demonstration, including your
The notification, recordkeeping, and in lieu of the MACT technology-based
reporting requirements that we determination of whether the 1-hour
emission standards established under average HCl-equivalent emission rate
promulgate today for solid fuel boilers, §§ 63.1216, 63.1217, 63.1219, 63.1220, limit may be exceeded absent an hourly
liquid fuel boilers, and hydrochloric and 63.1221. See § 63.1215. To identify rolling average limit on the feedrate of
acid production furnaces are identical to and comply with the limits, you must: total chlorine and chloride, for review
those that are applicable to incinerators, (1) Identify a total chlorine emission and approval;
cement kilns, and lightweight aggregate rate for each on-site hazardous waste
kilns. See §§ 63.1210 and 63.1211. We combustor. You may select total (7) Document during the
note, however, that today’s final rule chlorine emission rates as you choose to comprehensive performance test the
revises some of these requirements as demonstrate eligibility for the health- total chlorine system removal efficiency
based limits, except the total chlorine for each combustor and use this system
they apply to all or specific HWCs.
emission rate limits for incinerators, removal efficiency to calculate chlorine
You must submit notifications feedrate limits. Also, document that
including the following to the cement kilns, and lightweight aggregate
kilns cannot result in total chlorine total chlorine emissions during the test
permitting authority in addition to those do not exceed the 1-hour average HCl-
emission concentrations exceeding the
required by the NESHAP General equivalent emission rate limit during
Interim Standards provided by
Provisions, subpart A of 40 CFR part 63: §§ 63.1203, 63.1204, and 63.1205;19 any run of the test. In addition, establish
1. Notification of changes in design, (2) Calculate the HCl-equivalent operating limits on the emission control
operation, or maintenance emission rate for the total chlorine device based on operations during the
(§ 63.1206(b)(5)(i)); emission rates you select, considering comprehensive performance test; and
2. Notification of performance test long-term exposure and using Reference (8) Comply with the requirements for
Concentrations (RfCs) as the health changes in the design, operation, or
and continuous monitoring system
threshold metric. This emission rate is maintenance of the facility which could
evaluation, including the performance
called the annual average HCl- affect the HCl-equivalent emission rate
test plan and continuous monitoring limits or system removal efficiency for
system performance evaluation plan equivalent emission rate;
(3) Perform an eligibility total chlorine, and changes in the
(§ 63.1207(e)); vicinity of your facility over which you
demonstration to determine if your
3. Notification of compliance, annual average HCl-equivalent emission do not have control (e.g., new receptors
including results of performance tests rate meets the national exposure locating proximate to the facility).
and continuous monitoring system standard (i.e., Hazard Index not B. HCl-Equivalent Emission Rates
evaluations (§§ 63.1210(b), 63.1207(j); exceeding 1.0 considering the maximum
63.1207(k), and 63.1207(l)); and annual average ambient concentration of You must express total chlorine
4. Various notifications if you request hydrogen chloride and chlorine at an emission rates (lb/hr) from each on-site
or elect to comply with alternative off-site receptor location which hazardous waste combustor, including
requirements at § 63.1210(a)(2). concentrations are attributable to all on- hydrochloric acid production
site hazardous waste combustors) and furnaces 20, as an annual average HCl-
You must submit the following thus is below the annual average HCl- equivalent emission rate and a 1-hour
reports to the permitting authority in equivalent emission rate limit; average HCl-equivalent emission rate.
addition to those required by the (4) Calculate the HCl-equivalent See § 63.1215(b). The annual average
NESHAP General Provisions, subpart A emission rate for the total chlorine HCl-equivalent emission rate equates
of 40 CFR part 63: emission rates you select, considering chlorine emission rates to hydrogen
1. Startup, shutdown, and short-term exposure and using acute chloride (HCl) emission rates using
malfunction plan, if you elect to comply Reference Exposure Levels (aRELs) as Reference Concentrations (RfCs) as the
with § 63.1206(c)(2)(ii)(B)); the health threshold metric. This health risk metric for long-term
emission rate is called the 1-hour exposure. The 1-hour average HCl-
2. Excessive exceedances report average HCl-equivalent emission rate. equivalent emission rate equates
(§ 63.1206(c)(3)(vi)); and (5) Determine whether your 1-hour chlorine emission rates to HCl emission
3. Emergency safety vent opening HCl-equivalent emission rate may rates using 1-hour Reference Exposure
reports (§ 63.1206(c)(4)(iv)). exceed the national exposure standard
Finally, you must keep records (i.e., Hazard Index not exceeding 1.0 20 Although hydrochloric acid production
documenting compliance with the considering the maximum 1-hour furnaces are not eligible for the health-based total
average ambient concentration of chlorine emission limits (because control of total
requirements of Subpart EEE. chlorine is a surrogate for control of metal HAP),
Recordkeeping requirements are hydrogen chloride and chlorine at an you must consider total chlorine emissions from
hydrochloric acid production furnaces when
prescribed in § 63.1211(b), and include 19 Note that the final rule sunsets the Interim demonstrating that total chlorine emissions from all
requirements under the NESHAP Standards on the compliance date of today’s rule on-site hazardous waste combustors will not exceed
General Provisions, subpart A of 40 CFR but codifies the Interim Standards for total chlorine the Hazard Index limit of 1.0 at an off-site receptor
under § 63.1215(b)(7). location.
Levels (aRELs) as the health risk metric for the health-based total chlorine for your hazardous waste combustors
for acute exposure. standard is determined by comparing and the distance between the stack and
To calculate HCl-equivalent emission the annual average HCl-equivalent the property boundary. If any of these
rates, you must apportion total chlorine emission rate for the total chlorine values for stack height, stack diameter,
emissions (ppmv) between chlorine and emission rate you select for each and distance to nearest property
HCl using the volumetric ratio of combustor to the annual average HCl- boundary do not match the exact values
chlorine to hydrogen chloride (Cl2/HCl). equivalent emission rate limit. in the look-up table, you must use the
• To calculate the annual average The annual average HCl-equivalent next lowest table value. If you have
HCl-equivalent emission rate (lb/hr) and emission rate limit is the HCl-equivalent more than one hazardous waste
the emission rate limit, you must use emission rate, determined by equating combustor on site, you must adjust the
the historical average Cl2/HCl the toxicity of chlorine to HCl using emission rate limits provided by the
volumetric ratio from all regulatory RfCs as the health risk metric for long- tables such that the sum of the ratios for
compliance tests and the gas flowrate term exposure, which ensures that all combustors of the adjusted emission
(and other relevant parameters) from the maximum annual average ambient rate limit to the emission rate limit
most recent RCRA compliance test or concentrations of HCl equivalents do provided by the table cannot exceed 1.0.
MACT performance test. not exceed a Hazard Index of 1.0, See § 63.1215 (c)(3)(v).
• To calculate the 1-hour average rounded to the nearest tenths decimal
HCl-equivalent emission rate (lb/hr) and 2. Site-Specific Compliance
place (0.1) and considering all on-site
emission rate limit, you must use the Demonstration
hazardous waste combustors. See
highest Cl2/HCl volumetric ratio from § 63.1215(b)(2). You may use any scientifically-
all regulatory compliance tests and the Your facility is eligible for the health- accepted peer-reviewed risk assessment
gas flowrate from the most recent RCRA based compliance alternatives for total methodology for your site-specific
compliance test or MACT performance chlorine if either: (1) The annual compliance demonstration to calculate
test. average HCl-equivalent emission rate for an annual average HCl-equivalent
• If you believe that the Cl2/HCl each on-site hazardous waste combustor emission rate limit for each on-site
volumetric ratio for one or more is below the HCl-equivalent emission hazardous waste combustor. An
historical compliance tests is not rate limit determined from the example of one approach for performing
representative of the current ratio, you appropriate value for the emission rate the demonstration for air toxics can be
may request that the permitting limit in the applicable look-up table and found in the EPA’s ‘‘Air Toxics Risk
authority allow you to screen those the proration procedure for multiple Assessment Reference Library, Volume
ratios from the analysis of historical combustors discussed below; or (2) the 2, Site-Specific Risk Assessment
ratios. annual average HCl-equivalent emission Technical Resource Document,’’ which
• If the permitting authority believes rate for each on-site hazardous waste may be obtained through the EPA’s Air
that too few historical Cl2/HCl ratios are combustor is below the annual average Toxics Web site at http://www.epa.gov/
available to establish a representative HCl-equivalent emission rate limit you ttn/atw.
average ratio and a representative calculate based on a site-specific To determine the annual average HCl-
maximum ratio, the permitting authority compliance demonstration. equivalent emission rate limit for each
may require you to conduct periodic
1. Look-Up Table Analysis on-site hazardous waste combustor,
testing to establish representative ratios.
• You must include the Cl2/HCl your site-specific compliance
Look-up tables for the eligibility demonstration must, at a minimum: (1)
volumetric ratio demonstrated during demonstration are provided as Tables 1
each performance test in your data base estimate long-term inhalation exposures
and 2 to § 63.1215. Table 1 presents through the estimation of annual or
of historical Cl2/HCl ratios to update the annual average HCl-equivalent emission
ratios for subsequent calculations of the multi-year average ambient
rate limits for sources located in flat concentrations; (2) estimate the
annual average and 1-hour average HCl- terrain. For purposes of this analysis,
equivalent emission rates (and emission inhalation exposure for the actual
flat terrain is terrain that rises to a level individual most exposed to the facility’s
rate limits). not exceeding one half the stack height emissions from hazardous waste
C. Eligibility Demonstration within a distance of 50 stack heights. combustors, considering locations
Table 2 presents annual average HCl- where people reside and where people
You must perform an eligibility equivalent emission rate limits for
demonstration to determine whether the congregate for work, school, or
sources located in simple elevated recreation; (3) use site-specific, quality-
total chlorine emission rates you select terrain. For purposes of this analysis,
for each on-site hazardous waste assured data wherever possible; (4) use
simple elevated terrain is terrain that health-protective default assumptions
combustor meet the national exposure rises to a level exceeding one half the
standard (i.e., the Hazard Index of 1.0 wherever site-specific data are not
stack height, but that does not exceed available, and: (5) contain adequate
cannot be exceeded at an off-site the stack height within a distance of 50
receptor location considering maximum documentation of the data and methods
stack heights. used for the assessment so that it is
annual average ambient concentrations If your facility is not located in either
attributable to all on-site hazardous transparent and can be reproduced by
flat or simple elevated terrain, you must an experienced risk assessor and
waste combustors and the RfCs for HCl conduct a site-specific compliance
and chlorine) using either a look-up emissions measurement expert.
demonstration.
table analysis or a site-specific To establish the annual average HCl-
To determine the annual average HCl-
compliance demonstration.21 Eligibility equivalent emission rate limit for each
equivalent emission rate limit for a
combustor, you may apportion as you
source from the look-up table, you must
21 The total chlorine emission rates (lb/hr) for elect among the combustors the annual
incinerators, cement kilns, and lightweight use the stack height and stack diameter
average HCl-equivalent emission rate
aggregate kilns cannot result in total chlorine
emission concentrations (ppmv) exceeding the Standards on the compliance date of today’s rule
limit for the facility, which limit
Interim Standards provided by §§ 63.1203, 63.1204, but codifies the Interim Standards for total chlorine ensures that the RfC-based Hazard Index
and 63.1205. The final rule sunsets the Interim under § 63.1215(b)(7). of 1.0 is not exceeded.
D. Assurance That the 1-Hour HCl- for 1-hour average HCl-equivalent potential for the source to vary chlorine
Equivalent Emission Rate Will Not Be emission rate limits as Table 3 and feedrates substantially over the
Exceeded Table 4 to this section. Table 3 provides averaging period for the long-term
The long-term, RfC-based Hazard limits for facilities located in flat terrain. feedrate limit (i.e., 12-hours, or up to
Index will always be higher than the Table 4 provides limits for facilities annually) established to maintain
short-term, aREL-based Hazard Index for located in simple elevated terrain. You compliance with the annual average
a constant HCl-equivalent emission rate must use the Tables to establish HCl-equivalent emission rate limit.
because the health threshold levels for emission rate limits in the same manner
E. Review and Approval of Eligibility
short-term exposure are orders of as you use Tables 1 and 2 to establish
Demonstrations
magnitude higher than the health annual average HCl-equivalent emission
rate limits. The permitting authority will review
threshold levels for long-term
If you conduct a site-specific analysis and approve your eligibility
exposure.22 Even though maximum 1-
to establish a 1-hour average HCl- demonstration. Your eligibility
hour average ambient concentrations are
equivalent emission rate limit, you must demonstration must contain, at a
substantially higher than maximum
follow the risk assessment procedures minimum, the information listed in
annual average concentrations, the
you used to establish an annual average § 63.1215(d)(1).
higher short-term ambient
concentrations do not offset the much HCl-equivalent emission rate limit. The 1. Review and Approval for Existing
higher health threshold levels for short- 1-hour HCl-equivalent emission rate Sources
term exposures. Thus, the long-term, limit, however, is the emission rate than
ensures that the Hazard Index If you operate an existing source, you
RfC-based Hazard Index will always must submit the eligibility
associated with maximum 1-hour
govern regarding whether a source can demonstration to your permitting
average exposures is not greater than
make an eligibility demonstration. authority for review and approval not
1.0.
Accordingly, eligibility for the health- You must consider criteria including later than 12 months prior to the
based emission limits is based solely on the following to determine if a source compliance date. You must also submit
whether a source can comply with the may exceed the 1-hour HCl-equivalent a separate copy of the eligibility
annual average HCl-equivalent emission emission rate limit absent an hourly demonstration to: U.S. EPA, Risk and
rate limit. rolling average chlorine feedrate limit: Exposure Assessment Group, Emission
Nonetheless, some sources may have Standards Division (C404–01), Attn:
(1) The ratio of the 1-hour average HCl-
highly variably chlorine feedrates (and Group Leader, Research Triangle Park,
equivalent emission rate based on the
corresponding highly variable HCl- total chlorine emission rate you select North Carolina 27711, electronic mail
equivalent emission rates) such that for each hazardous waste combustor to address REAG@epa.gov.
they may feed chlorine at very high the 1-hour average HCl-equivalent Your permitting authority should
levels for short periods of time and still emission rate limit for the combustor; notify you of approval or intent to
remain in compliance with the chlorine and (2) the potential for the source to disapprove your eligibility
feedrate limit established to ensure vary total chlorine and chloride demonstration within 6 months after
compliance with the annual average feedrates substantially over the receipt of the original demonstration,
HCl-equivalent emission rate limit.23 To averaging period for the feedrate limit and within 3 months after receipt of any
ensure that the 1-hour HCl-equivalent you establish to ensure compliance with supplemental information that you
emission rate limit will not be exceeded the annual average HCl-equivalent submit. A notice of intent to disapprove
during these periods of peak emissions, emission rate limit. your eligibility demonstration will
you must establish a 1-hour average If you determine that a source may identify incomplete or inaccurate
HCl-equivalent emission rate and 1-hour exceed the 1-hour average HCl- information or noncompliance with
average HCl-equivalent emission rate equivalent emission rate limit, you must prescribed procedures and specify how
limit for each combustor and consider establish an hourly rolling average much time you will have to submit
site-specific factors including prescribed chlorine feedrate limit as discussed additional information or to comply
criteria to determine if the 1-hour below in Section G. with the MACT total chlorine standards.
average HCl-equivalent emission rate You must include the following If your eligibility demonstration is
limit may be exceeded absent an hourly information in your eligibility disapproved, the permitting authority
rolling average limit on chlorine demonstration to document your may extend the compliance date of the
feedrate. If the 1-hour average HCl- determination whether an hourly rolling total chlorine standard to allow you to
equivalent emission rate limit may be average feedrate limit is needed to make changes to the design or operation
exceeded, you must establish an hourly maintain compliance with the 1-hour of the combustor or related systems as
rolling average feedrate limit on HCl-equivalent emission rate limit: (1) quickly as practicable to enable you to
chlorine. Determination of the Cl2/HCl achieve compliance with the MACT
You must calculate the 1-hour average volumetric ratio established for 1-hour standard for total chlorine.
HCl-equivalent emission rate from the average HCl-equivalent emission rate If your permitting authority has not
total chlorine emission rate you select determinations as provided by approved your eligibility demonstration
for each source. § 63.1215(b)(6)(ii); (2) determination of by the compliance date, and has not
You must establish the 1-hour average the 1-hour average HCl-equivalent issued a notice of intent to disapprove
HCl-equivalent emission rate limit for emission rate calculated from the total your demonstration, you may
each affected source using either a look- chlorine emission rate you select for the nonetheless begin complying, on the
up table analysis or site-specific combustor; (3) determination of the 1- compliance date, with the annual
analysis. Look-up tables are provided hour average HCl-equivalent emission average HCl-equivalent emission rate
rate limit; (4) determination of the ratio limits you present in your eligibility
22 USEPA, ‘‘Technical Support Document for
of the 1-hour average HCl-equivalent demonstration.
HWC MACT Standards, Volume III: Selection of
MACT Standards,’’ September 2005, Section 24.2. emission rate to the 1-hour HCl- If your permitting authority issues a
23 See discussion below in Section F regarding the equivalent emission rate limit for the notice of intent to disapprove your
requirement to establish chlorine feedrate limits. combustor; and (5) determination of the eligibility demonstration after the
compliance date, the authority will standard until October 14, 2008, after under § 63.1209(o), except that feedrate
identify the basis for that notice and which you must comply with the final limits on total chlorine and chloride
specify how much time you will have to standard. must be established as described below.
submit additional information or to
F. Testing Requirements 1. Feedrate Limit to Ensure Compliance
comply with the MACT total chlorine
You must comply with the with the Annual Average HCl-
standards. The permitting authority may
requirements for comprehensive Equivalent Emission Rate Limit
extend the compliance date of the total
chlorine standard to allow you to make performance testing under § 63.1207. For sources subject to the feedrate
changes to the design or operation of the limit for total chlorine and chloride
1. Test Methods for Stack Gas
combustor or related systems as quickly under § 63.1209(n)(4) to ensure
Containing Alkaline Particulate
as practicable to enable you to achieve compliance with the semivolatile metals
compliance with the MACT standard for If you operate a cement kiln or a standard, the feedrate limit (and
total chlorine. combustor equipped with a dry acid gas averaging period) for total chlorine and
scrubber, you must use EPA Method chloride to ensure compliance with the
2. Review and Approval for New and 320/321 or ASTM D 6735–01, or an annual average HCl-equivalent emission
Reconstructed Sources equivalent method, to measure rate limit is the same as required by that
The procedures for review and hydrogen chloride, and the back-half paragraph. Thus, the chlorine feedrate
approval of eligibility demonstrations (caustic impingers) of Method 26/26A, limit is the average of the run averages
applicable to existing sources discussed or an equivalent method, to measure during the comprehensive performance
above also apply to new or chlorine. test, and is established as a 12-hour
reconstructed sources, except that the 2. Test Methods for Stack Gas rolling average.
date you must submit the eligibility Containing High Levels of Bromine or That chlorine feedrate limit cannot
demonstration is as discussed below. Sulfur exceed the numerical value (i.e., not
If you operate a new or reconstructed considering the averaging period) of the
source that starts up by April 12, 2007, If you operate an incinerator, boiler, feedrate limit that ensures compliance
or a solid fuel-fired boiler or liquid fuel- or lightweight aggregate kiln and your with the annual average HCl-equivalent
fired boiler that is an area source that feedstreams contain bromine or sulfur emission rate limit, however. Therefore,
increases its emissions or its potential to during the comprehensive performance the numerical value of the total chlorine
emit such that it becomes a major source test at the levels indicated below, you and chloride feedrate limit must not
of HAP before April 12, 2007, you must must use EPA Method 320/321 or exceed the value you calculate as the
either: (1) Submit an eligibility ASTM D 6735’01, or an equivalent annual average HCl-equivalent emission
demonstration for review and approval method, to measure hydrogen chloride, rate limit (lb/hr) divided by [1 ¥ system
by April 12, 2006 and comply with the and Method 26/26A, or an equivalent removal efficiency]. You must calculate
HCl-equivalent emission rate limits and method, to measure chlorine and a total chlorine system removal
operating requirements you establish in hydrogen chloride combined. You must efficiency for each test run of the
the eligibility demonstration; or (2) determine your chlorine emissions to be comprehensive performance test as [1 ¥
comply with the final total chlorine the higher of: (1) The value measured by total chlorine emission rate (g/s)/
emission standards under §§ 63.1216, Method 26/26A, or an equivalent chlorine feedrate (g/s)], and calculate
63.1217, 63.1219, 63.1220, and 63.1221, method; or (2) the value calculated by the average system removal efficiency of
by October 12, 2005, or upon startup, the difference between the combined the test run averages. If your source does
whichever is later, except for a standard hydrogen chloride and chlorine levels not control total chlorine, you must
that is more stringent than the standard measured by Method 26/26a, or an assume zero system removal efficiency.
proposed on April 20, 2004 for your equivalent method, and the hydrogen If emissions during the comprehensive
source. If a final standard is more chloride measurement from EPA performance test exceed the annual
stringent than the proposed standard, Method 320/321 or ASTM D 6735–01, average HCl-equivalent emission rate
you may comply with the proposed or an equivalent method. limit, eligibility for the health-based
standard until October 14, 2008, after These procedures apply if you feed emission limits is not affected. This is
which you must comply with the final during the comprehensive performance because the emission rate limit is an
standard. test bromine at a bromine/chlorine ratio annual average limit. Compliance is
If you operate a new or reconstructed in feedstreams greater than 5 percent by based on a 12-hour rolling average
source that starts up on or after April 12, mass, or sulfur at a sulfur/chlorine ratio chlorine feedrate limit (rather than an
2007, or a solid fuel-fired boiler or in feedstreams greater than 50 percent (up to) an annual averaging period) for
liquid fuel-fired boiler that is an area by mass.24 sources subject to the 12-hour rolling
source that increases its emissions or its Finally, you should precondition the average feedrate limit for total chlorine
potential to emit such that it becomes a M26/26A filter for one hour prior to and chloride under § 63.1209(n)(4) to
major source of HAP on or after April beginning the performance test to ensure compliance with the semivolatile
12, 2007, you must comply with either minimize the potential for a low bias metals standard given that the more
of the following. You may submit an caused by adsorption/absorption of stringent feedrate limit (i.e., the feedrate
eligibility demonstration for review and hydrogen chloride on the filter. limit with the shorter averaging period)
approval 12 months prior to startup. G. Monitoring Requirements would apply.
Alternatively, you may comply with the For sources exempt from the feedrate
final total chlorine emission standards You must establish and comply with limit for total chlorine and chloride
under §§ 63.1216, 63.1217, 63.1219, limits on the same operating parameters under § 63.1209(n)(4) because they
63.1220, and 63.1221 upon startup. If that apply to sources complying with comply with § 63.1207(m)(2) (which
the final standard is more stringent than the MACT standard for total chlorine allows compliance with the semivolatile
the standard proposed for your source 24 USEPA, ‘‘Technical Support Document for
metals emission standard absent
on April 20, 2004, however, and if you HWC MACT Standards, Volume IV: Compliance
emissions testing by assuming all metals
start operations before October 14, 2008, with the HWC MACT Standards,’’ September 2005, fed are emitted), the feedrate limit for
you may comply with the proposed Chapter 15.1.2. total chlorine and chloride to ensure
compliance with the annual average rate limits you present in your eligibility equivalent emission rate to the annual
HCl-equivalent emission rate must be demonstration; (4) performance test average HCl-equivalent emission rate
established as follows: emission rates for total chlorine and limit cannot not exceed 1.0. This will
• You must establish an average HCl-equivalent emissions; and (5) long- ensure that the RfC-based Hazard Index
period for the feedrate limit that does term and hourly rolling average chlorine of 1.0 is not exceeded, a principle
not exceed an annual rolling average; feedrate limits. criterion of the eligibility
• You must calculate a total chlorine demonstration.
system removal efficiency for each test 1. Total Chlorine Emission Rate, Annual If you use site-specific risk analysis to
run of the comprehensive performance Average HCl-Equivalent Emission Rate, demonstrate that a Hazard Index of 1.0
test as [1 ¥ total chlorine emission rate and Annual Average HCl-Equivalent is not exceeded, you would generally
(g/s)/chlorine feedrate (g/s)], and Emission Rate Limit identify for each combustor the
calculate the average system removal For the eligibility demonstration, you maximum annual average HCl-
efficiency of the test run averages. If must select a total chlorine emission equivalent emission rate that the risk
your source is not equipped with a concentration (ppmv) for each assessment estimates would result in an
control system that consistently and combustor, determine the Cl2/HCl RfC-based Hazard Index of 1.0 at any
reproducibly controls total emissions volumetric ratio, calculate the annual off-site receptor location (i.e.,
(e.g., wet or dry scrubber), you must average HCl-equivalent emission rate considering locations where people
assume zero system removal efficiency. (lb/hr), and document that the emission reside and where people congregate for
If emissions during the comprehensive rate does not exceed the annual average work, school, or recreation.25 This
performance test exceed the annual HCl-equivalent emission rate limit. emission rate would be the annual
average HCl-equivalent emission rate You select a total chlorine (i.e., HCl average HCl-equivalent emission rate
limit, eligibility for emission limits and chlorine combined) emission limit for each combustor.
under this section is not affected. The concentration (ppmv) for each
hazardous waste combustor expressed 2. 1-Hour Average HCl-Equivalent
emission rate limit is an annual average
as chloride (Cl(-)) equivalent. For Emission Rate and Emission Rate Limit
limit and compliance is based on an
annual average feedrate limit on total incinerators, cement kilns, and As discussed in Section D above, you
chlorine and chloride (or a shorter lightweight aggregate kilns, this must determine in your eligibility
averaging period if you so elect under emission concentration cannot exceed demonstration whether the 1-hour HCl-
paragraph (g)(2)(ii)(A) of this section); the Interim Standards for total chlorine. equivalent emission rate limit may be
and You then determine the average Cl2/HCl exceeded absent an hourly rolling
• You must calculate the feedrate volumetric ratio considering all average chlorine feedrate limit. To make
limit for total chlorine and chloride as historical regulatory emissions tests and this determination, you must establish a
the annual average HCl-equivalent apportion total chlorine emissions 1-hour average HCl-equivalent emission
emission rate limit (lb/hr) divided by [1 between Cl2 and HCl accordingly. You rate and a 1-hour average HCl-
¥ system removal efficiency] and use these apportioned volumetric equivalent emission rate limit.
comply with the feedrate limit on the emissions to calculate the Cl2 and HCl You calculate the 1-hour average HCl-
averaging period you establish. emission rates (lb/hr) using the average equivalent emission rate from the total
gas flowrate (and other relevant chlorine emission rate, established as
2. Feedrate Limit To Ensure Compliance parameters) for the most recent RCRA discussed above, using the equation in
With the 1-Hour Average HCl- compliance test or MACT performance § 63.1215(b)(3).
Equivalent Emission Rate Limit test for total chlorine. Finally, you use You establish the 1-hour average HCl-
You must establish an hourly rolling these Cl2 and HCl emission rates to equivalent emission rate limit by either
average feedrate limit on total chlorine calculate an annual average HCl- using Tables 3 or 4 in § 63.1215 to look-
and chloride to ensure compliance with equivalent emission rate, which cannot up the limit, or conducting a site-
the 1-hour average HCl-equivalent exceed the annual average HCl- specific risk analysis. Under the site-
emission rate limit unless you equivalent emission rate limit that you specific risk analysis option, the 1-hour
determine that the hourly rolling establish as discussed below. average HCl-equivalent emission rate
average feedrate limit is waived as To establish the annual average HCl- limit would be the highest emission rate
discussed under Section D above. If equivalent emission rate limit, you may that the risk assessment estimates would
required, you must calculate the hourly either use Tables 1 or 2 in § 63.1215 to result in an aREL-based Hazard Index
rolling average feedrate limit for total look-up the limit, or conduct a site- not exceeding 1.0 at any off-site receptor
chlorine and chloride as the 1-hour specific risk analysis. Under the site- location (i.e., considering locations
average HCl-equivalent emission rate specific risk analysis option, the annual where people reside and where people
limit (lb/hr) divided by [1 ¥ system average HCl-equivalent emission rate congregate for work, school, or
removal efficiency] using the system limit would be the highest emission rate recreation).
removal efficiency demonstrated during that the risk assessment estimates would 3. Performance Test Emissions
the comprehensive performance test. result in a Hazard Index not exceeding
1.0 for the actual individual most During the comprehensive
H. Relationship Among Emission Rates, exposed to the facility’s emissions performance test, you must demonstrate
Emission Rate Limits, and Feedrate considering off-site locations where a system removal efficiency for total
Limits people reside and where people chlorine as [1 ¥ TCl emitted (lb/hr)/
We summarize here the relationship congregate for work, school, or chlorine fed (lb/hr)]. During the test,
among: (1) the total chlorine emission recreation. however, the total chlorine emission
rate you select in your eligibility If you have more than one on-site rate you select for each combustor and
demonstration; (2) the annual average hazardous waste combustor, and if you the annual average HCl-equivalent
and 1-hour average HCl-equivalent use the look-up tables to establish the 25 Note again, however, that the total chlorine
emission rates you present in your annual average HCl-equivalent emission emission concentration (ppmv) is capped by the
eligibility demonstration; (3) the annual rate limits, the sum of the ratios for all Interim Standards for incinerators, cement kilns,
average and 1-hour average emission combustors of the annual average HCl- and lightweight aggregate kilns.
emission rate limit can exceed the levels than the feedrate during the If you plan to change the design,
you present in the eligibility performance test. This is because the operation, or maintenance of the facility
demonstration. This is because those averaging period for this chlorine in a manner than would increase the
emission rates are annual average rates feedrate limit (that ensures compliance annual average or 1-hour average HCl-
and need not be complied with over the with the annual average HCl-equivalent equivalent emission rate limit, and you
duration of three runs of the emission rate limit) is up to an annual elect to increase your total chlorine and
performance test, which may be rolling average. See § 63.1215(g)(2). chloride feedrate limits, prior to the
nominally only 3 hours. Thus, the chlorine feedrate, and total change you must submit to the
The 1-hour average HCl-equivalent chlorine emissions, can be higher than permitting authority a revised eligibility
emission rate limit cannot be exceeded the limit during the relatively short demonstration documenting the
during any run of the comprehensive duration of the comprehensive increased emission rate limits and
performance test, however. This limit is performance tests. calculations of the increased feedrate
based on an aREL Hazard Index of 1.0; Hourly Rolling Average Chlorine limits prior to the change.
an exceedance of the limit over a test Feedrate Limit. If you determine under Changes That Affect System Removal
run with a nominal 1-hour duration § 63.1205(d)(3) that the 1-hour average Efficiency. If you plan to change the
would result in a Hazard Index of HCl-equivalent emission rate limit may design, operation, or maintenance of the
greater than 1.0. be exceeded, you must establish an combustor in a manner than could
hourly rolling average chlorine feedrate decrease the system removal efficiency,
4. Chlorine Feedrate Limits
limit. That feedrate limit is established you are subject to the requirements of
To maintain compliance with the § 63.1206(b)(5) for conducting a
annual average HCl-equivalent emission as the 1-hour HCl-equivalent emission
rate limit divided by [1 ¥ system performance test to reestablish the
rate limit, you must establish a long- combustor’s system removal efficiency.
term average chlorine feedrate limit. In removal efficiency]. The hourly rolling
average chlorine feedrate limit is not You also must submit a revised
addition, if you determine under eligibility demonstration documenting
§ 63.1205(d)(3) that the 1-hour average established based on feedrates during
the lower system removal efficiency and
HCl-equivalent emission rate may be the performance test because
the reduced feedrate limits on total
exceeded (i.e., because your chlorine performance test feedrates may be
chlorine and chloride.
feedrate may vary substantially over the substantially lower than the feedrate If you plan to change the design,
averaging period for the long-term needed to ensure compliance with the operation, or maintenance of the
chlorine feedrate limit), you must 1-hour average HCl-equivalent emission combustor in a manner than could
establish an hourly rolling average rate. Note, however, that the hourly increase the system removal efficiency,
chlorine feedrate limit. rolling average feedrate limit cannot be and you elect to document the increased
Long-Term Chlorine Feedrate Limit. exceeded during any run of the system removal efficiency to establish
The chlorine feedrate limit to maintain comprehensive performance test. This higher feedrate limits on total chlorine
compliance with the annual average chlorine feedrate limit is based on the and chloride, you are subject to the
HCl-equivalent emission rate is either: 1-hour average HCl-equivalent emission requirements of § 63.1206(b)(5) for
(1) The chlorine feedrate during the rate limit, which is based on an aREL conducting a performance test to
comprehensive performance test if you Hazard Index of 1.0. Thus, an reestablish the combustor’s system
demonstrate compliance with the exceedance of the hourly rolling average removal efficiency. You must also
semivolatile metals emission standard feedrate limit (and the 1-hour lHCl- submit a revised eligibility
during the test (see § 63.1209(o)); or (2) equivalent emission rate limit) over a demonstration documenting the higher
if you comply with the semivolatile test run with a nominal 1-hour duration system removal efficiency and the
metals emission standard under would result in a Hazard Index of increased feedrate limits on total
§ 63.1207(m)(2) by assuming all metals greater than 1.0. chlorine and chloride.
in the feed to the combustor are emitted, I. Changes 2. Changes Over Which You Do Not
the HCl-equivalent emission rate limit
Your requirements will change in Have Control
divided by [1 ¥ system removal
efficiency] where you demonstrate the response to changes that affect the HCl- If you use site-specific risk assessment
system removal efficiency during the equivalent emission rate or HCl- in lieu of the look-up tables to establish
comprehensive performance test. equivalent emission rate limit for a the HCl-equivalent emission rate limit,
If you establish the chlorine feedrate source. you must review the documentation you
limit based on the feedrate during the use in your eligibility demonstration
1. Changes Over Which You Have
performance test to demonstrate every five years from the date of the
Control
compliance with the semivolatile metals comprehensive performance test and
emission standard, the averaging period Changes That Affect HCl-Equivalent submit for review and approval with the
for the feedrate limit is a 12-hour rolling Emission Rate Limits. If you plan to comprehensive performance test plan
average. If you establish the chlorine change the design, operation, or either a certification that the
feedrate limit based on the system maintenance of the facility in a manner information used in your eligibility
removal efficiency during the that would decrease the annual average demonstration has not changed in a
performance test, the averaging period is or 1-hour average HCl-equivalent manner that would decrease the annual
up to an annual rolling average. See emission rate limit (e.g., reduce the average HCl-equivalent emission rate
discussion in Part Four, Section VII.B of distance to the property line; reduce limit, or a revised eligibility
this preamble. stack gas temperature; reduce stack demonstration. Examples of changes
If you comply with the semivolatile height), prior to the change you must beyond your control that may decrease
metals emission standard under submit to the permitting authority a the annual average HCl-equivalent
§ 63.1207(m)(2), however, the long-term revised eligibility demonstration emission rate limit (or 1-hour average
chlorine feedrate limit is based on the documenting the lower emission rate HCl-equivalent emission rate limit) are
system removal efficiency during the limits and calculations of reduced total construction of residences at a location
comprehensive performance test rather chlorine and chloride feedrate limits. exposed to higher ambient
concentrations than evaluated during combustion unit, and the efficiency of mercury and dioxin/furan emissions
your previous risk analysis, or a control of the unit’s outputs. This data from source 327, and have instead
reduction in the RfCs or aRELs. methodology reasonably selects the best used data from an older test condition
If, in the interim between the dates of performing (and for new sources, best to represent this source’s emissions
your comprehensive performance tests, controlled) sources, and reasonably because the source encountered
you have reason to know of changes that assesses their level of performance. problems with its carbon injection
would decrease the annual average HCl- When HAP feed control is not feasible, system during the most recent test. See
equivalent emission rate limit, you must notably where HAP is contributed by part four, section I.F. Emissions data
submit a revised eligibility raw material and fossil fuel inputs, EPA from source 3006 has been excluded for
demonstration as soon as practicable but determined best performers and their purposes of calculating the semivolatile
not more frequently than annually. level of performance using a metal standard because this source did
If you determine that you cannot methodology that selects the lowest not measure cadmium emissions during
demonstrate compliance with a lower emitters using the best air pollution its emissions test. See part four, section
annual average HCl-equivalent emission control technology. This methodology I.F. We have added mercury emissions
rate limit (dictated by a change over reasonably estimates the best data from source 901 (DSSI) to the
which you do not have control) during performing sources’ level of incinerator mercury database because
the comprehensive performance test performance, and better accounts for this source (which is otherwise subject
because you need additional time to total variability in emissions levels of to standards for liquid fuel boilers) is
complete changes to the design or the best performing sources. burning a waste which is unlike that
operation of the source or related EPA carefully examined approaches burned by any other liquid fuel boiler
systems, you may request that the selecting lowest emitters as best with respect to mercury concentration
permitting authority grant you performers. Examination of other test and waste provenance, but typical of
additional time to make those changes conditions from the same best waste burned by incinerators with
as quickly as practicable. performing sources shows, however, respect to those factors. See part four,
X. Overview on Floor Methodologies that this approach results in standards section VI.D.1. This change
not achievable even by the best correspondingly affects the liquid fuel
The most contentious issue in the performers. Indeed, in order to meet
rulemaking involved methodologies for boiler standard by removing that data
such standards, even ‘‘best performing’’ from the liquid fuel boiler database.
determining MACT floors, namely, sources (lowest emitting in individual
which sources are best performing, and tests) would have to add additional air B. Hazardous Waste Cement Kilns
what is their level of performance. pollution control technology. EPA views
Superficially, these questions have a 1. Use of Emissions Data From Ash
this result as an end run around the Grove Cement Company
ready answer: the best performers are section 112(d)(2) beyond-the-floor
the lowest emitters as measured by process, because floor standards would The emissions data from Ash Grove
compliance tests, and those tests fix force industry-wide technological Cement Company, which operates a
their level of performance. But changes without consideration of the recently constructed preheater/
compliance tests are snapshots which factors (cost and energy in particular) precalciner kiln located in Chanute,
do not fully capture sources’ total which Congress mandated for Kansas, are considered when calculating
operating variability. Since the consideration when establishing MACT floors for new hazardous waste
standards must be met at all times, beyond-the-floor standards. burning cement kilns. In the proposal,
picking lowest compliance test data to we did not consider their emissions data
set the standard results in standards best Part Three: What Are the Major in the floor analyses for existing sources
performing sources themselves would Changes Since Proposal? because Ash Grove Cement used the
be unable to meet at all times. I. Database data to demonstrate compliance with
To avoid this impermissible result, the new source interim standards, and
EPA selected approaches that A. Hazardous Burning Incinerators
did not address the data for purposes of
reasonably estimate best performing Five incinerators have been removed new source standards. See 69 FR at
sources’ total variability. Certain types from the database because they have 21217 n. 35. Consistent with our
of variability can be quantified initiated or completed RCRA closure.26 position on use of post-1999 emissions
statistically, and EPA did so here (using Two incinerators have been added to data, we are including Ash Grove
standard statistical approaches) in all of the list of sources used to calculate the Cement’s emissions data in the floor
the floor methodologies used in the rule. floor levels.27 Emissions data from analyses for new sources. See also Part
There are other components of source 3015 has been excluded for Four, Section I.B of the preamble.
variability, however, which cannot be purposes of calculating the particulate
fully quantified, but nonetheless must matter floor because the source was 2. Removal of Holcim’s Emissions Data
be accounted for in reasonably processing an atypical waste stream From EPA’s HWC Data Base
estimating best performing sources’ from a particulate matter compliance Following cessation of hazardous
performance over time. EPA selected perspective. See part four, section I.F. waste operations in 2003, we are
ranking methodologies which best We have excluded the most recent removing all emissions data from both
account for this total variability. wet process cement kilns at Holcim’s
Where control of the feed of HAP is 26 See ‘‘Final Technical Support Document for
Holly Hill, South Carolina, plant from
feasible and technically assessable (the HWC MACT Standards, Volume II: HWC Database’’ our hazardous waste combustor data
for a list of the sources that have initiated or
case for HAP metals and for total completed RCRA closure. base. This is consistent with our
chlorine), EPA used a methodology that 27 We noticed the data from these sources but did approach in both this rule and the 1999
ranked sources by their ability to best not include them in the MACT standard rule to base the standards only on
control both HAP feed and HAP calculations at proposal. Note that inclusion of performance of sources that actually are
these sources did not affect any of the calculated
emissions. This methodology thus MACT standards. See ‘‘Final Technical Support
operating (i.e., burning hazardous
assesses the efficiency of control of both Document for HWC MACT Standards, Volume II: waste). See also Part Four, Section I.A
the HAP inputs to a hazardous waste HWC Database’’ for more discussion. and 64 FR at 52844.
3. Use of Mercury Data D. Liquid Fuel Boilers as a boiler based on comments received
In the proposed rule, we classified at proposal.
As discussed below, we are using a
commenter-submitted dataset as the liquid fuel boilers as one category. The F. Total Chlorine Emissions Data Below
basis of the mercury standards for final rule classifies them into two for 20 ppmv
existing and new cement kilns. This purposes of the mercury, semivolatile
We corrected all the total chlorine
metals, chromium, and total chlorine
comprehensive dataset documents the measurements in the data base that were
standards: one for liquid fuel boilers
day-to-day levels of mercury in below 20 ppmv to account for potential
burning lower heating value hazardous
hazardous waste fired to all cement systemic negative biases in the Method
waste (hazardous waste with a heating
kilns for a three year period covering 0050 data in response to comments on
value less than 10,000 Btu/lb), and
1999 to 2001. We have determined that the proposed rule. See the discussion in
another for liquid fuel boilers burning
the commenter-submitted data are more Part Four, Section I.C.1 below.
higher heating value hazardous waste
representative than data used at To account for the bias, we corrected
(hazardous waste with a heating value
proposal. See Part Four, Section I.D of all total chlorine emissions data that
of 10,000 Btu/lb or greater).
the preamble for our rationale. were below 20 ppmv to 20 ppmv. We
We also made other, minor changes to
accounted for within-test condition
C. Hazardous Waste Lightweight the data base because some sources have
emissions variability for the corrected
Aggregate Kilns initiated closure, were misclassified as
data by imputing a standard deviation
other sources in the proposed rule, or
that is based on a regression analysis of
We are incorporating mercury data were inadvertently not considered in
run-to-run standard deviation versus
submitted by a commenter into the the floor calculations although the
emission concentration for all data
MACT floor analysis for existing and sources’ test reports were in the docket
above 20 ppmv. This approach of using
new lightweight aggregate kilns. These at proposal.
a regression analysis to impute a
data document the day-to-day levels of E. HCl Production Furnaces standard deviation is similar to the
mercury in hazardous waste fired to approach we used to account for total
Six of the 17 hydrochloric acid
lightweight aggregate kilns located at variability (i.e., test-to-test and within
production furnaces have ceased
Solite Corporation’s Arvonia plant test variability) of PM emissions for
burning hazardous waste since
between October 2003 and June 2004. proposal. Consequently, we do not use sources that use fabric filters.
We have determined that the emissions data from these sources to II. Emission Limits
commenter-submitted data are more establish the final standards. All six of
representative than the data used at these sources were equipped with waste A. Incinerators
proposal. See Part Four, Section I.E of heat recovery boilers and had relatively The changes in the incinerator
the preamble for our rationale. high dioxin/furan emissions. In standards for existing sources since
addition, we reclassified source #2020 proposal are:
Standard Proposed limit Final limit
Dioxin/Furans (ng TEQ/dscm) ........................... Sources with dry air pollution control systems For all sources, 0.20 or 0.40 and temperature
or waste heat boilers: 0.28; For others: 0.2 control < 400 °F at the air pollution control
or 0.4 and temperature control at inlet of air device inlet.
pollution control device < 400 °F.
Particulate Matter (gr/dscf) ................................ 0.015 ................................................................ 0.013.
Semivolatile Metals (µg/dscm) ........................... 59 ..................................................................... 230.
Low Volatile Metals (µg/dscm) ........................... 84 ..................................................................... 92.
Total Chlorine (ppmv) ........................................ 1.5 .................................................................... 32.
Alternative to the particulate matter standard: 59 ..................................................................... 230.
Combined emissions of lead, cadmium and
selenium (µg/dscm).
Alternative to the particulate matter standard: 84 ..................................................................... 92.
Combined emissions of arsenic, berrylium,
chrome, antimony, cobalt, manganese, and
nickel (µg/dscm).
The changes in the incinerator

standards for new sources since
proposal are:
Proposed
Standard Final limit
limit
Particulate Matter (gr/dscf) .................................................................................................................................................. 0.0007 0.0015

Mercury (µg/dscm) ............................................................................................................................................................... 8.0 8.1
Semivolatile Metals (µg/dscm) ............................................................................................................................................. 6.5 10
Low Volatile Metals (µg/dscm) ............................................................................................................................................ 8.9 23
Total Chlorine (ppmv) .......................................................................................................................................................... 0.18 21
Alternative to the particulate matter standard: Combined emissions of lead, cadmium and selenium (µg/dscm) ............ 6.5 10
Alternative to the particulate matter standard: Combined emissions of arsenic, berrylium, chrome, antimony, cobalt,
manganese, and nickel (µg/dscm) ................................................................................................................................... 8.9 23
Hazardous Waste Burning Cement Kilns

The changes in the standards for
existing cement kiln since proposal are:
Mercury (µg/dscm) ............................................. 64 1 ................................................................... Both 3.0 ppmw 2 and either 120 µg/dscm
(stack emissions) or 120 µg/dscm (ex-
pressed as a hazardous waste MTEC) 3.
Particulate matter ............................................... 0.028 gr/dscf .................................................... 0.028 gr/dscf and 20% opacity 4.
Semivolatile metals ............................................ 4.0E–04 lb/MMBtu 5 ......................................... 7.6E–04 lb/MMBtu 5 and 330 µg/dscm.
Low volatile metals ............................................. 1.4E–05 lb/MMBtu 5 ......................................... 2.1E–05 lb/MMBtu 5 and 56 µg/dscm.
Total chlorine (ppmv) 6 ....................................... 110 ................................................................... 120.
1 The proposed mercury standard was an annual limit.
2 Feed concentration of mercury in hazardous waste as-fired.
3 HW MTEC means maximum theoretical emissions concentration of the hazardous waste and MTEC is defined at § 63.1201(a).
4 The opacity standard does not apply to a source equipped with a bag leak detection system under § 63.1206(c)(8) or a particulate matter de-
tection system under § 63.1206(c)(9).
5 Standard is expressed as mass of pollutant stack emissions attributable to the hazardous waste per million British thermal unit heat input of
the hazardous waste.
6 Combined standard, reported as a chloride (Cl(-)) equivalent.
The changes in the standards for new

cement kilns since proposal are:
Mercury (µg/dscm) ............................................. 35 1 ................................................................... Both 1.9 ppmw 2 and either 120 µg/dscm
(stack emissions) or 120 µg/dscm (ex-
pressed as a hazardous waste MTEC) 3.
Particulate matter ............................................... 0.0058 gr/dscf .................................................. 0.0023 gr/dscf and 20% opacity 4.
Semivolatile metals ............................................ 6.2E–05 lb/MMBtu 5 ......................................... 6.2E–05 lb/MMBtu 5 and 180 µg/dscm.
Low volatile metals ............................................. 1.4E–05 lb/MMBtu 5 ......................................... 1.5E–05 lb/MMBtu 5 and 54 µg/dscm.
Total chlorine (ppmv) 6 ....................................... 78 ..................................................................... 86.
1 The proposed mercury standard was an annual limit.
2 Feed concentration of mercury in hazardous waste as-fired.
3 HW MTEC means maximum theoretical emissions concentration of the hazardous waste and MTEC is defined at § 63.1201(a).
4 The opacity standard does not apply to a source equipped with a bag leak detection system under § 63.1206(c)(8) or a particulate matter de-
tection system under § 63.1206(c)(9).
5 Standard is expressed as mass of pollutant stack emissions attributable to the hazardous waste per million British thermal unit heat input of
6 Combined standard, reported as a chloride (Cl(-)) equivalent.
C. Hazardous Waste Burning

Lightweight Aggregate Kilns
The changes in the standards for
existing lightweight aggregate kilns
since proposal are:
Dioxins and furans (ng TEQ/dscm) .................... 0.40 .................................................................. 0.20 or rapid quench of the flue gas at the
exit of the kiln to less than 400 °F.
Mercury (µg/dscm) ............................................. 67 1 ................................................................... 120 µg/dscm (stack emissions) or 120 µg/
dscm (expressed as a hazardous waste
MTEC) 2.
Semivolatile metals ............................................ 3.1E–04 lb/MMBtu 3 and 250 µg/dscm ............ 3.0E–04 lb/MMBtu 3 and 250 µg/dscm.
1 The
proposed mercury standard was an annual limit.
2 HWMTEC means maximum theoretical emissions concentration of the hazardous waste and MTEC is defined at § 63.1201(a).
3 Standard
is expressed as mass of pollutant stack emissions attributable to the hazardous waste per million British thermal unit heat input of
The changes in the standards for new

lightweight aggregate kilns since
proposal are:
Dioxins and furans (ng TEQ/dscm) .................... 0.40 .................................................................. 0.20 or rapid quench of the flue gas at the
exit of the kiln to less than 400 °F.
Particulate matter ............................................... 0.0099 gr/dscf .................................................. 0.0098 gr/dscf.

Mercury (µg/dscm) ............................................. 67 1 ................................................................... 120 µg/dscm (stack emissions) or 120 µg/
dscm (expressed as a hazardous waste
MTEC) 2.
Semivolatile metals ............................................ 2.4E–05 lb/MMBtu 3 and 43 µg/dscm .............. 3.7E–05 lb/MMBtu 3 and 43 µg/dscm.
1 The
proposed mercury standard was an annual limit.
2 HWMTEC means maximum theoretical emissions concentration of the hazardous waste and MTEC is defined at § 63.1201(a).
3 Standard
is expressed as mass of pollutant stack emissions attributable to the hazardous waste per million British thermal unit heat input of
D. Solid Fuel Boilers

The changes in the solid fuel boiler
standards for existing sources since
proposal are:
Proposed Final
Standard limit limit
Mercury (µg/dscm) ................................................................................................................................................................... 10 11

Semivolatile Metals (µg/dscm) ................................................................................................................................................. 170 180
Low Volatile metals (µg/dscm) ................................................................................................................................................ 210 380
Alternative to the particulate matter standard: Combined emissions of lead, cadmium and selenium (µg/dscm) ................ 170 180
Alternative to the particulate matter standard: Combined emissions of arsenic, beryllium, chromium, antimony, cobalt,
manganese, and nickel (µg/dscm) ....................................................................................................................................... 210 380
The changes in the solid fuel boiler

proposal are:
Proposed Final
Standard limit limit
Mercury (µg/dscm) ................................................................................................................................................................... 10 11

Semivolatile Metals (µg/dscm) ................................................................................................................................................. 170 180
Low Volatile metals (µg/dscm) ................................................................................................................................................ 210 380
Alternative to the particulate matter standard: Combined emissions of lead, cadmium and selenium (µg/dscm) ................ 170 180
E. Liquid Fuel Boilers of the hazardous waste they burn: Those today’s preamble for a complete
that burn waste below 10,000 Btu/lb, discussion.
We redefined the liquid fuel boiler those that burn hazardous waste with a The additional changes to the liquid
subcategory into two separate boiler heating value of 10,000 Btu/lb or fuel boiler standards for existing sources
subcategories based on the heating value greater. See Part Four, Section VI.D.2 of since proposal are:
Final limit
Proposed
Standard limit HW Fuel < HW Fuel ≥
10,000 Btu/lb 10,000 Btu/lb
Mercury (lb/MM Btu) ........................................................................................................................... 3.7E–6 ......... 19 µg/dscm 4.2E–5

Particulate matter (gr/dscf) ................................................................................................................. 0.032 ........... 0.035
Semivolatile metals (lb/MM Btu) ......................................................................................................... 1.1E–5 ......... 150 µg/dscm 8.2E–5
Chromium (lb/MM Btu) ........................................................................................................................ 1.1E–4 ......... 370 µg/dscm 1.3E–4
Total chlorine (Lb/MM Btu) ................................................................................................................. 2.5E–2 ......... 31 ppmv ...... 5.1E–2
Alternative to the particulate matter standard: Combined emissions of lead, cadmium and sele- 1.1E–5 ......... 150 µg/dscm 8.2E–5
nium (lb/MM Btu).
Alternative to the particulate matter standard: Combined emissions of arsenic, beryllium, chro- 1.1E–4 ......... 370 µg/dscm 1.3E–4
mium, antimony, cobalt, manganese, and nickel (lb/MM Btu).
The changes in the liquid fuel boiler

proposal are:
Final limit
Standard Proposed limit HW fuel < HW fuel >
10,000 Btu/lb 10,000 Btu/lb
Dioxin and Furan, dry APCD (ng TEQ/dscm) ............................................... 0.015 or temp control <400F for dry 0.40
APCD.
Mercury (lb/MM Btu) ...................................................................................... 3.8E–7 .............................................. 6.8 µg/dscm 1.2E–6
Particulate matter (gr/dscf) ............................................................................ 0.0076 ............................................... 0.0087
Semivolatile metals (lb/MM Btu) ................................................................... 4.3E–6 .............................................. 78 µg/dscm 6.2E–6
Chromium (lb/MM Btu) .................................................................................. 3.6E–5 .............................................. 12 µg/dscm 1.4E–5
Total chlorine (lb/MM Btu) ............................................................................. 7.2E–4 .............................................. 31 µg/dscm 5.1E–2
Alternative to the particulate matter standard: Combined emissions of 4.3E–6 .............................................. 78 µg/dscm 1 6.2E–6 1
lead, cadmium and selenium (lb/MM Btu).
Alternative to the particulate matter standard: Combined emissions of ar- 3.6E–5 .............................................. 12 µg/dscm 2 1.4E–5 2
senic, beryllium, chromium, antimony, cobalt, manganese, and nickel
(lb/MM Btu).
1 New or reconstructed liquid fuel boilers that process residual oil or liquid feedstreams that are neither fossil fuel nor hazardous waste and that
operate pursuant to the alternative to the particulate matter standard must comply with the alternative emission concentration standard of 4.7 µg/
dscm, which is applicable to lead, cadmium and selenium emissions attributable to all feedstreams (hazardous and nonhazardous).
2 New or reconstructed liquid fuel boilers that process residual oil or liquid feedstreams that are neither fossil fuel nor hazardous waste that op-
erate pursuant to the alternative to the particulate matter standard must comply with the alternative emission concentration standard of 12 µg/
dscm, which is applicable to arsenic, beryllium, chrome, antimony, cobalt, manganese, and nickel emissions attributable to all feedstreams (haz-
ardous and nonhazardous).
F. Hydrochloric Acid Production

Furnaces
The changes in the hydrochloric acid
production furnace standards for
existing sources since proposal are:
Dioxin and Furans ................ 0.4 ng TEQ/dscm ............................................................ Carbon Monoxide/Total Hydrocarbons and DRE stand-
ards as surrogates.
Total chlorine ....................... 14 ppmv or 99.9927% system removal efficiency .......... 150 ppmv or 99.923% system removal efficiency.
The changes in the hydrochloric acid

production furnace standards for new
sources since proposal are:
Dioxin and Furans ................ 0.4 ng TEQ/dscm ............................................................ Carbon Monoxide/Total Hydrocarbons and DRE stand-
ards as surrogates
Total chlorine ....................... 1.2 ppmv or 99.9994% system removal efficiency ......... 25 ppmv or 99.987% system removal efficiency
G. Dioxin/Furan Testing for Sources Not explained in Volume III of the Technical nondetect measurements are assumed to
Subject to a Numerical Standard Support Document, Sections 12 and 15. be present at the detection limit.
Today’s final rule requires that all We note that sources not subject to a At proposal, we assumed that
sources not subject to a numerical numerical dioxin/furan emission nondetects (i.e., HAP levels in stack
dioxin and furan standard perform a one standard are subject to the carbon emissions below the level of detection
time test to determine their dioxin and monoxide or hydrocarbon standards and of the applicable analytic method) are
furan emissions. See the discussion in the DRE standard as surrogates. invariably present at the detection limit.
Part Four, Section VII.L. We are making no changes to the Commenters on the proposed rule
In the proposed rule, this requirement implementation of this requirement. See stated, however, that assuming
was limited to solid fuel boilers and the proposed rule at 69 FR at 21307 for nondetects are present at the detection
those liquid fuel boilers with a wet or more information. limit dampens emissions variability—a
no air pollution control system. The consideration necessary to reasonably
final rule expands this requirement to III. Statistics and Variability ascertain sources’ performance over
include hydrochloric acid production A. Using Statistical Imputation To time. This could have significant
furnaces and those lightweight aggregate Address Variability of Nondetect Values practical consequence for those data sets
kilns that elect to comply with the (such as the data base for liquid fuel
temperature limit at the kiln exit in lieu In the final rule, we use a statistical boilers) dominated by nondetected
of the 0.20 ng TEQ/dscm dioxin/furan approach to impute the value of values. We agree with these
standard. Those sources are not subject nondetect emissions and feedrate commenters, and instead of making the
to a numerical dioxin/furan standard measurements to avoid dampening of arbitrary assumption that all
under the final rule for reasons the variability of data sets when nondetected values are identical (which
in fact is highly unlikely), we are using runs to identify the t-statistic rather than specific operating parameter limits for
a statistical methodology to impute the assuming nine runs is appropriate given compliance assurance for sources
value of nondetect measurements. that the true test condition average is equipped with electrostatic precipitators
The imputation approach assigns a less certain for sources with only three and ionizing wet scrubbers. The rule
value for each nondetect measurement runs, and thus there is less certainty in explicitly allows you to maximize
in a data set within the possible range the imputed standard deviation. The controllable operating parameters
of values that results in maximizing the higher t-statistic associated with a three- during the comprehensive performance
99th percentile upper prediction limit run data set reflects this uncertainty. test to account for variability by, for
for the data set. For example, the In addition, we include emissions example, detuning the APCD or spiking
possible range of values for a data classified as ‘‘normal’’ in the ash. To establish the alarm set-point,
measurement that is 100% nondetect is regression analysis for the final rule. At you may either establish the set-point as
between zero and the detection limit. proposal, we used only data classified the average of the test condition run
On February 4, 2005 we distributed a as CT (i.e., highest compliance test average detector responses during the
direct request for comments on the condition in a test campaign) or IB (i.e., comprehensive performance test or
imputation approach to major a compliance test condition that extrapolate the detector response after
stakeholders. We respond to the achieved lower emissions than another approximating the correlation between
comments we received in Part Four, compliance test condition in the test the detector response and particulate
Section IV.D of today’s notice. campaign). We conclude that normal matter emission concentrations. You
B. Degrees of Freedom When Imputing data (i.e., emissions data that were not may extrapolate the detector response
a Standard Deviation Using the used to establish operating limits and up to a response value that corresponds
Universal Variability Factor for thus do not reflect variability in to 50% of the particulate matter
Particulate Matter Controlled by a controllable operating parameters) emission standard or 125% of the
Fabric Filter should also be considered in the highest particulate matter concentration
regression analysis because particulate used to develop the correlation,
The use of the universal variability
factor to impute a standard deviation for matter emissions are relatively whichever is greater. To establish an
particulate emissions from sources insensitive to baghouse inlet loading approximate correlation of the detector
controlled with a fabric filter takes and operating conditions.29 Including response to particulate matter emission
advantage of the empirical observation normal emissions in the analysis concentrations you should use as
that the standard deviation of provides additional data to better guidance Performance Specification-11
particulate emissions from sources is quantify these devices’ performance for PM CEMS (40 CFR Part 60,
positively correlated to the average variability. Appendix B), except that you need
particulate emissions of sources. Based conduct only 5 runs to establish the
IV. Compliance Assurance for Fabric
on this observation, we use regression initial correlation rather than a
Filters, Electrostatic Precipitators, and
analysis to determine the best fitting minimum of 15 runs required by PS–11.
Ionizing Wet Scrubbers The final rule also notes that an
curve to explain the relationship of
average value to standard deviation. The final rule provides additional exceedance of a detector response that
In the final rule, we use the actual requirements to clarify how you corresponds to the particulate matter
sample size, rather than an assumed determine the duration of periods of emission standard is not evidence that
sample size of nine used at proposal, to operation when the alarm set point has the standard has been exceeded because
determine the degrees of freedom for the been exceeded for a bag leak detection the correlation is an approximate
t-statistic to calculate the floor using the system or a particulate matter detection correlation used for the purpose of
standard deviation imputed from the system: compliance assurance to determine
universal variability factor for 1. You must keep records of the date, when corrective measures must be
particulate matter controlled by a fabric time, and duration of each alarm, the taken. The correlation, however, does
filter. time corrective action was initiated and not meet the requirements of PS–11 for
At proposal, we used eight degrees of completed, and a brief description of the compliance monitoring.
freedom to identify the t-statistic to cause of the alarm and the corrective In addition, if you elect to use a
account for within-test condition action taken. particulate matter detection system in
variability (i.e., run-to-run variability) 2. You must record the percent of the lieu of site-specific control device
for standard deviations imputed from operating time during each 6-month operating parameter limits on the
the universal variability factor period that the alarm sounds. electronic control device, the ash
regression.28 This is because, on 3. In calculating the operating time feedrate limit for incinerators and
average, about three test conditions with percentage, if inspection of the fabric boilers under § 63.1209(m)(3) is waived.
nine individual test runs are associated filter, electrostatic precipitator, or The ash feedrate limit is waived because
with each source used to develop the ionizing wet scrubber demonstrates that the particulate matter detection system
regression curve. no corrective action is required, no continuously monitors relative
A commenter states, however, that alarm time is counted. particulate matter emissions and the
this approach can dramatically 4. If corrective action is required, each alarm set point provides reasonable
understate variability when imputing a alarm shall be counted as a minimum of assurance that emissions will not
standard deviation for a source with 1 hour. exceed the standard.30
only three runs because the t-statistic is The final rule also establishes revised
substantially higher for 2 degrees of procedures for establishing the alarm set 30 Note that if your incinerator or boiler is
freedom than 8 degrees of freedom. point if you elect to use a particulate equipped with a fabric filter and you elect under
§ 63.1206(c)(8)(i) to use a particulate matter
We agree with the commenter. matter detector system in lieu of site- detection system in lieu of a bag leak detection
Moreover, using the actual number of system for compliance assurance, the ash feedrate
29 USEPA, ‘‘Technical Support Document for limit is waived. The ash feedrate limit is not waived
28 USEPA,
‘‘Draft Technical Support Document HWC MACT Standards, Volume III: Selection of if you use a bag leak detection system, however,
for HWC MACT Standards, Volume III: Selection of MACT Standards,’’ September 2005, Section 5.3. because the alarm level may not ensure compliance
MACT Standards,’’ March 2004, p. 5–4. See also Part Four, Section III.C of this preamble. with the emission standard when you follow the
Finally, you must submit an excessive although your eligibility for the health- for total chlorine. Under the final rule,
exceedance notification within 30 days based compliance alternatives is based if your permitting authority has not
of the date that the alarm set-point is on annual average HCl-equivalent approved your eligibility demonstration
exceeded more than 5 percent of the emissions, you must determine by the compliance date, and has not
time during any 6-month block period considering prescribed criteria whether issued a notice of intent to disapprove
of time, or within 30 days after the end your 1-hour HCl-equivalent emission your demonstration, you may
of the 6-month block period, whichever rate may exceed the national exposure nonetheless begin complying, on the
is earlier. The proposed rule would have standard (i.e., Hazard Index not compliance date, with the annual
required you to submit that notification exceeding 1.0 considering the maximum average HCl-equivalent emission rate
within 5 days of the end of the 6-month 1-hour average ambient concentration of limits you present in your eligibility
block period. hydrogen chloride and chlorine at an demonstration. In addition, if your
off-site receptor location32) and thus permitting authority issues a notice of
V. Health-Based Compliance
may exceed the 1-hour average HCl- intent to disapprove your eligibility
Alternative for Total Chlorine
equivalent emission rate limit absent an demonstration, the authority will
The final rule includes the following hourly rolling average limit on the identify the basis for that notice and
major changes to the proposed health- feedrate of chlorine. If the acute specify how much time you will have to
based compliance alternative for total exposure standard may be exceeded, submit additional information or to
chlorine: you must establish an hourly rolling comply with the MACT total chlorine
(1) You must use 1-hour Reference average chlorine feedrate limit as the 1- standards. The permitting authority may
Exposure Levels (aRELs) rather than 1- hour HCl-equivalent emission rate limit extend the compliance date of the total
hour acute exposure guideline levels divided by [1 ¥ system removal chlorine standards to allow you to make
(AEGL–1) as the acute health risk efficiency]. You establish the system changes to the design or operation of the
threshold metric when calculating 1- removal efficiency during the combustor or related systems as quickly
hour HCl-equivalent emission rates; comprehensive performance test. as practicable to enable you to achieve
(2) You must establish a long-term (4) When calculating HCl-equivalent compliance with the MACT total
average chlorine feedrate limit (i.e., 12 emission rates, rather than partitioning chlorine standards;
hour rolling average or an (up to) annual total chlorine emissions between
rolling average) as the annual average chlorine and HCl (i.e., the Cl2/HCl (7) We have revised the approach for
HCl-equivalent emission rate limit volumetric ratio) based on the determining chlorine emissions if you
divided by [1 ¥ system removal comprehensive performance test as feed bromine or sulfur during the
efficiency]. You establish the total proposed, you must establish the Cl2/ comprehensive performance test at
chlorine system removal efficiency HCl volumetric ratio used to calculate levels higher than those specified in
during the comprehensive performance the annual average HCl-equivalent § 63.1215(e)(3)(ii)(B). Under the final
test. The proposed rule would have emission rate based on the historical rule, you must use EPA Method 320/321
required you to establish the long-term average ratio from all regulatory or ASTM D 6735’01, or an equivalent
average chlorine feedrate limit as the compliance tests. You must establish method, to measure hydrogen chloride,
average of the test run averages of the the Cl2/HCl volumetric used to calculate and Method 26/26A, or an equivalent
comprehensive performance test.31 the 1-hour average HCl-equivalent method, to measure chlorine and
(3) At proposal, we requested emission rate as the highest of the hydrogen chloride. You must determine
comment on whether and how to historical ratios from all regulatory your chlorine emissions to be the higher
establish a short-term chlorine feedrate compliance tests. The rule allows you to of: (1) The value measured by Method
limit to ensure that the acute exposure exclude ratios from historical 26/26A, or an equivalent method; or (2)
Hazard Index of 1.0 is not exceeded. See compliance tests where the emission the value calculated by difference
69 FR at 21304. We conclude for the data may not be representative of the between the combined hydrogen
final rule that a 1-hour rolling average current Cl2/HCl ratio for reasons such as chloride and chlorine levels measured
feedrate limit may be needed for some changes to the design or operation of the by Method 26/26a, or an equivalent
situations (i.e., if chlorine feedrates can combustor or biases in measurement method, and the hydrogen chloride
vary substantially during the averaging methods. The rule also explicitly allows measurement from EPA Method 320/
period for the long-term feedrate limit the permitting authority to require 321 or ASTM D 6735–01, or an
and potentially result in an exceedance periodic emissions testing to obtain a equivalent method; and
of the 1-hour average HCl-equivalent representative average and maximum (8) The proposed rule would have
emission rate limit). Accordingly, ratio; required you to conduct a new
(5) The look-up table analysis has comprehensive performance test if you
concepts in the Agency’s guidance document on been refined by presenting annual
bag leak detection systems to establish the alarm
planned to make changes to the facility
level.
average and 1-hour HCl-equivalent that would lower the annual average
31 Note that, as a practical matter, most sources emission rate limits as a function of HCl-equivalent emission rate limit.
must establish the chlorine feedrate limit as the stack height, stack diameter, and Under the final rule, you would be
average of the test run average feedrate limit during distance to property line. In addition, required to conduct a performance test
the comprehensive performance test to demonstrate separate look-up tables are presented for
compliance with the semivolatile emission as a result of a planned change only for
standard. This is because chlorine feedrate is a flat terrain and simple elevated terrain; a change to the design, operation, or
compliance assurance parameter for the (6) The proposed rule required maintenance of the combustor that
semivolatile metal emission standard. That feedrate approval of the eligibility demonstration could affect the system removal
limit is based on a 12-hour rolling average. To before you could comply with the
ensure compliance with the annual average HCl- efficiency for total chlorine if the change
equivalent emission rate limit, however, that alternative health-based emission limits could reduce the system removal
feedrate limit cannot exceed the value calculated as efficiency, or if the change would
the annual average HCl-equivalent emission rate 32 Under the site-specific risk assessment
limit divided by [1 ¥ system removal efficiency], approach to demonstrate eligibility, you must
increase the system removal efficiency
where you demonstrate the total chlorine system consider locations where people reside and where and you elect to increase the feedrate
removal efficiency during the performance test. people congregate for work, school, or recreation. limits on total chlorine and chloride.
Part Four: What Are the Responses to Response: We agree with the emissions data for the source. Thus, the
Major Comments? commenter. In today’s rule, Dow Clean Air Act requires EPA to use the
I. Database Chemical Company’s boiler F–2820, most recently generated data available
located in Freeport, Texas, is to it and precludes the Agency from
A. Revisions to the EPA’s Hazardous reclassified in our data base as a boiler. using older, out-of-date performance
Waste Combustor Data Base This source is identified as unit number data.
Comment: Several commenters 2020 in our data base. EPA also received several comments
identify sources which have ceased B. Use of Data From Recently Upgraded stating that the language of section
operations as a hazardous waste Sources 112(d)(3)(A) of the Clean Air Act
combustor and should be removed from informs how the Agency should
Comment: Many commenters consider emissions data from sources
EPA’s data base. recommend that EPA remove from the
Response: We agree with commenters that conducted testing after that 1999
data base (or not consider for standards- rule was promulgated. One commenter
that data and information from sources setting purposes) emissions data from
no longer burning hazardous waste states that the only data which should
sources that upgraded their emissions not be used in calculating the MACT
should not be included in our controls to comply with the
hazardous waste combustor data base floors are from sources that are subject
promulgated emission standards of to lowest achievable emission rates
and should not be used to calculate the either the 1999 rule or the 2002 interim
MACT standards. We consider any (LAER). Thus, the commenter states,
standards. Several commenters also Congress considered the possibility of
source that has initiated RCRA closure state that any emissions data that were
procedures and activities as a source significant and recent upgrades, and
obtained or used to demonstrate
that is no longer burning hazardous concluded that EPA should use up-to-
compliance with the promulgated
waste. This data handling decision is date data to reflect source’s
standards of 1999 or 2002 should not be
consistent with the approach we used in performance, but must exclude certain
used for standard-setting purposes by
the 1999 final rule. See 64 FR at 52844. sources from the floor calculation if
the Agency. That is, EPA must evaluate
As we stated in that rule, ample their upgrades were of a specific degree
the source category as it existed at the
emissions data remain to support and were accomplished within a
beginning of the rule development
calculating the MACT standards specific period of time. Another
process and not after emissions controls
without using data from sources that no commenter states that Congress did not
are later added to comply with the 1999
longer burn hazardous waste. or 2002 standards. Several commenters intend to pile technology upon
As a result, we removed the following also state that EPA is only partly correct technology as confirmed by section
former hazardous waste combustors in claiming that the interim standards 112(d)(3)(A) that specifically excludes
from the data base: the Safety-Kleen are not MACT standards because the sources that implemented LAER from
incinerator in Clarence, New York, the interim standards were established and consideration when establishing section
Dow Chemical Company incinerators in considered to be MACT until the Court 112(d) standards. Thus, the commenter
Midland, Michigan, and LaPorte, Texas, issued its opinion in July 2001. Until states, considering data from sources
the two Holcim wet process cement that time, sources proceeded to upgrade that have upgraded violates both the
kilns in Holly Hill, South Carolina, the their facilities to achieve the standards language and intent of the Clean Air
Dow Chemical Company liquid fuel- promulgated in 1999. The rationale for Act. Another commenter states that,
fired boiler in Freeport, Texas, the these recommendations is threefold: (1) while Congress no doubt contemplated
Union Carbide liquid fuel-fired boilers Use of the data unfairly ignores the that EPA should use all available
in Hahnville, Louisiana, and Texas City, MACT-driven reductions already emissions information in setting initial
Texas, and six Dow Chemical Company achieved by some sources; (2) it is MACT standards, neither the statute nor
hydrochloric production furnaces in contrary to sound public policy to use the legislative history suggest that
Freeport, Texas. data from upgraded facilities to ‘‘ratchet follow-up MACT rulemakings require
We are retaining, however, Solite down’’ the MACT floors to a level more the use of data reflecting compliance
Corporation’s lightweight aggregate stringent because these sources would efforts with previous MACT standards
facility in Cascade, Virginia, in the data not have increased their level of or interim standards.
base. Even though the facility recently performance but for the legal obligation Response: As proposed, EPA
initiated RCRA closure procedures, this to comply with the standards; and (3) maintains its position on use of post-
data handling decision differs from EPA’s reliance on National Lime Ass’n 1999 emissions data. The statute
those listed in the preceding paragraph v. EPA, 233 F.3d 625, 640 (D.C. Cir. indicates that EPA is to base MACT
because Solite Corporation provided 2000), for the proposition that the floors on performance of sources ‘‘for
this new information in February 2005 motivation for a source’s performance is which the Administrator has emissions
while information on the other closures legally irrelevant in developing MACT information.’’ Section 112(d)(3)(A);
was reported or available to us in 2004. floor levels is misplaced because that CKRC, 255 F. 3d at 867. There can be
Because we cannot continually adjust case involved the initial MACT standard no dispute that post-1999 performance
our data base and still finalize this setting process, and not a subsequent data in EPA’s possession fits this
rulemaking by the court-ordered rule. description. We also reiterate that the
deadline, we stopped making revisions One commenter agrees with EPA’s motivation for the control reflected in
to the data base in late 2004. Additional proposed position and states that use of data available to us is irrelevant. See 69
facility changes after that date, like data from sources that have upgraded is FR at 21217–218. We further agree with
Solite Corporation’s Cascade facility not only appropriate, but also required those commenters who pointed out that
closure, simply could not be by the Clean Air Act. This commenter Congress was explicit when it wanted
incorporated. states that the actual performance of certain emissions information (i.e.,
Comment: One commenter identifies sources that have upgraded their sources operating pursuant to a LAER
a source in EPA’s data base that should emissions equipment—to meet the 1999 standard) excluded from consideration
be classified as a boiler instead of a standards or for any reason—is reflected in establishing floors. There is, of
hydrochloric acid production furnace. only by the most recently generated course, no such enumerated exception
for sources that have upgraded their indeterminate quality and are chlorine measurements below 20 ppmv
performance for other reasons. unachievable. Commenters assert that when establishing the standards.
We also do not agree with those Method 26A and its RCRA equivalent, Response: For the reasons discussed
commenters arguing (with respect to the SW 846 Method 0050, have a negative below, we corrected all total chlorine
standards for the Phase 1 sources bias at concentrations below 20 ppmv measurements in our data base for all
(incinerators, cement kilns, and when used on stacks controlled with source categories that were below 20
lightweight aggregate kilns)) in effect wet scrubbers. Commenters cite two ppmv to 20 ppmv to establish the total
that the present rulemaking involves recurring situations when this bias is chlorine floors. Moreover, to address
revision of an existing MACT standard. likely to occur: (1) hydrogen chloride run-to-run variability given that all runs
If this were indeed a revision of a MACT dissolving in condensed moisture in the for several data sets are now corrected
standard under section 112(d)(6), then sampling train; and (2) hydrogen to 20 ppmv, we impute a run standard
EPA would not redetermine floor levels. chloride reacting with alkaline deviation based on a regression analysis
See 70 FR at 20008 (April 15, 2005). compounds from the scrubber water that of run standard deviation versus total
However, EPA has not to date are collected on the filter ahead of the chlorine concentration for sources with
promulgated valid MACT floors or valid impingers. total chlorine measurements greater
MACT standards for these sources. The Commenters are particularly than 20 ppmv. This is the same
1999 standards do not reflect MACT, as concerned about the negative bias approach we used to impute variability
held by the CKRC court. The interim associated with stack gas containing from sources using fabric filters when
standards likewise do not reflect MACT, substantial water vapor. Commenters determining the particulate matter
but were designed to prevent a note that EPA found in a controlled MACT floors.
regulatory gap and were described as laboratory study by Steger 33 that the Effect of Moisture Vapor. Commenters
such from their inception. 67 FR at 7693 bias is between 17 and 29 percent at imply that stack gas with high levels of
(Feb. 13, 2002); see also Joint Motion of stack gas moisture content of 7 to 9 gas phase water vapor will inherently be
all Parties for Stay of Issuance of percent. This stack gas moisture is much problematic, particularly at emissions
Mandate in case no. 99–1457 (October less than the nominal 50% moisture less than 20 ppmv. There is no basis for
19, 2001), pp. 11–12 (‘‘The Parties contained in some hazardous waste claiming that water vapor, per se, causes
emphasize that the contemplated combustor stacks according to the a bias in SW–846 Method 0050 or its
interim rule is in the nature of a remedy. commenters. Commenters believe this is equivalent, Method 26A. Condensed
It would not respond to the Court’s why EPA’s Method 0050, which was moisture (i.e., water droplets), however,
mandate regarding the need to used to gather most of the data in the can cause a bias because it can dissolve
demonstrate that EPA’s methodology HWC MACT data base, states in Section hydrogen chloride in the sampling train
reasonably predicts the performance of 1.2 that ‘‘this method is not acceptable and prevent it from being captured in
the average of the best performing for demonstrating compliance with HCl the impingers if the sampling train is
twelve percent of sources (or best- emission standards less than 20 ppm.’’ not properly purged. Water droplets can
performing source). EPA intends to Moreover, commenters state that the potentially be present due to
address those issues in a subsequent procedures in Method 0050 to address entrainment from the wet scrubber,
rule, which will necessarily require a the negative bias caused by condensed condensation in cooler regions of the
longer time to develop, propose, and moisture were not followed for many stack along the stack walls, and
finalize.’’) EPA consequently believes RCRA compliance tests. The method entrainment from condensed moisture
that it is adopting in this rule the initial uses an optional cyclone to collect dripping down the stack wall across the
section 112(d) MACT standards for moisture droplets, and requires a 45 inlet duct opening.
hazardous waste burning incinerators, Although Method 0050 addresses the
minute purge of the cyclone and
cement kilns, and lightweight aggregate water droplet issue by use of a cyclone
sampling train to recover hydrogen
kilns, and that the floor levels for and 45 minute purge, the Steger paper
chloride from water collected by the (Ibid.) concludes that a 45 minute purge
existing sources are based, as provided cyclone and any condensed moisture in
in section 112(d)(3), on performance of is not adequate to evaporate all water
the train. The cyclone is not necessary collected by the cyclone in stacks with
those sources for which EPA has if the stack gas does not contain water
‘‘emissions information.’’ a total moisture content (vapor and
droplets. According to commenters, the condensed moisture) of 7 to 9%. At
However, we disagree with the
cyclone and subsequent purge were those moisture levels, Steger
comment that we must make exclusive
often not used in the presence of water documented the negative bias that
use of the most recent information from
droplets because a potential low bias commenters reference. Steger’s
hazardous waste combustion sources.
below 20 ppmv was irrelevant when recommendation was to increase the
There is no such restriction in section
demonstrating compliance with heat input to the sample train by
112(d)(3). EPA has exhaustively
emission standards on the order of 100 increasing the train and filter
examined all of the data in its
ppmv. There was no need for the extra temperature from 120C (248F) to 200C
possession for all source categories
complexity and expense of using a (392F). We agree that increasing the
covered by this rule, and determined
cyclone and train purge given the probe and filter temperature will
(and documented) which data are
purpose of the test. Although the data provide a better opportunity to
suitable for evaluating sources’
were acceptable for their intended evaporate any condensed moisture, but
performance.
purpose, commenters conclude that the another solution to the problem is to
C. Correction of Total Chlorine Data to data are not useful for establishing require that the post-test purge be run
Address Potential Bias in Stack standards below 20 ppmv. long enough to evaporate all condensed
Measurement Method For these reasons, commenters moisture. That is the approach used by
Comment: Several commenters state suggest that EPA not consider total Method 26A, which EPA promulgated
that EPA’s proposed total chlorine 33 Steger, J.L., et al, ‘‘Laboratory Evaluation of
after Method 0050, and which sources
standards of 1.5 ppm for existing Method 0050 for Hydrogen Chloride’’, Proc of 13th
must use to demonstrate compliance
incinerators and 0.18 ppm for new Annual Incineration Conference, Houston, TX, May with the final standards. Method 26A
incinerators are based on biased data of 1994. uses an extended purge time rather than
elevating the train temperature to Materials indicate that the filter used in Decision Unique to Hazardous Waste
address condensed moisture because the Method 0050 train (and the M26/ Combustors. We note that the rationale
that approach can be implemented by 26A trains) may adsorb/absorb hydrogen for our decision to correct total chlorine
the stack tester at the site without using chloride and cause a negative bias at data below 20 ppmv to account for the
nonstandard equipment. low emission levels. (See ASTM D6735– biases discussed above is unique to the
We attempted to quantify the level of 01, section 11.1.3 and ‘‘note 2’’ of hazardous waste combustor MACT rule.
condensed moisture in the Steger study section 14.2.3) This inherent affinity for Some sources apparently did not follow
and to compare it to the levels of hydrogen chloride can be satisfied by Method 0050 procedures to minimize
condensed moisture that may be present preconditioning the sampling train for the low bias caused by condensed
in hazardous waste combustor stack gas. one hour. None of the tests in our moisture for understandable reasons.
This would provide an indication if the database were preconditioned in such a Even if sources had followed Method
bias that Steger quantified with a 45 manner. 0050 procedures to minimize the bias
minute purge might also be applicable We are normally not concerned about (i.e., cyclone and 45 minute purge) there
to some hazardous waste combustors. this type of bias because we would still may have been a substantial bias
We conclude that this comparison expect the bias to apply to all sources because of insufficient purge time, as
would be problematic, however, equally (e.g., wet or dry gas) and for all Steger’s work may indicate. We note
because: (1) given the limited subsequent compliance tests. In other that the total chlorine stack test method
information available in the Steger words, we are ordinarily less concerned used by sources other than hazardous
paper, it is difficult to quantify the level if a standard is based on biased data, as waste combustors—Method 26A—
of condensed moisture in his gas long as the means by which the requires that the cyclone and sampling
samples; and (2) we cannot estimate the standard was developed and the means train be purged until all condensed
levels of condensed moisture in of compliance would experience moisture is evaporated. We believe it is
hazardous waste combustor stack gas identical bias. necessary to correct our data below 20
because, even though condensed However, we did correct the wet gas ppmv data because of issues associated
moisture may have been present during measurements below 20 ppmv to exclusively with Method 0050 and how
a test, method protocol is to report the address the potential low bias caused by it was used to demonstrate compliance
saturation moisture level only (i.e., the condensed moisture. This correction with these sources.
amount of water vapor present), and not would also correct for any potential bias Determining Variability for Data at 20
the total moisture content (i.e., both caused by the filter’s inherent affinity
ppmv. Correcting those total chlorine
condensed and vapor phase moisture). data below 20 ppmv to 20 ppmv brings
for hydrogen chloride. This results in a
We can conclude, however, that, if about a situation identical to the one we
data set that is partially corrected for
hazardous waste combustor stack gas confronted with nondetect data. See
were to contain the levels of condensed this issue—sources with wet stacks
Part Four, Section V.B. below. The
moisture present in the gas that Steger would be corrected for this potential
MACT pool of best performing source(s)
tested, the 45 minute purge required by bias while sources with dry stacks
for some data sets is now comprised of
Method 0050 would not be sufficient to would not be corrected. To address this
largely the same values. This has the
avoid a negative bias. We also conclude unacceptable mix of potentially biased
effect of understating the variability
that this is potentially a practical issue and unbiased data (i.e., dry gas data
associated with these data.
and not merely a theoretical concern biased due to affinity of filter for To address this concern, we took an
because, as commenters note, hazardous hydrogen chloride and wet gas data approach similar to the one we used to
waste combustors that use wet scrubbers corrected for condensed moisture and determine variability of PM emissions
are often saturated with water vapor that affinity of filter for hydrogen chloride), for sources equipped with a fabric filter.
will condense if the flue gas cools. we also correct total chlorine In that case, we performed a linear
Data from Wet Stacks When a Cyclone measurements from dry gas stacks (i.e., regression on the data, charting
Was Not Used. Commenters state that sources that do not use wet scrubbers). variability against emissions, and used
Method 0050 procedures for addressing Deposition of Alkaline Particulate on the variability that resulted from the
water droplets (adequate or not, as the Filter. Commenters are also linear regression analysis as the
discussed above) were not followed in concerned that hydrogen chloride may variability for the sources average
many cases because a low bias below 20 react with alkaline compounds from the emissions. In this case, most or all of the
ppmv was not relevant to demonstrating scrubber water droplets that are incinerator and liquid fuel boiler
compliance with standards on the order collected on the filter ahead of the sources in the MACT pool have average
of 100 ppmv. We do not know which impingers. Commenters suggest this emissions at or near 20 ppmv. We
data sets may be problematic because, as potential cause for a low bias at total therefore performed a linear regression
previously stated, the moisture chlorine levels below 20 ppmv is on the total chlorine data charting
concentration reported was often the another reason not to use measurements average test condition results above 20
saturation (vapor phase only) moisture below 20 ppmv to establish the ppmv against the variability associated
level and not the total (vapor and liquid) standards. with that test condition. The variability
moisture in the flue gas. We also have Although alkaline particulate associated with 20 ppmv was the
no documentation that a cyclone was deposition on the method filter causing variability we used for incinerator and
used—even in situations where the a negative bias is a much greater liquid fuel boiler data sets affected by
moisture content was documented to be concern for sources that have stack gas the 20 ppmv correction.
above the dew point. We therefore containing high levels of alkaline We also considered using the
conclude that all data below 20 ppmv particulate (e.g., cement kilns, sources statistical imputation approach we used
from sources controlled with a wet equipped with dry scrubbers), we agree for nondetect values. See discussion in
scrubber are suspect and should be with commenters that this may be of Section IV.B below. The statistical
corrected. concern for all sources equipped with imputation approach for correcting data
Potential Bias Due to Filter Affinity wet scrubbers. Our approach to correct below 20 ppmv without dampening
for Hydrogen Chloride. Studies by the all data below 20 ppmv addresses this variability would involve imputing a
American Society of Testing and concern. value between the reported value and 20
ppmv because the ‘‘true’’ value of the discussion regarding stack gas moisture, used to fire the kiln over the course of
biased data would lie in this interval. filter affinity for hydrogen chloride, and a day’s operation (or longer period).36 In
This approach would be problematic, alkaline compound reactions with comparison, the data used to support
however, given that many of the hydrogen chloride in the sampling train. the proposed floor level consisted of a
reported values were much lower than Sources must demonstrate much smaller dataset of approximately
20 ppmv; our statistical imputation compliance using a stack test method 50 test conditions representing a
approach would tend to overestimate that also addresses these issues. Sources snapshot of performance somewhere in
the run to run variability. Consequently, with wet stacks must use Method 26A the range of normal operations, with
we conclude that a regression analysis and follow those procedures regarding each test condition representing a
approach is more appropriate. A the use of a cyclone and the purging of relatively short period of time (e.g.,
regression analysis is particularly the system whenever condensed several hours).37 As discussed at
pertinent in this situation because: (1) moisture may be present in the proposal, we were concerned regarding
We consider data above 20 ppmv used sampling system. the representativeness of this smaller
to develop the regression to be Finally, all sources—those with either dataset. See 69 FR at 21251. In addition,
unbiased; and (2) all the corrected data wet or dry gas—should precondition the the commenter-submitted dataset allows
averages for which we are imputing a sampling train for one hour prior to us to better evaluate the only mercury
standard deviation from the regression beginning the test to satisfy the filter’s
control technique used by existing
curve are at or near 20 ppmv. Thus, any affinity for hydrogen chloride. The
hazardous waste burning cement kilns—
potential concern about downward permitting authority will ensure that
controlling the feed concentration of
extrapolation from the regression would sources precondition the sample train
mercury in the hazardous waste. The
be minimized. (under authority of § 63.1209(g)(2))
when they review and approve the commenters have demonstrated
We note that, although a regression convincingly that the mercury dataset
analysis is appropriate to estimate run- performance test plan.
used at proposal does not properly show
to-run variability for the corrected total D. Mercury Data for Cement Kilns the range of performance and variability
chlorine data, we could not use a linear in performance these cement kilns
Comment: Several commenters state
regression analysis to address variability actually experience, while the
that EPA’s data base of mercury
of nondetect values. To estimate a significantly more robust dataset
emissions data (and associated feed
standard deviation from a regression submitted by commenters does illustrate
concentrations of mercury in the
analysis, we would need to know the this variability. Thus, we conclude the
hazardous waste) are unrepresentative
test condition average emissions. This larger commenter-submitted dataset is
and unsuitable for use in determining
would not be feasible, however, because superior to EPA’s smaller testing
MACT standards for cement kilns.
some or all of the run measurements for dataset.
These comments are supported by an
a test condition are nondetect. In extensive amount of data submitted by
addition, we are concerned that a We note that our MACT floor analysis
the cement manufacturing industry of the commenter-submitted dataset to
regression analysis would not accurately including three years of data
estimate the standard deviation at low determine which sources are the best
documenting day-to-day levels of performers and to identify a mercury
emission levels because we would have mercury in hazardous waste fuels fired
to extrapolate the regression downward standard for cement kilns is discussed
to all 14 hazardous waste burning in the background document.38
to levels where we have few measured cement kilns.35 The commenters
data (i.e., data other than nondetect). Additional discussion of issues related
recommend that EPA use the to the mercury standard for cement
Moreover, the statistical imputation commenter-submitted data as the basis
approach is more suitable for handling kilns is found in Part Four, Section VI.B
for assessing cement kilns’ performance of the preamble.
nondetects because the approach for control of mercury because it is the
calculates the run-to-run variability by most complete and representative data
taking into account the percent available to EPA.
36 Mercury is a volatile compound at the typical
nondetect for the emissions for each operating temperatures of the air pollution control
Response: We agree that the devices used by cement kilns (i.e., baghouses and
run.34 A regression approach would be commenter-submitted mercury data are electrostatic precipitators). Most of the mercury
difficult to apply particularly in the case more representative than those we used exits the cement kiln system as volatile stack
of test conditions containing partial at proposal. First, these data represent a
emissions, with a smaller fraction partitioning to
nondetects or a mix of detect and the clinker product or cement kiln dust. Thus, in
significantly larger and more general, there is a proportional relationship
nondetect values. Given these concerns comprehensive dataset compared to the between the mercury concentration in the
with using a regression analysis to one used to support the proposed hazardous waste and stack emissions of mercury
estimate the standard deviation of test mercury standard. The commenter-
(i.e., as the mercury concentration in hazardous
conditions with runs that have one or waste increases (assuming mercury concentrations
submitted data document the day-to-day in other inputs such as raw materials and fossil
more nondetect (or partial nondetect) levels of mercury in hazardous waste fuels (coal) and other factors remain constant),
measurements, we conclude that the fired to all cement kilns for a three year emissions of mercury will correspondingly
statistical imputation approach best period covering 1999 to 2001. In total,
increase).
37 EPA’s dataset for mercury for cement kilns is
assures that the calculated floor levels approximately 20,000 measurements of not like the RCRA compliance test emission data for
account for run-to-run emissions the concentration of mercury in other HAPs where each source designs the
variability. hazardous waste are included in the compliance test such that the operating limits it
Compliance with the Standards. The dataset. When considered in whole, establishes account for the variability it expects to
final standards are based on data that encounter during its normal operations (e.g., semi-
these data describe the performance and low volatile metals). This is not necessarily true
were corrected to address specific issues (and variability thereof) of all cement for mercury for cement kilns as shown in our
concerning these data. See the above kilns for the three year period because analysis of our mercury dataset at proposal. See 69
FR at 21251.
34 For multi-constituent HAP (e.g. SVM) the
each measurement represents the 38 USEPA, ‘‘Technical Support Document for
emissions for a run could be comprised of fully mercury concentration in the burn tank HWC MACT Standards, Volume III: Selection of
detected values for some HAP and detection limits MACT Standards,’’ Sections 7.5.3 and 11.0,
for other HAP that were nondetect. 35 See docket item OAR–2004–0022–0049. September 2005.
E. Mercury Data for Lightweight Comment: One commenter opposes expenditure of significant resources,
Aggregate Kilns the use of the commenter-submitted both in time and level of effort, to
Comment: One commenter, an owner mercury data because EPA would be collect several more months of data, we
and operator of seven of the nine uncritically accepting a limited and conclude that obtaining additional
operating lightweight aggregate kilns, select data set from a commenter with mercury measurements is unnecessary
states that the mercury dataset used by a direct interest in the outcome of its because the available eight months of
EPA at proposal is a limited and use. Instead, the commenter suggests data—including over 310 individual
unrepresentative snapshot of EPA use its section 114 authority to measurements—represent a significant
performance of their seven kilns. To obtain all data that are available, not just amount of data that we judge to be
support their position that the snapshot the data selected by that commenter. adequately reflective of the source’s
emissions data are unrepresentative, the Response: We disagree that we performance and variability in
commenter submitted eight months of uncritically accepted the commenter- performance.
data documenting levels of mercury in submitted mercury data. The reason the
commenter submitted data collected F. Incinerator Database
hazardous waste fuels fired to their
lightweight aggregate kilns.39 between October 2003 and June 2004 is
Comment: Commenters state that
Response: We agree with the that the facility was, prior to October
many of the top performers (e.g., 3011,
commenter that their mercury data 2003, in the process of upgrading its on-
3015, 3022, 349) dilute emission
submission is more representative than site analysis equipment. One outcome of
concentrations in the stack by burning
those used at proposal. As discussed in this laboratory upgrade was its
natural gas to initiate reactive waste
a notice for public comment sent capability to detect mercury in
(e.g., explosives, inorganic hydrides) or
directly to certain commenters,40 the hazardous waste at lower
to decontaminate inert material.
commenter-submitted dataset concentrations. Prior to the upgrade, the
Commenters do not believe these units
documents the day-to-day levels of facility’s on-site laboratory was capable
of detecting mercury in the hazardous should be considered ‘‘representative’’
mercury in hazardous waste fuels fired of the overall incinerator source
to Solite Corporation’s Arvonia kilns waste at a concentration of
approximately 2 ppmw, which is a level category and should not be used to
between October 2003 and June 2004. establish standards for incinerators
The dataset consists of over 310 such that the vast majority of
measurements would neither be combusting primarily organic wastes.
measurements of the concentration in
mercury in hazardous waste. Each detected nor useful for identifying best Response: Source 3022 has closed and
measurement represents the mercury performers and their level of has been removed from the database.
concentration of the burn tank used to performance.42 The June 4, 2004 cutoff Emission data from source #3015 (ICI
fire the kiln over the course of a day’s date represents a practicable date that explosives) has been excluded for
operation (or longer period). In measurements could still be purposes of calculating the particulate
comparison, the data used to support incorporated into the commenter’s matter floor because the test report
the proposed floor level consisted of a public comments to the proposed rule, indicates this source was primarily
smaller dataset of 15 test conditions. which were submitted on July 6, 2004. feeding scrap metal, which we conclude
The nature of the mercury data Finally, the commenter provided all to be an atypical waste stream from a
submitted by the commenter is the same waste fuel measurements during this particulate matter compliance
as we received for the cement kiln period and states reliably that no perspective.44
category discussed in the preceding measurements made during this period The sources identified by the
section. For similar reasons, we accept were selectively excluded.43 commenter are among the best
the more comprehensive commenter- We also reject the commenter’s
performing sources in two instances.
suggestion that we use our authority
submitted dataset as one that better Source 3011 is the second ranked best
under section 114 of the Clean Air Act
shows the range of performance and performer for the particulate matter
to obtain additional hazardous waste
variability in performance for these standard. This source is among the best
mercury concentration data from the
lightweight aggregate kilns. One notable performers for particulate matter
facility. There is no obligation for us to
difference, however, is that the because it uses a state-of-the art
gather more performance data, given
commenter submitted mercury data baghouse that is equipped with Teflon
that the statute indicates that we are to
only for its company (representing coated bags. There is no evidence to
base floor levels on performance of
seven of nine lightweight aggregate suggest that this source was diluting its
sources ‘‘for which the Administrator
kilns). Thus, we received no data particulate matter emissions. We
has emissions information.’’ Section
documenting day-to-day levels of the 112(d)(3)(A); CKRC, 255 F. 3d at 867. In acknowledge that we do not have ash
concentration of mercury in hazardous addition, given our concerns about the feed data for the test conditions that
waste fuels for the other two lightweight usefulness of measurements with high were used in the particulate matter
aggregate kilns owned by a different detection limits discussed above, the standard analysis. However, this source
company. For these two lightweight collection of additional data prior to the had the third and fourth highest metal
aggregate kilns, we continue to use laboratory upgrade would not be feed control levels among all the sources
available data available in our productive. When balanced against the used in the MACT analysis for the
database.41 semivolatile and low volatile metal
submitted data, see the document ‘‘Technical
39 See docket items OAR–2004–0022–0270 and Support Document for HWC MACT Standards, 44 We did not have ash feed data for source 3015.
OAR–2004–0022–0333. Volume III: Selection of MACT Standards,’’ Section We acknowledge that ash feed control levels do not
40 See docket item OAR–2004–0022–0370. 7.5.3 and 12.0, September 2005. significantly affect particulate matter emissions
41 Unlike that is available for the commenter’s 42 A mercury concentration of 2 ppmw in the from sources equipped with baghouses. However,
kilns, we note that we have compliance test hazardous waste corresponds to a stack in this instance, the particulate matter emissions
emissions data, which is designed to maximize concentration of approximately 200 µg/dscm, which from this source may not be representative because
operating parameters (e.g., HAP feedrates) that is well above the interim standard of 120 µg/dscm this source may not have been feeding any
affect emissions, for the other two kilns. For for mercury. appreciable levels of ash given that scrap metal
additional discussion on how these data were 43 See also docket items OAR–2004–0022–0233 feeds generally would not contribute to the ash
analyzed in conjunction with the commenter- and OAR–2004–0022–0367. loading into the baghouse.
standards.45 We therefore conclude that are also not influenced by the size of the subcategorizing commercial incinerators
it is appropriate to include this source combustor.47 and on-site incinerators.48
in the MACT analysis that determines Emission limitations are similarly We have identified source 341 as the
the relevant best performers for best performing source for particulate
normalized to remove the influence of
particulate matter. matter and low volatile metals. It is the
combustion unit size by expressing the
Source 349 is the eighth ranked (out single best performing source for these
standards as emission concentration standards because it is equipped with a
of 11) best performer for the particulate
limits rather than as mass emission rate state-of-the-art baghouse.49 This source,
matter standard. We acknowledge that
limits. See section III.D. This is which simultaneously feeds hazardous
the ash feed level for this source is
illustrated in the following example. and nonhazardous wastes, conducted
lower than most incinerators equipped
with baghouses. However, particulate Assume there are two cement kilns side several emission tests that reflected
matter emissions from sources equipped by side with similar designs, the only different modes of operation. The
with baghouses are not significantly difference being one is twice the size of amount of nonhazardous waste that was
affected by the ash inlet loading to the the other, producing twice as much processed in the combustion unit varied
baghouse.46 This is further supported by clinker. They both have identical types across test conditions. We could not
the fact that this source is ranked eighth of air pollution control systems (the ascertain the exact amount of hazardous
among the best performers. We larger source is equipped with a larger waste processed in the test condition
conclude source 349 is a best performer control device that is appropriately that was used in the MACT analysis for
not because of its relatively low ash feed sized to accommodate the larger low volatile metals because the test
level, but rather because it is equipped volumetric gas flow rates and achieves report stated the wastes that were
with a well designed and operated the same control efficiency as the processed were a mixture of hazardous
baghouse. It is therefore appropriate to smaller control device). If we were to and nonhazardous wastes, although we
include this source in the MACT assess performance based on HAP mass estimate that at least 26% of the waste
analysis. emission rates (e.g., pounds per hour), processed was nonhazardous.50 We note
Comment: Commenters state that the smaller source would be the better that we are aware of several other
source 341 should not be considered in performer because its mass emission incinerators that processed
the MACT analysis because it is a small rates would be half of the mass emission nonhazardous waste at levels greater
laboratory waste burner that processes rate of the larger source, even though than 26 percent during their emission
only 900 lbs/hr of waste. Commenters they both are achieving the same back- tests. We therefore do not believe this to
claim that more than 80 percent of the end control efficiency. Emission be atypical operation that warrants
waste profile is non-hazardous waste. concentrations, on the other hand, are subcategoriztion.
Response: We approached this Moreover, the fact that this source
calculated by dividing the HAP mass
was feeding nonhazardous wastes does
comment by asking if it would be emission rate (e.g., pounds per hour) by not result in atypically low hazardous
appropriate to create a separate the volumetric gas flowrate (e.g., cubic waste low volatile metal feed control
subcategory for source 341. We feet per hour). In the above example, levels, as evidenced by the relative feed
conclude it is not necessary to both sources would have identical HAP control ranking for this source of
subcategorize hazardous waste emission concentrations (the larger thirteenth among the 26 sources
incinerators based on the size of source has twice the mass emission rate, assessed in the MACT analysis. It also
combustion units. This is because the but twice the volumetric gas flow rate), has the highest normalized hazardous
ranking factors used to identify the accurately reflecting their identical waste feed control level among the best
relevant best performing sources are control efficiency. Emission performing sources, and has the fifth
normalized in order to remove the concentrations normalize the size of best low volatile metal system removal
influence that combustion unit size each source by accounting for efficiency among those same 26 sources.
would otherwise have when identifying volumetric gas flowate, which is We repeat that this source is being
best performing sources. See part 4 directly tied to the amount of raw identified as the best performing source
section III.D below. Air pollution material each source processes (and primarily because it is equipped with a
control system types (a ranking factor subsequently the amount of product that highly efficient baghouse, not because it
for particulate matter) are generally is produced). This is a reason we point is feeding low levels of HAP metals
sized to match the corresponding out that normalization eliminates the attributable to its hazardous waste.
volumetric gas flow rate in order to need to create subcategories based on Furthermore, this source is not the
achieve a given control efficiency. The lowest emitting source in the database.
unit size. See part four section III.D.
size of the combustor therefore does not There are two sources with similar, but
influence a source’s ability to achieve a Further, it would be difficult to
slightly lower low volatile metal
given control efficiency. System determine an appropriate minimum size
compliance test emissions (one
removal efficiency and hazardous waste cutoff in which to base such a commercial incinerator and one onsite,
feed control MTECs (ranking factors subcategorization determination. Such a non-commercial incinerator). This
used by the SRE/Feed methodology as subcategorization scheme could also provides further evidence that the
described in part 4 section III.B below) yield nonsensical floor results, as was
the case when we assessed 48 See USEPA, ‘‘Technical Support Document for
45 We note that feed control levels are normalized the HWC MACT Standards, Volume III: Selection of
based on each source’s gas flowrate. The feed 47 System removal efficiency is a measure of the MACT Standards’’, September 2005, Section 4.3.2
control levels used to assess performance are amount of the pollutant that is removed from the for further discussion.
therefore appropriate indicators that directly flue combustion gas prior to being emitted and 49 See USEPA, ‘‘Final Technical Support
address whether emissions of these pollutants are likewise is not influenced by the size of the Document for the HWC MACT Standards, Volume
in fact being diluted by the combustion of natural combustor because back-end control systems are I: Description of Source Categories’’, September
gas. sized to achieve a given performance level. 2005, Section 3.2.1, for further discussion.
46 See USEPA, ‘‘Technical Support Document for Hazardous waste feed control levels are normalized 50 See USEPA, ‘‘Final Technical Support
the HWC MACT Standards, Vol I: Description of to remove the influence of combustor size by Document for the HWC MACT Standards, Volume
Source Categories,’’ September 2005, Section 3.2.2, dividing each source’s mass feed rate by its I: Description of Source Categories’’, September
for further discussion. volumetric gas flowrate. 2005, Section 2.1 for further discussion.
emissions from this source indicative of additional variability lead us to conclude that the emission
appropriately represent emissions of a above and beyond the run-to-run data are inappropriate for purposes of
relevant best performing source. variability and should be assessed if the calculating MACT standards. However,
Regarding the particulate matter data are deemed usable at all. since proposal, we have determined that
standard, source 341 does not have Response: We conclude both of these this source did not measure cadmium
atypically low ash feed rates as sources are in fact unique sources that emissions during this emissions test. As
compared to other sources equipped should be assessed as individual a result, we conclude the semivolatile
with baghouses. Out of the nine best sources for purposes of the MACT metal emissions data from this source
performing particulate matter sources analysis. Although these sources are of should not be used in the MACT
for which we have ash feed information, similar design, we do not believe they standard calculation for semivolatile
this source ranks fourth (a ranking of are identical, in part because: (1) The metals because the data do not represent
one is indicative of the lowest ash feed facility itself conducted separate the source’s combined emissions of lead
rate). Nonetheless, as previously emission tests for the two units (rather and cadmium.
discussed, particulate matter emissions than trying to avail itself of the ‘data in
from sources equipped with baghouses lieu’ option, which could save it the II. Affected Sources
are not significantly affected by the ash expense of a second compliance test, the A. Area Source Boilers and
inlet loading to the baghouse. We note obvious inference being that the source Hydrochloric Acid Production Furnaces
that particulate matter emissions from or regulatory official regards the two
units as different); and (2) discussions Comment: Five commenters state that
the second and third best performing
with facility representatives indicated the area sources subject to the proposed
source are not significantly different
these units are similar, but not rule are negligible contributors to
from this source, providing further
identical.53 As a result, it would be 112(c)(6) HAP emissions and should not
evidence that this source is
inappropriate to assess emissions be subject to major source standards for
representative of the range of emissions
variability by combining the emissions 112(c)(6) HAP. Commenters note that
exhibited by other well designed and
of these two sources into one test requiring compliance with MACT for
operating incinerators equipped with
condition given they are not identical 112(c)(6) HAP and RCRA for other toxic
baghouses.51
Comment: Commenters state that units. pollutants is more complicated and
sources 3018 and 3019 are identified as Comment: Commenters state that burdensome for sources than complying
best performers for mercury emissions emissions data from source 327 should only with RCRA. Although an area
for incinerators. After evaluating the not be used to calculate dioxin/furan source can choose to become regulated
trial burn plans for these sources, the and mercury floors because they claim as a major source in order to reduce
commenter believes the data should not the carbon injection system did not some RCRA requirements, they would
be used to calculate the MACT floor appear to function properly during the become subject to more onerous
because the spiking rate for mercury test. emissions limits under Subpart EEE and
was extremely low for a compliance Response: We agree with the the other MACT requirements.
commenters. We have determined that One of these commenters states that
test. The ranking for feedrate is therefore
this source encountered problems with subjecting an area source to major
unrepresentative. The commenter
its carbon injection system during the source standards under 112(c)(6) sends
suggests that these test results should be
emissions test from which the data were a negative message to industry that EPA
characterized as ‘‘normal’’.
Response: We have verified that the obtained and subsequently used in does not value emissions reduction and/
emission tests performed for sources EPA’s proposed MACT analysis. We or chemical substitution, or other
3018 and 3019 reflect the upper range have also verified that this source did methods used by area sources to achieve
of mercury emissions that are not to be not establish operating parameter limits that status. EPA is no longer providing
exceeded by these sources, and that for the carbon injection system as a any incentive for sources to take such
their spiked mercury feed rates were result of this test.54 We therefore have difficult yet environmentally beneficial
back-calculated from a risk assessment. excluded this mercury and dioxin data steps to become an area source.
We therefore conclude that we properly from the MACT analysis, and have Imposing Title V permitting
characterized these emissions as instead used emissions data from an requirements on an entire facility that
compliance test emissions data because older test condition to represent this operates as an area source of hazardous
they reflect the emissions resulting from source’s emissions. air pollutants (HAPs) will impose an
the upper bound of hazardous waste Comment: Commenters state that the unfair and undue burden on the facility.
mercury feedrates from these sources.52 emissions data from source 3006 were Another of these commenters states
Consequently, these data are properly based on a miniburn to determine how that section 112(c)(6) requires in
included with the other data used to close the unit was to achieving the pertinent part that EPA list categories
calculate floor standards for mercury for interim MACT standards. The and subcategories of sources assuring
incinerators. commenter questions whether these that sources accounting for not less than
Comment: Commenters state the trial data should be used for purposes of 90% of the aggregate emissions of each
burn plan for sources 3018 and 3019 calculating MACT standards. pollutant (specified in 112(c)(6)) are
describes these units to be of similar Response: The fact that a source subject to standards under Section
design. Thus the difference in results conducts a voluntary emissions test 112(d)(2) or (d)(4). In 1998, EPA
between these two similar sources is (e.g., a miniburn) to determine how published a notice identifying the list of
close it is operating to upcoming source categories accounting for the
51 Source 341 particulate matter emissions, after emission standards does not necessarily section 112(c)(6) HAP emissions and to
accounting for variability, equated to 0.0015 gr/dscf. be regulated under section 112(d) to
The second and third ranked particulate matter 53 Also see February 11, 2005 memo to docket
meet the 90% requirement. (63 FR
sources emissions, considering variability, equated titled ‘‘October 20 Conference Call with Squibb 17838) At the time, EPA acknowledged
to 0.0018 and 0.0023 gr/dscf, respectively. Manufacturing regarding Source # 3018 and 3019’’.
52 See February 11, 2005 memo to docket titled 54 See July 15, 2005 memo to docket titled that MACT standards for a number of
‘‘October 20 Conference Call with Squibb ‘‘Telephone Conversation with Utah DEQ Regarding the source categories had not yet been
Manufacturing regarding Source # 3018 and 3019’’. 2001 Clean Harbor Emission Test.’’ promulgated, and stated that when the
regulations for each of those categories standards). Congress singled out the products, or destroy these HAP if they
are developed, EPA will analyze the pollutants in section 112(c)(6) as being are present in the wastes being
data specific to those sources and of ‘‘’specific concern’’’ not just because combusted.55 See discussion in Part
determine, under Section 112(d), in of their toxicity but because of their Four, Sections V.A and V.B of this
what manner requirements will be propensity to cause substantial harm to preamble.
established. EPA also stated that: human health and the environment via The HWC NESHAP also applies to
‘‘Some area categories may be negligible indirect exposure pathways (i.e., from hazardous waste-burning boilers and
contributors to the 90% goal, and as such the air through other media, such as hydrochloric acid production furnaces.
pose unwarranted burdens for subjecting to water, soil, food uptake, etc.). In particular, for these boilers and
standards. These trivial source categories will Furthermore, these pollutants have furnaces, this rule addresses emissions
be removed from the listing as they are exhibited special potential to of dioxin/furan, mercury, POM and
evaluated since they will not contribute bioaccumulate, causing pervasive PCBs either through specific numeric
significantly to the 90% goal.’’ (63 FR 17841) environmental harm in biota and, standards for the identified HAP, or
The commenter believes the ‘‘two or ultimately, human health risks. through standards for surrogate
fewer’’ area source boilers identified by Section 112(c)(6) of the CAA requires pollutants which control emissions of
EPA in the present rulemaking are EPA to list categories and subcategories the identified HAP.
‘‘negligible contributors’’ to the 90% of sources of seven specified pollutants We estimate that approximately 620
goal and therefore, should not be to assure that sources accounting for not pounds of mercury are emitted annually
required to adopt the same MACT less than 90 percent of the aggregate in aggregate from hazardous waste
emission limitations and requirements emissions of each such pollutant are burning boilers in the United States.56
as major sources of the 112(c)(6) subject to standards under CAA section Also, we estimate that hazardous waste
pollutants. The commenter believes 112(d)(2) or 112(d)(4). In 1998, EPA burning boilers and hydrochloric acid
EPA’s decision to subject area source issued the list of source categories production furnaces emit in aggregate
boilers and hydrochloric acid pursuant to section 112(c)(6), and that approximately 2.3 and 0.2 grams TEQ
production furnaces is incorrect, list is published at 63 Fed. Reg. 17838, per year of dioxin/furan, respectively.
unsupported by the administrative 17849, Table 2 (April 10, 1998). Controlling emissions of these HAP
record, and therefore arbitrary and In the 1998 listing, EPA identified the from area sources consequently reduces
capricious. following three subcategories of the emissions of these HAP through
One commenter states that, if EPA HWC source category that emit one or application of MACT standards. We
regulates area sources, it should more of the seven section 112(c)(6) note that only major source boilers and
significantly reduce the administrative pollutants: (1) Hazardous waste hydrochloric acid furnaces are subject to
burden for area sources by: exempting incinerators—(emit mercury, dioxin, the full suite of subpart EEE emission
them from Title V provisions for furans, polycyclic organic matter (POM) standards.57 Section 112(c)(3) of the
Subpart EEE requirements; exempting and polychlorinated biphenyls (PCBs)); CAA requires us to subject area sources
them from compliance with the General (2) Portland cement manufacture: to the full suite of standards applicable
Provisions of 63 Subpart A; limiting hazardous waste kilns—(emit mercury, to major sources if we find ‘‘a threat of
them to a one-time comprehensive dioxin, furans, and POM); and (3) adverse effects to human health or the
performance test; or limiting other lightweight aggregate kilns: hazardous environment’’ that warrants such action.
applicable requirements. waste kilns—(emit dioxin, furans, and We cannot make this finding for area
Response: We continue to believe that mercury). These three subcategories are source boilers and halogen acid
boiler and hydrochloric acid furnace all subject to today’s rule, which is production furnaces. 69 FR at 21212.
area sources warrant regulation under issued pursuant to CAA section Consequently, as proposed, area sources
the major source MACT standards for 112(d)(2). As explained below, the HWC in these categories would be subject
mercury, dioxin/furan, carbon NESHAP effectively controls emissions only to the MACT standards for
monoxide/hydrocarbons, and of the identified section 112(c)(6) mercury, dioxin/furan, and polycyclic
destruction and removal efficiency pollutants from the identified
pursuant to section 112(c)(6). subcategories. Accordingly, EPA 55 Courts have repeatedly upheld EPA’s authority
As discussed at proposal (69 FR at considers the sources in these three under CAA section 112(d) to use a surrogate to
21212), section 112(c)(6) of the CAA regulate hazardous pollutants if it is reasonable to
subcategories as being ‘‘subject to do so. See, e.g., National Lime, 233 F. 3d at 637
requires EPA to list and promulgate standards’’ for purposes of section (holding that EPA properly used particulate matter
section 112(d)(2) or (d)(4) standards 112(c)(6). as a surrogate for HAP metals).
(i.e., standards reflecting MACT) for Specifically, with regard to hazardous 56 See USEPA ‘‘Technical Support Document for
categories and subcategories of sources waste-burning incinerators, cement HWC MACT Standards, Volume V: Emission
Estimates and Engineering Costs,’’ September, 2005,
emitting seven specific pollutants. Five kilns, and lightweight aggregate kilns, Section 3.
of those listed pollutants are emitted by EPA is adopting in this final rule MACT 57 We note that as a practical matter, however, the
boilers and hydrochloric acid standards for mercury and dioxins/ same MACT standards apply to both major and area
production furnaces: mercury, 2,3,7,8- furans. EPA has already adopted MACT source HCl production furnaces. This is because
tetrachlorodibenzofuran, 2,3,7,8- standards for control of POM and PCBs major sources are subject to the following
standards: CO/HC, DRE, and total chlorine. Because
tetrachlorodibenzo-p-dioxin, polycyclic emitted by these sources in the 1999 the CO/HC and DRE standards are surrogates to
organic matter, and polychlorinated rule, which standards were not control dioxin/furan, and the total chlorine
biphenyls. reopened or reconsidered in this standard is a surrogate to control metal HAP, area
As discussed below, EPA must assure rulemaking. These standards are the sources are subject to the same standards that
address dioxin/furan, polycyclic organic matter,
that source categories accounting for not CO/HC standards, which in polychlorinated biphenyls, and mercury. There is
less than 90 percent of the aggregated combination with the Destruction an enforcement difference between the
emissions of each enumerated pollutant Removal Efficiency (DRE) requirement, requirements, however. For area sources, an
are subject to MACT standards (and of assure that these sources operate exceedance of the total chlorine standard (or failure
to ensure that compliance is maintained) relates to
course is not prohibited from requiring continuously under good combustion control of mercury only while for a major source,
more than 90 percent of aggregated conditions which inhibit formation of the same failure relates to control of mercury, other
emissions to be controlled by MACT POM and PCBs as combustion by- metal HAP, and HCl and chlorine.
organic matter and polychlorinated monoxide/hydrocarbons and regulating LRWE units under Subpart
biphenyls (through the surrogate destruction and removal efficiency EEE is unnecessary and inconsistent
standards for carbon monoxide/ (DRE)—should also be minimal for area with RCRA subtitle C and more
hydrocarbons and destruction and source liquid fuel boilers. The dioxin/ importantly, appears to be controlling
removal efficiency) to control the HAP furan standard applicable to the 90% of LRWE units for control’s sake.
enumerated in section 112(c)(6). RCRA liquid fuel boilers with wet or no air Commenters also state that EPA has
standards under Part 266, Subpart H for pollution control equipment is not properly addressed the requirements
particulate matter, metals other than compliance with the carbon monoxide/ of CAA section 112(n)(7) regarding the
mercury, and hydrogen chloride and hydrocarbon standard and the DRE inconsistency between the requirements
chlorine gas would continue to apply to standard. Liquid fuel boilers already for Low Risk Waste Exempt (LRWE)
these area sources unless an area source comply with these same standards units under RCRA and those of Subpart
elects to comply with the major source under RCRA. The surrogate standards to EEE. The purported purpose of section
standards in lieu of the RCRA standards. control other polycyclic organic matter 112(n)(7) is to allow EPA to avoid
See § 266.100(b)(3) and the revisions to are also the carbon monoxide/ imposing additional emission
§§ 270.22 and 270.66. hydrocarbon and DRE standards. limitations on a source category
Commenters refer to the ‘‘two or Finally, we note that the DRE subcategory when such limitations
fewer’’ potential area source boilers we requirement under Subpart EEE is less would be unnecessary and duplicative.
identified at proposal as ‘‘negligible burdensome than the DRE requirement In addition, commenters state that the
contributors’’ and, therefore, conclude under RCRA. Under Subpart EEE, a costs associated with this MACT are
that these area sources should not be source needs to conduct a one-time only much more than improved feed control
subject to major source standards for DRE test, provided that design and or better back-end control. This
emission of these HAPs. Commenters operation does not change in a manner proposed rule also requires substantial
did not quantify the amount of than could adversely affect DRE. Under dollar investment in improved data
emissions from area sources, and did RCRA, the DRE test must be conducted acquisition, computer controls and
not even identify how many area each time the RCRA permit is renewed. recordkeeping systems, performance
sources are at issue. We do not know The incremental compliance burden testing, training, development of plans,
how many boilers and hydrochloric acid associated with the other Subpart EEE and other regulatory requirements.
furnaces are area sources. We major source requirements, such as the Response: Boilers and hydrochloric
apparently underestimated the number operations and maintenance plan, the acid production furnaces that currently
given that four companies commented startup, shutdown, and malfunction qualify for the RCRA § 266.109 low risk
on the proposed rule saying that area plan, operator training, and the waste exemption are not exempt from
sources should not be subject to major automatic waste feed cutoff system Subpart EEE under the final rule.
source standards for mercury, dioxin/ should also be minimal for liquid fuel The Administrator does not have the
furan, PCBs, and polycyclic organic boilers without an emission control authority under CAA section 112(d) to
matter, and one of those companies device. In addition, most of the exempt sources that comply with RCRA
indicates it operates multiple area requirements are either identical to or § 266.109. Indeed, there is no necessary
sources. Consequently, we continue to very similar to requirements under connection between the two provisions,
believe that area sources in these RCRA with which these area sources are since one is technology-based and the
categories may have the potential to already complying.58 other is risk-based. CAA section
emit more than negligible levels of these 112(d)(2) requires the Administrator to
HAP. B. Boilers Eligible for the RCRA Low establish technology-based emission
We also note that the major source Risk Waste Exemption standards, standards that require the
standards are tailored to minimize the Comment: Several commenters state maximum degree of reduction in
compliance burden for sources that emit that EPA should exempt those boilers emissions that is deemed achievable.
low levels of HAP. Commenters raise that qualify as Low Risk Waste Although section 112(d)(4) gives the
concerns about applying the major Exemption (LRWE) burners under the Administrator the authority to establish
source standards for HAP enumerated in RCRA Boiler and Industrial Furnace health-based emission standards in lieu
section 112(c)(6) to liquid fuel boiler Rule at § 266.109 from the MACT of the MACT standards for pollutants
area sources. The emission standard particulate matter and destruction and for which a health threshold has been
compliance burden for liquid fuel removal efficiency (DRE) standards established, we cannot use that
boilers that have the potential to emit because EPA has not: (1) Made a authority to develop health-based
only low levels of mercury, dioxin/ demonstration that the data used to standards for sources that comply with
furan, and polycyclic organic matter is provide the exemption to low risk RCRA § 266.109 because those sources
minimal. For example, sources that emit burners under RCRA is no longer valid; emit HAP for which a health threshold
low levels of mercury because their or (2) established in the affirmative that has not been established.
feedstreams have low levels of mercury regulating these units will provide any The final rule complies fully with
can elect to comply with the mercury benefit to human, health and the CAA section 112(n)(7) by coordinating
emission standard by documenting that environment. Commenters believe that applicability of the RCRA and CAA
the mercury in feedstreams will not requirements and precluding dual
exceed the standard assuming zero 58 RCRA, 40 CFR Part 264 requirements that are requirements. For example, RCRA
removal by emission control equipment. similar to MACT requirements include: the general requirements that are duplicative of
We note that 75% of the liquid fuel inspection requirements and personnel training MACT requirements will be removed
requirements of Subpart B; the preparedness and
boilers in our data base, and the two prevention requirements of Subpart C, including from the RCRA operating permit when
boilers cited by commenters, do not design and operation of facility, testing and the permitting authority issues a
have emission control devices. maintenance of equipment, and access to certification of compliance after the
The compliance burden for the major communications or alarm system; the contingency source submits a Notification of
plan and emergency procedures requirements of
source standards for dioxin/furan and Subpart D; and the operating requirements and Compliance.
for the surrogates to control other monitoring and inspection requirements of Subpart We also note that the MACT
polycyclic organic matter—carbon O. standards are tailored to impose
minimal burden on sources that have may emit any air pollutant. A mobile results in high particulate matter
low emissions of HAP. The particulate incinerator at a remediation site meets emissions. They operate only for a short
matter emission standard and associated this definition. duration at a site (usually less than 6
testing can be waived (similar to the Comment: One commenter states that months), and have no flexibility with
§ 266.109 exemption) for boilers that a subcategory with different standards regard to their waste feed.
elect to document that emissions of total must be created for mobile incinerators, Response: We recognize that there is
metal HAP do not exceed the limits or the standards for incinerators must be variability between various sources’
provided by § 63.1206(b)(14). calculated using actual emissions data with regard to size, capacity, operating
Hydrochloric acid production furnaces from mobile units. temperatures etc., and so we applied a
are not subject to a particulate matter Response: EPA did not have any statistical test to assess the need of
emission standard. emissions data from mobile incinerators subcategorization, as has been discussed
The compliance burden with the in the database for the proposed rule. above. The emissions data provided by
destruction and removal efficiency That data base was developed over the commenter also indicate the source
(DRE) standard is also minimal given many years with ample opportunity for can achieve the final standards. The soil
that it is a one-time test, provided that public comment. We developed a data entrained in desorber off-gases of mobile
the source does not change its design or base for incinerators to support the 1996 incinerators has a relatively large
operation in a manner that would proposed rule (61 FR 17358) and particle size, and is very easy to capture
adversely affect DRE. In addition, the noticed that data base for public with conventional particulate control
compliance burden for sources with low comment on January 7, 1997 (64 FR systems (such as a fabric filter) used by
levels of metals in their feedstreams is 52828). We updated that data base in the incinerators.
minimal. Sources can document July 2002, and noticed the revised data Comment: Since mobile incinerators
compliance with the metals emission base for public comment (67 FR 44452). are relocated from site to site, the new
standards by assuming all metals in the We used that revised data base to source standard should not apply based
feed are emitted (i.e., by assuming zero support the proposed rule. We did not on the erection date of the mobile unit.
system removal efficiency). Under this receive comments providing data for Response: We are not changing the
procedure, boilers burning relatively mobile incinerators as a result of either applicability of a new or reconstructed
clean wastes are not required to conduct public notice. source designation in this rulemaking.
a performance test to document One commenter on the proposed rule The relocation issue is addressed in the
compliance with the metals emission provided a summary of emissions data definition of ‘‘construction’’ in 40 CFR
standards. from one test at a mobile incinerator. Section 63.2, which states:
Further, we note that the MACT The commenter suggested that the data ‘‘Construction does not include the
standard to control organic HAP support its view that its mobile removal of all equipment comprising an
emissions other than dioxin/furan is the incinerator is unique and that EPA affected source from an existing location
same as the RCRA standard— should consider subcategorizing and the reinstallation of such equipment
demonstrating good combustion incinerators according to mobile at a new location * * *’’ (emphasis
incinerators versus other incinerators. added). Therefore, the relocation of an
conditions by complying with a carbon
We analyzed these data and conclude existing Subpart EEE affected source,
monoxide standard of 100 ppmv.
Finally, we note that the ancillary that the final standards are readily such as a mobile incinerator, would not
requirements under MACT (e.g., achievable by this source. Moreover, as result in that mobile incinerator
personnel training; operating and explained elsewhere, EPA’s approach to becoming a ‘‘new’’ source. Keep in mind
assess the need for subcategorization is also that the relocation exemption only
maintenance plan; startup, shutdown,
to apply a statistical test to determine applies to affected sources. If a mobile
and malfunction plan) should not pose
whether the emissions data are incinerator is relocated from an R&D
substantially higher costs than similar
statistically different from the remaining facility (where the unit is not an affected
requirements under RCRA. See response
group. Given that owners and operators source per Table 1 to Section 63.1200)
to comment in Section A above. To the
of mobile incinerators have not to a location where the mobile
extent that compliance costs increase,
provided emissions data prior to incinerator would become an affected
we have accounted for those costs in our
proposal, and that the commenter source, the relocation exemption within
estimates of the cost of the final rule.59
provides summarized data for only one the definition of ‘‘construction’’ would
C. Mobile Incinerators mobile incinerator (which also indicate not apply and the mobile incinerator
Comment: A mobile incinerator used that the source can achieve the emission would be a ‘‘new’’ source. Also, with
as a directly-fired thermal desorption standards in the final rule); we are not regard to leased sources, the owner/
unit at a Superfund remediation site compelled to gather additional operator of the facility is responsible for
should not be an affected source under information, particularly given our time all affected sources operating at his/her
this rule. constraints to promulgate the final rule facility regardless of whether the
Response: EPA is not determining or under a court-ordered deadline. sources are owned or leased. The owner
Comment: In support of or operator should obtain from the
changing the applicability of any
subcategorizing mobile incinerators, leasing company all relevant
hazardous waste burning unit under
commenters state that mobile thermal information pertaining to the affected
today’s rule. A combustion unit that
treatment systems are substantially source in order to be able to
treats hazardous waste and meets the
different from hazardous waste demonstrate that the affected source is
definition of incinerator at 40 CFR
incinerators. They are much smaller in operating in compliance with the
260.10 is an affected source under this
size, firing capacity rate, refractory appropriate standards.
rule. 40 CFR part 63 also defines a
lining, and operating temperatures.
source as any building, structure, III. Floor Approaches
Most of them treat contaminated soil, so
facility, or installation which emits or
have very high particulate feedrate In this section we discuss comments
59 USEPA ‘‘Technical Support Document for loading with high ash content, rapid addressing methodologies used in this
HWC MACT Standards, Volume V: Emission kiln rotation rate, and counter-current rule for determining MACT floors. We
Estimates and Engineering Costs,’’ September, 2005. flow design like cement kilns. This address comments relating both to
general, overarching issues and to the Response: Our response explains our in practice’ is interpreted to mean
specific methodologies used in the rule. approach to estimating best performing ‘achieved under the worst foreseeable
Our most important point is that the sources’ variability and addresses the circumstances’; see also National Lime
methodologies EPA selected reasonably following issues: (1) Considering the Ass’n v. EPA, 627 F. 2d 416, 431 n. 46
estimate the performance of the best variability in each source’s performance (D.C. Cir. 1980) (where a statute requires
performing sources by best accounting is necessary to identify the best that a standard be ‘achievable,’ it must
for these sources’ total variability. performing sources and their level of be achievable under ‘‘the most adverse
performance; (2) EPA reasonably circumstances which can reasonably be
A. Variability considered variability in ranking expected to recur’’);
1. Authority To Consider Emissions sources to identify the best performers The court has further indicated that
Variability and in considering the range of best EPA is to account for variability in
performing sources’ performance over assessing sources’ performance for
Comment: Many commenters concur
time to identify an emission level that purposes of establishing floors, and
with our approach to account for stated that this assessment may require
the average of those sources can
emissions variability while several EPA to make reasonable estimates of
achieve; (3) considering variability at
commenters believe that our approach performance of best performing sources.
the 99th percentile level is reasonable;
does not adequately account for CKRC, 255 F. 3d at 865–66; Mossville
(4) considering intersource variability
emissions variability. See discussions Environmental Action Now v. EPA, 370
by pooling run-to-run variability is
on separate topics below. One F. 3d 1232, 1242 (D.C. Cir.
appropriate; and (5) compliance test
commenter, however, states that use of 2004)(maximum daily variability must
conditions do not fully reflect all of best
variability factors (however derived) is performing sources’ performance be accounted for when establishing
inherently unlawful and arbitrary and variability. MACT floors).60 Indeed, EPA’s error in
capricious. The commenter notes that, a. Variability Must Be Considered. CKRC was not in estimating best
because floors for existing sources must Variability in each source’s performance performing sources’ variability, but in
reflect the ‘‘average’’ emission level must be considered at the outset in using an unreasonable means of doing
achieved by the relevant best identifying the best performing sources. so. CKRC, 255 F. 3d at 866; Mossville,
performing sources, they cannot reflect This is simply another way of saying 370 F. 3d at 1241.
any worse levels of performance from that best performers are those that Since the emission standards in
the best performers. Indeed, the perform best over time (i.e. day-in, day- today’s rule must be met at all times, the
argument is that the Clean Air Act out), a reasonable approach. This standards need to account for
already accounts for variability by approach not only reasonably reflects performance variability that could occur
requiring EPA to base existing source the statutory language, but also furthers on any single day of these sources’
floors on the average emission level the ultimate objective of section 112 operation (assuming proper design and
achieved by the best performing which is to reduce risk from exposure operation). See Mossville, 370 F. 3d at
sources. to HAP. Since most of the risk from 1242 (upholding MACT floor because it
The commenter continues by stating exposure to emissions from this source was established at a level that took into
that EPA has added variability factors category is associated with chronic account sources’ long term performance,
both to each individual source’s exposure to HAP (see Part 1 section VI not just performance on individual
performance and to the collective above), assessing a source’s performance days). Moreover, since EPA’s database
performance of the alleged best over time by accounting for variability consists of single data points (because
performers, in each case purporting to is reasonable and appropriate. there are no continuous emission
find an emission level that the For similar reasons, variability must monitors for HAPs in stack emissions),
individual or group would meet ninety- be considered in ascertaining these EPA must of necessity estimate long-
nine times out of 100 future emission sources’ level of performance. Floors for term performance, including daily
tests. Thus, EPA ignores sources’ existing sources must reflect ‘‘the maximum performance, from this
measured performance in favor of the average emission limitation achieved by limited set of short term data.
theoretical worst performance that the best performing 12 percent’’ of b. EPA Reasonably Considered
might ever be expected from them. By sources, and for new sources, must Variability in Ranking Sources to
looking to the best performers’ worst reflect ‘‘the emission control that is Identify the Best Performers and in
performance rather than their average achieved in practice by the best Considering the Range of Best
performance, EPA would set weaker controlled source.’’ Section 112 (d) (3). Performing Sources’ Performance Over
floors than the Clean Air Act allows. EPA construes these requirements as Time to Identify an Emission Level that
In addition, the commenter notes that meaning achievable over time, since the Average of Those Sources Can
EPA’s approach to account for sources are required to achieve the Achieve. (1) Selecting Best Performing
emissions variability is arbitrary and standards at all times. This Sources. Each of the floor
capricious because EPA never explains interpretation has strong support in the methodologies used in the rule
why it chose the 99th percentile for its case law. See Sierra Club v. EPA, 167 F. considers various factors in ranking
variability adjustments rather than some 3d 658, 665 (D.C. Cir. 1999), stating that which sources are the best performing.
other percentile. ‘‘EPA would be justified in setting the For each methodology, we therefore
Finally, the commenter notes that floors at a level that is a reasonable consider the quantifiable variability of
EPA appears to indicate that its estimate of the performance of the ‘best
variability analysis would either be controlled similar unit’ under the worst 60 See also Chemical Manufacturers Ass’n v. EPA,
applied to variation between sources or reasonably foreseeable circumstances. It 870 F. 2d 177, 228 (5th Cir. 1989) (‘‘The same plant
using the same treatment method to remove the
would affect EPA’s statistical analysis of is reasonable to suppose that if an same toxic does not always achieve the same result.
the variation between sources. The emissions standard is as stringent as ‘the Tests conducted one day may show a different
commenter states that any attempt by emissions control that is achieved in concentration of the same toxic than are shown by
the same test the next day. This variability may be
EPA to add a variability factor to adjust practice’ by a particular unit, then that due to the inherent inaccuracy of analytical testing,
for intersource variability is unlawful particular unit will not violate the (i.e. ‘analytical variability,’ or to routine
and arbitrary and capricious. standard. This only results if ‘achieved fluctuations in a plant’s treatment performance.’’)
the ranking factors in determining control device 63, and the variability Test-to-test variability must be
which are the best performing sources. caused by other uncontrollable factors accounted for in other instances as well,
69 FR at 21230–31. Specifically, we such as using a different stack testing however. It follows that if the
assess run-to-run variability (normally crew or different analytical laboratory, performance of most efficient fabric
the only type of variability which we and by different weather conditions filters varies over time relative to
can quantify) of the factors used under (e.g., ambient moisture and temperature) particulate matter emissions, then so
each methodology to rank best that may affect measurements. does their performance relative to the
performers. Where SRE/Feed is the We are able to quantify run-to-run non-mercury metal HAP emissions. We
ranking methodology, we thus assess variability. We do so by applying a 99th also believe that particulate matter
run-to-run variability of hazardous percentile modified upper prediction emissions variability from sources
waste HAP feedrate and of system limit to the averaged emissions of the equipped with back-end controls other
removal efficiency. Where ranking is best performing sources. Id. at 21233 than fabric filters also exists, and is
based on sources’ emissions (the straight and Technical Support Document furthermore likely to be higher than
emissions methodology), we assess the Volume III section 7.2. The modified what was calculated for fabric filters
run-to-run variability of emission levels. upper prediction limit accounts for run- because there are more uncertainties
Where we use the air pollution control to-run variability of the best performers associated with the correlations between
device methodology for ranking, we by pooling their run variance (i.e., operating parameter limits and control
assess the run-to-run variability of within-test condition variability).64 See efficiency for these devices.65 Again, it
emissions of the lowest-emitting sources Chemical Manufacturer’s Ass’n v EPA, clearly follows that if the performance
(as we do for straight emissions) using 870 F. 2d 177, 228 (5th Cir. 1989) of these other control devices varies
the best air pollution control devices. (upholding use of a variability factor relative to particulate matter emissions
For hydrochloric acid production derived, as here, by pooling the (perhaps even more than what has
furnaces, we assess the run-to-run performance variability of the best already been quantified for fabric
variability of total chlorine system performing plants). Using this approach, filters), then so does their performance
removal efficiency. Id.61 we ensure that the average of the best relative to the non-mercury metal HAP
To account for run-to-run variability performing sources will be able to emissions.
in these ranking factors, we rank sources achieve the floor in 99 of 100 future Although we cannot quantify this test-
by the 99th percentile upper prediction performance tests, assuming these best to-test variability, we can document its
limit (UPL99). The UPL99 is an estimate performing sources could replicate their existence and its significance. We
of the value that the source would performance when attempting to operate conducted two parallel analyses
achieve in 99 of 100 future tests if it under identical conditions to those used examining all situations where we had
could replicate the operating conditions for the compliance test establishing the multiple test conditions for the sources
of the compliance test. Id. at 21231. source as best performing. As just noted, ranked as best performing performing
(2). Assessing the Best Performers’ we call this value the modified UPL 99. (examining separate pools for best
Level of Performance Over Time. Once performing sources under both the
The only instance in which we are
we identify the best performing sources, straight emissions and SRE/feed ranking
able to quantify test-to-test variability
we need to consider their emissions methodologies). These analyses showed
(as noted above, the other significant
variability to establish a floor level that that these sources’ emissions do in fact
component of total operating variability)
the average of the best performing vary over time, sometimes significantly.
is for fabric filters (baghouses) when
sources can achieve day-in, day-out. In many instances sources had poorer
used to control emissions of particulate
There are two components of emissions system removal efficiencies and higher
matter. The modified UPL 99 in these
variability that must be considered: run- emission levels than those in the
instances reflects not only run-to-run
to-run variability and test-to-test compliance test used to identify the
variability, but test-to-test variability as
variability. Run-to-run emissions source as best performing. We further
well. That total variability is expressed
variability encompasses variability in projected that in many instances these
by the Universal Variability Factor
individual runs comprising the best performing sources would not
which is derived from analyzing long-
compliance tests, and includes achieve their own UPL 99, the
term variability in particulate matter
uncertainties in correlation of statistically determined prediction limit
emissions for best performing sources
monitoring parameters and emissions, which captures 99 out of 100 future
across all of the source categories
and imprecision of stack test methods three-run test averages for the source, if
sources that are equipped with fabric
and laboratory analyses. See 69 FR at they were to operate at the poorer
filters. 69 FR at 21233. See also the
21232.62 Test-to-test emissions system removal efficiency of its earlier
discussion below in Section III.A.2.
variability is the variability that exists test and used the federate of its later
between multiple compliance tests (best-performing) compliance test. This
63 There are myriad factors that affect
conducted at different times and performance of an emissions control device. These
is significant because the UPL 99
includes the variability in control factors change over time, including during the reflects all of a source’s run-to-run
device collection efficiency caused by maintenance cycle of the device, such that it is
testing at different points in the virtually impossible to conduct future compliance 65 For example, sources equipped with
tests under conditions that replicate the electrostatic precipitators generally establish
maintenance cycle of the emission performance of the control device. See USEPA, multiple operating limits to best assure compliance
‘‘Technical Support Document for HWC MACT with the emission standard (feed control limits,
61 These ranking methodologies are discussed Standards, Volume III: Selection of MACT power input limits, etc.). There is not an exact
later in this section of the preamble, and in USEPA, Standards,’’ September 2005, Section 5.3. correlation between emission levels and operating
‘‘Technical Support Document for HWC MACT 64 We note that the Agency used a statistical levels because there are several factors that can
Standards, Volume III: Selection of MACT approach when proposing the NESHAP for Electric affect the control efficiency of these air pollution
Standards,’’ September 2005, Section 7. Utility Steam Generating Units. See memo from control systems, such as variations in inlet loads,
62 Analytic variability exists, and normally must William Maxwell, EPA, to Utility MACT Project power inputs, spark rates, humidity, as well as
be accounted for in establishing technology-based Files, entitled, ‘‘Analysis of variability in particle resistivity. See USEPA, ‘‘Technical Support
standards based on performance of the best- determining MACT floor for coal-fired electric Document for the HWC MACT Standards, Volume
performing plants. Chemical Manufacturers Ass’n utility steam generating units,’’ dated Nov. 26, 2003, III: Selection of MACT Standards,’’ September 2005,
v. EPA, 870 F. 2d at 230. Docket A–92–55. Sections 16 and 17.
variability. Failure to meet the UPL 99 tests conducted under the same intersource run-to-run variability of the
thus shows both that further variability conditions as its performance test.. The best performing sources (which is all
exists, namely test-to-test variability, commenter thus sharply our floor calculation does, by
and that it is a significant component of mischaracterizes a 99% confidence level considering the pooled run-to-run
total variability. We obtained similar as the worst performance of a best variability of the best performing
results when we projected best performing source.: the level in fact sources). Section 112(d)(3) states that
performing sources’ performance based assumes identical operating conditions MACT floors are to reflect the ‘‘average
on each of these sources’ overall system as those of the performance test. emission limitation achieved’’ but does
removal efficiency obtained by pooling EPA routinely establishes not-to- not specify any single method of
the removal efficiencies of all of its exceed standards (daily maximum ascertaining an average. Considering the
tests. In many instances, moreover, values which cannot be exceeded in any average run-to-run variability among the
these projected levels exceeded floor compliance test) using the 99% group of best performing sources is well
levels calculated by using the straight confidence level. National Wildlife within the language of the provision
emissions approach, which ranks best Federation v. EPA, 286 F. 3d 554, 572 (and was upheld in CMA, as noted
performers as those with the lowest (D.C. Cir. 2002).69 At a confidence level above; see 870 F. 2d at 228). The
emission levels. This point is discussed of only 97% for example, the average of commenter’s further argument that
further in Section III.B below. EPA’s the best performing sources could ‘average’ can only mean average of
analysis is set out in detail in chapters expect to achieve the floor in only 97 of emission levels achieved in
16 and 17 of Volume III of the Technical 100 future performance tests. performance tests is inconsistent with
Support Document.66 We note that the choice of a the holding in Mossville, 370 F. 3d at
EPA’s conclusion is that total confidence level is not a choice 1242, that EPA must account for
variability includes both run-to-run and regarding the stringency of the emission variability in developing MACT floors
test-to-test variability, and that both standard. Although the numerical value and that individual performance tests
must be accounted for in determining of the floor increases with the do not by themselves account for such
which are the best performing sources confidence level selected it only appears variability.
and what are their levels of performance to become less stringent. If EPA selected We believe that it is reasonable and
over time. As explained in the following a lower confidence interval, we would necessary to account for intersource
Sections B and C, EPA has accordingly necessarily adjust the standard variability of the best performing
adopted floor methodologies which downward due to the expectation that a sources by taking the pooled average of
account for this total variability either source would not be expected to achieve the best performing sources’ run-to-run
quantitatively or qualitatively. The the standard for uncontrollable reasons variability. This is an aspect of
approach advocated by the commenter a larger per cent of the time. We would identifying the average performance of
simply ignores that variability exists. then have to account in some manner those sources. Emissions data for each
Since this approach is contrary to both for this inability to achieve the standard. best performing source are random in
fact and law, EPA is not adopting it. See Weyerhaeuser v. Costle, 590 F. 2d nature, and this random nature is
c. Quantifying Run-to-Run Variability 1011, 1056–57 (D.C. Cir. 1978) (also characterized by a stochastic
at the 99th Percentile Level Is upholding standards established at 99 distribution. The stochastic distribution
Reasonable. We selected the 99% % confidence level). The governing is defined by its central tendency
prediction limit to ensure a reasonable issue is what level of confidence should (average value) and the amount of
level ‘‘ namely the 99th percentile—of the average of the best performing dispersion from the point of central
achievability for sources designed and sources, and similarly designed and tendency (variance or standard
operated to achieve emission levels operated sources, have of passing the deviation). Consequently, to define the
equal to or better than the average of the performance test demonstrating performance of the average of the best
best performing sources.67 Because of compliance with the standard. We performing sources, we must consider
the randomness of the emission values, believe that the 99% confidence level is the average of the average emissions for
there is an associated probability of the a confidence level within the range of the best performing sources as well as
average of the best performing sources, values we could have reasonably the pooled variance for those sources.
and similarly designed and operated selected.70 Hence, we must consider intersource
sources, not passing the performance d. Considering Intersource Variability variability to identify the floor—the
test conducted under the same by Pooling Run-to-Run Variability is average performance of the best
conditions.68 At a 99% confidence Appropriate. The commenter believes performing sources.
level, the average of the best performing that any attempt by EPA to add a The commenter further states that
sources could expect to achieve the variability factor to adjust for EPA’s attempt to adjust for intersource
floor in 99 of 100 future performance intersource variability is unlawful and variability is unlawful, arbitrary, and
arbitrary and capricious. We see no capricious. EPA set floors at the 99th
66 We explain in those sections that these
statutory prohibition in considering percentile worst emission level that it
projections assume that system removal efficiencies believed any source within the group of
are constant across differing HAP federates and that
the sources’ historical (poorer) system removal 69 The opinion notes further that percentiles for best performers could achieve,
efficiencies were not the primary result of operating standards expressed as long-term average typically according to the commenter. The 99th
at poorer ‘‘controllable’’ conditions relative to the use a lower confidence level (usually 95 %c) due percentile worst performance that could
most recent test condition. These are reasonable to the opportunity to lower the overall distribution be expected from a source within the
assumptions, as explained in section 17. 3 of with multiple measurements. 286 F. 3d at 573. The
Volume III of the Technical Support Document, standards in this rule are necessarily daily
best performers is, simply put, not the
although these assumptions also create a measure maximum standards because continuous emissions average performance of the sources in
of uncertainty regarding the emissions projections. monitors for HAP do not exist or have not been that group, according to the commenter.
67 Note, again, that the variability we quantify by demonstrated on all types of Subpart EEE sources. The commenter misunderstands our
these analyses is within-test condition variability 70 See also Chemical Mfrs. Ass’n v. EPA, 870 F.
approach to calculate the floor—the
only. We cannot quantify test-to-test variability and 2d at 229 (99th percentile daily variability factor is
thus cannot quantify sources’ total variability. reasonable); 227 (‘‘the choice of statistical methods
floor is not the 99th percentile highest
68 See Volume III of the Technical Support is committed to the sound discretion of the emission level that any best performing
Document, Section 7.2 . Administrator’’). source could achieve. The floor for
existing sources is calculated as the 99th operations of the average of the best performing sources) to operate under
percentile modified upper prediction performing sources. We disagree. their full range of normal operations,
limit of the average of the best The compliance test is designed to and thus impermissibly would fail to
performing sources. It represents the mirror the outer end of the controllable account for their total variability.
average of the best performing sources’ variability occurring in normal As discussed throughout this
emissions levels plus the pooled within- operations. These controllable factors preamble, emissions variability—run-to-
test condition variance of the best include the amount of HAP fed to a run and test-to-test variability—is real
performing sources. The floor for source in hazardous waste, and and must be accounted for if a best
existing sources is not the highest 99th controllable operating parameters on performing source is to be able to
percentile upper prediction limit for any pollution control equipment (such as replicate the emissions achieved during
best performing source as the power input to ESPs, or pressure drop the initial compliance test. We
commenter states. across wet scrubbers, factors which are consequently conclude that we must
e. Why isn’t Total Variability Already reflected in the parametric operating account for variability in establishing
Accounted for by Compliance Test limits written into the source’s permit floor levels, and that merely considering
Conditions? and which are based on the results of the average of compliance test data fails
Comment: One commenter states that the compliance testing). However, this to do so. We have therefore quantified
EPA’s use of variability factors along is plainly not all of the variability a run-to-run variability using standard
with worst-case data is unlawful and source experiences. Other components statistical methodologies, and accounted
arbitrary and capricious. EPA has stated of run-to-run variability, including for test-to-test variability either by
that its use of worst case ‘‘compliance’’ variability relating to measuring (both quantifying it (in the case of fabric filter
data accounts for variability. EPA stack measurements and measurements particulate matter removal performance)
admits that compliance data reflect at analytic laboratories) are not or accounting for it qualitatively (in the
special worst case conditions created reflected, for example. Nor is test-to-test case of the SRE/feed ranking
artificially for the purpose of obtaining variability reflected, notably the point in methodology).
the maintenance cycle that testing is Comment: The commenter notes that
lenient permit limits, according to the
conducted and the variability associated if EPA believes that single performance
commenter. EPA provides no reason
with those inherently differing test test results do not accurately capture
whatsoever to believe that a source
conditions even though the source source’s variability, the solution is to
would continue to operate under such
attempts to replicate the test conditions gather more data, not to avoid using a
conditions even one percent of the time.
(e.g., measurement variability straight emissions methodology. EPA
Thus, the commenter concludes, by
attributable to use of a different test cannot use this as an excuse for basing
applying a 99 percent variability factor
crew and analytical laboratory and floor levels on a chosen technology
to compliance test data, EPA ensures
different weather conditions such as rather than the performance of the best
that the adjusted data do not accurately performing sources.
ambient temperature and moisture).
reflect the performance of any source. Response: There is no obligation for
Other changes that occur over time are
Accordingly, EPA’s use of a variability EPA to gather more performance data,
due to a wide variety of factors related
factor is unlawful. since the statute indicates that EPA is to
to process operation, fossil fuels, raw
The commenter also states that, to materials, air pollution control base floor levels on performance of
increase compliance data with the equipment operation and design, and sources ‘‘for which the Administrator
reality that sources will not be operating weather. Sampling and analysis has emissions information.’’ Section
under the worst case conditions except variations can also occur from test to 112(d)(3)(A); CKRC, 255 F. 3d at 867
during permit setting tests, the Agency’s test (above and beyond those accounted (upholding EPA’s decision to use the
use of a variability factor with for when assessing within-test compliance test data in its possession in
compliance data is arbitrary and variability) due to differences in establishing MACT standards). Indeed,
capricious. emissions testing equipment, sampling the already-tight statutory deadlines for
Response: All but two standards in crews, weather, and analytical issuing MACT standards would be even
the final rule are based on compliance laboratories or laboratory technicians. less feasible if EPA took further time in
test data—when sources maximized Thus, there is some need for a data gathering. EPA notes further that
operating parameters that affect standard to account for this additional because particulate matter continuous
emissions to reflect variability of those variability, and not simply expect for a emission monitors are not widely used,
parameters and to achieve emissions at single performance test to account for it. even further data gathering would be
the upper end of the range of normal The analyses in Sections 16 and 17 of limited to snapshot, single performance
operations. Use of these data is Volume III of the Technical Support test results, still leaving the problem of
appropriate both because they are data Document confirm these points. estimating variability from a limited
in EPA’s possession for purposes of Moreover, the best performing sources data set.71 See also Sierra Club v. EPA,
section 112(d)(3) and because these data (and the average of the best performers) 167 F. 3d at 662 (‘‘EPA typically has
help account for best performing must be able to replicate the compliance wide latitude in determining the extent
sources’ operating variability. CKRC, test if they are to be able to continue of data-gathering necessary to solve a
255 F. 3d at 867. operating under their full range of problem’’).
The main thrust of the comment is normal operations. It is thus no answer Thus, EPA has no choice but to assess
that total variability is accounted for by to say that the best performing sources best performers and their level of
the conditions of the performance test, could operate under a more restricted performance on the basis of limited
so that making further adjustments to set of conditions in subsequent amounts of data per source. As
allow for additional variability is performance tests and still demonstrate explained in the previous response to
improper. The commenter believes that compliance, so that there is no need to
71 Performance tests take an average of 5–8 days
the floor should be calculated simply as assure that results of initial performance
to conduct, and cost approximately from
the average emissions of the best tests can be replicated. To do so would $200,000—$500,000 per test. The commenter’s off-
performing sources and that this floor no longer allow the best performing hand suggestion appears to have ignored these
would encompass the range of sources (and thus the average of the best realities.
comments, EPA has selected a across source categories.75 Put another (i.e., run-to-run variability) for the other
methodology that reasonably do so. way, fabric filter particulate matter floors, which is only one component of
EPA notes further that it has carefully reduction is relatively independent of total variability. This is one reason we
examined those instances where there inlet loadings to the fabric filter. 69 FR use the SRE/Feed approach wherever
are multiple test conditions (usually 21233. This is confirmed by the fact that possible rather than a straight emissions
compliance tests conducted at different there are no operating parameters that approach to rank the best performing
times) for sources ranked as best can be readily changed to increase sources to calculate the floor—the SRE/
performing. This analysis confirms emissions from fabric filters, id., so Feed ranking approach derives floors
EPA’s engineering judgment that total control efficiencies reflected in test that better estimate the levels of best
variability is not fully encompassed in conditions from different source types performing sources’ performance. See
the single test condition results used to will still accurately reflect fabric filter also discussion in Part Four, Section
identify these sources as best control efficiency. III.A, and the discussion below
performing, and that without taking this documenting that test-to-test variability
3. Test-to-Test Variability
additional variability into account, best can be substantial.
performing sources would be unable to Comment: Several commenters state Comment: One commenter states that
achieve the floor standard reflecting that EPA seems to have ignored test-to- EPA should use the universal variability
their own performance in those single test variability resulting from changes factor (UVF) that accounts for total
test conditions.72 that occur over time such as: normal variability for particulate matter
and natural changes in a wide variety of controlled with a fabric filter to derive
2. Universal Variability Factor for factors related to process operation, a correction factor to account for the
Particulate Emissions Controlled with a fuels, raw materials, air pollution missing test-to-test variability
Fabric Filter control equipment operation and component of variability for
Comment: One commenter states that, design, and differences in emissions semivolatile metals and low volatile
in calculating the universal variability testing equipment, sampling crews, metals. The commenter then suggests
factor (UVF) to account for total weather, analytical laboratories or that the within-test variability for
variability—test-to-test variability and laboratory technicians. All these sources semivolatile and low volatile metals be
within-test variability—for sources of variation are expected in that they are adjusted upward by the correction factor
controlling particulate matter with a typical and are not aberrations. In to correct for the missing test-to-test
fabric filter, it appears that EPA addition, there are unexpected sources variability component.
considered the variability of sources of variability that occur in real-world The commenter focused on cement
that are not best performing sources. If operations, which also must be kilns and compared the total variability
so, EPA has contravened the law. accommodated according to imputed from the UVF for the three
commenters. cement kiln facilities used to establish
The commenter also states that EPA’s
Commenters state that using the UVF to the within-test variability
attempt to use a variability factor
compliance test data and assessing (i.e., run variance) for each facility. The
derived from an analysis of variability of
within-test condition variability (i.e., commenter determined that, on average
multiple sources is unlawful. If EPA
run variance) do not fully account for for the three facilities, total variability
considers variability at all, it must
test-to-test variability and thus was a factor of 4.2 higher than within-
consider the relevant source’s
understates total variability. test variability. Because semivolatile
variability.
Consequently, the average of the best and low volatile metals are also
Response: We developed the performing sources may not be able to
particulate matter UVF for sources controlled with a fabric filter, the
achieve the same emission level under commenter suggested that the total
equipped with a fabric filter using data a MACT performance test when
from best performing sources only.73 variability of particulate matter could be
attempting to operate under the same used as an estimate of the total
It is reasonable to aggregate conditions as it did during the
particulate matter emissions data across variability for semivolatile and low
compliance test EPA used to establish volatile metals. Thus, the commenter
source categories for all best performing the floor. Even though sources generally
sources equipped with a fabric filter suggested that the within-test condition
operated at the extreme high end of the variability for semivolatile and low
because the relationship between range of normal operations during the
standard deviation and emissions of volatile metals be increased by a factor
compliance tests EPA uses to establish of 4.2 to account for total variability
particulate matter is not expected to be the standards, the average of the best
impacted by the source category type.74 when calculating floors.
performing sources would need to
Rather, particulate emissions from fabric Response: As stated throughout this
operate under those same compliance
filters are a function of seepage (i.e., preamble, we believe that there is
test conditions to establish the same
migration of particles through the filter variability in addition to within-test
operating envelope—the operating
cake) and leakage (i.e., particles leaking condition (i.e., run-to-run) variability
envelope needed to ensure the source
through pores, channels, or pinholes that we cannot quantify—that we refer
can operate under the full range of
formed as the filter cake builds up). The to as test-to-test variability. We also do
normal emissions.
effect of seepage and leakage on not believe this test-to-test variability is
Response: We agree with commenters
emissions variability should not vary captured by compliance test operating
that we have not quantified test-to-test
conditions as discussed above, and thus
variability when establishing the floors
establishing the floor using emissions
72 USEPA, ‘‘Technical Support Document for the
for standards other than particulate
HWC MACT Standards, Volume III: Selection of data representing the extreme high end
matter where a best performing source
MACT Standards,’’, September 2005, Sections 16 of the range of normal emissions does
and 17. uses a fabric filter. We are able to
not account for test-to-test variability.
73 USEPA, ‘‘Draft Technical Support Document quantify only within-test variability
We disagree, however, with the
for HWC MACT Standards, Volume III: Selection of
MACT Standards,’’, March 2004, p. 5–4. 75 See USEPA, ‘‘Technical Support Document for commenter’s attempts to quantify the
74 In addition, emissions are not generally the HWC MACT Standards, Volume III: Selection of remaining test-to-test variability for
affected by particulate inlet loading. MACT Standards,’’ September 2005, Section 5.3. floors other than particulate matter
where all best performing sources are (1) What is the appropriate number of level (adjusted statistically to account
equipped with fabric filters. runs to use to identify the degrees of for quantifiable variability) of the source
We generally agree with the freedom and the t-statistic in the floor with the best combined ranking. A more
commenter’s approach for extracting the calculations (e.g., should we use the detailed description of the methodology
test-to-test component of variability number of runs available for metals is found in Volume III of the Technical
using the UVF curve for particulate emissions for the source or the number Support Document, section 7.3.
matter controlled with a fabric filter.76 of runs available for particulate matter This methodology provides a
The commenter has documented that for emissions from which the correction reasonable estimate of the best
cement kilns, test-to-test variability of factor is derived); and (2) should we use performing sources and their level of
particulate emissions controlled with a a generic correction factor for all source performance for HAP susceptible to
fabric filter is on average a factor of 4.2 categories or calculate source category- hazardous waste feed control. As
higher than within-test variability. specific or source-specific correction required by section 112(d)(2), EPA has
We believe the commenter’s factors. considered measures that reduce the
suggestion to adopt this correction For these reasons, we believe the volume of emissions through process
factor to semivolatile and low volatile approach we use for quantifying changes, or that prevent pollutant
metals is technically flawed and for baghouse particulate matter collection release through capture at the stack, and
several reasons would present statistical variability is not readily transferable to assessed how these control measures are
difficulties. First, total variability for other types of control devices and other used in combination. Section
semivolatile metals and low volatile HAP. We therefore are not applying a 112(d)(2)(A), (C) and (E). Hazardous
metals controlled with a fabric filter can quantified correction factor in the final waste feed control is clearly a process
be different from the total variability of rule but rather are using a MACT change that reduces HAP emissions; air
particulate matter controlled with a ranking methodology that qualitatively pollution control systems collect
fabric filter because: (1) The test accounts for total emission variability, pollutants at the stack. These are the
methods are different (i.e., Method 5 for notably test-to-test variability. best systems and measures for
particulate matter and Method 29 for controlling HAP emissions from
metals) and thus sample extraction and B. SRE/Feed Methdology hazardous waste combustors. 69 FR at
analysis methods differ; (2) the factors 1. Description of the Methodology 21226. In considering these factors, EPA
that affect partitioning of particulate has necessarily considered such factors
matter to combustion gas (i.e., As proposed, we are using the System as design of different air pollution
entrainment) are different from the Removal Efficiency (SRE)/Feed control devices, waste composition,
factors that affect semivolatile metal approach to determine the pool of best pollution control operator training and
partitioning to the combustion gas (i.e., performing sources for those HAP behavior, and use of pollution control
metal volatility); and (3) the volatility of whose emissions can be controlled in devices and methodologies in
semivolatile metals is affected by part by controlling the hazardous waste combination. CKRC, 255 F. 3d at 864–
chlorine feedrates. feed of the HAP—that is, controlling the 65 (noting these as factors, in addition
Second, adopting a variability factor amount of HAP in the hazardous waste to a particular type of air pollution
applicable to fabric filters for use on fed to the source. These are HAP metals control device, that can influence
electrostatic precipitators 77 is and chlorine. Our basic approach is to pollution control performance); 69 FR at
problematic because both test-to-test determine the sources in our database 21223 n. 47 (system removal efficiency
and within-test variability of these with the lowest hazardous waste measures all internal control
emission control devices can be vastly feedrate of the HAP in question (semi- mechanisms as well as back-end
different. Factors that affect emissions volatile metals, low volatile metals, emission control device performance).
variability for sources equipped with a mercury, or chlorine), and the sources EPA also believes that this
fabric filter include: (1) Bag wear and with the best system removal efficiency methodology reasonably estimates the
tear due to thermal degradation and for the same HAP. The system removal best performing sources’ level of
chemical attack; and (2) variability in efficiency is a measure of the percentage performance by accounting for these
flue gas flowrate. Factors that affect of HAP that is removed prior to being sources’ total variability, including their
emissions variability for sources emitted relative to the amount fed to the performance over time. The
equipped with an electrostatic unit from all inputs (hazardous waste, methodology quantifies run-to-run
precipitator are different (see discussion fossil fuels, raw materials, and any other variability. See 69 FR at 21232–33. It
in Section III.B above) and include: input). The pool of best performing does not quantify test-to-test variability
variations in particle loading and sources are those with the best because we are unable to do so for these
particle size distribution, erosion of combination of hazardous waste pollutants. (See sections A. 2.a.(2) and
collection plates, and variation in fly feedrate and system removal efficiency 3 above.) Although all variability must
ash resistivity due to changes as determined by our ranking be accounted for when calculating
atmospheric moisture and in sulfur procedure, separate best performer floors, the only definitive way to
feedrate (e.g. different type of coal). pools being determined for each HAP in accurately quantify this test-to-test
Finally, the approach raises several question (SVM, LVM, mercury, and emissions variability is through
difficult statistical questions including: chlorine), reflecting the variability evaluation of long-term continuous
inherent in each of these ranking factors emissions monitoring data, which do
76 We note, however, that an argument could be
(see A.2.a.(1) above). We then use the not presently exist. We believe,
made for using a source or condition-specific emission levels from these sources to
correction factor rather than averaging the
however, that SRE/Feed methodology
correction factors for all sources within a source calculate the emission level achieved by provides some margin for estimating
category. the average of the best performing this additional, non-quantifiable
77 We infer that the commenter suggests that we sources, as also explained in the variability. This is illustrated in the
use this correction factor for semivolatile and low previous section. This is the MACT technical support document (volume III
volatile metals controlled by both electrostatic
precipitators and fabric filters since the majority of
floor for the HAP from the source type. section 17), which clearly shows that
cement kilns are equipped with electrostatic For new sources, we use the same the straight emissions approach
precipitators. methodology but select the emission underestimates (indeed, fails to account
for) lower emitting sources’ long-term performing incinerators have the first Report) stating that MACT is not
emissions variability. These lower and second best system removal intended to drive sources out of
emitting sources that would otherwise efficiencies (and the highest two metal business). Standards that could force
not meet the floor levels on individual feedrates). It is noteworthy that the commercial sources to reduce the
days under the straight emission highest feed control level among these overall scope of their operations are also
approach would be able (or otherwise best performing sources is over three inconsistent with requirements and
are more capable) to do so under the orders of magnitude higher than the objectives of the Resource Conservation
SRE/feed approach. feed control level of the lowest feeding and Recovery Act to require treatment of
EPA further believes that the SRE/ best performing source.80 Establishing hazardous wastes before the wastes can
Feed methodology appropriately limits dominated by both superior feed be land disposed, and to encourage
accounts for design variability that control sources and back-end controlled hazardous waste treatment. RCRA
exists across sources for categories, like sources would result in floor levels that sections 3004 (d), (e), (g) and 1003 (a)
those here, which consist of a diverse are not reflective of the range of (6); see also section 112 (n) (7) of the
and heterogeneous mixture of sources. emissions exhibited by either low CAA, stating that section 112 (d) MACT
This is especially true of incinerators feeding sources or high feeding sources standards are to be consistent with
and boilers, for which there are smaller and would more resemble new source RCRA subtitle C emission standards for
on-site units that are located at widely standards for both of these different the same sources to the maximum
varying industrial sectors that types of combustors. Such floors could extent practicable (consistent with the
essentially combust single, or multiple lead to situations, for example, where requirements of section 112 (d));
wastestreams that are specific to their commercial sources could find it moreover, EPA doubts that a standard
industrial process, and off-site impracticable to achieve the standards which precludes effective treatment
commercial units dealing with many without reducing the overall scope of mandated by a sister environmental
different wastes of different origins and their operations (since the standard statute must be viewed as a type of best
HAP metal and chlorine composition. could operate as a direct constraint on performance under section 112 (d). The
EPA believes that these variations are the amount of hazardous waste that SRE/Feed methodology avoids this
best encompassed in the SRE/Feed could be fed to the device, in effect result by always considering hazardous
approach, rather than with a depriving a combustion source of its waste feed control in combination with
subcategorization scheme that could raw material). Similarly, low feeding system removal efficiency and
result in anomalous floor levels because sources that cannot achieve this floor according equal weight to both means of
there are fewer sources in each source level may be required to add expensive control in the ranking process.
subcategory from which to assess back-end control equipment that would It is also important to emphasize what
relative performance.78 See Mossville, result in minimal emission reductions, the SRE/Feed methodology does not
370 F. 3d at 1240 (upholding floor likely forcing the smaller on-site source evaluate: Feed control of HAP in fossil
methodology involving reasonable to cease hazardous waste treatment fuel or raw material inputs to these
estimation, rather than use of emissions operations and to instead send the waste devices. Emission reduction of these
data, when sources in the category have to a commercial treatment unit. HAP are controllable by back-end
heterogeneous emission characteristics The inappropriateness of a straight pollution control devices which remove
due to highly variable HAP emissions-based approach for feed a given percentage of pollutants
controlled pollutants for commercial irrespective of their origin and is
concentrations in feedstocks).
hazardous waste combustors is further assured by the system removal
Use of the SRE/Feed approach also
highlighted by the fact that several efficiency portion of the methodology,
avoids basing the floor standards on a
commercial hazardous waste as well as through the particulate matter
combination of the lowest emitting low
combustors that are achieving the standard (see section IV.A below). Feed
feeding sources and the lowest emitting
design level of the particulate matter control of these inputs is not a feasible
high feeding sources. For example, the
standard are not achieving the means of control, however. HAP content
five lowest emitting incinerators for
semivolatile and/or low volatile metals in raw materials and fossil fuel can be
semivolatile metals that would comprise
straight emissions based design level, highly variable, and so cannot even be
the MACT pool using a straight
replicated by a single source. Raw
emissions methodology include three and, in some instances, floor level.81
material and fossil fuel sources are also
sources that are the first, second, and This provides further evidence that low
normally proprietary, so other sources
fourth lowest feeding sources among all feeding sources are in fact biasing some
would not have access to raw material
the incinerators.79 The other two best of the straight emissions-based floors to
and fossil fuel available (in its
the extent that even the sources with the
performance test) to a source with low
78 At proposal, we conducted a technical analysis most efficient back-end control devices
to determine potential subcategorization options. HAP fossil fuel and raw material inputs.
would be incapable of achieving the
We then conducted an analysis to determine if Such sources would thus be unable to
emission standards calculated on a
these different types of sources exhibited duplicate these results. Moreover, there
statistically different emissions. Although EPA in straight emission basis.
These results are inconsistent with are no commercial-scale pretreatment
the end determined that these source categories
should not be subcategorized further, this decision the intent of the section 112 (d) (see 2 processes available for removing or
was based in part because the SRE/Feed
Legislative History at 3352 (House reducing HAP content in raw materials
methodology better accounts for the range of or fossil fuels to these units. See
emissions from the best performing sources for technical support document volume III
these diverse combustion types. See USEPA, Table ‘‘E_INC_SVMCT’’ and, to determine relative
‘‘Technical Support Document for the HWC MACT feed control and SRE rankings for these sources, section 17.5 and 25; see also 69 FR at
Standards, Volume III: Selection of MACT Appendix E Table ‘‘SF_INC_SVMCT’’. 21224 and n. 48.
Standards,’’ September 2005, Section 4, for an 80 Source 340 had a semivolatile metal feed
explanation of the subcategorization assessment, control MTEC of 892 µg/dscm, whereas source 327 2. Why Aren’t the Lowest Emitters the
which includes examples of anomalous floor results had a semivolatile metal feed control MTEC of Best Performers?
for certain subcategorization approaches. 3,080,571 µg/dscm.
79 See USEPA, ‘‘Technical Support Document for 81 See USEPA, ‘‘Technical Support Document for Some commenters nonetheless argue
the HWC MACT Standards, Volume III: Selection of the HWC MACT Standards, Volume III: Selection of that best performing sources can only
MACT Standards,’’ September 2005, Appendix C, MACT Standards,’’ September 2005, Section 17.4 mean sources with the lowest HAP
emissions, and that the SRE/Feed (d) best performing sources under the 1240–41 (collecting cases). In Mossville,
methodology is therefore flawed straight emissions approach are sources had varying levels of vinyl
because it does not invariably select projected, based on two separate chloride emissions due to varying
lowest emitters as best performers.82 analyses using reasonable assumptions, concentrations of vinyl chloride in their
The statute does not compel this result. not to achieve the straight emissions feedstock. Individual measurements
There is no language stating that lowest floor standard based on these sources’ consequently did not adequately
emitting sources are by definition the demonstrated variations in system represent these sources’ performance
best performers. The floor for existing removal efficiencies over time (i.e., from over time. Not-to-exceed permit limits
sources is to be based on the average test-to-test); and (e) SRE/feed thus reasonably estimated sources’
emission limitation achieved by the methodology yields floor levels (i.e. the performance, corroboration being that
‘‘best performing’’ 12 per cent of floor standards in the rule) that better individual sources with the lowest long-
sources. Section 112(d)(3)(A). This estimate the emission levels reflecting term average performance occasionally
language does not specify how ‘‘best the performance over time of the best came close to exceeding those permit
performing’’ is to be determined: by performing sources. See Mossville, 370 limits. Id. at 1241–42. The facts are
means of emission level, emission F. 3d at 1242 (floor standard is similar here, since our examination of
control efficiency, measured over what reasonable because it accommodated best performing sources with multiple
period of time, etc. See Sierra Club v. best performing source’s highest level of test conditions likewise shows instances
EPA, 167 F. 3d at 661 (language of floor performance (i.e. its total variability), where these sources would be unable to
requirement for existing sources ‘‘on its even though the level of the standard meet floors established based solely on
own says nothing about how the was higher than any individual lowest emissions (including their own).
performance of the best units is to be measurement from that source). As here, EPA was not compelled to base
calculated’’). Put another way, this As noted earlier, the straight the floor levels on the lowest measured
language does not answer the question emissions methodology can also limit emission levels.
of which source is the better performing: operation of commercial units because Comment: The same commenter
one that emits 100 units of HAP but also the standard reflects a level of maintains that it is clear from the
feeds 100 units of that HAP, or one that hazardous waste feed control which caselaw that MACT floors must reflect
emits 101 units of the HAP but feeds could force commercial units to burn the relevant best performing sources’
10,000 units. See 69 FR at 21223. less hazardous waste because such ‘‘actual performance’’, and that this
Moreover, new source floors are to be standards more resemble new source must refer to the emissions level it
based on the performance of the ‘‘best standards. The straight emissions achieves.
controlled’’ similar source achieved in methodology also arbitrarily reflects Response: As just stated, the D.C.
HAP levels in raw materials and fossil Circuit has repeatedly stated that EPA
practice. Section 112(d)(3). ‘‘Best
fuels, an infeasible means of control for may make reasonable estimates of
controlled’’ can naturally be read to
any source. sources’ performance in assessing both
refer to some means of control such as
Another arbitrary, and indeed which sources are best performing and
system removal efficiency as well as to
impermissible, result of the straight the level of their performance. The court
emission level.
emissions methodology is that in some has further indicated that EPA is to
Use of a straight emissions approach
instances (noted in responses below) the account for variability in assessing
to identify floor levels can lead to
methodology results in standards which sources’ performance for purposes of
arbitrary results. Most important, as
would force sources identified as best establishing floors, and this assessment
explained above, it leads to standards
performing to install upgraded air may require that EPA make reasonable
which cannot be achieved consistently
pollution control equipment. This result estimates of performance of best
even by the best performing sources undermines section 112 (d) (2) of the performing sources. CKRC, 255 F. 3d at
because operating variability is not statute, by imposing what amounts to a 865–66; Mossville, 370 F. 3d at 1241–42.
accounted for. This is shown in section beyond the floor standard without See discussion in A.1.a above.
17 of volume III of the technical support consideration of the beyond the floor Comment: The commenter generally
document. These analyses show that (a) factors: the cost of achieving those maintains that EPA’s floor approaches
emissions from these sources do in fact reductions, as well as energy and nonair consider only the performance of back-
vary from test-to-test, and that no two environmental impacts. end pollution control technology and so
snapshot emission test results are Comment: The commenter states that fail to capture other means of HAP
identical; (b) our statistical approach because MACT floors must reflect the emission control that otherwise would
that quantifies within test, run-to-run ‘‘actual performance’’ of the relevant be captured if EPA were to assess
variability underestimates the best best performing hazardous waste performance based on the emission
performing sources’ long term, test-to- combusters, this means that the lowest levels each source achieved.
test variability; 83 (c) best performing emitters must be the best performers. Response: EPA agrees that factors
sources under the straight emissions The commenter cites CKRC v. EPA, 255 other than end-of-stack pollution
approach advocated by the commenter F. 3d at 862 and other cases in support. control can affect metal HAP and
(i.e. the lowest emitting sources) had Response: As explained in the chlorine emissions. This is why EPA
other test conditions that did not introduction above, the statute does not assesses performance for these HAP by
achieve straight emission floor levels; specify that lowest emitters are considering combinations of system
invariably best performers. Nor does the removal efficiency (which measures
82 In fact, many of the sources identified as best
caselaw cited by the commenter support every element in a control system
performing under the SRE/Feed methodology are
also the lowest emitting, although this is not this position. The D.C. Circuit has held resulting in HAP reduction, not limited
invariably the case. repeatedly that EPA may determine to efficiency of a control device), and
83 Best performing sources pursuant to the which sources are best performing and hazardous waste HAP feed control.
straight emissions methodology are projected to be may ‘‘reasonably estimate’’ the Standards for dioxins and other organic
unable to achieve the levl of their of their
performance test emissions even after they are
performance of the top 12 percent of HAP (which have no hazardous waste
adjusted upward to account for run-to-run these sources by means other than use feed control component) likewise assess
variability. of actual data. Mossville, 370 F. 3d at every element of control.
EPA also accounts for the variability materials and fossil fuels are effectively 112(d)(3) says that sources have to use
of HAP levels in the (essential) use of controlled with the particulate matter the means of achieving a level of
raw materials and fossil fuels by standard, a standard that is based on the performance that other best performing
assessing performance of back-end sources with best back-end control sources used.
control but not evaluating fuel/raw devices. The only element which is not Response: The thrust of this comment
material substitution, which, as controlled is what cannot be: HAP is essentially to impermissibly bypass
discussed later in the response to levels in feeds for which fuel or raw the beyond-the-floor factors set out in
comments section, are infeasible means material switching is simply not an section 112(d)(2) under the guise of
of control. Mossville, 370 F. 3d at 1241– available option. adopting a floor standard. Suppose that
42, is instructive on this point. The Comment: The commenter further EPA were to adopt a floor standard
court held that the constant change in maintains, however, that the means by dominated by emission levels reflecting
raw materials justified EPA’s use of a which sources may be achieving levels HAP concentrations present in a few
regulatory limit to estimate a floor level. of performance are legally irrelevant sources’ raw materials and fossil fuels
The reasonableness of this level was (citing National Lime Ass’n v. EPA, 233 during their test conditions. Suppose
confirmed by showing that the highest F. 3d 625 , 634 and 640 (D.C. Cir. 2000)). further that some sources have to
individual data point of a best The fact that sources with ‘‘cleaner’’ raw upgrade their back-end control
performing source was nearly at the material and fossil fuel inputs may not equipment to operate at efficiencies
level of the regulatory limit. Under the intend to have resulting lower HAP better than the average level
commenter’s approach, the court would emissions is therefore without legal demonstrated by the best performing
have had no choice but to hold that the bearing. sources, because test results based on
level the source achieved in a single test Response: The issue here is not one of fossil fuel and raw material levels are
result using ‘clean’ raw materials—i.e. intent. The Court, in National Lime, neither replicable nor duplicable. In this
the ‘level achieved’ in the commenter’s rejected the argument that sources’ lack situation, EPA believes that it would
language—dictated the floor level. of intent to control a HAP did not have improperly adopted a beyond-the-
See part four, section III.C for EPA’s preclude EPA from establishing a floor standard because EPA would have
response to this comment as it relates to section 112(d) standard for that HAP. failed to consider the beyond-the-floor
the methodologies for the particulate See 233 F. 3d at 640, rejecting the factors (cost, energy, and nonair
matter standard and total chlorine argument that HAP metal control environmental impacts) set out in
standard for hydrochloric acid achieved by use of back-end control section 112(d)(2).85
production furnaces. devices (baghouses) could not be Comment: EPA has not substantiated
Comment: The commenter notes that assessed by EPA because the sources its claim that sources cannot switch
the SRE/Feed methodology does not used the back-end control devices to fossil fuels or raw materials.
account for all HAP emissions, failing to control emissions of particulate matter. Response: At proposal we evaluated
account for metal and chlorine feedrates The case did not consider the facts fuel switching and raw material
in raw materials and fossil fuels. present here, where the issue is not a substitution as beyond-the-floor
Response: The methodology does not source’s intent, but rather a means of technologies for cement kilns and
assess the effect of feed ‘‘control’’ of control which involves happenstance lightweight aggregate kilns and stated
HAP levels in raw materials or fossil (composition of HAP in raw materials these technologies would not be cost
fuels which may be inputs to the and fossil fuel used the day the test was effective.86 We also discussed why fuel
combustion units. This is because such conducted) and so is neither replicable switching is not an appropriate floor
nor duplicable. control technology for solid fuel-fired
control may not be replicable by an
National Lime also held that EPA boilers. 69 FR at 21273. Upon further
individual source, or duplicable by any
must establish a section 112(d) emission evaluation, we again conclude that fuel
other source. See 69 FR at 21224 and n.
standard for every HAP emitted by a switching and raw material substitution
48; Sierra Club v. EPA, 353 F. 3d 976,
major source. 233 F. 3d at 634. EPA is are not floor control technologies and
988 (‘‘substitution of cleaner ore stocks
establishing emission standards for all are not cost effective beyond-the-floor
was not * * * a feasible basis on which
HAP emitted by these sources. In technologies for cement kilns,
to set emission standards. Metallic
establishing these standards, EPA is not lightweight aggregate kilns, and solid
impurity levels are variable and
evaluating emission reductions fuel-fired boilers.87
unpredictable both from mine to mine Comment: EPA has failed to
attributable to the type of fossil fuel and
and within specific ore deposits, document the basis for its SRE ranking.
raw material used in the performance
thereby precluding ore-switching as a
tests, because this is not a ‘‘feasible
predictable and consistent control 85 Analysis of the levels of HAP in raw matrial
basis on which to set emission
strategy’’).84 EPA’s methodology does and nonhazardous waste fuels suggests that this is
standards.’’ Sierra Club, 353 F. 3d at
account for HAP control of all inputs by a realistic outcome. Our analysis shows that
988. emissions attributable to raw material and fossil
assessing system removal efficiency, EPA thus does not agree with this fuel can be significant relative to the level of the
which measures reductions of HAPs in comment because the issue is not a straight emissions-based floor design level and floor
all inputs (including fossil fuel and raw source’s intent but rather whether or not (the methodology advocated by the commenter),
materials) to a hazardous waste and therefore could inappropriately impact a
to assess emission reductions from sournce’s ability to comply with such a floor
combustion unit. Further, nonmercury individual test results which reflect an standard. See USEPA, ‘‘Technical Support
metal HAP emissions attributable to raw infeasible means of control. Document for the HWC MACT Standards, Volume
Comment: The commenter maintains, III: Selection of MACT Standards,’’ September 2005,
84 Although this language arose in the context of Section 17.6.
a potential beyond-the-floor standard, EPA believes
however, that even if individual sources 86 See, for example, 69 FR at 21252, where we
that the principle stated is generally applicable. (including those in the pool of best discuss the use of fuel-switching or raw material
MACT standards, after all, are technology-based, performing sources) cannot reduce HAP substitution as a possible beyond-the-floor control
and if there is no technology (i.e. no avaialble concentrations in raw materials and for mercury at cement kilns.
means) to achieve a standard—i.e. for a soruce to 87 See USEPA, ‘‘Technical Support Document for
achieve a standard whenever it is tested (as the

fossil fuels, they may achieve the same the HWC MACT Standards, Volume III: Selection of
rules require)—then the standard is not an reductions by adding back-end MACT Standards, September 2005, Sections 11 and
achieveable one. pollution control. Nothing in section 25, for further discussion.
Specifically, EPA has not stated how it Response: System removal efficiency database that is calculated by the
measured sources’ SREs, or how it is a parameter that is included in our following formula:
knows those rankings are accurate.
( total HAP mass feedrate ) − ( stack gas HAP mass emissiion rate ) 
SRE = 100 × 
total HAP mass feedrate
The HAP feedrate and emission data Response: Emission levels are used to gives a number of reasons for its
are components of the database that calculate system removal efficiencies in criticisms, which we answer in the
were extracted from emission test order to assess each source’s relative following sequence of comments listed
reports for each source. We use system back-end control efficiency. Also, as a though f.
removal efficiency for each relevant explained in the introduction to this a. Comment: The commenter states
pollutant or pollutant group (e.g., comment response section, the SRE/ that EPA claims emission levels do not
semivolatile metals, low volatile metals, Feed methodology uses the stack fully reflect variability in part because
mercury, total chlorine) whenever the emission levels of the sources using the they are sometimes based on tests where
data allows us to calculate a reliable best combinations of hazardous waste the source was feeding low levels of
system removal efficiency. For example, feed control and system-wide air HAP during the test. The commenter
we generally do not use system removal pollution control (expressed as HAP claims this is inconsistent with the fact
efficiencies that are based on normal percent removal over the entire system) that EPA preferentially uses worst-case
emissions data because of the concern to calculate the floors. The data are emissions obtained from tests where the
that normal feed data are too sensitive adjusted statistically to account for sources spiked their feedstreams with
to sampling and measurement error. See quantifiable forms of variability (run-to- metals, and that the mere possibility
69 FR at 21224.88 run variability). This methodology that these emissions do not reflect test
reasonably selects best performing data from conditions where variability
The system removal efficiencies used sources (for HAP amenable to these was not maximized does not mean those
in our ranking process are reliable and means of control), and reasonably data fail to represent a source’s actual
accurate because the feed and emissions estimates these sources’ performance performance. The commenter also states
data originate from compliance tests over time. As further stated in section that ‘‘EPA’s apparent suggestion that the
that demonstrate compliance with B.2 above, using a straight emissions best performing sources could not
existing emission standards (primarily approach to identify best performers replicate the average performance of the
RCRA requirements). As such, the data and their level of performance can lead sources with the lowest emissions is
are considered to have excellent to standards for these HAP that do not unsubstantiated and unexplained.
accuracy and quality. RCRA trial burn fully account for variability (including Assuming that EPA accurately assesses
and certification of compliance reports variability resulting from varying and/or a source’s actual performance, the
are typically reviewed in detail by the uncontrollable amounts of HAP in raw source can replicate that performance.’’
permitting authority. The compliance materials and fossil fuels) and could Response: HAPs in raw materials and
tests and test reports generally contain force installation of de facto beyond-the- fossil fuels contribute to a source’s
the use of various quality assurance floor controls without consideration of emissions. EPA has concerns that a
procedures, including laboratory, the section 112(d)(2) beyond-the-floor straight emissions approach to setting
method, and field blanks, spikes, and factors. floors may not be replicable by the best
surrogate samples, all of which are EPA thus does not see the performing sources nor duplicable by
designed to minimize sampling and contradiction expressed by the other non-best performing sources
analytical inaccuracies. EPA also commenter. Use of the straight because of varying concentration levels
noticed the data base for this rule for emissions approach as advocated by the of HAP in raw material and
multiple rounds of comment and has commenter would lead to standards that nonhazardous waste fuels. The best
made numerous changes in response to do not reasonably estimate sources’ performing sources operated under
comment to assure accuracy of the performance and which could not be compliance test conditions as the
underlying data. Thus, EPA concludes achieved even by the best performers commenter suggests. However, raw
the calculated system removal with individual test conditions below material and nonhazardous fuel HAP
efficiencies used in the ranking process the average of the 12 percent of best concentrations for the best performing
are both reliable and accurate. performing sources. These problems sources will change over time, perhaps
Comment: EPA’s approach with would be compounded many-fold if the due to a different source of fuel or raw
regard to use of stack data is internally data were not normalized and adjusted material quarry location, which could
contradictory. EPA uses stack data in to at least account for quantifiable affect their ability to achieve the floor
establishing floors, but does not use variability, steps the commenter also level that was based on emissions
stack data to determine which opposes. EPA’s use of emissions data obtained while processing different
performers are best. EPA has failed to (suitably adjusted) after identifying best fossil fuel or raw materials. EPA takes
explain this contradiction. performers through the ranking sharp issue with the commenter’s
methodology avoids these problems and statement that a single performance test
88 See USEPA, ‘‘Technical Support Document for reasonably estimates best performers’ result is automatically replicable so long
the HWC MACT Standards, Volume II: Database,’’ level of performance. as it is measured properly in the first
September 2005, Section 2, for further discussion Comment: The commenter rejects instance. This statement is incorrect
on system removal efficiencies, which includes EPA’s finding (69 FR at 21226) that even disregarding HAP contributions in
sample calculations and references to the database
that contain the calculated system removal
individual test results in the data base raw materials and fossil fuels since, as
efficiencies for each source and each HAP or HAP do not fully express the best performing noted previously in section A.2.e, there
ER12OC05.000</MATH>
group. sources’ performance. The commenter are many other sources of variability
which will influence sources’ methodology is flawed. The standards e. Comment: The commenter
performance over time (i.e., in are not mutually dependent, so the fact criticizes EPA’s discussion at 69 FR
subsequent performance tests). that they are not achieved 21227–228 indicating that both
A straight emissions approach for simultaneously is irrelevant. There is no hazardous waste feed control and back-
establishing semivolatile and low reason a best performer for one HAP end pollution control are superior
volatile metal floors may result in should be a best performer for other means of HAP emission control and
instances where the best performing HAP. treatment standards should be
sources would not be capable of Response: EPA agrees with this structured to allow either method to be
achieving the standards if their raw comment. On reflection, EPA believes the dominant control mechanism.
material and nonhazardous waste fuel that because all our standards are not Response: EPA is not relying on this
HAP levels change over time. For each technically interdependent (i.e., part of the proposed preamble
cement kiln and lightweight aggregate implementation of one emission control discussion as justification for the final
kiln, we estimated the emissions technology does not prevent the source rule, with the one exception noted in
attributable to these raw materials and from implementing another control the response to the following comment.
fossil fuels assuming each source was technology), the fact that sources are not f. Comment: Considerations of proper
operating with hazardous waste HAP achieving all the standards waste disposal policy are not relevant to
feed and back-end control levels simultaneously does not indicate a flaw MACT floor determinations. In any case,
equivalent to the average of the best in a straight emissions approach. See the possibility that some commercial
performing sources (the difference in Chemical Manufacturers Ass’n, 870 F. waste combustors may upgrade their
emissions across sources only being the 2d at 239 (best performing sources can back-end pollution control systems to
result of the differing HAP levels in the be determined on a pollutant-by- meet standards reflecting low hazardous
nonhazardous waste feeds). The pollutant basis so that different plants waste HAP feedrates, or divert wastes to
analysis shows that emissions can be best performers for different better-controlled units, is positive, not
attributable to these nonhazarous waste pollutants). negative.
feedstreams (raw materials and fossil d. Comment: Several commenters Response: As discussed in section B.1
fuels) varies across sources, and can be took the opposite position that EPA above, there are instances where
significant relative to the level of the must assure that all existing source standards derived by using a straight
straight emissions-based floor design standards must be achievable by at least emissions approach are based on a
level and floor, and therefore could 6 percent of the sources, and that all combination of lowest emitting low
inappropriately impact a source’s ability new source standards must be feeding sources and lowest emitting
to comply with the floor standard.89 achievable by at least one existing higher feeding sources. Resulting floor
b. Comment: The commenter states source. standards would thus reflect these low
Response: As discussed above, we are
that EPA must consider contributions to hazardous waste feedrates and could
not obligated to establish a suite of
emissions from raw materials and fossil put some well-controlled commercial
floors that are simultaneously
fuels, that it is irrelevant if sources from incinerators in the untenable situation
achievable by at least six percent of the
outside the pool of best performing of having to reduce the amount of
sources because the standards are not
sources can duplicate emission levels hazardous waste that is treated at their
technically interdependent.
reflecting ‘‘cleaner’’ raw materials and source. Our database verifies that such
Nonetheless, the SRE/Feed methodology
fossil fuels used by the best performing an outcome is in fact realistic.92
does result in existing floor levels (when
sources, and that sources unable to This type of standard would operate
combined with the other floor levels for
obtain such ‘‘cleaner’’ inputs may as a direct constraint on the amount of
sources in the source category) that are
always upgrade other parts of their hazardous waste that could be fed to the
simultaneously achievable by at least
systems to achieve that level of device, in effect depriving a combustion
six percent of the sources (or, for source
performance. source of its raw material. In this
categories that have fewer than 30
Response: As previously discussed, instance, hazardous wastes could not be
sources, by at least two or three
EPA’s methodology does account for readily diverted to other units because
sources).90 However, for the new source
HAP control of all inputs by assessing the low feeding hazardous waste
standards, three of the source categories
system removal efficiency, which sources tend not to be commercial units.
do not include any sources that are
measures reductions of HAPs from all In these circumstances, there would be
simultaneously achieving all the
inputs. Further, nonmercury metal HAP a significant adverse nonair
standards (incinerators, cement kilns,
emissions attributable to raw materials environmental impact. Hazardous waste
and lightweight aggregate kilns). Again,
and fossil fuels are effectively controlled is required to be treated by Best
similar to existing sources, EPA is not
with the particulate matter standard, a Demonstrated Available Technology
obligated to establish a suite of new
standard that is based on the sources (BDAT) before it can be land disposed.
source floors that are simultaneously
with lowest emissions from best back- RCRA sections 3004 (d), (e), (g), and (m);
achievable by at least one existing
end control devices. We are not basing Hazardous Waste Treatment Council v.
source because these standards are not
any standards on performance of EPA, 866 F. 2d 355, 361 (D.C.Cir. 1990)
technically interdependent. We
sources not ranked as among the best (upholding Best Demonstrated Available
conclude that a new source can be
performing. Technology treatment requirement).
c. Comment: The commenter disputes designed (from a back-end control
perspective) to achieve all the new Most treatment standards for organic
EPA’s conclusions that failure of pollutants in hazardous waste can only
sources to meet all of the standards source standards.91
be achieved by combustion. Leaving
based on a straight emissions 90 These achievability analyses did not account some hazardous wastes without a
methodology at once shows that the for the additional test-to-test variability that we
cannot quantify. such a new source could be designed to achieve the
89 See USEPA, ‘‘Final Technical Support 91 See USEPA, ‘‘Technical Support Document for new source standards.
Document for the HWC MACT Standards, Volume the HWC MACT Standards, Volume V: Emission 92 See USEPA, ‘‘Technical Support Document for
III: Selection of MACT Standards,’’ September 2005, Estimates and Engineering Costs,’’ September 2005, the HWC MACT Standards, Volume III: Selection of
Section 17.6. . Section 4.2.3 for a discussion that explains how MACT Standards’’, September 2005, Section 17.4.
treatment option is in derogation of aggregated score (as proposed), as matter and for total chlorine, we discuss
these statutory requirements and goals, opposed to the source with the lowest the approaches separately below.
and calls into question whether a emissions among the best performing
1. Air Pollution Control Device
treatment standard that has significant existing sources (an approach on which Methodology for Particulate Matter
adverse nonair environmental impacts we requested comment).
must be viewed as best performing. See Our approach to establishing floor
Response: We agree with the standards for particulate matter raises
Portland Cement Ass’n v. Ruckelshaus, commenters because this is consistent
486 F. 2d 375 , 386 (D.C. Cir. 1973); three major issues.
with our methodology for defining best The first issue is whether particulate
Essex Chemical Co. v. EPAEPA, 486 F.
performers for existing sources and matter is an appropriate surrogate for
2d 427, 439 (D.C. Cir. 1973). The
assessing their level of performance. We non-enumerated HAP metals from all
commenter’s statement that waste
note, however, that with respect to the inputs, and for all non-mercury HAP
disposal policy is not relevant to the
new source standards, we encountered metals in raw material and fossil fuel
MACT standard-setting process is not
two instances where the SRE/Feed inputs. This issue is discussed at section
completely correct, since section 112 (n)
methodology identified multiple IV.A of this part, where we conclude
(7) of the Clean Air Act directs some
sources with identical single best that particulate matter is indeed a
accommodation between MACT and
RCRA standards for sources combusting aggregated scores, resulting in a tie for reasonable surrogate for these metal
hazardous waste. Part of this the best performing source. This HAP.
accommodation is using a methodology occurred for the mercury and low The second issue is why EPA is not
to evaluate best performing sources that volatile metal new source standards for evaluating some type of feed control for
evaluates as best performers those using incinerators. In these instances, EPA the particulate matter floor. There are
the best combination of hazardous waste applied a tie breaking procedure that two potential types of feed control at
feed control (among other things, an resulted in selecting as the single best issue: hazardous waste feed control of
existing control measure under RCRA performing source as that source (of the nonenumerated metals, and feed control
rules) and system-wide removal. tied sources) with the lowest emissions. of non-mercury HAP metals in raw
We assessed whether we could We believe this is a reasonable material and fossil fuel inputs. With
address this issue by subcategorizing interpretation of section112(d)(3), which respect to feed control of non-
commercial incinerators and on-site states the new source standard shall not enumerated metals in hazardous waste,
incinerators. Applying the straight be less stringent than the emission as discussed in more detail in section
emission approach to such a control that is achieved in practice by IV.A of this part, we lack sufficient
subcategorization scheme, however, the best controlled similar source reliable data on non-enumerated metals
yields anomalous results due to the (‘‘source’’ being singular, not plural). to assess their feedrates in hazardous
scarcity of available and complete Moreover, we believe use of the waste. In addition, there are significant
compliance test data from commercial emission level as the tie-breaking questions about whether feedrates of the
incinerators. Calculated floor levels for criteria is reasonable, not only because non-enumerated metals can be
semivolatile metals and low volatile it is a measure of control, but because optimized along with SVM and LVM
metals for the commercial incinerator we have already fully accounted for feedrates. We also have explained
subcategory equate to 2,023 and 111 µg/ hazardous waste feedrate control and elsewhere why control of hazardous
dscm, respectively (both higher than the system removal efficiency in the ranking waste ash feedrate would be technically
current interim standards).93 We methodology. To choose either of these inappropriate, since it would not
conclude that the SRE/Feed factors to break the tie would give that properly assess feed control of
methodology better addresses this issue factor disproportionate weight. nonenumerated metals in hazardous
because it yields floor levels that better waste. See also 69 FR at 21225.
C. Air Pollution Control Technology We have also explained why we are
represent the performance of the best
Methodologies for the Particulate Matter not evaluating control of feedrates of
performing commercial incinerators and
Standard and for the Total Chlorine HAP metals in raw materials and fossil
onsite incinerators alike by applying
Standard for Hydrochloric Acid fuels to hazardous waste combusters: it
equal weights to hazardous waste feed
Production Furnaces is an infeasible means of control. See
control and back-end control in the
ranking process. At proposal, EPA used what we section B of this part. We consequently
EPA notes, however, that its choice of termed ‘‘air pollution control are not evaluating raw material and
the SRE/Feed methodology is justified technology’’ methodologies to estimate fossil fuel ash feed control in
independent of considerations of floor levels for particulate matter from determining the level of the various
adverse impact on hazardous waste all source categories as a surrogate for floors for particulate matter.
treatment and disposal. a. The methodology. The final issue is
non-mercury HAP metals, and for total
Comment: The commenter reiterates the means by which EPA is evaluating
chlorine from hydrochloric acid furnace
its comments with respect to floor levels back-end control. Essentially, after
production furnaces. 69 FR at 21225–
for new sources. determining (as just explained) that
226. Under this approach, we do not
Response: EPA’s previous responses back-end control is the means of
to comments apply to both new and estimate emission reductions
controlling non-mercury metal HAP and
existing source standards. attributable to feed control, but instead
that particulate matter is a proper
Comment: Two commenters assess the performance of back-end
surrogate for these metals, EPA is using
recommend that EPA define the single control technologies.94 We are adopting
its engineering judgment to determine
best performing source as that source the same methodologies for these HAP
what the best type of air pollution
with the lowest aggregated SRE/Feed in the final rule. Because the details of
control device (i.e., back-end control) is
the approaches differ for particulate
to control particulate matter (and, of
93 See USEPA, ‘‘Technical Support Document for
course, the contained HAP metals). We
the HWC MACT Standards, Volume III: Selection of 94 See generally USEPA, ‘‘Technical Support
MACT Standards’’, September 2005, Section 4. and Document for the HWC MACT Standards, Volume
then ascertain the level of performance
Appendix C, Table ‘‘E–INC–SVM–CT–COM’’ and III: Selection of MACT Standards’’, September 2005, by taking the average of the requisite
Table ‘‘E–INC–LVM–CT-COM’’ Section 7.4 and 7.5. number of sources (either 12 % or five,
depending on the size of the source of which are supported by our data, are section 112(d)(3). We therefore conclude
category) equipped with the best back- found in section 16 of volume III of the that a straight emissions floor would not
end control with the lowest emissions.95 technical support document. be achievable for any source feeding
These floor standards are therefore We therefore conclude sources appreciable levels of ash, even if they all
essentially established using a straight equipped with baghouses are the best were to upgrade with baghouses, or
emissions methodology. We have performers for particulate matter control baghouses in combination with HEPA
determined that baghouses (also termed not only based on engineering filters, and that a rote selection of lowest
fabric filters) are generally the best air judgment, but because we are able to emitters as best performers can lead to
pollution control technology for control reliably quantify their likely the nonsensical result of uncontrolled
of particulate matter, and that performance over time. The straight units being classified as best performers.
electrostatic precipitators are the next emissions methodology ignores the Comment: Commenter claims end-of-
best. presence of long-term emissions stack control technology is not the only
b. Why not select the lowest emitters? variability from sources not equipped factor affecting emissions of particulate
Although sources with baghouses with baghouses, and assumes without matter, stating that EPA admits that
tended to have the lowest emission basis that these sources are always particulate matter emission levels are
levels for particulate matter, this was better performing sources in instances affected by the feedrate of ash.
not invariably the case. There are where they achieved lower snapshot Accordingly, the performance of a
certain instances when sources emissions relative to the emissions from source’s end-of-stack control technology
controlled with electrostatic baghouses, emissions that have notably is not a reasonable estimate of that
precipitators (or, in one instance, a already been adjusted to account for source’s total performance.
venturi scrubber) had lower emissions long-term emissions variability. Response: The particulate matter
in individual test conditions than A straight emissions approach also standard serves as a surrogate control
sources we identified as best performing results in inappropriate floor levels for for the non-enumerated metals in the
which were equipped with baghouses.96 particulate matter because it improperly hazardous waste streams (for all source
Under the commenter’s approach, we reflects/includes low ash feed when categories), and all nonmercury metal
must always use these lowest emitting identifying best performing sources for HAP in the nonhazardous waste process
sources as the best performers. particulate matter. 69 FR at 21228. For streams (essentially, raw materials and
We again disagree. We do not know example, the MACT pool of best fossil fuels) for cement kilns,
if these sources equipped with control performing liquid fuel boilers for lightweight aggregate kilns, and liquid
devices other than baghouses with particulate matter under the straight fuel boilers. The commenter suggests
lower emissions in single test emissions approach includes eight that the APCD approach inappropriately
conditions would actually have lower sources, only one of which is equipped ignores HAP feed control in the
emissions over time than sources with a back-end control device. These assessment of best performing sources.
equipped with baghouses because we sources have low particulate matter We conclude that it would not be
cannot assess their uncontrollable emissions solely because they feed low appropriate to use a methodology that
emissions variability over time. Our levels of ash. The average ash inlet directly assesses feed control, such as
data suggests that they likely are not loadings for these sources are well over the SRE/Feed methodology, to
better performing sources. We further two orders of magnitude lower than the determine particulate matter floors.
conclude that our statistical procedures average ash inlet loading for the best First, direct assessment of total ash feed
that account for these sources’ within performing sources that we identify control would inappropriately assess
test, run-to-run emissions variability with the Air Pollution Control and seek to control (even though
underestimates these sources long-term Technology approach. (Of course, since variability of raw material and fossil
emissions variability. This is not the ash loadings are not a proper surrogate fuel inputs are uncontrollable) raw
case for sources equipped with for HAP metals, these sources’ material and fossil fuel HAP input, as
baghouses, where we have completely emissions are lowest for particulate well as raw material and fossil fuel
assessed, quantified, and accounted for matter but not necessarily for HAP input. Controlling raw material and
long-term, test-to-test emissions metals.) The straight emissions fossil fuel HAP input is infeasible, as
variability through application of the approach would yield a particulate previously discussed. It also
universal variability factor.97 The matter floor level of 0.0025 gr/dscf (with inappropriately limits theses sources’
sources equipped with control devices a corresponding design level of 0.0015 feedstocks that are necessary for their
other than baghouses with lower gr/dscf). There is not one liquid fuel associated production process.
snapshot emissions data could therefore boiler that is equipped with a back-end Second, we do not believe that
have low emissions in part because they control that achieved this floor level, developing a floor standard based on
were operating at the low end of the much less the design level. The best hazardous waste feed control of
‘‘uncontrollable’’ emissions variability performing source under the air nonenumerated metals (as opposed to
profile for that particular snapshot in pollution control technology approach, feed control of these metals in raw
time. The basis for these conclusions, all which is equipped with both a fabric material and fossil fuels) is appropriate
filter and HEPA filter, did not even or feasible. In part four, section IV.A, we
95 As explained in the responses below, the make the pool of best performing explain that we lack the data to reliably
approach varies slightly if the requisite number of sources for the straight emissions assess direct feedrate of these metals in
sources do not all use the best back-end pollution approach. Yet this unit has an excellent hazardous waste. In addition, we also
control technology. In that case, EPA includes in its
pool of best performers the lowest emission levels ash removal efficiency of 99.8% and the discuss that it is unclear (the lack of
from sources using the next best pollution control lower emitting devices’ removal certainty resulting from the sparse
technology. efficiencies are, for the most part, 0% available data) that hazardous waste
because they do not have any back-end feed control of the nonenumerated
the HWC MACT Standards, Volume III: Selection of controls. EPA believes that it is arbitrary metals is feasible. The majority of these
MACT Standards,’’ September 2005, Section 22.
97 See USEPA, ‘‘Technical Support Document for to say that these essentially metals are not directly regulated under
the HWC MACT Standards, Volume III: Selection of uncontrolled devices must be regarded existing RCRA requirements, so sources
MACT Standards,’’ September 2005, Section 5.3. as ‘‘best performing’’ for purposes of have optimized control of the other HAP
metals, raising issues of whether levels each source achieved because, as hazardous industrial waste combustors,
simultaneous optimization of feed EPA admits, it fails to account for the and non-hazardous waste Portland
control of the remaining metals is effect of ash feed rate. cement kilns.100
feasible. Moreover, even if one were to Response: We explain above why the After we assign a ranking score to
conclude that hazardous waste feed Air Pollution Control Technology
each back-end control class, we
control is feasible for the approach properly identifies the
determine the number of sources that
nonenumerated metal HAPs, hazardous relevant best performing sources for
are using each of these control
waste ash feedrates are not reliable purposes of controlling non-mercury
technology classes. We then identify the
indicators of nonmercury metal HAP metal HAP (measured as particulate
MACT control technology or
feed control levels and are therefore matter), irrespective of ash feed rates.
Typically, this results in selecting the technologies to be those best ranked
inappropriate parameters to assess in
sources with the lowest particulate back-end controls that are being used by
the MACT evaluation process. For
matter emission rates, the result the 12 percent of the sources (or used by
example, a source could reduce its ash
commenter advocates. This is because five sources in instances where there are
feed input by reducing the amount of
we evaluate sources with the best- fewer than 30 sources). We then look
silica in its feedstreams. This would not
performing (e.g. lowest emitting) only at those sources using MACT back-
result in feed control or emission
baghouses, and particulate matter end control and rank order all these
reductions of metal HAP.98
Finally, hazardous waste ash feed emissions from baghouses are not sources first by back-end control type,
control levels do not significantly affect significantly affected by inlet particulate and second by emissions. For example,
particulate matter emissions from matter loadings. Where the pool of best in instances where there is more than
cement kilns, lightweight aggregate performing sources includes sources one MACT back-end control, we array
kilns, and solid fuel-fired boilers operating some other type of back-end the emissions from the sources
because the majority of particulate control device (because insufficient equipped with the top ranked back-end
matter that is emitted originates from numbers of sources are equipped with controls from best to worst (i.e., lowest
the raw material and nonhazardous fuel. baghouses to comprise 12% of sources, to highest), followed by the emissions
Hazardous waste ash feed control levels or five sources (depending on the size from sources equipped with the second
also do not significantly affect of the source category)), we again use ranked back-end controls from best to
particulate matter emissions from the lowest particulate matter emission worst, and so on. We then determine the
sources equipped with baghouses level from the sources equipped with appropriate number of sources to
because these control devices are not second best technology. Although these represent 12 percent of the source
sensitive to particulate matter inlet data do not reflect test-to-test variability, category (5 in instances where there are
loadings.99 they are the best remaining data in fewer than 30 sources). If 10 sources
Thus, even if one were to conclude EPA’s possession to estimate represented 12% of the sources in the
that the nonenumerated metal HAPs are performance and EPA is therefore, as source category, we would then select
amenable to hazardous waste feed required by section 112 (d) (3) (A) and the emissions from best ranked 10
control, explicit use of ash feed control (B), using the data to fill out the sources in accordance with this ranking
in a MACT methodology would not requisite percentage of sources for procedure to calculate the MACT floor.
assure that each source’s ability to calculating floors. This methodology results in selection of
control either nonmercury metal HAP or Comment: Commenter states that EPA lowest emitters using best back-end air
surrogate particulate matter emissions is has failed to demonstrate how it pollution control as pool of the best
assessed. The Air Pollution Control reasonably estimated the actual performing sources.
Device methodology identifies and performance of each source’s end-of- The commenter is correct that there
assesses (with the surrogate particulate stack control technology because: (1) It can be differences between the
matter standard) the known technology failed to acknowledge that there can be performance of different models of the
that always assures metal HAP substantial differences between the same type of technology. We are not
emissions are being controlled to MACT performance of different models of the capable of thoroughly assessing
levels—that technology being back-end same type of technology; and (2) it did differences in designs of each air
control. not explain or support its rankings of pollution control device in a manner
Comment: Commenter claims the Air pollution control devices. that could be used in the MACT
Pollution Control Device approach to Response: As discussed in sections evaluation process, so that we would
calculate particulate matter floors is 7.4 and 16.2 of volume III of the only select, for example, baghouses of a
flawed because the performance of back- technical support document and C.1 of certain type. Each baghouse, for
end control technology alone does not this comment response section, we rank example, will be designed differently
reflect the performance of the relevant associated back-end air pollution and thus will have different
best sources that otherwise would be control device classes (e.g., baghouses, combinations of design aspects that may
reflected if EPA were to assess electrostatic precipitators, etc.), after
or may not make that baghouse better
performance based on the emission assessing particulate matter control
than other baghouses (e.g., bag types, air
efficiencies from hazardous waste
to cloth ratios, control mechanisms to
98 For the same reason, even if feed control of combustors that are equipped with the
collect accumulated filter cake and
total inputs (i.e. raw material and fossil fuel as well associated back-end control class. The
maintain optimum pressure drops). We
as hazardous waste fuel) were feasible, it would be data used to make this assessment are
technically inappropriate to use ash feedrates as a also do not have detailed design
included in our database. We also
surrogate: ash feed control allows sources to information for each source’s air
selectively reduce the ash feeds without reducing evaluated particulate matter control
pollution control system; such an
the metal HAP portion of that feed. Back-end efficiencies from other similar source
assessment would therefore not be
control, in contrast, unselectively removes a categories that also use these types of
percentage of everything that is fed to the control systems, such as municipal
combustor. 100 See USEPA, ‘‘Technical Support Document for
99 See USEPA, ‘‘Technical Support Document for waste combustors, medical waste th HWC MACT Standards, Volume III: Selection of
the HWC MACT Standards, Volume III: Selection of incinerators, sewage sludge combustors, MACT Standards,’’ September 2005, Section 5.3
Mact Standards,’’ September 2005, Section 3.1. coal-fired boilers, oil fired boilers, non- and 16.2, for further discussion.
possible even if the information could that ‘‘could not necessarily be achieved standards were based on levels of
be used to assess relative performance. by sources using the chosen end-of- chlorine in feedstock to these units, less
We instead account for this difference stack technology,’’ citing 69 FR at product would be produced since there
by selecting sources with the lowest 21228. Commenter claims that it is would be less chlorine to recover. EPA
emissions that are using the defined settled law that standards do not have has instead reasonably chosen to
MACT back-end controls to differentiate to be achievable through the use of any evaluate best performing/best controlled
the performance among those sources given control technology, and that it is sources for this source category by
that are using that technology (the best also erroneous to establish floors at measuring the efficiency of the entire
performer being the source with the levels thought to be achievable rather chlorine emission reduction system.
lowest emissions, as just explained). For than levels sources actually achieve. Indeed, the situation here is similar to
example, in situations where more than Response: EPA is not establishing that in Mossville, where polyvinyl
12% of the sources are using the single floor levels based on assuring the chloride production units fed raw
best control technology (e.g., more than standards are achievable by a particular materials containing varying amounts of
12% of incinerators use baghouses to type of end-of-stack technology (or, for vinyl chloride depending on the
control particulate matter), we use the that matter, any end-of-stack product being produced. This led to
emissions from the lowest emitting technology). The floor levels in today’s variable levels of vinyl chloride in plant
sources equipped with baghouses to final rule reasonably estimate average emissions. Rather than holding that EPA
calculate the MACT floor. In instances performance of the requisite percent of must base a floor standard reflecting the
where there are two defined MACT best performing sources without regard lowest amount of vinyl chloride being
technologies (i.e., 12% of sources do not for whether the levels themselves can be fed to these units, the court upheld a
use the single best control technology), achieved by a particular means. Floor standard estimating the amount of
we use all the emissions data from standards for particulate matter are pollution control achievable with back-
sources equipped with the best ranked based on the performance of those end control. 370 F. 3d at 1240, 1243. In
control class, and then subsequently use sources with the lowest emissions using the present case, as in Mossville, the
only the lowest emissions from the the best back-end control technology standard is based on actual performance
sources equipped with the second (most often baghouses, and sometimes of back-end pollution control (although
ranked back-end controls. electrostatic precipitators). EPA uses here EPA is assessing actual
Comment: EPA did not say how it this approach not to assure that the performance of the control technology
picked the best performers if more than floors are achievable by sources using rather than estimating performance by
twelve percent used the chosen these control devices, but to best use of a regulatory limit, making the
technologies. If EPA used emissions estimate performance of the best situation here a fortiorari from that in
data to differentiate performance, the performing sources, including these Mossville), and does not reflect
Agency is necessarily acknowledging sources’ variability. ‘‘emission variations not related to
that emissions data are a valid measure technological performance’’. 370 F. 3d
of sources’ performance—in which case 2. Total Chlorine Standard for
Hydrochloric Acid Production Furnaces at 1240.
the Agency’s claims to the contrary are It also should be evident that EPA is
arbitrary and capricious. We are adopting the methodology we not establishing a standard to assure its
Response: We did use emissions data proposed to estimate floor levels for achievability by a type of pollution
to select the pool of best performers total chlorine from hydrochloric acid control technology, as the commenter
where over 12% use the best type of production furnaces. 69 FR at 21225– mistakenly asserts. The standard for
emissions control technology, as 226. As stated there, we are defining total chlorine is based on the average of
explained in the previous response. best performers as those sources with the best five sources ‘‘ best meaning
Emissions data is obviously one means the best total chlorine system removal those sources with greatest (most
of measuring performance. EPA’s efficiency. We are not assessing a level efficient) system removal efficiencies.
position is that it need not be the of control attributable to control of EPA did not, as in CKRC, establish the
exclusive means, in part because doing chlorine in feedstocks because this standard using the highest emission
so leads to arbitrary results in certain would simply prevent these furnaces limit achieved by a source operating a
situations. Our use of emission levels to from producing their ultimate product. particular type of control.
rank sources that use the best Further details are presented in Comment: The commenter generally
particulate matter control (i.e., responses below. maintains that EPA’s methodology to
baghouses) does not lead to arbitrary Comment: Basing the standard for determine total chlorine floors for
results, however. First, we are assessing hydrochloric acid production furnaces hydrochloric acid production furnaces
emission levels here as a means of on the basis of system removal fails to capture other means of HAP
differentiating sources using a known efficiency rather than chlorine emission emission control that otherwise would
type of pollution control technology. reduction is impermissible. Even though be captured if EPA were assess
More importantly, the adjusted emission these devices’ purpose is to produce performance based on the emission
levels from sources equipped with chlorinated product, the furnaces can levels each source achieved.
baghouses are the most accurate use less chlorinated inputs. EPA’s Response: As discussed above, the
measures of performance because these proposed approach is surreptitious, an standard for total chlorine is based on
emissions have been statistically impermissible attempt to assure that the the sources with the best system
adjusted to accurately account for long- standards are achievable by all sources removal efficiencies. System removal
term variability through application of using EPA’s chosen technology, the efficiency encompasses all means of
the universal variability factor. approach already rejected in CKRC. MACT floor control when assessing
Comment: Commenter states that Response: EPA disagrees. There is relative performance because: (1)
EPA, in its support for its Air Pollution nothing in the text of the statute that Chlorine feed control is not a MACT
Control Technology Approach used to compels an approach that forces sources floor technology for these sources; and
calculate particulate matter floors, to produce less product to achieve a (2) the measure of system removal
claims that an emissions-based MACT floor standard. Yet this is the efficiency accounts for every other
approach would result in floor levels consequence of the comment. If controllable factor that can affect
emissions (e.g., operating practices, See Weyerhaeuser v. Costle, 590 F. 2d BTUs hazardous waste but emitting 101
worker training, proper maintenance, 1011, 1059 (D.C. Cir. 1978) (technology- pounds overall.
pollution control device type, etc). based standards are typically expressed The approach also is reasonable. First,
in terms of volume of pollutants emitted as with other standards expressed in
D. Format of Standards normalized terms, by normalizing the
per volume of some type of unit of
1. Thermal Emissions production). emissions standard we ensure the same
EPA proposed, and is finalizing There is no legal bar to this approach percentage of emission reduction per
standards for HAP metals and chlorine since the statute does not directly unit of raw material processed by each
(the HAPs amenable to hazardous waste address the question of whether a source, thus allowing meaningful
feed control) emitted by energy recovery source emitting 100 units of HAP per comparison among sources. For
units (cement kilns, lightweight unit of production but 100 units of HAP example, emission concentration-based
aggregate kilns, and liquid fuel boilers) overall is a better performer (or, for new standards normalize the size of each
expressed in terms of pounds of HAP sources, better controlled) than a source source by accounting for volumetric gas
attributable to the hazardous waste fuel emitting 10 units of HAP per unit of flowrate, which is directly tied to the
per million british thermal units (BTUs) production but emitting 101 units amount of raw material each source
of hazardous waste fired. 69 FR at overall.102 One commenter appeared to processes (and subsequently to the
21219–20. EPA received many suggest that we should assess amount of product that is produced),
comments on this issue to which we performance on mass feedrates and and assures equal levels of control per
respond below and in the Response to mass emission rates, without amount of product. Normalization on
Comment Document. Some initial normalizing. Such an approach would the basis of HAP amount in hazardous
discussion of the issue is appropriate, yield nonsensical results because the waste per BTU level in the hazardous
however. best performing sources would more waste similarly assures equal levels of
a. Expressing Standards in Terms of a likely be the smallest sources in the control across sources per amount of
Normalizing Parameter is Reasonable. source category (smaller sources raw material that is processed. Here, the
First, using a thermal emissions form of generally have lower mass emission raw material is the hazardous waste
a standard is an example of expressing rates because they process less fuel, expressed as units of energy. It is
standards in terms of a normalizing hazardous waste). This would likely reasonable to regard a hazardous waste
parameter. EPA routinely normalizes yield emission standards that would not fuel as a raw material to an energy
emission standards either by expressing be achievable by the larger sources that recovery device. Indeed, fuels are the
them as stack HAP concentrations or by more likely are better controlled sources only input to boilers, so fuels are
expressing the standards in units of based on a HAP removal efficiency necessarily such units’ sole raw
allowable mass emissions per amount of basis.103 Normalization by unit of material.104 105 Hazardous waste burning
production or raw material processed. production is another way of expressing cement kilns and lightweight aggregate
Emission concentration-based standards unit size, so that normalizing on this kilns produce a product in addition to
normalize the size of each source by basis is a reasonable alternative to recovered energy and so process other
accounting for volumetric gas flowrate, subcategorization on a plant size-by- raw materials. However, the reason
which is directly tied to the amount of plant size basis. See section 112(d)(1) these units use hazardous waste as
raw material each source processes (and (size is an enumerated basis for inputs is typically to recover usable
subsequently the amount of product that subcategorizing). energy from the wastes. Hence, the
is produced). Metal and particulate b. Using Hazardous Waste Thermal hazardous waste fuel is reasonably
matter emission standards for Input as the Normalizing Parameter is viewed as a raw material to these
commercial and industrial solid waste Permissible and Reasonable. devices.
incinerators are expressed in emission Normalization of standards based on In this regard, we note that our choice
concentration format. See § 60.2105. thermal input is analogous. For energy of normalizing parameter essentially
The particulate matter standard for recovery units (in this rule, kilns and says that best performers with respect to
Portland cement kilns is expressed as most liquid fuel boilers), normalizing on hazardous waste fuel burned in energy
mass of allowable emissions per mass of the basis of thermal input uses a key recovery units are those using the
raw material processed. See § 63.1342. feed input as the normalizing parameter, lowest HAP feedrate (for metals and
The particulate matter, mercury, and allowing comparison of units with chlorine) per amount of energy
hydrogen chloride standards for different inputs rather than separately
nonhazardous waste industrial boilers evaluating these units by size and type 104 EPA thus has expressed the MACT standards
are expressed as pounds of allowable (see section 112(d)(1)). Again, this for particulate matter, mercury, and hydrogen
approach is legally permissible. The chloride standards for nonhazardous waste
emissions per million British thermal industrial boilers as pounds of allowable emissions
units (BTUs). See § 63.7500. statute does not answer the question of per million BTUs. § See 63.7500. This
Technology-based standards typically which source is better performing, the normalization considers the total heat input into the
normalize emissions because such a source emitting 100 pounds of HAP per combustion device. Normalizing by total heat input
million BTUs hazardous waste but 100 would not be appropriate for hazardous waste
format assures equal levels of control combustors for metals and chlorine because this
across sources per amount of raw pounds of HAP overall or the source would implicitly account for, and in turn require
material that is processed, and allows emitting 10 pounds of HAP per million the use of, feed control of HAP in non hazardous
waste fuels. This is inappropriate for the reasons
EPA to equally assess source categories discussed in Section III.B of this Part.
innings pitched in order to make comparisons
that comprise units that differ in size. among pitchers possible. 105We distinguish (i.e., subcategorize) liquid fuel
By normalizing the emissions standard 102 Or, put another way, the statute does not boilers that process hazardous waste with heating
we better ensure the same percentage of directly address the question of whether a small values less than 10,000 BTU/lb from those
emission reduction per unit of raw source that emits 10 units of HAP is better than a processing hazardous wastes with heating content
much larger source with better back-end control greater than 10,000 BTU/lb. Although boilers that
material processed by each source.101 (but feeding the same raw material at a higher mass process hazardous waste with heating values less
feedrates) that emits 100 units of HAP. than 10,000 BTU/lb are still considered to be energy
101 A more familiar example of normalization is 103 See USEPA, ‘‘Technical Support Document for recovery units, we conclude a thermal emissions
the Earned Run Average (ERA), which normalizes the HWC MACT Standards, Volume III: Selection of normalization approach for these sources is not
a baseball pitchers’ earned runs on the basis of nine MACT Standards,’’ September 2005, Section 6.0. appropriate. See Part Four, Section VI.D.
recovered.106 This approach accords devices to achieve a standard, without performance of the best performing
well with the requirement in section decreasing concentrations of HAP in sources (on a mass of HAP per million
112(d)(2) that EPA take energy their hazardous waste fuels, by diluting BTU basis). Such a requirement assures
considerations into account in the HAP contribution of hazardous that these sources are processing the
developing MACT, and also that the waste with emissions from fossil fuel. A cleanest hazardous waste fuels to
Agency consider front-end means of thermal emissions format prevents this recover energy and are reducing HAP
control such as input substitution type of dilution from happening because emissions to MACT levels.
(section 112(d)(2)(A)). In addition, our it ignores additions of stack gases We note that it would not be
choice furthers the RCRA goal of attributable to burning fossil fuels. appropriate to express the emission
encouraging properly conducted Weyerhaeuser, 590 F. 2d at 1059 (use of standards for incinerators, hydrochloric
recycling and reuse (RCRA section production of a unit as a normalizing acid production furnaces, and solid fuel
1003(b)(6)), which is of relevance here parameter serves ‘‘the commendable boilers in terms of thermal emissions.
in that Congress directed EPA to purpose’’ of preventing plants from As just explained, the choice of a
consider the RCRA emission controls for achieving emission limitations via normalizing parameter is fitted to the
hazardous waste combustion units in dilution). nature of the device to which it is
developing MACT standards for these For example, assume there are two applied in order to allow the most
units, and to ensure ‘‘to the maximum identical energy recovery units with meaningful comparisons between
extent possible, and consistent with identical back-end control devices (that devices of like type. We therefore
[section 112 ]’’ that section 112 reflect the performance of the average of conclude that a thermal emissions
standards are ‘‘consistent’’ with the the best performing sources). Source A format (i.e., normalizing parameter) for
RCRA scheme. CAA section fulfills 25% of its energy demand from incinerators is not appropriate because
112(n)(7).107 Conversely, emission the combustion of hazardous waste; the primary function of incinerators is
concentration-based standards, the source B fulfills 50% of its energy to thermally treat hazardous waste (as
methodology that otherwise would be demand from the combustion of opposed to recovering energy from the
used to calculate emission hazardous waste. Also assume that the hazardous waste). See 67 FR at 17362
concentration-based standards, may hazardous waste for these two sources (April 19, 1996). Our database indicates
result in standards that are biased have equivalent energy contents. If these that most incinerators processed
against sources that recover more energy sources were required to comply with hazardous waste during their emissions
from hazardous waste. This may an emission concentration based- tests that had, on average, heating
discourage sources from recovering standard (e.g., µg/dscm), source A values below 10,000 BTU/lb.109 We
energy from hazardous waste because would be allowed to feed hazardous have emission test hazardous waste
such standards do not normalize each waste containing twice the metal heating value information for 62
source’s allowable emissions based on content (on a mass concentration basis, incinerators in our database. Of these 62
the amount of hazardous waste it e.g., ppm), and would be allowed to sources, 40 sources processed hazardous
processes for energy recovery purposes. emit metal HAP at the same mass waste with an average heating value of
See 69 FR at 21219 and responses emission rate relative to source B. This less than 10,000 BTU/lb. The other 22
below. is because this source is effectively sources processed hazardous waste with
Second, use of this normalizing diluting its emissions with the heating values greater than 10,000 BTU/
parameter makes it much more likely emissions that are being generated by lb in at least one test condition,
that hazardous waste feed controls will the fossil fuels.108 A thermal emissions although we note that 14 of these 22
be utilized by these devices as an aspect standard format does not allow sources sources also processed hazardous waste
of emissions control. See section to dilute their emissions with the in different test conditions with heating
112(d)(2)(A) (use of measures reducing emissions from fossil fuel inputs values lower than 10,000 BTU/lb.110
the volume of pollutants emitted because it directly regulates the We assessed whether we should
through ‘‘substitution of materials’’); emissions and feeds associated with the subcategorize incinerators, similar to
CKRC, 255 F. 3d at 865 (EPA to consider hazardous waste fuel. Under a thermal how we subcategorize liquid fuel
means of control in addition to back-end emissions format both sources would be boilers, based on the BTU content of the
pollution control technology when required to feed hazardous waste with hazardous waste. Incinerators do
establishing MACT floors). As explained the same thermal feed concentrations recover energy from processing high
in our discussion of the SRE/Feed (on a lb HAP per million BTU BTU wastes. Some incinerators are
methodology, the MACT floor level for hazardous waste basis), and source A equipped with waste heat boilers, and
metals and chlorine reflects the best would be required to process hazardous high BTU hazardous waste can displace
combination of hazardous waste waste with an equivalent concentration fossil fuels that otherwise would have to
feedrate, and total HAP removal of metal HAP (on a mass basis) and also be burned to thermally treat low BTU
efficiency. See section III.B. However, if be required to emit half as much metal wastestreams. However, such energy
standards for energy recovery units are HAP (on a mass emission rate basis) recovery is considered to be a secondary
expressed in terms of mass of HAP per relative to source B, because source A is product because their primary function
volume of stack gas, then it would be processing half as much hazardous is to thermally treat hazardous waste. A
relatively easy for these energy recovery waste fuel, thus vindicating the
109 As discussed later, the heating values of
hazardous waste feed control aspect of
106 As explained earlier, the ultimate ranking of hazardous wastes processed at cement kiln and
the standard (see also note below lightweight aggregate kilns are primarily 10,000
best performers then further evaluates system
removal efficiency, best performers then being
regarding the likelihood of sources BTU/lb or greater.
defined in terms of the combination of hazardous using hazardous waste feed control). 110 These data are based on a compilation of
waste thermal feed and system removal efficiency. Further, the thermal feed concentration heating contents for every incinerator test condition
See USEPA, ‘‘Technical Support Document for the with which these sources must comply in the database where the source reported such
HWC MACT Standards, Volume III: Selection of heating content, and include both the most recent
MACT Standards’’, September 2005, Section 7.3. reflects the feed control of the average test conditions as well as older test conditions.
107 EPA would adopt the thermal format for the Incinerator test condition heating values range from
standards, however, whether or not the approach 108 This example assumes there are no HAP a low of 790 to a high of 19,800 BTU/lb, with a
furthered RCRA objectives. emissions attributable to the fossil fuels. median value of 7800 BTU/lb.
thermal emissions normalization energy value.112 No commenters constant per the normalizing unit. As
approach for incinerators that combust suggested that we apply a thermal explained in the introduction to this
hazardous wastes with heating values emissions format to hydrochloric acid section, this approach is both routine
greater than 10,000 BTU/lb would production furnaces. and permissible.
therefore not be appropriate because the We consider the processing of Cement kilns, lightweight aggregate
normalized parameter would not be tied hazardous waste in solid fuel boilers to kilns, and liquid fuel boilers combust
to the primary production output that be more reflective of energy recovery hazardous waste to recover valuable
results from the processing of hazardous (relative to incinerators and energy. Recovering energy is an integral
waste (i.e., treated hazardous waste). In hydrochloric acid production furnaces) part of their production process. As
confirmation, no commenters suggested because these sources directly recover discussed at proposal, emission
that we apply a thermal emissions the heat that is released from the concentration-based standards (and the
format to incinerators. combustion of the waste streams. methodology that otherwise would be
We also conclude that a thermal However, as stated at proposal, not all used to calculate emission
emission format is inappropriate for these sources are processing hazardous concentration-based standards) may
hydrochloric acid production furnaces. wastes for energy recovery. 69 FR at result in standards that are biased
These devices recover chlorine, an 21220. These boilers are generally not against sources that recover more energy
essential raw material in the process, commercial units, and so tend to burn from hazardous waste. 69 FR at 21219.
from hazardous waste. The classic whatever hazardous wastes are This may discourage sources from
normalizing parameter of amount of generated at the facility where they are recovering energy from hazardous waste
product (HCl) produced is therefore the located. Heating values for this source because such standards do not
obvious normalizing parameter for these category range from 1,300 to 10,500 normalize each source’s allowable
sources. It is true that some BTU/lb, with a median value of 8,000 emissions based on the amount of
hydrochloric acid production furnaces BTU/lb. We therefore conclude that hazardous waste it processes for energy
recover energy from high BTU thermal emission standards for these recovery purposes. A source that fulfills
hazardous wastes. See 56 FR at 7141/1 sources are not appropriate because 100 percent of its energy demand from
and 7141–42 (Feb. 21, 1991). Some most of these sources are processing hazardous waste would be required to
sources are equipped with waste heat hazardous waste with energy content limit its mass HAP emissions to the
boilers, and high BTU wastes help lower than 10,000 BTU/lb. As discussed same levels as an identical source that
sustain the combustion process, which in section VI.D, we conclude that 10,000 satisfies, for example, only 10 percent of
is necessary to liberate the chlorine from BTU/lb is an appropriate level that its energy demand from hazardous
the wastestreams prior to recovering the distinguishes whether thermal emission waste and 90% from coal. This would
chlorine in the scrubbing systems. standards or mass emission inappropriately discourage the safe
Again, energy recovery is not the concentration-based standards are recovery of energy from hazardous
primary function of these types of appropriate. We also note that no waste, and could in turn result in
sources.111 Hydrochloric acid commenters suggested that we apply a greater consumption of valuable fossil
production furnace hazardous waste thermal emissions format to solid fuel fuels that otherwise would be
heating values range from 1,100 to boilers. consumed.
11,000 BTU/lb (the median energy Comment: Commenters state that Sources which fulfill a greater
content for these sources is slightly thermal emission standards are percentage of their energy demand from
above 6,000 BTU/lb). The range of inappropriate because sources burning hazardous waste (either by processing
hazardous waste heating contents from hazardous waste with a higher energy hazardous wastes that are higher in
these sources is much lower than the content or higher percent hazardous energy content, or by simply processing
ranges for cement kilns, lightweight waste firing rate (i.e., one that fulfills a more hazardous waste) will be allowed
aggregate kilns, and liquid fuel boilers, greater percentage of its total energy to emit more HAP (on a mass emission
supporting the premise that energy demand from the hazardous waste) rate basis) than an identical source that
recovery is of secondary importance. In would be allowed to emit more HAP. satisfies less of its total energy demand
Response: Part of this comment would from hazardous waste. This is
addition, and critically, the hazardous
apply regardless of what normalizing appropriate because: (1) The source
waste that is processed in these units
parameter is used. Technology-based fulfilling a greater percentage of its
contains high concentrations of
standards (including MACT standards) energy demand from hazardous waste is
chlorine, confirming that the wastes are almost always expressed in terms of
serve as feedstock for hydrochloric acid processing more raw material than the
some type of normalizing parameter, other source (the raw material being the
production, even if the wastes also have i.e., ‘‘X’’ amount of HAP may be emitted energy content of the waste); and (2)
111 EPA notes that when first adopting RCRA air
per unit of normalizing parameter. This The source fulfilling a lower percentage
emission standards for hydrochloric acid recovery allows a meaningful comparison of its energy demand requirements from
furnaces (then called ‘halogen acid furnaces’), EPA between units of different size and hazardous waste would not be allowed
indicated that those furnaces designed as boilers production capacity. A consequence is to dilute its emissions with
would be subject to the emission standards for that the overall mass of HAP emissions
boilers. 56 FR at 7040. This determination did not nonhazardous waste fuels, and we
have regulatory consequence, since all hydrochloric varies, but the rate of control remains would thus assure that all sources
acid production furnaces were subject to the same implement hazardous waste feed control
emission standards whether they were classified as 112 Hazardous waste chlorine feedrates that are
boilers or as industrial furnaces. Thus, EPA was not included in our database (expressed as MTECs) to levels consistent with MACT.113 This
concluding that some hydrochloric acid furnaces range from a low of 46,000,000 µg/dscm to a high
existed for the primary purpose of recovering of 294,000,000 µg/dscm. On a mass chlorine 113 Although the rule does not require use of feed
energy in the 1991 rulemaking. 56 FR at 7139 percentage basis, these wastes range from 17% to control (or any particular means of control to
(‘‘[Hydrochloric acid recovery furnaces] are 82%, noting that these percentages did not include achieve a standard), the rule assures that all
typically modified firetube boilers that process the chlorine that was also spiked during the sources’ emissions will reflect the emissions of the
secondary waste streams containing 20 to 70 per emissions tests). See USEPA, ‘‘Technical Support sources with the best hazardous waste federates
cent chlorine or bromine to produce a halogen acid Document for the HWC MACT Standards, Volume expressed in terms of amount of HAP per BTU of
product by scrubbing acid from the combustion III: Selection of MACT Standards’’, September 2005, hazardous waste. Because this format eliminates
gases’’). Section 15. Continued
was illustrated in the example provided performance demonstrated by the best energy content is overstated for these
in the introduction to this comment performing sources); and (3) source D is sources.
response section. required to have lower mass Energy content of hazardous wastes
Similarly, two sources that combust concentrations of metals in its processed in liquid fuel boilers and
hazardous waste with the same energy hazardous waste because it is firing lightweight aggregate kilns varies more
content and the same metal poorer quality hazardous waste fuel than energy content of hazardous wastes
concentrations (on both a thermal (from an energy recovery perspective) processed by cement kilns, and sources
concentration and mass-based and because it is feeding less of the with higher energy content wastes
concentration basis), but at different same raw material (measured by energy would be allowed to emit more metals
hazardous waste firing rates, would be content). Thus, the thermal emissions than identical sources burning identical
required to achieve identical back-end format appropriately encourages and volumes of lower energy content wastes
control device operating efficiencies to promotes the processing of clean, high (although the degree of control is
comply with a thermal emissions-based identical per BTU of hazardous waste
energy content hazardous waste fuels
standard. Holding these factors fuel processed).115 Again, these are
(consistent with evaluating hazardous
constant, thermal emission standards hypothetical examples. Each energy
waste feed control as an aspect of
require sources to achieve identical recovery unit will have an upper bound
percent reductions of the HAP that is MACT, and not just relying on control on the amount of energy it can process
processed within the combustor via solely through use of back end from the hazardous waste. Sources that
removal with an air pollution control technology), and does so equally for all process higher energy content
device. A thermal emission standard sources because it normalizes the hazardous wastes would not necessarily
format is thus equally stringent for these allowable emissions based on the feed the same volume of hazardous
sources on a percent HAP removal basis, amount of energy each source recovers waste as compared to sources
irrespective of the amount of hazardous from the hazardous waste. Put another processing lower energy content
waste it processes for energy recovery, way, source C in the above example is hazardous wastes because they cannot
and better assures that sources burning controlling HAP emissions to the same exceed the thermal capacity of their
smaller amounts of hazardous waste extent as the average of the best combustion unit. Under a thermal
(from an energy recovery perspective) performing sources per every BTU of emission standard format, the mass
are also controlling emissions as well as hazardous waste fuel it processes (as is emission rates that would be allowed for
the average of the best performing source D). identical sources that fulfill 100 percent
sources. We note that this is a hypothetical of their energy demand from hazardous
Sources processing higher energy example. In practice the average energy waste and that have differing hazardous
content hazardous wastes would be content of hazardous waste processed at waste energy contents would be
allowed to feed hazardous wastes with cement kilns does not vary significantly identical. Although the source with the
higher metal and chlorine mass-based across sources. Cement kilns burn higher energy content hazardous waste
concentrations relative to other sources would have a higher allowable mass-
hazardous wastes with relatively
combusting lower energy content based hazardous waste feed
consistent energy contents because that
wastes. To illustrate this, assume there concentration, this source would have
is what their production process
are two sources (named C and D) with to process less hazardous waste (on a
identical back-end control systems and necessitates. This is supported by our
mass basis) to remain within its thermal
identical mass feedrates of hazardous database and by comments received
capacity. This helps to ensure that its
waste. Also assume the hazardous waste from the Cement Kiln Recycling mass HAP emission rate is similar to
of source C has twice the energy content Coalition.114 Heating values of other sources that process lower energy
as compared to the hazardous waste hazardous wastes processed at cement content hazardous waste.
processed by source D. A thermal kilns during compliance tests One commenter’s apparent concern
emission standard will allow Source C (information which is included in our with thermal emissions seems to center
to feed a hazardous waste that has twice database) range from 10,300 to 17,600 on an assertion that sources will
the metals concentration (as measured BTU/lb, with a median value of 12,400 intentionally blend nonhazardous, high
on a mass basis) as compared to source BTU/lb. We note that these are snapshot heating value wastes or fuels with low
D, even though both sources would be representations of hazardous waste energy, high metal bearing hazardous
required to comply with equivalent heating content from these sources that wastes in order to increase the energy
thermal feed rates limitations. Notably, originate from compliance tests. We also content of these metal bearing wastes so
however: (1) Source C is displacing (i.e., have long term average hazardous waste that they will be subject to higher
not using) twice as much valuable fossil heating measurements from cement allowable emissions via thermal
fuel as the source with the lower energy kilns indicating that the heating content emission standards. We specifically
content hazardous waste, and is feeding of the hazardous wastes on average address that comment later as it relates
twice as much raw material—the raw range from 9,900 to 12,200 BTU/lb, with to commercial energy recovery units
material being energy content contained a median value of 11, 500 BTU/lb. We (lightweight aggregate kilns and cement
in the hazardous waste; (2) source C thus conclude that the commenter’s kilns). We note here, however, that we
cannot exceed the feed control levels concern regarding sources being do not consider that comment to be of
(expressed on a lbs of HAP per million allowed to emit more HAP if they practical concern for liquid fuel boilers
BTU basis) that was achieved by the process hazardous waste with higher
average of the best performing sources 115 The hazardous waste heating values of liquid
(assuming its back-end control fuel boilers range from 2,200 to 21,000 BTU/lb, with
114 See comment submitted by the Cement Kiln
a median value of 14,800. Heating values of
efficiency is equivalent to the average Recycling Coalition, USEPA, ‘‘Comment Response lightweight aggregate kilns range from 4,900 to
Document to the Proposed HWC MACT Standards, 16,900 BTU/lb, with a median value of 14,800. We
consideration of stack gas attributable to fossil fuel Volume 1: MACT Standards,’’ September 2005, note that the low end heating value for lightweight
emissions, and thus eliminates the dilutive effect of Section 3.3. Also see USEPA, ‘‘Technical Support aggregate kilns reflects one source and is not typical
these emissions, the likelihood that sources will in Document for the HWC MACT Standards, Volume of heating values used by the other commercial
fact use hazardous waste feed control as part of III: Selection of MACT Standards,’’ September 2005, lightweight aggregate kiln facilities, and are similar
their control strategy is great. Section 23. to the heating values of cement kilns.
because they do not engage in not related to our choice to use thermal reality that the hazardous waste fuels
commercial fuel blending practices. content of hazardous waste as a that are currently processed safely and
Comment: A commenter states that normalizing parameter. Rather, the issue efficiently in energy recovery units to
EPA’s assessment of thermal emissions is whether feed control of fossil fuels displace valuable fossil fuel do in fact
to identify the relevant best sources is and raw materials is a feasible means of contain metal HAP. From a feed control
inappropriate because thermal control at all. We have determined that perspective, the thermal emissions
emissions are not emission levels, but it is not, and that only back-end control format appropriately requires sources to
rather a ratio of emissions to the heat (expressed as system removal efficiency) process high energy content hazardous
content in a source’s hazardous waste. is feasible. Moreover, today’s rule waste fuels that reflect the thermal feed
Response: This comment challenges controls emissions from HAP in raw control levels achieved by the average of
the basic idea of normalization, since material and fossil fuels. All non- the best performing sources, and does so
the comment would be the same mercury metal HAP emissions equally for all sources because it
regardless of the normalizing parameter attributable to fossil fuels or raw normalizes the allowable emissions
being used. Thermal emissions are material are effectively and efficiently based on the amount of energy each
emission levels that are normalized to controlled to the level of the average of source recovers from the hazardous
account for the amount of energy (i.e., the best performing sources with the waste.
raw material) these sources recover by surrogate particulate matter standard, as Comment: A commenter states that
processing hazardous waste. Similarly, a well as the system removal efficiency EPA should be concerned that fuel
mass emission concentration (i.e., µg/ component of the SRE/Feed blenders and kilns will use the thermal
dscm) is a ratio of the emissions to the methodology. emission standard format to increase the
volume of combustion gas that is Comment: EPA has failed to allowable metals feedrates for their
generated, which normalize emissions document sources’ actual feedrates. units. The commenter claims that
to account for differences in the size of Feedrates are presented either as MTECs sources could inappropriately convert
the combustion units (as well as (where hazardous waste HAP feedrates non-hazardous waste fuel to hazardous
differences in production capacity). are divided by gas flow rates) or as waste fuel by simply putting coal in a
This rulemaking assesses performance thermal feedrates, (where feedrate is bunker in which hazardous waste was
and expresses emission standards in expressed as the mass of HAP per once stored, or mixing nonhazardous
both of these formats; both formats million BTUs of hazardous waste fired). waste fuel oil with hazardous waste.
normalize the emissions so that we may This is impermissible, since it does not The commenter states that a facility
better assess emission control measure actual feed levels. with a low hazardous waste firing rate,
efficiencies equally across sources based Response: This comment essentially and relatively low allowable emissions
on the percent of HAP in the feed takes the position that it is legally can become a facility with a high
(whether thermal feed or feed impermissible to normalize standards, hazardous waste percent firing rate,
normalized based on combustor size) 116 i.e., express standards on a common with higher allowable emissions, simply
that is controlled or removed from the basis. EPA rejects this comment for the by ‘creative’ use of the hazardous waste
stack gas prior to being emitted into the reasons stated in the introduction to this mixture rule. The commenter suggests
atmosphere. As discussed above, section. that EPA clearly state that the hazardous
technology-based standards have Comment: A commenter states that an waste thermal emission standards apply
historically assessed performance after increasing number of fuel blenders are only to the hazardous waste portion of
normalizing emissions based on the producing fuels with a minimum the fuel blend mixture. The commenter
amount of raw material processed by the heating content and maximum metals further suggests that EPA require fuel
given industry sector. Thermal content in order to maximize revenues blenders to report the amount of
emissions normalize each source’s because high metal bearing wastes nonhazardous waste fuel that is
emissions based on the amount of raw command a higher revenue on the contained in the fuel blend, and that
material (hazardous waste fuel) it commercial waste market. The cement kilns use this to determine
processes, and are therefore appropriate commenter states that thermal emission allowable metal feed rates based on the
to assess and identify the relevant best standards are not appropriate because original hazardous waste energy
performers. Finally, as previously they are based on the implicit content.
explained, this approach is consistent assumption that energy recovery entails Response: We do not believe
with both the language of section 112 metals feed. hazardous waste combustors will engage
(d) (2) and (3), and the purpose of these Response: Contrary to what the in the practice of redesignating their
provisions. commenter suggests, the thermal fossil fuels, i.e., coal, as hazardous
Comment: A commenter states that emissions format will more likely wastes with creative use of the mixture
EPA’s assessment of thermal emissions discourage the alleged practice of fuel rule in order to increase their allowable
to identify the relevant best sources is blenders producing fuels with a metal HAP emission rate. That would
inappropriate because it ignores HAP minimum heat content and maximum require large quantities of coal to be
emissions attributable to the metals content because the standard newly classified as hazardous waste.
nonhazardous fuel and raw material. limits the allowable metal emissions The coal, and the unit where the coal is
Response: Thermal emission based on the amount of energy stored, would subsequently become
standards do not directly control HAP contained in the hazardous waste. Thus, subject to all applicable subtitle C
emissions attributable to the fossil fuels a source with a lower energy waste requirements, which include storage
and raw material, in the sense that we would have to ensure that the mass and closure/post closure requirements.
did not assess feed control of fossil fuels concentration of metals is also lower to We believe this disincentive will
or raw materials. However, this issue is comply with the thermal emission discourage this hypothetical practice.
formatted standard. The source would Moreover, as previously discussed,
116 For emission concentration-based standards
consequently emit less metals (on a today’s rule does not allow cement kiln
we normalize hazardous waste feed control levels
by calculating what we call maximum theoretical
mass basis) because of the lower metal or lightweight aggregate kiln emissions
emission concentrations, which are equivalent to mass concentration in the waste fuel. to exceed the interim standards. The
the HAP mass feed rate divided by gas flow rate. Thermal emission standards reflect the fact that we are issuing emission
standards for some pollutants in the difference, this would not result in an that are mixtures of solids and liquids
thermal emissions standard format will arbitrary increase of allowable tend be biased high, which would
not encourage fuel blenders to send hazardous waste metals emissions. The inappropriately give these sources
more metals to these commercial energy thermal emission standards were higher allowable metal emission
recovery sources because their calculated using thermal emissions data limitation.
allowable emission concentrations are, that are based on each source’s Response: There are standard ASTM
by definition, either equivalent to or compliance test. These tests were procedures that reliably measure the
more stringent than the current conducted at hazardous waste feed energy content of the hazardous waste.
limitations with which they are control levels that represented the upper Any parameter that is measured for
complying. Thus, even if the fuel bound of feed control levels these compliance purposes is subject to
blenders and energy recovery units sources see on a day-to-day basis. To method imprecision and variability. We
engaged in this practice, they could not accomplish this, sources spiked metals do not believe that hazardous waste
emit more metals than they are into the hazardous waste prior to energy content measurements result in
currently allowed to emit. We therefore combusting the wastes. The amount of imprecision and variability above and
conclude that it is not necessary to metals that were contained in the beyond the measurement methods that
promulgate complicated regulatory hazardous waste streams, after are currently used to assure compliance
provisions that would increase the accounting for these spiked metals, far with emission concentration-based
reporting and recordkeeping exceeded the metal levels that were standards.
requirements of fuel blenders and contained in the raw material. Thus the The commenter did not provide
energy recovery units in order to differences in partitioning, if any, would evidence that supports the claim that
address a hypothetical scenario that likely be overshadowed by the fact that energy content measurement and/or
likely would never occur. the majority of the metals were sampling methods consistently result in
Finally, we note that combustion of contained in the hazardous waste. a positive bias. If a bias were
certain high HAP metal content wastes Notably, any partitioning bias that consistently present for these types of
is already prohibited under RCRA rules. that may be present would also have wastes, then one would expect it to be
See 40 CFR 268.3. Such wastes remain been present during these compliance also reflected in the measured data for
prohibited from combustion even if they tests. As a result, this potential bias which we based the emission standards,
are mixed with fossil fuel so that the would be built into the emission which would fully address the
mixture has a higher energy content. standard and thus would not result in commenter’s concern. Nonetheless, we
U.S. v. Marine Shale Processors, 81 F. an arbitrary increase in allowable note that all hazardous waste sampling
3d 1361, 1366 (5th Cir. 1996) (an hazardous waste metals emissions and analysis procedures must be
unrecyclable hazardous waste is not because these sources will again prescribed in each source’s feedstream
recycled when it is mixed with a usable demonstrate compliance under testing analysis plan, which can be reviewed by
non-waste and the mixture is conditions similar to those used to the permitting authority upon request.
processed). Thus, the dilution generate the data used to calculate the These feedstream analysis plans must
prohibition in § 268.3 serves as a further MACT floors. We conclude that it is not ensure that sampling and analysis
guard against the commenter’s concern. necessary to provide additional procedures are unbiased, precise, and
Comment: A commenter states that prescriptive regulatory language that that the results are representative of the
the thermal emissions format may be would require sources to demonstrate feedstream. See § 63.1208(b)(8). More
problematic because it is based on a system removal efficiencies under information on obtaining a
flawed assumption that metal HAP from testing conditions that exhibit a high representative samples can be found in
the cement kiln raw material and ratio of hazardous waste metal content EPA’s SW–846 publication.119 These
hazardous waste partition in equal to raw material metal content because procedures involve acquiring several
proportions to the total stack gas the regulations implicitly require sub-samples that provide integration
emissions. The commenter believes that sources to demonstrate hazardous waste over the breadth, depth and surface area
metal retention in the raw materials is metal feed control levels that represent of the waste container and obtaining
higher than the hazardous waste, the upper range of their allowable feed replicate samples (see Ch. 13.3.1 of SW–
suggesting that thermal emission control levels.118 846).
standards allow an arbitrary increase in Comment: A commenter states that Comment: A commenter states that
allowable hazardous waste metals compliance with standards expressed in BTU measurements can be reported as
emissions. The commenter suggests that a thermal emissions format is either a higher heating value or a lower
EPA require that compliance problematic because the measurement heating value, and suggests that EPA
demonstrations be conducted only of energy content of hazardous waste require sources to use the lower heating
under conditions where the metals fuel blends is subject to significant value calculation when determining
content in the hazardous waste is variability due to the nature of the test. allowable hazardous waste feed control
significantly higher than the metal The commenter also claims that heating levels. The commenter seems to imply
content in the raw material to minimize value measurements of waste streams that use of higher heating values will
this bias. inappropriately result in higher
Response: The commenter has not very point. See USEPA, ‘‘Comment Response allowable metal feed rates for fuel
provided any emissions data to support Document to the Proposed HWC MACT Standards, blends that contain aqueous waste.
Volume 1: MACT Standards,’’ September 2005,
this claim, nor does the EPA know of Section 3.3. We have evaluated these comments and
Response: The BTU data in our
data available that reaches this find them persuasive on this issue. database that we use to calculate the
conclusion. We do not believe there is 118 Although today’s final rule allows sources to emission standards reflect higher
a significant difference in the extrapolate their allowable hazardous waste feed heating values. It is standard practice in
partitioning rates of these metals in a control levels to levels that are higher than the level the incineration/combustion industry to
demonstrated in the comprehensive performance
cement kiln.117 Even if there is a test, sources must still spike metals into the report the gross heat of combustion (or
hazardous waste during the test in order to assure
117 We reference comments submitted by the that the system removal efficiency used for the 119 SW–846, ‘‘Test Methods for Evaluating Solid
cement kiln recycling coalition that address this extrapolation procedure is reliable and accurate. Waste, Physical/Chemical Methods.’’
higher heating value). We conclude that some other reason) is irrelevant in and 63.1205, respectively). As reflected
sources should use the higher heating determining levels of MACT floors. in the comment, making a judgment as
value rather than the lower heating National Lime v. EPA, 233 F. 3d at 640. to whether a replacement standard is
value for all compliance determinations What matters is the level of more stringent than the interim standard
because these are method-based performance, not what motivated that for the HAP is not always a straight-
emission standards. Fuel blends that level. forward calculation. As we discussed in
contain aqueous wastes will not be As a result, the replacement standards the proposed rule 122 and echoed by the
inappropriately rewarded with higher promulgated today ensure that sources commenter, comparing standards in the
allowable feed rates because any fuel will emit HAP at levels no higher than thermal emissions format to those in a
mixture that contain aqueous mixtures levels achieved under current mass concentration format involves
will have lower reported heating values, regulations. We do this in this rule, assumptions that vary on a site-specific
irrespective of whether they are when necessary, by either capping a basis and can vary over time, including
reported as higher heating values or calculated floor level by the interim the hazardous waste fuel replacement
lower heating values.120 standard (when both the calculated floor rate, contributions to emissions from
level and interim standard are expressed nonhazardous waste inputs such as raw
E. Standards Can Be No Less Stringent in the same format of the standard) or
Than the Interim Standards materials and nonhazardous waste fuels
by adopting dual standards in cases such as coal, how close to the standard
Comment: Several commenters where formats of the standard vary (so a source elects to comply, the system
oppose EPA’s position in the proposed that comparison of stringency cannot be removal efficiency demonstrated during
rule that the replacement standards can uniformly determined (as for cement testing, and the type and composition,
be promulgated at a level no less kilns and lightweight aggregate kilns, as including heating value, of fuels burned.
stringent than the interim standards for explained in the preceding section
incinerators, cement kilns, and To ensure that sources operating
above and in the following response). In
lightweight aggregate kilns. In instances under standards expressed in a thermal
this case, the sources are subject to both
where the calculated replacement emissions format will not emit HAP
the replacement and interim standards.
standard is less stringent than the Comment: One commenter states that metals at levels higher than currently
interim standard, the commenters some proposed standards expressed in a achieved under the interim standards,
oppose EPA’s position of ‘‘capping’’ the thermal emissions format would allow we adopt a dual standard to prevent
replacement standard at the level of the some sources to emit semivolatile emissions increasing to levels higher
interim standard to prevent backsliding metals at levels higher than the interim than the interim standards. The dual
from those levels. Instead, commenters standard. The commenter states that standard structure includes both the
recommend that EPA calculate and EPA reached incorrect conclusions standard expressed in a thermal
finalize the existing and new source when making relative stringency emissions format and the interim
floor levels without regard to the comparisons between standards standard, which is expressed in a mass
interim standards. One commenter also expressed in a thermal emissions and concentration format. We apply this
notes that the interim standards are mass concentrations format because, in concept to several standards including
simply a placeholder without the part, EPA assumed an average F-factor semivolatile metals, low volatile metals,
necessary statutory basis to qualify as (e.g., semivolatile metals for cement and mercury 123 for cement kilns and
emission limitations for purposes of kilns).121 In addition, the commenter semivolatile metals and low volatile
establishing MACT floors. Another notes that the actual relationship metals for lightweight aggregate kilns.
commenter, however, supports EPA’s between standards expressed in terms of This approach ensures that sources are
position to prevent backsliding to levels thermal emissions and mass not emitting HAP metals above the
less stringent than the interim concentrations is complex and depends levels of the interim standards because
standards. on a number of factors. As a result, the we cannot reliably determine that
Response: We maintain that the commenter urges EPA to adopt dual emissions under a standard expressed in
replacement standards can be no less standards (i.e., promulgate the MACT a thermal emissions format would not
stringent than existing standards, standard as both the standard expressed exceed the interim standard for all
including the interim standards under in a thermal emissions format and also sources in the category. See
§§ 63.1203–1205, for incinerators, the interim standard expressed in a §§ 63.1220(a)(2)–(a)(4), and (b)(2)–(b)(4)
cement kilns, and lightweight aggregate mass concentration format) to prevent and 63.1221(a)(3)–(a)(4) and (b)(3)–
kilns. These standards were backsliding. (b)(4).
promulgated on February 13, 2002, and Response: Even though a source may We evaluated the relative stringency
sources were required to comply with operate in compliance with a standard of the standards expressed in the
them no later than September 30, 2003, expressed in a thermal emission format, thermal emissions format compared to
unless granted a one-year extension (see a source may or may not also be in the interim standards for the entire
§ 63.1206(a)). Thus, all hazardous waste compliance with the corresponding source category in order to determine if
combustors are currently complying mass concentration interim standard the dual standard scheme could be
with the interim standards. The (e.g., the semi- and low volatile metal avoided. We determined that we could
comment that the standards lack some emission standards for cement and not. For some HAP groups we found
type of requisite statutory pedigree lightweight aggregate kilns of §§ 63.1204 that many sources in the category would
misses the central point of our have the potential to exceed the interim
interpretation of the statute: motivation 121 An F-factor is an estimate of the amount of
for achieving a standard (be it regulatory combustion gas volume that is generated per fuel
122 For example, see 69 FR at 21255–258, 267–
heat input for a given type of fuel, expressed in
compulsion, statutory requirement, or units, for example, cubic feet of combustion gas per 271.
123 Although the mercury standard promulgated
million British thermal units (BTU) of fuel burned.
120 The difference between the higher heating In the proposal, EPA used F-factors to convert the for cement kilns is not expressed using a thermal
value and lower heating value of an aqueous waste emission standards expressed on a thermal basis to emission format basis, the same concept applies
is insignificant relative to the difference in heating mass concentrations in order to make a judgment because the mercury standard is a hazardous waste
value between an aqueous waste and an organic as to the relative stringency of the proposed MACT feed concentration standard, which is a different
liquid waste fuel. standards relative to the interim standards. format than the interim standard.
standards for that HAP.124 In this case, Response: In those few situations and a small amount of data can skew the
we considered simply ‘‘capping’’ the where we have established floor levels result. See § 112(d)(3)(B).127
standard expressed in the thermal at the level of the interim standards, we For example, many of the calculated
emission format by the interim standard have done so as the best means of new source chlorine floors were slightly
(i.e., the promulgated standard would estimating performance of the best higher than the calculated existing
only be expressed in a mass performing sources. Based on the source standards because we assumed
concentration format). However, we available data to us, the average of the
all sources with measured emissions
conclude that this approach would not best performing sources exceeds the
be appropriate because the standard below 20 ppmv were in fact emitting at
level of the interim standards in a few
expressed in a thermal emission format 20 ppmv (see part four, section I.C). We
instances. Under these circumstances,
would likely be more stringent than the generally are unable to differentiate a
the binding regulatory limit becomes the
mass concentration for some sources, best means available to us to estimate single best performing source among
and the statute requires that MACT performance. See Mossville, 370 F. 3d at these best performers because many/all
floors reflect this superior level of 1241–42 (accepting regulatory level as a of the best performing sources emissions
performance. floor standard where sources’ measured are adjusted to the same emission level.
In other cases we found that the performance is not a valid means of The calculated new source floor can be
standards expressed in the thermal determining floor levels, and where slightly higher than the existing source
emissions format would not likely such data contains results as high as floor because the variability factor that
exceed the interim standards by the those regulatory levels). is applied to the single best performing
majority of sources operating under source is based on only one test
typical conditions.125 While our F. How Can EPA’s Approach to condition (with three emission test
analysis (based on information in our Assessing Variability and its Ranking runs). This results in a higher level of
data base) shows in these cases that the Methodologies Be Reasonable When uncertainty relative to the existing
emission standard expressed in a They Result in Standards Higher Than source standard, which is based on a
thermal emission format would not the Interim Standards? compilation of emissions data from
likely result in an exceedance of the several sources that have essentially the
A commenter argued that EPA’s floor
interim standard, this conclusion may same projected emissions as a result of
methodologies, in particular its
not be true because the assumptions the method bias correction factor. The
consideration of variability beyond that
may not be valid for a particular source variability factor that is applied to the
demonstrated in single test conditions,
or site-specific factors may change in emissions of the single best performing
future operations. For example, HAP the SRE/feed and Air Pollution Control
Device methodologies, must be arbitrary source is therefore higher than the
metal emissions could increase over
because in a few instances projected variability factor for the existing source
time due to increases in HAP
standards using these approaches were floor because there are fewer degrees of
contributions from raw materials or
alternative raw materials. Given this higher than the current interim freedom in the statistical analysis.128
potential, we adopt dual standards for standards, a level every source (not just Likewise, many of the calculated solid
the HAP metal standards in order to the best performers) are achieving. fuel boiler new source standards were
ensure that standards expressed in a Commenters also noted that one of the slightly higher than the calculated
thermal emissions format will not new source standards calculated under existing source standards because, as
exceed emission levels achieved under these approaches was higher than an discussed above, there are fewer degrees
the interim standards.126 existing source standard, another of freedom when assessing the
Comment: Several commenters state arbitrary result. variability from a single best performing
that the interim standards do not reflect EPA believes that these seeming source. The solid fuel boiler
the average performance of the best anomalies (which are infrequent) result ‘‘anomalies’’ also occur using a straight
sources, and so cannot be the basis for from the database used to calculate emissions methodology. See USEPA,
floor levels. performance and standards, rather than ‘‘Technical Support Document for the
from the approaches to assessing HWC MACT Standards, Volume III:
124 An example for each category is semivolatile variability or the two questioned floor Selection of MACT Standards,’’
metals thermal emissions standard for existing methodologies. The data base is from September, 2005, Section 19, for further
cement and lightweight aggregate kilns. See USEPA,
‘‘Final Technical Support Document for the HWC
test results which preceded EPA’s discussion that summarizes and
MACT Standards, Volume III: Selection of MACT adoption of the interim standards. Thus, explains these so-called anomalies.
Standards,’’ Section 23.1, September 2005. the level of performance required by the
125 An example is the emission standards for low
later rule is not necessarily reflected in
volatile metals for existing and new cement kilns pre-rule test data. In confirmation, some
and new lightweight aggregate kilns. See USEPA,
‘‘Final Technical Support Document for the HWC of the standards computed using 127 See USEPA, ‘‘Technical Support Document for
MACT Standards, Volume III: Selection of MACT straight emission approaches also are the HWC MACT Standards, Volume III: Selection of
Standards,’’ Section 23.1, September 2005. higher than the interim standards. Other MACT Standards,’’ September 2005, Section 19, for
126 In response to a comment regarding the further discussion.
anomalies arise simply due to scarcity 128 For a single test condition the t factor used in
implementation of dual standards, we note the
promulgation of a new provision allowing sources
of data (floor levels for certain HAP variability factor calculation has n–1 degrees of
to petition the Administrator to waive the HAP emitted by lightweight aggregate kilns freedom where n is the number of runs for that
metal feedrate operating parameter limits for either especially, where there are only nine condition. For the MACT floor calculation the t
the emissions standards expressed in a thermal sources total). In these situations there factor has X–N degrees of freedom where X is the
emissions format (or the mercury feed
concentration standard for cement kilns) or the
is a greater likelihood that one or more total number of runs from all sources in the MACT
of the best performing sources will have pool and N is the number of sources in the pool.
interim standards based on documentation that the
feedrate operating parameter limit is not needed to relatively high emissions because we are See USEPA, ‘‘Technical Support Document for the
ensure compliance with the relevant standard on a HWC MACT Standards, Volume III: Selection of
required to use data from five sources to MACT Standards,’’ September, 2005, Section 7.1 for
continuous basis. See new § 63.1209(g)(1)(iv) and
Comment Response Document, Volume I, Section
comprise the MACT pool whenever we more information on the floor calculation
3.5. have data from fewer than 30 sources, procedure.
IV. Use of Surrogates and that when particulate matter is hazardous waste mercury feed control
removed from emissions the data that support their assertion.
A. Particulate Matter as Surrogate for
nonmercury HAP metals are removed Although these stakeholders did not
Metal HAP
with it.130 Nonmercury metal HAP submit long-term normal hazardous
Comment: A commenter states that emissions are therefore reduced waste feed control data for the
EPA’s use of particulate matter as a whenever particulate matter emissions nonenumerated metals, we can still see
surrogate for nonenumerated metals is are reduced. The particulate matter that use of the normal nonenumerated
unlawful and arbitrary and capricious standard thus is an effective and metal snapshot emissions in our
because although particulate matter appropriate surrogate that assures database to determine MACT floors
emissions may provide some indication sources are controlling these metal HAP could raise similar concerns with
of how good a source’s end-of stack with an appropriate back-end control respect to whether the normal data in
control of such metals is, it does not technology. National Lime v. EPA, 233 fact represents average emissions at
indicate what its actual metal emission F. 3d at 639. The nonenumerated metal these sources, and their level of
levels are.129 The commenter states that HAP are no different than other performance.
emissions of these metals can vary semivolatile or low volatile metals in Use of particulate matter emissions
based on metal feed rate without having that they also will be effectively data to assess the relevant best
any appreciable effect on particulate controlled with a back-end particulate performers for nonenumerated metal
matter emission levels. Thus a matter air pollution control device. HAP is therefore more appropriate for
particulate matter standard does not We also considered the possibility of two reasons. Compliance test data better
necessarily ensure that metal emissions developing a standard for account for emissions variability and
are reduced to the metal emission levels nonenumerated HAP metals instead of a avoid the normal emissions bias
achieved by the relevant best PM standard (i.e., regulating these discussed above. We also have much
performing sources. To support this metals directly, rather than through use more particulate matter emissions data
assertion, the commenter states that of a surrogate). We conclude for several from more sources, which better allows
EPA is on record saying ‘‘low reasons, however, that issuing emission us to evaluate the true range of
particulate matter emissions do not standards for these nonenumerated emissions from all the sources within
necessarily guarantee low metal HAP metals in lieu of a particulate matter the source category and to assess and
emissions, especially in instances where standard would not adequately control identify the relevant top performing 12
the hazardous waste feeds are highly nonmercury metal HAPs to levels percent of the sources.
concentrated with metal HAP.’’ 69 FR at achieved by the relevant best
performing sources. It would be inappropriate to assess
21221. total stack gas emissions of
Response: The final rule uses a We generally lack sufficient
compliance test emissions data for the nonenumerated metals for cement kiln
particulate matter standard as a and lightweight aggregate kilns when
surrogate to control: (1) Emissions of noneneumerated metals to assess the
relevant best performing sources, determining the relevant best
nonenumerated metals that are performers because these emissions
attributable to all feedstreams (both because, as discussed below, most of
these metals were not directly regulated would, in part, reflect the metal feed
hazardous waste and remaining inputs); levels in these sources’ nonhazardous
and (2) all nonmercury metal HAP pursuant to RCRA air emission
standards.131 Although we have more waste process feedstreams. This is not
emissions (both enumerated and appropriate because nonhazardous
nonenumerated metal HAP) from the emissions data for these metals that are
based on (so called) normal operations, process feedstream control is not a
nonhazardous waste process feeds at feasible means of control. See part four,
cement kilns, lightweight aggregate we still lack sufficient emissions data to
establish nonenumerated metal section III.B.1. A potential solution to
kilns, and liquid fuel boilers (e.g., this problem would be to identify the
standards for all the source categories.
emissions attributable to coal and raw relevant best performers by assessing
Use of normal data may also be
material at a cement kiln, and emissions each source’s hazardous waste thermal
problematic because of the concern
attributable to fuel oil for liquid fuel emissions for these nonenumerated
raised by the cement kiln and
boilers). Incinerators, liquid and solid metals (given that hazardous waste
lightweight aggregate kiln stakeholders
fuel boilers may elect to comply with an thermal emissions exclude by definition
that our normal metals emissions data
alternative to the particulate matter emissions attributable to inputs other
obtained from compliance tests are not
standard that would limit emissions of than hazardous waste, i.e. raw materials
representative of the range of actual
all the semivolatile metal HAPs and low and fossil fuels). This, however, would
emissions at their sources. Cement kiln
volatile metal HAPs. See § 63.1219(e). be problematic because, aside from the
and lightweight aggregate kiln
The particulate matter standard is a data limitation issues, the majority of
stakeholders submitted long-term
necessary, effective, and appropriate the nonenumerated metals data reflect
surrogate to control nonmercury metal 130 This statement is equally true for any emitting normal emissions which often do not
HAPs. The record demonstrates source, not just hazardous waste combustors. It is contain the highest feed rates used by
overwhelmingly that when a hazardous well established that semivolatile and low volatile the source. As a result, we cannot assess
waste combustor emits particulate metals exist in solid particulate form at typical air
pollution control device operating temperatures. performance on a thermal emissions
matter, it also emits nonmercury HAP This is supported by (1) known operating basis because of the uncertainty
metals as part of that particulate matter, temperature ranges of air pollution control devices associated with system removal
used by hazardous waste combustors; (2) known efficiencies at such low metal feedrates.
129 ‘‘Enumerated’’ metals are those HAP metals metal volatility equilibrium relationships; and (3)
directly controlled with an emission limit, i.e., lead, extensive technical literature. See USEPA, Furthermore, even if we could issue
cadmium, chromium, arsenic and beryllium. The ‘‘Technical Support Document for the HWC MACT hazardous waste thermal emissions
remaining nonmercury metal HAP (i.e., antimony, Standards, Volume III: Selection of MACT standards for these metals, a particulate
cobalt, manganese, nickel, and selenium) are called Standards,’’ September 2005, Section 3.1. matter emission standard would still be
‘‘nonenumerated’’ metal HAP (note that arsenic and 131 At best, we may have enough compliance test
berrylium are nonenumerated metals for liquid fuel data for antimony and selenium to adequately
necessary to control nonmercury metal
boilers because the low volatile metal emission assess relevant best performers for only incinerators HAP emissions from the nonhazardous
standard applies only to chrome). and lightweight aggregate kilns. waste process feedstreams.
Emission standards for these the relevant best performers for metal emissions have traditionally been
nonenumerated metals could require nonenumerated HAP metals and regulated in volatility groupings because
sources to implement hazardous waste estimating these sources’ level of the volatility of the metal affects the
feed control (for these metals) to comply performance. efficiency of back-end control (i.e.,
with the standard.132 We are less Comment: A commenter states that semivolatile metals are more difficult to
assured that these sources were EPA’s rationale for use of particulate control than low volatile metals because
implementing hazardous waste feed matter as a surrogate for nonenumerated they volatilize in the combustor and
control for these nonenumerated metals metals is flawed because EPA has then condense as small particulates
at the time they conducted the provided no data in the proposal to prior to or in the emission control
emissions tests (which serve as the basis justify its hypothesis that particulate device). When identifying the best
for floor calculations) because most of matter is an appropriate surrogate for performing sources, we previously have,
these metals were never directly non-enumerated metal HAP. The in general, only evaluated sources that
regulated pursuant to the RCRA commenter also states that the proposed have metal emissions information for
emission standards.133 This means that emission standards for particulate every metal in the volatility grouping.
sources tended to optimize (or at least matter for existing sources discriminate This approach could prove to be
concentrate their efforts on) control of against boilers and process heaters that problematic since it is not likely many
the metals that are regulated. Although burn clean (i.e., little or very low sources will have emissions data for all
these metals were being controlled with concentrations of HAP metals) the metals.
each source’s back-end control device, hazardous waste fuels. The commenter Although we could not calculate
sources may not have been controlling suggests that if there are sufficient data, alternative total metal emission floor
these metal feedrates because they EPA should consider developing an standards based on the available
probably were not subject to specific alternative emission standard for total emissions data we have, we agree with
feedrate limitations (feed control of the HAP metals for new and existing liquid the commenters’ view that sources that
enumerated metal HAP does not ensure fuel boilers, as was done for the Subpart burn hazardous waste fuels with low
feed control of these nonenumerated DDDDD National Emission Standards levels of nonenumerated metals should
metal HAP). Furthermore, simultaneous for Hazardous Air Pollutants for be allowed to comply with a metals
feed control of all these metals, when Industrial/Commercial/Institutional standard rather than the particulate
combined with enumerated semivolatile Boilers and Process Heaters. matter standard. We proposed an
and low volatile metals, may not be Response: As previously discussed in alternative to the particulate matter
possible because the best performing this section, particulate matter reflects standard (see 69 FR at 21331) for
sources for all these metals may emissions of nonmercury metal HAPs incinerators, liquid, and solid fuel
collectively represent a hazardous waste because these compounds comprise a boilers that was a simplified version of
feedstream that does not exist in percentage of the particulate matter the alternative particulate matter
practice (from a combined metal (provided these metals are fed into the standard that is currently in effect for
concentration perspective) because combustion unit). The technologies that incinerators pursuant to the interim
there likely would be different best have been developed and implemented standards (see § 63.1206(b)(14)). We
performers for each of the metal HAP or to control particulate matter also control received no adverse comment and are
metal HAP groups.134 We thus conclude nonmercury metal HAP. Since non- promulgating this alternative as
that back-end control as measured and mercury metal HAP is a component of proposed. The alternative metal
assessed by each source’s particulate particulate matter, we can use standards apply to both enumerated and
matter emissions is the appropriate floor particulate matter as a surrogate for nonenumerated metal HAP, excluding
technology to assess when identifying these metals. Further justification for mercury. For purposes of these
the use of particulate matter as a alternative requirements, each
132 Sources that otherwise would be equipped
surrogate to control metal HAP is nonenumerated metal is classified as
with what is considered to be a MACT back-end either a semivolatile or a low volatile
control devices (i.e., a control device achieving the included in the technical support
final rule particulate matter standard) may not be document.135 metal and subsequently grouped with
able to achieve these metal emissions standards due We conclude that we do not have the associated semivolatile and low
to varying metal feed levels (both within sources enough nonenumerated metal emissions volatile enumerated metals. The
and across sources). Such an outcome may require semivolatile and low volatile metals
a source to limit the amount of metal that is fed into data to calculate alternative total metal
the combustion unit to achieve the standard. emission floors for liquid fuel boilers. standards under this alternative are the
133 Antimony is the only nonenumerated metal The most problematic of these metals same as those that apply to other liquid
that is directly regulated pursuant to the boilers and are manganese and cobalt, where we fuel boilers, but the standard would
industrial furnace regulations. See § 266.106.
have emission data from only three apply to all metal HAP, not just those
134 We generally cannot combine these
sources. We have much more enumerated in the generic low volatile
nonenumerated metals into the associated
semivoltile or low volatile metal volatility compliance test particulate matter metal and semivolatile metal standards.
groupings promulgated in this final rule for emissions data from liquid fuel boilers, See §§ § 63.1216(e), 63.1217(e) and
purposes of establishing ‘‘grouped’’ emission
and thus conclude that the particulate 63.1219(e).
standards because we cannot mix compliance test
data with normal emissions data when calculating matter standard best reflects the B. Carbon Monoxide/Hydrocarbons and
floors (the majority of the standards included in this emission levels achieved by the relevant DRE as Surrogates for Dioxin/Furan
final rule are based on compliance test data, and the best performers.
majority of the data we have for nonenumerated Comment: One commenter states that
Similar to the above discussion,
metals being normal). Furthermore, if we were to the dioxin/furan floors for new and
separately group the normal nonenumerated metal calculating an alternative total metal
existing solid fuel boilers is unlawful
emission data into their associated semivolatile or emissions floor raises questions
low volatile metal group, we may encounter data and arbitrary and capricious. EPA
regarding the method used to calculate
limitation issues because each source would need established the floor for dioxin/furan for
to have measured each of the nonenumerated
such floors. Hazardous waste combustor
these sources as compliance with the
metals in that associated metal volatility group in
order for us to conclude that the emission data 135 See USEPA, ‘‘Technical Support Document for carbon monoxide or hydrocarbon
adequately represents the sources combined the HWC MACT Standards, Volume III: Selection of standard and the destruction and
emissions of semivolatile or low volatile metals. MACT Standards,’’ September 2005, Section 3.1. removal efficiency (DRE) standard. The
commenter states that EPA has not demonstrated that setting an emission aggregate kilns, Portland cement plants,
shown that carbon monoxide or standard is infeasible under section and industrial boilers; and section 129
hydrocarbon emissions correlate to 112(h)(1). standards for commercial and industrial
dioxin/furan emissions, and, Carbon Monoxide and Hydrocarbons waste incinerators, municipal waste
accordingly, has not shown that the Are Adequate Surrogates to Control combustors, and medical waste
carbon monoxide or hydrocarbon Dioxin/Furan when Other Controls Are incinerators. Finally, hydrocarbon
standard, together with the DRE Not Effective or Achievable. Carbon emissions are an indicator of organic
standard, are valid surrogates. monoxide and hydrocarbons (coupled hazardous air pollutants because
This commenter also states that it is with the DRE standard) are the best hydrocarbons are a direct measure of
inappropriate for EPA to use carbon available surrogates to control dioxin/ organic compounds.
monoxide or hydrocarbons and DRE as furan emissions when a numerical floor Commenters on our proposed MACT
surrogates to establish dioxin/furan would not be achievable and when standards for hazardous waste
floors for liquid fuel boilers with wet or other indirect controls, such as control incinerators, cement kilns, and
no air pollution control devices and for of the gas temperature at the inlet of a lightweight aggregate kilns stated that
hydrochloric acid production furnaces. dry particulate matter control device to EPA’s own surrogate evaluation 138 did
The commenter believes EPA 400F, are not applicable or effective.136 not demonstrate a relationship between
inappropriately justifies these surrogates As we explained at proposal, carbon monoxide or hydrocarbons and
by claiming that a numerical dioxin/ operating under good combustion organic HAP at the carbon monoxide
furan floor would not be replicable by conditions to minimize emissions of and hydrocarbon levels evaluated. See
the best sources or duplicable by the organic compounds such as 64 FR at 52847 (September 30, 1999).
others. The commenter states that EPA polychlorinated biphenyls, benzene, Several commenters on that proposed
has no discretion to avoid setting floors and phenol that can be precursors to rule noted that this should not have
for a HAP just because it believes that dioxin/furan formation is an important been a surprise given that the carbon
HAP is not controlled with a requisite to control dioxin/furan monoxide and hydrocarbon emissions
technology. Rather, EPA must set floors emissions.137 See 69 FR at 21274. data evaluated were generally from
reflecting the relevant best sources’ Minimizing dioxin/furan precursors by hazardous waste combustors operating
actual performance. Such floors operating under good combustion under good combustion conditions (and
necessarily will be duplicable by the practices plays a part in controlling thus, relatively low carbon monoxide
relevant best sources themselves. That dioxin/furan emissions, and that role is and hydrocarbon levels). Under these
they cannot be replicated by other substantially enhanced when there are conditions, emissions of HAP were
sources is irrelevant according to the no other dominant factors that relate to generally low, which made the
commenter. dioxin/furan formation and emission demonstration of a relationship more
In addition, the commenter states that (e.g., operating a dry particulate matter difficult. These commenters noted that
EPA does not claim or demonstrate that control device at temperatures above there may be a correlation between
the carbon monoxide and hydrocarbon 400F). carbon monoxide and hydrocarbons and
floors for solid fuel boilers reflect the Carbon monoxide and hydrocarbons
organic HAP, but it would be evident
average emission levels achieved by the are widely accepted indicators of
primarily when actual carbon monoxide
relevant best sources. combustion conditions. The current
and hydrocarbon levels are higher than
Finally, the commenter also notes that RCRA regulations for boilers and
the regulatory levels. We agreed with
EPA appears to argue that its carbon hydrochloric acid production furnaces
those commenters, and concluded that
monoxide or hydrocarbon standard and use emissions limits on carbon
carbon monoxide and hydrocarbon
DRE standard could be viewed as work monoxide and hydrocarbons to control
levels higher than those we established
practice standards under section 112(h) emissions of toxic organic compounds.
which allows EPA to establish work See 56 FR 7150 (February 21, 1991) as emission standards for hazardous
practice standards in lieu of emission documenting the relationship between waste burning incinerators, cement
standards only if it is not be feasible to carbon monoxide, combustion kilns, and lightweight aggregate kilns
set the former. Because EPA has made efficiency, and emissions of organic are indicative of poor combustion
no such demonstration, setting work compounds. In addition, carbon conditions and the potential for
practice standards to control dioxin/ monoxide and hydrocarbons are used by increased emissions organic HAP. We
furan emissions from boilers would be many CAA standards for combustion continue to believe that carbon
unlawful according to the commenter. sources to control emissions of organic monoxide and hydrocarbons are
Response: The commenter raises four HAP, including: MACT standards for adequate surrogates for organic HAP
issues: (1) Are the carbon monoxide/ hazardous waste burning incinerators, which may be precursors for dioxin/
hydrocarbon standard and the DRE hazardous waste burning cement kilns, furan formation and note that the
standard adequate surrogate floors to hazardous waste burning lightweight commenter did not explain why our
control dioxin/furan; (2) floors for technical analysis is problematic.
existing sources must be established as 136 As discussed in Part Two, Section V, we view Emissions that Are Not Replicable or
the average emission limitation the carbon monoxide, hydrocarbon, and destruction Duplicable Are Not Being ‘‘Achieved’’.
achieved by the best performing sources removal efficiency standards as unaffected by the The commenter believes that floors
Court’s vacature of the September 1999 challenged must be established as the average
irrespective of whether the limitation is regulations for incinerators, cement kilns, and
duplicable by the best performing lightweight aggregate kilns. We are therefore not re- emission limitation of the best
sources or replicable by other sources; promulgating and reopening consideration of these performing sources irrespective of
(3) EPA has not explained how the
standards in today’s final rule for these source whether they are replicable by the best
categories. performing sources or duplicable by
carbon monoxide and hydrocarbon 137 Operating under good combustion conditions
floors reflect the average emission also helps minimize soot formation on boiler tubes. other sources. To the contrary, emission
limitation achieved by the relevant best Research has shown that operating under
conditions that can form soot followed by operating 138 See Energy and Environmental Research
sources; and (4) EPA cannot establish under good combustion conditions can lead to Corporation, ‘‘’Surrogate Evaluation of Thermal
work practice standards for dioxin/furan dioxin/furan formation. See Section 2.4 of Volume Treatment Systems,’’’ Draft Report, October 17,
under section 112(h) because it has not III of the Technical Support Document. 1994.
levels that are not replicable by the best counter-productiveness of encouraging sources operate under the good
performing sources are not being use of higher sulfur coal. Hence, that combustion conditions required to
‘‘achieved’’ by those sources and cannot reported emission level is not being achieve carbon monoxide levels in the
be used to establish the floor. ‘‘achieved’’ for the purpose of range of zero to 100 ppmv.141 (See also
For solid fuel boilers, we explained at establishing a floor. the discussion below regarding the
proposal why dioxin/furan emissions Finally, we note that beyond-the-floor progression of hydrocarbon oxidation to
are not replicable by the best performing controls such as activated carbon can carbon dioxide and water). As such,
sources (or duplicable by other sources): control dioxin/furan to a particular lowering the carbon monoxide floor
there is no dominant, controllable emission level. If a source were to below 100 ppmv may not provide
means that sources are using that can install activated carbon, it could achieve significant reductions in organic HAP
control dioxin/furan emissions to a the level demonstrated in a compliance emissions. Moreover, it would be
particular level. See 69 FR at 21274–75. test, after adjusting the level to account inappropriate to establish the floor
We explained that data and information for emissions variability to ensure the blindly using a mathematical
lead us to conclude that rapid quench measurement was replicable. The approach—the average emissions for the
of post-combustion gas temperatures to commenter argues that such a result is best performing sources—because the
below 400 °F—the control technique mandatory under the straight emissions best performing sources may not be able
that is the basis for the MACT standards approach (the only way the commenter to replicate their emission levels (and
for dioxin/furan for hazardous waste believes best performers can be other sources may not be able to
burning incinerators, and cement and determined). Doing so, however, would duplicate those emission levels) using
lightweight aggregate kilns—is not the amount to a surreptitious beyond-the- the exact types of good combustion
dominant dioxin/furan control floor standard (forcing adoption of a practices they used during the
mechanism for coal-fired boilers. We control technology not used by any compliance test documented in our data
believe that sulfur contributed by the existing source), without considering base. This is because there are myriad
coal fuel is a dominant control the beyond-the-floor factors set out in factors that affect combustion efficiency
mechanism by inhibiting formation of section 112(d)(2). In fact, we considered and, subsequently, carbon monoxide
dioxin/furan. Nonetheless, we do not beyond-the-floor standards based on use emissions. Extremely low carbon
know what minimum level of sulfur of activated carbon for these sources— monoxide emissions cannot be assured
provides significant control. Moreover, solid fuel boilers, liquid fuel boilers by controlling only one or two operating
sulfur in coal causes emissions of sulfur with wet or no emission control device, parameters.
oxides, a criteria pollutant, and and hydrochloric acid production We proposed a floor level for
particulate sulfates. For this reason, as furnaces—but rejected them for reasons hydrocarbons of 10 ppmv even though
well as reasons stated at 69 FR 21275, of cost. The cost-effectiveness ranged the currently enforceable standard for
we are not specifying a level of sulfur from $2.5 million to $4.9 million per boilers and hydrochloric acid
in coal for these sources as a means of gram TEQ of dioxin/furan removed. In production furnaces is 20 ppmv
dioxin/furan control. contrast, the cost-effectiveness of the because: (1) Although very few sources
The same rationale applies to liquid beyond-the-floor standard we elect to comply with the RCRA standard
fuel boilers with no air pollution promulgate for liquid fuel boilers for hydrocarbons rather than the
controls or wet air pollution control equipped with dry emission control standard for carbon monoxide, those
systems and to hydrochloric acid devices is $0.63 million per gram TEQ that comply with the hydrocarbon
production furnaces—there is no of dioxin/furan removed.140 standard have hydrocarbon levels well
dominant, controllable means that Consequently, we are not below 10 ppmv; and (2) reducing
sources are using that can control promulgating a beyond-the-floor hydrocarbon emissions within the range
dioxin/furan emissions to a particular standard for dioxin/furan for these of 20 ppmv to 10 ppmv may reduce
emission level.139 Thus, best performer sources, and do not believe we should emissions of organic HAP.
Although all sources are likely to be
dioxin/furan emissions are not adopt such a standard under the guise
achieving hydrocarbon levels below 10
replicable by the best performing of determining floor levels.
ppmv, it is not appropriate to establish
sources (or duplicable by other sources). The Carbon Monoxide and
a lower floor level because
For these sources, the predominant Hydrocarbon Floors Are Appropriate
hydrocarbons are a surrogate for organic
dioxin/furan formation mechanism for MACT Floors. We explained at proposal
HAP. Although total hydrocarbons
other source categories—operating a why the carbon monoxide standard of
would be reduced at a floor level below
fabric filter or electrostatic precipitator 100 ppmv and the hydrocarbon
10 ppmv, we do not know whether
above 400F—is not a factor. standard of 10 ppmv are appropriate
Given that these sources are not using floors. See 69 FR at 21282. The floor 141 We note, however, that this general principle
controllable means to control dioxin/ level for carbon monoxide of 100 ppmv may not always apply. There are data that indicate
furan to a particular emission level, is a currently enforceable Federal that even though carbon monoxide levels are below
there is no assurance that the best standard. Although some sources are 100 ppmv, hydrocarbon levels may not always be
below 10 ppmv. See 64 FR at 52851 and Part Four,
performers can achieve in the future the achieving carbon monoxide levels Section IV B. and C. of this preamble. An example
emission level reported in the below 100 ppmv, it is not appropriate to of how this might occur, although not a likely
compliance test in our data base. Put establish a lower floor level because practical scenario, is if combustion is quenched
another way, the test data do not reflect carbon monoxide is a conservative before substantial carbon monoxide can be
generated, leaving unburned hydrocarbons in the
these sources’ variability, and the surrogate for organic HAP. Organic HAP stack gas. Because of this potential (although
variability is largely unquantifiable emissions may or may not be substantial unlikely) concern, the rule requires sources that
given the uncertainties regarding control at carbon monoxide levels greater than elect to monitor carbon monoxide rather than
hydrocarbons to conduct a one-time test to
mechanisms plus the environmental 100 ppmv, and are extremely low when document that hydrocarbons are below 10 ppmv
and to establish operating limits on parameters that
139 We note that the same rationale also applies 140 See USEPA, ‘‘Technical Support Document for affect combustion conditions (i.e., the same
to incinerators with wet or no air pollution control the HWC MACT Standards, Volume III: Selection of operating parameters that we use for compliance
equipment and that are not equipped with a waste MACT Standards,’’ September 2005, Sections 12, assurance with the DRE standard). See
heat boiler. 13, and 15. § 63.1206(b)(6).
organic HAP would be reduced C. Use of Carbon Monoxide and Total kilns. In that rule we proposed to
substantially. As combustion conditions Hydrocarbons as Surrogate for Non- require compliance with both the total
improve and hydrocarbon levels Dioxin Organic HAP 143 hydrocarbon standard and the carbon
decrease, the larger and easier to Comment: A commenter states that monoxide standard. We requested
combust compounds are oxidized to neither the total hydrocarbon nor carbon comment on whether these
form smaller compounds that are, in monoxide standard alone provides requirements were redundant, and we
turn, oxidized to form carbon monoxide adequate surrogate control for organic later requested comment on whether we
and water. As combustion continues, HAP. Accordingly, EPA must include should allow sources to comply with
carbon monoxide is then oxidized to standards for both. Hazardous waste either the carbon monoxide standard or
form carbon dioxide and water. Because combustors could have total the total hydrocarbon standard. We
carbon monoxide is a difficult-to- hydrocarbon levels below the standard clarified, however, that allowing sources
destroy refractory compound (i.e., during the carbon monoxide compliance to comply with the carbon monoxide
oxidation of carbon monoxide to carbon tests, but higher total hydrocarbon standard would be contingent on the
dioxide is the slowest and last step in levels at other times during normal source demonstrating compliance with
the oxidation of hydrocarbons), it is a operation because there are many the hydrocarbon standard during the
conservative surrogate for destruction of variables that can affect total compliance test. We believed this was
hydrocarbons, including organic HAP, hydrocarbon emissions, and these will necessary because we had limited data
as discussed above. As oxidation not all be represented during the carbon that showed a source could have total
progresses and hydrocarbon levels monoxide compliance test. The hydrocarbon levels exceeding 10 ppmv
decrease, the larger, heavier compounds commenter states that EPA is on record even though their carbon monoxide
are destroyed to form smaller, lighter stating that carbon monoxide limits emission levels were below 100 ppmv.
compounds until ideally all alone may not by itself minimize EPA subsequently promulgated this
hydrocarbons are oxidized to carbon organic emissions because products of approach in the September 1999 Final
incomplete combustion can result from Rule. 62 FR 52829.
monoxide (and then carbon dioxide)
and water. Consequently, the small pockets within the combustion Today’s rule adopts the same
relationship between total hydrocarbons zone where adequate time, temperature, approach for liquid and solid fuel
and organic HAP becomes weaker as turbulence and oxygen have not been boilers and hydrochloric acid
provided to completely oxidize these production furnaces. We again conclude
total hydrocarbon levels decrease to
organics. The commenter also states that that it is not necessary to require
form compounds that are not organic
EPA is on record stating that total sources to verify compliance with both
HAP, such as methane and acetylene.142
hydrocarbon levels can exceed good of these standards on a continuous basis
Moreover, as discussed above for combustion condition levels when with two separate continuous emission
carbon monoxide, it would be carbon monoxide levels are below 100 monitors, given the redundancy of these
inappropriate to establish the floor ppmv. measurement techniques. Total
blindly using a mathematical Response: The final rule requires hydrocarbon emission measurements
approach—the average emissions for the compliance with destruction and are a more direct indicator of organic
best performing sources—because the removal efficiency and carbon HAP emissions than carbon monoxide.
best performing sources may not be able monoxide or hydrocarbon standards as Hence, continuous compliance with this
to replicate their emission levels (and surrogates to control non-dioxin organic standard always assures that organic
other sources may not be able to HAP emissions 144 from liquid fuel HAP are well controlled. Carbon
duplicate those emission levels) using boilers, solid fuel boilers, and monoxide is a conservative indicator of
the exact types of good combustion hydrochloric acid production furnaces. combustion efficiency because it is a
practices they used during the These are effective and reliable product of incomplete combustion and
compliance test documented in our data surrogates to control organic HAP. We because it is a refractory compound that
base. This is because there are myriad conclude that simultaneous is more thermally stable than
factors that affect combustion efficiency measurement of both total hydrocarbons hydrocarbons. The hydrocarbon
and, subsequently, hydrocarbon (and and carbon monoxide with continuous products of incomplete combustion that
carbon monoxide) emissions. Extremely emission monitors is not necessary are simultaneously formed during
low hydrocarbon emissions cannot be because each serves as a reliable incomplete, or inefficient, combustion
assured by controlling only one or two surrogate to control organic HAP conditions can be subsequently
operating parameters. emissions. The commenter has cited oxidized later in the combustion
EPA preamble language that was process. In such instances carbon
The Standards for CO and HC Are included in the April 19, 1996 proposed
Not Work Practice Standards. The floor monoxide will likely still be prevalent
rule for hazardous waste incinerators, in the exhaust gas even though the
standards for CO or HC for boilers and cement kilns, and lightweight aggregate
hydrochloric acid production furnaces products of incomplete combustion
are quantified emission limits. The were later oxidized. The conservative
143 As discussed in part two, section V, we view
standards consequently are not work nature of carbon monoxide as an

carbon monoxide, hydrocarbon, and destruction
removal efficiency standards as unaffected by the indicator of good combustion practices
practice standards (even though they
Court’s vacature of the September 1999 challenged is supported by our data. At carbon
represent levels showing good regulations for incinerators, cement kilns, and monoxide levels less than 100 ppmv,
combustion control). CAA section lightweight aggregate kilns. We are therefore not re- our data indicates that there is no
302(k). EPA’s reference to section promulgating and did not reconsider these
standards in today’s final rule for these source apparent relationship between carbon
112(h)(1) at proposal (69 FR at 21275)
categories. monoxide and hydrocarbons (other than
was consequently erroneous. 144 As discussed in the previous section, these that hydrocarbon levels are generally
standards are also used as surrogates to control below 10 ppm when carbon monoxide
142 USEPA, Technical Support Document for dioxin/furans for hydrochloric acid production
HWC MACT Standards, Volume III: Selection of furnaces, solid fuel-fired boilers, and liquid fuel-
levels are below 100 ppm). For example,
MACT Standards and Technologies, July 1999, fired boilers that are not equipped with dry air a source with a carbon monoxide level
Section 12.1.2. pollution control devices. of 1 ppm is no more likely to have lower
measured hydrocarbons than a source that are achieving carbon monoxide of sources to identify the best performers
achieving a carbon monoxide emission emission levels of 100 ppm or a and emissions variability is understated
level of 100 ppm. 145 hydrocarbon emission levels of 10 ppm when calculating the floor.
We consider the few instances where are known to be operating pursuant to Response: We agree with the
the data showed total hydrocarbon good combustion practices. This is commenter. For the final rule, we use an
levels above 10 ppmv while carbon supported by an extensive data analysis approach to address nondetects
monoxide levels are below 100 ppmv to we used to support identical standards whereby a value is assigned to each
be anomalies. Even so, we have for incinerators, cement kilns, and nondetect within its possible range such
accounted for this by requiring lightweight kilns which were that the 99th percentile upper
compliance with the hydrocarbon promulgated in the September 1999 prediction limit for the data set (i.e., test
standard during the compliance test if a Final Rule. We are applying the same condition runs for each source) is
source elects to comply with the carbon rationale to support these standards for maximized. Although this approach
monoxide standard. See boilers and hydrochloric acid maximizes the deviation among runs
§§ § 63.1216(a)(5)(i), 1217(a)(5)(i), and production furnaces. containing nondetect measurements, the
1218(a)(5)(i). Today’s rule requires continuous test condition average is lower because
We disagree with the commenter’s compliance with either a carbon we no longer assume the nondetect
assertion that the total hydrocarbon monoxide and hydrocarbon standard, in analyte is present at the level of
compliance demonstration during the combination with a destruction and detection. See response to comments
compliance test is insufficient. Sources removal efficiency standard, as discussion below for more information
are required to establish numerous surrogates to control organic HAP. We on this statistical approach to address
operating requirements based on conclude that sources which comply variability of nondetects.
operating levels that were demonstrated with these standards are operating We use this measurement imputation
during the test, including minimum under efficient combustion conditions, approach to address variability of
operating temperature, maximum feed assuring non-dioxin organic HAP are feedrate data sets containing nondetects
rates, minimum combustion zone being oxidized, thus limiting emissions for source ranking purposes and to
residence time, and operating to levels reflecting MACT. Efficient address variability of emissions data
requirements on the hazardous waste combustion of hazardous waste sets containing nondetects when
firing system that control liquid waste minimizes emissions of organic HAP calculating floors. We do not apply the
atomization efficiency. Sources must that are fed to the combustion chamber measurement implementation approach
comply with these operating as well as emissions attributable to to system removal efficiency (SRE) data
requirements on a continuous basis. products of incomplete combustion that sets where feedrates or emissions
Compliance with these requirements, in may form within the combustion contain nondetects, however. Statistical
addition to the requirements to comply chamber or post combustion. We are not imputation of nondetect SREs is
with the carbon monoxide and capable of issuing emission standards complicated given that SRE is derived
destruction and removal standards, for each organic HAP because of data from feedrate and emissions data, both
adequately assure sources are limitations and because such emission of which could contain nondetect
controlling organic HAP emissions to standards may not be replicable by measurements.147 Our inability to apply
MACT levels. individual sources or duplicable by the the imputation approach to SREs is not
Comment: A commenter states that other best performing sources because of a major concern, however, because
EPA’s proposed use of surrogates for the complex nature of combustion and system removal efficiency is used as a
organic HAP do not ensure that each of post combustion formation of products source ranking criterion only (i.e., it is
the organic HAP (e.g., polychlorinated of incomplete combustion. not used as the standard, except for
biphenyls and polyaromatic hydrochloric acid production furnaces
V. Additional Issues Relating to
hydrocarbons) are reduced to the level where there are no nondetect feedrate or
Variability and Statistics
of the HAP emitted by the relevant best emissions measurements), and there are
Many commenters raised issues few instances where system removal
performing sources. EPA has not shown relating to emissions variability and
the necessary correlation between either efficiencies are derived from nondetect
statistics other than those discussed feedrate or emissions data.
the total hydrocarbon or carbon above in Section III.A: (1) Variability
monoxide standards and organic HAP, dampening for data sets containing B. Using Statistical Imputation To
and neither is a reasonable surrogate nondetects; (2) imputation of variability Address Variability of Nondetect Values
according to the commenter. to address variability dampening for On February 4, 2005, EPA distributed
Response: Carbon monoxide and total data sets containing nondetects; and (3) by email to major commenters on the
hydrocarbon monitoring are widely our analysis of variance procedures to proposed rule a direct request for
used and accepted indicators of identify subcategories. We present comments on a limited number of issues
combustion efficiency, and hence comments and responses on the that were raised by the public comments
control organic HAP, which are remaining topics below. on the proposed rule. The nondetect
destroyed by combustion.146 Sources measurement imputation approach
A. Data Sets Containing Nondetects
145 See
discussed above was one of the issues
USEPA, ‘‘Technical Support Document for Comment: One commenter states that
the HWC MACT Standards, Volume III: Selection of
for which we requested comment. We
MACT Standards,’’ September 2005, Section 3.2
EPA’s approach of assuming discuss below the major comments on
and USEPA, ‘‘Final Technical Support Document measurements that are below detection the approach.
for the HWC MACT Standards, Volume III: limits are present at the detection limit Comment: Most commenters state that
Selection of MACT Standards and Technologies,’’ dampens the variability of the data set.
July 1999, Section 5.1.
they agree with either the concept or the
146 This is why almost all of the RCRA Land
Thus, the variability of ranking approach in principle but cannot
Disposal Restiction treatment standards for organic parameters is understated when ranking
waste, which standards are for the most part 147 See USEPA, ‘‘Technical Support Document for
established at an analytic detection level for the are based on the performance of combustion the HWC MACT Standards, Volume III: Selection of
organic HAP in question plus a variability factor, technology. See 40 CFR Part 268.40–43. MACT Standards,’’ September 2005 Section 7.3.
provide substantive comments. These standard for which we have nondetect developed the statistical imputation
commenters indicate they cannot data, and use the relationship to impute approach discussed in Section IV.A
provide substantive comments because the standard deviation for a data set above.
they cannot determine the implications containing nondetects.148 For 22 separate floors, we compared
of using the approach given that we did We could not perform this analysis, the results of the approaches we
not provide the resulting floor however, because: (1) We have very few considered for nondetects: (1)
calculations. One commenter suggests detected measurements for the data sets Nondetects present at the detection
that, before blindly applying this for several standards and could not limit (i.e., full detection limit approach);
arbitrary estimate of a nondetect value, establish the relationship between (2) MLE; (3) MLE combined with an
a reality check should be done to relative standard deviation and approximation approach (i.e., MLE/
validate that this is reasonable by emission concentration for those data Approximation approach; and (4)
consulting what is published on the sets; and (2) moreover, for many data statistical imputation.150 The MLE
method variability, as well as by sets where detected measurements approach was only applicable to 2 of the
checking variability factors derived for would have been adequate to establish 22 floor data sets, and the numerical
other data in the database that are above the relationship, it would have been algorithm failed to converge on an
the detection limit. problematic statistically to extrapolate answer for one of those. The MLE/
Another commenter voiced significant the relationship to the very low values Approximation approach sometimes
concerns with the approach. The assigned to the nondetect measurements results in floors that are unrealistically
commenter states that EPA contradicts (e.g., 100% of the detection limit; the high (i.e., it calculated 5 of 22 floors that
its assumption at proposal that all data value assigned by our statistical were higher than the statistical
that are reported as nondetect are imputation approach).149 imputation approach, which always
present at the detection limits by now This commenter also suggests that we produces floors that are equal to or
admitting that the true value is between check the resultant standard deviation higher than assuming nondetects are
zero and the level of detection. The after imputation by consulting what is present at the full detection limit), and
commenter concludes that EPA now published on the method variability. sometimes fails to converge on an
proposes to retreat from its assumption The commenter did not explain, answer. Because of these limitations, we
that undetected pollutants are always however, how method variability relates do not use either the MLE or MLE/
present at the detection limits not to the variability of nondetect data. Approximation approach.
because that assumption is false but Moreover, we believe that the We believe the statistical imputation
because it does not generate sufficiently imputation approach is one approach approach is preferable to the full
lenient floors. The commenter believes we could have reasonably used to detection limit approach because it: (1)
that this underscores that EPA’s estimate variability of nondetect data. Accounts for variability of data sets
statistical analysis approach cannot We first attempted to apply standard containing nondetects; (2) can be
possibly give an accurate picture of any statistical techniques to address the applied to all data sets containing
source’s actual emission levels. nondetect issue. We investigated nondetects; and (3) results in reasonable
Accordingly, it cannot possibly satisfy standard interval censoring techniques floor levels. In most cases, floors
EPA’s obligation to ensure that its floors to calculate maximum likelihood calculated using statistical imputation
reflect the average emission levels estimates (MLE) of the average and are close to those calculated by the full
achieved by the relevant best standard deviation that provide the best detection limit approach. The statistical
performing sources. imputation approach can produce
fit for a normal distribution for the data
The commenter also states that EPA’s substantially higher floors than the full
containing nondetect values, taking into
imputation approach is independently detection limit approach, however,
flawed because it assumes—again account that each nondetect data point
can be anywhere within its allowable when a relatively high nondetect is
inaccurately—that the value for a reported because of a high detection
nondetect is always either the highest interval. These techniques are not
applicable, however, to data sets where limit. Nonetheless, the statistical
value or lowest value in the allowable imputation approach calculated floors
range. In reality the undetected values all data are nondetects, as is the case for
many of our data sets. In that situation, that were 30% higher than the full
will necessarily fall in a range between detection limit approach for only 2 of
the highest and lowest, and thus yield we approximated the mean as the
average of the midpoints of the the 22 floors.
less variability than EPA would assume. We reject the comment that our
Response: We agree in theory with the nondetect intervals, and the standard
deviation as one half of the possible approach to handling nondetect data is
commenter who suggests that the results a mere manipulation to raise the floor.
of the imputation approach should be range of the data.
After working with this MLE/ The commenter observes that EPA
checked to see if it overstates variability appears to determine that its initial
for nondetect data by comparing the Approximation approach for some time
and iteratively developing complicated approach of assuming the worst-case for
results of the imputation approach with nondetect data—that the data are
the actual variability for detected algorithms to address problems as they
arose, we concluded that we needed a present at the detection limit—did not
measurements in the data set. We produce floors that were high enough,
considered comparing the relative simpler approach that could be applied
to all data sets. Accordingly, we and consequently applies another
standard deviation derived from the manipulation—statistical imputation of
imputation approach for data sets with nondetect measurements—that assumes
148 Note that, under this approach, we would
nondetects, to the relative standard the nondetect data are present at lower
continue to assume that the nondetect analyte is
deviation for the data set using a present at the detection limit. levels but nonetheless generates floors
regression analysis. Under the 149 Note that this was not the case where we use
that are even higher than before.
regression analysis approach, we a regression analysis of relative standard deviation
Although the commenter is correct
considered relating the relative standard versus total chlorine measurements to impute a
standard deviation for values below 20 ppmv that
deviation of detected data sets to the we corrected to 20 ppmv to address the low bias 150 See USEPA, ‘‘Technical Support Document for
average measurement. We would of Method 0050. In that situation, we have several the HWC MACT Standards, Volume III: Selection of
determine this relationship for each total chlorine measurements very close to 20 ppmv. MACT Standards,’’ September 2005, Section 5.4.
about the outcome of our handling of the highest and lowest, and thus yield emissions data causing the right-hand
nondetect data’the floors are generally less variability than EPA would assume. skew will be truncated when we
higher after statistically imputing Although the commenter is correct that identify the best performing sources—
nondetect measurements than if the true value of a nondetect those with the lowest emissions—to
nondetects are simply assumed to be measurement is likely to be in the range calculate floors. This moves the
present at the detection limit—our between the highest or lowest value appearance of a skewed distribution
rationale for handling nondetects is possible rather than at either extreme, toward one that is more symmetric and
sound. At proposal, we assumed that we do not know where the true value is thus, more representative of a normal
nondetects are present at the detection within that range. To ensure that distribution.
limit. We do not know (nor does anyone variability is adequately considered in In addition, our analyses showed: (1)
else) whether a nondetect value is establishing a floor, the statistical The probability plots do not suggest that
actually present at 1% or 99% of the imputation approach, by design, either assumed distribution is
detection limit. We thought that maximizes the deviation by assuming significantly or consistently better; (2)
assuming that all values were at the the nondetect value is at one end of the the data set arithmetic averages tend to
limit of detection would reasonably range or the other, whichever results in be in the neighborhood of the medians,
estimate the range of performance a a higher average for the data set. indicating the data sets are not
source could experience for these significantly skewed and more closely
C. Analysis of Variance Procedures To
nondetect measurements. This approach normal than lognormal; and (3) in some
Assess Subcategorization
inherently maximizes the average cases, neither assumed distribution
emissions but minimizes emissions We use analysis of variance (ANOVA) could be statistically rejected.151
variability. to determine whether subcategories of Comment: Some of the data sets used
Commenters on the proposed rule sources have significantly different for comparison have very few members.
state that assuming nondetects are emissions. For two subsets of emissions, This means that the within-group
present at the detection limit dampens the variance of the data between the two variance for a small data set would have
emissions variability—a consideration subsets is compared to the variance to be very low for the two groups to be
necessary to ensure that a source’s within the subsets. The ratio of these judged as separate.
performance over time is estimated two variances is called the F-statistic. Response: We agree, but note that as
reasonably. Mossville, 370 F. 3d at 1242 The larger the F-statistic the more likely the sample sizes change, the critical
(daily maximum variability must be the underlying data distributions are values are also changing depending on
accounted for in MACT standards different. To make a decision regarding the degrees of freedom.
[including floors] which must be the difference between the two subsets, Comment: Only emissions data were
achieved continuously). See also CMA, we compare this calculated F-statistic to considered in the ANOVA tests. Feed
870 F. 2d at 232 (EPA not even obligated an F-value associated with a particular rate and removal efficiency should have
to use data from plants that consistently confidence level. been considered as well.
reported nondetected values in One commenter has raised several Response: Differences between
calculating variability factors for best concerns with our use of the ANOVA subcategories in feedrates or system
performing plants). We agree with these procedure in the selection of incinerator removal efficiency are irrelevant if there
commenters, and are using the subcategories. is no significant difference in emissions
statistical imputation approach to Comment: The ANOVA procedure is between the subcategories. The purpose
address the concern. Relative to our based upon the assumption that the of considering subcategorization is to
proposed approach of assuming underlying distribution of both data sets determine if there are design, operation,
nondetect measurements are present at has a normal shape. For incinerator or maintenance differences between
the detection limit, the statistical emissions data this assumption is not subcategories that could affect the type
imputation approach reduces the valid. A log-probability plot shows that or concentration of HAP emissions and
average of the data set for a source while particulate emission data is better thus sources’ ability to achieve the floor
maximizing the deviation of the data described by a lognormal distribution. absent subcategorization. Consequently,
set. These are competing and somewhat Prior to conducting the ANOVA it is appropriate to consider emissions
offsetting factors when calculating the procedure, the data should be log- only when evaluating subcategorization.
floor for existing sources given that we transformed. Comment: The confidence level used
use a modified 99th percentile upper Response: We use probability plots, by EPA for the F-statistic in all cases
prediction limit to calculate the floor— Skewness Coefficients, and Correlation was 95 percent. If the calculated F-
the floor is the average of the test Coefficient/Shapiro-Wilks testing to statistic were equal to this 95 percent
condition averages for the best evaluate whether it is more appropriate confidence value, it would mean that
performers plus the pooled variance of to analyze emissions data for ANOVA there is only a 5 percent chance that
their runs. See CMA, 870 F. 2d at 232 and floor calculations assuming the data data for the two subsets were drawn
(upholding approach to variability for represent a normal or lognormal from the same parent distribution. A
datasets with nondetect values where distribution. We believe it is reasonable less stringent (lower) confidence level
various conservative assumptions in to assume the data represent a normal would be more appropriate for this
methodology offset less conservative distribution for several reasons. analysis.
assumptions). The purpose of the ANOVA
The commenter evaluated particulate
We further disagree with this subcategorization analysis is to
emissions for specialty incinerators (i.e.,
commenter’s view that the statistical determine if there is a significant
munitions, chemical weapons and
imputation approach is independently difference in emission levels between
mixed waste incinerators) and non-
flawed because it assumes that the value potential subcategories to warrant
specialty incinerators (all others). The
for a nondetect is always either the establishing separate floors for the
commenter log-transformed the data and
highest value or lowest value in the subcategories. Although in some cases it
allowable range. The commenter states may appear that a data set in its entirety 151 USEPA, ‘‘Technical Support Document for the
that, in reality, the undetected values may be better represented by a HWC MACT Standards, Volume III: Selection of
will necessarily fall in a range between lognormal distribution, the high MACT Standards,’’ September 2005, Section 8.2.
determined that there was only a 30 equipment are not different ‘‘classes,’’ electrostatic precipitators, etc.
percent chance that the two data sets ‘‘types,’’ or ‘‘sizes’’ of source. The Incinerators that are designed in this
could come from the same parent commenter implies that EPA justifies manner have the potential for elevated
distribution. This result, together with this subcategorization by stating that dioxin/furan emissions because dry air
the vastly different operating these sources have different emission pollution control systems provide
characteristics for the two types of characteristics, which is no less locations where surface-catalyzed
incinerators, argues for their being unlawful and arbitrary than reactions can occur (e.g., on particles on
treated as separate categories, according subcategorizing based on the pollution fabric filter bags or electrostatic
to the commenter. control devices they use. precipitator plates). Thus, for purposes
Response: A confidence level of 95% Response: We agree that it would not of dioxin/furan formation and control,
assigns a probability of 0.95 of accepting be appropriate to subcategorize source incinerators equipped with dry air
the hypothesis when there is no categories based on a given air pollution pollution systems are in fact different
difference between subcategories and control technique. See 69 FR at 403 (Jan. ‘‘types’’ of incinerators because of their
hence a probability of 0.05 of rejecting 4, 2004). As stated at proposal, we do unique pollutant generation
a true hypothesis. This reduces the not subcategorize incinerators with characteristics.
probability to 5% of rejecting a true respect to dioxin/furans based on the On the other hand, incinerators with
hypothesis. A less stringent confidence type of air pollution control device wet air pollution control systems are
level would increase the chances of used. 69 FR at 21214. For example, with generally designed to effectively reduce
rejecting a true hypothesis. The farther respect to dioxin/furans, it would not be total chlorine emissions (with the use of
apart the averages of the two potential appropriate subcategorize based on wet scrubbers) and metals and
subcategories are, the more likely they whether a source is using: (1) Good particulate matter emissions. There
are to be statistically different and the combustion practices; (2) a carbon bed; generally is a tradeoff, however, in that
more likely you are to be wrong if you (3) an activated carbon injection system; these types of incinerators may not be
hypothesize that they are not different. or (4) temperature control at the inlet to as efficient in reducing particulate
A 95% confidence level is most often its dry air pollution control device. matter and metal emissions compared to
used for ANOVA because it is generally These devices and practices are what incinerators that are equipped with
believed that being wrong one time out control dioxin/furan emissions. Today’s baghouses and dry electrostatic
of 20 is an acceptable risk for purposes final rule does not subcategorize based precipitators. These types of
of ANOVA. In addition, statisticians are on these control devices and practices. incinerators generally do not have the
comfortable with a 95% confidence Instead, our subcategorization approach potential to have elevated dioxin/furan
level because, in a normal distribution, recognizes the potential of some emissions because they do not provide
95% of the data fall within 2 (actually emission control equipment to create locations where surface catalyzed
1.96) standard deviations of the mean. pollutant emissions that subsequently reactions can occur. For purposes of
Other confidence levels could be used must be addressed.152 dioxin/furan emission formation and
for ANOVA—99% or 90%—if there is a Dioxin/furans are unique in that these control, sources with wet air pollution
good reason to deviate from the general pollutants are not typically present in control systems are thus likewise
default of 95%. A 99% confidence level the process inputs, but rather are formed different types of incinerators.154
is the second most commonly used in the combustor or in post combustion Subcategorizing dry air pollution
confidence level and is generally used equipment. The primary cause of systems and wet air pollution control
when it is very important that you be dioxin/furan emissions from systems for purposes of establishing a
sure that you are right (i.e., where you incinerators not equipped with waste dioxin/furan standard is no different
can only accept the risk of being wrong heat boilers is post combustion than subcategorizing incinerators
1 time out of 100) before you classify the formation by surface-catalyzed reactions equipped with waste heat boilers. The
populations (in this case subcategories) that occur within the dry air pollution waste heat boiler is the origin of the
system.153 This is evidenced by the dioxin/furan that is generated. These
as different. Occasionally, but much less
statistically significant higher dioxin incinerators are designed to efficiently
frequently, confidence levels of 90% or
furan emissions for incinerators with recover heat from the flue gas to
less are used. But, we note that these
dry air pollution control systems produce useful energy. A result of this
situations are so infrequent that some
compared to those without dry systems. type of incinerator design, however, is
statistics books provide tables for the
Incinerators with dry air pollution that it also provides a location where
ANOVA F-statistic only at the 95% and
systems are designed to effectively surface catalyzed reactions can occur
99% confidence levels.
control metal and particulate matter (i.e., the boiler tubes), potentially
For these reasons, we believe that the
emissions through use of baghouses, resulting in elevated dioxin/furan
95% confidence level is an appropriate
formation (and emissions if not properly
level among those we could have
152 Although we subcategorize between controlled).
reasonably selected. An alternative approach that does not
incinerators with wet or no air pollution control
VI. Emission Standards device and incinerators equipped with dry air subcategorize these sources, but rather
pollution control devices or waste heat boilers for identifies best performing sources as
A. Incinerators the floor analysis, the calculated dioxin furan floors
those sources with the lowest emissions
for both subcategories for existing sources were
Comment: A commenter states that determined to be less stringent than the current irrespective of whether they have a wet
EPA’s subcategorization (and interim standard. Subsequently, the final rule
assignment of differing dioxin/furan emission limitations for both subcategories are, for 154 A similar analogy applies to incinerators that
the most part, identical, and equivalent to the are not equipped with air pollution systems. These
standards as a result) between interim standard. See USEPA, ‘‘Technical Support incinerators are not designed to control emissions
incinerators with wet or no air pollution Document for the HWC MACT Standards, Volume of metals, chlorine, and particulate matter (perhaps
control device and incinerators III: Selection of MACT Standards,’’ September 2005, because emission levels are low due to low HAP
equipped with dry air pollution control Section 10.1, for further discussion. feed levels). Similar to incinerator types with wet
153 See USEPA, ‘‘Technical Support Document for systems, this design does not provide the locations
devices or waste heat boilers is unlawful the HWC MACT Standards, Volume IV: Selection for surface catalyzed reactions to occur, which leads
because incinerators equipped with a of MACT Standards,’’ September 2005, Section 3, us to conclude that these are different types of
given type of pollution control for further discussion. incinerator with respect to dioxin/furan control.
or dry air pollution control device, fuels fired to all hazardous waste 50 percent hazardous waste as fuel (i.e.,
would yield floors that would not be burning cement kilns, to identify a replacing at least half its fossil fuel with
achievable unless all the sources, MACT floor for existing and new hazardous waste) would exceed the
including the best performers, adopted cement kilns. Several commenters state interim standard, emitting mercury
beyond-the-floor technology. The that existing cement kilns should have higher than the levels allowed under
calculated dioxin/furan floor for the option to comply with either of the §§ 63.1204(a)(2) and 63.1206(b)(15) of
existing incinerators and liquid fuel following mercury standards: (1) A the interim standards.157 The hazardous
boilers using such an approach would hazardous waste feed concentration waste MTEC of 150 µg/dscm calculated
be 0.008 and 0.009 ng TEQ/dscm, limit, expressed in ppmw, based on an by the commenters is also higher than
respectively.155 All of the best evaluation of the five best performing the level currently allowed under
performing sources for these calculated sources within the commenter- § 63.1206(b)(15) of the interim
floors had either wet air pollution submitted dataset (documenting day-to- standards. Since sources cannot
systems or no air pollution control day levels of mercury in the hazardous backslide from the levels of the interim
systems. The floor technology used by waste over a three year period); or (2) a standards, if we were to accept the
these sources is good combustion hazardous waste maximum theoretical commenters’ floor analysis results as
practices. As a result, these floor levels emissions concentration (MTEC), presented (which we are not), then we
would not be replicable by these best expressed in units of µg/dscm, would ‘‘cap’’ each calculated standard
performing sources nor duplicable by developed by projecting emissions of (i.e., 3.3 ppmw hazardous waste feed
other sources through use of the same the best performing sources assuming concentration and 150 µg/dscm in stack
good combustion practices because of mercury concentrations in the emissions) at the interim standard level.
the uncertainties associated with hazardous waste were at the source’s This would result in a mercury standard
dioxin/furan generation mechanisms 99th percentile level in the commenter- for existing sources of 3.3 ppmw
and rates that can vary both within submitted dataset. To identify the best hazardous waste feed and a hazardous
sources and across sources, potentially performing sources, the commenter waste feed MTEC of 120 µg/dscm or 120
leading to significant variability in suggests selecting the five sources with µg/dscm as a stack gas concentration
emission levels.156 Sources equipped the lowest median mercury limit. We note this is similar to the
with wet or no air pollution systems concentrations in the dataset. For mercury standard adopted today: a
would thus likely be required to install existing sources, the commenters’ hazardous waste feed concentration
carbon systems to comply with these evaluation yields a hazardous waste limit of 3.0 ppmw and a hazardous
standards, a technology used by only feed concentration limit of 3.3 ppmw waste feed MTEC of 120 µg/dscm or 120
four incinerators (none of which were and a stack concentration emission limit µg/dscm as a stack gas concentration
best performers in the above discussed of 150 µg/dscm (rounded to two limit. For an explanation of why we
floor analysis). Such an outcome should significant figures and considering derived a level of 3.0 ppmw from the
be viewed as a beyond-the-floor mercury contributions only from the data, see Section 7.5.3 of Volume III of
technology and therefore assessed hazardous waste). For new cement kilns, the Technical Support Document.
pursuant to the factors enumerated in the commenters recommend a mercury The commenters’ suggested new
section 112(d)(2). Furthermore, it is standard in the format of a hazardous source mercury standard of 1.9 ppmw in
unclear, and perhaps doubtful, that waste feed concentration limit only, the hazardous waste has the same
these floors would be achievable by expressed in ppmw, based on the single deficiency. New sources with a
these sources even if they were to install source with the lowest 99th percentile hazardous waste fuel replacement rate
beyond-the-floor controls such as level of mercury in hazardous waste. of approximately 75% could emit
activated carbon systems because no The commenters recommend a mercury mercury at levels higher than currently
sources using activated carbon are standard of 1.9 ppmw for new sources. allowed under the interim standards.
currently achieving those floor levels. After capping the calculated standard at
We therefore conclude that it is Response: We agree with commenters
the interim standard level, we would
appropriate, and necessary, to that the commenter-submitted dataset
identify the mercury standard for new
subcategorize these types of incinerators documenting the day-to-day levels of
sources as a hazardous waste
for purposes of calculating dioxin/furan mercury in hazardous waste fuels fired
concentration limit of 1.9 ppmw in the
floor standards. to all hazardous waste burning cement
hazardous waste and a hazardous waste
kilns is the best available data to
B. Cement Kilns feed MTEC of 120 µg/dscm or 120 µg/
identify floor levels for existing and new
dscm as a stack gas concentration limit.
1. Hg Standard cement kilns. See discussion in Part
For reasons discussed in Section 7.5.3 of
Four, Section I.D. However, we disagree
Comment: Several commenters Volume III of the Technical Support
with the commenters’ suggested format
recommend that EPA use a commenter- Document, this is indeed the mercury
of the mercury standard for existing
submitted dataset, which includes three standard we are promulgating for new
sources. Establishing the mercury
years of data documenting day-to-day cement kilns.
standard as the commenters’ suggest
levels of mercury in hazardous waste The commenters also suggest that the
(i.e., 3.3 ppmw in the hazardous waste
best performing sources should be
feed or 150 µg/dscm as a hazardous
155 See USEPA, ‘‘Technical Support Document for identified as those with the lowest
the HWC MACT Standards, Volume III: Selection of waste MTEC) fails to consider the
three-year median concentration of
MACT Standards,’’ September 2005, Section 20 and interim mercury standards. As
Appendix C, tables labeled ‘‘E-INC-all-DF’’ and ‘‘E-
mercury in hazardous waste. Although
discussed in Part Four, Section III.E,
LFB-all-DF’’. this approach would be permissible, we
there can be no backsliding from the
156 Dioxin/furan formation mechanisms are conclude that it is more appropriate to
complex. Sources equipped with wet or no air levels of performance established in the
identify the best performers (or single
pollution control systems cannot rely on good interim standards. While not every
combustion practices alone to achieve these floor
best performer for new sources) by
source feeding hazardous waste with a
levels because they cannot ‘‘dial in’’ to a specific
emission level, as is the case with typical back-end
maximum mercury concentration of 3.3 157 USEPA, ‘‘Technical Support Document for
control systems that control particulate matter and ppmw would exceed the interim HWC MACT Standards, Volume III: Selection of
metals, for example. See Part Four, Section IV.B. standard, most sources using more than MACT Standards,’’ Section 23.4, September 2005.
selecting those with the lowest 99th 2. Total Chlorine as we do for other sources. One
percentile upper level mercury Comment: One commenter states that commenter objects to our determination
concentrations. (This is not a the proposed MACT floor approach is that an SRE of 90 percent is
statistically determined upper inconsistent with the statutory representative of the best performing
prediction limit; there is sufficient data definition of MACT because EPA’s sources because we have not established
for an arithmetically calculated 99th selection of a routinely achievable a MACT SRE—the average SRE
percentile to reliably reflect sources’ system removal efficiency (SRE) was achieved by the best performing
performance.) We believe that this arbitrary and not representative of the sources.
approach best accounts for the best performing sources. Instead, the There is no doubt that the cement
variability experienced by best commenter suggests EPA identify a manufacturing process is capable of
performing sources over time. MACT SRE based on the five sources capturing significant quantities of
with the best SREs and apply that SRE chlorine when favorable conditions
A detailed discussion of the MACT exist within the kiln system. Our usual
floor analysis for existing and new to the MACT chlorine feed level. Later,
in supplemental comments, the same approach of establishing an SRE by
cement kilns is presented in Section ranking the most efficient SREs taken
7.5.3 of Volume III of the Technical commenter suggests two alternative
approaches to identify a floor level. One from individual compliance tests,
Support Document. In summary, the however, would result in a standard that
mercury standard for existing cement approach applies a ranking
methodology based on emissions and would not be achievable because it may
kilns is 3.0 ppmw in the hazardous not be duplicable by the best performers
chlorine feed, and the second suggested
waste feed and 120 µg/dscm as a or certainly would not be replicable by
approach applies a triple ranking
hazardous waste maximum theoretical others, given that it is a function of
method based on emissions, feed, and
emission concentration feed limit or 120 various highly variable parameters,
chlorine SRE. Other commenters,
µg/dscm as a stack gas concentration however, supported EPA’s proposed especially levels of alkali metals (e.g.,
limit. For new sources the mercury approach. sodium and potassium) and volatile
standard is 1.9 ppmw in the hazardous Response: We are adopting the same compounds (e.g., chlorine and sulfur) in
waste feed and 120 µg/dscm as a approach we proposed at 69 FR at the raw materials. Alkalis and volatiles
hazardous waste maximum theoretical 21259. As we explained, this is a variant vary at a given best performer facility (in
emission concentration feed limit or 120 of the SRE/Feed approach, the variant fact, at all facilities) as different strata
µg/dscm as a stack gas concentration involving the degree of system removal are mined in the quarry, and across
limit.158 efficiency achieved by the best facilities due to different sources of raw
Comment: Two commenters oppose performing sources. In summary, to materials. Raw material substitution is
EPA’s proposed approach to base determine the floor level we first infeasible and counter to the objective of
compliance with the mercury standard identify the best performing sources producing quality product (i.e., a
on averaged annual emissions. The according to their hazardous waste product with low alkali content).
commenters state an annual average chlorine feedrate. The best performing Cement kilns thus are not able to
would allow mercury emissions to sources are those that have the lowest design or operate to achieve a specific
exceed the interim standard because a maximum theoretical emissions SRE at the high (most efficient) end of
source could burn high concentrations concentration (MTEC), considering the range of test conditions. This is
of mercury waste over a short period variability. We then apply an SRE of 90 demonstrated by our calculations of
percent (the specific point in system removal efficiency data, which is
and still comply with an annual limit by
contention) to the best performing essentially a collection of performance
burning low concentration wastes at
sources’ total MTEC (i.e., thus ‘‘snapshots.’’ See SRE data summarized
other times. These commenters support
evaluating removal of total chlorine in Table 1 at the end of this response;
the concept of a 12-hour rolling average
across the entire system, including see also Mossville, 370 F. 3d at 1242
feedrate limit (i.e., the current
chlorine contributions to emissions (maximum emission variability
requirement under the interim
from all feedstreams such as raw associated with raw material variability
standards) in conjunction with an
materials and fossil fuels) to identify the needs to be accounted for in MACT
emission standard no less stringent than
MACT floor, which is expressed as a floor determination since the standard
the interim standard.
stack gas emissions concentration in must be met at all times under all
Response: We agree with these parts per million by volume. This operating conditions). The performance
comments. Cement kilns must establish approach defines the MACT floor as an data of the ‘‘apparent’’ best performers—
a 12-hour rolling average feedrate limit emission level that the best performing upwards of 99 percent—identified by
of mercury to comply with these sources could achieve if the source the commenter are simply a snapshot in
standards. The mercury standards for limits the feedrate of chlorine in the the possible range of performance and
cement kilns are ‘‘capped’’ at the hazardous waste to the MACT level (i.e., are not replicable in the future due to
interim standard level to prevent the level achieved by the average of the factors which are uncontrollable by the
backsliding from the current level of best performing five sources) while also source, as just explained. In
performance. This is accomplished by achieving an SRE that accounts for the confirmation, cement kilns achieving
expressing the standard as a limit on the inherent variability in raw material this level of removal in one test proved
mercury concentration in the hazardous alkalinity and (to a lesser degree) incapable of replicating their own result
waste (with the rolling average) and cement kiln dust recycle rates, and in other tests even though individual
either an emission concentration limit production requirements. 69 FR at sources each have their own proprietary
or hazardous waste maximum 21259. source of raw materials. See results in
theoretical emission concentration feed Under this approach, we are table for Giant (SC), Essroc (IN), Holcim
limit. See § 63.1209(l)(1)(iii). evaluating hazardous waste feed control (MO), Giant (PA), and LaFarge (KS) all
158 Please note that we do not regard this standard emission limit which is measured at the stack. EPA 112(h)(1) apply when a work practice is the
as a work practice standard under section 112(h)(1) believes the special requirements of section exclusive standard.
of the Act, because part of the standard includes an
of whom would violate a 99 + percent standard based on their own operating

results.
TABLE 1.—SUMMARY OF SYSTEM REMOVAL EFFICIENCY DATA FOR WET PROCESS CEMENT KILNS 159
Average SRE
Number Runs Low SRE Run High SRE Run
Facility of All Runs
in Data Base (%) (%) (%)
LaFarge (OH) ................................................................................................... 3 99.1 99.4 99.3

Giant (SC) ........................................................................................................ 24 95.5 99.8 99.0
Essroc (IN) ....................................................................................................... 13 97.3 99.9 98.7
Holcim (MO) ..................................................................................................... 6 96.4 99.9 98.4
LaFarge (KS) ................................................................................................... 12 95.7 99.3 98.1
Giant (PA) ........................................................................................................ 17 87.7 99.4 97.1
Continental (MO) ............................................................................................. 3 95.7 97.0 96.5
Ash Grove (AR) ............................................................................................... 37 85.1 98.8 95.1
Texas Industries (TX) ...................................................................................... 6 88.8 97.0 93.6
Holcim (MS) ..................................................................................................... 9 76.5 99.2 90.0
159 See Section 3.6 of Volume II (Specific MACT Standards) of Comment Response Document, September 2005.
However, the data indicate that SRE is a separate and unique source in the are called ‘‘data in lieu of’’ sources in
reasonably quantifiable to a point. Based SRE/Feed MACT floor analysis for our data based on the RCRA provisions
on our data base of system removal cement kilns.161 Commenters state that under § 266.103(c)(3)(i). We
efficiency information from 130 test the approach is an improper way to acknowledge that co-located sources
conditions where total chlorine was perform a statistical analysis and may in fact share certain similar
evaluated, we conclude that a system reduces the variability in emissions that operation features (e.g., use of raw
removal efficiency of 90 percent is a otherwise would be observed in a material from the same quarry, use of
reasonable estimate of MACT SRE.160 MACT pool of five unique sources. the same coal and hazardous waste burn
We also reject the commenter’s three Variability is reduced because co- tank to fire the kilns); however, given
suggested alternative approaches to located kilns at the same plant share that the co-located sources (except those
identify a MACT SRE to apply to the many of the factors that comprise front- designated as data in lieu of) are not
MACT feed level. The commenter’s end and back-end controls. As a result, designed identically, and given their
methods all suffer a common flaw: They the calculated MACT floors for SVMs hazardous waste feed control levels
fail to recognize and take into account and LVMs for cement kilns are too were not identical during testing, we
the limitations of the total chlorine SRE stringent. The commenters’ conclude we must consider each source
data. For example, as just demonstrated, recommended solution (in instances as a unique source in the floor
available data show that considering the where co-located kilns are among the analyses.162
SRE data associated with the most top five performers) is to use only the
Comment: Commenter states that
recent compliance test as a ranking data from the best performing co-located
EPA’s proposed standards for new
factor will result in unachievable kiln, exclude any lesser performing
standards due to the varying kilns at the plant site, and then include cement kilns are unachievable due to
effectiveness of chlorine capture (which the next-best performing non-co-located problems with its accounting for
impacts emissions) depending on the kiln in the MACT pool. Implementing variability, in part because EPA did not
raw material mix characteristics. their recommendation, the commenters consider geographic differences when
Considering only the most recent state that the MACT floor for SVMs assessing feed control levels. The
compliance test data as suggested yields increases from 4.0 × 10¥4 to 7.4 × 10¥4 concentrations of hazardous
results that are unachievable because lbs/MMBtu and the floor for LVMs constituents in the waste in a particular
the best performer’s SRE data are likely increases from 1.4 × 10¥5 to 1.8 × 10¥5 region are likely to be different than in
biased high (e.g., sources that happen to lbs/MMBtu. Another commenter the waste from another geographical
test under favorable conditions are generally supports EPA’s approach region due to types of industrial sectors
likely to be identified as best noting that the variability factor applied located within each region. Sources
performers), which would not be to the emissions data already accounts cannot reasonably arrange for
replicable by even that source on a day- for variability. transportation of lower HAP wastes
to-day basis. Response: We consider sources that generated across the country and cannot
are not identical as unique sources and treat the hazardous waste to remove or
3. Semivolatile and Low Volatile Metals reduce HAP concentrations. The
emissions data and information from
Comment: Commenters oppose EPA’s unique sources are considered separate commenter cites several court decisions
proposed approach to treat each kiln as sources in the floor analyses. An that support their assertions.
example of an ‘‘identical’’ source in our Commenter believes that while this
160 As discussed a number of times earlier, we are
data base is compliance test data from represents a problem for developing
not basing any standards on feed control of HAP in a similar on-site combustion unit used both the new and existing source floors,
raw material and fossil fuel input. We instead are it is a greater predicament for the new
controlling HAP attributable to those inputs by in place of a compliance test for another
means of end-of-stack emission standards which unit (i.e., emissions testing of an
reflect removal of HAP by some type of control identical unit was not conducted). 162 Nonetheless, we analyzed the SVM and LVM
device. This approach is consistent with the These sources and their associated data floors for cement kilns as suggested by the
discussion above, since we are not basing the commenter. Results of the analysis are presented in
cement kiln chlorine standard on control of any raw ‘‘Technical Support Document for HWC MACT
material input, but rather on some type of back-end 161 It is common for cement manufacturing plants Standards, Volume III: Selection of MACT
removal efficiency. to operate multiple cement kilns at the same plant. Standards,’’ Section 8.8, September 2005.
source floor because this floor level is from many of the same generators that spite of a source using properly
based on test data for only one source. provide hazardous waste fuel to the designed and operated MACT floor
Response: We are not obligated to cement kilns. Consequently, the control technologies, including
account for varying hazardous waste commenter states that the cement controlling the levels of metals in the
feed control levels occurring because of industry’s data set of actual mercury hazardous waste. The commenter
differing HAP generation rates in feed concentrations in the hazardous opposes the proposed alternative
different locations (for commercial waste best represents the full range of standard of 42 µg/dscm, which is
sources), or because different hazardous waste fuel concentrations expressed as a hazardous waste
production process types generate that exist in the waste fuel market (see maximum theoretical emissions
higher or lower levels HAP also Part Four, Sections I.D and E). concentration. Instead, the commenter
concentration wastes. Hazardous waste Response: We disagree with the suggests that EPA maintain the
feed control is a legitimate control commenter. Although the cement alternative standard options of
technology. The commenter seems to industry’s set of mercury feed §§ 63.1206(b)(15) or 63.1206(b)(9).
suggest that we should subcategorize concentration data in the hazardous Response: We agree with the
low feeding sources and high feeding waste may represent the full range of commenter that the mercury standard
sources based on their hazardous waste concentrations for the cement kiln should address the concern of raw
feed control level. This would source category, we cannot conclude the material contributions causing an
inappropriately subcategorize sources same for lightweight aggregate kilns exceedance of the emission standard.
based on differing levels of controls, because the commenter states that the We also agree that the proposed
which we do not do. See 69 FR at 403 mercury dataset are only applicable to alternative standard of a hazardous
(January 5, 2004). Nonetheless, as its kilns.163 Further, the commenter waste maximum theoretical emissions
previously discussed, the SRE/Feed provides no specific information or data concentration of 42 µg/dscm is an
methodology lessens the impact of feed to support the conclusion that its improper standard because the
control variations across commercial suggested approach is justified for the underlying data are unrepresentative.
units because it results in fewer other lightweight aggregate kiln facility. See discussion in Part Four, Section I.E.
situations where best performing back- We also disagree with the commenter We note that the mercury standard
end controlled sources (from a as to the appropriateness of establishing promulgated today is 120 µg/dscm as a
particulate matter emissions the mercury standard in the format of a stack gas concentration limit or 120 µg/
perspective) cannot achieve the hazardous waste feed concentration dscm as a hazardous waste maximum
semivolatile and low volatile metal (i.e., 3.3 ppmw for existing sources and theoretical emission concentration feed
design levels and floors. 1.9 ppmw for new sources) for limit. The alternative mercury standard
For new source standards, the single lightweight aggregate kilns. A hazardous sought by the commenter under
best performing cement kiln sources for waste feed concentration standard is § 63.1206(b)(15) is a limit of 120 µg/
semivolatile metals and low volatile improper for this source category dscm as a hazardous waste maximum
metals were not the lowest hazardous because one lightweight aggregate kiln theoretical emission concentration,
waste feed controlled source (both floors facility’s sources (although not the which is included in the mercury
were based on sources with the fourth commenter’s) controls mercury standard promulgated today. This
best, (i.e., lowest, hazardous waste feed emissions using wet scrubbing. Thus, a should address the commenter’s
control level). We therefore do not hazardous waste feed concentration concern.
believe these sources are atypically low standard would inappropriately limit Comment: One commenter supports a
hazardous waste feeders relative to the the mercury concentration in hazardous mercury standard with short-term
other best performing sources in the waste for sources that use control compliance limits (e.g., 12-hour rolling
existing source MACT pools. equipment capable of capturing average feedrate limits) as opposed to
mercury. A source with control the annual limit proposed.
C. Lightweight Aggregate Kilns equipment should not be restricted to a Response: For reasons discussed in
1. Mercury Standard hazardous waste feed concentration Part Four, Section I.E, we are using a
standard that is based on sources that different mercury dataset than at
Comment: One commenter, an can only control mercury emissions proposal. We solicited comment on a
operator of lightweight aggregate kilns through limiting the amount of mercury floor approach using these data in a
subject to this rule, recommends that in the hazardous waste. notice 164 sent directly to certain
EPA establish the mercury standard for In any case, as explained earlier in commenters. We are adopting that
lightweight aggregate kilns at a our discussion of cement kiln mercury approach today. The monitoring
hazardous waste feed concentration standard, we believe that it is preferable requirements of the mercury standard
limit of 3.3 ppmw for existing sources to establish an emission standard to for lightweight aggregate kilns includes
and 1.9 ppmw for new sources, which assure that the actual amount of short-term averaging periods (i.e., not to
is the same standard suggested in public mercury emitted by these sources is exceed a 12-hour rolling average), as
comments by a trade organization controlled by means of a numerical recommended by the commenter.
representing hazardous waste burning standard for stack emissions.
cement kilns. The commenter notes that Comment: One commenter agrees that 2. Total Chlorine Standard
these mercury limits are appropriate for a source may not be able to achieve the Comment: One commenter supports
lightweight aggregate kilns because the mercury standard due to raw material excluding from the floor analysis all
commenter’s two lightweight aggregate contributions that might cause an lightweight aggregate kiln sources that
manufacturing facilities participate in exceedance of the emission standard in lack air pollution control devices for
the same hazardous waste fuel market as chlorine, such as scrubbing technology.
the majority of cement kilns. Moreover, 163 We note that the commenter-submitted dataset
The floor analysis should simply
the commenter maintains that its parent is not amenable for use in establishing standards exclude sources without back-end
expressed in a thermal emission format because
company also owns and operates two sufficient information on the characteristics of the controls according to the commenter.
cement kilns and that its lightweight hazardous waste (e.g., heating value of hazardous
aggregate kilns receive hazardous waste waste) were not provided. 164 See docket item OAR–2004–0022–0370.
Response: We disagree. For the final these changes is a lower national The commenter suggests that their
rule, we are using the SRE/Feed MACT emissions estimate for dioxin/furans for boiler be subject to the incinerator
floor approach which defines best lightweight aggregate kilns. We now mercury standard because the mixed
performers as those sources with the estimate an incremental reduction in waste has far higher concentrations of
best combined front-end hazardous dioxin/furan emissions of 1.06 grams mercury than wastes burned by other
waste feed control and back-end air TEQ per year with costs ranging boilers and, as a consequence, the boiler
pollution control efficiency. The between $1.6 and $2.2 million per is more incinerator-like with respect to
commenter’s suggestion would exclude additional gram of dioxin/furan TEQ the feedrate of mercury.
emissions data from two of the three removed. Based on these costs and Response. We agree with the
facilities in this source category even consideration of the non-air quality commenter’s suggestion. The final rule
though valid emissions data from these impacts and energy requirements subjects this commercial liquid fuel
sources are available (and therefore (including more waste generated in the boiler to the mercury standard for
ordinarily to be used, see CKRC, 255 F. form of spent activated carbon, and incinerators. We are classifying this
3d at 867), and these sources achieved more energy consumed), we conclude source as a separate type of source for
the best front-end hazardous waste feed that a beyond-the-floor standard for purposes of the mercury standard,
control in the category. We note that the existing and new lightweight aggregate because the type of mercury-containing
best feedrate controlled sources have kilns is no longer justified. For an waste it processes is dramatically
hazardous waste thermal feed levels that explanation of the beyond-the-floor different from that processed by other
are approximately one-fifth the level of analysis, see Section 12.1.2 of Volume liquid fuel boilers, effectively making
the source’s with back-end controls. III of the Technical Support Document. this a different type of source for
These data describe the level of We note that EPA also retains its purposes of a mercury standard 165. The
performance of sources in the category authority under RCRA section 3005(c) source thus feeds mercury at
and must be evaluated in the MACT (the so-called omnibus permitting concentrations exceeding that of any
floor analysis. We also note that even if authority) by which permit writers can boiler but at concentrations within the
we were to implement the commenter’s adopt more stringent emission standards range processed by hazardous waste
suggestion, the MACT floor results in RCRA permits if they determine that incinerators. The maximum test
would not change for existing and new today’s standards are not protective of condition average MTEC 166 for mercury
lightweight aggregate kilns because the human health and the environment. for the remaining liquid fuel boilers is
total chlorine emissions data of the 20 µg/dscm. All the liquid fuel boiler
source with back-end air pollution D. Liquid Fuel Boilers mercury data represent ‘‘normal’’ data,
controls (after considering variability) i.e., data that were not spiked. (The lack
1. Mercury Standard Not Achievable
are higher than the standards of spiked data in the liquid fuel boiler
When Burning Legacy Mixed Waste
promulgated today. Thus, the data base, in and of itself, indicates that
commenter’s suggestion also would Comment: One commenter states that these sources do not process mercury-
result in a standard that would be the proposed liquid fuel boiler mercury bearing waste and do not need the
capped by the interim standard. standard is not achievable by a operational flexibility gained by spiking
commercial boiler, DSSI (Diversified to account for occasional higher
3. Beyond-the-Floor Standards concentration mercury wastes.) DSSI’s
Scientific Services, Inc.) that burns
Comment: One commenter opposes mercury-bearing low level radioactive 2002 mercury test condition average
EPA’s proposed decision to promulgate waste that is also a hazardous waste (so- MTEC was spiked to 3500 µg/dscm. In
a beyond-the-floor standard for dioxin/ called ‘mixed waste’) that was generated other words, DSSI needs the operational
furans for existing and new lightweight flexibility to feed 175 times more
years ago (so-called, legacy waste). The
aggregate kilns based on performance of mercury than any other liquid fuel
waste is an organic liquid containing
activated carbon injection. boiler. Incinerators, on the other hand,
Response: For the final rule, we high concentrations of mercury. The
boiler is equipped with a wet scrubber had mercury MTECs that ranged to
conclude that a beyond-the-floor
which provides good mercury control— 110,000 µg/dscm in 2002. In fact, DSSI’s
standard for lightweight aggregate kilns mercury feed rate is the eighth highest
is not warranted. The Clean Air Act 93%, system removal efficiency
according to the commenter. of the 40 incinerators, including DSSI,
requires us to consider costs and non- for which we have 2002 mercury feed
air quality impacts and energy The commenter states that the rate data. DSSI’s process feed is thus
requirements when considering more proposed liquid fuel boiler mercury within the upper range of mercury feed
stringent requirements than the MACT standard is not achievable using found at incinerators.
floor. In the proposed rule, we estimated feedrate control and/or additional back- We believe it is well within the broad
that the incremental annualized end control. Waste minimization is not discretion accorded us in section
compliance costs for lightweight an option because the waste has already 112(d)(1) to subcategorize among
aggregate kilns to achieve the beyond- been generated. Further, available ‘‘types’’ and ‘‘classes’’ of sources within
the-floor standard would be national treatment capacity for mercury- a category. See also Weyerhaeuser v.
approximately $1.8 million and would bearing, low-level radioactive organic Costle, 590 F. 2d at 254, n. 70 (D.C. Cir.
provide an incremental reduction in hazardous waste is very limited. The 1978) (similar raw waste characteristics
dioxin/furan emissions of 1.9 grams only other hazardous waste combustion justify common classification) and
TEQ per year (see 69 FR at 21262). At facility authorized to treat such waste is Chemical Manufacturers Ass’n v. EPA,
proposal we judged costs of the Department of Energy incinerator at 870 F. 2d 177, 253–54 and n. 340 (5th
approximately $950,000 per additional Oak Ridge, Tennessee. Waste treatment
gram of dioxin/furan TEQ removed as volumes at that facility are restricted by 165 See CAA section 112 (d) (1)), authorizing EPA
justified, and, therefore, we proposed a the mercury feed rate limitation for the to distinguish among different ‘‘types * * * of
beyond-the-floor standard. Since incinerator. In addition, the feedrate of sources within a category or subcategory’’ in
developing MACT standards.
proposal, we made several changes to the waste cannot be practicably reduced 166 Maximum theoretical emission concentration
the dioxin/furan data base as the result because of the large back-log of waste is the feedrate normalized by gas flowrate assuming
of public comments. One implication of that must be treated. zero system removal efficiency.
Cir. 1989) (same). We note that this sources. It is not practical or feasible to most relevant sources for the purposes
boiler will be subject to the liquid fuel investigate how to remove the mercury of clarification in this case are
boiler standards for all HAP other than from wastes of such varied and unique incinerators, not liquid fuel boilers.
mercury (the only HAP where the issue origins. Accordingly, we have classified DSSI
of appropriate classification arises). Only one other facility could as an incinerator for purposes of a
Not surprisingly, given the disparity potentially treat this mixed waste, mercury standard (i.e., made it subject
in waste concentration levels, the DSSI DOE’s incinerator at Oak Ridge, to the mercury standard for
boiler, even though equipped with back Tennessee, whose permit allows the incinerators), and have included the
end control comparable to best incinerator to manage mixed waste. DSSI mercury data with the incinerator
performing commercial incinerators, However, waste treatment volumes for data when assessing mercury standards
achieves mercury emission levels less mercury-bearing wastes at that facility for incinerators.
than an order of magnitude higher than are restricted by the mercury feed rate Comment. In something of a
the other hazardous waste-burning limitation in the incinerator’s permit. contradiction, the same commenter
liquid fuel boilers, few of which use The DOE incinerator alone cannot argues that the mixed waste boiler
back end control that is effective for assure national capacity for mercury- source (DSSI) does not claim that it
mercury.167 This emission disparity bearing, low-level radioactive organic cannot meet the relevant mercury
likewise indicates that DSSI is treating hazardous waste. In addition, the back- standard for liquid fuel boilers, but only
a different type of waste than other end emission controls of the mixed that it cannot do so ‘‘using either
liquid fuel boilers. waste boiler are superior to those used feedrate control or MACT floor back end
The nature of the mercury-bearing by most incinerators, including the Oak emission control.’’ Floors must reflect
waste further confirms that it is of a Ridge incinerator. This boiler uses a the emission levels that the relevant best
different type than that processed by highly effective wet scrubbing system— sources actually achieve, not what is
other hazardous waste burning liquid the principal MACT floor back-end achievable through the use of a chosen
fuel boilers. The waste is a remediation control for mercury used by emission control technology. It is flatly
waste, a type of waste burned routinely incinerators—that achieves over 93% unlawful—and essentially
by commercial hazardous waste system removal efficiency. This is contemptuous of court—for EPA even to
incinerators but almost never by a liquid superior control compared to most entertain the source’s argument that the
fuel boiler. incinerators, including the one at Oak source should be subject to a less
Moreover, the waste is a legacy, Ridge which achieves 75 to 85% stringent emission standard based on
mixed waste generated decades ago in removal.168 the levels they believe would be
support of the United States’ strategic Thus, this mixed waste boiler is achievable through the use of one
nuclear arsenal. It is not amenable to the reasonably classified a different type of chosen control technology.
types of control all other liquid fuel source with respect to mercury waste The commenter also states that the
boilers use to reduce mercury than other hazardous waste-burning source acknowledges that it could
emissions—some type of feed control or liquid fuel boilers, based on the nature achieve a better emission level, and
other minimization technique. We of the waste burned and confirmed by apparently meet the relevant standards,
investigated whether any waste the source’s mercury emissions. We by using activated carbon. Their
minimization options are feasible for note that, although the final rule argument that doing so would generate
this waste, and find that they are not. subjects only the DSSI mixed waste large quantities of spent radioactive
Normally, waste minimization is boiler to the incinerator mercury carbon does not support its attempt to
accomplished by one of three means: standard, we would conclude that any avoid Clean Air Act requirements; the
eliminating the use of mercury in the other liquid fuel boiler with the same alternative to the source accumulating
process to prevent it from being in the fact pattern (i.e., that met the same large quantities of radioactive carbon is
waste; pretreating the waste before criteria as the DSSI boiler as discussed releasing large quantities of radioactive
burning to remove the mercury; or above) should also be subject to the and toxic pollution into the
sending it to another facility better incinerator mercury standard rather environment.
suited to handle the waste. Changing the than the liquid fuel boiler mercury Response. DSSI cannot meet the
production process to eliminate or standard. liquid boiler mercury standard because
reduce the mercury content of the waste Comment. One commenter states that it burns a unique waste that resembles
is not an option because this waste has EPA’s standards for all sources must wastes processed by hazardous waste
already been generated. Pretreatment is reflect the actual emission levels incinerators (in terms of mercury
already practiced to the maximum achieved by the relevant best sources. If concentration and provenance) and is
extent feasible by settling out and EPA wishes to subject the boiler source unlike any mercury-containing waste
separating the heavier mercury from the and incinerators to the same emission burned by the remaining liquid fuel
liquid components after thermal standards, however, it is entirely within boilers. See the earlier discussion
desorbtion. The remaining organic the Agency’s power to do so. showing that DSSI needs the operational
liquid that is burned by the mixed waste Response. We agree. There is no flexibility to feed 175 times more
boiler contains concentrations of functional difference between this boiler mercury than any other liquid fuel
mercury (in organo-mercury and other and incinerators with respect to boiler, but that DSSI’s process feed is
organic soluble forms) that are orders of mercury feed rate and the type of waste within the upper range of mercury feed
magnitude higher than burned by other processed (incinerators often treat found at incinerators.
remediation wastes). Therefore, the We agree that DSSI is processing
liquid fuel boilers. Much of the waste
different types of mercury-bearing
cannot be feasibly pretreated to remove
168 For more explanation concerning mixed waste wastes than those combusted by all
mercury because this legacy, mixed
sources, limitations on the concentrations of other liquid fuel boilers. We believe that
waste comes from many highly diverse mercury fed to these sources, and the system establishing a different mercury
removal efficiency achieved, see USEPA,
167 USEPA, ‘‘Technical Support Document for ‘‘Technical Support Document for HWC MACT
standard for DSSI is warranted, as it
HWC MACT Standards, Volume I: Description of Standards, Volume III: Selection of Standards,’’ would for any source with demonstrably
Source Categories,’’ September 2004, Section 2.4.4. September 2005, Section 8.7. unique, unalterable feedstock which is
more difficult to treat than that is that it would be very difficult to meet reasons just discussed. This source is
processed by other sources otherwise in this standard by any means other than one of two available sources that is
the same category. combustion. Moreover, as an organic authorized to treat mixed waste, and the
How DSSI chooses to comply with the liquid, the waste is readily amenable to other source is not likely to have the
incinerator mercury standard (for treatment by combustion. In addition, ability to burn mercury-bearing organic
example, whether it must use some combustion is a legal form of treatment waste in the future due to permit
other type of emissions control for the waste. EPA did not propose to limitations and size constraints.
technology) is not germane to this change or otherwise reconsider these Comment. The commenter states that
decision. We note that today’s mercury treatment standards in this rulemaking, mixed legacy waste should not be
standard for incinerators will force this and is not doing so here. We note, burned at all. If there are truly no other
source to lower its mercury emissions, however, that 40 CFR 268.42 and 268.44 facilities that are currently permitted to
since it is unlikely that it can meet provide means by which generators and dispose of mixed legacy waste, such
today’s 120 µg/dscm standard at all treatment facilities can petition the waste should be stored until a facility
times without some changes in Agency to seek different treatment that can treat such waste safely—e.g.,
operations. standards from those specified by rule, through chemical oxidation—can be
Comment. The source argues that and set out requirements for evaluating permitted.
waste minimization is not feasible for such petitions. Response. The commenter’s
legacy mixed waste that has already We note further that, because this suggestion is beyond the scope of
been generated. It is not possible to waste is radioactive, exceptional today’s rulemaking. The suggestion is
travel back in time and unmake mixed precautions need to be taken in its also illegal, since RCRA prohibits the
legacy waste that already has been handling. The nonthermal treatment storage of hazardous waste for extended
created. That obvious fact, however, alternatives mentioned by the periods. See RCRA section 3004(j); and
lends no support to their argument that commenter ignore the potential for Edison Electric Inst. v. EPA, 996 F. 2d
it should be allowed to burn mixed radiation exposure if nonthermal 326, 335–37 (DC Cir. 1993) (illegal
legacy waste with less stringent treatment is used. Concerns (some of under RCRA section 3004(j) to store
emission standards, according to one which are mentioned in DSSI’s hazardous waste pending development
commenter. comment) include: Nonthermal of a treatment technology). EPA also
Response. As discussed above, the treatment would (or could) increase notes that it retains authority under
mercury standard for liquid fuel boilers worker exposure; desire to reduce RCRA section 3005(c) (the so-called
is not achievable for this source because handling of radioactive materials in omnibus permitting authority) by which
it is a different type and class of boiler, general; need to avoid contaminating permit writers can adopt more stringent
based on the type of mercury-containing equipment that subsequently requires emission standards in RCRA permits if
hazardous waste it processes. Because decontamination or handling as they determine that today’s standards
this boiler has mercury feed rates that radioactive material; minimizing the are not protective of human health and
resemble those of incinerators—not generation of additional radioactive the environment.
liquid fuel boilers—and waste waste residues; reducing the amount of 2. Different Mercury, Semivolatile
minimization is not possible, subjecting analysis of radioactive materials, which Metals, Chromium, and Total Chlorine
the boiler to the mercury incinerator causes potential exposure, generation of Standards for Liquid Fuel Boilers
standard is a reasonable means of sub- radioactive wastes and equipment; Depending on the Heating Value of the
categorization pursuant to the wastes are varied and often of small Hazardous Waste Burned
discretionary authority provided us by volumes, which makes it difficult to
section 112(d)(1) of the Clean Air Act. Comment. Several commenters state
develop routine procedures. that liquid fuel boilers should have an
Comment. The commenter states that Nonthermal treatment alternatives are
it is entirely possible to dispose of alternative concentration-based
also not currently available to DOE to standard in addition to the thermal
mixed legacy waste without burning it. manage the diversity and volume of
Specifically, currently available emission-based standard. Liquid fuel
DOE mixed waste. It is thus our belief boilers are typically ‘‘captive’’ units that
technologies such as chemical oxidation that the commenter has not fully
and precipitation can be used to treat burn waste fuels generated from on-site
explored the implications of its or nearby manufacturing operations,
mixed legacy waste without burning it— position, especially with regard to
and without releasing mercury into the rather than accepting wastes from a
radiation exposure. wide variety of other sources. Because
air. Therefore, mixed legacy waste If the commenter wishes to pursue
should not be burned at all; it should be they have captive fuel sources, operators
this issue, EPA believes the appropriate generally do not have fuel blending
disposed of safely through the context is through the Land Disposal
application of one of these more capabilities. Liquid fuel boilers ‘‘burn
Restriction mechanisms described what they have,’’ and as such have very
advanced technologies. above.
Response. First, these wastes must be limited operational flexibility. EPA
Comment. The commenter states that
treated before they can be land should not penalize boilers that have
the source argues that feedrate control is
disposed. RCRA sections 3004(d), (g)(5), the same mass concentrations of metals
not ‘‘practical.’’ There appears to be no
and (m). They also must meet a standard or chlorine in their waste compared to
record evidence indicating what would
of 0.025 mg/l measured by the Toxicity other boilers, but which wastes have a
make feedrate control impractical and
Characteristic Leaching Procedure lower heating value than wastes burned
why any such obstacle could not be
before land disposal is permissible. 40 by other boilers. (The ‘‘penalty’’ is that
overcome.
CFR 268.40 (standard for ‘‘all other emissions limits that are normalized by
Response. Feedrate control to the
nonwastewaters that exhibit the the heating value of the hazardous waste
extent necessary to achieve the liquid
characteristic of toxicity for require that less volume of lower
fuel boiler standards is not practical for
mercury’’).169 EPA’s technical judgment heating value waste can be burned
due to its high organic content, it is in fact a liquid
compared to higher heating value fuel.)
169 Althoughthe legacy waste that DSSI is matrix, meaning that the treatment standard of This problem is made worse by the
burning is nominally classified as a nonwastewater 0.025 µg/l is effectively a total standard. limited data base for liquid fuel boilers,
the lack of historical data to verify that different means of accounting for Section 23.2 of Volume III of the
these standards are achievable over sources of different size), where the Technical Support Document for more
time, and having most or all of the standard is expressed as amount of HAP information.)
measured emissions below detection per volume of gas flow (the same Moreover, liquid fuel boilers are not
limits. In addition, most of the mercury normalizing parameter used for most of irrevocably placed in one or the other of
and semivolatile metal data EPA has in the other standards promulgated in these subcategories. Rather, the source
the data base were obtained during today’s final rule.) is subject to the standard for one or the
normal operations and while the source The commenters requested that liquid other of these subcategories based on
demonstrated compliance with RCRA’s fuel boilers be able to select the the as-fired heating value of the
chromium standard—the other metals applicable standard (i.e., to choose hazardous waste it burns at a given
data were available only because stack between normalizing parameters) and time. Thus, when the source is burning
method Method 29 reports data for all further requested that we assess the for energy recovery, then the thermal
RCRA metals, even ones that are not at performance of these units (for the emissions-based standard would apply.
issue for the compliance test. (Sources purpose of establishing concentration- When the source is burning at least in
generally elected to comply with the BIF based MACT floor levels) by using the part for thermal destruction, then the
Tier I metals emissions levels, but Tier same MACT pool of best performing concentration based standard would
III for chromium. Thus, the Method 29 sources expressed on a thermal apply. This approach is similar to how
test for chromium will give emissions emissions basis. we have addressed the issue of
results for all the metals—even those Neither of these suggestions is normalization in other rules where
not subjected to stack testing—not just appropriate. Choice of normalizing single sources switch back and forth
chromium.) parameter is not a matter of election, but among inputs which are sufficiently
Response. As explained earlier in Part rather reflects an objective different to warrant separate
Four, Section V.A., EPA has selected determination of what parameter is classification. 171
normalizing parameters that best fit the reasonably related to the activity We next considered what an
input to the combustion device. A conducted by the source. Moreover, the appropriate as-fired heating value
thermal normalizing parameter (i.e., commenter’s suggestion to use thermal would be for each liquid fuel boiler
expressing the standards in terms of emissions to measure best performance subcategory. Although we have used
amount of HAP contributed by for a concentration-based standard does 5000 Btu/lb (the heating value of lowest
hazardous waste per thermal content of not make sense. It arbitrarily assumes grade fuels such as scrap wood) in past
hazardous waste) is appropriate where that the best performers with respect to RCRA actions as a presumptive measure
hazardous waste is being used in low and high heating value wastes are of when hazardous waste is burned for
energy-recovery devices as a fuel, since identical. destruction (see, e.g. 48 FR 11159
the waste serves as a type of fuel. Using Instead, we have established two (March 16, 1983)), we do not think that
a thermal normalizing parameter in subcategories among the liquid fuel measure is appropriate here. We used
such instances avoids the necessity of boilers: those burning high and those the 5,000 Btu/lb level to delineate
subcategorizing based on unit size. burning low heating value hazardous burning for destruction from burning for
The commenters raise the other side waste. The normalizing parameter for energy recovery at a time when that
of the same issue. As the commenters sources burning lower energy hazardous determination meant the difference
point out, some liquid fuel boilers burn waste is that used for the other between regulation and nonregulation.
lower Btu hazardous waste because that hazardous waste treatment devices, gas See 50 FR 49166–167 (Nov. 29, 1985).
is the waste available to them, and those flow rate, so that the standard is This is a different issue from choosing
with waste that has a low heating value expressed as concentration of HAP per the most reasonable normalizing
are, in their words, ‘‘penalized,’’ volume of gas flow (a concentration- parameter for regulated units (i.e., units
compared to those with a high(-er) based form of the standard.) The which will be subject to a standard in
heating value. Also, since these are not normalizing parameter for sources either case).
commercial combustion units, they burning higher energy content Instead, we are adopting a value of
normally lack the opportunity to blend hazardous waste is the thermal 10,000 Btu/lb as the threshold for
wastes of different heating values to parameter used for energy recovery subcategorization. This is approximately
result in as-fired high heating value devices, such as cement kilns and the heating value of commercial liquid
fuels. If boiler standards are normalized lightweight aggregate kilns. For the fossil fuels. 63 FR 33782, 33788 (June
by hazardous waste heating value, purposes of calculating MACT floors, 19, 1998) It is also typical of current
sources with lower heating value waste the best performers are then drawn from hazardous waste burned for energy
must either reduce the mass those liquid fuel boilers burning lower recovery. Id. Moreover, EPA has used
concentration of HAP or increase the energy hazardous waste for the lower this value in its comparable fuel
waste fuel heating value (or increase the heating value subcategory, and from specification as a means of
system removal efficiency) compared to those liquid fuel boilers burning higher differentiating fuels from waste. See id.
sources with wastes having the same energy hazardous waste for the higher and Table 1 to 40 CFR section 261.38,
mass concentration of HAP but higher heating value subcategory 170. (See showing that EPA normalizes all
heating value.
Moreover, the thermal normalizing 170 We also agree that liquid fuel boilers present hand—namely that not all liquid fuel boilers burn
parameter is not well suited for a several unique circumstances, namely: they are for energy recovery—they are secondary issues that
hazardous waste that is not burned often unable to blend fuel and have limited we need to closely consider to make sure we do not
operational flexibility as a result; our data base on estimate what the best performing 12% of sources
entirely for its fuel value. In cases where these sources’ performance is relatively small; are achieving in an unreasonable manner.
the lower heating value waste is burned, much of our mercury and semivolatile metals data 171 See NESHAP for Stationary Combustion
the boiler is serving—at least in part— is at or near detection limits; and much of the Turbines, 40 CFR section 63.6175 (definitions of
as a treatment device for the lower mercury and semivolatile metals data was obtained ‘‘diffusion flame gas-fired stationary combustion
for other purposes, namely from risk burns or as a turbine’’, ‘‘diffusion flame oil-fired stationary
heating value hazardous waste. When result of Method 29 testing to demonstrate combustion turbine’’, ‘‘lean pre-mix gas-fired
this occurs, the better normalizing compliance with a RCRA chromium standard. stationary combustion turbine’’ and ‘‘lean premix
parameter is the unit’s gas flow (a While not immediately important to the topic at oil-fired stationary combustion turbine’’).
constituent concentrations to a 10,000 operating requirements between the processed, when identifying best
Btu/lb level in its specification for subcategories, you may elect to comply performing sources.
differentiating fuels from wastes. with the more stringent operating The methodology we use to identify
We next examined the waste fuel requirements that ensure compliance best performing sources for particulate
being burned at cement kilns and with the standards for both matter emissions is not affected by the
lightweight aggregate kilns, which burn subcategories. firing rate bias in the manner that metal
hazardous waste fuels to drive the Comment: EPA’s attempt to give and chlorine emissions are. This is
process chemistry to produce actual performance two different primarily because we define best
products172, to cross-check whether meanings within a single floor approach performing sources as those with the
10,000 Btu/lb is a reasonable is unlawful, unexplained, internally best back-end air pollution control
demarcation value for subcategorizing. inconsistent, and arbitrary. If EPA technology; feed control is not assessed
10,000 Btu/lb is the minimum heating believes that mass-based emissions (specifically ash feed control) for raw
value found in burn tank and test report constitute sources’ actual performance, materials, fossil fuel, or unenumerated
data we have for cement kilns and the best performing sources must be HAP metal in the hazardous waste. The
lightweight aggregate kilns 173. We those with the best mass based hazardous waste firing rate bias is
believe the cement kiln and light weight emissions—not thermal emissions. therefore not present when we identify
aggregate kiln data confirm that this is Response: As just explained, we agree the best performing particulate matter
an appropriate cutpoint, since these with this comment, and have developed sources because a source’s hazardous
sources are energy recovery devices that MACT floors independently for the two waste firing rate is not a direct factor in
blend hazardous wastes into a subcategories of liquid fuel boilers. the ranking procedure.
consistent, high heating value fuel for We also note that four of the nine best
Thus, we have defined two separate
energy recovery in their manufacturing performing liquid fuel boilers for
MACT pools based on the thermal input
process. particulate matter are equipped with
of the waste fuel and derived two
We then separated the liquid fuel fabric filters. Particulate matter
separate and consistent MACT
boiler emissions data we had into two emissions from sources equipped with
standards for sources when they burn
groups, sources burning hazardous fabric filters are not significantly
solely for energy recovery, and when
waste fuel with less than 10,000 Btu/lb affected by ash inlet loading. This is not
they do not.
and all other liquid fuel boilers, and true for metals and chlorine, given metal
performed separate MACT floor We also note that a source cannot and chlorine emissions from fabric
analyses. (See Sections 13.4, 13.6, 13.7, ‘‘pick and choose’’ the less stringent of filters tend to increase at increased feed
13.8, and 22 of Volume III of the the two standards and comply with rates. See Volume III of the Technical
Technical Support Document.) We those. The source must be in Support Document, Sections 5.3 and
calculated concentration-based MACT compliance with the set of standards 7.4. We conclude that the hazardous
standards for these sources from their that apply. waste firing rate issue is not a concern
respective mercury, semivolatile metals, 3. Alternative Particulate Matter for these sources given their particulate
chromium, and total chlorine data. Standard for Liquid Fuel Boilers matter emissions would not be
Liquid fuel boilers will need to significantly affected by increased
determine which of the two Comment: A commenter requested hazardous waste firing rates.
subcategories the source belongs in at that EPA establish standards that allow
any point in time. Thus, you must boilers the option to comply with either 4. Long-term, Annual Averaging Is
determine the as-fired heating value of a concentration-based particulate matter Impermissible
each batch of hazardous waste fired so standard or thermal emissions-based Comment: Standards expressed as
that you know the heating value of the particulate matter standard. long-term limits are legally
hazardous waste fired at all times.174 If Response: We determined that it is impermissible because those levels, by
the as-fired heating value of hazardous appropriate to express the particulate definition, would sometimes be greater
wastes varies above and below the matter emission standard as a than the average emission levels
cutpoint (i.e., 10,000 Btu/lb) at times, concentration-based standard achieved by the best performing
you are subject to the thermal emissions consistently across source categories sources. Compliance also must be
standards when the heating value is not and not to give boilers the option to measured on a continuous basis, under
less than 10,000 Btu/lb and the mass comply with a thermal emissions-based section 302(k) of the Act. Thus, floor
concentration standards when the particulate matter standard. As levels (and standards) for mercury
heating value is less than 10,000 Btu/lb. discussed in Part Four, Section III.D as expressed as long-term limits are illegal.
To avoid the administrative burden of well as the preceding section, metal and Response: The commenter maintains
frequently switching applicable chlorine concentration-based emission that the statutory command in section
standards can be biased against sources 112(d)(3)(A) to base floor standards for
172 The Norlite light-weight aggregate kiln was not that process more hazardous waste existing sources on ‘‘the average
included in this analysis because they claim they (from an energy demand perspective), in emission limitation achieved by the best
are not burning for energy recovery. The waste
Norlite burns is 4,860 Btu/lb or lower. This is
part because the SRE/Feed methodology performing 12 percent of * * * existing
indicative of a source burning solely for thermal assesses feed control of each source sources’’ precludes establishing
treatment of the waste and not, at least in part, for when identifying the best performing standards expressed as long term
energy recovery. See 40 CFR 266.100(d)(2)(ii). sources; the ranking procedure thus averages because certain daily values
173 The cement kiln burn tank data and test report
favors sources with lower percentage could be higher. We do not accept this
data shows the minimum heating values of 9,900
and 10,000 Btu/lb, respectively, for the hazardous hazardous waste firing rates (keeping all position. The statute does not state what
waste. The minimum lightweight aggregate kiln other assessment factors equal). The type of ‘‘average’’ performance EPA
heating values for hazardous waste was 10,000 Btu/ thermal emission standard format must assess. Long term, i.e., annual,
lb, excluding the Norlite source. eliminates this firing rate bias, which averaging of performance is quite
174 If you burn hazardous waste in more than one
firing nozzle, you must determine the mass-

amounts to a limitation on the amount evidently a type of average, and so is
weighted average heating value of the as-fired of raw material (hazardous waste fuel to permissible under the statutory text.
hazardous waste across all firing nozzles. an energy recovery device) that may be Moreover, it is reasonable to establish
standards on this basis (the standards because such limits allow sources to the liquid fuel boiler and take measures
being the average of the best performing average their varying feed control levels to address the issue.
sources, expressed as a long-term over time while still assuring average
5. Gas Fuel Boilers
average), where sufficient data exist. emissions over this period are below the
Indeed, since the principal health levels demonstrated by the best Comment: How can a boiler burning
concern posed by the emitted HAP is performing sources. only gaseous waste also be burning
from chronic exposure (i.e. cumulative Indeed, under the commenter’s hazardous waste? Uncontained gases are
exposure over time), long-term approach where no averaging of intra- not considered hazardous waste under
standards (which reduce the long-term source data would be allowed, sources RCRA. Why are boilers that burn only
distribution of emitted HAP) arguably would not be in compliance with the gasses part of the liquid fuel boiler
would be preferable in addressing the standards during the performance tests subcategory?
chief risks posed by these sources’ themselves. The tests consist of the Response: We agree with the
emissions. average of three data runs, so half of the commenter that boilers that burn gasses
We establish standards with long-term emissions-weighted data points would are unlikely to burn hazardous wastes.
averaging limits whenever we use be impermissibly higher than the However, gas fuel hazardous waste
normal data to estimate long-term average during the test used to derive boilers have existed in the past,177 and
performance. We do this in the few today’s emission standards. we believe we need to define a MACT
instances where there are insufficient standard for them. Therefore, we
EPA also does not see that section
data (whether normal data or included gas fuel boilers in the liquid
302(f) of the Act, cited by the
compliance test data) to estimate each fuel boiler subcategory for reasons cited
commenter, supports its position. That
source’s short term emission levels (e.g., in the proposed rule. See 69 FR at
provision indicates that the emission
mercury and semivolatile metal 21216.
standards EPA establishes must limit
standards for liquid fuel boilers).175 One the quantity, rate, or concentration of air E. General
or two snapshot data based on normal pollutants on a continuous basis. A
operations are not likely to reflect a 1. Alternative to the Particulate Matter
standard expressed as a long-term Standards
source’s short-term operating levels in average does so by constraining the
part because feed control levels can vary overall distribution of emissions to meet Comment: Commenters state that
over time.176 See Mossville, 370 F. 3d at a long-term average. Also, long term some incinerators are currently
1242 (varying feed rates lead to different limits result in emission standards that complying with the alternative to the
emission levels, and this variability are lower than those that otherwise particulate matter standard provision
must be encompassed within the floor would be implemented on a short-term pursuant to the interim standards. See
standard because the standard must be § 63.1206(b)(14). The eligibility and
basis. The short-term limit would have
met at all times). As a result, snapshot operating requirements for the
to reflect the best performing sources’
normal emissions, when averaged alternative to the particulate matter
short term emissions variability (i.e., the
together, better reflect a source’s long standard in the Interim Standards are
maximum amount of variability a source
term average emissions. An emission different than the proposed alternative
could experience during a single test
standard based on normal data that is to the particulate matter standard in the
period). National Wildlife Federation,
averaged together, but expressed as a replacement rule. Specifically, the
286 F. 3d at 571–73.
short-term limit, would not be proposed alternative to the particulate
achievable by the best performing Comment: Other commenters argued
the opposite point, that ERA has no data matter standard would no longer require
sources because it would not adequately sources to demonstrate a 90% system
account for their emissions variability. to show that an annual average is
achievable, and EPA should establish a removal efficiency or a minimum
See National Wildlife Federation v. hazardous waste metal feed control
EPA, 286 F. 3d at 572–73 (‘‘[c]ontinuous longer averaging period.
level to be eligible for the alternative.
operation at or near the daily maximum Response: We believe that all sources
Commenters request that EPA clarify in
would in fact result in discharges that can achieve the mercury and
the final rule that the proposed
exceed the long-term average. Likewise, semivolatile metals standards for liquid
alternative to the particulate matter
setting monthly limitations at the 99th fuel boilers on an annual basis using
standard supersedes the requirements in
percentile would not insure that the some combination of MACT controls, the Interim Standards.
long-term average is met’’). Long-term i.e., feed control, back end control, or Response: We are finalizing the
limits better account for this variability some combination of both. We agree alternative to the particulate matter
that we have a small data set for these standard for incinerators as proposed,
175 Two emission standards in this rulemaking are standards, but also believe that it is with the exception that the alternative
based on normal data but are expressed as short intuitive that a liquid fuel boiler can metal emission limitations have been
term limits (the mercury standards for lightweight meet these standards on an annual
aggregate and cement kilns). However, in these revised as a result of database changes
instances we had enough normal data to reasonably basis, because one year is sufficiently since proposal. See § 1219(e) and part
estimate each source’s maximum emissions, thus more than any seasonal (i.e., several three, section II.A. We considered
allowing us to express the standard as a short term month long) production of certain items superseding the interim standard
limit. See USEPA, ‘‘Technical Support Document that may not be represented by the tests
for HWC MACT Standards, Volume III: Selection of alternative to the particulate matter
MACT Standards,’’ September 2005, Sections 11.2 we have. standard requirements (63.1206(b)(14))
and 12.2. This informs us that an average of less immediately (upon promulgation) by
176 This is not the case for floors that are based
than a year may not be achievable. It replacing it with the revised alternative
on compliance tests because sources spiked their does not inform us that averaging of
hazardous wastes to account for variability in
hazardous waste feedrate. See Part Four, Section more than a year is required, since 177 For example, sources 2014 and 2015 owned by
III.C above. Normal data, however, are a snapshot variations that occur with a year are Environmental Purification Industries in Toledo,
of what occurred on that day and are not likely to averaged together. An annual average is Ohio, were considered hazardous waste boilers at
be representative over the long term, especially for the time the Phase II data base was noticed in the
mercury and semivolatile metals for liquid fuel
sufficient for a source to determine June 27, 2000, despite the fact that these boilers
boilers, where these limited data were almost whether an individual waste stream burned only gasses. These boilers have since
entirely below the analytic detection limit. impacts negatively on the compliance of stopped burning hazardous waste.
standard provisions finalized in today’s 2. Assessing Risk as Part of are established under Section 112(d) of
rule. Although the eligibility Consideration of Nonair Environmental the Clean Air Act. A precedent-setting
requirements for the alternative to the Impacts change of the magnitude that EPA has
particulate matter standard finalized Comment: Commenter states that EPA raised should be discussed openly and
today are less stringent than the interim has inappropriately failed to consider carefully with all affected parties, rather
standard requirements, the metal emissions of persistent bioaccumulative than being buried in several individual
emission limitations that are also pollutants in its beyond-the-floor proposed standards.
required by the alternative finalized analysis despite EPA’s acknowledgment Response: Including health-based
that these HAPs have non-air quality compliance alternatives for hazardous
today are by definition equivalent to or
health and environmental impacts. waste combustors does not mean that
more stringent than the metal
Response: EPA has taken the EPA will automatically provide such
limitations in the interim standard alternatives for other source categories.
alternative. We therefore cannot consistent position that considerations
of risk from air emissions have no place Rather, as has been the case throughout
completely supersede the interim the MACT rule development process,
standard provisions immediately (upon when setting MACT standards, but
rather are to be considered as part of the EPA will undertake in each individual
promulgation) because sources have rule to determine whether it is
residual risk determination and
three years to comply with more appropriate to exercise its discretion to
standard-setting process made under
stringent standards. We are instead use its authority under CAA section
section 112 (f) of the statute. EPA thus
revising the interim standard provisions interprets the requirement in section 112(d)(4) in developing applicable
of § 63.1206(b)(14) to only reflect the 112 (d) (2) that we consider ‘‘non-air emission standards. Stakeholders for
revised alternative standard eligibility quality health and environmental those affected rules will have ample
criteria (specifically, we have removed impacts’’ as applying to the by-product opportunity to comment on the
the requirements to achieve a given outputs from utilization of the pollution Agency’s proposals.
system removal efficiency and control technology, such as additional Comment: One commenter states that
hazardous waste metal HAP feed control amount of waste generated, and water the proposed approach is contrary to the
level).178 These eligibility criteria discharged.179 EPA’s interpretation was intent of the CAA which explicitly calls
revisions become effective immediately upheld as reasonable in Sierra Club v. for a general reduction in HAP
with respect to the interim standards EPA, 353 F. 3d 976, 990 (D.C. Cir. 2004) emissions from all major sources
(Roberts, J.). nationwide through the establishment of
because they are less stringent than the
MACT standards based on technology,
current requirements. Sources should VII. Health-Based Compliance rather than risk, as a first step.
modify existing Notifications of Alternative for Total Chlorine Response: For pollutants for which a
Compliance and permit requirements as health threshold has been established,
necessary prior to implementing these A. Authority for Health-Based
Compliance Alternatives CAA section 112(d)(4) allows the
revised procedures. Administrator to consider such
Comment: One commenter is opposed Comment: One commenter states threshold level, with an ample margin
to the alternative to the particulate there is no established health threshold of safety, to establish emission
for either HCl or chlorine. standards.
matter standard because it ignores the
Response: Although EPA has not Comment: One commenter states that
health effects/benefits that are developed a formal evaluation of the
attributable to particulate matter. the proposed approach would take the
potential for HCl or chlorine national air toxics program back to the
Response: Particulate matter is not carcinogenicity (e.g., for IRIS), the time-consuming NESHAP process that
defined as a hazardous air pollutant evaluation by the International Agency existed prior to the Clean Air Act
pursuant the NESHAP program. See for Research on Cancer stated that there Amendments of 1990.
CAA 112(b)(1). We control particulate was inadequate evidence for Response: We disagree that allowing a
matter as a surrogate for metal HAP. See carcinogenicity in humans or health-based compliance alternative in
part four, section IV.A. As a result, a experimental animals and thus the final rule will alter the MACT
particulate matter standard is not concluded that HCl and chlorine are not program or affect the schedule for
necessary in instances where metal HAP classifiable as to their carcinogenicity to promulgation of the remaining MACT
emission standards can alternatively humans (Group 3 in their categorization standards. Today’s rule is the last
and effectively control the nonmercury method). Therefore, for the purposes of MACT rule to be promulgated, and the
metal HAP that is intended be this rule, we have evaluated HCl and health-based compliance alternative did
controlled with the surrogate particulate chlorine only with regard to non-cancer not delay promulgation of the rule.
matter standard. The alternative to the effects. In the absence of specific Comment: A commenter is concerned
scientific evidence to the contrary, it has that the proposal would remove the
particulate matter standard in the final
been our policy to classify non- benefit of the ‘‘level-playing field’’ that
rule accomplishes this. We acknowledge
carcinogenic effects as threshold effects. would result from the proper
that particulate matter emission RfC development is the default
reductions result in health benefits. implementation of technology-based
approach for threshold (or nonlinear) MACT standards.
That in itself does not give EPA the effects. Response: Providing health-based
authority under § 112(d)(2) to directly Comment: One commenter states that compliance alternatives in the final rule
regulate particulate matter, however. the proposal is an inappropriate forum for sources that can meet them will
for bringing forward such a significant assure the application of a uniform set
change in the way that MACT standards of requirements across the nation. The
178 Sources can only use § 63.1206(b)(14) for
final rule and its criteria for
purposes of complying with the interim standards. 179 See USEPA, ‘‘Technical Support Document for
After the compliance date for today’s rule,
demonstrating eligibility for the health-
the HWC MACT Standards, Volume V: Emission
incinerators electing to comply with the alternative Estimates and Engineering Costs,’’ September 2005,
based compliance alternatives apply
to the particulate matter standard must comply with Section 6, for a discussion of the non-air impact uniformly to all hazardous waste
the provisions found in § 63.1219(e). that were assessed for this final rule. combustors except hydrochloric acid
production furnaces. The final rule design, equipment, work practice or (including floor standards) would be.180
establishes two baseline levels of operational standard promulgated under EPA’s use of such standards is not
emission reduction for total chlorine, this chapter. EPA’s alternate risk-based limited to situations where every source
one based on a traditional MACT emission standard will limit the in the category or subcategory can
analysis and the other based on EPA’s quantity, rate or concentration of the comply with them. As with technology-
evaluation of the health threat posed by emissions. The commenter states that based standards, a particular source’s
emissions of HCl and chlorine. All there is no requirement in the definition ability to comply with a health-based
hazardous waste combustor facilities that specifies where the emission standard will depend on its individual
must meet one of these baseline levels, standard is to be measured, nor is there circumstances, as will what it must do
and all facilities have the same such a requirement anywhere in the to achieve compliance.
opportunity to demonstrate that they statute. In developing health-based
can meet the alternative health-based Second, the commenter notes that compliance alternatives under section
emission standards. We also note that EPA’s proposed exposure-based limit 112(d)(4), EPA seeks to ensure that the
additional uniformity is provided by will result in facilities establishing concentration of the particular HAP to
limiting the health-based compliance operating parameter limitations, or which an individual exposed at the
alternatives for incinerators, cement OPLs. These OPLs qualify as emission upper end of the exposure distribution
kilns, and lightweight aggregate kilns to limitations because they are is exposed does not exceed the health
the emission levels allowed by the ‘‘operational standards’’ being threshold. The upper end of the
Interim Standards. promulgated under section 112, exposure distribution is calculated
Comment: Several commenters state according to the commenter. They will using the ‘‘high end exposure estimate,’’
that site-specific emission limits are be measured at the facility, not at the defined as ‘‘a plausible estimate of
inappropriate under section 112(d)(4) point of exposure. Finally, the individual exposure for those persons at
because they are not emission commenter reasons that the limitations the upper end of the exposure
standards. One commenter asserts that EPA is establishing are uniform. They distribution, conceptually above the
the Agency’s position that the limits are uniformly protect the individual most 90th percentile, but not higher than the
based on uniform procedures is flawed exposed to emission levels no higher individual in the population who has
because the process allows ‘‘any than a hazard index of 1.0. the highest exposure’’ (EPA Exposure
scientifically-accepted, peer-reviewed Consequently, the commenter believes Assessment Guidelines, 57 FR 22888,
risk assessment methodology for your that there is nothing in the statute that May 29, 1992). Assuring protection to
site-specific compliance prevents the Agency from promulgating persons at the upper end of the
demonstration.’’ This is not a ‘‘uniform’’ exposure-based emission standards. exposure distribution is consistent with
procedure, according to the commenter. Response: We agree with the the ‘‘ample margin of safety’’
There are a host of variables that commenters who believe the Agency requirement in section 112(d)(4).
influence the results of an accepted has the authority to establish health- We agree with the view of several
methodology. The commenter reasons based compliance alternatives under a commenters that section 112(d)(4) is
that, without some standardization of national exposure standard. In appropriate for establishing health-
those variables, there is no uniform or particular, we agree with the commenter based compliance alternatives for total
standard analysis. Each permitting that the health-based compliance chlorine for hazardous waste
authority could establish its view of alternatives are national standards since combustors other than hydrochloric
appropriate variables; there would be no acid production furnaces. Therefore, we
they provide a uniform and national
national consistency. have established such compliance
measure of risk control, and also that
Several other commenters assert that alternatives for affected sources in those
EPA has the authority to establish an the health-based compliance
alternatives are ‘‘emission standards’’ categories. Affected sources which
exposure-based emission limit for total believe that they can demonstrate
chlorine. One commenter notes that one because they limit the quantity, rate or
concentration of total chlorine compliance with the health-based
issue that often arises when considering compliance alternatives may choose to
risk-based standards is whether EPA has emissions.
Section 112(d)(4) authorizes EPA to comply with those compliance
authority under section 112 to establish alternatives in lieu of the otherwise
an exposure-based emission limit. The bypass the mandate in section 112(d)(3)
in appropriate circumstances. Those applicable MACT-based standard.
commenter states that the concern Comment: One commenter states that
seems to be that some stakeholders circumstances are present for hazardous
waste combustors other than the risk assessments would not provide
construe the Act’s statutory provisions an ample margin of safety because
as requiring uniform emission hydrochloric acid production furnaces.
Section 112(d)(4) provides EPA with background exposures are not taken into
limitations at all facilities, rather than account. There is no accounting for
emissions that are measured at places authority, at its discretion, to develop
health-based compliance alternatives for other chlorine compounds from other
away from the source and that vary from sources at the facility, or from other
facility to facility. The commenter does HAP ‘‘for which a health threshold has
been established,’’ provided that the neighboring facilities. The commenter
not see any legal impediment to believes that there is no evidence in the
establishing exposure-based limits. standard reflects the health threshold
‘‘with an ample margin of safety.’’ section 112(f) residual risk assessments
The commenter notes that, first, under produced thus far that emissions from
section 112, EPA has authority to Both the plain language of section
112(d)(4) and the legislative history collocated sources will actually be
establish ‘‘emission standards.’’ pursued by EPA. The commenter also
Emission standards are defined to be a indicate that EPA has the discretion
under section 112(d)(4) to develop notes that the Urban Air Toxics program
requirement established by the State or cannot be relied upon to address
the Administrator which limits the health-based compliance alternatives for
some source categories emitting ambient background. This program,
quantity, rate or concentration of
emissions of air pollutants on a threshold pollutants, and that those 180 See also Legislative History at 876 (section
continuous basis * * * to assure standards may be less stringent than the 112(d)(4) standard may be less stringent than
continuous emission reduction, and any corresponding MACT standard MACT).
required under section 112(k), was to be thresholds below which there would be protective for greater than once in a
completed by 1999. However, the no adverse health effects while AEGL– lifetime exposures. The acute exposure
strategy has not been finalized and the 1 values are health thresholds below levels are called acute Reference
small amount of activity in this area is which there may be mild adverse Exposure Levels and are available at
focused on voluntary emission effects. http://www.oehha.ca.gov/air/acute_rels/
reductions rather than federal Acute exposures are relevant (in acuterel.html.
requirements. Finally, the commenter addition to chronic exposures) and the The 1-hour REL values for hydrogen
notes that control of criteria pollutants acute exposure hazard index of 1.0 chloride and chlorine are 2.1 mg/m3 and
via State Implementation Plans to could be exceeded multiple times over 0.21 mg/m3, respectively. The AEGL–1
achieve compliance with the NAAQS is an individual’s lifetime. Although we values for hydrogen chloride and
problematic. For particulate matter (PM) concluded at proposal that the chronic chlorine are 2.7 mg/m3 and 1.4 mg/m3,
and ozone, new NAAQS were set in exposure Hazard Index would always be respectively. Although there is little
1997 and seven years later the higher than the acute exposure Hazard difference between the 1-hour REL and
nonattainment designations are still Index, and thus would be the basis for AEGL–1 values for hydrogen chloride,
being determined. The designation the total chlorine emission rate limit, the 1-hour REL for chlorine is
process will be followed by a 3 year this conclusion relates to acute versus substantially lower than the AEGL–1
period to prepare State Implementation chronic exposure to a constant, value.
Plans and several more years to carry maximum average emission rate of total In summary, we believe that aRELs
out those plans. In the meantime, there chlorine from a hazardous waste are a more appropriate health threshold
will be high levels of PM and ozone in combustor. See 69 FR at 21300. We metric than AEGL–1 values for
the air near many hazardous waste explained that acute exposure must establishing health-based compliance
combustors in New Jersey which will nonetheless be considered when alternatives for hazardous waste
exacerbate exposures to chlorine and establishing operating requirements to combustors because aRELs are ‘‘no
hydrogen chloride. ensure that short-term emissions do not adverse effect’’ threshold levels that are
Response: Total chlorine missions result in an acute exposure Hazard intended to be protective for multiple
from collocated hazardous waste Index of greater than 1.0. This is exposures.
combustors must be considered in because total chlorine and chloride Comment: One commenter states that
establishing health-based compliance feedrates to a hazardous waste the health-based compliance alternative
alternatives under § 63.1215. Ambient combustor (e.g., commercial incinerator) is unlawful because the proposal does
levels of HCl or chlorine attributable to can vary substantially over time. not address ecological risks that may
other on-site sources, as well as off-site Although a source may remain in result from uncontrolled HAP
sources, are not considered, however. compliance with a feedrate limit with a emissions, including risks posed to
As we indicated in the Residual Risk long-term averaging period (e.g., 12- those areas where few people currently
Report to Congress and in the recent hour, monthly, or annual) based on the live, but sensitive habitats exist.
residual risk rule for Coke Ovens, the chronic Hazard Index, the source could Response: An ecological assessment is
Agency intends to consider facility-wide feed chlorine during short periods of normally required under CAA section
HAP emissions as part of the ample time that substantially exceed the long- 112(d)(4) to assess the presence or
margin of safety determination for CAA term feedrate limit. This could result absence of ‘‘adverse environmental
section 112(f) residual risk actions. 70 potentially in emissions that exceed the effects’’ as that term is defined in CAA
FR at 19996–998 (April 15, 2005); see one-hour (i.e., acute exposure) Hazard section 112(a)(7). To identify potential
also, 54 FR at 38059 (Sept. 14, 1989) Index. Consequently, we discussed at multimedia and/or environmental
(benzene NESHAP). proposal the need to establish both concerns, EPA has identified HAP with
Comment: Several commenters state short-term and long-term total chlorine significant potential to persist in the
that acute exposure guideline levels and chloride feedrate limits to ensure environment and to bioaccumulate. This
(AEGLs) are once-in-a-lifetime exposure that neither the chronic exposure nor list does not include hydrogen chloride
levels. They assert that, because short the acute exposure Hazard Index or chlorine.
term exposures at a Hazard Index exceeds 1.0.181 We also note that health-based total
greater than 1.0 may occur more than We conclude that 1-hour Reference chlorine emission limits for
once in a lifetime, using AEGLs for the Exposure Levels (aRELs) are a more incinerators, cement kilns, and
purpose of setting risk-based short-term appropriate health threshold metric lightweight aggregate kilns cannot be
limits for HCl and chlorine does not than AEGL–1 values for hazardous higher than the current Interim
provide an ‘‘ample margin of safety.’’ waste combustors given that the acute Standards. See § 63.1215(b)(7). Thus,
Response: To assess acute exposure, Hazard Index limit of 1.0 may be the ecological risk from total chlorine
we proposed to use acute exposure exceeded multiple times over an emissions from these sources will not be
guideline levels for 1-hour exposures individual’s lifetime, albeit resulting increased under the health-based limits.
(AEGL–1) as health thresholds. We have from uncontrollable factors. The In addition, we note that only 2 of 12
investigated commenters’ concerns, California Office of Health Hazard solid fuel boilers have total chlorine
however, and conclude that AEGLs are Assessment has developed acute health emissions higher than 180 ppmv, and
not likely to be protective of human threshold levels that are intended to be only 1 liquid fuel boiler has emissions
health because individuals may be higher than 170 ppmv. Thus, boilers
subject to multiple acute exposures at a 181 Note that we conclude for the final rule that
generally have low total chlorine
Hazard Index greater than 1.0 from most sources are not likely to exceed the acute emissions which would minimize
Hazard Index because they will establish a 12-hour
hazardous waste combustors. rolling average chlorine feedrate limit and their ecological risk.
Consequently, we use acute Reference chlorine feedrates are not likely to vary Consequently, we do not believe that
Exposure Levels (aRELs) rather than substantially over that averaging period. Thus, we emissions of hydrogen chloride or
acute exposure guideline levels (AEGLs) believe that most sources will not be required to chlorine from hazardous waste boilers
establish an hourly rolling average chlorine feedrate
as acute exposure thresholds for the limit. The owner/operator must determine whether
will pose a significant risk to the
final rule. See also Part Two, Section the hourly rolling average chloride feedrate limit environment, and facilities attempting
IX.D above. Acute RELs are health can be waived under § 63.1215(d). to comply with the health-based
alternatives for these HAP are not analyses—using the look-up tables. source category to ensure the HCl-
required to perform an ecological However, some facilities will require equivalent emission rate limits reflected
assessment. more detailed modeling. The criteria for stack parameters representative of each
demonstrating eligibility for the source category. Similarly, another
B. Implementation of the Health-Based
compliance alternatives are clearly commenter notes that a look-up table
Standards
defined in the final rule. Moreover, designed to be applicable to all
Comment: Several commenters are under authority of RCRA section hazardous waste combustors is very
concerned that the health-based 3005(c)(3), multi-pathway risk conservative and will have limited
compliance alternative will place an assessments will typically have already utility. This commenter does not suggest
intensive resource demand on state and been completed for many hazardous that EPA develop look-up tables for
local agencies to review and approve waste combustors to document that each class of hazardous waste
facilities’ eligibility demonstrations, and emissions of toxic compounds, combustors, however. Rather, the
State and local agencies may not have including total chlorine, do not pose a commenter suggests that since look-up
adequate expertise to review and hazard to human health and the tables have already been developed for
approve the demonstrations. One environment. Thus, state permitting industrial boilers that do not burn
commenter states that permitting officials have already reviewed and hazardous waste 182 hazardous waste
authorities do not have the expertise to approved detailed modeling studies for combustors should be allowed to use
review eligibility demonstrations that many hazardous waste combustors. The those look-up tables instead of the look-
are based on procedures other than results of these studies could be applied up tables proposed for hazardous waste
those included in EPA’s Reference to the eligibility demonstration required combustors.
Library, as would be allowed. The by this final rule. Response: We noted at proposal that
commenter also states that, if the health- Because these requirements are the emission rates provided in the look-
based compliance alternative is clearly defined, and because any up table for hazardous waste
promulgated, EPA should establish one standards or requirements created under combustors are more stringent than
standard method for the analyses so CAA section 112 are considered those promulgated for solid fuel
there is consistency nationwide. If EPA applicable requirements under 40 CFR industrial boilers that do not burn
offers more than one method, EPA part 70, the compliance alternatives hazardous waste. This is because the
should do all of the risk assessment would be incorporated into title V key parameters used by the SCREEN3
reviews, instead of passing the programs, and states would not have to atmospheric dispersion model (i.e.,
responsibility, without clear direction, overhaul existing permitting programs. stack diameter, stack exit gas velocity,
to the permitting authorities, according Finally, with respect to the burden and stack exit gas temperature) to
to the commenter. associated with ongoing assurance that predict the normalized air
Response: The health-based facilities that opt to do so continue to concentrations that EPA used to
compliance alternatives for total comply with the health-based establish HCl-equivalent emission rates
chlorine that EPA has adopted in the compliance alternatives, the burden to for solid fuel industrial boilers that do
final rule should not impose significant states will be minimal. In accordance not burn hazardous waste are
resource burdens on states. The required with the provisions of title V of the CAA substantially different for hazardous
compliance demonstration methodology and part 70 of 40 CFR (collectively ‘‘title waste combustors. Thus, the maximum
is structured in such a way as to avoid V’’), the owner or operator of any HCl-equivalent emission rates for
the need for states to have significant affected source opting to comply with hazardous waste combustors would
expertise in risk assessment the health-based compliance generally be lower than those EPA
methodology. We have considered the alternatives is required to certify established for solid fuel industrial
commenters’ concerns in developing the compliance with those standards every boilers that do not burn hazardous
criteria defining eligibility for these five years on the anniversary of the waste.
compliance alternatives, and the comprehensive performance test. In Nonetheless, we agree with the
approach that is included in the final addition, if the facility has reason to commenter’s concerns that the look-up
rule provides clear, flexible know of changes over which the facility tables would have more utility if they
requirements and enforceable does not have control, and these better reflected the range of stack
compliance parameters. The final rule changes could decrease the allowable properties representative of hazardous
provides two ways that a facility may HCl-equivalent emission rate limit, the waste combustors. Accordingly, we
demonstrate eligibility for complying facility must submit a revised eligibility examined the stack parameters for all
with the health-based compliance demonstration. Further, before changing hazardous waste-burning sources in our
alternatives. First, look-up tables allow key parameters that may impact an data base (except for hydrochloric acid
facilities to determine, using a limited affected source’s ability to continue to production furnaces that are not eligible
number of site-specific input meet the health-based emission for the health-based emission
parameters, whether emissions from standards, the source is required to standards). After analyzing the
their sources might cause the Hazard evaluate its ability to continue to relationships among the various stack
Index limit to be exceeded. Second, if a comply with the health-based parameters (i.e., stack height, stack
facility cannot demonstrate eligibility compliance alternatives and submit diameter, stack gas exhaust volume, and
using a look-up table, a modeling documentation to the permitting exit temperature), we concluded that the
approach can be followed. The final rule authority supporting continued look-up table should be modified to
presents the criteria for performing this eligibility for the compliance treat both stack diameter and stack
modeling. alternative. Thus, compliance height as independent variables rather
Only a portion of hazardous waste requirements are largely self- than relying on stack height alone.
combustors will submit eligibility implementing and the burden on states We developed separate tables for
demonstrations for the health-based will be minimal. short-term (i.e., 1-hour) HCl-equivalent
compliance alternatives. Of these Comment: One commenter suggests
sources, several should be able to that the look-up tables would have more 182 See Table 2 of Appendix A to Subpart
demonstrate eligibility based on simple utility if EPA developed tables for each DDDDD, Part 63.
emissions limits to protect against acute cannot be used in areas of complex Comment: Several commenters state
health effects and long-term (i.e., terrain (which we define generally as that facilities should be allowed to
annual) emission limits to protect terrain that rises above stack top). establish an averaging period for the
against chronic effects from exposures Sources located in complex terrain (i.e., total chlorine and chloride feedrate
to chlorine and hydrogen chloride. As as a practical matter, sources other than limit that is shorter than an annual
discussed above, we used the acute those that are located in flat or simple rolling average. Commenters are
Reference Exposure Level (aREL) elevated terrain as discussed below and referring to the feedrate limit to ensure
developed by Cal-EPA as the benchmark thus cannot use the look-up tables) must compliance with the annual average
for acute health effects. We used EPA’s use site-specific modeling procedures to HCl-equivalent emission rate limit.
Reference Concentrations (RfC) as the establish HCl-equivalent emission rates. Commenters are concerned with the
benchmark for chronic health effects We looked at two generic terrain data handling issues that could arise
from exposures occurring over a scenarios for purposes of the look-up from calculating, recording, and
lifetime. table. In one we assumed the terrain reporting an annual rolling average
Emission limits in the look-up table rises at a rate of 5 meters for every 100 feedrate level that is updated hourly,
are expressed in terms of HCl-toxicity meter run (i.e., a slope of 5 percent) and and note that a shorter averaging period
equivalent emission rates (lbs/hr). To that terrain is ‘‘chopped off’’ above stack would make the limit more stringent.
convert your total chlorine emission rate top (following the convention for such Response: We agree with commenters,
(lb/hr) to an HCl-equivalent emission analyses in simple elevated terrain). In and conclude, moreover, that a 12-hour
rate, you must adjust your chlorine the other we assumed flat terrain. As averaging period rather than an annual
emission rate by a multiplicative factor can be seen from the tables in § 63.1215, averaging period will be imposed on the
representing the ratio of the HCl health the emission limits with flat terrain are vast majority of sources as a practical
risk benchmark to the chlorine health significantly higher than those with matter. This is because sources must
risk benchmark. For 1-hour average HCl- simple elevated terrain. To reasonably establish a limit on the feedrate of total
equivalent emission rates, the ratio is ensure that the emission limits are not chlorine and chloride to ensure
the ratio of the aREL for HCl (2100 substantially over-stated (e.g., by a compliance with the semivolatile metals
micrograms per cubic meter) to the factor of 2), the simple elevated terrain emission standards. See § 63.1209(n).
aREL for chlorine (210 micrograms per table must be used whenever terrain The feedrate limit for total chlorine and
cubic meter), or a factor of 10.183 For rises to an elevation of one half (1⁄2) the chloride is established under
annual average emissions, the ratio is stack height within a distance of 50 § 63.1209(n) as the average of the hourly
the ratio of the RfC for HCl (20 stack heights. rolling averages for each test run, and
micrograms per cubic meter) to the RfC For both the simple elevated terrain the averaging period is 12 hours. Thus,
of chlorine (0.2 micrograms per cubic and flat terrain scenarios, we performed the averaging period for the feedrate
meter), or a factor of 100. See model runs for urban and rural limit for semivolatile metals—12-hour
§ 63.1215(b). dispersion conditions, with and without rolling average updated hourly—trumps
We used the SCREEN3 air dispersion building downwash. We selected the the annual rolling average averaging
model to develop the emission limits in highest (ambient air concentration) period that would otherwise apply
the look-up tables. SCREEN3 is a values at each distance from among the here.184
screening model that estimates air four runs for each of the terrain Sources may also demonstrate
concentrations under a wide variety of scenarios. compliance with the semivolatile metals
meteorological conditions in order to As can be seen from the tables in standard by assuming all semivolatile
identify the meteorological conditions § 63.1215, the HCl-equivalent emission metals in feedstreams are emitted. See
under which the highest ambient air rate limits range from 0.13 pounds per § 63.1207(m)(2). Sources that do not
concentrations are likely to occur and hour on an annual average (for a 0.3 have emission control equipment, such
what the magnitude of the ambient air meter diameter stack that is 5 meters tall as most liquid fuel boilers, are
concentrations are likely to be. The that lies within 30 meters of the particularly likely to use this approach.
SCREEN3 model implements the property boundary) to 340 pounds per Under this approach, there is no
procedures in EPA’s ‘‘Screening hour (for a 4.0 meter diameter stack that concern regarding increased volatility of
Procedures for Estimating the Air is 100 meters tall that lies 5000 meters metals as chlorine feedrates increase,
Quality Impact of Stationary Sources, from the property boundary) when and such sources are not subject to a
Revised’’ (EPA–454/R–92–019, U.S. located in simple elevated terrain. In flat feedrate limit for chlorine for
Environmental Protection Agency, terrain, the range is from 0.37 to 1100 compliance assurance with the
Office of Air Quality Planning and pounds per hour on an annual average. semivolatile metal standard. These
Standards, Research Triangle Park, NC, This contrasts with the look-up table at sources may establish an averaging
October 1992). Included are options for proposal, where the comparable range period for the feedrate of total chlorine
estimating ambient air concentrations in was from 0.0612 pounds per hour (for and chloride for compliance with the
simple elevated terrain and complex a 5 meter stack height at a distance of health-based compliance alternative for
terrain. Simple elevated terrain refers to 30 meters) to a maximum of 18 pounds total chlorine of not to exceed one
terrain elevations below stack top. We per hour (for stack heights of 50 meters year.185
did not use the complex terrain option or greater, at distances of 500 meters or
in the development of the look-up tables greater). 184 To also ensure compliance with the annual
because of the site-specific nature of If you have more than one hazardous average HCl-equivalent emission rate limit,
waste combustor on site, the sum of the however, the numerical value of the feedrate limit
plume impacts in areas of complex established during the semivolatile metals
terrain. Therefore, the look-up tables ratios for all combustors of the HCl- performance test cannot exceed the value calculated
equivalent emission rate to the HCl- as the annual average HCl-equivalent emission rate
183 We note that this factor of 10 ratio of the equivalent emission rate limit cannot limit divided by [1 ¥ system removal efficiency],
aRELs of HCl to chlorine is based on current aREL exceed 1.0. See § 63.1215 (c)(3)(v). This where you demonstrate the total chlorine system
values and is subject to change. You must use removal efficiency during the comprehensive
current aREL (and RfC) values when you conduct
will ensure that the Hazard Index of 1.0 performance test.
your eligibility demonstration. See § 63.1215(b)(4 is not exceeded considering emissions 185 We note that we have also applied this ‘‘not-
and 5). from all on-site combustors. to-exceed’’ approach to establishing the duration of
Comment: Several commenters Only long-term exposures—maximum limit may be exceeded. That feedrate
offered suggestions on whether a short- annual average exposures—need be limit is established as the 1-hour HCl-
term feedrate limit was needed for total considered when confirming that the equivalent emission rate limit divided
chlorine and chloride (i.e., chlorine) as chlorine feedrate during the by [1 ¥ system removal efficiency].
EPA suggested, and if EPA continues to comprehensive performance test (i.e., Under § 63.1215(d)(3), you must
consider it necessary, how the limit average of the hourly rolling averages establish an hourly rolling average
should be established. for each run) is acceptable because the chlorine feedrate limit unless you
One commenter states that it is not annual exposure Hazard Index limit determine considering specified criteria
necessary to set short-term limits for (i.e., not to exceed 1.0) would always be that your chlorine feedrates will not
chlorine feedrates. If EPA concludes exceeded before the 1-hour Hazard increase over the averaging period for
that short-term limits are necessary, Index limit (i.e., not to exceed 1.0). the long-term chlorine feedrate limit
however, the commenter recommended Thus, the feedrate limit associated with (i.e., 12-hour rolling average or (up to)
these options: (1) Cap the feedrate at a annual exposures would always be more annual rolling average) to a level that
level that is extrapolated up to the stringent than the feedrate limit may result in an exceedance of the 1-
feedrate associated with Interim associated with 1-hour exposures. See hour average HCl-equivalent emission
Standard for incinerators; (2) if the 69 FR at 21299. rate limit. The criteria that you must
facility uses the site-specific option to We further explained at proposal, consider are: (1) The ratio of the 1-hour
set emission limits, the dispersion however, the need to establish a short- average HCl-equivalent emission rate
models can easily be used to set a 1- term feedrate limit for chlorine to based on the total chlorine emission rate
hour (or longer) limit; and (3) if the ensure that the 1-hour HCl-equivalent you select for each combustor to the 1-
facility uses the look up table (which at emission rate did not exceed the 1-hour hour average HCl-equivalent emission
proposal provided only annual average average HCl-equivalent emission rate rate limit for the combustor; and (2) the
HCl-equivalent emission rate limits), a limit due to variability in the chlorine potential for the source to vary chlorine
short-term limit can be set based on a feedrate during the annual averaging feedrates substantially over the
multiplier of the annual limit’10 times period for the feedrate limit. We averaging period for the long-term
the annual limit as recommended by requested comment on approaches to chlorine feedrate limit.
documents in EPA’s Air Toxics Risk establish this 1-hour chlorine feedrate For example, if a source’s primary
Assessment Reference Library. limit, including extrapolating feedrates chlorine-bearing feedstreams have a
Another commenter states that, if EPA to 100% of the 1-hour average HCl- relatively constant chlorine
were to promulgate a short-term feedrate equivalent emission rate limit. See 69 concentration over the averaging period
limit, the EPA-endorsed factor of 0.08 FR at 21304. for the chlorine feedrate limit to ensure
employed to translate maximum hourly In the final rule we have corrected compliance with the annual average
concentrations to annual concentrations and refined these procedures. The final HCl-equivalent emission rate limit (e.g.,
could be used to identify the maximum rule requires you to establish a long- generally 12-hours), as may be the case
hourly feedrate limit. term chlorine feedrate limit to maintain for commercial sources feeding from
Finally, another commenter states that compliance with the annual average large burn tanks or on-site sources
extrapolation of the chlorine feedrate HCl-equivalent emission rate limit as where chlorine levels in wastes are
(from the level during the either: (1) The chlorine feedrate during fairly constant, you may conclude that
comprehensive performance test when the comprehensive performance test if there is little probability that 1-hour
the source documents compliance with you demonstrate compliance with the feedrates would vary substantially over
the annual average HCl-equivalent semivolatile metals emission standard the averaging period. Thus, a 1-hour
emission rate limit) should be allowed during the test (see § 63.1209(o)); or (2) rolling average chlorine feedrate limit
to 100% of the 1-hour average HCl- if you comply with the semivolatile may not be warranted. Even if chlorine
equivalent emission rate limit because metals emission standard under feedrates could vary substantially over
numerous safety factors have already § 63.1207(m)(2) by assuming all metals the long-term feedrate averaging period,
been included in the health risk in the feed to the combustor are emitted, however, an hourly rolling average
threshold values, look-up tables, and the annual average HCl-equivalent feedrate limit still may not be warranted
modeling demonstration. emission rate limit divided by [1 ¥ if the source’s 1-hour average HCl-
Response: At proposal, we explained system removal efficiency] where you equivalent emission rate is well below
that sources would establish an annual demonstrate the system removal the 1-hour HCl-equivalent emission rate
average feedrate limit on chlorine as the efficiency during the comprehensive limit. See Part Two, Section IX.H, of this
feedrate level during the comprehensive performance test. See discussion in Part preamble for a discussion of the
performance test demonstrating Two, Section IX.H, of this preamble. If relationship between emission rates,
compliance with the annual average you establish the chlorine feedrate limit emission rate limits, and feedrate limits.
HCl-equivalent emission rate limit. 186 based on the feedrate during the We disagree with the commenter who
performance test to demonstrate states that short-term chlorine feedrate
averaging periods for the limits on all operating compliance with the semivolatile metals limits are not necessary. The 1-hour
parameters established under § 63.1209. See new average HCl-equivalent emission rate
§ 63.1209(r) and USEPA, ‘‘Final Technical Support
emission standard, the averaging period
Document for HWC MACT Standards, Volume IV: for the feedrate limit is a 12-hour rolling limit could potentially be exceeded for
Compliance with HWC MACT Standards, average. If you establish the chlorine sources with highly variable chlorine
September 2005, Section 2.4.6. feedrate limit based on the system feedrates and where the 1-hour HCl-
186 We discussed at proposal that the feedrate
removal efficiency during the equivalent emission rate is relatively
limit to ensure compliance with the long-term
Hazard Index limit of not to exceed 1.0 would be performance test, the averaging period is high compared to the 1-hour HCl-
the average of the hourly rolling averages for each up to an annual rolling average. equivalent emission rate limit. The 1-
test run, with compliance based on an annual The final rule also requires you to hour average HCl-equivalent emission
average. Note that, under the final rule however, the establish an hourly rolling average rate limit could be exceeded even
long-term chlorine feedrate limit is established as
the annual average HCl-equivalent emission rate
chlorine feedrate limit if you determine though the source remains in
limit divided by [1 ¥ system removal efficiency]. under § 63.1215(d)(3) that the 1-hour compliance with the annual average
See § 63.1215(g)(2). average HCl-equivalent emission rate HCl-equivalent emission rate limit (and,
moreover, the 12-hour rolling average or emission rate and the annual average Response: We agree with commenters
(up to) annual rolling average chlorine HCl-equivalent emission rate limit at the that requiring prior approval of the
feedrate limit). level equivalent to the Interim Standard eligibility demonstration may be
We agree with commenters that for total chlorine. Thus, the long-term unworkable for the reasons commenters
suggest that the hourly rolling average chlorine feedrate limit (12-hour rolling suggest. We also agree with commenters
chlorine feedrate limit should be average or (up to) an annual rolling that sources who make a good-faith
extrapolated from performance test average) is capped at the level eligibility demonstration but whose
feedrates up to 100% of the 1-hour corresponding to the Interim Standards demonstration is denied by the
average HCl-equivalent emission rate for incinerators, cement kilns, and permitting authority may need
limit. The final rule requires you to lightweight aggregate kilns. The hourly additional time to install controls or
establish the hourly rolling average rolling average feedrate limit to take other measures to comply with the
feedrate limit (if a limit is required maintain compliance with the 1-hour MACT emission standards.
under § 63.1215(d)(3)) as the 1-hour average HCl-equivalent emission rate Accordingly, the final rule does not
HCl-equivalent emission rate limit limit, however, can exceed the require prior approval of the eligibility
divided by [1 ¥ system removal numerical value of the long-term demonstration for existing sources. If
efficiency]. Establishing the hourly chlorine feedrate limit because the 1- your permitting authority has not
rolling average feedrate in this manner hour average HCl-equivalent emission approved your eligibility demonstration
ensures that the 1-hour HCl-equivalent rate limit is substantially higher than by the compliance date, and has not
emission rate limit is not exceeded, and the annual average HCl-equivalent issued a notice of intent to disapprove
thus that the aREL-based Hazard Index emission rate limit. Thus, capping at the your demonstration, you may
of 1.0 is not exceeded. interim standard level is inappropriate nonetheless begin complying, on the
We also agree in principle with unless the interim standard were compliance date, with the HCl-
commenters that suggest that the hourly somehow re-expressed as a 1-hour limit. equivalent emission rate limits and
rolling average feedrate limit be based Comment: Many commenters state associated chlorine feedrate limits you
on the 1-hour average HCl-equivalent that requiring prior approval of the present in your eligibility
emission rate limit which is based on eligibility demonstration would be demonstration.
emissions modeling. These commenters unworkable. Commenters are concerned In addition, the final rule states that
suggested that we use a multiplier of 10 that the permitting authority may not the permitting authority should notify
or 12.5 (i.e., 1/0.08) to project 1-hour approve the demonstration prior to the you of approval or intent to disapprove
average HCl-equivalent emission rate compliance date even though the source your eligibility demonstration within 6
limits from the annual average HCl- has submitted complete and accurate months after receipt of the original
equivalent emission rate limits. Rather information and has responded to any demonstration, and within 3 months
than use these approaches to project 1- requests for additional information in after receipt of any supplemental
hour average emissions from annual good faith. Commenters are also information that you submit. A notice of
average emissions, however, we use concerned that the permitting authority intent to disapprove your eligibility
emissions modeling to develop look-up may disapprove the demonstration too demonstration, whether before or after
tables for both 1-hour average HCl- late for the source to take other the compliance date, will identify
equivalent emission rate limits and measures to comply with the total incomplete or inaccurate information or
annual average HCl-equivalent emission chlorine MACT standard. Once noncompliance with prescribed
rate limits. For sources that use site- commenter recommends the following procedures and specify how much time
specific risk assessment to demonstrate alternative approach: (1) If the you will have to submit additional
eligibility, they will use the same regulatory agency does not act on a risk information or comply with the total
models to estimate 1-hour average demonstration within the 6-month chlorine MACT standards. The
maximum ambient concentrations. period, it is conditionally deemed permitting authority may extend the
Thus, the final rule uses modeling to approved; and (2) if a risk compliance date of the total chlorine
establish directly 1-hour average HCl- demonstration is disapproved, the MACT standards to allow you to make
equivalent emission rate limits rather source would have to comply with the changes to the design or operation of the
than approximating those limits from MACT emission standards no later than combustor or related systems as quickly
annual average HCl-equivalent emission three years after notice of disapproval as practicable to enable you to achieve
rate limits as commenters suggest. In and, in the interim, sources would compliance with the total chlorine
summary, the final rule requires you to comply with current emission limits for MACT standards.
establish the 1-hour average HCl- total chlorine. Comment: One commenter states that
equivalent emission rate limit by either Another commenter suggests that, if proposed § 63.1215(f)(1)(A) should have
using Tables 3 or 4 in § 63.1215 to look- the permitting authority has neither required sources to conduct a new
up the limit, or conducting a site- approved nor disapproved the eligibility comprehensive performance test only if
specific risk analysis. Under the site- demonstration by the compliance date, there are changes that would decrease
specific risk analysis option, the 1-hour the source may begin complying on the the HCl-equivalent emission rate limit
average HCl-equivalent emission rate compliance date with the alternative below the HCl-equivalent emission rate
limit would be the highest emission rate health-based limits specified in the demonstrated during the comprehensive
that the risk assessment estimates would eligibility demonstration. performance test. Similarly, the
result in an aREL-based Hazard Index Finally, another commenter states that commenter suggests that a retest should
not exceeding 1.0 at any off-site receptor facilities should be granted a three-year not be required if a change increases the
location. extension of the compliance date if the HCl-equivalent emission rate limit but
We do not agree that the short-term Agency denies a good-faith eligibility the source elects to maintain the current
feedrate limit should be capped at the demonstration. The commenter is feedrate limit.
level corresponding to the Interim concerned that sources will not have Another commenter states that the
Standards for incinerators, cement kilns, time to install additional controls or Agency should clarify that if there are
and lightweight aggregate kilns. The take other measures after a denial is any changes that are not controlled by
final rule caps the total chlorine issued but prior to the compliance date. the facility owner/operator, and the
facility is required to change its design variability in the hydrogen chloride or current scientific knowledge, we have
or operation to lower chlorine emissions total chlorine measurements due to committed to consider alternative
to address the changes, the facility may method imprecision or other factors credible and readily available
request up to three years to make such could result in inaccurate estimations of assessments (e.g., the acute Relative
changes. chlorine emission levels. Exposure Levels established by the
Response: We generally agree with the We do not agree, however, that using California Office of Health Hazard
commenters and have revised the rule Method 26/26A for chlorine Assessment). For our use, these
as follows: (1) A new comprehensive measurements for combustors feeding alternatives need to be grounded in
performance test is required to high levels of bromine or sulfur is publicly available, peer-reviewed
reestablish the system removal acceptable–the chlorine measurement information. We agree with the
efficiency for total chlorine only if you may be biased low. Chlorine emission commenter that the issue of changing
change the design, operation, or levels must be determined as accurately toxicity values is a general challenge in
maintenance of the source in a manner as possible given that the long-term setting health-based regulations.
that may decrease the system removal health threshold for chlorine is 100 However, we are committed to
efficiency (e.g., the emission control times the threshold for HCl, and the establishing such regulations that reflect
system is modified in a manner than short-term health threshold for chlorine current scientific understanding, to the
may decrease total chlorine removal is 10 times the threshold for HCl (i.e., extent feasible.
efficiency); and (2) if you use the site- using current RfCs and aRELs). To
specific risk analysis option for your ensure that a conservative estimate of C. National Health-Based Standards for
eligibility demonstration and changes the chlorine emission rate is used to Cement Kilns
beyond your control (e.g., off-site establish the alternative health-based Comment: One commenter states that
receptors newly residing or congregating emission limits and to address our suggestion at proposal that it would
at locations exposed to higher ambient commenters’ concerns, the final rule be appropriate to establish a single
levels than originally estimated) dictate requires that you determine chlorine national emission rate type standard
a lower HCl-equivalent emission rate emissions to be the higher of: (1) The applicable to all cement kilns based on
limit and you must make changes to the chlorine value measured by Method 26/ the worst-case scenario cement kiln is
design, operation, or maintenance of the 26A, or an equivalent method; or (2) the unduly burdensome as it discounts the
combustor or related systems to comply chlorine value calculated by difference benefits of improved dispersion realized
with the lower limit, you may request between the combined hydrogen by facilities that have invested in taller
that the permitting authority grant you chloride and chlorine levels measured stacks that minimize downwash effects.
additional time to make those changes by Method 26/26A, or an equivalent The commenter recommends a dual
as quickly as practicable. method, and the hydrogen chloride limit for cement kilns such that the HCl
Comment: Several commenters state measurement from EPA Method 320/ equivalent emission rate is limited to
that the proposed approach for 321 or ASTM D 6735–01, or an both: (1) A 130 ppmv total chlorine
calculating chlorine emissions to equivalent method. emission standard (the Interim
address the potential bias using Method Comment: Several commenters state Standard) coupled with a chlorine
26/26A attributable to high bromine or the procedures for calculating HCl- feedrate limit based on a 12-hour rolling
sulfur levels in feedstreams is not equivalent emission rates cannot merely average; and (2) a Hazard Index of 1.0.
statistically valid. They indicate that the reference an outside source, such as a Response: We have decided not to
approach could lead to collection of Web site, unless that reference specifies include a separate national standard for
total chlorine, hydrogen chloride and that the contents of the source are as of cement kilns in the final rule for several
chlorine data that are contradictory and a date certain. To specify use of health reasons: (1) We have no assurance that
difficult to apply in a compliance threshold values that can change over the Cl2/HCl volumetric ratio exhibited
situation. One commenter suggests that time provides inadequate opportunity during the most recent compliance test,
using Method 26/26A results for sources for notice and comment on the and that was the basis for the
with bromine and sulfur dioxide, while regulation. commenter documenting in a study 187
recognizing that there is bias in the Response: We believe that the best that the Hazard Index of 1.0188 was not
sampling method, will result in a valid available sources of health effects exceeded, is representative of ratios in
compliance approach. information should be used for risk or the past or future; (2) the commenter’s
Response: We agree with commenters hazard determinations. To assist us in recommended emission standard for
that the proposed approach to avoid the identifying the most scientifically cement kilns—130 ppmv total chlorine
bias when feedstreams contain high appropriate toxicity values for our emission limit and a Hazard Index of
levels of bromine or sulfur (bromine/ analyses and decisions, the Web site to 1.0—is equivalent to the requirements
chlorine ratio in feedstreams of greater be used for RfCs identifies pertinent under § 63.1215 applicable to other
than 5 percent, or sulfur/chlorine ratio toxicity values using a default hierarchy hazardous waste combustors to establish
in feedstreams of greater than 50 of sources, with EPA’s Integrated Risk site-specific emission limits; (3) the
percent) during the comprehensive Information System (IRIS) being the MACT standard for total chlorine for
performance test may be problematic. preferred source. The IRIS process cement kilns is 120 ppmv such that the
The proposed approach would have contains internal and external peer health-based standard that the
required you to use Method 320/321 or review steps and IRIS toxicity values commenter recommends—130 ppmv,
ASTM D 6735–01 for hydrogen chloride represent EPA consensus values. When
measurements, to use Method 26/26A adequate toxicity information is not 187 See Trinity Consultants, ‘‘Analysis of HCl/Cl2
for total chlorine (i.e., hydrogen available in IRIS, however, we consult Emissions from Cement Kilns for 112(d)(4)
Consideration in the HWC MACT Replacement
chloride and chlorine combined) other sources in a default hierarchy that Standards,’’ September 17, 2003.
measurements, and to calculate chlorine recognizes the desirability of these 188 The HCl/Cl ratio for the total chlorine
2
levels by difference. The potential qualities in ensuring that we have measurement is important because the current RfC
problem is that chlorine emission levels consistent and scientifically sound for chlorine is 0.2 µg/m3 while the current RfC for
HCl is 20 µg/m3. Thus, when calculating HCl-
are generally a very small portion of assessments. Furthermore, where the equivalent emission rate limits, chlorine emissions
total chlorine measurements, and IRIS assessment substantially lags the are currently multiplied by a factor of 100.
the Interim Standard—would provide VIII. Implementation and Compliance detection system need not have a
little compliance relief; and (4) even detection limit as low as 1.0 mg/acm to
A. Compliance Assurance Issues for
though the final rule does not provide detect increases in normal emissions.
both Fabric Filters and Electrostatic
a separate national health-based One commenter believes that bag leak
Precipitators (and Ionizing Wet
standard for cement kilns, cement kilns detection systems installed on cement
Scrubbers)
may apply for the health-based kilns should be allowed to have a
compliance alternatives applicable to 1. Implementation Issues detection limit of 10 mg/acm because:
other hazardous waste combustors. Comment: Several commenters state (1) A detection limit requirement of 10
that design and performance mg/acm is more than sufficient to
Prior to publication of the proposed protect the particulate matter emission
rule, the commenter submitted results of specifications and explicit detailed test
procedures to determine conformance limit and to detect increases in
site-specific risk assessments for all particulate matter concentration given
with the specifications are needed so
cement kiln facilities showing that both that the current particulate matter
that manufacturers can certify that their
the long-term and short term Hazard bag leak detection systems and emission limit for existing kilns is 63
Index of 1.0 would not be exceeded at particulate matter detection systems mg/dscm; (2) a detection limit
any facility assuming: (1) Sources emit meet applicable criteria. Absent design requirement of 10 mg/acm is consistent
total chlorine at the Interim Standard and performance specifications and test with the requirement for bag leak
level of 130 ppmv; and (2) total chlorine procedures, commenters assert that the detection systems in Subpart LLL, Part
emissions are apportioned between HCl ‘‘manufacturer’s certification’’ cannot 63, for cement plants that choose to
and chlorine according to the ensure the performance capabilities of install bag leak detection systems on
apportionment exhibited during the the devices. finish mills and raw mills, for bag leak
most recent compliance test. Response: In general, we believe detection systems and particulate matter
At proposal, we requested comment adherence to manufacturer’s written detection systems installed on lime
on how to ensure that the 130 ppmv specifications and recommendations is kilns under Subpart AAAAAA, and for
an appropriate approach to reasonably industrial boilers under Subpart
concentration-based standard would
ensure performance of a bag leak DDDDD; (3) a 10 mg/acm detection limit
ensure that total chlorine emission rates
detection system or particulate matter is achievable using state-of-the-art
(lb/hr) would not increase to levels that detection system, and we have retained transmissometers (the actual instrument
may exceed the Hazard Index limit of that provision in the final rule. We used in a continuous opacity monitoring
1.0 given that: (1) The partitioning ratio agree, however, that there may be cases system (COMS) at cement plants having
between HCl and chlorine could change where other procedures are more kiln stack diameters of 2–3 meters, or
over time such that a larger fraction of appropriate than the manufacturer’s greater; and (4) it is unclear if any bag
total chlorine could be emitted as recommendations to ensure leak detection system device can
chlorine, which has a much lower performance of a bag leak detection actually be demonstrated to achieve a
health risk threshold; and (2) the mass system or particulate matter detection 1.0 mg/acm detection limit except by
emission rate of total chlorine could system. Consequently, the rule allows extrapolation from tests conducted at
increase. See 69 FR at 21306. you to request approval for alternative higher dust loadings and theoretical
The commenter has addressed the monitoring procedures under arguments based on signal-to-noise
concern about the mass emission rate of § 63.1209(g)(1).189 We note that you may ratios or other parameters. This
total chlorine potentially increasing by use references other than EPA’s commenter also recommends that EPA
suggesting that the health-based Guidance Document, ‘‘Fabric Filter Bag establish a 10 mg/am3 detection limit
Leak Detection Guidance,’’ September for all cement kilns rather than provide
standard include a limit on the feedrate
1997 to identify appropriate for site-specific determinations because
of total chlorine and chloride at the
performance specifications for the bag allowing site-specific determinations is
level used in their risk assessment leak detection system or particulate likely to create confusion in the
supporting a separate national standard matter detection system, including: PS– selection of monitoring devices and
for cement kilns. The commenter has 11 for PM CEMS; PS–1 for opacity further complicate the manufacturer’s
also addressed the concern about the monitors; and CPS–001 for opacity certification of performance
HCl and chlorine apportionment ratio monitoring below 10% opacity. You requirements.
changing over time by suggesting that may use these references to support Response: The current requirement
the standard also include a requirement your request for additions to, or for the bag leak detection system
that the Hazard Index of 1.0 not be deviations from, manufacturer’s sensitivity/detection limit applicable to
exceeded. We agree that sources need to specifications. incinerators and lightweight aggregate
account for variability in the chlorine to Comment: One commenter states that kilns is 1.0 mg/acm unless you
HCl ratio (see § 63.1215(b)(6)) and that bag leak detection systems and demonstrate under § 63.1209(g)(1) that a
periodic checks to ensure that the particulate matter detection systems lower sensitivity (i.e., higher detection
Hazard Index of 1.0 is not exceeded are should have a detection limit of 1.0 mg/ limit) would detect bag leaks. We
needed. We believe the best way to acm to ensure peak performance is proposed to apply the bag leak detection
ensure that the health-based compliance maintained rather than explicitly system requirements to all hazardous
alternatives for total chlorine for cement allowing sources to request approval for waste combustors equipped with fabric
kilns are protective with an ample a detection limit on a site-specific basis filters and promulgate that requirement
margin of safety is through the as the rule currently allows. Several today. Although we also requested
procedures of § 63.1215 where site- other commenters state that the bag leak comment whether detection limits
specific emission rate limits are detection system or particulate matter higher than 1.0 mg/acm should be
established rather than under a separate 189 See discussion in Part Five, Section III.C, for
allowed, none of the comments has
national standard for cement kilns. an explanation of how the alternative monitoring
convinced us to alter our view that the
provisions of § 63.1209(g)(1) relate to those of rule should allow higher detection
§ 63.8(f). limits on a site-specific basis. Similarly,
we believe that the same detection limit detection limit requirement as discussed operating parameter limits (which are
requirement should apply to particulate above (i.e., 1.0 mg/acm or a higher linked to the automatic waste feed
matter detection systems that you may detection limit that you document cutoff system). APCD operating
elect to use for compliance monitoring under an alternative monitoring petition parameters often have an uncertain
for your electrostatic precipitator or under § 63.1209(g)(1) would routinely relationship to PM emissions while the
ionizing wet scrubber in lieu of site- detect particulate matter loadings BLDS and PMDS provide real-time
specific operating parameter limits. during normal operations) as the information on actual PM mass
Both bag leak detection systems and detector for your bag leak detection emission levels.192
particulate matter detection systems system or particulate matter detection Comment: One commenter states that
must be able to detect particulate system. requiring a notification if the bag leak
emission in the range of normal detection system or particulate matter
concentrations. For example, to 2. Compliance Issues detection system set point is exceeded
establish the alarm level for the bag leak Comment: One commenter states that, more than 5% of the time in a 6-month
detection system, you must first adjust if the bag leak detection system or period is not cost-effective. Sources
detector gain/sensitivity and response particulate matter detection system using bag leak detection systems have
time based on normal operations. exceeds the alarm level or an operating not linked exceedances to the data
Although the alarm level for particulate parameter limit (OPL) is exceeded, the logging system and would incur an
matter detection systems will be automatic waste feed cutoff (AWFCO) expense to do so.
established based on operations during system must be initiated. Allowing a Response: We continue to believe that
the comprehensive performance test or source to exceed the alarm level for 5% limiting the aggregate duration of
higher (see discussion below), the of the time in a 6-month period does not exceedances in a 6-month period is a
detector must be responsive within the ensure continuous compliance. reasonable approach to gage the
range of normal operations for you to Response: Although the AWFCO effectiveness of the operation and
effectively minimize exceedances of the system must be initiated if an OPL is maintenance procedures for the
alarm level. exceeded, we believe that allowing combustor. We note that recent MACT
The range of normal emission exceedances of the bag leak detection standards for several other source
concentrations will generally be well system or particulate matter detection categories use this approach, including
below both the particulate matter system alarm level up to 5% of the time standards for industrial boilers and
standard and emissions during the in a 6-month period is reasonable. process heaters and standards for lime
comprehensive performance test. Requiring initiation of the AWFCO for kilns.
Consequently, we disagree with an exceedance of an OPL is reasonable Comment: One commenter states that
commenters that believe the detection because sources generally can control EPA did not present a rationale for
limit need only be within the range of directly the parameter that is limited. requiring a notification within 5
emissions at the particulate matter Examples are the feedrate of metals or working days if the bag leak detection
emission standard. On the other hand, chlorine, or pressure drop across a wet system or particulate matter detection
normal emissions may be well above 1.0 system set point is exceeded more than
scrubber. Bag leak detection systems
mg/acm such that a higher detection 5% of the time during a 6-month period.
and particulate matter detection
limit (e.g., 10 mg/acm) may be The commenter notes that this notice is
systems, however, measure mass
appropriate on a site-specific basis. not required under the Subpart DDDDD
emissions of particulate matter, a
We also disagree with the comment boiler and process heater MACT. The
parameter that is affected by many
that bag leak detection systems (or commenter also notes that the source is
interrelated factors and that is not
particulate matter detection systems) required to take corrective measures
directly controllable. We believe that
may not be able actually to achieve a 1.0 under both the operation and
the 5 percent alarm rate is a reasonable
mg/acm detection limit. EPA is aware of maintenance plan and bag leak
allowance for sources due to difficult-to-
bag leak detection system instruments detection systems and particulate matter
certified to meet levels of 0.2 mg/m3 and control variations in particulate matter detection systems requirements. The
particulate matter detection systems can emissions. More important, although the commenter believes that requiring a
readily achieve detection limits well bag leak detection system and report to the permitting authority is
below 1.0 mg/acm.190 particulate matter detection system duplicative, unnecessary, and increases
Comment: One commenter states that measure mass emissions of particulate the burden on regulated facilities
a continuous opacity monitoring system matter, the detector response is not without providing additional protection
(COMS) that can achieve a detection correlated to particulate matter emission to human health or the environment.
level of 10 mg/acm or less can be used concentrations to the extent necessary Response: If a source exceeds the
to monitor electrostatic precipitator for compliance monitoring.191 Thus, alarm set point more than 5% of the
performance. The commenter believes triggering the alarm level is not time in a 6-month period, it is an
that allowing a COMS for compliance evidence that the particulate matter indication that the operation and
under Subpart EEE is also appropriate emission standard has been exceeded. maintenance plan may need to be
because cement kilns will be operating The purpose of a BLDS or PMDS is to revised. Requiring the source to report
under the requirements of Subpart LLL alert the operator that the PM control the excess exceedances to the permitting
(for cement kilns that do not burn device is not functioning properly and
hazardous waste) at times, which that corrective measures must be 192 Moreover, for FFs, we are not aware of any
requires compliance with an opacity undertaken. We believe that using a APCD operating parameters that correlate well with
standard using a COMS. BLDS or PMDS for compliance PM emissions. Thus, sources must use a BLDS or
assurance better minimizes emissions of PMDS for compliance assurance. For ESPs and
Response: You may use a COMS (i.e., IWSs, we are not aware of generic APCD parameters
transmissometer) that meets the PM (and metal HAP) than use of that correlate well with PM emissions. See
discussion below in Section VIII.C of the text.
190 USEPA, ‘‘Technical Support Document for 191 Actually, the BLDS is not correlated at all to Consequently, although the rule allows sources
HWC MACT Standards, Volume IV: Compliance PM concentrations, and the alarm level for a PMDS with ESPs and IWSs to establish site-specific
with the HWC MACT Standards,’’ September 2005, may or may not be approximately correlated to PM operating parameter limits, sources are encouraged
Appendix C, Section 4.0. concentrations. See § 63.1206(c)(9). to use a PMDS.
authority serves as a notification that commenter did not indicate the cannot determine whether it is
the authority may need to review the deficiencies or suggest additional operating within the alarm set point.
operation and maintenance plan with requirements. If you need additional Accordingly, it is reasonable to consider
the source to determine if changes are guidance on compliance with this periods when the bag leak detection
warranted. provision, you should contact the system or particulate matter detection
We agree with the commenter, permitting authority. system is malfunctioning as
however, that it is not necessary to Comment: One commenter supports exceedances of the set point.
require that the report be submitted the approach of listing the shutting
within five working days of the end of down of the combustor as a potential— B. Compliance Assurance Issues for
the 6-month block period. but not mandatory—corrective measure Fabric Filters
Consequently, the final rule requires in response to exceeding an alarm set Comment: One commenter states that
you to submit the report within 30 days point. Several commenters suggest, establishing the set point for the bag
of the end of the 6-month block period. however, that EPA should specify that leak detection system at twice the
Allowing 30 days to submit the report corrective measures could include detector response achieved during bag
rather than 5 days as proposed is shutting off the hazardous waste feed cleaning as recommended by EPA
reasonable. We are concerned that 5 rather than shutting down the guidance would not be sensitive enough
days may not be enough time to combustor. Other commenters state that to detect gradual degradation of the
complete the report given that several it is inappropriate to imply that shutting fabric filter, nor would it be low enough
exceedances toward the end of the 6- down the combustor must be part of a to require the operator of the source to
month block period may cause you to corrective measures program for take corrective measures that would
exceed the 5% time limit and that there responding to exceedance of a set point. ensure effective operation of the
may be many individual exceedances These commenters believe that the baghouse over time.
that need to be included in the report. requirement to take corrective action Response: The commenter expresses
We acknowledge that it may take some upon the alarm is sufficiently the same concern that EPA raised at
time to prepare the report given that you protective. The facility should proposal. See 69 FR at 21347. We have
must describe the causes of each determine if shutting down the concluded, however, that it may be
exceedance and the revisions to the combustor is a necessary response to problematic to establish an alarm set
operation and maintenance plan you avoid noncompliance with a standard. point for fabric filters based on
have made to mitigate the exceedances. Response: You must operate and operations during the comprehensive
Comment: One commenter notes that maintain the fabric filter, electrostatic performance test. This is because, as
there is no guidance on how to calculate precipitator, or ionizing wet scrubber to noted in earlier responses and at 69 FR
when the set-point has been exceeded ensure continuous compliance with the at 21233, it is much more difficult to
more than 5 percent of the operating particulate matter, semivolatile metals, ‘‘detune’’ a fabric filter than an
time within a 6 month period. The and low volatile metals emission electrostatic precipitator to maximize
commenter notes that the MACT for standards. Your response to exceeding emissions during the performance
industrial boilers and process heaters the alarm set point should depend on test.193 Consequently, emissions from
provides minimal instruction on how whether you may be close to exceeding fabric filters that have not been detuned
this is to be done, but it is not specific an operating parameter limit (e.g., ash during the performance test may not be
enough to enable facilities to ensure that feedrate limit for an incinerator or representative of the range of normal
they are in compliance with this liquid fuel boiler equipped with an emissions caused by factors such as bag
requirement. electrostatic precipitator) or an emission aging. Baghouse performance degrades
Response: For the final rule, we have standard. If so, corrective measures over time as bags age. In addition,
adopted the procedures specified in the should include, as commenters suggest, establishing the alarm set point based
industrial boiler and process heater cutting off the hazardous waste feed. on operations during the performance
MACT for calculating the duration of Corrective measures could also include, test for baghouses that have not been
exceedances of the set point. Those however, shutting down the combustor detuned would establish more stringent
procedures are as follows: as the ultimate immediate corrective compliance requirements on sources
1. You must keep records of the date, measure if an emission standard may that perform the best—the lower the
time, and duration of each alarm, the otherwise be exceeded. Consequently, emissions, the lower the alarm set point.
time corrective action was initiated and the final rule continues to require you This would unfairly penalize the best
completed, and a brief description of the to alleviate the cause of the alarm by performing sources.
cause of the alarm and the corrective taking the necessary corrective For these reasons, the final rule
action taken. measure(s) which may include shutting requires you to establish the alarm set-
2. You must record the percent of the down the combustor. This provision point for bag house detection systems
operating time during each 6-month does not imply that shutting down the using principles provided in USEPA,
period that the alarm sounds. combustor is the default corrective ‘‘Fabric Filter Bag Leak Detection
3. In calculating the operating time measure. Rather, it implies that the Guidance,’’ (EPA–454/R–98–015,
percentage, if inspection of the fabric ultimate immediate response, absent September 1997).
filter, electrostatic precipitator, or other effective corrective measures, Comment: One commenter states that
ionizing wet scrubber demonstrates that would be to shut down the combustor. the bag leak detection system
no corrective action is required, no Comment: One commenter states that requirement should not apply to the
alarm time is counted. periods of time when the combustor is coal mill baghouse for cement kilns
4. If corrective action is required, each operating but the bag leak detection with indirect-fired coal mill systems
alarm shall be counted as a minimum of system or particulate matter detection where a fraction of kiln gas is taken
1 hour. system is malfunctioning should not be
193 One approach to detune a fabric filter to
Although the commenter indicates considered exceedances of the set-point.
simulate the extreme high range of normal
that these procedures are not specific Response: If the bag leak detection operations would be to install a butterfly valve that
enough to ensure that sources are in system or particulate matter detection allows a portion of the combustion gas to by-pass
compliance with the requirements, the system is malfunctioning, the source a section of the baghouse.
from the preheater and routed to the must be linked to the automatic waste determine the best approach is far wiser
coal mill and subsequently to a feed cutoff system. than regulating from a distance.
baghouse before entering the stack. The In response to commenters’ concern Response: We agree with the
commenter notes that the PM in this gas that high moisture stack gas may be commenters that state that it is not
is nearly exclusively coal dust, and the problematic for particulate matter practicable to establish operating
baghouse is substantially smaller than detection systems, we note that parameter limits that would effectively
the baghouse for the kiln. extractive light-scattering detectors and ensure performance of all electrostatic
Response: We believe that a bag leak beta gauge detectors can effectively devices. Accordingly, the final rule
detection system is a reasonable operate in high moisture environments. continues to allow sources to establish
approach to monitor emissions for the We acknowledge, however, that the cost site-specific operating parameter limits
coal mill baghouse to ensure of these extractive detector systems is for these devices.
substantially higher than We disagree with the commenter’s
compliance with the particulate matter
transmissometers or in situ light- assertion that site-specific operating
(and semivolatile and low volatile
scattering detectors. parameter limits obviate the
metals) emission standards. These Comment: One commenter states that requirements for federal standards. The
systems are inexpensive to install and EPA must set minimum total power site-specific operating parameter limits
operate. Annualized costs are requirements for both ionizing wet merely reflect the truism that no two
approximately $24,000.194 Although the scrubbers and electrostatic precipitators sources are identical and so what each
commenter did not suggest an because allowing permit officials to needs to do to comply with the uniform
alternative monitoring approach, and establish compliance operating standards may differ. The final rule
we are not aware of a less expensive and parameters on a site-specific basis provides consistent, federally-
effective approach, we note that sources frustrates the intention of the CAA by enforceable emission standards.
may petition the permitting authority obviating the requirements for federal Necessary flexibility in compliance
under § 63.1209(g)(1) to request an standards. The commenter asserts that a assurance for those emission standards
alternative monitoring approach. minimum total power requirement is does not undermine the uniformity of
C. Compliance Issues for Electrostatic monitorable, recordable, and reportable, those standards. In addition, we
Precipitators and Ionizing Wet three requirements that are necessary for disagree with the commenter’s concern
Scrubbers these facilities to come into, and remain that without a minimum power limit,
in compliance with, their Title V there will be no monitorable,
Comment: Several commenters operating permits. recordable, and reportable Title V
believe that a particulate matter Other commenters state that permit limits for electrostatic devices.
detection system may not be necessary electrostatic devices are not easily To the contrary, site-specific operating
for monitoring of electrostatic characterized by operating parameters parameter limits can and will be
precipitators and ionizing wet in a ‘‘one-size-fits-all’’ fashion. The monitored, recorded, reported, and
scrubbers. Commenters state that site- significant operating parameters for linked to the automatic waste feed cut-
specific operating parameter limits electrostatic devices are secondary off system. And, if a source elects to use
linked to the automatic waste feed voltage, secondary current, and a particulate matter detection system in
cutoff system can effectively monitor secondary power (the product of the lieu of establishing site-specific
and ensure the performance of first two items). The relationship operating parameter limits, the detector
electrostatic precipitators and ionizing between these parameters and response will be monitored, recorded,
wet scrubbers. Particulate matter performance of the unit differ between reported, and linked to requirements to
detection systems on cement kilns applications and unit types. For take corrective measures if the alarm set
would have to operate in a high example, inlet field power can increase point is exceeded.
moisture stack environment (all kilns as unit performance appears to decrease. Comment: One commenter asserts
burning hazardous waste that are In this case, an operating parameter that the use of electrostatic precipitator
equipped with electrostatic precipitators other than secondary power by field total power input data (sum of the
are wet process kilns), with the would be more appropriate. The product of kilovolts times milliamps for
potential for condensation and/or water commenter notes that, in its various each electrostatic precipitator field) is
droplet interference. Commenters state proposals over the years, EPA has one acceptable approach as a site-
that when water droplets are present, discussed a number of approaches to specific parameter to monitor
many of these devices are not establish operating parameter limits for electrostatic precipitator performance.
applicable. electrostatic devices, including: Limits on power input for each field (or
Response: The final rule provides Minimum total secondary power; particular fields) are not warranted.
minimum secondary power by field; Response: A limit on total power
sources equipped with electrostatic
pattern of increasing power from inlet to input to a multifield electrostatic device
precipitators or ionizing wet scrubbers
outlet field; and minimum secondary is generally not an acceptable operating
the alternative of using a particulate
power of the last 1⁄3 of fields (or the last parameter for compliance assurance. We
matter detection system or establishing
field). Commenters have also proposed: have documented that when total power
site-specific operating parameter limits
minimum specific power (secondary input was held constant for a four-field
for compliance assurance. If a
power divided by flue gas flow rate); electrostatic precipitator while the
particulate matter detection system is
minimum secondary voltage and/or power input to the fourth field was
used, corrective measures must be taken
secondary current; and total secondary decreased, emissions of particulate
if the alarm set point is exceeded. If the
voltage and/or secondary current. The matter doubled from 0.06 gr/dscf to 0.12
source elects to establish site-specific
commenter concludes that it is not gr/dscf. See 66 FR at 35143 (July 3,
operating parameter limits, the limits
surprising that there is so little 2001). Thus, if the total power input
194 USEPA, ‘‘Technical Support Document for
agreement on the right approach, during the comprehensive performance
HWC MACT Standards, Volume IV: Compliance
because different units and applications test were used as the operating
with the HWC MACT Standards,’’ September 2005, respond differently. EPA’s proposal to parameter limit, particulate matter
Appendix C. let facilities and local regulators emissions could exceed the emission
standard because of changes in other in the bag leak detection system maximize controllable operating
parameters that were not limited even guidance. 195 parameters during the comprehensive
though total power input did not exceed Neither the suggestion to establish the performance test to simulate the full
the parametric limit. alarm level at the 99th percentile upper range of normal operations (e.g., by
Notwithstanding our concern that a prediction limit (UPL99) based on the spiking the ash feedrate and/or detuning
limit on total power input to a average response during the the electrostatic device).196
multifield electrostatic device is comprehensive performance test runs You may establish the alarm set-point
generally not an effective operating nor the suggestion to establish the alarm as the average detector response of the
parameter for compliance assurance, level at the maximum test run response test condition averages during the
this does not preclude you from would control PM emissions at the level comprehensive performance test.
documenting to the permitting authority achieved during the performance test or Alternatively, you may establish the
that total power input is an effective provide some assurance that the PM alarm set point by extrapolating the
compliance assurance parameter for standard was not being exceeded, unless detector response. Under the
your source. See § 63.1209(m)(1)(iv). the detector response is correlated to extrapolation approach, you must
Comment: Several commenters PM concentrations. For example, if the approximate the correlation between the
suggest that the rule should offer detector response does not relate detector response and particulate matter
various approaches to establish an linearly to PM concentration (or if the emission concentrations during an
achievable particulate matter detection response changes w/changes in initial correlation test. You may
system alarm level on a site-specific particulate characteristics), the UPL99 extrapolate the detector response
basis in lieu of the approach we detector response could relate to a much achieved during the comprehensive
proposed (i.e., average detector response higher (e.g., 99.9th percentile) PM performance test (i.e., average of the test
during the comprehensive performance concentration. In addition, even if the condition averages) to the higher of: (1)
test): (1) Use the 2 times the maximum detector response were correlated to PM A response that corresponds to 50% of
peak height or 3 times the baseline concentration, there is no assurance that the particulate matter emission
concepts developed in EPA’s bag leak the correlation would be consistent over standard; or (2) a response that
detection guidance documents; (2) allow the range of the average detector correlates to 125% of the highest
spiking to set the alarm set point given response during the performance test to particulate matter concentration used to
that PS 11 allows for spiking as a way the UPL99 detector response. Note that develop the correlation.
to calibrate PM CEMs; (3) establish the under PS–11 for PM CEMS, even after To establish an approximate
limit as the 99th percentile upper complying with rigorous procedures to correlation of the detector response to
prediction limit of the average response correlate the detector response to PM particulate matter emission
during each performance test run concentrations, the detector response concentrations, you should use as
instead of the average of the test run may be extrapolated only to 125% of the guidance Performance Specification–11
averages; (4) allow upward highest PM concentration used for the for PM CEMS (40 CFR Part 60,
extrapolation from the average of the correlation. Thus, the final rule does not Appendix B), except that you need only
test run averages to some percentage of use these approaches to establish the conduct only 5 runs to establish the
the particulate matter emissions alarm level. initial correlation rather than a
standard (fraction could be variable If you elect to use a particulate matter
minimum of 15 runs required by PS–11.
depending upon how close to the detection system in lieu of site-specific
In addition, the final rule requires you
standard the facility is during the operating parameters for your
to conduct an annual Relative Response
compliance test); or (5) set the alarm electrostatic precipitator or ionizing wet
Audit (RRA) for quality assurance as
point at the maximum test run. scrubber, you must establish the alarm
Response: We agree with several of level using either of two approaches. required by Procedure 2—Quality
the commenters’ suggestions: explicitly See Appendix C of USEPA, ‘‘Technical Assurance Requirements for Particulate
allowing spiking (and emission control Support Document for HWC MACT Matter Continuous Emission Monitoring
device detuning) during the Standards, Volume IV: Compliance with Systems at Stationary Sources,
comprehensive performance test to the HWV MACT Standards,’’ September Appendix F, Part 60.197 The RRA is
maximize controllable operating 2005. Under either approach, you may required on only a 3-year interval,
parameters to simulate the full range of however, after you pass two sequential
normal operations; and upward 195 Note that a bag leak detection system is a type annual RRAs.
extrapolation of the detector response. of particulate matter detection system for purposes The rule requires only minimal
See discussion below. of this discussion. A triboelectric detector is correlation testing because the
normally used for a bag leak detector system particulate matter detection system is
The final rule is consistent with because it is very inexpensive and has a low
commenters’ suggestion to establish the detection limit. A triboelectric detector meets the used for compliance assurance only—as
alarm level for particulate matter criterion for a particulate matter detector in a an indicator for reasonable assurance
particulate matter detection system in that it detects that an emission standard is not
detection systems on fabric filters based relative mass emissions of particulate matter within
on the concepts in the Agency’s the range of normal emission concentrations. (Note
exceeded. The particulate matter
guidance document on bag leak further, however, that a triboelectric detector cannot detection system is not used for
detection systems. Commenters made be correlated to particulate matter concentrations compliance monitoring—as an indicator
and thus cannot be used as a particulate matter of continuous compliance with an
this suggestion in response to our CEMS. Note also that a triboelectric detector cannot
request for comments on requiring be used on sources equipped with electronic
196 Note, however, that bypassing or detuning an
particulate matter detection systems on control devices.) The alarm level for a bag leak
detection system would be established using the emission control system could cause PM
fabric filters and establishing the alarm stratification and could make it difficult to pass the
concepts discussed in the Agency’s guidance
level based on the detector response document on bag leak detection systems. The alarm PS–11 performance criteria you use as guidelines
during the comprehensive performance level for a particulate matter detection system used for a PMDS.)
on a fabric filter, however, (preferable with a 197 You perform an RRA by collecting three
test. See 69 FR at 21347. The final rule
detector other than a tribolectric device that could simultaneous reference method PM concentration
requires bag leak detection systems on be correlated to PM concentrations) would be measurements and PM CEMS measurements at the
all fabric filters and suggests that you established based on the detector response during as-found source operating conditions and PM
establish the alarm level using concepts the comprehensive performance test. concentration.
emission standard. Because particulate are conducted in a manner protective of instantaneous deviations from the
matter detection system correlation human health and the environment.201 negative pressure requirements; and (2)
testing and quality assurance is much It is therefore appropriate to clarify that these requirements, though not
less rigorous than the requirements of the intent of this requirement is to standards themselves, effectively
PS–11 for a PM CEMS, the particulate control fugitive emission releases from increase the stringency of the standard
matter detection system response cannot the combustion of hazardous waste. itself beyond what even the best
be used as credible evidence of Furthermore, MACT requirements for available technology can achieve.
exceedance of the emission standard. source categories that do not combust Response: As previously discussed,
hazardous waste (e.g., industrial boilers, the fugitive emission control
D. Fugitive Emissions Portland cement kilns, and commercial requirements included in today’s rule
Comment: A commenter does not and industrial solid waste incinerators) originated from the RCRA hazardous
support EPA’s proposed approach to do not have combustion chamber waste combustion chamber fugitive
allow alternative techniques that can be fugitive emission control requirements emission control requirements. These
demonstrated to prevent fugitive for the non-hazardous waste inputs or provisions allow sources to control
emissions without the use of outputs (e.g., clinker product for cement fugitive emissions by ‘‘maintaining the
instantaneous pressure limits given that kilns or coal and natural gas fuels for combustion zone pressure lower than
the CAA requires continuous industrial boilers). We have previously atmospheric pressure, or an alternative
compliance with the standards and taken the position that emissions not means of control equivalent to
given positive pressure events can result affected by the combustion of hazardous maintenance of combustion zone
in fugitive emissions, irrespective of waste (e.g., clinker coolers, raw material pressure lower than atmospheric
facility design. handling operations, etc.) are regulated pressure.’’ All sources that must comply
Response: Rotary kilns can be pursuant to the applicable nonazardous with the provisions of this rule are, or
designed to prevent fugitive emissions waste MACT rules.202, 203 We conclude were, required to control fugitive
during positive pressure events. As the clarification that the fugitive emissions from the combustion unit
stated in the February 14, 2002 final emission control requirements applies pursuant to RCRA.
rule, and subsequently in the April 20, only to fugitive emissions that result The monitoring requirements in
2004 proposed rule, there are state-of- from the combustion of hazardous waste today’s rule do not increase the
the-art rotary kiln seal designs (such as is appropriate because it regulates stringency of the standard beyond what
those with shrouded and pressurized emissions attributable to nonhazardous the best available technology can
seals) which are capable of handling waste streams to the same level of achieve. Although we do not have data
positive pressures without fugitive stringency that otherwise would apply if that confirm negative pressure is being
releases. See 67 FR at 6973 and 69 FR the source did not combust hazardous maintained on an instantaneous basis
at 21340. We have included waste.204 (as we define it)205 for at least 12
documentation of such kiln designs in Comment: A commenter states that percent of the boilers, we believe this is
the docket.198 Instantaneous combustion the instantaneous monitoring current practice and readily achievable
zone pressure limits thus may not be requirements are inappropriate because by most sources.206 These requirements
necessary to assure continuous (1) EPA has not demonstrated that the have been in force for many years, and
compliance with these fugitive emission average of the top 12% of boilers are there is no basis for stating that they are
control requirements. Our approach to capable of operating with no unachievable (EPA is not aware of
allow alternative techniques that have industrywide noncompliance with these
been demonstrated to prevent fugitive 201 Section 3004(a) of RCRA requires the Agency provisions, the necessary premise of the
emissions is therefore reasonable and to promulgate standards for hazardous waste comment). First, maintaining negative
appropriate. We note that these treatment, storage, and disposal facilities as pressure is the option that most boilers
alternative techniques must be reviewed necessary to protect human health and the elect to implement to demonstrate
environment. The standards for hazardous waste
and approved by the appropriate incinerators generally rest on this authority. compliance with the RCRA fugitive
delegated regulatory official. 199
§ 3004(q) of RCRA requires the Agency to emission control requirements. Second,
Comment: A commenter disagrees promulgate standards for emissions from facilities negative pressure is readily achieved on
with EPA’s clarification that fugitive that burn hazardous waste fuels (e.g., cement and an instantaneous basis in boilers
emission control requirements apply lightweight aggregate kilns, boilers, and
hydrochloric acid production furnaces) as necessary through use of induced draft fans.
only to fugitives attributable to the to protect human health and the environment. Third, the requirements we are
hazardous waste, given that the CAA 202 See 69 FR at 21203 and 64 FR at 52871, and finalizing today for boilers are identical
does not distinguish between HAP § 63.1206(b)(1)(ii). to the fugitive emission control
emissions that come from hazardous 203 Portland cement manufacturing facilities that
requirements that hazardous waste
waste streams and other HAP emissions. combust hazardous waste are subject to both incinerators, cement kilns, and
Subpart EEE and Subpart LLL, and hydrochloric
Response: The fugitive emission acid production facilities that combust hazardous lightweight aggregate kilns are currently
control requirements in today’s final waste may be subject to both Subpart EEE and complying with pursuant to the EEE
rule originated from the RCRA Subpart NNNNN. In these instances Subpart EEE interim standard regulations. See
hazardous waste combustion fugitive controls HAP emissions from the cement kiln and
hydrochloric acid production furnace stack (and § 63.1206(c)(5). Most of these sources
emission control requirements for also fugitive emissions from the combustion maintain negative combustion chamber
incinerators and boilers and industrial chamber), while Subparts LLL and NNNNN would pressure through use of induced draft
furnaces. 200 The primary focus of these control HAP emissions from other operations that fans, providing further evidence that
RCRA requirements is to ensure are not directly related to the combustion of
hazardous waste (e.g., clinker cooler emissions for continuously maintaining combustion
hazardous waste treatment operations cement production facilities, and hydrochloric acid
product transportation and storage for hydrochloric 205 The February 14, 2002 Final Amendments
198 See USEPA, ‘‘Technical Support Document for acid production facilities). Rule clarifies that that a reasonable pressure
the HWC MACT Standards, Volume IV: Compliance 204 This issue has little relevance given that the monitoring frequency that could meet the intent of
With the HWC MACT Standards,’’ September 2005, measures taken to control the fugitive emissions ‘‘instantaneous’’ would be once every second. See
Section 10. from the combustion of hazardous waste will also 67 FR at 6974.
199 See § 63.1206(c)(5)(i)(C) and (D).
control the fugitive emission associated with other 206 Commenters did not provide data to the
200 See § 266.102(e)(7) and § 264.345(d). feedstreams. contrary.
zone pressure lower than ambient E. Notification of Intent To Comply and days for comment, and an opportunity
pressure is readily achievable by well Compliance Progress Report to request a hearing.
designed and operated boilers.207 While we do agree that the Title V re-
We note that use of instantaneous 1. Notice of Intent To Comply opening and renewal requirements
pressure monitoring is not a In the NPRM, we proposed to re- provide adequate information to the
requirement. A source can elect to institute the Notification of Intent to public and an opportunity for the public
implement any of the four compliance Comply (NIC) because we felt that it to comment and request a hearing, we
options to control combustion system offered many benefits in the early stages are concerned that the timing
leaks as well as request to use of MACT compliance. As discussed in requirements for the NIC may not
alternative monitoring approaches. See the 1998 ‘‘fast track’’ rule (63 FR 33782) correspond with the timing
§§ 63.1206(c)(5) and 63.1209(g). The and in the proposal, the NIC serves requirements for title V permit
instantaneous pressure monitoring several purposes: as a planning and reopenings, revisions, and renewals.
option offers sources a method that communication tool in the early The public review of the draft NIC and
satisfies the intent of the rule that can implementation stages, to compensate subsequent public meeting are
be applied at numerous sources. The for lost public participation scheduled to occur 9 and 10 months,
inclusion of this requirement in today’s opportunities when using the RCRA respectively, after the rule’s effective
rule is thus an attempt to simplify the streamlined permit modification date. On the other hand, Title V permits
review process for both regulators and procedure to make upgrades for MACT for major sources that have a remaining
affected sources; the absence of compliance, and as a means to share permit term of greater than 3 years from
prescriptive compliance options in this information and provide public the rule’s promulgation date will need
case may likely result in time- participation opportunities that would to be re-opened, but this re-opening may
consuming site-specific negotiations be lost when new units are not required not occur until 18 months beyond the
that would prolong the review and to comply with certain RCRA permit promulgation date of the rule. Also,
approval of comprehensive performance requirements and performance Title V permits that have a remaining
test workplans. standards. Please refer to the proposal at permit term of less than 3 years from the
Comment: A commenter believes that 69 FR 21313–21316 for additional rule’s promulgation date will need to be
requiring an instantaneous waste-feed discussion of the regulatory history, renewed, but the timing of the renewal
cutoff when these pressure excursions purpose, and implementation of the NIC is contingent upon the individual
occur is short-sighted and will result in provisions. permit term, not the timing
greater HAP emissions. The commenter requirements for public review of the
Overall, most commenters supported
recommends EPA instead allow the use draft NIC and public meeting. Thus, we
our decision to finalize NIC provisions.
of reasonable pressure averaging periods do not believe there is ample
in lieu of instantaneous pressure However, they also feel that the NIC
should only be required for sources that opportunity to combine the
requirements. requirements of the NIC and Title V
Response: As discussed in the have not completed a NIC previously
(i.e., Phase 2 sources or Phase 1 sources process for the vast majority of
February 14, 2002 Final Amendments
that did not meet the previous NIC sources.208 Also, those sources that need
Rule, automatic waste feed cutoffs are
deadline) and for sources that need to to make upgrades to comply with the
appropriate non-compliance deterrents,
make upgrades to comply with the final final standards and that need to modify
and are necessary whenever an
standards (i.e., either Phase 1 or Phase any applicable conditions in their RCRA
operating limit is exceeded. See 67 FR
2). They suggest that if sources do not permit will not be able to request the
at 6973. Pressure excursions that result
need to make upgrades, then they streamlined modification procedure (see
in combustion system leaks (and
should not be required to complete the 40 CFR 270.42(j)) until they meet the
subsequently lead to automatic waste
feed cutoffs) should be prevented by NIC process, if they had done so NIC requirements. So the earlier they
maintaining negative pressure in the previously. To require a second NIC comply with the NIC requirements, the
combustion zone. We agree that would only add to the administrative earlier they can begin upgrading their
needless triggering of automatic waste burden and is not in line with Agency combustion units.
efforts to reduce reporting burdens. We Another commenter suggested a
feed cutoffs due to short term pressure
agree that if Phase 1 sources do not need change to the regulations at
fluctuations that do not result in
to make upgrades to comply with the § 63.1210(c)(1) to account for sources
combustion system leaks would provide
less environmental protection, not more. Replacement Standards and if they that will cease burning hazardous waste
Today’s rule offers three alternative completed the NIC process before, then prior to or on the compliance date. The
options that do not require the use of it is not necessary to do so again. regulations, as proposed, require
instantaneous pressure monitoring to In addition to the comment discussed sources to hold an informal public
control combustion system leaks. See above, a few commenters proposed that meeting to discuss anticipated activities
§ 63.1206(c)(5). The use of averaging for sources who must still comply with described in the draft NIC even if they
periods in these alternatives is not the NIC because they wish to make plan to cease burning hazardous waste.
prohibited. Sources that elect to use an upgrades, that the NIC public notice be The commenter also suggested a similar
alternative compliance option must combined with the Title V re-opening or change to § 63.1210(b)(2) that requires
demonstrate that the alternative method renewal public notice. They point out the draft NIC be made available for
is equivalent to maintaining combustion that sources with existing Title V public review no later than 30 days
zone pressure lower than ambient permits will have their permits re- 208 We recognize that there may be instances
pressure or, that the alternative opened or renewed to incorporate the when states can coordinate the Title V permit re-
approach prevents fugitive emissions. new applicable requirements (i.e., Phase opening, revision, and renewal process with the
1 Replacement or even Phase 2 NIC timeframe requirements. Where this is possible,
207 The commenter did not provide information
Standards) shortly after the NIC public we encourage states (or other permitting authorities)
that would lead us to conclude that these to coordinate the two processes. By coordinating
requirements are harder to implement for boilers
notice and meeting are to occur. Title V the two, duplication with respect to material
than for incinerators, cement kilns, and lightweight permit re-openings and renewals content and public participation would be
aggregate kilns. require: public notice, a minimum of 30 eliminated for both sources and states.
prior to the public meeting. We agree commenter’s concerns that the term you standard is defined by 42 U.S.C.
with the commenter that it does not ‘‘will’’ implies that sources are required § 7602(k) as a standard that ‘‘limits the
make sense to require sources that meet their ‘‘estimated’’ dates for quantity, rate, or concentration of
intend to cease burning hazardous waste achieving key activities. We have emissions of air pollutants on a
to submit a NIC that discusses removed ‘‘will’’ and replaced it with continuous basis, including any
anticipated activities that will allow ‘‘anticipate’’ to more accurately requirement relating to the operation of
them to achieve compliance with the represent the objective of the NIC, maintenance of a source to assure
standards. We also agree that it is not which is for sources to communicate continuous emission reduction, and any
necessary for those sources to hold an their plans for complying with the design, equipment, work practice or
informal public meeting, since there are standards in two years. operational standard * * *.’’ The
no MACT compliance activities to commenter concludes that a standard
2. Compliance Progress Report
discuss. However, we believe that the that contains a malfunction exemption
public should be provided notice of the In the proposal, we explained why we does not apply ‘‘on a continuous basis’’
draft NIC so that they are aware of the thought a compliance progress report as required by the statute. Likewise, the
source’s intentions to cease burning and would be beneficial. In short, we commenter concludes that an
the steps (and key dates) the source will believed it would help regulatory exemption for startup and shutdown
undertake to stop hazardous waste agencies determine whether Phase 1 and periods would also violate this
combustion activities. Phase 2 sources were making sufficient unambiguous statutory language.
With regard to Phase 2 sources, we headway in their efforts to meet the The commenter also notes that,
had proposed that all Phase 2 sources compliance date. The progress report although some courts have held that a
comply with the same NIC requirements would be due to the regulatory agency technology-based standard must provide
as the Phase 1 sources. Commenters did at the midway point of the 3 year some kind of an exemption for
not express opinions in favor or against compliance period and would serve to unavoidable technology failures, the
the NIC for Phase 2 sources. We believe update the information the source rationale for such an exemption is that
that the NIC is beneficial in several provided in its NIC. However, because the underlying standard is based on the
respects. As mentioned previously, it we do not have any experience to draw performance of a particular control
serves as a planning and upon regarding the value of the progress technology that cannot be expected to
communication tool in the early report, we requested comment on function properly all of the time. The
implementation stages, it compensates whether or not it should be required. commenter believes that neither the
for lost public participation In response to our request for rationale nor the exemption apply to
opportunities when using the RCRA comment, all commenters were opposed section 112(d) standards, which are not
streamlined permit modification to the progress report. They cited based on the performance of any
procedure to make upgrades for MACT several reasons, with the most particular technology but instead must
compliance, and it is a tool to share consistent one being that the progress reflect the ‘‘maximum degree of
information and provide public report serves no useful purpose and reduction’’ that can be achieved,
participation opportunities that would imposes unnecessary additional irrespective of the measures used by a
be lost when new units are not required burdens on sources. As we discussed source to achieve that reduction. CAA
to comply with certain RCRA permit above, sources and regulatory agencies § 112(d)(2).
requirements and performance will be focusing on the NIC as well as The commenter states that, even
standards. Ultimately, it creates more initial Title V applications, re-openings, assuming for the sake of argument that
public confidence in the permitting revisions, and renewals during this EPA has authority to depart from the
process and so promotes a more stable three year compliance period. We agree statutory language and carve out a
regulatory environment. with the commenter who noted that startup, shutdown, and malfunction
For today’s rule, we are finalizing our there is already significant interaction exemption, any such exemption must be
decision to re-institute the NIC between sources and regulatory narrowly drafted to apply only where a
provisions for Phase 1 and Phase 2 authorities during this period. source demonstrates that a violation was
sources. We are including a few minor Furthermore, we learned through unavoidable. See, e.g., Marathon Oil,
changes and clarifications to improve implementation of the Interim 564 F.2d at 1272–73. As EPA
the proposed regulatory language based Standards that some regulatory agencies recognizes, emission exceedances that
on commenters’ suggestions. Section found it difficult to manage the notices, occur during SSM events are frequently
63.1210(b) is revised so that Phase 1 applications, requests, and test plans avoidable. See 69 FR at 21339/3 (noting
sources that previously complied with that were due prior to the compliance that ‘‘proper operation and maintenance
the NIC requirements, and that do not date. Therefore, we have decided not to of equipment’’ helps avoid exceedances
need to make upgrades to comply with finalize any compliance progress report during startup, shutdown, and
the Replacement Standards, are not requirements for today’s rule. malfunction events), 69 FR at 21339/2
required to comply with the NIC again. (describing the industry view that
Sections 63.1210(b)(1)(iv) and (b)(2) F. Startup, Shutdown, and Malfunction ‘‘some’’ exceedances that occur due to
have been revised and (c)(5) has been Plan malfunctions are unavoidable). Thus,
added so that sources that intend to Comment: One commenter states that the commenter concludes that, even if a
cease burning hazardous waste prior to an exceedance of a standard or Marathon Oil-type exemption applies to
or on the compliance date are only operating requirement during a a § 112(d) standard, it would be
required to prepare a (draft) NIC, make malfunction should be a violation not unlawful and arbitrary for EPA to
a draft of the NIC available for public only because source owners and exempt sources from liability for all
review no later than 9 months after the operators need an incentive to minimize emission exceedances occurring during
effective date of the rule, and submit a exceedances caused by malfunctions, startup, shutdown, and malfunction
final NIC to the Administrator no later but also because an exemption for events. Rather, such an exemption could
than one year following the effective malfunction periods would violate the only apply where a source demonstrates
date of the rule. Last, we have revised plain language of the CAA. The that a given exceedance was
language in § 63.1210(b) based upon a commenter notes that an emission unavoidable.
Many other commenters state that it standards during startup, shutdown and month block period when hazardous
would be illegal to require compliance malfunction periods was required. waste is in the combustion chamber. See
with the emission standards and We do not agree with the commenter § 63.1206(c)(2)(v). This will alert the
operating requirements during startup, who contends that the § 112(d) MACT permitting authority that the source’s
shutdown, and malfunction events. The standards are not technology-based operation and maintenance plan may
commenters note that EPA and the standards because they are not based on not be adequate to maintain compliance
courts have long recognized that the performance of any particular with the emission standards and that
technology fails at times, despite a technology but instead must reflect the the authority may need to direct the
source’s best efforts to maintain ‘‘maximum degree of reduction’’ that source to revise the plan under
compliance. For this reason, the courts can be achieved, irrespective of the § 63.6(e)(3)(vi). Finally, we note that
have recognized that technology-based measures used by a source to achieve sources must report all excess emissions
standards such as EPA’s § 112(d)(2) that reduction. On the contrary, the semiannually under § 63.10(e)(3) if an
MACT standards must account for such standards must reflect the average emission standard or operating limit is
unavoidable technology failures if the performance of the best performing exceeded, including during
standards are to be truly ‘‘achievable.’’ sources, which performance is achieved malfunctions.
Thus, the standards must excuse using technical controls—air pollution Comment: One commenter states that
noncompliance with the actual emission control devices, and for some any exemption for emission
standards during startup, shutdown, pollutants, hazardous waste feedrate exceedances during startup, shutdown,
and malfunction events. control. Those controls can fail for or malfunction events would violate the
These commenters also note that EPA reasons that are not reasonably RCRA mandate for standards necessary
took the position in the September 1999 preventable. We note further that the ‘‘to protect human health and the
final MACT rule for hazardous waste situation was the same in the Clean environment.’’ 42 U.S.C. 6924(a). The
combustors that exceedance of an Water Act cases which the commenter commenter reasons that, because EPA’s
operating requirement during startup, seeks to distinguish. Like section 112(d) RCRA standards are health-based rather
shutdown, or malfunction events was a standards, Clean Water Act standards than technology-based, no
violation if hazardous waste remained are technology-based (reflecting Best unavoidability defense is available.
in the combustion chamber. The Practicable Technology or Best Given that EPA concludes that the
commenters note that industry groups Available Technology, see CWA hazardous waste combustor MACT rule
sections 304 (b) and 301 (b)) and do not satisfies both its CAA and RCRA
challenged the rule, and while the D.C.
require use of any particular type of mandates, the emission standards and
Circuit did not reach this issue because
technology. See also Mossville, 370 F. operating requirements cannot be
it vacated the emission standards, it
3d at 1242 (EPA must account for waived during startup, shutdown, and
pointed out that ‘‘industry petitioners
foreseeable variability in establishing malfunction events.
may be correct that EPA should have Response: We agree that the RCRA
MACT floor standards).
exempted HWCs from regulatory limits We agree with the commenter who mandate to ensure protection of human
during periods of startup, shutdown, states that any exemption from the health and the environment applies at
and malfunction, permitting sources to emission standards and operating all times, including during startup,
return to compliance by following the requirements during malfunctions must shutdown, and malfunction events.
steps of a startup, shutdown, and apply only where a source demonstrates Accordingly, the existing MACT
malfunction plan filed with the that a violation was unavoidable. We requirements for incinerators, cement
Agency.’’ CKRC v. EPA, 255 F.3d 855, note that the term malfunction is kilns, and lightweight aggregate kilns
872 (2001). Commenters conclude that, defined in § 63.2 as ‘‘any sudden, give sources the option of continuing to
after reading this language, EPA officials infrequent, and not reasonably comply with RCRA permit requirements
wisely decided that hazardous waste preventable failure of air pollution to control emission during these events,
combustors should not be required to control and monitoring equipment, or to comply with special MACT
meet the MACT emission standards and process equipment, or a process to requirements that are designed to be
operating limits during startup, operate in a normal or usual manner proactive and reactive and intended to
shutdown, and malfunction events. which causes, or has the potential to be equivalent to the incentive to
Response: We agree with commenters cause, the emission limitations in an minimize emissions during these events
who state that sources must be exempt applicable standard to be exceeded. provided by the RCRA requirements.
from technology-based emission Failures that are caused in part by poor See existing § 63.1206(c)(2)(ii). The
standards and operating limits during maintenance or careless operation are special MACT requirements require
startup, shutdown, and malfunction not malfunctions.’’ We believe this sources to include proactive measures
events. Technology is imperfect and can definition largely addresses the in the startup, shutdown, and
malfunction for reasons that are not commenter’s concern. malfunction plan to minimize the
reasonably preventable. The regulations We acknowledge, however, that frequency and severity of malfunctions
must provide relief for such situations. emissions can increase during and to submit the startup, shutdown,
We believe that existing case law malfunctions and potentially exceed the and malfunction plan to the permitting
supports this position. See, e.g., standards and agree that exceedances authority for review and approval. We
Chemical Mfr’s Ass’n v. EPA, 870 F. 2d must be minimized. Accordingly, the proposed to require boilers and
at 228–230 (daily maximum limitations final rule (and the current rule for hydrochloric acid production furnaces
established at 99th percentile reasonable incinerators, cement kilns, and to comply with those same provisions
because rules also provide for upset lightweight aggregate kilns) requires that providing for equivalence between the
defense for unavoidable exceedances); sources maintain compliance with the two sets of requirements, and
Marathon Oil v. EPA, 541 F. 2d at 1272– automatic hazardous waste feed cutoff promulgate those provisions today.
73 (acknowledged by commenter). As system during malfunctions and notify Comment: One commenter states that
commenters noted, the D.C. Circuit also the permitting authority if they have 10 the rule should clarify the definitions of
intimated in CKRC that some type of or more exceedances of an emission startup, shutdown, and malfunctions to
exception from compliance with standard or operating limit during a 6- preclude sources from improperly
classifying as unavoidable exceedances expanding the scope of the startup, access to startup, shutdown, and
those exceedances that could have been shutdown, and malfunction plan on a malfunction plans, coupled with the
avoided had the source implemented an site-specific basis under authority of provisions of § 63.6(e)(3)(vii) under
appropriate operation and maintenance § 63.6(e)(3)(vii) if the source has which the Administrator must require
plan. Many other commenters state that excessive exceedances during the source to make changes to a
the current definitions in § 63.2 clearly malfunctions. See deficient plan, should ensure that
define these terms. § 63.1206(c)(2)(v)(A)(3) defining startup, shutdown, and malfunction
Response: We believe the definitions excessive exceedances during plans are complete and accurate. We
of startup, shutdown, and malfunction malfunctions and requiring reporting of note that under § 63.6(e)(3)(vii) the
are clearly defined in § 63.2, and the exceedances in the excess emissions Administrator must require the source
combined with the startup, shutdown, report required under § 63.10(e)(3). to revise the plan if the plan: (1) does
and malfunction plan requirements, will Comment: Two commenters state that not address a startup, shutdown, or
preclude sources from improperly all startup, shutdown, and malfunction malfunction event that has occurred; (2)
classifying as malfunctions events that plans should be submitted for review fails to operate the source (including
could have been reasonably prevented and approval by the delegated authority associated air pollution control and
by following appropriate procedures in and made available for a 60-day public monitoring equipment) during a startup,
the operation and maintenance plan. As review period. Review and approval of shutdown, or malfunction event in a
discussed above, the definition of the plans is needed in light of EPA’s manner consistent with the general duty
malfunction clearly states that failures acknowledgment that most excess to minimize emissions; (3) does not
that are caused in part by poor emissions would occur during startup, provide adequate procedures for
maintenance or careless operation are shutdown, and malfunctions. One of correcting malfunctioning process and/
not malfunctions. these commenters also believes that the or air pollution control and monitoring
Comment: One commenter states that regulations should provide for the equipment as quickly as practicable; or
all stack bypasses, automatic waste feed public review period to be extended as (4) includes an event that does not meet
cutoffs, and excursions from the necessary to accommodate a thorough the definition of startup, shutdown, or
operating parameter limits should be public review. The reviewing authority malfunction listed in § 63.2.
considered malfunctions. should be required to provide a written The commenter advocating that all
Response: All failures resulting in response to public comments explaining hazardous waste combustors should be
stack bypasses, automatic waste feed any decision to reject a public comment required to submit their startup,
cutoff, and excursions from the suggesting ways for a facility to limit shutdown, and malfunction plans for
operating parameter limits are not emissions during startup, shutdown, review and approval did not explain
malfunctions. As discussed above, and malfunction events. why the concerns the Agency expressed
failures caused in part by poor Many other commenters have in the General Provisions rulemaking
maintenance or careless operation are concerns with requiring review and (see 68 FR at 32589–93) are not valid for
not malfunctions. approval of startup, shutdown, and hazardous waste combustors.
Comment: One commenter states that malfunction plans, except as required
the rule should require sources to Accordingly, we do not believe it is
under § 63.1206(c)(2)(ii) for sources that appropriate to deviate from the General
expand the startup, shutdown, and elect to meet the RCRA mandate using
malfunction plan to address specific Provisions to require that all hazardous
provisions of the startup, shutdown, and waste combustors submit their startup,
proactive measures that the source has malfunction plan as discussed above.
considered and is taking to minimize shutdown, and malfunction plans for
Response: Commenters express the
the frequency and severity of review.
same views here that they expressed
malfunctions. Many other commenters under the rulemaking the Agency G. Public Notice of Test Plans
believe that it is not necessary to expand recently completed to revise the startup,
the scope of the startup, shutdown, and 1. What Are the Revised Public Notice
shutdown, and malfunction plan
malfunction plan beyond that required Requirements for Test Plans?
requirements of the General Provisions
under § 63.6(e)(3) for other MACT applicable to all MACT source Prior to the proposal, it was brought
source categories. categories. See 68 FR at 32589–93 (May to our attention that the Agency did not
Response: We do not believe that it is 30, 2003). provide any direction in the 1999 final
necessary to expand the scope of the EPA concluded in that final rule that rule regarding how and when sources
startup, shutdown, and malfunction the Administrator may at any time should notify the public, what the
plan generically for all hazardous waste request in writing that the owner or notification should include, or where
combustors to address specific proactive operator submit a copy of any startup, and for how long performance test plans
measures that the source has considered shutdown, and malfunction plan (or a should be made available.
and is taking to minimize the frequency portion thereof). Upon receipt of such a Consequently, we proposed to add
and severity of malfunctions. Imposing request, the owner or operator must clarifying language to the § 63.1207(e)(2)
additional requirements in particular promptly submit a copy of the requested public notification requirement for
situations is appropriate, however. For plan (or a portion thereof) to the approved performance test and CMS
example, as discussed above, this Administrator. In addition, the performance evaluation test plans
expanded plan is required for sources Administrator must request that the because we believe that providing
that elect to meet the RCRA mandate owner or operator submit a particular opportunities for timely and adequate
using provisions of the startup, startup, shutdown, or malfunction plan public notice is necessary to fully
shutdown, and malfunction plan. See (or a portion thereof) whenever a inform nearby communities of a
§ 63.1206(c)(2)(ii). In addition, the plan member of the public submits a specific source’s plans to initiate important
with expanded scope may be and reasonable request to examine or to waste management activities. The
appropriate for sources that have receive a copy of that plan or portion of proposed clarifications are based upon
demonstrated an inability to minimize a plan. the RCRA Expanded Public
malfunctions. Consequently, the These provisions to provide the Participation Rule (60 FR 63417,
permitting authority should consider Administrator and the public with December 11, 1995) requirements for
public notification of an impending trial In consideration of practicality, we public notice. Irrespective of the public
burn test. As a result, we did not feel believe that the second alternative notice requirements for noticing test
that the clarifications imposed any new provides an adequate solution. As we plans, we expect that sources will notify
or additional requirements upon sources mentioned, the drawback is that the their regulatory authority of their
that will conduct a MACT public may not have the opportunity to decision to proceed with their test in the
comprehensive performance test or view an approved test plan. However, absence of plan approval.
confirmatory performance test. we believe it is more important that the
2. What Are the Revised Public Notice
Commenters generally supported the public be aware of a source’s plans (i.e.,
Requirements for the Petition To Waive
clarifications to the public notice.209 how and when) for conducting the
a Performance Test?
However, they suggested a change to the performance test.210 This way, if they
proposed requirement to provide notice have questions, there will be 60 days in In the Final Amendments Rule (67 FR
of test plan approval no later than 60 which they may contact the regulatory 6968, February 14, 2002), the Agency
days prior to conducting the test. The authority or the source before the test is did not provide direction regarding
basis for suggesting a change is that scheduled to begin. This alternative will how, when, where, and what should be
many sources had not received approval also eliminate the conflict associated included in the public notice for a
of their test plans 60 days prior to the with the confirmatory performance test. petition for time extension if the
deadline for initiating their test under The regulations at § 63.1207(e)(1)(ii) Administrator fails to approve or deny
the Interim Standards. Moreover, specify that a source must submit to the test plans.211 In the proposal, we
several sources did not receive approval regulatory authority its notice of intent believed it important to provide
until well after the deadline for to conduct a confirmatory performance clarification regarding when the notice
initiating the test. The problem created test and the applicable test plans at least must be issued and what it should
for these sources is that the required 60 60 calendar days prior to the date the contain. Thus, we proposed to revise
day notification of the approved test test is to begin. Since we are no longer paragraph § 63.1207(e)(3)(iv).
requiring that the test plans be approved We received only one comment in
plan effectively determines when the
before issuing public notice, sources response to the proposed requirements.
source will be able to begin its test. In
would then provide notice of their The commenter did not express any
other words, its test would need to be
confirmatory performance test plan to concern over the requirements
postponed until the approved test plan themselves, but rather suggested a
had been noticed for 60 days. Thus, the public at the same time they submit
their notice of intent and test plans to change to terminology used. The
commenters provided several possible commenter feels that the terms ‘‘to
alternatives. the regulatory authority. Therefore, we
are requiring that sources issue the waive a performance test’’ or ‘‘waiver’’
One alternative that would avoid as used in § 63.1207(e)(3)(iv) could be
causing delays to testing is to require public notice of test plans 60 days in
advance of commencing the confusing to readers when we are
the public notice when the source actually referring to a time extension for
submits its test plan. Although this performance test, whether their test
plans have been approved or not. The commencing the test. Although we agree
fulfills the notification requirement, this the terminology could be confusing, 40
alternative has a shortfall: The notice regulations at § 63.1207(e)(2) have been
revised accordingly. CFR 63.1207(e)(3) clearly uses the term
would occur at least one year (barring ‘‘waiver’’ in the context of an extension
any extensions) in advance of the test One last concern related to the public
notice of approved test plans involves of time to conduct the performance test
and given this long period of time, the at a later date, implying that the
test plan is likely to be modified prior sources that choose to conduct a
performance test without an approved deadline can be waived in this specific
to approval. A second alternative is to situation. The use of the term waiver is
provide notice of the test plan 60 days test plan (e.g., both time extensions
provided by §§ 63.7(h) and 63.1207(e)(3) derived from the General Provisions
before the test as before, but regardless requirements for requesting a waiver of
of approval status. This alternative is have expired or due to other
circumstances, the source has elected to performance tests (§ 63.7(h)). Thus,
improved over the first, but still faces § 63.7(h)(3) provides the basis by which
the same problem of potentially not begin the test without approval).
Because we did not believe any sources sources may petition, in the form of a
offering the public an opportunity to waiver, for a time extension under
view a final approved plan. A third would choose or need to do so, we did
not propose any guidance or regulations § 63.1207(e)(3). In consideration of the
alternative is to issue notice of the test above and that the existing regulations
plan as soon as it is approved. With this specific to issuing notice to the public
of their test plans. Nevertheless, a few of § 63.1207(e)(3)(i)-(iii) consistently use
alternative, the public will have the the term waiver, we do not feel that a
most up-to-date information; however, it commenters raised this possibility
indirectly in their discussion of the change to § 63.1207(e)(3)(iv) is
may not be until a few days prior to warranted.
commencement of the test. Ideally, the problematic 60 day notice of approved
second and third alternatives could be test plan requirement. The revised H. Using Method 23 Instead of Method
combined to provide the best possible proposal addresses this concern by no 0023A
chance of providing the public with an longer requiring that test plans be Comment. Most commenters support
approved test plan in a reasonable approved before issuing public notice. our proposal to allow the use of Method
period of time prior to the test. On the Thus, sources that choose to begin their 23 instead of Method 0023A if a source
other hand, that would potentially test without an approved plan will have includes this request in the
require the facility to issue two notices complied with the requirement to issue comprehensive test plan to the
if the test plan is not approved 60 days 210 We expect that some source’s test plans may
permitting authority. Some commenters
prior to the test. We do not believe this be modified after notice is issued and prior to believe that Method 23 should be
would be reasonable given that sources approval or commencement of their test. However,
will be focused on activities associated even under the previous regulations, test plans 211 Sections 63.1207(e)(2) and (e)(3) each require
with the impending test. could be modified after they had been approved public notice, but neither had provided any
and public noticed. It is often a necessary direction on how, when, where, and what should
consequence as sources continue to prepare the be included in their respective notices until today’s
209 See 69 FR 21347–21349. combustion unit for the test. final rule.
allowed in all cases without prior a petition to EPA, but will be included approach because the emissions data
approval or on a source category basis. as a part of the performance test plan used to establish the standards are more
Response. We proposed to allow that is submitted to the delegated representative of normal emissions than
sources to use Method 23 for dioxin and regulatory authority for review and compliance test emissions.213
furan testing instead of SW–846 Method approval. This means that the expertise, As we explained at proposal, to
0023A in situations where the enhanced training, and decision-making will not ensure compliance with the mercury
procedures found in Method 0023A be consistent across the country. This is and semivolatile metal emission
would not increase measurement especially a problem because of the standards for liquid fuel boilers, you
accuracy. We proposed this change in severe resource, training and staff must document during the
the July 3, 2001, proposed rule, and reductions among the delegated comprehensive performance test a
again in the April 20, 2004, proposal. regulatory authorities across the country system removal efficiency for the metals
See 66 FR at 35137 and 69 FR at 21342. and from region to region. The decision and back-calculate from the emission
The final rule promulgates this to allow or disallow use of Method 23 standard a maximum metal feedrate
change as proposed. See should come specifically, for each case, limit that must not be exceeded on an
§ 63.1208(b)(1)(i). You may use Method from EPA consideration of the (not to exceed) annual rolling average.
23 in lieu of Method 0023A after submitted justification, based on the See 69 FR at 21311–12. If your source
justifying use of Method 23 as part of knowledge and expertise of trained and is not equipped with an emission
your performance test plan that must be experienced EPA staff. This is important control system (such as activated carbon
reviewed and approved the delegated for uniformly applying the testing to control mercury) for the metals in
permitting authority. You may be requirements all across the country.’’ question, however, you must assume
approved to use Method 23 considering Response: We disagree, and we zero system removal efficiency. This is
factors including whether previous believe the responses to comments in because, although a source that is not
Method 0023A analyses document that today’s rule make clear when Method 23 equipped with an emission control
dioxin/furan are not detected, are is an acceptable substitute for Method system may be able to document a
detected at low levels in the front half 0023A. If the source has carbon in the positive system removal efficiency in a
of Method 0023A, or are detected at flue gas, as is the case with coal-fired single test, that removal efficiency is not
levels well below the emission standard, boilers, boilers with carbon injection, likely to be reproducible. Rather, it is
and the design and operation of the and other sources likely to have a likely to be an artifact of the calculation
combustor has not changed in a manner substantial amount of carbonaceous of emissions and feeds rather than a
that could increase dioxin/furan particulate matter in the flue gas, removal efficiency that can reliably be
emissions. We note that coal-fired Method 0023A will generally be repeated.
boilers and combustors equipped with preferable because it includes To ensure that you can calculate a
activated carbon injection systems may procedures to account for dioxin and valid, reproducible system removal
not be able to support use of Method 23, furan bound to carbonaceous particulate efficiency for sources equipped with a
however, because these sources’ stack matter found in the probe and filter. In control system that effectively controls
gas is likely to contain carbonaceous other situations, Method 23 will the metal in question, you may need to
particulate. Thus, these sources are generally give the same results at a spike metals in the feed during the
likely to benefit the most from using lower cost. comprehensive performance test at
Method 0023A. levels that may result in emissions that
The final rule does not automatically I. Extrapolating Feedrate Limits for
are higher than the standard. This is
allow use of Method 23 for particular Compliance With the Liquid Fuel Boiler
appropriate because compliance with an
source categories because we cannot Mercury and Semivolatile Metal
emission standard derived from normal
assess whether all sources in a category Standards
emissions data is based on compliance
meet the conditions for use of Method Comment: One commenter questions with an (not to exceed) annual average
23—generally that quality assurance whether allowing sources to extrapolate feedrate limit calculated as prescribed
may not be improved—such as those metal feedrates downward from the here, rather than compliance with the
listed above. These determinations can levels achieved during the emission standard during the
only be made on a site specific basis by comprehensive performance test to comprehensive performance test.214
the permitting authority most familiar establish a metal feedrate limit will The commenter is concerned that
with the particular source. ensure compliance with the emission downward extrapolation from the levels
Comment: Commenters do not believe standards. achieved during the comprehensive
that an additional petition process (i.e., Response: The mercury and performance test to establish a metal
under § 63.1209(g)(1)) is necessary semivolatile metals standards for liquid feedrate limit may not ensure
before allowing use of Method 23. fuel boilers are annual average emission
Instead, EPA should require that the use limits where compliance is established to calculate the rolling average period you select.
of Method 23 should be submitted with by a rolling average mercury feedrate We note that this is an approach allowed for
limit with an averaging period not to calculating rolling averages under different modes
the test plan to the regulatory agency for of operation at § 63.1209(q)(2)(ii). At that time and
approval. exceed an annual rolling average thereafter, you update the rolling average feedrate
Response: We agree that a separate (updated hourly).212 We use this each hour with a 60-minute average feedrate.
petition is unnecessary. Sources should 213 See USEPA, ‘‘Technical Support Document for
include a justification to use Method 23 212 If you select an averaging period for the HWC MACT Standards, Volume III: Selection of
feedrate limit that is greater than a 12-hour rolling HWC MACT Standards,’’ September 2005, Section
in the performance test plan that is 13.
average, you must calculate the initial rolling
submitted for review and approval. This average as though you had selected a 12-hour 214 The emission standard accounts for long-term
will allow the permitting authority to rolling average, as provided by § 63.1209 (b)(5)(i). variability by incorporating an (not to exceed)
determine whether use of Method 23 is This is reasonable because allowing a longer period annual averaging period that is implemented by an
appropriate for the source. of time before calculating the initial rolling average (not to exceed) annual average chlorine feedrate
would not effectively ensure compliance with the limit. Thus, because the emission level achieved
Comment: Two commenters state that feedrate limit. You must calculate rolling averages during the performance test relates to daily (or
‘‘the justification of the use of Method thereafter as the average of the available one-minute hourly) variability, an exceedance of the emission
23 will not be by the existing system of values until enough one-minute values are available standard during the test is not a violation.
compliance with the standard because approach we use for semivolatile metals protective under Subpart EEE. For
system removal efficiency may be lower for dry scrubbers will also be example, the commenter notes that
at lower feedrates. appropriate to extrapolate a semivolatile § 63.1206(b)(1)(ii) could be incorrectly
This is a valid concern, and we have metal feedrate limit for wet scrubbers. interpreted to require a source to
investigated it since proposal. We To ensure that downward extrapolation comply with illogical requirements
conclude that downward extrapolation of the feedrate limit is conservative and when the source temporarily switches to
of feedrates for the purpose of to be practicable, you should limit alternative, otherwise applicable
complying with the mercury and semivolatile metal emissions during the standards, including standards testing
semivolatile metals emission standards comprehensive performance test to five and opacity monitoring under the
for liquid fuel boilers will ensure times the emission standard. alternative section 112 or 129
compliance with the emission standards For mercury, ensuring control with requirements. The commenter states
under the conditions discussed below. wet systems is more complicated that this interpretation makes little
We investigated the theoretical because the level of chlorine present sense because a source that temporarily
relationship between stack gas affects the formation of mercuric changes its mode of operation will
emissions and feedrate considering chloride which is soluble in water and continue to do testing under Subpart
vapor phase metal equilibrium, the easily controlled by wet scrubbers. EEE, Part 63, or, in the case of opacity,
chlorine, mercury, and semivolatile Elemental mercury has very low the alternative section 112 requirements
metal feedrates for liquid fuel boilers in solubility in scrubber water and is not for cement kilns would necessarily
our data base, and the mercury and controlled. The worst-case situation for require duplicate systems and
semivolatile emission standards for conversion of elemental mercury to compliance with redundant limits
liquid fuel boilers.215 We considered soluble mercuric chloride would be because a source may already be using
sources equipped with dry particulate when the chlorine MTEC is lowest and a bag leak detection system or a
matter controls and sources equipped the mercury MTEC is highest. We particulate matter detection system. The
with wet particulate matter controls. conclude that downward extrapolation commenter suggests only requiring
Sources Equipped with Dry Controls. of mercury feedrates is conservative for sources to comply with the otherwise
For sources equipped with dry controls feedstreams that contain virtually no applicable emission standards under the
other than activated carbon, mercury is chlorine, e.g., below an MTEC of 100 µg/ alternative section 112 or 129
not controlled. Thus, you must assume dscm. In addition, we conclude that requirements while still operating under
zero system removal efficiency. downward extrapolation is the various associated compliance
Consequently, if you are in the low Btu appropriate 217 for boilers feeding requirements of Subpart EEE, part 63.
subcategory and comply with the chlorinated feedstreams provided that Response: The commenter requests
mercury standard expressed as a mass during the performance test: (1) clarification of § 63.1206(b)(1)(ii), which
concentration (µg/dscm), the mercury Scrubber blowdown has been states that if a source is not feeding
feedrate limit expressed as an MTEC minimized and the scrubber water has hazardous waste to the combustor and
(maximum theoretical emission reached steady-state levels of mercury the hazardous waste residence time has
concentration, µg/dscm) is equivalent to prior to the test (e.g., by spiking the expired (i.e., the hazardous waste feed
the emission standard.216 If you are in scrubber water); (2) scrubber water pH to the combustor has been cut off for a
the high Btu subcategory and comply is minimized (i.e., you establish a period of time not less than the
with the mercury standard expressed as minimum pH operating limit based on hazardous waste residence time), then
a hazardous waste thermal emission the performance test as though you were the source may elect to comply
concentration (lb/MM Btu), the mercury establishing a compliance parameter for temporarily with alternative, otherwise
feedrate limit expressed as a hazardous the total chlorine emission standard); applicable standards promulgated under
waste thermal feed concentration (lb/ and (3) temperature of the scrubber the authority of sections 112 and 129 of
MM Btu) is also equivalent to the water is maximized (i.e., you establish a the Clean Air Act.218 As we have
emission standard. maximum scrubber water temperature explained in previous notices,219
For semivolatile metals, the limit). sources that elect to invoke
theoretical relationship between § 63.1206(b)(1)(ii) to become
J. Temporary Compliance With temporarily exempt from the emission
emissions and feedrate indicates that Alternative, Otherwise Applicable
downward extrapolation introduces standards and operating requirements of
MACT Standards Subpart EEE, Part 63, remain an affected
only a trivial error’0.17% at an emission
rate 100 times the standard irrespective Comment: One commenter requests source under Subpart EEE (and only
of the level of chlorine present. Id. clarification on the requirements Subpart EEE) until the source is no
Nonetheless, to ensure the error is applicable to a source that switches to longer an affected source by meeting the
minimal and to be practicable, you an alternative mode of operation when requirements specified in Table 1 of
should limit semivolatile emissions hazardous waste is no longer in the § 63.1200. Of course, a source can elect
during the comprehensive performance combustion chamber under the not to use the alternative requirements
test to five times the emission standard. provisions of § 63.1206(b)(1)(ii). The for compliance during periods when
Sources Equipped with Wet commenter suggests that
Scrubbers. For sources equipped with § 63.1206(b)(1)(ii) can imply that the 218 Examples include 40 CFR part 60, subparts
complete compliance strategy needs to CCCC and DDDD for commercial and industrial
wet scrubbers, we conclude that the solid waste incinerators, 40 CFR part 63, subpart
be switched over to the alternative LLL for Portland cement manufacturing facilities,
215 USEPA, ‘‘Technical Support Document for section 112 or 129 requirements, even 40 CFR part 63, subpart DDDDD for industrial/
HWC MACT Standards, Volume IV: Compliance though compliance with the Subpart commercial/institutional boilers and process
with the HWC MACT Standards,’’ September 2005, EEE requirements for monitoring, heaters, and 40 CFR part 63, subpart NNNNN for
Section 2.5 and Appendix B. hydrochloric acid production facilities.
216 Note, however, that you convert the MTEC
notification, reporting, and 219 This provision has been discussed in several
(µg/dscm) to a mass feedrate (lb/hr) by considering recordkeeping remains environmentally Federal Register notices including 64 FR at 52904
the average gas flowrate of the test run averages (September 30, 1999), 66 FR at 35090, 35145 (July
during the comprehensive performance test to 217 Mercury SRE is constant as the mercury 3, 2001), 67 FR at 6979 (February 14, 2002), and 69
simply implementation and compliance. feedrate decreases. FR at 21203 (April 20, 2004).
they are not feeding hazardous waste, EEE. An example would be a cement compliance with the emissions
but, if so, the source must comply with kiln source complying with the dioxin standards and the associated operating
all of the operating and monitoring and furan monitoring requirements limits of alternative sections 112 and
requirements and emission standards of under § 63.1209(k) of Subpart EEE for 129 standards.
Subpart EEE at all times.220 To the dioxin and furan standards under Finally, the commenter notes that
implement § 63.1206(b)(1)(ii) a source § 63.1343(d) under Subpart LLL. We did Subpart LLL (the alternative section 112
defines the period of compliance with not determine, when promulgating the standards for cement kilns) includes
the otherwise applicable sections 112 provisions of §§ 63.1206(b)(1)(ii) and opacity monitoring while Subpart EEE
and 129 requirements as an alternative 63.1209(q)(1), that the monitoring may not. The commenter states that this
mode of operation under § 63.1209(q). provisions under Subpart EEE are unnecessarily would require duplicate
In order to be exempt from the emission equivalent to the associated monitoring systems and compliance with redundant
standards and operating requirements of requirements under the otherwise limits because of the bag leak detection
Subpart EEE, a source documents in the applicable 112 and 129 standards, or and particulate matter detection system
operating record that they are indeed, whether they are even well- requirements under Subpart EEE. We
complying with the otherwise matched. Such a determination would respond that Subpart LLL specifies
applicable Section 112 and 129 require notice and opportunity for opacity as a standard (see
requirements specified under comment, which we have not provided. § 63.1343(b)(2)), and, therefore, cement
§ 63.1209(q). However, this should not be interpreted kilns subject to Subpart EEE must
The commenter recommends that the to mean that a similar determination comply with the opacity standard when
complete compliance strategy need not could not be made on a site-specific electing to comply temporarily with the
be switched over to the alternative basis given that the MACT general requirements of Subpart LLL. We note
section 112 and 129 requirements when provisions allow a source to request that the opacity standard under Subpart
temporarily switching to the alternative alternative monitoring procedures under EEE does not apply to cement kilns that
standards. In general, we disagree. The § 63.8(f)(4). Certainly, a source can are equipped with a bag leak detection
intent of § 63.1206(b)(1)(ii) is to ensure apply under this provision that the system under § 63.1206(c)(8) and to
that a source is complying with all compliance requirements under Subpart sources using a particulate matter
requirements of sections 112 and 129 as EEE satisfy the associated monitoring detection system under § 63.1206(c)(9).
an alternative mode of operation in lieu requirements under the otherwise However, a cement kiln may use an
of the requirements under Subpart EEE. applicable 112 and 129 standards. opacity monitor that meets the detection
In the 1999 final rule we stated that the limit requirements as the detector for a
We also disagree with the commenter bag leak detection system or particulate
source must comply with all otherwise that emissions testing under the
applicable standards under the matter detection system. See Part Four,
alternative standards of sections 112 Section VIII.A-C of the preamble.
authority of sections 112 and 129. and 129 is an example of an illogical
Specifically, the source must comply requirement under § 63.1206(b)(1)(ii). K. Periodic DRE Testing and Limits on
with all of the applicable notification Performance testing generally is Minimum Combustion Chamber
requirements of the alternative required to demonstrate compliance Temperature for Cement Kilns
regulation, comply with all of the with the emission standards and to Comment: Several commenters
monitoring, recordkeeping, and testing establish limits on specified operating oppose the need for cement kilns that
requirements of the alternative parameters to ensure compliance is burn at locations other than the normal
regulation, modify the Notice of maintained. In order to take advantage flame zone to demonstrate compliance
Compliance (or Documentation of of the alternative under with the destruction and removal
Compliance) to include the alternative § 63.1206(b)(1)(ii), a source needs to efficiency (DRE) standard during each
mode(s) of operation, and note in the show that compliance with and comprehensive performance test. These
operating record the beginning and end establish operating parameter limits for commenters recommend that EPA
of each period when complying with the the otherwise applicable standards of remove the requirement of
alternative regulation. See 64 FR at sections 112 and 129. Thus, testing in § 63.1206(b)(7)(ii) for cement kilns
52904. A source that elects to comply order to establish operating parameter citing that existing rule provisions (i.e.,
with otherwise applicable standards limits will be necessary. However, this the requirements under § 63.1206(b)(5)
under § 63.1206(b)(1)(ii) must specify all does not mean that a separate pertaining to changes that may
requirements of those standards, not performance test with the alternative adversely affect compliance) are
only the emission standards applicable sections 112 or 129 standards is sufficient to require additional DRE
under the sections 112 and 129 necessarily required. We note that a testing after changes are made that may
standards, but also the associated source can make use of the performance adversely affect combustion efficiency.
monitoring and compliance test waiver provision under § 63.7(h) of Commenters question EPA’s position
requirements and notification, the general provisions to request that that cement kilns that burn hazardous
reporting, and recordkeeping the performance test under the waste at locations other than the normal
requirements in the operating record alternative sections 112 and 129 flame zone demonstrate a variability in
under § 63.1209(q). standards be waived because the source DRE sufficient to justify the expense of
The commenter suggests that a source is meeting the relevant standard(s) on a re-testing for DRE with each
should be able to comply with the continuous basis by continuing to performance test. Commenters point to
otherwise applicable emission comply with Subpart EEE for the EPA’s data base that includes DRE
standards, while continuing to operate relevant HAP. This approach may be results from over 30 tests with nearly
under the associated compliance practicable for sources that can 250 runs showing consistent DRE
requirements for the HAP under Subpart demonstrate that their level of results, including sources burning
220 However, the operating requirements do not
performance during testing under hazardous waste at locations other than
apply during startup, shutdown, or malfunction
Subpart EEE, including the associated the normal flame zone, being achieved
provided that hazardous waste is not in the operating and monitoring limits, will by cement kilns. The commenters note
combustion chamber. See § 63.1206(b)(1)(i). undoubtedly ensure continuous several burdens associated with DRE
testing that do not result in improved the hazardous waste firing system, and standard. Of these, approximately one-
environmental benefit including the that those variations may not be quarter of the data are from cement kilns
purchase of expensive exotic virgin identical or limited through operating that burned hazardous waste at
chemicals for performance testing, the limits set during a single DRE test locations other than the normal flame
risks to workers and contractors (similar to what we concluded for zone (e.g., injecting waste at midkiln in
associated with the handling of these sources that burn hazardous waste only a wet process kiln), but we do not have
chemicals, and increasing the length of in the normal flame zone). See 64 FR at DRE results from every operating
operation at stressful kiln operating 52850. cement kiln. Considering available DRE
conditions necessary to conduct DRE Commenters now question the need data and the concerns of the
testing at minimum combustion for subsequent DRE testing at cement commenters, we believe that DRE
chamber temperatures. Alternatively, kilns that feed hazardous waste at testing during three consecutive
commenters recommend that EPA revise locations other than the normal flame comprehensive performance tests is
the DRE requirements such that periodic zone once a cement kiln demonstrates sufficient to provide needed certainty
testing is no longer required for cement compliance with the MACT DRE about DRE performance while reducing
kilns (that burn at locations other than standard. The regulatory requirement the overall costs and toxic chemical
the normal flame zone) after they have for the destruction and removal handling concerns to the regulated
successfully achieved the DRE standard efficiency standard has proved to be an source. Thus, we are revising the
over multiple testing cycles (e.g., two or effective method to determine requirements of § 63.1206(b)(7)(ii) such
three) under similar testing regimes. appropriate process controls necessary that cement kilns that feed hazardous
That is, the source should only be for the combustion of hazardous waste. waste at locations other than the normal
required to demonstrate compliance We are not convinced that only one DRE flame zone need only demonstrate
with the DRE standard a maximum of test is sufficient to ensure that a cement compliance with the DRE standard
two or three times until the source (that kiln that burns hazardous waste at during three consecutive comprehensive
burns at locations other than the normal locations other than the normal flame performance tests provided that the
flame zone) modifies the system in a zone will continue to meet the DRE source has successfully demonstrated
manner that could affect the ability of it standard because temperatures are compliance with the DRE standard in
to achieve the DRE standard. lower and gas residence times are each test and that the design, operation,
Response: We are revising the shorter at the other firing locations. This and maintenance features of each of the
requirements of § 63.1206(b)(7)(ii) such is especially true given the industry three tests are similar. If a facility
that cement kilns that feed hazardous trend to convert to the more thermally wishes to operate under new operating
waste at locations other than the normal efficient preheater/precalciner kiln parameter limits that could be expected
flame zone need only demonstrate manufacturing process.222 Precalciner to affect the ability to meet the DRE
compliance with the DRE standard kilns use a secondary firing system (i.e., standard, then the source would need to
during three consecutive comprehensive flash furnace) at the base of the conduct another DRE test. Once the
performance tests provided that the preheater tower to calcine the raw facility has conducted another three
source has successfully demonstrated material feed outside the rotary kiln. DRE tests under the new operating
compliance with the DRE standard in This results in two separate combustion limits, then subsequent DRE testing
each test and that the design, operation, processes that must be controlled ‘‘ one would not be required. Accordingly, we
and maintenance features of each of the in the kiln and the other in the flash are revising the requirements of
three tests are similar. These revisions furnace. The gas temperature necessary § 63.1206(b)(7)(ii).
do not affect sources that burn for calcining the limestone raw material
hazardous waste only in the normal Comment: Several commenters
in the flash furnace is lower than the support EPA’s proposal to delete the
flame zone.221 temperature required making the clinker
Prior to today’s change, we required requirement to establish an operating
product. We conclude, therefore, that it limit on the minimum combustion
sources that feed hazardous waste in is necessary, in spite of the concerns
locations other than the flame zone to chamber temperature for dioxin/furans
raised by commenters, to retain periodic under § 63.1209(k)(1) for cement kilns.
perform periodic DRE testing every 5 DRE testing to ensure continued
years to ensure that the DRE standard These commenters point to the high
compliance with the DRE standard temperatures of approximately 2500°F
continues to be achieved over the life of necessary for the control of nondioxin/
the unit. See § 63.1206(b)(7)(ii). We required to make the clinker product.
furan organic HAP. These high temperatures are fixed by
justified this requirement because of We also acknowledge, however, the
concerns that sources that feed the reaction kinetics and
concerns raised by the commenters. Our thermodynamics occurring in the
hazardous waste at locations other than DRE data base of operating cement kilns
the flame zone have a greater potential burning zone and cannot be reduced
includes results from approximately 25 below minimum values at the whim of
of varying DRE performance due to their DRE tests and nearly 200 runs.223 All
hazardous waste firing practices. As we the operator and still make a marketable
data show compliance with the DRE product. In addition to deleting the
stated in the 1999 rule, we were
concerned that the DRE may vary over 222 For example, Ash Grove Cement in Chanute,
minimum combustion chamber
time due to the design and operation of KS replaced their two wet process cement kilns temperature limit for dioxin/furans,
with one preheater/precalciner kiln in 2001. Holcim commenters also recommend, for
221 The DRE demonstration for these sources need Inc in Holly Hill, SC has also recently constructed similar reasons, that EPA delete the
be made only once during the operational life of a a new preheater/precalciner kiln to replace two wet minimum combustion chamber
source, either before or during the initial process cement kilns. Keystone Cement Company
comprehensive performance test, provided that the in Bath, PA is considering replacing their two wet temperature requirement under
design, operation, or maintenance features do not process cement kilns with a new preheater/ § 63.1209(j)(1) associated with the
change in a manner that could reasonably be precalciner kiln. See docket item OAR–2004–0022– destruction and removal efficiency
expected to affect the ability to meet the DRE 0384. standand. Commenters note that
standard. See §§ 63.1206(b)(7) and 63.1207(c)(2)(ii). 223 U.S. EPA, ‘‘Final Technical Support Document
The source would ensure continued compliance by for HWC MACT Standards, Volume III: Selection of
demonstrating the minimum
operating under the operating parameter limits MACT Standards and Technologies,’’ Section 23.4, temperature requires operating under
established during this DRE test. September 2005. stressful operating conditions that can
lead to upset conditions and potentially the DRE standard to ensure that other than limiting kiln exit gas
damage the integrity of the combustion efficiency within the entire temperatures may be influencing
manufacturing equipment. Other kiln system is maintained for the control significantly dioxin/furan formation in
commenters oppose, however, deletion of nondioxin/furan organic HAP. lightweight aggregate kilns. As such, we
of the minimum combustion chamber However, we acknowledge the conclude that removing the minimum
temperature limit for cement kilns. difficulties that cement kiln operators combustion chamber temperature limit
These commenters state that all face in establishing a minimum would not be appropriate at this time
combustion sources, including cement combustion chamber temperature limit, due to the uncertain nature of dioxin/
kilns, must meet a minimum including the stressful operating furan formation in lightweight aggregate
combustion chamber temperature limit conditions necessary to establish the kilns. Thus, we are retaining the
to control dioxin/furans and organic limit. As we stated at proposal, our data requirement to establish a minimum
HAP emissions given that some cement indicate that limiting the gas combustion chamber temperature limit
kilns feed hazardous waste at locations temperature at the inlet to the for dioxin/furans under § 63.1209(k)(2)
other than the high temperature clinker- particulate matter control device is a and § 63.1209(j)(1) for lightweight
forming zone of the kiln. critical parameter in controlling dioxin/ aggregate kilns.
Response: We are deleting as furan emissions in cement kilns. See 69
proposed the requirement to establish a L. One Time Dioxin and Furan Test for
FR at 21344. Therefore, we believe that
minimum combustion chamber Sources Not Subject to a Numerical
an operating limit on the minimum
temperature limit for dioxin/furan Limit for Dioxin and Furan
combustion chamber temperature is less
under § 63.1209(k)(2) for cement kilns. important to ensure compliance with Comment. Commenters support the
See 69 FR at 21343. However, we retain the dioxin/furan standard than to ensure one-time dioxin/furan test for sources
the requirement for cement kilns to compliance with the DRE standard. not subject to a numerical dioxin and
establish and comply with a minimum Thus, we remove the requirement to furan standard. Commenters agree that
combustion chamber temperature limit establish a minimum combustion previous testing should be allowed to
for the destruction and removal chamber temperature limit for dioxin/ document the one time test.
efficiency standard under furan under § 63.1209(k)(2) for cement Response. The final rule requires
§ 63.1209(j)(1).224 kilns. This change does not affect the sources that are not subject to a standard
As discussed in the 1999 rule, other operating parameter limits under with numerical dioxin and furan
nondioxin/furan organic hazardous air § 63.1209(k) that must be established for levels 226 to conduct a one-time dioxin
pollutants are controlled by the DRE dioxin/furans, including a limit on the and furan test as part of their initial
standard and the carbon monoxide and gas temperature at the inlet to the comprehensive performance testing:
hydrocarbon standards. See 64 FR at particulate matter control device. lightweight aggregate kilns that elect to
52848–52852. This standard was not Comment: One commenter supports control the gas temperature at the kiln
reopened in the present rulemaking. We the use of previous minimum exit rather than comply with a dioxin/
note, however, that the DRE standard combustion zone temperature data, furan standard of 0.20 ng TEQ/dscm,
determines appropriate process controls regardless of the test age, in lieu of solid fuel boilers, liquid fuel boilers
necessary for the combustion of conducting new, stressful DRE testing. with wet or no air pollution control
hazardous waste. Establishing and That is, if a cement kiln is required to systems, and HCl production furnaces.
monitoring a minimum temperature of conduct future DRE tests, then the We will use these data as part of the
the combustion chamber is a principal source should not have to re-establish a process of addressing residual risk
factor in ensuring combustion efficiency minimum combustion chamber under CAA section 112(f) and
and destruction of toxic organic temperature limit during the new test. evaluating future MACT standards
compounds. As discussed in the Rather, the source should have the under section 112(d)(6). The results may
previous response, we believe this is option to submit minimum combustion also be used as part of the RCRA
especially true given the industry trend chamber temperature results in lieu of omnibus permitting process.
to convert to the more thermally Comment. EPA proposed that source
re-establishing the limit.
efficient preheater/precalciner kiln Response: We reject the commenter’s not subject to a numerical dioxin and
manufacturing process, which use two suggestion for reasons discussed above. furan limit conduct a dioxin and furan
separate combustion processes. We We believe that it is necessary to retain test under worst-case conditions.
conclude that it is necessary, in spite of Commenters state that operating under
the link between the minimum
worst-case conditions is inconsistent
the concerns raised by commenters, to combustion chamber temperature limit
with the CAA Section 112(f) process,
retain the minimum combustion and the DRE test itself, which will
which is to consider actual (i.e., normal)
chamber temperature limit as related to ensure that the combustion efficiency of
emissions. Commenters suggest that we
the entire system will be maintained for
require the tests be conducted under
224 Under the interim standards, cement kilns
the control of nondioxin/furan organic
must establish and continuously monitor limits on normal to above normal conditions.
HAP. Response. Section 112 (f) standards
minimum gas temperature in the combustion zone
Comment: One commenter supports
for both the dioxin/furan and DRE standards. As evaluate allowable emission levels,
discussed in the preceding paragraph, a source may deletion of the minimum combustion
although actual emissions levels may
not need to conduct DRE testing during each chamber temperature requirement for
comprehensive performance test. If DRE testing is
also be considered. See 70 FR at 19998–
dioxin/furan under § 63.1209(k)(2) for
required, then the source will need to establish a
minimum combustion zone temperature limit as
lightweight aggregate kilns. high of 57.9 ng TEQ/dscm. See ‘‘Source Category
required under the DRE standard. However, if DRE Response: We reject the commenter’s Summary Sheets’’ available in the docket or
testing is not required, then (according to the suggestion. Our data base of dioxin/ USEPA, ‘‘Final Technical Support Document for
changes made today) the cement kiln will not be furan emissions data shows substantial HWC MACT Standards, Volume II: HWC Data
required to establish the minimum combustion Base,’’ September 2005.
variability in test results at each
chamber temperature limit under the dioxin/furan 226 These sources do, however, need to comply
standard during a subsequent comprehensive source.225 This may indicate that factors with the carbon monoxide or hydrocarbon
performance test. The minimum combustion standards, as well as the DRE standard as surrogates
chamber temperature operating limit established 225 For example, dioxin/furan emissions from to comply with today’s dioxin and furan emissions
during previous testing remains in effect, however. source number 307 range from a low of 0.024 to a control requirements.
19999 (April 15, 2005). Although we results from controllable operating sulfur compounds are routinely fed to
agree with the commenter that, in conditions. the boler, dioxin/furan testing should
general, emissions in the range of For these reasons, the final rule not be performed after a period of firing
normal to maximum are considered for requires sources to test under feed and high sulfur fuel or injection of sulfur
section 112(f) determinations, we operating conditions that are most likely additives. See 69 FR at 21308 for more
believe that dioxin/furan testing to to reflect maximized expected daily information.
provide information of use in section variability of dioxin/furan emissions, as Comment: Commenters state that we
112(f) residual risk determinations proposed. Such testing is similar to a should delete the one-time testing
should be conducted under conditions comprehensive performance test to requirement for dioxin and furans. The
where controllable operating conditions demonstrate compliance with a Clean Air Act at Section 114(a)(1)(D)
are maximized to reflect the full range numerical dioxin/furan emission allows EPA to request ‘‘any person’’ to
of expected variability of those standard where operating limits would sample emissions. Applying the Section
parameters which can be controlled. be established based on operations 114 authority to an entire subcategory of
This is because dioxin/furan emissions during the test. As a practical matter, sources is overly broad, particularly in
may relate exponentially with the however, we note that many of the the context of having already
operating conditions that affect operating parameters discussed below, established appropriate surrogates for
formation. We believe that dioxin/furan although controllable to some extent, dioxin and furan in a MACT rule.
emissions relate exponentially with gas cannot be quantified and cannot be Commenters are not aware of EPA
temperature at the inlet to an ESP or controlled to replicate the condition in taking this approach in previous efforts.
fabric filter,227 and are concerned that a future test. In addition, some operating (Section 114 requests have focused on
emissions may also relate exponentially parameters we identify may not have as collecting existing information from
with the operating parameters strong a relationship to dioxin/furan sources facing future MACT standards).
(discussed below) that affect emissions emissions as others. Consequently, the Commenters oppose this approach
from sources subject to the one-time operating conditions are generally because it established a precedent they
dioxin/furan emissions test. Emissions described subjectively. do not favor, and will bring about
testing under operating conditions that Based on currently available research, significant costs and difficulties to
are in the range of ‘‘normal to above you should consider the following provide the data. They suggest that we
normal’’ may be exponentially lower factors to ensure that you conduct the delete the proposed requirements for a
than emissions under operating test under operating conditions that one-time dioxin and furan test.
conditions reflecting maximum daily seek to fully reflect maximum daily Response: We believe that section
variability of the source. Since testing variability of dioxin/furan emissions: (1) 114(a)(1)(D) of the Clean Air Act
under normal operating conditions Dioxin/furan testing should be provides us the authority to require
conducted at the point in the sources to conduct a one time test to
makes no effort to assess operating
maintenance cycle for a boiler when the generate data which can be used in
variability, emissions during such
boiler tubes are more fouled and soot- making later section 112 (f)
testing would fail to reflect expected
laden, and not after maintenance determinations for the source category.
daily maximum operating variability
involving soot or ash removal from the The results of the testing may also
and so would not represent time-
tubes; (2) dioxin/furan testing should be inform the section 112(d)(6) review and
weighted average emissions and would
performed following (or during) a the RCRA omnibus permitting
under-represent health risk from
period of feeding normal or greater processes. The fact that section 114
chronic exposure.
quantities of metals; (3) dioxin/furan specifically indicates that a purpose of
Although we acknowledge that
testing should be performed while gathering information under section 114
sources will not exhibit maximum feeding normal or greater quantities of
operating variability each day of is to assist in developing national rules
chlorine; (4) the flue gas temperature in indicates that the provision can have
operation, we believe that it is some portion of the heat recovery
important to assess the upper range of wide sweep extending to all sources in
section of a boiler should be within the a category. See 69 FR at 21307–308 for
emissions that these sources may emit dioxin formation temperature window
to properly evaluate under section a full explanation.
of 750 to 400°F during the testing; (5) We believe a dioxin and furan test
112(f) whether the MACT standards for the testing should not be conducted costs approximately $10,000 when
dioxin/furan for these sources (i.e., under optimal combustion conditions conducted along with other testing. We
absent a numerical emission standard) (e.g., combustion chamber temperature do not believe this cost is significant,
protect public health with an ample should be in the range of normal to the and sources must only perform this test
margin of safety.228 operating limit; hazardous waste once, not more frequently as would be
In addition, we note that emissions feedrate and combustor through put the case to ensure compliance with a
reflecting daily maximum variability should be in the range of normal to standard. We also allow sources to use
would be most useful for section maximum); (6) for units equipped with prior testing to meet this requirement,
112(d)(6) determinations in the future wet air pollution control systems, the and allow sources to use ‘‘data in lieu’’
because they would represent the full testing should be conducted after a high so they can test one source if they have
range of emissions variability that solids loading has developed in the more than one of the same identical
scrubber system (consistent with normal sources.
operating cycles); and (7) for solid fuel We do not believe that obtaining these
HWC MACT Standards, Volume IV: Compliance,’’
July 1999, Chapter 3.
boilers, the sulfur content of the coal data will be difficult, and note that the
228 Dioxin/furan are some of the most toxic should be equivalent to or lower than permitting authority can assist sources
compounds known due to their bioaccumulation normal coal sulfur levels (within the in planning their tests.
potential and wide range of health effects, including range of sulfur levels that the source
carcinogenesis, at exceedingly low doses. Exposure utilizes), and the gas temperature at the M. Miscellaneous Compliance Issues
via indirect pathways is a chief reason that
Congress singled out dioxin/furan for priority
inlet to the electrostatic precipitator or Comment: Several commenters state
MACT control in CAA section 112(c)(6). See S. Rep. fabric filter should be close to the that § 63.1206(c)(3)(iv) requiring an
No. 128, 101st Cong. 1st Sess. at 154–155. operating limit. In addition, unless automatic waste feed cutoff (AWFCO) if
a parameter linked to the AWFCO is We note, however, that the Interim under §§ 63.1206(b)(9)(i) and
exceeded should be revised to reflect Standards for total chlorine continue to 63.1206(b)(10)(i) is inconsistent with the
§ 63.1206(c)(2)(v)(A)(1). Section apply to sources that establish health- proposed preamble, which states that
63.1206(c)(2)(v)(A)(1) states that, if the based limits for total chlorine under sources should be allowed to petition
AWFCO is affected by a malfunction § 63.1215. Consequently, we have for alternative standards provided they
such that the malfunction itself prevents incorporated the total chlorine Interim submit information showing that HAP
immediate and automatic cutoff of the Standards in § 63.1215 as they apply as contributions to emissions from the raw
hazardous waste feed, you must cease a cap to the health-based emission materials are preventing the source from
feeding hazardous waste as quickly as limits. achieving the emissions standard
possible. Comment: Several commenters state though the source is using MACT
Response: We agree with commenters that the rule should allow extrapolation control.229 The commenters state that
in principle, but note that the automatic of ash and chlorine feedrates to the proposed regulatory language,
waste feed cutoff system may fail for establish feedrate limits corresponding despite the intent signaled in the
reasons other than a malfunction. That to the particulate matter and total proposed preamble, inappropriately
is, equipment or other failures are chlorine standards. Commenters believe excludes the provisions of
malfunctions only if they meet the the rationale we use to allow §§ 63.1206(b)(9)(i) and 63.1206(b)(10)(i)
definition of malfunction at § 63.2. extrapolation of metals feedrates is also as an alternative option when
Failures that result from improper applicable to ash and chlorine. complying with the replacement
maintenance or operation are not Response: The final rule does not emission standards under §§ 63.1220
malfunctions. Consequently, the final allow you to extrapolate ash and and 63.1221.
rule revises § 63.1206(c)(3)(iv) to state chlorine feedrates achieved during the Response: We agree with the
that if the AWFCO is affected by a comprehensive performance test to commenters. The proposed regulatory
failure such that the failure itself establish feedrate limits comparable to text inadvertently excluded the
prevents immediate and automatic the particulate matter and total chlorine alternative standard provisions from use
emission standards. by cement and lightweight aggregate
cutoff of the hazardous waste feed, you
We do not allow extrapolation of ash kilns under the replacement standards.
must cease feeding hazardous waste as
to the particulate matter emission Accordingly, we are revising the
quickly as possible. Revised standard because particulate matter (i.e.,
§ 63.1206(c)(3)(iv) does not refer to introductory text of §§ 63.1206(b)(9)(i)
soot) may form in the combustor, and 63.1206(b)(10)(i) by making the
malfunctions, however, because the particularly at times of unstable
AWFCO system may fail for reasons alternative standards available under
combustion conditions. Consequently, the replacement standards.
other than a malfunction. The reference extrapolating from ash feedrates may Comment: One commenter states that
in § 63.1206(c)(2)(v)(A)(1) to underestimate particulate matter the availability of the alternative
malfunctions is appropriate because that emissions and may not ensure standard for mercury under
paragraph addresses requirements compliance with the particulate matter § 63.1206(b)(10)(i) should not be
during malfunctions. emission standard. conditioned upon mercury being
Comment: Several commenters note We do not allow extrapolation of present only at levels below the
that the proposed rule did not include chlorine feedrates to the total chlorine detection limit in raw materials, as
a sunset provision for the Interim emission standard because chlorine specified under § 63.1206(b)(10)(i)(B).
Standards applicable to incinerators, feedrate is an operating parameter limit The commenter suggests that the
cement kilns, and lightweight aggregate to ensure compliance with the approach for mercury should be the
kilns after the compliance date of the semivolatile metal emission standard. same as for other HAP such as semi- and
standards we promulgate today (i.e., the Because an increase in chlorine feedrate low volatile metals under
‘‘permanent replacement standards’’). can increase the volatility of § 63.1206(b)(10)(i)(A).
Commenters are concerned that, semivolatile metals and we do not know Response: The commenter misreads
although the Agency intends for the the precise relationship among chlorine the alternative standard provisions
replacement standards to be more feedrate, metal volatility, and metals under § 63.1206(b)(10)(i). We note that
stringent than the Interim Standards, emissions, extrapolating the chlorine § 63.1206(b)(10) includes two separate
that may not be the case in all situations feedrate achieved during the provisions for cement kilns. The first
because of the different format used for comprehensive performance test to a provision allows sources to petition for
some of the replacement standards. For feedrate comparable to the total chlorine an alternative standard when a source
example, several of the replacement emission standard may not ensure cannot achieve a standard because of
standards for cement kilns and compliance with the semivolatile metal HAP metal or chlorine concentrations in
lightweight aggregate kilns are emission standard. If a source complies their raw material feedstocks cause an
expressed as hazardous waste thermal with the semivolatile metals emission exceedance of a standard despite the
emissions. standard under § 63.1207(m)(2) where source’s use of MACT control. See
Response: Although we are the performance test is waived, § 63.1206(b)(10)(i)(A). The term
promulgating the replacement standards however, by assuming zero system ‘‘regulated metals’’ specified in
in a format that ensures they are not less removal efficiency and limiting the § 63.1206(b)(10)(i)(A) includes mercury,
stringent than the Interim Standards, we semivolatile feedrate (expressed as a semivolatile metals, and low volatile
agree with commenters that not maximum theoretical emission metals. The second provision allows a
sunsetting the Interim Standards may concentration) to the level of the source to petition for an alternative
lead to confusion as to which standards emission standard, the source may mercury standard when mercury is not
apply. Consequently, we include a request under § 63.1209(g)(1) to present at detectable levels in the
sunset provision in today’s rule for the extrapolate chlorine feedrates during the source’s raw materials.
Interim Standards. The Interim comprehensive performance test up to § 63.1206(b)(10)(i)(B). These two
Standards will be superseded by the the total chlorine emission standard. provisions are indeed separate as
final rule promulgated today on the Comment: Several commenters state
compliance date. that the proposed regulatory language 229 For example, see 69 FR at 21268.
discussed in the 1999 rule. See 64 FR at objecting to the permitting authority’s Mandate? Although we have concluded
52962–967. Also note that the conclusions regarding the need for a risk that the Phase 1 Replacement and Phase
conjunction separating paragraphs assessment, the permitting authority 2 standards are generally protective, as
(b)(10)(i)(A) and (b)(10)(i)(B) is ‘‘or,’’ not must respond fully to the comments. we discussed in the 2004 proposal (69
‘‘and.’’ Any permit conditions determined to be FR 21325), there may be instances
Given the potential confusion of the necessary based either on the SSRA, or where we cannot assure that emissions
term ‘‘regulated metals,’’ we are because the facility declined to conduct from each source will be protective of
clarifying the regulatory text by an SSRA, also must be documented and human health and the environment, and
specifying the three metal HAP supported in the administrative record. therefore an SSRA may be necessary.
volatility groups that comprise the term Today, we are codifying additional Furthermore, it should be noted that,
‘‘regulated metals.’’ See revised regulatory language providing authority just as for the risk assessment for the
§ 63.1206(b)(10)(i)(A). Finally, given for SSRAs while maintaining the same 1999 rule, the comparative analysis does
that the alternative standard provisions basic SSRA policy. It is important to not account for cumulative emissions at
are similar for lightweight aggregate note that all of the requirements of Part a source or background exposures from
kilns, we are also clarifying 124 referred to above will continue to other sources.
§§ 63.1206(b)(9)(i)(A) and (b)(9)(iv). apply to actions taken in accordance Before discussing factors that may
with the additional regulatory language lead permit authorities to consider
IX. Site-Specific Risk Assessment we are codifying. The SSRA regulatory
Under RCRA whether or not to conduct an SSRA, it
provisions, which establish that the should be noted that the Agency
A. What Is the Site-Specific Risk need for an SSRA should be determined
generally does not expect that facilities
Assessment Policy? on a case-by-case basis, apply equally to
that have conducted risk assessments
both Phase 1 and Phase 2 sources.
The Site-Specific Risk Assessment will have to repeat them. As we
(SSRA) Policy has undergone several B. Why Might SSRAs Continue To Be explained in the 1999 final rule
revisions since its inception in the 1993 Necessary for Sources Complying With preamble, changes to comply with the
draft Combustion Strategy. Currently, it Phase 1 Replacement Standards and MACT standards should not cause an
is the same policy as we expressed in Phase 2 Standards? increase in risk for the vast majority of
the 1999 final rule preamble. In the EPA conducted a national evaluation facilities given that the changes will
1999 rule, we recommended that for of human health and ecological risk for likely be the addition of pollution
hazardous waste combustors subject to the MACT standards as proposed in the control equipment or a reduction in the
the Phase 1 MACT standards, permitting 1996 NPRM and then revised the hazardous waste being burned (see 64
authorities should evaluate the need for evaluation to include more facilities for FR 52842). Instances where a facility
an SSRA on a case-by-case basis. the 1999 final rulemaking. Based on the may need to repeat a risk assessment
Further, while SSRAs are not results of the final national risk would be related to changes in
anticipated to be necessary for every evaluation for hazardous air pollutants conditions that would likely lead to
facility, they should be conducted (excluding non-dioxin products of increased risk. For example, if the only
where there is some reason to believe incomplete combustion), we concluded changes at a facility relate to the
that operation in accordance with the that sources complying with the MACT exposed population (a new housing
MACT standards alone may not be standards generally would not pose an development is constructed within a
protective of human health and the unacceptable risk to human health or few square miles of the source), what
environment. For hazardous waste the environment. For today’s final rule, was once determined to be protective
combustors not subject to the Phase 1 we did not conduct another national under a previous risk assessment may
standards, we continued to recommend risk assessment as we did for the 1999 now be beyond acceptable levels.
that SSRAs be conducted as part of the rule. Rather, for both the April 20, 2004 Another example would be where a
RCRA permitting process. See 64 FR NPRM and today’s final rule we hazardous waste burning cement kiln
52841. Since 1999, we have provided conducted a comparative risk analysis, that previously monitored hydrocarbons
additional clarification of the comparing the Phase 1 Replacement and in the main stack elects to install a mid-
appropriate use of the SSRA policy and Phase 2 Standards to the 1999- kiln sampling port for carbon monoxide
technical guidance in an April 10, 2003 promulgated Phase 1 Standards, to or hydrocarbon monitoring to avoid
memorandum from OSWER’s Assistant determine if there were any significant restrictions on hydrocarbon levels in the
Administrator to the EPA Regional differences that might influence or main stack. Thus, the stack hydrocarbon
Administrators entitled, ‘‘Use of the impact the potential risk. Similar to the emissions may increase (64 FR 52843,
Site-Specific Risk Assessment Policy proposal, the comparative analysis footnote 29). In such situations, we
and Guidance for Hazardous Waste conducted for today’s final rule focused would anticipate that the risk
Combustion Facilities’’ (see Docket # on several key characteristics: emission assessment would not have to be
OAR–2004–0022–0083). Most rates, stack height, stack gas buoyancy, entirely redone. It may be as limited as
importantly, in this memorandum we meteorological conditions (which collecting relevant new data for
reiterated that where a permitting include a number of variables), comparison purposes, leading to a
authority concludes that a risk population parameters including decision not to repeat any portion of a
assessment is necessary for a particular density and radial distribution, and risk assessment. Or, it may be more
combustor, the basis for this decision correlations among the characteristics inclusive such that modifications would
must be substantiated in each case. The themselves. The results of the be made to specific inputs to or aspects
factual and technical basis for any comparative analysis suggest that the of the risk assessment using data from
decisions to conduct a risk assessment MACT standards for both Phase 1 and a previous risk assessment, risk burn or
must be included in the administrative Phase 2 sources are generally protective. comprehensive performance test. In
record for the facility per 40 CFR 124.7, Therefore, separate national emissions recognition of this, we have added an
124.8, 124.9, and 124.18. In addition, if standards under RCRA are unnecessary. additional factor to the list of factors at
the facility, or any other party, files See Part Seven: How Does the Final § 270.10(l)(1) to indicate that a
comments on a draft permit decision Rule Meet the RCRA Protectiveness previously conducted risk assessment
would be relevant in evaluating changes standard (carbon monoxide or total combustion, remain today and may
in conditions that may lead to increased hydrocarbon as surrogate controls, influence a permitting authority’s
risk. The factor reads as follows: versus a numerical standard) for HCl decision. Last, we are finalizing the
‘‘Adequacy of any previously conducted production furnaces; and the dioxin/ option for Phase 2 area sources to
risk assessment, given any subsequent furan standard for liquid fuel-fired comply with specific MACT standards
changes in conditions likely to affect boilers (LFBs) with dry APCDs. In as provided by CAA § 112(c)(6) specific
risk.’’ The following discussion is addition, dioxin/furan emissions data pollutants authority. These area sources
intended mainly to address facilities for LFBs with wet or no APCDs indicate may need to conduct an SSRA for the
that have not yet conducted an SSRA an observed level (1.4 ng TEQ/dscm) of remaining RCRA standards that they
(i.e., where it has been determined that more than three times the highest choose to comply with (i.e., since they
one is needed). dioxin/furan standard evaluated in the do not address the potential risk from
In the proposal we discussed our 1999 national risk assessment (69 FR indirect exposures to long-term
conclusion that almost all of the 21285).230 Thus, these standards may deposition of metals onto soils and
proposed standards for Phase 1 sources warrant site-specific risk consideration, surface waters).233
were equivalent to or more stringent especially with respect to the dioxin/ In addition to the examples provided
than the 1999 final standards, with the furan standards. That is, due to the in the previous paragraph, we also
exception of the mercury standard for complexity of the dioxin/furan expressed that an SSRA may be
new and existing LWAKs and the total formation mechanism and given the necessary with respect to the proposed
chlorine standard for new LWAKs. toxicity of dioxin/furans,231 an SSRA thermal emission standards. With
However, there are additional standards may be needed based on the specific respect to Phase 1 sources, we had
for Phase 1 sources finalized in today’s emission levels of each source not noted in the proposal that the thermal
rulemaking that are less stringent than subject to a numerical standard. For emission standards for semi-volatile and
the 1999 final standards. In addition to additional discussion on the low volatile metals for cement kilns and
those discussed in the proposal, the protectiveness of standards, please refer LWAKs may be of concern because they
following standards are less stringent to Part Seven: How Does the Final Rule directly address emissions attributable
than the 1999 final standards: mercury Meet the RCRA Protectiveness Mandate? to hazardous waste versus a source’s
for new cement kilns and semi-volatile There are also site-specific factors total HAP metal emissions. See 69 FR
metals for existing cement kilns; dioxin/ beyond the standards that can be 21326. However, we are requiring
furan for existing and new LWAKs, important to the SSRA decision making sources to comply with both the thermal
mercury for existing and new LWAKs, process. As discussed in the proposal, emission standards and the Interim
and total chlorine for existing and new examples include a source’s proximity Standards in today’s final rulemaking,
LWAKs. Because these standards exceed to a water body or endangered species since compliance with the thermal
the levels which were evaluated in the habitat, repeated occurrences of emission standards may not always
1999 national risk assessment, contaminant advisories for nearby water assure compliance with the Interim
especially with respect to mercury and bodies, the number of hazardous air Standards. As a result, the thermal
dioxin/furan standards for which the pollutant emission sources within a emission standards for cement kilns and
national risk assessment showed high facility and the surrounding LWAKs no longer pose the uncertainties
end risks at or near levels of concern, community, whether or not the waste that they had in the proposal.234 In
permit authorities may decide on a case- feed to the combustor is made up of regard to Phase 2 sources, the concern
by-case basis that an SSRA is persistent, bioaccumulative or toxic at the time of proposal was with respect
appropriate to determine whether the contaminants, and sensitive receptors to the thermal emission standards for
less stringent Replacement standards are with potentially significantly different liquid fuel-fired boilers. However, the
protective. In addition, the comparative exposure pathways, such as Native comparative analysis for today’s final
analysis results suggest concern Americans (69 FR 21326). Also, there rulemaking for liquid fuel-fired boilers,
regarding the dioxin/furan standard for are several uncertainties inherent in the which is based on total stack emissions
LWAKs and thus, permit authorities 1999 national risk assessment.232 Thus, from these sources while assuming
may consider site-specific factors in the same uncertainties related to the fate compliance with the thermal standards,
determining whether the standard is and transport of mercury in the does not suggest that risks for LFBs are
sufficiently protective. environment and the biological cause for concern (except as otherwise
Specific to Phase 2 sources, we significance of mercury exposures in
mentioned earlier that we conducted the noted, e.g., dioxins).
fish (i.e., once mercury has been
same comparative risk analysis for transformed into methylmercury, it can C. What Changes Are EPA Finalizing
Phase 2 sources as we did for Phase 1 be ingested by the lower trophic level With Respect to the Site-Specific Risk
sources (i.e., by comparing the Phase 2 organisms where it can bioaccumulate Assessment Policy?
standards to the 1999 final standards for in fish tissue), as well as the risk posed In the 1999 final rule preamble, we
Phase 1 sources). Although several by non-dioxin products of incomplete included a revised site-specific risk
MACT standards for Phase 2 sources are assessment (SSRA) policy
more stringent than the BIF standards 230 The comparative analysis did not specifically
recommendation to account for
under RCRA, there are a few MACT suggest concern as it has for other source categories,
promulgation of the new technology-
standards that may be cause for concern but per the reference to the proposal, we have some
concern regarding the protectiveness of the based CAA MACT standards for Phase
on a case-by-case basis, as they are
standard.
either less stringent than some of the 231 There is ongoing uncertainty in cancer and 233 Currently, there are only five area sources that
1999 final standards or the comparative other health effects levels for chlorinated dioxins this may apply to; they are interim status units in
risk analysis suggests concern. They are: and furans. the process of conducting an SSRA as part of their
The particulate matter standard (and 232 Uncertainties stem from a lack of information final permits.
certain metals such as antimony and regarding the behavior of mercury in the 234 An exception would be the semivolatile metal
environment and a lack of sufficient emissions data Interim standard for existing cement kilns, which
thallium), mercury standard, and total and parameter values (e.g., bioaccumulation values) is less stringent than the 1999 final standard. As we
chlorine standard for solid fuel-fired for nondioxin products of incomplete combustion. noted, permit authorities may consider the need for
boilers (SFBs); the dioxin/furan See 64 FR 52840–52841. an SSRA as a result.
1 sources. We recommended that proposed codification. The comment some facilities have RCRA risk-based
permitting authorities evaluate the need most frequently presented was that the permit conditions that establish more
for an SSRA on a case-by-case basis for proposed regulatory language is not frequent sampling or limits on feed rate
hazardous waste combustors subject to helpful to anyone (i.e., regulated for specified metals to ensure that
the Phase 1 MACT standards. For community, the public or permitting ecologically sensitive areas are not
hazardous waste combustors not subject agencies), is redundant with the adversely impacted.
to the Phase 1 standards, we continued omnibus authority, and sets an Many commenters also state that CAA
to recommend that SSRAs be conducted extremely low hurdle for regulators to § 112(f) residual risk process is the
as part of the RCRA permitting process require SSRAs. appropriate method to assess risk for
if necessary to protect human health We disagree that the new regulatory hazardous waste combustors complying
and the environment. We indicated that language is not helpful and that it sets with MACT, not RCRA risk assessments.
the RCRA omnibus provision authorized an extremely low hurdle for regulators Specifically, one commenter argued that
permit authorities to require applicants to require SSRAs. We believe that the EPA lacked statutory authority to rely
to submit SSRA results where an SSRA new provisions are beneficial in two on the omnibus provisions to require
was determined to be necessary. For the ways: (1) They provide notice to the SSRA and SSRA-based controls on the
reasons described in the previous regulated community and public that an grounds that § 112(f) of the Clean Air
subsection, we believe that additional SSRA may be necessary to support a Act establishes a specific provision to
controls may be necessary on a site- source’s permit; and (2) they remind the control any residual risk from
specific basis to ensure that adequate permitting agencies of the importance of combustor emissions. We disagree with
protection is achieved in accordance evaluating whether an SSRA would be commenters for two reasons. First, as we
with RCRA. necessary on a site-specific basis. The explained in the 1999 final rule
Consequently, because SSRAs are new regulatory provision in no way preamble, the omnibus provision is a
likely to continue to be necessary at expands or supplements the authority RCRA statutory requirement and the
some facilities (mainly those that have on which EPA had previously relied— CAA does not override RCRA.
not previously conducted an SSRA), we i.e., omnibus and § 270.10(k), thus it Promulgation of the MACT standards,
concluded that it is more appropriate to does not provide any more or less therefore, does not duplicate, supersede,
include a regulatory provision that authority to permit authorities (i.e., or otherwise modify the omnibus
explicitly provides for the permit lower or raise the hurdle) to require provision or its applicability to the
authority to require SSRAs on a case-by- SSRAs. We agree that, because the sources covered by today’s rule. Second,
case basis and add conditions to RCRA proposed language provides permitting
the SSRA under RCRA is usually
permits based on SSRA results. authorities with no greater authority
conducted prior to issuance of the final
Therefore, instead of relying on RCRA than the omnibus authority, it is
permit. The CAA residual risk
§ 3005(c)(3) and its associated somewhat duplicative of § 270.10(k).
determination is generally made eight
regulations at § 270.10(k) when However, as noted, EPA believes this
years after promulgation of the MACT
permitting authorities conduct or provision offers important benefits to
standards for a source category.
require a risk assessment on a site- both the agency and the regulated
Accordingly, a permit authority
specific basis (i.e., as applicable to those community, and as explained further
currently facing a permit decision could
newly entering the RCRA permit below, EPA has adopted a slightly
not rely on these yet unwritten residual
process), we had proposed to codify the modified version of the proposal
authorities provided by sections 3004(a) pursuant to RCRA § 3004(a) and risk standards to resolve its identified
and (q) and 3005(b). See proposed § 3005(b). See also discussion in concern that the MACT standard may
regulations at 69 FR 21383–21384, subsection F. not be sufficiently protective at an
§§ 270.10(l) and 270.32(b)(3). In Another common view expressed by individual site. In addition, even though
proposing to codify these authorities, commenters is that, although extensive we believe that § 3005(c)(3) and its
we stated that we were not requiring risk assessments that have been associated regulations provide the
that SSRAs automatically be conducted performed for more than a decade, authority to require and perform SSRAs
for hazardous waste combustion units, showing lack of risk to human health and to write permit conditions based on
but that the decision of whether or not and the environment, EPA continues to SSRA results, we are not relying on
a risk assessment is necessary must be require SSRAs without a technical these provisions as the authority for
made based upon relevant factors evaluation of the historical results. To § 270.10(l). Rather, we are relying on
associated with an individual the contrary, EPA Regional permit §§ 3004(a) and (q) and 3005(b). See 69
combustion unit and that there are writers have found that certain FR 21327.
combustion units for which an SSRA chemicals (especially dioxin and With respect to the costs incurred
will not be necessary. Further, we mercury)235 pose excess risk in certain when conducting an SSRA, several
explained that the proposed language circumstances—even under the Interim commenters raised the concern that our
would provide notice to the regulated Standards—and consequently find it approximations do not include portions
community that an SSRA may be necessary to assess risk to human health of actual costs (e.g., data gathering, QA/
necessary to support a source’s permit, and the environment based on site- QC, and third party consultants, risk
while reminding the permit agency of specific conditions at the facility. In assessors, and plant personnel time to
the need to evaluate whether an SSRA EPA Regions 7 and 10 for example, coordinate and review SSRA efforts and
would be necessary on a site-specific collect facility data), thus resulting in
basis. 235 Dioxin is a common risk driver due to ongoing artificially low costs. Commenters cited
Despite our efforts to explain that by uncertainty in cancer and other health effects levels additional reasons why they feel that
for chlorinated dioxins and furans. Mercury is also
codifying these provisions, we are only a common risk driver due to uncertainties implicit EPA’s cost estimates are too low
modifying the statutory authority under in the quantitative mercury analysis. See discussion including our assumptions that: (1)
which we implement the SSRA policy in Part Seven, Section II. and 65 FR 52997. Thus, SSRAs are a one-time or infrequent cost;
while maintaining the same SSRA it is not uncommon for permit authorities to require (2) most SSRAs fall under ‘‘normal’’
risk-based RCRA permit limits (based on risk
policy from a substantive standpoint, assessment results) to control emissions of these versus ‘‘unusual’’ situations; and (3) the
commenters generally opposed EPA’s pollutants. cost of conducting a risk burn during a
trial burn adds only 20% more to the have revised our cost analysis to reflect appears as though commenters believe
cost. inclusion of these higher-cost facilities. that the procedures (and procedural
Regarding the comment that we did See background document, Assessment protections) currently applicable
not include actual costs for our of the Potential Costs, Benefits, and whenever an SSRA is conducted are
estimates of overall costs to conduct an Other Impacts of the Hazardous Waste unique to circumstances in which the
SSRA, we agree that some costs were Combustion MACT Replacement permitting authority proceeds under the
overlooked. We did include the costs Standards—Final Rule, October 12, authority of RCRA § 3005(c)(3)—the
related to conducting an SSRA under 2005. ‘‘omnibus’’ provision. This is incorrect.
‘‘normal’’ and ‘‘unusual’’ conditions, Also, we maintain our assumption All of the specific procedural
SSRA data collection in conjunction that SSRAs generally represent a one- requirements the commenters have
with a regular performance burn, and a time cost unless a facility significantly raised would be applicable whether the
full independent risk burn including changes its operations or if receptors permitting authority proceeded under
protocol, sampling, analysis, and report. change such that an increase in risk is § 270.10(l), as EPA proposed, or
However, we did not capture facility anticipated as a result. Even so, as pursuant to RCRA § 3005(c)(3) and
time associated with data collection and explained earlier in subsection B., we § 270.10(k), as is the current practice.
management related to the SSRA. would anticipate that the risk All of the requirements established in
Consequently, we have revised our cost assessment would not have to be Part 124 continue to apply, whether
estimate for performing these activities; entirely redone. It may be as limited as EPA proceeds under § 270.10(l) or under
see chapter 4 of the background collecting relevant new data for § 270.10(k). As we discussed in the
document entitled, Assessment of the comparison purposes, leading to a proposal, the basis for the decision to
Potential Costs, Benefits, and Other decision not to repeat any portion of a conduct a risk assessment, or to request
Impacts of the Hazardous Waste risk assessment. Or, it may be more additional information to evaluate risk
Combustion MACT Replacement inclusive such that modifications would or determine whether a risk assessment
Standards—Final Rule, October 12, be made to specific inputs to or aspects is necessary, must be included in the
2005. of the risk assessment using data from administrative record for the facility and
In response to the broader comment a previous risk assessment, risk burn or made available to the public during the
that our cost estimates are too low (for comprehensive performance test. With comment period for the draft permit.
several reasons mentioned previously), respect to chemical weapons See 40 CFR 124.7 [statement of basis];
we agree that our estimate of a 20% demilitarization facilities, we recognize 124.9 [administrative record for draft
additional cost to conduct a risk burn that due to their specialized waste permit]; 124.18 [administrative record
with a trial burn may have been streams and multiple treatment units, for final permit]. If the facility, or any
conservative and therefore, we have SSRAs, in many cases, are not one-time other party, files comments on a draft
adjusted our previous estimate to events and as a result, their SSRA costs permit decision objecting to the
include a range of 20% to 40%. The are relatively high. The high costs can permitting authority’s conclusions
total SSRA cost range has also been be attributed to the necessity for each regarding the need for a risk assessment,
updated from $141K–$370K to $157K– chemical weapons demilitarization the permitting authority must respond
$815K.236 With respect to our facility to perform surrogate trial burns fully to the comments. Any permit
assumption that the majority of SSRAs and then agent trial burns for each conditions determined to be necessary
are conducted under ‘‘normal’’ furnace and each agent campaign (e.g., based either on the SSRA, or because
conditions (lending to overall lower cost GB (Sarin), VX, and HD (Sulfur the facility declined to conduct an
estimates), we do believe that the Mustard)). For example, a chemical SSRA, also must be documented and
majority of future SSRAs will fall under weapons demilitarization facility would supported in the administrative record.
the ‘‘normal’’ conditions.237 We believe conduct GB trial burns on all the The commenters’ concern that
this is appropriate due to: lack of new furnaces and then complete destruction § 270.10(l) allows the permitting
facilities coming on-line for which there of the GB stockpile, followed by VX trial authority to require the SSRA prior to
is no previous test data; availability of burns and VX stockpile and finally, the the issuance of a draft permit, and
commercial modeling software; and HD trial burns and the HD stockpile. therefore the applicant would have no
finalization of the ‘‘Human Health Risk This effectively extends the input to the opportunity to comment or challenge
Assessment Protocol for Hazardous risk assessment of the trial burn data that determination, is equally
Waste Combustion Facilities’’ guidance, over most of the operational life of the unfounded. There is effectively no
or ‘‘HHRAP’’ guidance. However, we do facility. practical or substantive distinction
recognize that some facilities can be Last, several commenters raised the between the circumstance when a
more complex than others in the concern that EPA’s proposal to codify permit authority communicates the
hazardous waste combustion universe. the authority to require SSRAs on a decision that an SSRA is necessary to
Therefore, we have identified a portion case-by-case basis and add conditions to issue the permit prior to issuing the
of facilities that are likely to incur RCRA permits based on SSRA results, draft permit, or as part of the draft
‘‘unusual’’ costs for a future SSRA and violates the due process protections permit. In either case, if a facility
afforded under the current structure, refuses to provide a risk assessment or
236 The high end of this range applies only to where SSRAs are required and data to support a risk assessment
those systems operating under ‘‘unusual performed pursuant to RCRA requested under this provision, the
conditions’’ (the available data suggest that there § 3005(c)(3) omnibus authority. regulations at part 124 make clear that
are only five such facilities).
237 Normal conditions assume use of previously
Commenters were further concerned the appropriate recourse is for the
collected performance burn data, use of standard
that the proposed language in § 270.10(l) permit authority to deny the permit (See
commercial modeling software that meet Agency would remove existing procedural 40 CFR 124.3(d); 124.6(b) and 270.10(c).
guidance, and limited interactions with State and safeguards by allowing the Agency to The basis for the denial would
Federal oversight authorities. Unusual conditions require a very expensive SSRA before essentially be the same in either case—
assume the need for site-specific modeling,
extensive interactions with stakeholders and
the draft permit is even issued, thus that the information before the agency
regulators, an extended time frame, and targeted violating EPA’s own procedural gives rise to a concern that the MACT
ecological analyses. standards as well as due process. It may not be sufficiently protective,
which the agency is unable to dispel are promulgated under non-HSWA However, EPA has adopted some
based on the information before it. authority. Consequently, when it is further clarifications to the final
Consequently, the permit authority determined that an SSRA is needed, the provisions in response to comments. In
cannot determine that the permit meets applicability of these provisions will response to comments that the
RCRA’s standard for permit issuance. vary according to the type of regulatory language EPA had proposed
An as noted above, all of the combustion unit (whether it is regulated still fails to provide the regulated
requirements of Part 124 would apply to under 3004(q), or only 3004(a) and community with adequate notice that an
actions taken in accordance with 3005(b)), and the authorization status of SSRA might be required, and what that
§ 270.10(l). For additional discussion on the state. Depending on the facts, the might entail, EPA has included
this issue, please refer to the Response new authority would be applicable, or additional language to address those
to Comments background document for the omnibus provision would remain issues. Specifically, EPA has included a
this final rule.238 the principal authority for requiring sentence stating that the information
Despite the many reasons offered by SSRAs and imposing risk-based required under § 270.10(l) can include
commenters opposing our proposal, we conditions where appropriate. See 69 the information necessary to evaluate
continue to believe that our proposed FR 21327. the potential risk to human health and/
approach is appropriate. As discussed According to the state authorization or the environment resulting from both
in the proposal (69 FR 21327) and in the section of this preamble (see Part Five, direct and indirect exposure pathways.
previous subsection, although the Phase Section IV.), EPA does not consider EPA has also added language to remind
1 Replacement and Phase 2 standards these provisions to be either more or permit authorities that the
provide a high level of protection (i.e., less stringent than the pre-existing determination that the MACT standards
they are generally protective) to human federal program, since they simply make may not be sufficiently protective is to
health and the environment, thereby explicit an authority that has been and be based only on factors relevant to the
allowing us to nationally defer the remains available under the omnibus potential risk from the hazardous waste
RCRA emission requirements to MACT, authority and its implementing combustion unit at the site, and has
additional controls may be necessary on regulations. Thus, states with provided a list of factors to guide the
an individual source basis to ensure that authorized equivalents to the federal permit authority in making that
adequate protection is achieved in omnibus authority will not be required determination. See subsections E. and F.
accordance with RCRA. Until today, we to adopt these provisions, so long as for further discussion. The applicability
have relied exclusively upon RCRA they interpret their omnibus authority language of §§ 270.19, 270.22, 270.62,
§ 3005(c)(3) and its associated broadly enough to require risk and 270.66 also has been amended to
regulations at § 270.10(k) when assessments where necessary.239 allow a permit authority that has
conducting or requiring an SSRA. We determined that an SSRA is necessary to
The provisions of §§ 270.10(l) and
continue to believe that § 3005(c)(3) and continue to apply the relevant
270.32(b)(3) adopted in today’s rule are
its associated regulations provide the requirements of these sections on a case-
substantially similar to the provisions
authority to require and perform SSRAs by-case basis and as they relate to the
EPA proposed. Section 270.10(l)
and to write permit conditions based on performance of the SSRA after the
continues to explicitly provide that a
SSRA results. In fact, as the next source has demonstrated compliance
permit authority has the authority to
subsection will explain, EPA will likely with the MACT standards.
evaluate, on a case-by-case basis, the
continue to include permit conditions As previously noted, the requirements
need for an SSRA. EPA has also retained
based on the omnibus authority in some at 40 CFR Part 124 continue to apply to
its proposed language that explicitly
circumstances when conducting these actions taken to implement § 270.10(l).
provides that, where an SSRA is
activities, and state agencies in states Thus, if the permitting authority
determined to be necessary, the permit
with authorized programs will continue concludes that a risk assessment or
authority may require a permittee or an
to rely on their own authorized additional information is necessary for a
applicant to conduct an SSRA, or to
equivalent. However, because SSRAs particular combustor, the permitting
provide the regulatory agency with the
are likely to continue to be necessary at authority must provide the factual and
information necessary to conduct an
some facilities, we are finalizing the technical basis for its decision in the
SSRA on behalf of the permittee/
authority to require them on a case-by- permit’s administrative record and must
applicant. The final provision also
case basis and add conditions to RCRA make it available to the public during
essentially retains the standard laid out
permits based on SSRA results under the comment period for the draft permit.
in the proposal: that a permit authority
the authority of RCRA §§ 3004(a) and (q) If the facility or any other party files
may decide that an SSRA is warranted
and 3005(c). Therefore, we are finalizing comments on a draft permit decision
based on a conclusion that additional
§§ 270.10(l) and 270.32(b)(3) with some objecting to the permitting authority’s
controls beyond those required pursuant
minor modifications to provide further conclusions regarding the need for an
to 40 CFR parts 63, 264, 265, or 266 may
clarification of the Agency’s intent. SSRA, the authority must respond fully
be needed to ensure protection of
D. How Will the New SSRA Regulatory human health and the environment to the comments. In addition, the SSRA
Provisions Work? under RCRA. In § 270.32(b)(3), EPA has must be included in the administrative
also explicitly codified the authority for record and made available to the public
The new regulatory provisions are during the comment period. Any
finalized under both base program permit authorities to require that the
applicant provide information, if additional conditions and limitations
authority (§ 3004(a) and § 3005(b)) and determined to be necessary as a result
HSWA authority (§ 3004(q)). That is, needed, to make the decision of whether
an SSRA should be required. of the SSRA must be documented and
changes made to regulations applicable supported in the administrative record
to boilers are promulgated under HSWA as well.240
239 Authorized states are required to modify their
authority, whereas changes made to
programs only when EPA enacts federal
regulations applicable to incinerators requirements that are more stringent or broader in 240 Additional clarification on the appropriate use
scope than existing federal requirements. This of the SSRA policy and technical guidance is
238 See final Response to Comment to the HWC applies to regulations promulgated under both provided in the April 10, 2003 memorandum from
MACT Standards, Volume 5, Miscellaneous. HSWA and non-HSWA authorities. Marianne Lamont Horinko entitled ‘‘Use of the Site-
E. What Were Commenters’ Reactions to preambles. See 61 FR 17372, 64 FR at 64 FR 52842 (September 30, 1999) did
EPA’s Proposed Decision Not To 52842, and 69 FR 21328. However, EPA not adequately focus on the central
Provide National Criteria for has also incorporated a few minor question of whether there are likely to
Determining When an SSRA Is or Is Not revisions to reflect the standards be emissions that would be
Necessary? promulgated today, and to reflect the uncontrolled under the Subpart EEE
In the proposal, we stated that we fact that the factors will be codified. final rule. They argued that, as an
were not proposing national criteria EPA has revised the language of the example, under guiding factor #5, if the
(e.g., guiding factors) for determining factors so that the language is consistent waste containing highly toxic
when an SSRA is necessary. Although between the provisions. Consistency of constituents are being addressed by the
we had developed a list of qualitative phrasing is generally more important in Subpart EEE standards, the fact that
guiding factors for permit authorities to regulations, which are binding, than in there might be such wastes should not
consult when considering the need for guidance. For example, some of the justify an SSRA. The commenter
an SSRA in the September 1999 final factors listed in the 1999 preamble used apparently misunderstands that the
rulemaking (revised from the April 1996 the phrase ‘‘presence or absence’’ while factors were not intended to function as
NPRM), we never intended for them to other used the phrase ‘‘identities and stand-alone criteria for requiring an
comprise an exclusive list for several quantities.’’ EPA has adopted the phrase SSRA—i.e., to use their example, the
reasons. Mainly, we felt that the ‘‘identities and quantities,’’ on the commenter believes that the mere fact
complexity of multi-pathway risk grounds that it more precisely expresses that highly toxic constituents are
assessments precluded the conversion the concept intended by both phrases. present in the waste would justify an
of the qualitative guiding factors into EPA has also made minor revisions to SSRA without consideration of whether
more definitive criteria. See 69 FR reduce redundant text, and to shorten the MACT emission standards were
21328. the provisions, in the interests of clarity. sufficiently protective. This is an
Commenters generally agreed that the For example, rather than addressing the incorrect reading of EPA’s proposed
risk assessment guidance and policy proximity of receptors in two factors, regulation. Rather, the factors were
should not be codified. They agreed in EPA addresses this issue in a single always intended to function as
principle that it is important to keep the factor. However, nothing contained in considerations that might be relevant to
decision to require an SSRA flexible either of the original factors was deleted the determination of whether the MACT
because factors vary from facility to as part of this revision. None of the was sufficiently protective. However,
facility. However, several commenters revisions described here substantively the regulatory structure EPA has
raised the concern that the proposed change the issues to be considered from adopted in the final rule makes perfectly
language of § 270.10 (l) was too vague. those contained in the original eight clear that the critical determination is
For example, one commenter suggested guiding factors. that ‘‘compliance with the standards of
that any additional guidance clarifying In addition to these minor technical
40 CFR part 63, Subpart EEE alone may
how risk assessments should be revisions, EPA has included language to
not be protective of human health or the
performed and that providing standards clarify that one potentially relevant
environment.’’ Further, the provision
or goals to be achieved by the operating factor for consideration is the ‘‘identities
states that this determination is to be
conditions would be helpful. Another and quantities of persistent,
based only on factors relevant to the
commenter felt that EPA should identify bioaccumulative or toxic pollutants
potential risk from the hazardous waste
specific factors that the regions and considering enforceable controls in
combustion unit, including, as
authorized states should consider, and place to limit those pollutants.’’ This
appropriate, the listed factors. EPA
specific criteria that should be met, reflects changes made between the
believes that these provisions make
before requiring an SSRA or additional proposed and final MACT standards
(e.g., the proposed rule called for clear that the determination of whether
emission controls or other standards. to require an SSRA is to be based on
We agree with commenters that beyond-the-floor dioxin limits for some
sources; those were not promulgated in consideration of the conditions at the
additional guidance would be beneficial facility site, including, for example, an
and have taken a number of actions in the final rule).
Another change is the EPA has evaluation of all enforceable controls in
this regard. First, EPA is adopting a place to limit emissions. Further
more detailed regulatory provision that deleted the factor that listed ‘‘concerns
raised by the public.’’ The regulation discussion of EPA’s revised provisions
provides a non-exclusive list of guiding can be found in subsection F.
factors for permit authorities to use in will allow the decision to be based on
any one of the listed factors, and public Second, as discussed in more detail
determining whether the MACT will be below, EPA is issuing a revised risk
sufficiently protective at an individual concern, unaccompanied by an
identifiable risk, would not provide an assessment guidance document that we
site, and consequently, whether an
adequate basis for determining that an believe will provide additional insight
SSRA is warranted. Section 270.10(l)
SSRA was warranted. to help users. While clearly delineating
now requires that the permit writer’s
Finally, as discussed previously in between risk management and risk
evaluation of whether compliance with
subsection B., EPA has added an assessment, the HHRAP explains in
the standards of 40 CFR part 63, Subpart
additional factor to indicate that a great detail a recommended process for
EEE alone is protective of human health
previously conducted risk assessment performing and reporting on cost-
or the environment be based on factors
would be relevant in evaluating changes effective, scientifically defensible risk
relevant to the potential risk from a
in conditions that may lead to increased assessments. It includes numerous
hazardous waste combustion unit,
risk. The factor reads as follows: recommended defaults, while at the
including, as appropriate, any of the
‘‘Adequacy of any previously conducted same time is flexible enough to
specifically enumerated factors. These
risk assessment, given any subsequent incorporate site-specific values.
factors reflect the eight guiding factors
changes in conditions likely to affect Although the HHRAP provides
that EPA has discussed in several rule
risk.’’ See § 270.10(l)(1). numerous recommendations, it remains
Specific Risk Assessment Policy and Guidance for
One commenter raised the concern merely guidance and consequently
Hazardous Waste Combustion Facilities.’’ (See that the eight guiding factors the Agency leaves the final decisions up to the
Docket # OAR–2004–0022–0083). specified in its Federal Register notice permitting authority. We believe that
the revised HHRAP guidance will facilities that are in full compliance today than the 1999 standards) in
provide further assistance to permit with the MACT EEE standards. CKRC subsection B., present a reasonable basis
writers, risk assessors and facilities in also states that ‘‘[our] Petition for permitting authorities to consider
determining whether or not to conduct challenged EPA to explain why, if there whether or not risk should be evaluated.
an SSRA and what and how much is any need for SSRAs at all under In support of our position that the
information is required for the SSRA. RCRA, there is a rational basis for why examples we have provided in the 1999
it has limited the entire SSRA program final rule preamble, the 2004 proposed
F. What Are EPA’s Responses to the
to hazardous waste combustors.’’ They rule preamble, and this final rule, are
Cement Kiln Recycling Coalition’s
argue that, ‘‘The point is that if the more than ‘‘what-if’’ hypotheticals, we
Comments on the Proposal and What is
‘‘omnibus’’ words in RCRA mean what have placed copies of completed risk
EPA’s Final Decision on CKRC’s
EPA says they mean for hazardous assessments where risk-based limits
Petition?
waste combustors, why do they not were found to be necessary in the
In the proposal, we provided a mean the same thing for all of the other docket for today’s final rule (see OAR–
lengthy discussion in response to TSD facilities that also pose the same 2004–0022).
CKRC’s petition for rulemaking (69 FR kind of ‘‘what-if’’ hypotheticals that The CKRC fails to acknowledge that
21325–21331). In its petition, CKRC EPA throws out in its preamble?’’ there are many aspects of hazardous
presented two requests with respect to As discussed above in subsection B., waste combustors and the combustion
SSRAs: (1) That EPA repeal the existing and in greater detail below, EPA process itself, which make this category
SSRA policy and technical guidance believes that risk assessments will of TSD facilities different from others,
because CKRC believes that the policy continue to be necessary at some and which factor heavily into our SSRA
and guidance ‘‘are regulations issued facilities. For example, based on the policy. Consider that many combustion
without appropriate notice and inconclusive results from the national facilities feed a wide array of waste
comment rulemaking procedures’’; and risk assessment conducted for the 1999 streams comprised of many hazardous
(2) after EPA repeals the policy and final rule and the comparative risk constituents. The combustion of these
guidance, ‘‘should EPA believe it can analysis conducted for today’s rule, EPA constituents results in complex
establish the need to require SSRAs in is not able to conclude that all MACT chemical processes (which are difficult
certain situations, CKRC urges EPA standards will be sufficiently protective to predict) occurring throughout the
undertake an appropriate notice and for every facility (e.g., non-dioxin PICs combustion unit. The end product is
comment rulemaking process seeking to not previously modeled, no numerical stack emissions comprised of a variety
promulgate regulations establishing dioxin/furan emission standard for solid of compounds different from those that
such requirements.’’ Additionally, fuel-fired boilers, liquid fuel-fired enter the process, and thus are difficult
CKRC stated that it does ‘‘not believe boilers with wet or no APCDs, and to predict because they can vary greatly
that these SSRAs are in any event hydrochloric acid production furnaces, based on the many variables of the
necessary or appropriate’’ and that they etc.). EPA also provided examples of individual combustion unit, making
disagree with EPA’s use of the RCRA site-specific factors that might lead risk them difficult to address (i.e., there are
omnibus provision as the authority to assessors to decide that the MACT no specific emissions standards to limit
conduct SSRAs. Finally, CKRC raised standards may not be sufficiently certain compounds such as products of
three general concerns: (1) Whether an protective, and therefore an SSRA may incomplete combustion). For example,
SSRA is needed for hazardous waste be necessary (e.g., if a source’s in attempting to maximize the
combustors that will be receiving a emissions are comprised of persistent destruction of organic compounds,
RCRA permit when the combustor is in bioaccumulative or toxic contaminants). products of incomplete combustion are
full compliance with the RCRA boiler EPA also discussed this issue at length often generated as a consequence.
and industrial furnace regulations and/ in both the 2004 proposal, and the 1999 Further, due to stack dispersion,
or with the MACT regulations; (2) how rule preamble. See 69 FR 21326 and 64 hazardous waste combustors have the
an SSRA should be conducted; and (3) FR 52842. Given these uncertainties, the potential to affect several square miles.
what is the threshold level for a ‘‘yes’’ SSRA provides significant support for Other types of TSD facilities’ operations
or ‘‘no’’ decision that additional risk- the Agency’s 1006(b) determination typically do not encompass such
based permit conditions are necessary. supporting the elimination of separate complex processes or have the potential
We believe our tentative decision in the RCRA emission standards for MACT to adversely affect receptors for several
proposal addressed each request and EEE facilities. square miles.
concern presented in their petition. We disagree that our discussion of It should be noted that hazardous
However, in its comments, CKRC has standards (and site-specific factors) that waste combustors are not the only type
restated many of the same issues with may warrant a risk evaluation at certain of TSD subjected to site-specific
new emphasis. Thus, we believe it is types of facilities are mere ‘‘what-if’’ evaluations of risk. We take a site-
appropriate to address their major hypotheticals. The examples that we specific approach to regulating
comments in the following discussed in both the earlier preambles miscellaneous units under Part 264,
paragraphs.240a and above were based on the 1999 subpart X. Because it is not possible to
national risk assessment and a develop performance standards and
1. Whether SSRAs Are Necessary for comparative risk analysis, which emission limits for each type of
Facilities in Full Compliance With BIF concluded that either there was not treatment unit that may fall under this
or MACT Regulations enough information to make a definitive broad category, we rely on general
In its comments, CKRC continues to protectiveness determination or that environmental performance standards to
question the need for any SSRAs at uncertainty in cancer and other health meet our mandate under §§ 3004 (a) and
effects levels of dioxin and furans, for (q) that standards governing the
240a CKRC provided numerous comments instance, make it difficult to draw operation of hazardous waste facilities
organized by subtitles. Rather than relying on this conclusions about potential risks. be protective of human health and the
format in the preamble, we have organized the
comments and responses according to the concerns
Furthermore, the discussions with environment. For example, § 264.601(c)
initially raised in the petition, and consistent with respect to the protectiveness of certain requires ‘‘Prevention of any release that
the discussion presented in the proposal. standards (i.e., some are less stringent may have adverse effects on human
health or the environment due to Rather than establish separate facility, they should be conducted
migration of waste constituents in the emission standards under RCRA, EPA where there is some reason to believe
air, considering: * * * (6) the potential decided to coordinate its revisions to that operation in accordance with the
for health risks caused by human the RCRA emissions standards for MACT standards alone may not be
exposure to waste constituents; and hazardous waste combustors with the protective of human health and the
* * *’’ For all intents and purposes, adoption of the MACT standards environment. For hazardous waste
subparts X units are subject to SSRAs as pursuant to § 112(d) of the CAA. See 64 combustors not subject to the Phase 1
well. FR 52832. In the rulemaking standards, we continued to recommend
In addition, the question of whether establishing the MACT standards for that SSRAs be conducted as part of the
an SSRA continues to be necessary is incinerators, cement kilns and RCRA permitting process. See 64 FR
partly a function of the fact that EPA is lightweight aggregate kilns (Phase 1 52841. As discussed in subsection B.,
seeking to rely on CAA MACT standards sources), relying on RCRA § 1006(b), EPA believes that SSRAs may continue
in order to eliminate RCRA emissions EPA determined that in most cases, the to be necessary for some Phase 1
standards for these facilities. As noted MACT standards would be sufficiently facilities. For the Phase 2 sources, our
above, because the MACT is technology- protective that separate RCRA emission comparative risk analysis generally
based, and because of uncertainties in standards and operating conditions indicates that, although the MACT
our national risk assessments, permit would not need to be included in the standards for Phase 2 sources are
writers’ ability to conduct an SSRA in facility’s RCRA permit. However, for a appreciably more stringent than the
individual cases provides important variety of reasons, EPA lacked sufficient current RCRA BIF standards, an SSRA
support for our deferral. factual basis to conclude that a complete may be necessary to confirm that a
RCRA §§ 3004(a) and (q) mandate that deferral of RCRA requirements could be facility will operate in a way that is
standards governing the operation of supported for all facilities. protective of human health and the
hazardous waste combustion facilities Section 1006(b) conditions EPA’s environment.
Thus, for both Phase 1 and Phase 2
be protective of human health and the authority to reduce or eliminate RCRA
sources, we continue to believe that
environment. To meet this mandate, we requirements on the Agency’s ability to
SSRAs may be necessary for some
originally developed national demonstrate that the integration meets
facilities.241 We generally believe the
combustion standards under RCRA, RCRA’s protectiveness mandate (42
MACT standards will be protective; in
taking into account the potential risk U.S.C. 6005(b)(1)). See Chemical Waste
most cases they are substantially more
posed by direct inhalation of the Management v. EPA, 976 F.2d 2, 23, 25
protective than the existing RCRA part
emissions from these sources. With (D.C. Cir. 1992). To support its RCRA
264, 265, and 266 requirements.
advancements in risk assessment § 1006(b) determination, EPA conducted
However, because HWCs manage
science since promulgation of the a national evaluation of both direct and
hazardous waste and process it by
original national standards (i.e., 1981 for indirect human health and ecological
burning and emitting the by-products
incinerators and 1991 for boilers and risks to determine if the MACT
into the air, a multitude of potential
industrial furnaces), it became apparent standards would satisfy the RCRA
exposure pathways exist. These
that the risk posed by indirect exposure mandate to protect human health and
exposure pathways can also vary
(e.g., ingestion of contaminants in the the environment. That evaluation,
substantially based on site-specific
food chain) to long-term deposition of however, did not quantitatively assess
factors associated with an individual
metals, dioxins/furans and other organic the proposed standards with respect to
combustion unit and the surrounding
compounds onto soils and surface mercury and nondioxin products of
site. Such factors make it difficult for
waters should be assessed in addition to incomplete combustion. This was due to
the Agency to conclude that a single,
the risk posed by direct inhalation a lack of adequate information regarding
national risk assessment provides
exposure to these contaminants. We also the behavior of mercury in the
adequate factual support for its
recognized that the national assessments environment and a lack of sufficient
determination that the technology-based
performed in support of the original emissions data and parameter values
MACT standards will be sufficiently
hazardous waste combustor standards (e.g., bioaccumulation values) for
protective. This is further complicated
did not take into account unique and nondioxin products of incomplete
by the fact that, for certain parameters,
site-specific considerations which might combustion. Since it was not possible to
the Agency lacked sufficient
influence the risk posed by a particular suitably evaluate the proposed
information to quantitatively assess the
source. Therefore, until EPA was able to standards for the potential risk posed by
risk, but is relying on a combination of
revise its regulations, to ensure the mercury and nondioxin products of
quantitative and qualitative assessments
RCRA mandate was met on a facility- incomplete combustion, in order to
of the MACT standards’ protectiveness.
specific level for all hazardous waste support our 1006(b) determination, we
Nonetheless, EPA does not believe
combustors, we strongly recommended continued to recommend that SSRAs be
that the uncertainty is so great that it
that site-specific risk assessments conducted for some facilities as part of
would preclude a deferral under 1006(b)
(SSRAs), including evaluations of risk the permitting process until we could
for the affected categories of facilities;
resulting from both direct and indirect conduct a further assessment once final
nor does EPA believe that these
exposure pathways, be conducted as MACT standards were promulgated and
uncertainties necessarily support
part of the RCRA permitting process. In implemented. Specifically, we
requiring a risk assessment for all such
those situations where the results of an recommended that for hazardous waste
facilities. Conditions at the facility
SSRA showed that a facility’s operations combustors subject to the Phase 1
could pose an unacceptable risk (even MACT standards—hazardous waste 241 As discussed in section B., we expect that
after compliance with the RCRA burning incinerators, cement kilns and facilities that have previously conducted an SSRA
national regulatory standards), light-weight aggregate kilns—permitting will not need to conduct another in consideration
additional risk based, site-specific authorities should evaluate the need for of today’s final standards. Only those facilities
newly subject to the RCRA permitting requirements,
permit conditions could be imposed an SSRA on a case-by-case basis. We or existing sources where changes in conditions
pursuant to RCRA’s omnibus authority, further stated that while SSRAs are not could lead to increased risk, may need to conduct
§ 3005(c)(3). anticipated to be necessary for every or modify an existing SSRA.
might confirm that the MACT standards specific consideration of risk into its Given the complexity of multi-pathway
are sufficiently protective, without the regulatory framework. and indirect exposure assessments and
need for a facility-wide risk assessment. Next, CKRC comments that EPA has the fact that risk science is continuously
For example, if the results of the MACT a non-discretionary duty under CAA evolving, it would be difficult and
testing demonstrated that the facility’s § 112(f) to address and take care of any again, overly constraining, to codify risk
dioxin emissions fall below the levels ‘‘residual risk’’ from MACT facilities in parameters today. We note as well, in
estimated in the database EPA used for the future in any event. We discussed this regard, that several commenters
its comparative risk assessment, the why we do not believe that the residual agreed that codification of EPA’s risk
uncertainties in EPA’s comparative risk risk process should or can take the place assessment guidance would be too
assessment would not, by itself, support of an SSRA under RCRA in subsection constraining for both the agency and the
a decision to require an SSRA. Such C. of this SSRA preamble, as well as in regulated community.
decisions require an evaluation of the the 1999 rule preamble (64 FR 52843). We also believe that a guidance
conditions at the site, and EPA believes In short, because the residual risk approach is consistent with the fact that
it important to retain the flexibility for standards have not yet been established, permit authorities must make site-
permit authorities to take these permit writers cannot rely on this specific decisions whether to do risk
conditions into account. Accordingly, process in reaching current permitting assessments at all. We think that it
EPA believes that the regulatory decisions or in acting on currently makes little sense to allow this kind of
structure adopted in today’s rule strikes pending permit applications. flexibility regarding whether to do a risk
the appropriate balance between these assessment and for what purposes,
2. Codification of EPA’s Technical
competing factors. while prescribing how one must be
Guidance
conducted if one is required. In fact,
In response to EPA’s statement in the In response to our explanation in the permitting authorities, in some cases,
proposal that non-HAP emissions, proposal that risk assessment guidelines have developed their own guidance
which were beyond the direct scope of should be flexible and reflect current methodologies responsive to the specific
MACT, may pose risk which could science, CKRC gave three comments: (1) needs associated with their facilities.
necessitate an SSRA (69 FR 21326), Not a word of the current SSRA For example, North Carolina, Texas, and
CKRC pointed out that the same could guidelines has been changed in 3 years; New York have each developed their
be said for other types of TSDs, such as (2) it is easy to write regulations that own risk assessment methodologies.
landfills, land treatment systems, have provisions that might be applied Further, facilities that choose to conduct
etcetera, and EPA has not addressed this differently in different situations, and at SSRAs themselves can choose
point in its preamble. As previously least many basic, fundamental points alternative approaches in applying
noted, combustion units are distinct can go in regulations, while some methodologies as well. We think this
from other types of TSDs due to the details can be in guidance—EPA writes flexibility employed in the field
wide array of waste streams being fed to regulations accompanied by ‘‘fill in the supports our judgment that risk
the unit, the complex chemical small details’’ guidance all the time; and assessment methodologies should not be
processes throughout the combustion (3) EPA seems to have no real problems codified. CKRC’s comments failed to
unit, stack emissions comprised of a with regulatory fixes anyway. In address any of these issues.
wide variety of compounds that are addition, CKRC provides several Turning to the remainder of CKRC’s
difficult to address, and the potential to comments related to the previous three specific points—CKRC’s assertion that
impact receptors for several square throughout their comment document, the technical guidance has not been
miles due to stack dispersion. A further which are addressed below. amended in the past three years is
distinction is that EPA is seeking to rely None of these comments address the inaccurate. A revised HHRAP guidance,
on the MACT standards to eliminate specific issue EPA raised, which is that, that has been amended to take into
national RCRA stack emissions while it certainly is possible to codify account the technical recommendations
standards under § 1006(b). Unless EPA our risk assessment guidance, for a from both the public comments and
can affirmatively demonstrate that variety of reasons, we disagree that it peer review, is published in conjunction
RCRA’s protectiveness standards are would be appropriate to issue these with this rule. In addition, as noted
met, the Agency cannot eliminate RCRA technical recommendations as a above, in some cases, permitting
requirements. A number of uncertainties regulation. As we previously explained, authorities have developed their own
remain concerning the protectiveness of risk assessment—especially multi- methodologies responsive to the specific
the MACT standards based on the pathway, indirect exposure needs associated with their facilities.
uncertainties remaining in the assessment—is a highly technical and With respect to CKRC’s third point,
supporting national risk assessment and evolving field. Any regulatory approach the regulatory corrections made to the
comparative analysis, and the variability EPA might codify in this area is likely MACT rules were necessary either to fix
of site-specific factors from one facility to become outdated, or at least an error or omission or to resolve
to another. Permitting authorities’ artificially constraining, shortly after potential legal issues. To codify
ability to resolve these uncertainties promulgation in ways that EPA cannot technical tools and chemical
through the use of the SSRA, where anticipate now. In support of this, we information pertinent to the risk process
appropriate, provides important support noted specific examples of problems we simply is not prudent, as this
for the Agency’s 1006(b) finding. experienced in implementing the BIF information is continually changing and
Furthermore, as we have noted, under regulations. See 69 FR 21330. Further, would almost always be out of date.
omnibus, to the extent permitting we explained that at the time of Granted, when this information is
authorities believe there are problems codification, BIF risk assessments were presented in guidance, it can just as
with other types of TSDs, they can not intended to address indirect routes easily become outdated, however,
impose requirements and request of exposure, thus making the parameters facilities and risk assessors are free to
additional information, including an easier to implement. Today, however, use the most up-to-date air modeling
SSRA in accordance with § 270.10(k). risk assessments are more complex due tools and toxicity values available (i.e.,
Also as previously noted, Part 264, to the necessary inclusion of multi- they would not be bound to regulations
subpart X specifically incorporates site- pathway and indirect exposure routes. requiring the use of obsolete tools and
information). We continue to believe documents to make clear that they are we have solicited public review and
that publishing our technical non-binding. See 69 FR 21329. We comment. As a result, it has been
recommendations as regulation would specifically noted that CKRC indicated improved over the years by including
remove much of the flexibility that is in its petition that, in its view, the revisions to the guidance based upon
important in evaluating risk on a site- documents contain language that could feedback from users of the guidance and
specific basis. be construed as mandatory. While EPA from experience in the field. A response
CKRC discounts EPA’s statement that does not necessarily agree, and believes to comments document has been
codification of risk assessment is the that, in context, it is clear that the prepared and released along with the
exception arguing that ‘‘Neither TSCA recommendations in the documents are final HHRAP and final MACT rules,
or CERCLA, however, specifically discretionary, EPA nonetheless even though the Agency was not
commands EPA to define the type of reviewed the documents to ensure that required to do so. More to the point,
information necessary for a permit they are carefully drafted. Consequently, because it is only guidance, sources will
application through the rulemaking under the standards articulated in have the opportunity to raise questions
process as RCRA does. Moreover, the Appalachian Power Co. v. EPA, 208 or comments on anything in the
TSCA and CERCLA examples EPA cites F.3d 1015 (D.C. Cir. 2000) and guidance as part of the permitting
are not analogous to the situation where subsequent case law, the final HHRAP process and the permitting authority
a permit applicant can be denied a guidance is truly guidance and does not will be required to respond to those
permit—or at least strung through require notice-and-comment comments as part of the permitting
months or years of tortuous and costly rulemaking. The HHRAP explains in process. See 40 CFR part 124. Sources
submissions, revision, and great detail an acceptable process for will also have the right to challenge the
resubmission—to obtain a permit.’’ performing and reporting on cost- responses or use of the guidance as part
Even if TSCA and CERCLA were not effective, scientifically defensible risk of the permitting process.
considered to be analogous, that does assessments. It includes numerous
not change EPA’s fundamental rationale 3. Codification of Criteria for
recommended defaults, while at the
that codification of highly technical risk Determining That Additional Risk-
same time provides the risk assessor or
assessment guidance is not appropriate. Based Permit Conditions or an SSRA Is
facility full opportunity to incorporate
EPA does not believe that RCRA Necessary
site-specific values in place of the
§ 3005(b) requires EPA to codify an defaults. The HHRAP offers numerous CKRC argues that EPA’s proposed
exhaustive list of every possible piece of recommendations, but requires nothing. regulatory changes should not be
information that might be required in a EPA has placed a copy of the final considered as a partial grant because
permit. To some extent, that is the guidance document in the docket for EPA has not codified specific criteria in
reason for having a permit process—to today’s action (see OAR–2004–0022). the proposed regulations for permit
allow site specific conditions to be CKRC believes that EPA’s technical authorities to use to decide whether to
taken into account. Nevertheless, EPA guidance imposes information require an SSRA; to set the risk levels
has revised part 270, pursuant to RCRA requirements upon the RCRA permit that are deemed protective; or to
§ 3004(a) and § 3005(b) to specifically applicant that are not contained in any otherwise provide any further definition
provide that a risk assessment may be regulations and in fact exceed by orders as to what it means to protect human
necessary, where there is reason to of magnitude any information health and the environment.
believe that the MACT standards may requirements contained in the part 270 In its petition, CKRC requested that
not be sufficiently protective. This was regulations. We disagree that anything after we repeal the policy and guidance
done wholly to address the petitioner’s contained in HHRAP is ‘‘required’’ in (per the first request), ‘‘should EPA
concern that the current regulations do any way. Moreover, to the extent any believe it can establish the need to
not adequately provide notice that an individual facility believes the require SSRAs in certain situations,
SSRA might be necessary as part of a information requested is inappropriate CKRC urges EPA to undertake an
permit application. This provision, or unnecessary, they can challenge that appropriate notice and comment
while it does not provide as much detail as part of the permitting process. rulemaking process seeking to
as the petitioner wishes, clearly Lastly, CKRC argues that ‘‘The promulgate regulations establishing
‘‘defines the type of information procedures EPA has been using to issue such requirements.’’ As discussed at
necessary for a permit application.’’ and revise the SSRA guidance do not by length in both the proposal (69 FR
CKRC complains that the Agency did any measure comply with the full 21325–21327) and the preceding
not address in its proposed response the panoply of procedures and protections paragraphs, we believe that we have
petitioner’s discussion of the ‘‘strong offered by the APA process. Most established certain circumstances where
case law compelling the conclusion that critically, when EPA merely solicits the MACT standards may not be
‘guidance’ documents EPA has issued comments on draft guidance documents, protective and that an SSRA may be
for conducting SSRAs must be subjected it has no duty to respond to comments warranted, based on relevant site-
to notice-and-comment rulemaking.’’ and provide a rational basis and specific factors associated with an
EPA has chosen not to respond to justification in defense of its choices in individual combustion unit.
CKRC’s legal interpretation because we the face of comments. EPA is essentially Consequently, we are finalizing
believe that it is clear that the guidance running its entire SSRA program on the regulations that explicitly authorize
documents do not impose mandatory basis of ‘‘draft’’ guidance versions for permitting authorities to conduct or
requirements, and therefore need not be which EPA has never to this day require an SSRA on a site-specific basis.
issued by notice and comment prepared any response to comments.’’ This, in our view, grants the second of
rulemaking. Nevertheless, EPA notes As previously noted, EPA believes the CKRC’s requests. Our response directly
that in the proposal, the Agency final HHRAP is merely guidance and addresses a number of CKRC’s concerns:
explained that we were in the process therefore, EPA is not required to (1) Through a notice and comment
of reviewing the guidance documents, proceed through notice and comment rulemaking process, EPA has
and, to the extent we found language rulemaking pursuant to § 553 of the established circumstances in which an
that could be construed as limiting APA. However, because we want the SSRA may be necessary; and (2) EPA’s
discretion, we committed to revise the HHRAP guidance to be useful and clear, regulations will now explicitly
acknowledge that an SSRA might be struck down as a ‘‘standardless combustion unit at the site. EPA has
necessary as part of the permitting regulation.’’ also added language to § 270.10(l) to
process, thereby addressing the EPA disagrees that the provisions at identify guiding factors for permitting
petitioner’s concern that EPA’s past § 270.10(l) are impermissibly vague, or authorities to consult in determining
approach of relying on RCRA’s omnibus otherwise inconsistent with the cases whether the MACT will be sufficiently
authority to implement this policy the petitioner cites. In the cited cases protective at an individual site.
violates the requirements of RCRA the courts found that the regulated Although the list of guiding factors is
§ 3005(b). And as discussed further entity bore the entire burden of not all-inclusive, they offer a structure
below, EPA has codified criteria for determining how to comply with the for risk managers (as well as the
permit authorities to use to determine challenged regulation in the complete regulated community) to use to frame
whether to require an SSRA. absence of a government-generated the evaluation of whether a combustor’s
While it does not provide exactly standard or guidance. See Maryland v. potential risk may or may not be
what CKRC requested, the regulated EPA, 530 F.2d 215, 220 (4th Cir. 1975); acceptable.
community has had a full opportunity South Terminal Corp v. EPA, 504 F.2d Finally, we note that, unlike the
to comment on the need for an SSRA 646, 670 (1st Cir. 1974). This is entirely circumstances in the cited cases,
both as part of the 1999 rulemaking and, distinct from the regulations codified at § 270.10 is promulgated in the context
again, as part of this rulemaking to § 270.10(l). of an existing permitting regime. The
adopt the provisions of § 270.10(l), In § 270.10(l) EPA identified the regulatory standards at 40 CFR part 124
which contain an explicit reference to standard for when a risk assessment provide further structure for both the
the potential need for an SSRA as part may be necessary: where the regulatory regulated community and the permit
of the permitting process pursuant to authority identifies factors or conditions authority. For similar reasons, EPA
at the facility that indicate that the disagrees that the cited cases compel the
RCRA § 3004(a) and § 3005(b). As
MACT standards may not be sufficiently Agency to establish risk levels that are
previously explained, § 270.10(k) does
protective, and defers the articulation of deemed protective, or to otherwise
not explicitly mention the potential for
the more precise requirement to the provide any further definition as to
an SSRA to be required. Although the
permitting process, where the onus falls what it means to protect human health
rule does not identify a priori that an
on the permitting authority to identify and the environment. We discussed at
SSRA will be required in an individual
the basis for its determination. Until the length throughout the proposal the
circumstance, but defers that
permitting authority provides this reasons we believe it would not be
determination to the permitting process,
further guidance, the regulated entity appropriate to codify either an exclusive
the final rule reflects EPA’s findings that
incurs no obligation. The mere fact that set of national criteria for determining
an SSRA is not anticipated to be specific factors or facility conditions that an SSRA (or additional risk-based
necessary in every circumstance—only that form the basis for the determination permit conditions) would be necessary,
where site-specific conditions give the that an SSRA is warranted will be or a uniform risk level. The decision to
permit authority reason to believe that subsequently identified through the require an SSRA is inherently site
additional controls beyond those permitting process does not invalidate specific, thus permitting authorities
required pursuant to 40 CFR parts 63, the regulation. See Ethyl Corp v. EPA, need to have the flexibility to evaluate
264, 265, or 266 may be necessary to 306 F.3d 1144, 1149–1150 (D.C. Cir. a range of factors that can vary from
protect human health and the 2002). facility to facility. See 69 FR 21328–
environment. The regulation also identifies the 21331. CKRC has neither presented new
CKRC argues that EPA’s decision not categories of information that might be factual or policy reasons that would
to codify national criteria renders the required for MACT EEE facilities: The cause the Agency to reconsider the
regulation impermissibly vague, and information must be necessary to tentative decisions presented in the
therefore, ‘‘in their view totally deficient determine whether additional controls proposal, nor specifically addressed the
as a legal matter.’’ The petitioner argues are needed to ensure protection of issues underlying EPA’s decision.
that the rule is essentially ‘‘a bootstrap human health and the environment; it Instead, the petitioner has merely
attempt to avoid rulemaking can include the information necessary reiterated the concerns presented in its
requirements by establishing ‘rules’ that to evaluate the potential risk from both petition and its general disagreement
give no more guidance or direction than direct and indirect exposure pathways; with EPA’s decision.
general terms in the statute and in no or it can include the information EPA also disagrees that its new
way channel the decision maker’s necessary to determine whether such an regulatory structure grants permit
discretion or put the public on notice of assessment is necessary. Here as well, writers unbridled discretion for many of
anything.’’ According to CKRC, this EPA’s reliance on the permitting process the same reasons that EPA does not
unbridled discretion is manifest in three to provide further specification of the believe that § 270.10(l) is impermissibly
ways: (1) No criteria explain how a required information is not improper. vague. As EPA has previously
permit writer is to decide whether to Moreover, as discussed above in explained, the requirements at Part 124
require an SSRA; need merely to subsection C., in response to continue to apply to actions taken to
conclude ‘‘reason to believe’’; (2) there commenters’ concerns, EPA has revised implement § 270.10(l). Moreover, the
are absolutely no limits on what type of § 270.10(l) to provide more detail, both language of § 270.10(l) makes clear that
information or assessments the permit with respect to the basis for the the onus initially falls on the permitting
writer may demand and the proposed determination that an SSRA is authority to identify the basis for its
reg. does not even hint at what type of necessary, and with respect to the type conclusion that the MACT standards
information or assessments might be of information the permit authority may not be sufficiently protective. As
demanded; and (3) there is not a word might need. EPA has added language to both part 124 et. seq., and EPA’s
of guidance or specification as to what remind permit authorities that the preamble discussions make clear,
it means to ‘‘ensure protection of human determination that the MACT standards facilities will continue to have the
health and the environment.’’ The may not be sufficiently protective is to opportunity to comment on and
petitioner argues that as a consequence, be based only on factors relevant to the challenge the determination. See
the proposed § 270.10(l) would be potential risk from the hazardous waste §§ 124.10, 124.11, and 124.19. The
regulatory structure adopted in comply with the requirement in RCRA assessment is necessary. While it does
§ 270.10(l) mirrors the structure § 3005(b) to specify in regulations, the not provide as much detail as the
Congress established in sections 3004 information necessary to obtain a petitioner wishes, this provision
and 3005; although 3004 directs EPA to permit. They compare the level of detail unquestionably ‘‘defines the type of
establish national standards, section in § 270.10(l) to the lengthy regulations information necessary for a permit
3005 recognizes that those standards (codified in 40 CFR part 270) specifying application.’’
will be applied on a case-by-case basis in great detail the information required Thus, the issue turns on the level of
through the permitting process, to allow when one is submitting a RCRA permit specificity that RCRA § 3005(b) requires,
site-specific conditions to be taken into application, arguing that ‘‘these and EPA does not believe that RCRA
account, and to supplement those regulations cover 75 pages of fine print § 3005(b) requires EPA to publish a list
standards as necessary. in Code of Federal Regulations,’’ to of every possible piece of information
EPA has also provided demonstrate that this regulation would that might be required in a permit.
recommendations through guidance on be insufficient under RCRA § 3005(b). In Section 3005(b) merely establishes a
how an SSRA can be conducted. further support of this argument, CKRC broad directive that ‘‘each application
Although the recommendations are not cites Ethyl Corporation v. EPA, 306 F.3d for a permit under this section shall
binding, they provide risk managers (as 1144 (D.C. Cir. 2002). contain such information as may be
well as the facility) with a starting point EPA disagrees that its regulations are required under regulations promulgated
from which to determine whether a in any way inconsistent with the by the Administrator,’’ and that it shall
combustor’s potential risk may or may decision in Ethyl Corp. At issue in that include the information contained in
not be acceptable. case was a regulation issued pursuant to subsections (1) and (2), leaving to EPA’s
CKRC argues that it appears that section 206(d) of the CAA. Section discretion to determine the level of
rather than following the statutory 206(d) provides that EPA ‘‘shall, by specificity at which to promulgate
authorities and requirements to review regulation, establish methods and regulations. To some extent, this reflects
and amend regulations every 3 years as procedures for making tests under this the reason for having a permit process—
necessary (RCRA § 2002(b)), EPA section.’’ 42 U.S.C. 7525(d). The court to allow site specific conditions to be
decided to take the easy way out and found that ‘‘with CAP 2000, [the taken into account. The regulatory
impose, through non-rulemaking challenged regulation] the EPA does not structure adopted in § 270.10 mirrors
‘‘guidance’’, massive, costly, and claim to have itself articulated even a the structure Congress established in
confusing requirements leaving vague durability test. Rather CAP 2000 RCRA § 3004 and § 3005. Despite the
unbridled discretion to its permit requires that ‘the manufacturer shall petitioner’s comparison to the length of
writers. propose a durability program’ for EPA part 270, the length of these provisions
We disagree that the Agency has approval. 40 CFR 86.182301(a). It thus are not indicative of any determination
attempted to avoid rulemaking in this falls on the forbidden side of the line.’’ of the precise level of detail that
context. EPA has conducted several Ethyl Corp., 306 F.3d at 323–324. The § 3005(b) requires, but reflects the fact
rulemakings to amend our regulations. Court distinguished the challenged that EPA has adopted requirements
The first was in 1999, when we adopted regulation from the situation in which specific to individual types of units.
revised emission standards under the an agency issues a ‘‘vague’’ regulation, Moreover, notwithstanding the
authority of both § 112(d) of the CAA and relies on subsequent proceedings to petitioner’s characterization, the
and RCRA to more rigorously control flesh out the specific details. And as the language at § 270.10(l) is comparable to
toxic emissions from burning hazardous court explained, where ‘‘Congress had many other provisions in 40 CFR part
waste in incinerators, cement kilns, and not specified the level of specificity 270. See, for example: §§ 270.14(b)(8);
lightweight aggregate kilns. See 64 FR expected of the agency, we held that the 270.16(h)(1)–(2); 270.22(a)(6)(i)(C);
52828. At the time, we noted that agency was entitled to broad deference 270.22(c).
‘‘today’s rule fulfills our 1993 and 1994 in picking the suitable level.’’ 306 F.3d Lastly, CKRC argues that the proposed
public commitments to upgrade at 323 (citing American Trucking regulation is particularly problematic,
emission standards for hazardous waste Associations v. DOT, 166 F.3d 374 (D.C. because it extends beyond
combustors.’’ We have continued to Cir. 1999) and New Mexico v. EPA, 114 ‘‘information’’ that may already exist.
revise our regulations consistent with F.3d 290 (D.C. Cir. 1997). CKRC says that it is one thing to
and based on the facts before the In § 270.10(l) EPA has articulated the demand that a party go out and gather
Agency, taking into account the standard for when a risk assessment existing information, but another thing
arguments presented in CKRC’s petition. may be necessary: where the regulatory to demand that an applicant conduct
As explained above, we believe that the authority has identified factors or ‘‘assessments.’’ Moreover, nothing in the
facts do not support granting all of conditions at the facility that indicate regulations prohibits a permit authority
CKRC’s requests. Rather we believe that that the MACT standards may not be from demanding revised assessments,
the MACT standards will generally be sufficiently protective. EPA has also and even more revised assessments. We
protective, and that permit authorities adopted a list of factors on which permit agree that permit authorities have the
should reach the decision to require an writers are to rely in reaching this authority to require facilities to provide
SSRA based on a variety of factors and determination. EPA has also identified additional information beyond that
concerns specific to their sites. In the categories of information that might which already exists. However, based
addition, as previously addressed, we be required for MACT EEE facilities: on feedback from EPA Regional permit
believe that our risk assessment The information must be necessary to writers, SSRAs generally represent a
guidance should remain as guidance. determine whether additional controls one-time cost. We do not expect that
Several other commenters agree that the are needed to ensure protection of facilities that have conducted risk
guidance should not be codified. human health and the environment; it assessments will have to repeat them.
The petitioner argues that the can include the information necessary As discussed in the 1999 final rule
regulation EPA has proposed to adopt is to evaluate the potential risk from both preamble, changes to comply with the
so vague, that it is essentially not a direct and indirect exposure pathways; MACT standards should not cause an
regulation, and that consequently, even or it can include the information increase in risk for the vast majority of
if finalized, it would not be sufficient to necessary to determine whether such an facilities given that the changes, in all
probability, will be the addition of estimates developed by the Agency do estimates of SSRA costs. We disagree
pollution control equipment or a include data gathering costs, QA/QC, with this comment. The estimates of
reduction in the hazardous waste being and third-party consultants. (Refer to SSRA costs developed by the Agency
burned (see 64 FR 52842). Instances the proposed rule’s support document reflect total contractor costs for
where a facility may need to repeat a entitled: Preliminary Cost Assessment performing an SSRA at a facility under
risk assessment would be related to for Site Specific Risk Assessment, different sets of conditions, and are not
changes in conditions that would likely November 2003, Docket # OAR–2004– limited to stack emissions.
lead to increased risk.242 In such 0022; and the Assessment of the In the fifth cost-related issue, CKRC
situations, we would anticipate that the Potential Costs, Benefits, and Other asserts that EPA’s average estimates
risk assessment would not have to be Impacts of the Hazardous Waste might be reasonable if the SSRA process
entirely redone. It may be as limited as Combustion MACT Replacement were limited to the submission and
collecting relevant new data for Standards—Final Rule, October 12, acceptance of one SSRA effort. CKRC
comparison purposes, leading to a 2005, for a description of how the contends, however, that its members’
decision not to repeat any portion of a estimates were arrived at.) However, we experiences with SSRAs have involved
risk assessment. Or, it may be more agree with CKRC that the method used coordination with state and regional
inclusive such that modifications would to develop SSRA costs does not capture offices and multiple revisions and
be made to specific inputs to or aspects facility time associated with data submissions. Again, we do not question
of the risk assessment using data from collection and management related to the experiences and costs of specific
a previous risk assessment, risk burn or the SSRA. Consequently, we have facilities. However, we anticipate that
comprehensive performance test. As adjusted our SSRA cost estimates to the 2003 Memorandum, Use of the Site-
discussed in subsection B., we have account for these activities by Specific Risk Assessment Policy and
added a new regulatory provision to incorporating costs associated with time Guidance for Hazardous Waste
indicate a previously conducted risk needed for facility data collection and Combustion Facilities, and the Human
assessment would be relevant in management efforts associated with the Health Risk Assessment Protocol for
evaluating changes in conditions that SSRA, and will assume that engineering Hazardous Waste Combustion Facilities
may lead to increased risk. The factor staff are required to perform these tasks. guidance, which is finalized and
reads as follows: ‘‘Adequacy of any The second issue concerns the extent released in conjunction with today’s
previously conducted risk assessment, to which cement kiln SSRAs are rule, will provide facilities and
given any subsequent changes in consistent with EPA’s ‘‘normal’’ regulators with a clearer understanding
conditions likely to affect risk.’’ assumptions. We do not question the of SSRA policy and guidance and will
accuracy of the costs submitted by support a more efficient SSRA process.
4. EPA’s Cost Estimates for SSRAs CKRC. However, it is not clear that the EPA’s future SSRA cost estimates are
CKRC raised several objections to our costs submitted by CKRC represent based on current or recent cost data
cost estimates for conducting an SSRA, typical future costs for SSRA from multiple practitioners, and likely
and provided higher cost estimates implementation at all facilities in the reflect a more efficient process than that
($200K to $1M, with upper bound of universe. Certain of the CKRC cost experienced by some CKRC members in
$1.3M). We suggested in the proposal, estimates (e.g., those submitted by Ash the 1990s.
that the higher cost figures provided by Grove and Holcim) reflect
CKRC were likely incurred prior to the X. Permitting
implementation of SSRAs over a
1998 release of the Human Health Risk number of years in the 1990s, while As discussed in the proposal, we
Assessment Protocol (HHRAP) guidance SSRA implementation was in its early believe that the permitting approach we
document. We believe our lower cost stages. In other cases (e.g., estimates adopted in the 1999 final rule is still the
estimates can be attributed to the fact provided by Solite) costs appear to be most appropriate means to avoid
that we based them on the conduct of consistent with EPA estimates. While duplication to the extent practicable and
future SSRAs that will benefit from we do not dispute the accuracy of these to streamline requirements. Thus, both
substantially better guidance and costs, earlier costs are likely to reflect Phase 1 and Phase 2 sources will
commercially available software. the deliberative process common with comply with their RCRA emission limits
Multiple issues regarding the cost early SSRAs. and operating requirements until they
information we provided in the For the third issue, CKRC’s points out demonstrate compliance with the
proposal are raised by CKRC. The first that EPA’s estimate of 20 percent MACT standards by conducting a
of five issues is that CKRC believes that additional cost for adding a risk burn comprehensive performance test (CPT),
EPA’s methods for calculating costs during a trial burn may be low; CKRC submitting a Notification of Compliance
associated with future SSRAs do not asserts that additional test costs can add (NOC) documenting compliance to the
include data gathering costs, QA/QC, up to 40 percent depending on the Administrator or delegated state, and
third party consultants in addition to circumstances. We agree with this and then requesting to have their RCRA
risk assessors and plant personnel time have adjusted the range of total SSRA permits modified to remove the
to coordinate and review SSRA efforts costs as necessary to assure that a range duplicative RCRA requirements (unless
and collect facility data. We disagree of additional test costs for separate risk a sunset clause had been added
with this statement in part; the burns (20 to 40 percent incremental previously that inactivates specified
cost) are included. For revised figures, requirements upon compliance with
242 For example, hazardous waste burning cement
see background document, Assessment MACT).243 Ultimately, the MACT air
kilns that previously monitored hydrocarbons in of the Potential Costs, Benefits, and emissions and related operating
the main stack may elect to install a mid-kiln requirements will reside in the CAA
sampling port for carbon monoxide or hydrocarbon Other Impacts of the Hazardous Waste
monitoring to avoid restrictions on hydrocarbon Combustion MACT Replacement Title V permit, while all other aspects
levels in the main stack. Thus, their hydrocarbon Standards—Final Rule, October 12,
emissions may increase. (64 FR 52843, footnote 29.) 2005. 243 Although we expect that the vast majority of
Another example would be if the only change at a Phase 1 sources will have had their RCRA permits
facility relates to the exposed population; what was
CKRC’s fourth issue is that EPA does modified by the time this rule is promulgated, we
acceptable in a previous risk assessment may not not appear to include more than acknowledge that there may be a few permits yet
be any longer. evaluations of stack emissions in its to be modified.
of the combustion unit and the facility regarding the protectiveness of the risk-based conditions are present). We
(e.g., corrective action, general facility MACT standards, and pursuant to agree with the commenters that as long
standards, other combustor specific section 1006(b), EPA is generally as the technology upgrades (e.g.,
concerns such as material handling, eliminating the existing RCRA stack equipment changes to upgrade air
risk-based emission limits and operating emission standards to avoid duplication pollution control equipment) do not
requirements, and other hazardous with the new MACT standards. The affect any remaining conditions in the
waste management units) will remain in amendments made today to allow new RCRA permit, the regulations do not
the RCRA permit. A new pictorial combustion units to comply with the require a permit modification. For those
timeline has been provided to highlight MACT standards upon start-up, versus Phase 1 sources that need to make
milestones of the MACT compliance the RCRA stack emissions national upgrades to comply with the revised
process. See figure 1 at the end of this standards, are based on the principle of standards, they should address the
section. avoiding duplication between programs. specific upgrades in their draft
We are not stating that RCRA permit Notification of Intent to Comply (NIC)
A. What is the Statutory Authority for conditions to control emissions from
the RCRA Requirements Discussed in and during the informal NIC public
these sources will never be necessary,
this Section? meeting so that the regulatory authority
only that the national RCRA standards
EPA is finalizing amendments to and public are aware of the source’s
appear to be unnecessary. Under the
modify RCRA permits in today’s rule authority of RCRA’s ‘‘omnibus’’ clause activities and plans for compliance. We
pursuant to sections 1006(b), 2002, section 3005(c)(3); (see §§ 270.32(b)(2) encourage early communication
3004, 3005 and 7004(b) of RCRA. 42 and (b)(3)), RCRA permit authorities between the source and the RCRA
U.S.C. §§ 6905(b), 6912, 6924, 6905, and may impose additional terms and permit writer to ensure a common
6074. Our approach is likewise conditions on a site-specific basis as understanding of whether a RCRA
consistent with section 112(n)(7) of the may be necessary to protect human permit modification will be needed.
Clean Air Act which indicates that EPA health and the environment. Thus, if Additionally, Phase 1 sources must
should strive to harmonize requirements MACT standards are not protective in comply with the provisions of
under section 112 and RCRA an individual instance, RCRA permit § 63.1206(b)(5) for changes in facility
requirements for hazardous waste writers will establish permit limits that design. We do not anticipate that
combustion sources. With respect to the are protective. upgrades made to comply with the
regulatory framework that is discussed In RCRA, Congress gave EPA broad Replacement Standards will adversely
in this section, we are finalizing the authority to provide for public affect a source’s compliance with the
process to eliminate the existing RCRA participation in the RCRA permitting Interim Standards. Therefore, consistent
stack emissions national standards for process. Section 7004(b) of RCRA with § 63.1206(b)(5)(ii), sources must
hazardous air pollutant for Phase 2 requires EPA to provide for, encourage document the change in their operating
sources as we had done for Phase 1 and assist public participation in the record, revise their NOC and resubmit it
sources in the 1999 final rule. That is, development, revision, implementation, to the permitting authority (per
after submittal of the NOC established and enforcement of any regulation, § 63.9(h)), and, as necessary, revise their
by today’s rule and, where applicable, guideline, information, or program start-up, shutdown, and malfunction
once RCRA permit modifications are under the Act. plan.244
completed at individual facilities, RCRA
national stack emission standards will B. Did Commenters Express any Several commenters felt that we
no longer apply to these hazardous Concerns Regarding the Current should re-emphasize the importance of
waste combustors (unless risk-based Permitting Requirements? removing duplicative RCRA permit
permit conditions are determined Generally speaking, commenters favor conditions and requirements. We agree
necessary). maintaining the permitting approach with the commenters that this is an
We originally issued emission and requirements referred to above. This important action for regulatory agencies.
standards under the authority of section approach was finalized in the 1999 rule In addition to comments received, we
3004(a) and (q) of RCRA, which calls for and has been implemented, and in a few also have learned through the
EPA to promulgate standards ‘‘as may cases is currently being implemented, implementation process for the Interim
be necessary to protect human health for Phase 1 sources complying with the Standards, that some state agencies are
and the environment.’’ We believe that Interim Standards Rule. However, not removing duplicative requirements
the final MACT standards are generally several commenters raised similar from the RCRA permit. We have clearly
protective of human health and the concerns regarding certain aspects of the stated in several preambles and
environment, and that separate RCRA transition process from RCRA to MACT guidance documents that we believe it
emission standards are not needed to and Title V permitting. is appropriate to retain only the RCRA
protect human health and the risk-based conditions that are more
environment. See Part Seven, How Does 1. Removal of Duplicative RCRA Permit
Conditions stringent than the applicable MACT
the Final Rule Meet the RCRA limits (i.e., if the RCRA condition has
Protectiveness Mandate? for a One comment is in regard to Phase 1 been determined to limit risk to an
discussion of this topic. RCRA section sources that have been fully transitioned acceptable level and is necessary to
1006(b) directs EPA to integrate the (i.e., have had duplicative RCRA permit protect human health and the
provisions of RCRA for purposes of conditions and requirements removed environment) in the RCRA permit after
administration and enforcement and to or that have been ‘‘sunsetted’’) to
avoid duplication, to the maximum compliance with the Interim Standards 244 The requirements in § 63.1206(b)(5)(ii) call for
extent practicable, with the appropriate that may need to make upgrades to sources to revise (as necessary) the performance test
provisions of the Clean Air Act (and comply with the revised Phase 1 MACT plan, DOC, NOC, and start-up, shutdown, and
other federal statutes). This integration Standards. The concern is that Phase 1 malfunction plan. For sources complying with the
Interim Standards, it is not necessary to revise the
must be done in a way that is consistent sources needing to make upgrades for performance test plan or the DOC, since they were
with the goals and policies of these MACT should be able to do so without developed in preparation for compliance with the
statutes. Therefore, based on its findings a RCRA permit modification (unless Interim Standards.
compliance with MACT.245 However, Class 1 modification procedure at RCRA permit once the source has
we also acknowledge that in certain 270.42 Appendix I, section A.8) into demonstrated compliance with MACT.
cases it may not be clear which their state requirements (e.g., because We agree with commenters that every
compliance requirement is more the state’s procedures are broader in effort should be made to conserve
stringent. For example, standards under scope or more stringent than the federal resources and avoid duplication to the
MACT are expressed as concentration requirements) or is unable to reach an extent possible. However, we do not
based limits (micrograms/dry standard agreement between its RCRA and air believe it is appropriate to establish
cubic meter) while certain RCRA programs regarding which standards are policy or regulation that permitting
standards are expressed as mass more stringent, the Title V permitting authorities must suspend the RCRA
emission rate limits (grams/second). authority should document these issues, permit process (whether it pertains to
Also, averaging times between the two including any continuing RCRA permit interim status or renewals), especially in
programs differ: MACT requires hourly requirements, in the title V permit’s cases where considerable time and effort
rolling averages whereas RCRA requires statement of basis (40 CFR §§ 70.7(a)(5) has been invested and the permit is
instantaneous values. This is an and 71.7(a)(5)). This will help to ensure close to final issuance. As before, we
unfortunate consequence of moving that the source is clear regarding its strongly encourage sources and
compliance from a risk-based program compliance obligations, which is a main regulatory authorities to work together
to a technology-based program. Because goal of the Title V program. Further, for to establish an approach that will
we cannot definitively say when a purposes of clarification and as a matter provide for the most practical transition.
RCRA requirement is more stringent of courtesy, we urge regulatory For example, we strongly recommend
than a MACT requirement and authorities that choose to impose dual that sunset provisions be included in a
consistently apply it to all sources, we compliance requirements, to also permit that will be issued well in
are relying on sources and permitting provide a written justification to the advance of compliance with MACT to
agencies to work together to determine source explaining the reasons for their avoid duplication and a later
which requirement is more stringent. If decisions. modification to remove the duplicative
the MACT requirement is determined to RCRA conditions. Also, it would make
2. Transition of Interim Status Phase 2
be more stringent, the permitting agency more sense to transition a source to
Units From RCRA to CAA Permits
can remove the requirement from the MACT compliance prior to issuing the
RCRA permit. In response to our discussion in the RCRA permit if it will comply with
In adopting a permitting approach to proposal regarding RCRA permitting for MACT early.
place the MACT air emissions and interim status Phase 2 units (69 FR
related operating requirements in the 21324), two commenters suggest that 3. Transition From Compliance With the
CAA Title V permit and to keep all EPA establish policy and/or regulation Interim Standards to the Replacement
other aspects of the combustion unit that discourage further RCRA permitting Standards
and the facility in the RCRA permit, our work for interim status Phase 2 sources. A specific question that has been
intent was and still is, to minimize Their comments are directed our raised relates to the applicable
duplication to the extent practicable and statement in the proposal that the RCRA standards and operating parameters that
to eliminate the potential for dual combustion permitting procedures in 40 the source must comply with during the
enforcement. We view it as an CFR part 270 also continue to apply period between the rule’s effective date
unnecessary duplication of effort until you demonstrate compliance. As for the Phase 1 Replacement Standards
between programs as well as an noted in this statement, we intended for and submission of their new NOC. Upon
unnecessary expenditure of resources Phase 2 sources to continue to be the publication date of the rule, the
and costs for both facilities and subject to RCRA permitting Replacement Standards (and Phase 2
regulatory authorities to maintain a requirements for air emissions standards Standards) will become effective and
RCRA permit and a Title V permit that and related operating parameters, sources will have 3 years to come into
contain duplicative requirements, when including trial burn planning and compliance. During this 3-year period,
there are viable mechanisms (i.e., Class testing, until they have demonstrated Phase I sources’ existing title V permits
1 modification procedure at 270.42 compliance with the MACT standards will either be reopened to include the
Appendix I, section A.8, or Class 2 or by conducting a comprehensive Replacement Standards, or the
3 if a state has not adopted the Class 1 performance test and submitting an permitting authority will have
procedure) in place to avoid doing so. NOC to the Agency. However, we also incorporated the Replacement
Nevertheless, we believe that states provided several factors that should be Standards during permit renewal. In
should have the flexibility to decide taken into consideration when this example, a Phase 1 source’s Title V
how they will allocate their resources, determining whether to proceed with permit has been reopened, revised, or
which is why we did not include a the RCRA permit process such as: the renewed and includes the Replacement
single transition approach for facility’s permit status at the time the Standards, the compliance date has not
implementing agencies to follow in the MACT rule becomes final, the facility’s yet passed, no new documentation of
1999 rule or in today’s rule. So, in such anticipated schedule for MACT compliance (DOC) for the replacement
cases where a state agency chooses not compliance, the priorities and schedule standards has been included in the
to adopt the transition language (i.e. the of the regulatory agency, and the level operating record, and the source has not
of environmental concern at a given site yet conducted a comprehensive
245 As an example, a RCRA permit could specify
(69 FR 21324). performance test and submitted a new
a higher minimum operating temperature than what To support their position, the NOC (therefore it still has an NOC
is necessary for the facility to achieve compliance
with MACT. The lower minimum operating
commenters noted that time and containing the operating parameters for
temperature under MACT may be sufficient, unless resources would be conserved and compliance with the Interim Standards).
the RCRA permit authority determines that the duplicative and overlapping activities In the above scenario, the question
higher RCRA temperature is necessary to limit risk could be minimized if Phase 2 sources asked is whether the source should
to an acceptable level for that facility. There should
be a connection between the RCRA limit and
were permitted solely via Title V. Also, comply with the Interim Standards in
protection of human health and the environment they argued that it would avoid the current NOC or the Replacement
when retaining a RCRA limit. expending resources to modify the Standards in the Title V permit. The
source should comply with the Interim into the Title V permit is appropriate, test plan. While this still may be true,
Standards until the compliance date of this does not necessarily obviate the we have learned that sources who
the Replacement Standards. Although need for permit revisions if the material received extensions for testing (so that
the Title V permit now includes the incorporated by reference is they would have an approved plan),
Replacement Standards, the permit will subsequently revised. For more typically commenced their test shortly
also include the Replacement information on incorporation by after approval. Consequently, this still
Standards’ future compliance date. With reference, please refer to the Office of would not allow enough time to review
regard to the transition from the Interim Air Quality Planning and Standards’ and approve the permit modification
Standards NOC to the Replacement ‘‘White Paper Number 2 for Improved before the test begins. Thus, the new
Standards DOC, we are revising the Implementation of the Part 70 Operating Class 1 modification would be of no
regulations at § 63.1211(c) to render the Permits Program’’ (March 5, 1996), benefit to facilities that conduct their
NOC, which documented compliance Section II.E.2.c. This paper can be found tests without an approved test plan, or
with the Interim Standards, inapplicable at: http://www.epa.gov/ttn/oarpg/t5/ to facilities that received extensions and
upon inclusion of the DOC for the memoranda/wtppr-2.pdf. need to begin their tests upon or shortly
Replacement Standards in the operating after approval of the test plan. Also, we
C. Are There Any Changes to the
record by the compliance date. Thus, found one other circumstance where the
Proposed Class 1 Permit Modification
the source will not be placed in a timeframes could be problematic: If a
Procedure?
situation where it must continue to permitting agency has allowed sources
ensure compliance with the operating In the NPRM, we proposed a new to begin pretesting/testing upon
parameters established in the NOC for Class 1, with prior Agency approval, approval of the test plan. Again, a
the Interim Standards, while seeking to permit modification procedure to help source would not be able to have RCRA
comply with the Replacement Standards further minimize potential conflicts permit requirements waived in time to
and operating parameters in its DOC. between the RCRA permit requirements begin its test.
Although it can be assumed that the and MACT requirements. See 69 FR We agree with commenters that the
source would still be able to comply 21384 and proposed § 270.42(k). During proposed requirements in
with its Interim Standard-based NOC implementation of the Interim 270.42(k)(2)(ii) and (iii) do not provide
because the Replacement Standards are Standards for Phase 1 sources, it became any flexibility to waive RCRA permit
the same as or more stringent than the evident that there are two significant limits for sources that (1) do not have
Interim Standards, we believe that the instances where RCRA permit limits an approved test plan but choose to
revision to render the previous NOC may overlap with MACT requirements: conduct their test; (2) are granted an
inapplicable provides a clearer and during initial (and future) performance extension to their test date because they
more sensible approach. testing and during the period between do not yet have an approved test plan;
placement of the documentation of and (3) may begin testing upon approval
4. Changes to Title V Permits compliance (DOC) in the operating of their test plans. Our original intent to
Both the Replacement Standards and record and the final modification of the require prior Agency approval for the
the Phase 2 Standards will necessitate RCRA permit after receipt of the NOC. new Class 1 permit modification
permit reopenings or revisions to some We discussed several existing procedure was to ensure that the
existing title V permits; other permits approaches (e.g., a class 2 or 3 proposed test conditions would be
will incorporate the requirements upon modification, request for approval sufficiently protective when specific
renewal. 40 CFR §§ 70.7 and 71.7 submitted via the RCRA trial burn plan RCRA requirements are waived and that
include the requirements for Title V or coordinated MACT/RCRA test plan, a source has met the regulatory
permit revisions, reopenings, and or through a temporary authorization) requirements for performance test plans.
renewals. Also, approved Title V for addressing these instances, noting We still believe that review and
permitting authorities may have that none provided an optimal solution. approval is an important step; however,
additional requirements. Please refer to All commenters agreed that the new we also believe it should not be a barrier
the appropriate permitting authority and Class 1 modification procedure is the and therefore, should occur in advance
its individual Title V permits program appropriate and most efficient method of a source commencing its performance
to determine the necessary requirements to enable specific RCRA permit test. As a result, we have revised the
and procedures. conditions to be waived during proposed regulatory language in
With respect to incorporating minor instances of overlap referred to above. 270.42(k)(2)(i) to specify that sources
revisions into the Title V permit, one However, a few commenters were submit their permit modification
commenter had asked, for example, concerned with the requirements in requests with their test plans, to allow
whether revisions made to the NOC to proposed § 270.42(k)(2)(ii) and (k)(3), potentially up to one year for approval
reflect minor operating changes could that require sources to submit their (i.e., the performance test plan is due
be incorporated into the permit by permit modification request upon one year before the test is to begin).
reference rather than through the approval of the test plan and the Also, so that approval does not impede
reopening procedures. Determining the requirement for the Director to approve the commencement of the performance
appropriate Title V permit reopening or or deny the request within 30 days, or test, we have revised the proposed
revision requirements is based on the within 60 days with an extension. This language in 270.42(k)(2)(ii) so that the
nature of the change and the source timeframe is feasible only for those Director can choose whether to issue
specific permit terms and conditions, sources that have received approval of approval of the permit modification
and is therefore difficult to generalize. their test plans at least 60 days prior to request contingent upon approval of the
We recommend that sources work with their scheduled date for commencing performance test plan.246 In that respect,
their Title V permit authorities to their performance test. We
determine the appropriate requirements acknowledged the potential 246 In all likelihood, we anticipate that the RCRA
and procedures that are applicable to impracticality of this requirement in the permit authority will have reviewed the
modification request along with the test plans,
any specific situation. However, we proposal, but at the time believed that worked with its Air counterparts and the source to
would like to note that, when few sources, if any, would conduct their resolve any concerns, and have prepared the permit
incorporating requirements by reference performance tests without an approved Continued
the RCRA permit authority would remaining RCRA activities at the facility establishing combustor operations and
continue to have an extra measure of in their permit application (or emissions. Thus, new units are not
assurance in circumstances that may modification request) including subject to the combustor-specific RCRA
demand it. corrective action, general facility permit requirements and performance
standards, other combustor specific standards (i.e., to develop a trial burn
D. What Permitting Approach Is EPA
concerns such as materials handling, plan, provide suggested conditions for
Finalizing for New Units?
risk-based emission limits and operating the various phases of operation in their
1. Why Did EPA Propose a Separate requirements, and other hazardous permit application, and subsequently
Permitting Approach? waste management units. While this operate under those conditions).
As discussed in the proposal, the approach appears to be ideal from the However, because these units remain
current RCRA regulations at §§ 264.340, standpoint of reducing the regulatory hazardous waste treatment units, they
265.340, 266.100, 270.19, 270.22, burden to sources and RCRA permit are still required to obtain a RCRA
270.62, and 270.66 do not address how authorities, we noted that even though permit, or to modify an existing RCRA
or when new combustion units will a new unit will be required to meet the permit to include a new unit, prior to
comply with the MACT standards. RCRA public participation requirements construction. They need only address
Consequently, the part 270 regulations as part of the permit application the remaining hazardous waste
imply that a new unit must obtain a process, the operations and emission management activities at the facility in
complete RCRA permit before it can information specific to the combustor their permit application (or
demonstrate compliance with the would no longer be provided. Thus, we modification request) including
MACT standards. It was never our focused on certain compliance activities corrective action, general facility
intent for new units to develop a trial under the MACT/CAA framework (i.e., standards, other combustor specific
burn plan and provide suggested the Notification of Intent to Comply concerns such as materials handling,
conditions for the various phases of requirements) that would allow for risk-based emission limits and operating
operation in the RCRA permit combustor-specific information to be requirements, and other hazardous
application, given that these conditions made available to the public as it would waste management units. As we noted
will become inactive or need to be have been under the full RCRA permit in the previous section and will discuss
removed from their permits upon process. again more thoroughly in the next
demonstrating compliance with MACT. Regarding the three additional section, we are relying on the NIC
To rectify our previous omission, we approaches or ‘‘options’’, each process to provide the public with the
suggested several options that would considered a different point in the combustor-specific information that
allow units newly entering the RCRA RCRA permit process where a new unit previously would have been provided
permit process 247 (and that will comply could ‘‘transition’’ to compliance with under the full RCRA permit process.
with the Subpart EEE requirement upon the MACT standards (see 69 FR 21319).
Under the first option, a new unit could Almost all commenters supported our
start-up) to forego certain RCRA permit preferred approach to not require that
requirements and performance transition to MACT compliance after it
had submitted its RCRA Part B new units complete the full RCRA
standards. In developing the options permit process and to rely on the NIC
that would enable new units to forego application. The Part B however, would
not include the trial burn plan requirements and the MACT/CAA
certain RCRA requirements, we noted framework to provide a level of public
the importance of public participation information. The new unit would only
be required to discuss the compliance participation that is commensurate with
opportunities under the MACT/CAA the requirements under RCRA.
framework equivalent to those provided activities related to the combustor as
part of the RCRA informal public Commenters generally agreed that our
under the RCRA framework. Thus, each preferred approach achieves this goal
option was constructed in such a way meeting. In the second option, we
proposed that a new unit would while streamlining the RCRA permit
that would streamline the RCRA process for new units. One commenter
transition after its RCRA permit has
requirements, but continue to provide felt that the Title V and New Source
been issued. Here, the new unit would
early and frequent public participation Review programs (NSR) provide
be required to develop a trial burn plan
commensurate with the requirements of sufficient requirements to regulate new
which provided its proposed operations
the RCRA Expanded Public combustion units. We disagree that
and emissions information and to
Participation Rule (60 FR 63417, either or both of those programs fully
discuss its compliance activities via the
December 11, 1995). address the hazardous waste and public
RCRA informal public meeting. Then, a
2. What Options Did EPA Propose for permit would be issued, but it would participation components
Permitting New Units? not contain operating and emissions commensurate with that provided by
requirements in order to avoid a future the approach we are finalizing today.
In our preferred approach, we For instance, a unit may be constructed
proposed that new units not be required modification to remove them. For the
third option, the transition point would and operating before a Title V permit is
to develop a trial burn plan and provide issued, which directly conflicts with
suggested conditions for the various have been after the new unit places the
DOC in its operating record, which is RCRA’s early public participation
phases of operation in their RCRA requirements. Also, in some instances,
permit application. Instead, new units the compliance point for MACT. This
option is more inclusive than the public participation may not be a
would only be required to address the required component of state issued NSR
second because it requires the new unit
to have a draft permit that covers the permits (see footnote regarding public
modification approval prior to issuance of the test
plan approval. construction and shakedown period. participation and SIPs below). However,
247 Units ‘‘newly’’ entering the RCRA permit we do believe that the NSR program will
process refers to a newly constructed facility, thus 3. Which Option Is EPA Finalizing? play an important role regarding the
newly constructed hazardous waste combustion For today’s final rule, we are adopting exchange of information, as we will
unit; an existing facility that constructs a new unit;
or an existing facility that converts a non-hazardous
our preferred, proposed approach: new discuss in the section below. With
fuel combustion unit to a hazardous waste fuel units will not be required to follow the respect to the remaining three options
combustion unit. full RCRA permitting process for presented in the proposal (69 FR 21319–
21320) that suggested a transitional The steps associated with obtaining an meeting that encompasses the NIC, draft
approach (i.e., each option explored NSR permit, or a ‘‘pre-construction’’ CPT plan, and RCRA pre-application
progressive points in the RCRA permit permit, are similar, but not necessarily notice information, the public will be
process where facilities could transfer identical to that required under RCRA. provided with all information related to
over to MACT without fully completing They are: (1) Preparation of the permit the combustor’s compliance plans as
the RCRA process), nearly all application (sources must provide the well as its operating plans and
commenters were in agreement that they location, design, construction, and emissions estimates prior to burning
would require more work to implement operation information) and participation hazardous waste. See new requirements
than is necessary and consequently in pre-application meetings; (2) issuance in § 63.1212.
oppose them. of permit application completeness With respect to the requirements we
determination by the State; (3) are finalizing today, we received only
4. How Will Permitting for New Units development and negotiation of draft one comment that expressed concern.
Work? permit; (4) opportunity for public notice The concern is that the requirement to
In the proposed rule, we created an and comment on the draft permit; (5) submit the CPT plan is too early in the
approach that utilizes the NIC response of permitting authority to compliance process. For example, the
requirements and the MACT/CAA public comments; (6) possible RCRA application is submitted
framework with the intent of ensuring administrative and judicial appeals; and approximately 2–3 years before start-up
that the requirements of the RCRA (7) permit issuance/denial.249 whereas the CPT plan is required 1.5
Expanded Public Participation Rule A second component is that the NIC years after the final NIC is due.250 The
does not provide the information on the commenter feels that the facility would
would continue to be fulfilled. The four
proposed combustor operations or not have enough time to learn about the
requirements for public participation as
emissions information that would ‘‘detailed nuances of the system’’.
they relate to hazardous waste
normally be available as part of the However, the commenter does note that
combustion units are: (1) Permit
RCRA process. To address these gaps it is possible to submit the CPT plan,
applicants must hold an informal public
between RCRA and MACT, we are but it will not be as complete or refined
meeting before applying for a permit; (2)
requiring an approach similar to that as it would be if it was submitted
permit agencies must announce the
which was proposed. New sources according to the deadline for existing
submission of a permit application
must: (1) Prepare a draft NIC and make units. We agree with the commenter that
which will tell community members
it available to the public at the same a considerable amount of planning is
where they can view the application time as their RCRA pre-application required of the source to be able to draft
while the agency reviews it; (3) meeting notice; (2) provide a draft of the CPT plan at such an early stage, but
permitting agencies may require a their comprehensive performance test we are only requiring that a draft of the
facility to set up an information (CPT) plan (to the public) to coincide CPT plan be made available, with the
repository at any point during the with the draft NIC and RCRA pre- final CPT plan due 6 months prior to the
permitting process if warranted; and (4) application meeting notices; and (3) source’s compliance date. Moreover, at
permitting agencies must notify the hold their NIC public meeting with their this early stage, we liken the
public prior to a trial (or test) burn. RCRA informal public meeting. The first development of the draft CPT plan to
As discussed in the preamble to the two requirements ensure that the public the development of the trial burn plan.
proposal (69 FR 21318), we believe that is provided with most of the same Even though it may not be as complete
the NIC process addresses the first two information that would have been or refined as it will be when the final
RCRA public participation available via the RCRA trial burn plan CPT plan is due, we believe that it will
requirements. The NIC process requires prior to the source burning hazardous still be of benefit to the public and the
a source to make its draft NIC, which waste. Other information not required regulatory authority, but also to the
discusses the source’s plan for coming by the NIC or CPT plan, such as the source in terms of advance planning for
into compliance with the MACT combustion unit’s design specifications the design of the unit through start-up
standards, available for public review will, in most cases, be available to the of the unit.
and to hold an informal public meeting public through the NSR permit The components thus far, have
to discuss the activities contained in the application. We recommend that satisfied the first (2) two RCRA public
NIC. While the NIC process gives the sources submit a copy of their NSR participation requirements. The third
public an early opportunity to permit application to the RCRA permit RCRA public participation requirement
participate in the unit’s compliance authority so that this information is enables a regulatory authority to
planning process early on, a few readily available for development of the evaluate the need for and require a
components are still missing before we RCRA permit. The third requirement facility to establish and maintain an
can consider the first 2 RCRA public allows the public to inquire and information repository. The
participation requirements to be comment on both the new unit’s establishment of an information
fulfilled under the MACT framework. proposed activities and operations. By repository is typically required only
One component is that there is no requiring new sources to develop, when there are concerns or unique
permit action associated with the NIC notice, and hold a combined public information needs of a community. The
requirements. However, the NSR purpose of the information repository is
program can provide a permit be required to obtain an NSR permit if its potential to make information regarding the
to emit does not exceed the NSR threshold level. facility (and combustion unit) available
mechanism that will determine whether 249 With respect to numbers 4 and 5, many States
or not a source may be constructed.248 to the public during the permit issuance
omitted the public participation steps in their
federally approved SIPs. This was the reason why
process and during the life of the
248 We believe that the majority of new units will Sierra Club had been opposed to our efforts to permit. In the preamble, we noted that
be classified as major sources for NSR permitting simply rely on NSR permitting to provide public
(requiring either prevention of significant participation opportunities that would have been 250 Comprehensive performance test plans are
deterioration or nonattainment permits), however, otherwise provided under the traditional RCRA required to be submitted one year in advance of the
those that do not, will likely be required to obtain permit process for new units. Today, however, scheduled test. The submittal date would be as late
a minor NSR permit. In few cases, new sources (e.g., many SIPs have been revised to address public as 2.5 years after the effective date of the rule
newly constructed as opposed to modified) may not participation requirements. assuming no extensions are granted.
although the Title V permit process regulations and thus, no additional issue a draft permit. The public will
contains a provision that any materials regulatory revisions or amendments are then have an opportunity to comment
relevant to the permit decision be made needed. on the draft permit and request a public
available to interested persons (see hearing. Upon resolution of any issues
4.a. Process for New Units Seeking an
§ 70.7(h)(2) and § 71.11(d)), the surrounding the draft permit, a final
Initial RCRA Permit
information may not be made available RCRA permit will be issued. The RCRA
until well after the combustor is We anticipate that the process for new process is the same as before, but should
constructed and operating. units seeking an initial permit will work be reasonably shorter. Finally, the new
Consequently, we have chosen to adopt as follows. Any new unit would begin unit may begin burning hazardous waste
additional provisions under the NIC the process by developing and when it can assure it will operate in
requirements that parallel the compiling the information necessary for compliance with the MACT standards
requirements of § 124.33. the RCRA draft permit (e.g., information (i.e., by placing a documentation of
We had proposed two options that required for the part A application at compliance in its operating record on
would allow a regulatory authority to § 270.13, the relevant general the day it begins burning hazardous
require, on a case-by-case basis, a source information for the part B application waste). See new regulatory language at
to establish an information repository according to Part 270) and the § 63.1212(c). To aid readers in
specific to the combustor. The first applicable NSR permit.251 The understanding the above process, we
option was to place such a provision in information needed to compile the draft have included a pictorial timeline.
the NIC regulations and the second NIC and draft CPT plan would be Please see figure 2.
option was to amend the applicability gathered simultaneously, as if the Finally, it may also be feasible to
language in § 124.33 to include source were developing the trial burn combine an NSR pre-application
combustion sources that will comply plan. When the source has compiled its meeting and public notice of the draft
with Part 63, subpart EEE upon start-up. RCRA permit application, draft NIC and NSR permit with the process described
Two commenters felt that the second draft CPT plan, it would submit a RCRA above. Thus, we recommend that
option would create problems as far as pre-application meeting notice at least sources work closely with their Air and
organization (i.e., by modifying the 30 days prior to the date scheduled for RCRA permit agencies so that the NSR
RCRA regulations to include a provision the RCRA informal public meeting public notices and meetings may be
solely for new units complying with according to §§ 124.31(b) and (d). At the coordinated with the RCRA and NIC
MACT). We agree that the second option time of the RCRA pre-application notices and meetings so time and
could be confusing and that it would be meeting notice, the source would also resources are efficiently utilized.
more appropriate to keep all new issue notice of the NIC public meeting
(at least 30 days prior to the NIC 4.b. Process for New Units Modifying an
requirements for new units in one set of
meeting) according to § 63.1210(c)(3), so Existing RCRA Permit
regulations. Therefore, we are finalizing
a provision that will allow for an that the two meetings can occur at the The process of adding a new unit to
information repository to be established same time. In order for the public to be an existing permit is accomplished
specific to the combustor (recall that a able to view all information relevant to through a Class 3 permit modification
repository established pursuant to the the combustor before the combined (see § 270.42 (c) for requirements). The
RCRA permit will include documents RCRA pre-application and NIC public requirements governing public notices
relevant to the facility only), if deemed meeting, the source would make the of the draft NIC, draft CPT plan, and
appropriate, under the NIC regulations. draft NIC and draft CPT plan available holding a combined public meeting are
See new § 63.1212(c). Under the NIC to the public for review at the same time essentially the same as new units
regulations, the repository could the notices for the meetings are issued. seeking an initial permit. The process is
include the NIC, test plans, draft Title To aid the RCRA permit authority in its as follows. The source prepares and
V permit and application, reports, et development of the draft RCRA permit submits its RCRA permit modification
cetera. (i.e., mainly for purposes of evaluating request (and if applicable, NSR
The fourth and final RCRA public risk), we strongly recommend that the application). It must then publish a
participation requirement to be fulfilled source also provide copies of the draft notice of the modification request seven
is for the regulatory authority to notify NIC, draft CPT plan, and NSR days later, followed by a public meeting
the public of an impending trial burn or application (if applicable) to the RCRA no earlier than 15 days after publication
test burn. As discussed in the RCRA permit authority. It is our hope that the of the notice for the modification
Expanded Public Participation Rule, the availability of information will expedite request, and no later than 15 days before
RCRA permit authority will typically the development of the draft permit. All the close of the 60-day comment period.
provide the notice at least 30 days in notices should be presented to the As with new units that are submitting
advance of the test (60 FR 63426, public in sufficient time to allow for a an initial RCRA permit application, it is
December 11, 1995). Similarly, the combined RCRA informal public also important for sources seeking to
MACT regulations require an existing or meeting and NIC public meeting. modify their permit to coordinate their
new unit to provide notice to the public Following the combined public NIC public meeting with their RCRA
that the CPT plan (and the continuous meeting, the source will submit its permit modification public meeting.
monitoring system performance RCRA permit application and the RCRA This is made possible due to the
evaluation test plan) is available for regulatory authority will prepare and flexibility of the NIC public meeting; it
review. The regulations in can be held any time prior to the 10
§ 63.1207(e)(2) fulfill this requirement. 251 Because the information required for NSR
month deadline. After the combined
Although the CPT plan may not be permit is less comprehensive than a RCRA permit, public meeting and the close of the
it allows for a much shorter time period for
approved before the public is notified, issuance. The average time for issuing a PSD comment period, the permit authority
the intent is to provide notice to the permit, for example, after receiving an application will either grant or deny the
public of a future test. We believe that is slightly more than 7 months, but varies modification request. If approved, the
depending upon public involvement and
the MACT regulations provide public negotiation of the application content. USEPA.
source may then begin construction or
notice of the test plans that are Docket A–2001–19, Document II–A–01. NSR 90-Day modification of the unit. To aid readers
commensurate with the RCRA Review Background Paper, June 22, 2001. in understanding the timing of the
above process, we have included a streamlined RCRA permit modification MACT standards for Phase 2 major
pictorial timeline. Please see figure 2. procedure to allow sources to make sources or continue complying with the
Again, it may be feasible to combine upgrades to comply with MACT RCRA standards and requirements of
an NSR pre-application meeting and (§§ 270.42(j) and 270.42 appendix I, their RCRA permit.
public notice of the draft NSR permit section L.9), a second streamlined RCRA In the 2004 proposal, we stated that
with the process described above. Thus, permit modification procedure to we were not making a positive area
we recommend that sources work remove conditions from a permit that source finding for Phase 2 area sources
closely with their Air and RCRA permit are no longer applicable (§ 270.42
as we have for Phase 1 area sources (69
agencies so that the NSR public notices appendix I, section A.8), an addition to
and meetings may be coordinated with FR 21212 and 21325). Regardless of this,
§ 270.235 to specify conditions for start-
the RCRA and NIC notices and meetings however, the Phase 2 area sources are
up, shutdown, and malfunction plan
so time and resources are efficiently still subject to the requirement to obtain
and integrate them with the CAA
utilized. program, and an amendment to the a Title V permit because they are subject
interim status regulations at § 270.72 to to section 112 standards under this
E. What Other Permitting Requirements subpart. See § 502(a) of the CAA and 40
Were Discussed in the Proposal? exempt interim status facilities from the
reconstruction limitation when making CFR §§ 70.3(b)(2) and 71.3(b)(2).
At proposal, we discussed where most upgrades to comply with MACT. It is important to note that the Title
Phase 1 sources would be in terms of Also, we are finalizing three new V applications for the Phase 2 area
their transition from their RCRA permit permitting changes that are applicable sources will need to contain emissions
requirements to compliance with the to both Phase 1 and 2 sources. Two have information relative to all regulated air
MACT Interim Standards (see 69 FR been discussed previously in this pollutants (to determine applicable
21321). The transition process was section and are: (1) A new streamlined requirements, fees, etc.) that are being
discussed with respect to both the RCRA permit modification procedure emitted from the units subject to the
RCRA permit and the Title V permit. designed to reduce overlap during the MACT standards, not just the specific
However, when we discussed the Title transition from RCRA to MACT HAP pollutants regulated by the MACT
V permit requirements in the proposal, (§§ 270.42(k) and 270.42, appendix I, standards (see §§ 70.5(c)(3)(i) and
we did not elaborate on the transition L.10); and (2) regulatory provisions 71.5(c)(3)(i)). Although, the permit itself
between the Interim Standards and stating that new units are no longer would contain standards only for the
Replacement Standards. Because we subject to the full array of RCRA HAP subject to MACT standards (the
believe that such a discussion would be combustion permitting requirements. § 112(c)(6) HAP). A Phase 2 area source
helpful to readers, we have included The third change is discussed above in which chooses to control hydrogen
general information describing how the Section IX. Site-Specific Risk chloride, chlorine gas, and metals other
transition process would work for most Assessment Under RCRA and finalizes than mercury by continuing to comply
sources in Section B. Did Commenters our response to a petition for with the relevant RCRA standards and
Express any Concerns Regarding the rulemaking with respect to site-specific the requirements of its RCRA permit
Current Permitting Requirements?, risk assessments (SSRAs). As part of this
subsections 3 and 4. should note this choice in its Title V
change we have decided to adopt application and cite to the relevant
For Phase 2 sources, we proposed the regulatory language that specifically
same permitting approach as we did for requirements of this subpart. This will
provides clarification of authority for help ensure that the permitting
Phase 1 sources. Today, we are RCRA permit writers to evaluate the
finalizing as proposed, the following for authority is aware that these
need for and, where appropriate, require
Phase 2 sources: (1) the new Phase 2 requirements apply in lieu of the MACT
SSRAs and to add conditions to RCRA
emissions standards will be placed only standards for Phase 2 major sources.
permits that they determine, based on
in the CAA regulations at 40 CFR part The permitting authority should also
the results of an SSRA, are necessary to
63, subpart EEE, and be implemented protect human health and the document this choice in the statement
through the air program; (2) with few environment. of basis for the source’s Title V permit.
exceptions, the analogous standards in Last, as explained in part four section See §§ 70.7(a)(5) and 71.7(a)(5). Finally,
the RCRA regulations no longer apply II.A, we are finalizing our decision to for the units at a source which are
once a facility demonstrates compliance regulate emissions of dioxin/furans, subject to the subpart EEE MACT
with the MACT standards in subpart mercury, polycyclic organic matter, and standards, all CAA applicable
EEE and any duplicative requirements polychlorinated biphenyls from Phase 2 requirements to which these units are
have been removed from the RCRA area sources under section 112(d).252 subject, e.g., State Implementation Plan
permit; and (3) the new standards will This means that Phase 2 area sources are requirements, not just the relevant
be incorporated into operating permits subject to MACT standards only for Subpart EEE requirements, must be
issued under Title V of the CAA rather these hazardous air pollutants (HAP) in included in the Title V permits issued
than be incorporated into RCRA the final rule. To reiterate, they are: to these sources. See §§ 70.3(c)(2) and
permits. Consequently, we are finalizing Dioxin/furans, mercury, and polycyclic 71.3(c)(2). For more information
the proposed changes to §§ 270.22 and organic matter (controlled by the regarding § 112(c)(6) and how it relates
270.66 to implement the above. Also surrogates DRE and carbon monoxide/ to Phase 2 area sources, see Part Four,
applicable to Phase 2 sources via today’s hydrocarbon). For the remaining HAP Section II.A., ‘‘Area Source Boilers and
final rule are the changes and additions (hydrogen chloride and chlorine gas and Hydrochloric Acid Production
we finalized in the 1999 final rule for metals other than mercury), Phase 2 area Furnaces’’.
Phase 1 sources. These include a sources may either comply with the BILLING CODE 6560–50–P
252 As explained in the Comment Response would be (among other things) ‘‘unnecessarily of affording opportunity for public participation as
Document vol. V, although § 502(a) allows EPA to burdensome’’, we believe that Title V requirements provided for in the Title V permit issuance process.
exempt area sources from title V permitting remain appropriate for these sources given the
requirements if EPA finds that those requirements highly toxic nature of the HAP and the importance
ER12OC05.004</GPH>
ER12OC05.005</GPH>
Part Five: What Are the CAA more about the advantages of receiving following paragraphs will explain
Delegation Clarifications and RCRA delegation authority. which authorities in Subpart EEE are
State Authorization Requirements? Also, we would like to point out that delegable and are not delegable to
there are several delegation options that S/L/T agencies that have been delegated
I. Authority for This Rule
S/L/T agencies can receive. Regardless, authority and will provide some
Today’s rule amends the promulgated many S/L/T agencies choose the examples of or references to alternative
standards located at 40 CFR part 63, ‘‘straight delegation’’ option when requests associated with each delegable
subpart EEE. It amends the standards for applying for delegation approval. or non-delegable provisions authority.
the Phase 1 source categories— Straight delegation means that these To review, the regulations at 40 CFR
incinerators, cement kilns, and agencies have agreed to implement and 63.90 define three types of alternative
lightweight aggregate kilns that burn enforce federal MACT standards as they requests. Alternative requests or
hazardous waste, and it also amends have been written in the promulgated ‘‘changes’’ to a particular delegable or
subpart EEE to establish MACT requirements. As a result, many EPA non-delegable provision are classified as
standards for the Phase 2 source Regions and states have established major, intermediate, or minor
categories—boilers and hydrochloric memoranda of agreement that depending upon the degree (i.e.,
acid production furnaces that burn essentially provide automatic delegation potential to be nationally significance,
hazardous waste. Additionally, this rule of each future MACT, as opposed to the potential to reduce the stringency of the
amends several RCRA regulations state applying for delegation of each standard, etc.) of change being
located in 40 CFR part 270 to reflect future MACT, which requires a requested. An alternative request that
changes in applicability, addition of a rulemaking to implement. For more qualifies as a major change is not
new permit modification procedure, and information related to the delegation delegable to S/L/T agencies, even when
additions related to site-specific options and procedures, please refer to they have delegation authority. These
assessments and permitting. the fact sheet, Clean Air Act Delegation requests must be sent to the EPA Region
II. CAA Delegation Authority for the HWC NESHAP at: http:// or, if it concerns a test method under
www.epa.gov/epaoswer/hazwaste/ §§ 63.7(e)(2)(ii) and (f), 63.1208(b) and
Before discussing the clarifications 63.1209(a)(1) or a standard under
combust/toolkit/factshts.htm and EPA’s
being finalized today, it is important to §§ 63.1200, 63.1206(a), or 63.1216–
first highlight a few key aspects of delegation website at: http://
www.epa.gov/ttnatw01/112(l)/112- 63.1221, then it must be sent to our
delegation authority. Recall from the Office of Air Quality Planning and
proposal that a state, local, or tribal (S/ lpg.html.
Standards (OAPQS).255 An alternative
L/T) agency must be delegated authority III. Clarifications to CAA Delegation request that qualifies as an intermediate
under CAA section 112(l) before it can Provisions for Subpart EEE or minor change is delegable. However,
exercise the delegable provisions’ In the proposal, we discussed the the EPA Region may choose whether or
authorities. The delegable authorities need to provide additional clarification not they will delegate authority to
can be found in 40 CFR 63.91(g)(1)(i), for the delegable and non-delegable S/L/T agencies to approve intermediate
also known as Category I Authorities. A authorities within Subpart EEE based and, even some minor changes during
S/L/T agency that has applied for and upon our implementation experience the delegation approval process. In
received delegation authority can with the Phase 1 Interim Standards and addition to the regulations, the guidance
approve: test plans, requests for minor the Clarifications to Existing National document entitled, How to Review and
and in most cases, intermediate changes Emissions Standards for Hazardous Air Issue Clean Air Act Applicability
to monitoring and test methods, Pollutants Delegation’ Provisions final Determinations and Alternative
performance test waivers, and several rule published on June 23, 2003 (68 FR Monitoring (EPA 305–B–99–004,
other Category I Authorities. Please note 37334). Although the June 23, 2003 final February 1999) provides a listing of
that even though a S/L/T agency may rule provided clarification and delegable and non-delegable authorities
have an approved Title V permit streamlined the delegable provisions for in Tables 1 and 2, as well as
program, it cannot exercise delegable each existing NESHAP, it overlooked descriptions and examples of major,
authorities or be the primary several non-delegable and delegable intermediate, and minor changes in
enforcement authority if it has not authorities within Subpart EEE. It Attachment 1.
received delegation authority under provided clarification on the non-
CAA section 112(l). Moreover, when a A. Alternatives to Requirements
delegable authorities of Subpart EEE as
S/L/T agency has not taken delegation they relate to major alternatives to the Any change to a promulgated
of a section 112 standard, the agency standards themselves and to test standard is considered a major change
can only incorporate the section 112 methods, monitoring, or recordkeeping and as noted above, must be sent to
standard’s requirements into its Title V and reporting under the General OAQPS (see contact information in
permits, (and then implement and Provisions.254 However, it omitted footnote). The reason why a change to
enforce these requirements through its a standard must be sent to EPA
major alternatives specific to Subpart
title V permits) when it has adequate Headquarters is because the change
EEE such as: test methods under
authority under State, local, or tribal must be established through national
§§ 63.1208(b) and 63.1209(a)(1);
law which allows it to conduct the rulemaking, regardless of the degree of
monitoring under § 63.1209(a)(5) and;
above actions without delegation. See, change sought. Thus, only OAQPS can
recordkeeping and reporting under
e.g., the proposed Federal Plan for approve alternative requests for changes
§ 63.1211(a) through (d). Therefore, the
Commercial and Industrial Solid Waste to standards. Additionally, any change
Incinerators, November 25, 2002 (67 FR 254 For example, the final rule included approval to applicability requirements and
70640, 70652). Please also refer to 69 FR of alternatives to requirements in §§ 63.1200, compliance dates (e.g., requirements
21335 of the proposal and the fact sheet 63.1203, through 63.1205, and 63.1206(a); approval that ensure that the standards are
entitled, Clean Air Act Delegation for of major alternatives to test methods under
§ 63.7(e)(2)(ii) and (f); approval of major alternatives achieved as EPA intended) are also
the HWC NESHAP at: http:// to monitoring under § 63.8(f) and; approval of major
www.epa.gov/epaoswer/hazwaste/ alternatives to recordkeeping and reporting under 255 For contact information, please visit
combust/toolkit/factshts.htm to learn § 63.10(f). www.epa.gov/ttn/emc/staffdir.html.
considered major and also must be sent use to determine compliance with we view requests per § 63.1209(a)(5) to
to OAQPS for approval. Specific to subpart EEE. Also, we are adding the use PM CEMS as intermediate changes
Subpart EEE, alternative requirement CEMS monitoring requirement under to monitoring. Although the
requests including those pursuant to § 63.1209(a)(1). It is regarded as a test implementation of PM CEMS according
§§ 63.1200, 63.1206(a), or 63.1216– method because it serves as a to PS–11 (69 FR 1786 and 40 CFR part
63.1221 are considered major changes benchmark method for demonstrating 60, Appendix B; January 12, 2004) and
and consequently are non-delegable. compliance with the emission Procedure 2 (see also 40 CFR part 60,
The regulations at § 63.1214(c) correctly standards. Both sections are delegable to Appendix F) is largely ‘‘self-
identified the requirements in Subpart S/L/T agencies as long as they have implementing,’’ sources wishing to
EEE, however we have revised them been delegated authority and as long as apply to use of PM CEMS should
today (as we proposed) to reflect the the alternative requests comprise minor develop and submit QA/QC plans
new sections that house the Phase 1 or intermediate changes. However, a specifying audit frequencies to account
Replacement Standards and Phase 2 major change to either of these test for site-specific stack conditions. We
Standards. method sections must be sent to OAQPS believe that other site-specific issues
There are a few exceptions to the for approval.257 Only OAQPS can that may need to be addressed prior to
above, however. Subpart EEE approve major changes to test methods use of the CEMS, such as a source’s
incorporates specific provisions for because they are designated in the request to deviate from PS–11 or a
sources to request alternative standards standard as the means for determining source’s selection of the correct
which are delegable because they have compliance with an emission standard. correlation curve(s), are properly
been established through rulemaking. In The proposed revisions to § 63.1214 are addressed under EPA’s established
fact, several alternative standards are finalized today to include major policies and procedures for alternative
self-implementing meaning that the alternatives to test methods under method requests. We believe that a
source only need specify in their DOC §§ 63.1208(b) and 63.1209(a)(1) as non- petition to use PM CEMS under § 63.8(f)
which standard it will comply with. The delegable authorities. is still the appropriate mechanism, but
alternative to the particulate matter that sources can submit their petitions
standard in § 63.1206(b)(14) and the C. Alternatives to Monitoring
to their delegated S/L/T agency for
emissions averaging standards for For monitoring, the final delegations review and approval, and we
cement kilns with in line kiln raw mills rule stated that major alternatives to recommend that EPA Regional offices
and preheater or preheater/precalciner monitoring at § 63.8(f) were not work with these agencies to monitor
kilns with dual stacks in § 63.1204(d) delegable, but did not reference
implementation. Thus, with the
and (e) are three examples. There are monitoring specific to subpart EEE. In
exception of petitions to use PM CEMS
also alternative standards that sources subpart EEE, the monitoring
in lieu of an operating parameter which
may petition to comply with. They requirements are located in § 63.1209.
is considered an intermediate change,
include: Alternatives to the standards This section also includes two
we are finalizing our proposed revision
for existing and new LWAKs at provisions specific to alternative
to § 63.1214(c) to include major
§ 63.1206(9) and cement kilns at monitoring, thus removing some of the
alternatives to monitoring under
§ 63.1206(b)(10) and the alternative risk- ‘‘guesswork’’ when trying to discern
§ 63.1209(a)(5) as a non-delegable
based standard for total chlorine at whether a request for change is minor,
authority.
§ 63.1215. Sources choosing to comply intermediate, or major. One is located at
§ 63.1209(a)(5), Petitions to use CEMS Section 63.1209(g)(1), Alternative
with these alternative standards must
for other standards and the other is monitoring requirements other than
receive approval from their delegated S/
located at § 63.1209(g)(1), Alternative continuous emissions monitoring
L/T agency prior to implementing
monitoring requirements other than systems, contains the other alternative
them.256 With respect to changes to
continuous emissions monitoring monitoring provision. This provision
compliance dates, requests under
systems. Each is discussed in the allows sources to request alternative
§ 63.1213 specifically allow sources to
following paragraphs. monitoring methods to monitor
request an extension to the compliance
In the proposal, we explained that a compliance, except for those standards
date for the installation of pollution
request to use other monitoring in lieu that must be monitored with a CEMS
prevention or waste minimization
of a CEMS is always considered a major (e.g., those in § 63.1209(a)(1)), and to
controls. Again, because this provision
change due to CEMS generally being request a waiver of an operating
has been specified in subpart EEE, it is
considered a more accurate measure of parameter limit. We provided several
not considered a major change and is
compliance. However, if a source examples of alternative parameter
delegable.
requests to use a CEMS in lieu of a monitoring for which a request may be
B. Alternatives to Test Methods required operating parameter, it may be submitted under this section in the
With respect to test methods, we considered an intermediate change. proposal at 69 FR 21337. They include
noted above that the final delegations Since publication of the proposal, use of: a different detector, different
rule stated that major alternatives to the performance specifications have been monitoring location, a different method
test methods at §§ 63.7(e)(2)(ii) and (f) promulgated for PM CEMS (and as recommended by the manufacturer,
were not delegable. Therefore, as we mercury CEMS).258 Consequently, today or a different averaging period that is
proposed, it is necessary to add major more stringent than the applicable
alternatives to 63.1208(b), which
257 For contact information, please visit
standard. In the proposal, we stated that
www.epa.gov/ttn/emc/staffdir.html. we believe the majority of requests
specifies the test methods sources must 258 Although performance specifications have
been promulgated for mercury CEMS, there has not

submitted pursuant to § 63.1209(g)(1)
256 The alternative risk-based standard for total been as much experience in implementing these are not major and discussed in the
chlorine at § 63.1215 requires sources to submit devices for hazardous waste combustion sources (or preamble amending the language in
their eligibility demonstration to both the delegated similar sources) as there has been for PM CEMS at § 63.1209(g)(1) so that these types of
S/L/T agency and to the Risk and Exposure this time. Therefore, we believe it appropriate to
Assessment Group in Research Triangle Park, NC continue sending requests to use mercury CEMS in
changes could be reviewed and
for review, even though the delegated S/L/T agency lieu of an operating parameter to the appropriate approved by the delegated S/L/T
can grant or deny approval. EPA Region for review and approval. agency. However, when we added
language to § 63.1209(g)(1) to allow for E. Other Delegation Provisions In contrast, under RCRA section
the above, we inadvertently referred to 3006(g) (42 U.S.C. 6926(g)), which was
an approved Title V program instead of Although not discussed in the added by HSWA, new requirements and
a S/L/T agency which has taken proposal, it is important to note that prohibitions imposed under HSWA
delegation of subpart EEE. We have issuing applicability determinations is authority take effect in authorized states
another delegable authority. The EPA at the same time that they take effect in
corrected and finalized the proposed
document How to Review and Issue unauthorized states. EPA is directed by
language. Therefore, whether minor or
Clean Air Act Applicability the statute to implement these
intermediate, requests under
Determinations and Alternative requirements and prohibitions in
§ 63.1209(g)(1) may be sent to your
Monitoring (EPA 305–B–99–004, authorized states, including the
delegated S/L/T agency for review and February 1999) provides guidance
approval. issuance of permits, until the state is
regarding who has the lead for issuing granted authorization to do so. While
Please note that 63.1209(g)(1) cannot applicability determinations. In general, states must still adopt HSWA related
be used when requesting major changes Regions may delegate the authority to provisions as state law to retain final
to the monitoring required by the issue applicability determinations to S/ authorization, EPA implements the
standard. Such changes typically L/T agencies when the determinations HSWA provisions in authorized states
involve new unproven monitoring are routine in nature. However, until the states do so.
methods. Unproven monitoring delegation of authority for certain Authorized states are required to
methods refer to those where the applicability determinations should be modify their programs only when EPA
technology or procedures are not retained by the Regions. These include enacts federal requirements that are
generally accepted by the scientific applicability determinations that: (1) more stringent or broader in scope than
community (§ 63.90(a)). If you are Are unusually controversial or complex; existing federal requirements. RCRA
uncertain whether your request (2) have bearing on more than one state section 3009 allows the states to impose
constitutes a new unproven monitoring or district (are multi-Regional); (3) standards more stringent than those in
method, which is considered a major appear to create conflict with previous the federal program (see also 40 CFR
change, you should submit your request policy or determinations; (4) are a legal 271.1). Therefore, authorized states may,
to your EPA Region. The regulatory issue which has not previously been but are not required to, adopt federal
language in 63.1209(g)(1) has been considered (a matter of first impression); regulations, both HSWA and non-
or (5) raise new policy questions. It is HSWA, that are considered less
revised to reflect this clarification.
recommended that Regional offices stringent than previous federal
D. Alternatives to Recordkeeping and require notification when S/L/T regulations.
Reporting. agencies issue applicability We discussed in the proposal which
determinations. RCRA regulations we intended to
As with the others, the final amend and their impact on state
delegation provisions’ rule only cited IV. RCRA State Authorization and authorization procedures. Today, we are
the waiver of recordkeeping and Amendments to the RCRA Regulations finalizing those amendments in
reporting requirements of § 63.10(f) as a Under section 3006 of RCRA, EPA §§ 270.10, 270.22, 270.32, 270.42,
non-delegable provision. Thus, it is may authorize qualified states to 27062, 270.66, and 270.235. In addition,
necessary to add the relevant subpart administer their own hazardous waste we are amending the regulations in
EEE recordkeeping and reporting programs in lieu of the federal program §§ 264.340 and 266.100 to reflect
requirements of § 63.1211. Section within the state. Following changes that have been made based
63.1211 is delegable in its entirety to S/ upon comments. Today’s amendments
authorization, EPA retains enforcement
L/T agencies unless an alternative fall under both HSWA and non-HSWA
authority under sections 3008, 3013,
request is determined to be a major authorities. That is, changes made to
and 7003 of RCRA, although authorized
change. An alternative request that is a regulations applicable to boilers and
states have primary enforcement
major change, such as decreases in industrial furnaces are promulgated
responsibility. The standards and
record retention for all records, must be under HSWA authority, whereas
requirements for state authorization are
sent to your EPA Region for review and changes made to regulations applicable
found at 40 CFR Part 271.
approval. Similar to the monitoring to incinerators are promulgated under
Prior to enactment of the Hazardous non-HSWA authority. 259 All of the
section, § 63.1211 contains a specific and Solid Waste Amendments of 1984 amendments made today are considered
alternative provision. Section (HSWA), a State with final RCRA to be either less stringent or equivalent
63.1211(d) Data Compression, allows authorization administered its to the existing Federal program, which
sources to request to use data hazardous waste program entirely in means that states are not required to
compression techniques to record data lieu of EPA administering the federal adopt and seek authorization for these
from CMS and CEMS on a frequency program in that state. The federal provisions regardless of whether they
less than that required by § 63.1209. We requirements no longer applied in the are finalized under non-HSWA or
view the alternative request to be a authorized state, and EPA could not HSWA authorities. Nevertheless, we
minor change because available issue permits for any facilities in that strongly encourage states to become
guidance provides criteria for defining state, since only the state was authorized for today’s amendments.
fluctuation and data compression limits. authorized to issue RCRA permits.
See 64 FR 52961 and 52962, September When new, more stringent federal 259 When new requirements and prohibitions
30, 1999. Therefore, requests submitted requirements were promulgated, the (that are more stringent than the previous federal
regulations) are imposed under non-HSWA
under 63.1211(d) can be consistently state was obligated to enact equivalent authority, the new federal requirements do not take
evaluated by delegated S/L/T agencies. authorities within specified time frames. effect in an authorized state until the state adopts
Section 63.1214(c) has been revised to However, the new federal requirements the federal requirements as law. Conversely, when
specify that major alternatives to did not take effect in an authorized state imposed under HSWA authority, the new federal
requirements are federally enforceable in an
63.1211(a)—(c) are non-delegable until the state adopted the federal authorized state until the necessary changes to a
authorities. requirements as state law. state’s authorization are approved by EPA.
Experience has shown that when states could be between approximately 2,260 tons per year depending on the number
have been authorized for previous and 3,380 tons per year. A discussion of of sources that elect to comply with the
amendments (i.e., those finalized in the the emission estimates methodology and health-based compliance alternatives for
1999 rule) that were intended to results are presented in ‘‘Technical total chlorine. Of these totals,
facilitate the transition from the RCRA Support Document for HWC MACT approximately 250 tons per year will be
program to MACT and the CAA Title V Replacement Standards, Volume V: classified as hazardous waste subject to
program, the process has proven to be Emission Estimates and Engineering RCRA Subtitle C regulations. We
less cumbersome. For a more detailed Costs’’ that is available in the docket. estimate the remainder—between 8,450
discussion of non-HSWA and HSWA and 11,950 tons per year—will be
authorities with respect to how and TABLE 1.—NATIONWIDE ANNUAL EMIS- classified and managed as a non-
when they take effect, please refer to the SIONS REDUCTIONS OF HAP AND hazardous industrial waste subject to
proposal’s preamble discussion at 69 FR OTHER POLLUTANTS
21338. Subtitle D of RCRA. The costs
Several RCRA sections that have been associated with these disposal and
Estimated
enacted as part of HSWA apply to emission re- water requirements are accounted for in
Pollutant the annualized compliance cost
today’s rule: 3004(o), 3004(q), and ductions
3005(c)(3). Thus, if a state is not (tons per year) estimates. A discussion of the
authorized for the boiler and industrial methodology used to estimate impacts is
Dioxin/furans1 ....................... 0.20
furnace regulations, these provisions are All HAP metals ..................... 19.5
presented in ‘‘Technical Support
federally enforceable in an authorized Mercury ................................. 0.21 Document for HWC MACT Replacement
state until the necessary changes to a Semivolatile metals (Cd, Pb) 2.9 Standards, Volume V: Emission
state’s authorization are approved by us. Low volatile metals (As, Be, Estimates and Engineering Costs’’ that is
See RCRA section 3006, 42 U.S.C. 6926. Cr) ..................................... 6.5 available in the docket. We note that the
We are adding today’s requirements to Other metals (Co, Mn, Ni, nonair quality health and environmental
Table 1 in 271.1(j) where rulemakings Sb, Se) .............................. 9.9
impacts effects for both floor and
promulgated pursuant to HSWA HCl and chlorine gas2 .......... 1220
Particulate matter ................. 2,140 beyond-the-floor options are discussed
authority are identified. in the technical support document and
1 Dioxin/furan emission reductions are ex-
Part Six: Impacts of the Final Rule are part of our consideration of such
pressed as grams TEQ per year.
2 We are promulgating health-based compli- factors under section 112(d)(2).
I. What Are the Air Impacts?
ance alternatives for total chlorine for haz-
Table 1 below shows the emissions ardous waste combustors other than hydro- III. What Are the Energy Impacts?
reductions achieved by the final rule for chloric acid production furnaces in lieu of the
MACT technology-based emission standards We estimate that the national annual
all existing hazardous waste
(see Part Four, Section VII of the preamble for energy usage as a result of this rule will
combustors. For Phase I sources— details). Given that a number of sources may increase between approximately 73
incinerators, cement kilns, and elect to comply with the health-based compli-
lightweight aggregate kilns—the ance alternatives, the estimated reductions of million and 85 million kilowatt hours
emission reductions represent the total chlorine represent an upper bound (kWh) depending on the number of
estimate. sources that elect to comply with the
difference in emissions between sources
controlled to today’s standards and II. What Are the Water and Solid Waste health-based compliance alternatives for
estimated emissions when complying Impacts? total chlorine. The increase results from
with the interim MACT standards the electricity required to operate air
We estimate that water usage for
promulgated on February 13, 2002. pollution control equipment installed to
existing sources will increase between
Thus, the significant emissions 400 million and 1.6 billion gallons per meet the standards. The increase energy
reductions already achieved by the year as a result of today’s rule. The usage costs are accounted for in the
interim standards are not reflected in upper range estimate represents the annualized compliance cost estimates.
the estimates shown in Table 1.260 For water usage assuming no sources elect A discussion of the methodology used
Phase II sources—solid fuel boilers, to comply with the health-based to estimate impacts is presented in
liquid fuel boilers, and hydrochloric compliance alternatives for total ‘‘Technical Support Document for HWC
acid production furnaces—the chlorine, while the lower range estimate MACT Replacement Standards, Volume
reductions represent the difference in represents water usage assuming all V: Emission Estimates and Engineering
emissions between today’s standards sources elect the alternative. Water Costs.’’ We note that the energy effects
and the current baseline of control usage increases are estimated for for both floor and beyond-the-floor
provided by 40 CFR part 266, subpart H. reducing combustion gas temperatures options are discussed in the technical
Nationwide baseline HAP and with evaporated spray coolers for support document and are part of our
particulate matter emissions from dioxin/furan control as well as for new consideration of such factors under
hazardous waste combustors are particulate matter and acid gas air section 112(d)(2).
estimated to be approximately 12,650 pollution control equipment. The
tons per year at the current baseline increased water usage will also result in IV. What Are the Control Costs?
level of control. Depending on the an increase in wastewater generation.
number of facilities demonstrating Control costs, as presented in this
Depending on the number of sources section, refer only to engineering,
compliance with health-based that elect to comply with the health-
compliance alternatives for total operation, and maintenance costs
based compliance alternatives for total
chlorine, the total reduction of HAP and associated with unit/system upgrades
chlorine, we also estimate that up to 775
particulate matter for existing sources necessary to meet the final standards.
million gallons of wastewater may be
generated. These costs do not incorporate any
260 USEPA, ‘‘Final Technical Support Document
We estimate that the generation of market-based adjustments. All costs
for HWC MACT Standards, Volume V: Emission presented in this section are annualized
Estimates and Engineering Costs,’’ Section 3, July solid waste will increase between
1999. approximately 8,700 tons and 12,200 estimates in 2002 dollars.
We estimate there are a total of 267 market-adjusted compliance costs of Economic impacts presented in this
sources 261 that may be subject to approximately $274,500 per source. section are distinct from social costs,
requirements of this final rule. Of this Across all sectors covered by today’s which correspond only to the estimated
total, there are 116 boilers (104 liquid rule (Phase I and Phase II sources), total monetary value of market disturbances.
fuel boilers plus 12 solid fuel boilers), annualized compliance costs were
A. Market Exit Estimates
92 on-site incinerators, 25 cement kilns, found to average $150,500 per source.
15 commercial incinerators, nine (or Private sector engineering costs The hazardous waste combustion
seven) lightweight aggregate kilns, and (control) costs have also been assessed industry operates in a dynamic market,
ten hydrochloric acid (HCl) production on a per ton (U.S.) basis. Captive energy with systems entering and exiting the
furnaces. recovery sources (solid and liquid fuel- market on a routine basis. Our analysis
Total national private sector fired boilers, and hydrochloric acid defines ‘‘market exit’’ as ceasing to burn
engineering costs for the final standards production furnaces) burned a total of hazardous waste. We have projected
are estimated at $40.2 million per 944,667 tons of hazardous waste in post-rule hazardous waste combustion
year.262 This estimate reflects total non 2003. These facilities are projected to system market exits based on economic
market adjusted upgrade costs experience the highest average feasibility only. Social, liability, and
(engineering, plus administrative and incremental control costs, at informational issues are not
permitting), excluding chlorine control approximately $37 per ton of waste incorporated into our market exit
costs.263 All Phase II sources combined burned. Commercial energy recovery analysis.
(liquid fuel boilers, coal fired boilers, sources (cement kilns and LWAKs), Market exit estimates are derived from
and HCl production furnaces) represent burning an estimated 999,076 tons in a breakeven analysis designed to
86 percent of this total. The average 2003, are projected to experience determine system viability. This
private sector engineering cost, average incremental control costs of analysis is subject to several
excluding permitting and approximately of $3.00 per ton. Captive assumptions, including: Cost
administrative, is projected to be highest (on-site) and commercial incinerators assumptions concerning the per sector
for liquid fuel boilers, at $256,300 per burn an estimated 925,828 tons and baseline cost of hazardous waste
source. Coal fired boilers are second at 447,524 tons per year, respectively. burning, cost estimates for necessary
approximately $170,246 per source. These sources are estimated to pollution control devices (including
Total engineering costs to cement kilns experience average incremental operation and maintenance), prices for
and HCl production furnaces are engineering costs of $2.15 per ton and combustion services, and estimated
estimated to average $113,600, and $0.80 per ton, respectively. waste quantities burned at these
$16,645 per source, respectively. The aggregate control costs presented facilities. It is important to note that, for
Commercial incinerators are projected in this section do not reflect the most sectors, exiting the hazardous
to experience engineering costs anticipated real world cost burden on waste combustion market is not
averaging $12,300 per source. On-site the economy. Any market disruption, equivalent to closing a plant. (Actual
incinerators and LWAKs will face the such as the requirements in this final plant closure may occur only in the case
lowest engineering costs at $10,200 and rule, will cause a short-term of a commercial incinerator closing all
$3,330, respectively. disequilibrium in the hazardous waste systems.)
For all Phase I sources (141 sources; burning market, resulting in a natural We estimate that 39 systems,
commercial incinerators, on-site economic process designed to reach the representing about 15 percent of the
incinerators, cement kilns, and new market equilibrium. Actual cost total affected universe, may stop
lightweight aggregate kilns), total impacts to society are more accurately burning hazardous waste in response to
average annualized non market-adjusted measured by taking into account market the final standards. Approximately
compliance costs (including permitting adjustments in the targeted industry, 59,000 tons of hazardous waste may be
and administrative 264) are estimated at plus secondary (societal) costs. Total diverted from these closed systems.
$39,700 per source. The combined market-adjusted costs plus secondary These estimates assume no chlorine
Phase II sources (126 sources; solid and costs are commonly termed Social controls are put in place as a direct
liquid fuel-fired boilers and Costs, and are generally less than total result of the rule.266 Of the estimated 39
hydrochloric acid production furnaces) engineering costs due to efficiencies
market exits, 26 are projected to be on-
have total average annualized non implemented during the market
site incinerators and 8 are liquid fuel
adjustment process. Social Costs
boilers. Three commercial incinerator
261 For purposes of this discussion, a source is
theoretically represent the total real
defined as the air pollution control system systems may exit the market in response
world costs of all goods and services
associated with one or more hazardous waste to the final rule. However, these systems
combustion unit(s). A facility may operate one or society must give up in order to gain the
are considered economically marginal
more sources. Note that this total includes two added protection to human health and
LWAK units limited by system burn constraints.
in the baseline. Two coal-fired boiler
the environment. Social Costs are
Exclusion of these two units results in a total of 265 systems are also projected to exit the
presented in Part VI of this Section.265
independent sources. market. No cement kilns, lightweight
262 Not included here are total annual government
V. What Are the Economic Impacts? aggregate kilns, or HCl production
costs. These costs, with or without chlorine control, furnaces are projected to exit the market
are approximately $0.5 million/year. Economic impacts may be measured
263 We are finalizing the incorporation of section through several factors. This section as a result of the final rule. Market exit
112(d)(4) of the Clean Air Act to establish risk- presents estimated economic impacts estimates were found to be identical
based standards for total chlorine for hazardous
waste combustors (except for hydrochloric acid
relative to market exits, waste 266 Even though we are allowing sources (except
production furnaces). The low-end of this cost reallocations, and employment impacts. hydrochloric acid production furnaces) to invoke
range assumes all facilities emit total chlorine levels § 112(d)(4) in lieu of MACT chlorine control
below risk-based levels of concern. Under this 265 Beyond-the-Floor standards were assessed for requirements, we have not attempted to estimate
scenario, no total chlorine controls are assumed to all floors. These findings are available in Appendix the following: (1) The total number of sources that
be necessary. The total engineering cost with F and G of the engineering background document: may elect to implement this provision, and, (2)
chlorine control is estimated at $46.7 million/year.] See: Final Technical Support Document for HWC what level of control may be necessary following a
264 See Exhibit 4–3 in the economic assessment MACT Standards, Volume V—Emissions Estimates § 112(d)(4) risk-based determination, since this
background document. and Engineering Costs. would vary on a site-by-site basis.
when the cost of chlorine control is 1. Employment Impacts—Dislocations employment estimates reflect within
included in the model. (Losses) sector impacts only and therefore do not
Employment dislocations in the account for potential displacements
B. Waste Reallocations
combustion industry are projected to across sectors. This may occur if
Some on-site combustion systems occur when facilities consolidate waste investment funds are diverted from
(sources) may no longer be able to cover into fewer systems, or when a facility other areas of the larger economy.
their hazardous waste burning costs as exits the hazardous waste combustion VI. What Are the Social Costs and
a result of final rule requirements. These market altogether. Operation and Benefits of the Final Rule?
sources are projected to divert or reroute maintenance labor hours are expected to
their wastes to different hazardous The value of any regulatory action is
be reduced for each system that stops
waste combustion sources (usually some traditionally measured by the net
burning hazardous waste. For each
type of commercial unit).267 For change in social welfare that it
facility that completely exits the market,
multiple system facilities, this diversion generates. Our economic assessment
employment dislocations may also
may include on-site (non-commercial) conducted in support of today’s final
include supervisory and/or
waste consolidation among fewer rule evaluated compliance (control)
administrative personnel.
systems at the same facility. Under Total employment dislocations costs, and economic impacts, as
current market conditions, non- resulting from implementation of the discussed above. The Assessment also
combustion alternatives are generally final standards are estimated at 310 full- analyzed social costs, benefits, small
not economically feasible, and in any time-equivalent (FTE) jobs. On-site entity impacts, and other impacts (e.g.,
case, would normally be unable to incinerators account for about 62 children’s health, unfunded mandates).
achieve the RCRA Land Disposal percent of this total, followed by To conduct this analysis, we examined
Restriction Treatment standards, which commercial incinerators (about 24 the current combustion market and
are based on the performance of percent), and liquid-fuel boilers (about practices, developed and implemented a
combustion technology (which 12 percent). The large number of on-site methodology for examining compliance
optimizes destruction of organic HAP). incinerators drives the impacts within and social costs, applied an economic
As mentioned above, our economic this sector. model to analyze industry economic
model indicates that approximately impacts (discussed above), examined
59,000 tons (U.S.) of hazardous waste 2. Employment Impacts—Positive benefits, and followed appropriate
may be reallocated. This figure In addition to employment guidelines and procedures for
represents approximately 1.8 percent of dislocations, our analysis indicates that examining equity considerations,
the total 2003 quantity of hazardous today’s rule may also result in positive children’s health, and other impacts.
waste burned at all sources. On-site employment impacts. These positive The data applied in this analysis were
consolidations account for nearly 24 impacts are projected to occur to both the most recently available at the time
percent (13,915 tons) of all diverted the air pollution control industry and to of the analysis. Because our data were
waste. Commercial incinerators are combustion firms as they hire personnel limited, the findings from these analyses
projected to receive the vast majority to accommodate reallocated waste and/ should be more accurately viewed as
(42,722 tons, or 73 percent) of all off-site or comply with the various national estimates.
waste reallocations. Cement kilns and requirements of the rule. Hazardous A. Combustion Market Overview
LWAKs are projected to receive the waste combustion sources are projected
The hazardous waste industry
remaining reallocation (2,289 tons). to need additional operation and
consists of three key segments:
Currently, there is more than adequate maintenance personnel for the new
hazardous waste generators, fuel
capacity to accommodate all off-site pollution control equipment and other
blenders/intermediaries, and hazardous
hazardous waste diversions. compliance activities, such as new
waste burners. Hazardous waste is
reporting and record keeping
C. Employment Impacts combusted at four main types of
requirements.
The total annual positive employment facilities: commercial incinerators, on-
Today’s rule is projected to induce
impact associated with the final site incinerators, waste burning kilns
employment shifts across all affected
standards is estimated at 323 FTEs. (cement kilns and lightweight aggregate
sectors. These shifts may occur as
Positive employment impacts to the air kilns), and industrial boilers.
specific combustion facilities find it no
pollution control industry 268 are Commercial incinerators are generally
longer economically feasible to keep all
larger in size and designed to manage
of their systems running, or to stay in projected at 93 FTEs, or about 29
virtually all types of solids, as well as
the hazardous waste market at all. When percent of this total. At 183 jobs, liquid-
liquid wastes. On-site incinerators are
this occurs, workers at these locations fuel boilers are projected to experience
more often designed as liquid-injection
may lose their jobs or experience forced the greatest positive employment impact
systems that handle liquids and
relocations. At the same time, the rule among all combustors.
While it may appear that our analysis pumpable solids. Waste burning kilns
is projected to result in positive
suggests overall net positive and boilers generally burn hazardous
employment impacts, as new purchases
employment impacts, such a conclusion wastes to generate heat and power for
of pollution control equipment
would be inappropriate. Because the their manufacturing processes.
stimulate additional hiring in the As discussed above, we have
pollution control manufacturing sector, positive employment impacts and
identified a total of 267 hazardous waste
and as additional staff are required at employment dislocations occur in
burning sources (systems) currently in
selected combustion facilities to different sectors of the economy, they
operation in the United States. Liquid
accommodate reallocated waste and/or should not be added together. Doing so
fuel-boilers account for 104 sources,
various compliance activities. would mask important distributional
followed by on-site incinerators at 92
effects of the rule. In addition, these
267 This analysis includes the cost of waste
sources. Cement kilns, hydrochloric
transport to alternative combustion sources, 268 Manufacturers and distributors of air pollution acid production furnaces, and
burning fees, and purchase of alternative fuels (if control devices are projected to increase sales as a commercial incinerators account for 25,
appropriate). result of this action. 10, and 15 sources, respectively. Solid
fuel boilers and lightweight aggregate demand for hazardous waste C. Analytical Methodology and
kilns make up the remainder, at 12 and combustion services. Regulatory forces Findings—Social Cost Analysis
nine systems, respectively. These 267 influence the demand for combustion by Total social costs include the value of
sources are operated at a total of 145 mandating certain hazardous waste resources used to comply with the
different facilities. A single facility may treatment standards (land disposal standards by the private sector, the
have one or more combustion systems. restriction requirements, etc.). Liability value of resources used to administer
Facilities with multiple systems may concerns of waste generators affect the regulation by the government, and
have different types of hazardous waste combustion demand because the value of output lost due to shifts of
burning units. Combustion systems combustion, by destroying organic resources away from the current market
operating at chemical manufacturing wastes, greatly reduces the risk of future equilibrium. To evaluate these shifts in
facilities (NAICS 325) were found to environmental problems. Finally, if resources and changes in output
account for about 70 percent of the total alternative waste management options requires predicting changes in behavior
number of facilities and manage about are more expensive, hazardous waste by all affected parties in response to the
58 percent of all hazardous waste generators will likely choose to send regulation, including responses of
burned in 2003. their wastes to combustion facilities in directly-affected entities, as well as
The EPA Biennial Reporting System order to increase overall profitability. indirectly-affected private parties.
(BRS) reports a total demand for all Throughout much of the 1980s, For this analysis, social costs are
combusted hazardous waste, across all hazardous waste combustors enjoyed a grouped into two categories: Economic
facilities, at 3.32 million tons (U.S. ton) strong competitive position and welfare (changes in consumer and
in 2003. Commercial energy recovery generally maintained a high level of producer surplus), and government
(cement kilns and lightweight aggregate profitability. During this period, EPA administrative costs. The economic
kilns) burned about 30 percent of this regulations helped stimulate a greatly welfare analysis conducted for today’s
total. Hazardous waste destruction at expanded market. In addition, federal rule uses a simplified partial
on-site incinerators and commercial permitting requirements, as well as equilibrium approach. In this analysis,
incinerators accounted for 28 percent powerful local opposition to siting of changes in economic welfare are
and 13 percent, respectively. Captive new incinerators, constrained the entry measured by summing the changes in
energy recovery accounted for the of new combustion systems. As a result, consumer and producer surplus. This
remainder, at 29 percent of the total. combustion prices rose steadily, simplified approach bounds potential
About 65 percent of all hazardous ultimately reaching record levels in economic welfare losses associated with
waste burned in 2003 was organic 1987. The high profits of the late 1980s the rule by considering two scenarios:
liquids. This is followed by solids (14 induced many firms to enter the market, Compliance costs assuming no market
percent), inorganic liquids (11 percent), in spite of the difficulties and delays adjustments, and market adjusted
and sludges (10 percent). Hazardous anticipated in the permitting and siting compliance costs.
gases were found to represent a process. The annualized private sector
negligible portion, at about 0.08 percent Hazardous waste markets have compliance (engineering) costs of $40.2
of the total quantity burned in 2003. In changed significantly since the late million, as presented in Section IV,
terms of hazardous waste generating 1980s. In the early 1990s, substantial assume no market adjustments. Our best
sources, the Basic Organic Chemical overcapacity resulted in fierce estimate of total social costs
Manufacturing e sector (NAICS 325) competition, declining prices, poor incorporates rational market
generated approximately 32 percent of financial performance, numerous adjustments and all government costs.
all hazardous waste burned in 2001, project cancellations, system Under this scenario, increased
followed by pesticides and agricultural consolidations, and facility closures. compliance (engineering) costs are
chemicals, business services, organic Since the mid 1990s, several additional examined in the context of likely
fibers, medicinal chemicals, combustion facilities have closed, while incentives hazardous waste combustion
pharmaceuticals, plastics materials and many of those that have remained open facilities have to continue burning, and
resins, petroleum, and miscellaneous. have consolidated their operations. the competitive balance in the market.
Companies that generate large Available (prior to this final rule) excess Total annualized market-adjusted net
quantities of uniform hazardous wastes commercial capacity is currently private-sector costs are estimated at
generally find it more economical and estimated at about 21 percent of the $22.1 million. 270 In addition to the net
efficient to combust these wastes on-site total 2003 quantity combusted. private sector costs, total annual
using their own noncommercial B. Baseline Specification government costs are approximately
systems. Commercial incineration $0.50 million. Thus, our best estimate of
facilities manage a wide range of Proper and consistent baseline
total social costs of this final rule is
hazardous waste streams generated in specification is vital to the accurate
$22.6 million per year.
small to medium quantities by diverse assessment of incremental costs,
The $22.1 million figure incorporates
industries. Cement kilns, lightweight benefits, and other economic impacts
a net gain to selected Phase I sources
aggregate kilns, and boilers derive heat associated with today’s rule. The
and an estimated $3.6 million cost
and energy by burning high-Btu baseline essentially describes the world
(solvents and organics) liquid hazardous absent the rule. The incremental 270 We are finalizing alternative risk-based total
wastes.269 Sometimes these wastes are impacts of today’s rule are evaluated by chlorine standards for hazardous waste combustors
blended with fossil fuels where system predicting post MACT compliance (ecept for hydrochloric acid production furnaces).
responses with respect to the baseline. The net private sector costs of $22.1 million/year
operators choose to not derive all of may be considered a lower-bound estimate that
their energy input from hazardous The baseline, as applied in this analysis, assumes facilities emit total chlorine (TCI) below
waste. is the point at which today’s rule is risk-based levels of concern (i.e., no TCI controls
Regulatory requirements, liability promulgated. Thus, incremental cost are assumed to be necessary). Total net private
and economic impacts are projected sector market-adjusted costs would increase to
concerns, and economics influence the approximately $28.1 million per year if we were to
beyond the standards established in the assume all sources were to comply with technology-
269 Many cement kilns are also able to burn a February 13, 2002 Interim Standards based TCI standards (as opposed to the risk-based
certain level of non liquid waste. Final Rule. standards).
(price) increase to pre-existing facilities. This analysis applied avoided p-dioxins 275 and represents an upper
customers of commercial hazardous human health benefits factors from the bound 95th percentile confidence limit
waste combustion facilities. On-site March 2004 Assessment document,273 of the excess cancer risk from a lifetime
incinerators are projected to experience combined with more recent emissions exposure. For the past several years the
total market-adjusted cost increases of estimates for particulate matter. Agency has been conducting a
approximately $1.5 million/year. All Reduced PM emissions are estimated reassessment of the human health risks
phase II sources account for to result in monetized human health associated with dioxin and dioxin-like
approximately $31.9 million in benefits of approximately $6.29 million compounds. In October of 2004 this
increased costs. Our economic model per year. This is an undiscounted figure. reassessment 276 was delivered to the
indicates that, of the Phase I source Avoided PM morbidity cases account National Academy of Sciences (NAS) for
categories, commercial incinerators, for $3.42 million of this total, and review.
cement kilns, and LWAKs would Evidence compiled from this draft
include: respiratory illness,
experience net gains following all reassessment indicates that the
cardiovascular disease, chronic
market adjustments. The total net gain carcinogenic effects of dioxin/furans
bronchitis, work loss days, and minor
for these three source categories is may be six times as great as believed in
restricted activity. Chronic bronchitis
estimated at $14.8 million per year. 1985, reflecting an upper bound cancer
accounts for approximately 89 percent
Commercial incinerators would receive risk slope factor of 1 × 106 [mg/kg/
of the total value of avoided PM
about 98 percent of the total gain ($14.5 day]¥1 for some individuals. Agency
morbidity cases. All morbidity cases are
million/year). Gains to commercial scientists’ more likely (central tendency)
assumed to be avoided within the first
facilities occur due to marginally higher estimates (derived from the ED01 rather
year following reduced PM emissions
prices, increased waste receipts, and than the LED01) result in slope factors
and are not discounted under any
relatively low upgrade costs, when and risk estimates that are within 2–3
scenario.
compared to the other sources. times of the upper bound estimates (i.e.,
Avoided premature deaths (mortality) between 3 × 105 [mg/kg/day]¥1 and 5 ×
D. Analytical Methodology and are valued at $2.87 million per year,
Findings—Benefits Assessment 105 [mg/kg/day]¥1) based on the
undiscounted. Assuming a discount rate available epidemiological and animal
This section discusses the monetized of three and seven percent, PM cancer data. However, risks could be as
and non monetized benefits to human mortality benefits would be $2.52 low as zero for some individuals. Use of
health and the environment potentially million and $2.19 million, respectively. the alternative upper bound cancer risk
associated with today’s final rule. Our discounted analysis of PM mortality slope factor could result in a higher
Monetized human health benefits are benefits assumes that 30 percent of human health monetized health benefit
derived from reductions in particulate premature mortalities occur during the associated with premature cancer deaths
matter (PM) and dioxin/furan exposure, first year, 50 percent occur evenly from avoided in response to the final
and are based on a Value of Statistical the second through the fifth years, and standard for dioxin/furans. The
Life (VSL) estimate of $6.2 million. 271 the remaining 20 percent occur evenly assessment of upper bound cancer risk
Non monetized benefits are associated from the sixth through the twentieth using this alternative slope factor
with human health, ecological, and years.274 Due to limitations in the risk should not be considered current
waste minimization factors. analysis, this assessment of PM benefits Agency policy. The standards for dioxin
1. Monetized Benefits does not consider corresponding health in today’s final rule were not based on
benefits associated with the reduction of this draft reassessment.
Total monetized human health HAP metals carried by the PM. Total non-discounted human health
benefits for the final standards are
Dioxin/furan—Dioxin/furan benefits associated with projected
estimated to range from $5.61 million/
emissions are projected to be reduced by dioxin reductions are estimated at $0.02
year to $6.31 million/year. This estimate
a total of 0.2 grams per year under the million/year. These benefits may range
includes human health benefits
final standards. In the July 23, 1999 from $0.01 million/year to nearly zero,
associated with avoided PM and dioxin/
Addendum to the Assessment, cancer applying a discount rate of 3 percent
furans exposure. The range is driven by
alternative discount rate assumptions risk reductions linked to consumption and 7 percent, respectively. Our
(no discount rate, 3 percent, or 7 of dioxin-contaminated agricultural discounted estimates incorporate an
percent) for mortality valuation. PM products accounted for the vast majority assumed latency period of 21 and 34
benefits represent 99 percent of the total of the 0.36 cancer cases per year that years from exposure to death.
monetized human health benefits. were expected to be avoided due to the 2. Non-Monetized Benefits
1999 standards. Cancer risk reductions
Particulate Matter associated with the final standards are We examined, but did not monetize
Results from our risk assessment expected to be less than 0.36 cases per human health benefits potentially
extrapolation procedure 272 are used to year, but greater than zero. associated with reduced exposure to
evaluate incremental human health lead, mercury, and total chlorine. Non
At this time, the Agency is still using monetized ecological benefits
benefits potentially associated with a cancer risk slope factor of 1.56 × 105
particulate matter emission reductions [mg/kg/day]¥1 for dioxin. This cancer 275 USEPA, 1985. Health Assessment Document
from hazardous waste combustion slope factor is derived from the for Polychlorinated Dibenzo-p-Dioxins. EPA/600/8-
Agency’s 1985 health assessment 84/014F. Final Report. Office of Health and
271 Monetized benefits associated with avoided Environmental Assessment. Washington, DC.
document for polychlorinated dibenzo-
premature mortality reflect a VSL range of $1.1 September, 1985.
million to $11.4 million, with a central VSL 276 U.S.EPA. Exposure and Human Health
estimate of $6.2 million. These values are derived 273 Assessment of the Potential Costs, Benefits,
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-
from willingness-to-pay based VSL estimates and Other Impacts of the Hazardous Waste Dioxin (TCDD) and Related Compounds National
presented in U.S. EPA, Regulatory Impact Analysis Combustion MACT Replacement Standards: Academy Sciences (NAS) Review Draft, December
for the Final Clean Air Interstate Rule, March 2005. Proposed Rule, March 2004 (Chapter 6), and 2003. [Note: Toxicity risk factors presented in this
272 Inferential Risk Analysis in Support of Addendum to the Assessment. document should not be considered EPA’s official
Standards for Emissions of Hazardous Air 274 See: U.S. EPA. March 2005. Regulatory Impact estimate of dioxin toxicity, but rather reflect EPA’s
Pollutants from Hazardous Waste Combustors. Analysis for the Final Interstate Air Quality Rule. ongoing effort to reevaluate dioxin toxicity].
potentially associated with reductions mercury exposure in women and the 107 tons per year 281 (HCl production
in dioxin/furan; selected metals, total neuro-development of their unborn furnaces only). Hydrogen chloride is
chlorine, and particulate matter were children.277 The New Zealand and corrosive to the eyes, skin, and mucous
also examined. Finally, waste Faroe Islands studies both found a membranes. Acute inhalation can cause
minimization is examined as a non- statistically significant relationship eye, nose, and respiratory tract irritation
monetized benefit. between maternal methylmercury and inflammation, and pulmonary
Lead—The final standards are exposure and IQ decrements in the edema. Chronic occupational inhalation
expected to reduce lead emissions by unborn children of these women. In its has been reported to cause gastritis,
approximately 2.5 tons per year. In 2000 report on the toxicological effects bronchitis, and dermatitis in workers.
comparison, the 1999 standards were of methylmercury, the National Long term exposure can also cause
expected to reduce lead emissions by 89 Research Council suggested that dental discoloration and erosion.
tons per year, and were expected to integrating the results of all three Chlorine gas inhalation can cause
reduce cumulative lead exposures for studies could be useful for risk bronchitis, asthma and swelling of the
two children, ages zero to five, to less assessment purposes.278 Such an lungs, headaches, heart disease, and
than 10 µg/dL. The lead benefits integrative risk assessment, later meningitis. Acute exposure causes more
associated with these final standards are published by Ryan et al. in 2005, served severe respiratory and lung effects, and
therefore expected to be modest. The as the basis of the Agency’s health can result in fatalities in extreme cases.
final standards will also result in effects analysis for the Clean Air The exposure levels established under
reduced lead levels for children of sub- Mercury Rule (CAMR).279 The 112(d)(4) are expected to reduce
populations with especially high levels regulatory impact analysis for CAMR chlorine exposure for people in close
of exposure. Children of subsistence summarizes several of the adverse proximity to hazardous waste
fishermen, commercial beef farmers, health effects that may be linked to combustion facilities, and are therefore
and commercial dairy farmers who face mercury and reviews the likely to reduce the risk of all associated
the greatest levels of cumulative lead epidemiological literature examining health effects.
exposure may also experience the link between these effects and Ecological Benefits—We examined
comparable reductions in overall exposure to mercury.280 ecological benefits through a
exposure as a result of the MACT Total Chlorine—We were not able to comparison of the 1999 Assessment and
standards. quantify the benefits associated with today’s final standards. Ecological
Mercury—The HWC MACT final reductions in total chlorine emissions. benefits in the 1999 Assessment were
standards are expected to reduce Total chlorine is a combination of based on reductions of approximately
mercury emissions by approximately hydrogen chloride and chlorine gas. The 100 tons per year in dioxin/furans and
0.21 tons per year, approximately 93 final standards are projected to reduce selected metals. Lead was the only
percent less than the four-ton reduction total annual chlorine emissions by about pollutant of concern for aquatic
expected under the 1999 Standards. We ecosystems, while mercury appeared to
do not attempt to quantify the mercury- 277 Grandjean, P., K. Murata, E. Budtz-Jorgensen,
be of greatest concern for terrestrial
related benefits associated with today’s and P. Weihe. 2004. ‘‘Autonomic Activity in
Methylmercury Neurotoxicity: 14–Year Follow-Up ecosystems. Dioxin/furan and lead
final standards. However, because the of a Faroese Birth Cohort.’’ Journal of emission reductions also provided some
reduction in mercury emissions Pediatrics.144:169–76; Kjellstrom, T., P. Kennedy, potential benefits for terrestrial
represents a fraction of the reduction S. Wallis, A. Stewart, L. Friberg, B. Lind, P.
ecosystems. The final standards are
expected under the 1999 Standards, the Witherspoon, and C. Mantell. 1989. Physical and
mental development of children with prenatal expected to reduce dioxin/furan and
mercury-related benefits of the final exposure to mercury from fish. Stage 2: Interviews selected metal emissions by about 12
standards are likely to be less than the and psychological tests at age 6. National Swedish percent to 13 percent of the 1999
corresponding benefits under the 1999 Environmental Protection Board Report No. 3642;
Crump, K.S., T. Kjellstrom, A.M. Shipp, A. Silvers, estimate, resulting in fewer incremental
Standards.
To characterize the benefits associated and A. Stewart. 1998. ‘‘Influence of prenatal benefits than those estimated for the
with reduced mercury emissions, the
mercury exposure upon scholastic and 1999 Assessment (and later, for the 2002
psychological test performance: benchmark analysis Interim Standards). However, the 1999
1999 Assessment measured changes in of a New Zealand cohort.’’ Risk Analysis.
hazard quotients for populations living 18(6):701–713; Davidson, P.W., G.J. Myers, C. Cox, Assessment did not estimate the
near hazardous waste combustion C. Axtell, C. Shamlaye, J. Sloane-Reeves, E. ecological benefits of MACT standards
Cernichiari, L. Needham, A. Choi, Y. Wang, M. for hazardous waste burning industrial
facilities. For any given population, the Berlin, and T.W. Clarkson. 1998. ‘‘Effects of
hazard quotient is the ratio of the actual prenatal and postnatal methylmercury exposure
boilers and HCl production furnaces.
level of exposure to a safe level of from fish consumption on neurodevelopment: These systems were excluded from the
exposure. A hazard quotient greater outcomes at 66 months of age in the Seychelles universe in 1999 but are part of the
Child Development Study.’’ Journal of the universe addressed by today’s final
than one implies that a population is American Medical Association. 280(8):701–707;
potentially at risk. The exposure and Myers, G.J., P.W. Davidson, C. Cox, C.F.
standards. As a result, while the total
quotient analysis in the 1999 Shamlaye, D. Palumbo, E. Cernichiari, J. Sloane- ecological benefits of the final rule are
Assessment found that the measurable Reeves, G.E. Wilding, J. Kost, L.S. Huang, and T.W. likely to be modest, areas near facilities
Clarkson. 2003. ‘‘Prenatal methylmercury exposure with boilers may enjoy more significant
benefits of reduced mercury emissions from ocean fish consumption in the Seychelles
under the 1999 Standards were likely to child development study.’’ Lancet. 361(9370):1686–
ecological benefits under the final
be small because baseline exposures 92. standards than areas near facilities that
were relatively low. In addition, many 278 National Research Council of the National have already complied with the 2002
of the studies examining the adverse Academy of Sciences, Toxicological Effects of Interim standards.
Methylmercury. 2000, p. 299. Mercury, lead, and chlorides are
health effects of mercury are 279 Ryan, L.M. Effects of Prenatal Methylmercury
inconclusive. Over the past several on Childhood IQ: A Synthesis of Three Studies.
among the HAPs that can cause damage
years, however, scientists have Report to the U.S. Environmental Protection to the health and visual appearance of
conducted three large-scale studies of Agency, 2005; U.S. EPA. Regulatory Impact
Analysis of the Clean Air Mercury Rule: Final 281 This is a lower bound estimate that assumes
individuals in the Faroe Islands, New Report. March 2005. all other sources will implement 112(d)(4) and will
Zealand, and the Seychelles Islands 280 U.S. EPA. Regulatory Impact Analysis of the not move to reduce TCl emissions from current
examining the relationship between Clean Air Mercury Rule: Final Report. March 2005. baseline levels.
plants.282 While the total value of forest minimization analysis to inform the been assessed in our analysis but are
health is difficult to estimate, visible expected price change. The analysis likely to at least partially offset
deterioration in the health of forests and concluded that the demand for corresponding benefits.
plants can cause a measurable change in hazardous waste combustion is
4. Conclusion
recreation behavior. Several studies that relatively inelastic. While a variety of
measure the change in outdoor waste minimization alternatives are Total non-discounted monetized
recreation behavior according to forest available for managing hazardous waste human health benefits associated with
health have attempted to place a value streams that are currently combusted, the final standards are estimated at
on aesthetic degradation of forests.283 the costs of these alternatives generally $6.31 million/year. Annualized
Although these studies are available, exceed the cost of combustion. When discounted benefits were found to range
additional research is needed to fully the additional costs of compliance with from $5.61 million to $5.95 million/
understand the effects of these Haps on the MACT standards are taken into year. The range reflects an alternative
the forest ecosystem. Thus, these account, waste minimization discount rate of 3 percent and 7 percent
benefits are not quantified in this alternatives still tend to exceed the for mortality benefits.
analysis. higher combustion costs. This relative It is important to emphasize that
Emissions that are sufficient to cause inelasticity suggests that, in the short monetized benefits represent only a
structural and aesthetic damage to term, large reductions in the amount of portion of the total benefits associated
vegetation are likely to affect growth as hazardous waste requiring combustion with this rule. A significant portion of
well. Little research has been done on are not likely to occur. However, over the benefits are not monetized, as
the effects of compounds such as the longer term (i.e. as production discussed above, due to data and
chlorine, heavy metals (as air systems are updated), companies may analytical limitations. Specifically,
pollutants), and PM on agricultural continue to seek alternatives to ecological benefits, and human health
productivity.284 Even though the expensive hazardous waste- benefits associated with reductions in
potential for visible damage and management. This may include process chlorine, mercury, and lead are not
production decline from metals and adjustments that result, to some degree quantified or monetized. In some
other pollutants suggests the final in source reduction of hazardous waste regions these benefits may be
standards could increase agricultural and the increased generation of non significant. In addition, specific sub-
productivity, we have not monetized the hazardous waste. To the extent that populations near combustion facilities,
benefits of these changes. increases in combustion prices provide including children and minority
additional incentive to adopt more populations, may be disproportionately
3. Waste Minimization Benefits affected by environmental risks and may
efficient processes, the final standards
Facilities that burn hazardous waste may contribute to longer term process- therefore enjoy more significant
and remain in operation following based hazardous waste minimization benefits. Visibility benefits associated
implementation of the final standards efforts. with reduced PM are also expected from
are expected to experience marginally No hazardous waste minimization this final rule. For a complete
increased costs as a result of these impacts are captured in our quantitative discussion of the methodology, data,
standards. This will result in an analysis of costs and benefits.285 A findings, and limitations associated
incentive to pass these increased costs quantitative assessment of the benefits with our benefits analysis the reader is
on to their customers in the form of associated with waste minimization encouraged to review the Assessment
higher combustion prices. In the 1999 may result in double-counting of some document,286 and the Addendum to the
Assessment we conducted a waste of the benefits described earlier. For Assessment.
example, waste minimization may Part Seven: How Does the Final Rule
282 Although the primary pollutants which are reduce emissions of hazardous air Meet the RCRA Protectiveness
detrimental to vegetation aesthetics and growth are pollutants and therefore have a positive
tropospheric ozone, sulfur dioxide, and hydrogen Mandate?
fluoride (three pollutants which are not regulated effect on public health. Furthermore,
emission reductions beyond those As discussed in more detail below, we
in the MACT standards), some literature exists on
the relationship between metal deposition and necessary for compliance with the final believe today’s final standards are
vegetation health. (Mercury Study Report to standards are not addressed in the generally protective of human health
Congress Volume VI, 1997) (Several studies are
benefits assessment. In addition, waste and the environment. We therefore
cited in this report.) finalize and apply these standards, in
283 See, for example, Brown, T.C. et al. 1989, minimization is likely to result in
Scenic Beauty and Recreation Value: Assessing the specific types of benefits not captured in most instances, in lieu of the RCRA air
Relationship, In J. Vining, ed., Social Science and this Assessment. For example, waste emission standards applicable to these
Natural Resources Recreation Management,
generators that engage in waste sources.
Westview Press, Boulder, Colorado; this work
studies the relationship between forest minimization may experience a I. Background
characteristics and the value of recreational reduction in their waste handling costs Section 3004(a) of RCRA requires the
participation. Also see Peterson, D.G. et al. 1987, and could also reduce the risk related to
Improving Accuracy and Reducing Cost of Agency to promulgate standards for
Environmental Benefit Assessments. Draft Report to
waste spills and waste management. hazardous waste treatment, storage, and
the U.S. EPA, by Energy and Resource Consultants, Finally, waste minimization procedures disposal facilities as necessary to protect
Boulder, Colorado; Walsh et al. 1990, Estimating the potentially stimulated by today’s action human health and the environment. The
public benefits of protecting forest quality, Journal may result in additional costs to
of Forest Management, 30:175–189., and Homes et standards for hazardous waste
al. 1992, Economic Valuation of Spruce-Fir Decline
facilities that implement these incinerators generally rest on this
in the Southern Appalachian Mountains: A technologies. These factors have not authority. In addition, § 3004(q) requires
comparison of Value Elicitation Methods. Presented
at the Forestry and the Environment: Economic 285 Note that this rule does, in fact, consider
the Agency to promulgate standards for
Perspectives Conference, March 1, 1992 Jasper, hazardous waste feed control. Feed control can be emissions from facilities that burn
Alberta, Canada for estimates of the WTP of visitors implemented by each source through waste
and residents to avoid forest damage. minimization procedures. See: Final Technical 286 Assessment of the Potential Costs, Benefits,
284 MacKenzie, James J., and Mohamed T. El- Support Document for HWC MACT Standards, and Other Impacts of the Hazardous Waste
Ashry, Air Pollution’s Toll on Forests and Crops Volume V–Emissions Estimates and Engineering Combustion MACT Final Rule Standards.
(New Haven, Yale University Press, 1989). Costs. September 2005.
hazardous waste fuels (e.g., cement and degree of protection afforded by the conclusions. For example, for SFBs it
lightweight aggregate kilns, boilers, and MACT standards being finalized today. was not possible to conduct hypothesis
hydrochloric acid production furnaces) We have not conducted a tests that could be considered valid
as necessary to protect human health comprehensive risk assessment for this involving correlations among variables
and the environment. Using RCRA rulemaking as was done for incinerators, for a number of variables in the analysis
authority, the Agency has established cement kilns, and lightweight aggregate because of the small number of data
emission (and other) standards for kilns in the 1999 MACT rule where we points and the power of the tests to
hazardous waste combustors that are concluded that the promulgated detect differences for those that were
either entirely risk-based (e.g., site- standards were generally protective and conducted was very low, which greatly
specific standards for metals under the therefore, the RCRA standards need not diminishes the value of the results.
Boiler and Industrial Furnace rule), or be retained. However, we noted that in (Indeed, no differences in correlations
are technology-based but determined by certain instances, permit authorities were found for SFBs at the 0.1
a generic risk assessment to be may invoke the omnibus authority significance level—the level of
protective (e.g., the DRE standard for (RCRA § 3005(c)(3) and its significance that was used in the
incinerators and BIFs). implementing regulations at § 270.10(k)) analysis.) Similarly, for HCl production
The MACT standards finalized today if there is some reason to believe that furnaces the power of the tests to detect
implement the technology-based regime additional controls beyond those differences in correlations was quite
of CAA § 112(d). There is, however, a required pursuant to 40 CFR parts 63, low. It must be noted that the
residual risk component to air toxics 264, 265, and 266 may be needed to comparative analysis methodology was
standards. Section 112(f) of the Clean ensure protection of human health and not intended for comparisons that
Air Act requires the Agency to impose, the environment under RCRA. involve relatively few facilities because
within eight years after promulgation of For this final rule, we instead it is grounded in tests of hypotheses and
the technology-based standards compared the risk-related characteristics levels of statistical significance which
promulgated under § 112(d) (i.e., the of the sources covered by the 1999 rule generally require substantial amounts of
authority for today’s final standards), to the sources covered by today’s rule data to produce firm conclusions.
additional controls if needed to protect (e.g., estimated emissions, stack Nevertheless, in consideration of the
public health with an ample margin of characteristics, meteorology, and indications of possible risks for the
safety or to prevent adverse population). For a description of the aforementioned standards, permit
environmental effect. methodology and technical discussion authorities may want to consider site-
RCRA § 1006(b) directs that EPA of its application, see ‘‘Inferential Risk specific factors in determining whether
‘‘integrate all provisions of [RCRA] for Analysis in Support of Standards for or not the MACT standards are
purposes of administration and Emissions of Hazardous Air Pollutants sufficiently protective for facilities that
enforcement and * * * avoid from Hazardous Waste Combustors,’’ in fall into these categories.
duplication, to the maximum extent the docket for today’s rule. We The comparative analysis may also
possible, with the appropriate performed a large array of statistical raise possible concerns for lightweight
provisions of the Clean Air Act * * * ’’ comparisons and from these we aggregate kilns (LWAKs) and liquid
Thus, although considerations of risk attempted to make inferences about fuel-fired boilers (LFBs) with dry APCDs
are not ordinarily part of the MACT whether risks would be expected to be with regard to the dioxin/furan
process, in order to avoid duplicative about the same, less than, or greater standards, in view of the ongoing
standards where possible, we have than the risks estimated for 1999 rule. uncertainty in cancer and other health
evaluated the protectiveness of the We think the comparative analysis lends effects levels for chlorinated dioxins
standards finalized today. additional support to our view that and furans. In particular, some recent
As noted above, under RCRA, EPA today’s final standards are generally estimates of the carcinogenicity of these
must promulgate standards ‘‘as may be protective. We received no comments compounds that consider both human
necessary to protect human health and either in support of or in opposition to and animal data, are higher than earlier
the environment.’’ RCRA § 3004(a) and our use of the comparative analysis to estimates derived from animal data
(q). Technology-based standards evaluate the protectiveness of the alone. However, like SFBs and HCl
developed under CAA § 112 do not standards being finalized today or our production furnaces, LWAKs and LFBs
automatically satisfy this requirement, view that the standards are generally with dry APCDs both comprise small
but may do so in fact. See 59 FR at protective. source categories (3 LWAK facilities and
29776 (June 6, 1994) and 60 FR at 32593 While we regard the final standards as 7 dry APCD LFB facilities). This makes
(June 23, 1995) (RCRA regulation of generally protective, the comparative it very difficult to reach firm
secondary lead smelter emissions analysis suggests some concern for solid conclusions and suggests the need to
unnecessary at this time given fuel-fired boilers (SFBs) with regard to consider site-specific factors in
stringency of technology-based standard the particulate matter standard (and determining whether the MACT
and pendency of § 112(f) certain metals such as antimony and standards are sufficiently protective in
determination). If the MACT standards, thallium), mercury, and total chlorine these instances.
as a factual matter, are sufficiently standards (other than the alternative Except as noted, we believe today’s
protective to also satisfy the RCRA risk-based chlorine standards). The final standards provide a substantial
mandate, then no independent RCRA analysis also suggests some concern for degree of protection to human health
standards are required. Conversely, if hydrochloric acid (HCl) production and the environment. We therefore do
MACT standards are inadequate, the furnaces with regard to the dioxin/furan not believe that we need to retain the
RCRA authorities would have to be used standard, where carbon monoxide and existing RCRA standards for boilers and
to fill the gap. total hydrocarbon serve as surrogate hydrochloric acid production furnaces
control. However, because both SFBs (just as we found that existing RCRA
II. Evaluation of Protectiveness and HCl production furnaces comprise standards for incinerators, cement kilns,
For the purpose of satisfying the such small source categories (4 SFB and lightweight aggregate kilns were no
RCRA statutory mandates, the Agency facilities and 8 HCl production longer needed after the 1999 rule).
has conducted an evaluation of the facilities), it is difficult to reach firm However, as previously discussed in
more detail in Part Four, Section IX, support of today’s final rule. This processing and maintaining
site-specific risk assessments may be document is entitled: Assessment of the information, and disclosing and
warranted on an individual source basis Potential Costs, Benefits, and Other providing information; adjust the
to ensure that the MACT standards Impacts of the Hazardous Waste existing ways to comply with any
provide adequate protection in Combustion MACT Final Rule previously applicable instructions and
accordance with RCRA. Standards, September 2005. We have requirements; train personnel to be able
also prepared an Addendum to this to respond to a collection of
Part Eight: Statutory and Executive Assessment entitled: Addendum to the information; search data sources;
Order Reviews Assessment of the Potential Costs, complete and review the collection of
I. Executive Order 12866: Regulatory Benefits, and Other Impacts of the information; and transmit or otherwise
Planning and Review Hazardous Waste Combustion MACT disclose the information.
Final Rule Standards, September 2005. An agency may not conduct or
Under Executive Order 12866 [58 FR
This Addendum captures changes made sponsor, and a person is not required to
51735 (October 4, 1993)] the Agency, in
to the rulemaking following completion respond to, a collection of information
conjunction with OMB’s Office of
of the full Assessment document. The unless it displays a currently valid OMB
Information and Regulatory Affairs
Assessment and Addendum were control number. The OMB control
(OIRA), must determine whether a
designed to adhere to analytical numbers for EPA’s regulations are listed
regulatory action is ‘‘significant’’ and
requirements established under in 40 CFR part 9. When this ICR is
therefore subject to OMB review and the approved by OMB, the Agency will
Executive Order 12866, and
full requirements of the Executive publish a technical amendment to 40
corresponding Agency and OMB
Order. The Order defines ‘‘significant CFR part 9 in the Federal Register to
guidance; subject to data, analytical, and
regulatory action’’ as one that is likely display the OMB control number for the
resource limitations. Findings presented
to result in a rule that may: approved information collection
under Part Six of this Preamble were
(1) Have an annual effect on the requirements contained in this final
developed in accordance with this
economy of $100 million or more or rule.
guidance. The RCRA docket established
adversely affect in a material way the The EPA requested comments (see 70
for today’s rulemaking maintains a copy
economy, a sector of the economy, of the Assessment and Addendum for FR 20748, Apr. 21, 2005) on the need for
productivity, competition, jobs, the public review. Interested persons are this information, the accuracy of the
environment, public health or safety, or encouraged to read both documents to provided burden estimates, and any
State, local, or tribal governments or gain a full understanding of the suggested methods for minimizing
communities; analytical methodology, findings, and respondent burden, including through
(2) Create a serious inconsistency or limitations associated with this report. the use of automated collection
otherwise interfere with an action taken techniques.
or planned by another agency; II. Paperwork Reduction Act
(3) Materially alter the budgetary III. Regulatory Flexibility Act
We have prepared an Information
impact of entitlements, grants, user fees, Collection Request (ICR) document (ICR The Regulatory Flexibility Act (RFA)
or loan programs or the rights and No. 1773.08) listing the information as amended by the Small Business
obligations of recipients thereof; or collection requirements of this final Regulatory Enforcement Fairness Act of
(4) Raise novel legal or policy issues rule, and have submitted it for approval 1996 (SBREFA), 5 U.S.C. 601 et seq.,
arising out of legal mandates, the to the Office of Management and Budget generally requires an agency to prepare
President’s priorities, or the principles (OMB) under the provisions of the a regulatory flexibility analysis of any
set forth in the Executive Order. Paperwork Reduction Act, U.S.C. 3501 rule subject to notice and comment
Pursuant to the terms of Executive et seq. OMB has assigned a control rulemaking requirements under the
Order 12866, it has been determined number 2050–0171 for this ICR. This Administrative Procedure Act, or any
that this rule is a ‘‘significant regulatory ICR is available for public viewing in other statute. This analysis must be
action’’ because this action may raise the EPA Docket Center, Room B102, completed unless the agency is able to
novel legal or policy issues due to the 1301 Constitution Avenue NW., certify that the rule will not have a
methodology applied in development of Washington, DC. Copy may also be significant economic impact on a
the final standards. As such, this action obtained from the EDOCKET on the EPA substantial number of small entities.
was submitted to OMB for review. Web site, or by calling (202) 566–1744. Small entities include small businesses,
Changes made in response to OMB The information collection requirements small not-for-profit enterprises, and
suggestions or recommendations are are not enforceable until OMB approves small governmental jurisdictions.
documented in the public record. them. The RFA provides default definitions
The total social costs for this rule are The public burden associated with for each type of small entity. Small
estimated at $22.6 million per year 287. this final rule is projected to affect 238 entities are defined as: (1) A small
This figure is significantly below the HWC units and is estimated to average business as defined by the Small
$100 million threshold established 211 hours per respondent annually. The Business Administration’s (SBA)
under point number one above. Thus, reporting and recordkeeping cost regulations at 13 CFR 121.201; (2) a
this rule is not considered to be an burden is estimated to average $5,640 small governmental jurisdiction that is a
economically significant action. per respondent annually. government of a city, county, town,
However, in an effort to comply with Burden means total time, effort, or school district or special district with a
the spirit of the Order, we have financial resources expended by persons population of less than 50,000; and (3)
prepared an economic assessment in to generate, maintain, retain, disclose, or a small organization that is any not-for-
provide information to or for a Federal profit enterprise which is independently
287 This figure includes approximately $0.5 agency. That includes the time needed owned and operated and is not
million/year in total government costs. Total social to review instructions; develop, acquire, dominant in its field.
costs would increase to approximately $28.6
million per year if we were to assume all sources
install, and utilize technology and After considering the economic
were to comply with technology-based TC1 systems for the purposes of collecting, impacts of today’s final rule on small
standards. validating, and verifying information, entities, I certify that this action will not
have a significant economic impact on mandates’’ that may result in the officials in the development of
a substantial number of small entities. expenditure by State, local, and tribal regulatory policies that have tribal
We have determined that hazardous governments, in the aggregate, or by the implications.’’ Our Agency workgroup
waste combustion facilities are not private sector, of $100 million or more for this rule included Tribal
owned by small governmental in any one year. Today’s final rule does representation. We have determined
jurisdiction or nonprofit organizations. not result in $100 million or more in that this final rule does not have tribal
Therefore, only small businesses were expenditures for any of these categories. implications, as specified in the Order.
analyzed for small entity impacts. For The aggregate annualized social cost for No Tribal governments are known to
the purposes of the impact analyses, today’s rule is estimated at $22.6 own or operate hazardous waste
small entity is defined either by the million. combustors subject to the requirements
number of employees or by the dollar
V. Executive Order 13132: Federalism of this final rule. Furthermore, this rule
amount of sales. The level at which a
business is considered small is Executive Order 13132, entitled focuses on requirements for all
determined for each North American ‘‘Federalism’’ (64 FR 43255, August 10, regulated sources without affecting the
Industrial Classification System 1999), requires EPA to develop an relationships between tribal
(NAICS) code by the Small Business accountable process to ensure governments in its implementation, and
Administration. ‘‘meaningful and timely input by State applies to all regulated sources, without
Affected individual waste combustors and local officials in the development of distinction of the surrounding
(incinerators, cement kilns, lightweight regulatory policies that have federalism populations affected. Thus, Executive
aggregate kilns, solid and liquid fuel- implications.’’ ‘‘Policies that have Order 13175 does not apply to this rule.
boilers, and hydrochloric acid federalism implications’’ is defined in
production furnaces) will bear the the Executive Order to include VII. Executive Order 13045: Protection
impacts of today’s rule. These units will regulations that have ‘‘substantial direct of Children From Environmental Health
incur direct economic impacts (positive effects on the States, on the relationship Risks and Safety Risks
or negative) as a result of today’s rule. between the national government and
Executive Order 13045: ‘‘Protection of
Few of the hazardous waste combustion the States, or on the distribution of
power and responsibilities among the Children from Environmental Health
facilities affected by this rule were
various levels of government.’’ Risks and Safety Risks’’ (62 FR. 19885,
found to be owned by small businesses,
as defined by the Small Business Under Executive Order 13132, EPA April 23, 1997) applies to any rule that:
Administration (SBA). From our may not issue a regulation that has (1) Is determined to be ‘‘economically
universe of 145 facilities, we identified federalism implications, that imposes significant’’ as defined under E.O.
eight facilities that are currently owned substantial direct compliance costs, and 12866, and (2) concerns an
by small businesses. Four of these that is not required by statute, unless environmental health or safety risk that
facilities are liquid boilers, two are on- the Federal government provides the EPA has reason to believe may have a
site incinerators, one is a cement kiln, funds necessary to pay the direct disproportionate effect on children. If
and one is a lightweight aggregate kiln compliance costs incurred by State and the regulatory action meets both criteria,
(LWAK). Our analysis indicates that local governments, or EPA consults with the Agency must evaluate the
none of these facilities are likely to State and local officials early in the environmental health or safety effects of
incur annualized compliance costs process of developing the regulation. the planned rule on children, and
greater than one percent of gross annual This final rule does not have explain why the planned regulation is
corporate revenues. Cost impacts of the federalism implications. It will not have preferable to other potentially effective
final standards were found to range substantial direct effects on the States, and reasonably feasible alternatives
from less than 0.01 percent to 0.46 on the relationship between the national considered by the Agency. Today’s final
percent of annual gross corporate government and the States, or on the rule is not subject to the Executive
revenues. distribution of power and
Order because it is not economically
The reader is encouraged to review responsibilities among the various
significant as defined under point one of
our regulatory flexibility screening levels of government, as specified in the
the Order, and because the Agency does
analysis prepared in support of this Order. The rule focuses on requirements
for facilities burning hazardous waste, not have reason to believe the
determination. This analysis is
without affecting the relationships environmental health or safety risks
incorporated as Appendix H of the
Assessment document, and updated in between Federal and State governments. addressed by this action present a
the Addendum. Thus, Executive Order 13132 does not disproportionate risk to children.
apply to this rule. Although section 6 of VIII. Executive Order 13211: Actions
IV. Unfunded Mandates Reform Act of Executive Order 13132 does not apply
1995 Concerning Regulations That
to this rule, EPA did include various Significantly Affect Energy Supply,
Signed into law on March 22, 1995, State representatives on our Agency Distribution, or Use
the Unfunded Mandates Reform Act workgroup. These representatives
(UMRA) calls on all federal agencies to participated in the development of this This rule is not subject to Executive
provide a statement supporting the need rule. Order 13211, ‘‘Actions Concerning
to issue any regulation containing an Regulations That Significantly Affect
unfunded federal mandate and VI. Executive Order 13175:
Consultation and Coordination With Energy Supply, Distribution, or Use’’ (66
describing prior consultation with Fed. Reg. 28355 (May 22, 2001)). This
representatives of affected state, local, Indian Tribal Governments
rule, as finalized, will not seriously
and tribal governments. Executive Order 13175: Consultation
Today’s final rule is not subject to the and Coordination with Indian Tribal disrupt energy supply, distribution
requirements of sections 202, 204 and Governments (65 FR 67249, November patterns, prices, imports or exports.
205 of UMRA. In general, a rule is 9, 2000), requires EPA to develop an Furthermore, this rule is not an
subject to the requirements of these accountable process to ensure economically significant action under
sections if it contains ‘‘Federal ‘‘meaningful and timely input by tribal Executive Order 12866.
IX. National Technology Transfer and minority individuals living near a 40 CFR Part 63
Advancement Act relatively large sample of cement kilns Environmental protection, Air
Section 12(d) of the National and hazardous waste incinerators and pollution control, Hazardous
Technology Transfer and Advancement provide county, state, and national substances, Incorporation by reference,
Act of 1995 (‘‘NTTAA’’), Public Law population percentages for various sub- Reporting and recordkeeping
104–113, 12(d) (15 U.S.C. 272 note) populations. The demographic data in requirements.
directs EPA to use voluntary consensus these reports provide several important
findings when examined in conjunction 40 CFR Part 260
standards in its regulatory activities
unless to do so would be inconsistent with the risk reductions projected from Environmental protection,
with applicable law or otherwise today’s rule. Administrative practice and procedure,
impractical. Voluntary consensus We find that combustion facilities, in Confidential business information,
standards are technical standards (e.g., general, are not located in areas with Hazardous waste, Reporting and
materials specifications, test methods, disproportionately high minority and recordkeeping requirements.
sampling procedures, and business low-income populations. However,
there is evidence that hazardous waste 40 CFR Part 264
practices) that are developed or adopted
by voluntary consensus standards burning cement kilns are somewhat Environmental protection, Air
bodies. The NTTAA directs EPA to more likely to be located in areas that pollution control, Hazardous waste,
provide Congress, through OMB, have relatively higher low-income Insurance, Packaging and containers,
explanations when the Agency decides populations. Furthermore, there are a Reporting and recordkeeping
not to use available and applicable small number of commercial hazardous requirements, Security measures, Surety
voluntary consensus standards. waste incinerators located in highly bonds.
This rulemaking involves urbanized areas where there is a
disproportionately high concentration of 40 CFR Part 265
environmental monitoring or
measurement. Both Performance Based minorities and low-income populations Environmental protection, Air
Measurement System (PBMS) and within one and five mile radii. The pollution control, Hazardous waste,
specific measurement methods are reduced emissions at these facilities due Insurance, Packaging and containers,
finalized under this rule. The PBMS to today’s rule could represent Reporting and recordkeeping
approach is intended to be more flexible meaningful environmental and health requirements.
and cost-effective for the regulated improvements for these populations. 40 CFR Part 266
community; it is also intended to Overall, today’s rule should not result in
any adverse or disproportional health or Environmental protection, Energy,
encourage innovation in analytical Hazardous waste, Recycling, Reporting
technology and improved data quality. safety effects on minority or low-income
populations. Any impacts on these and recordkeeping requirements.
Where allowed, EPA is not precluding
the use of any method, whether it populations are likely to be positive due 40 CFR Part 270
constitutes a voluntary consensus to the reduction in emissions from
combustion facilities near minority and Environmental protection,
standard or not, as long as it meets the Administrative practice and procedure,
performance criteria specified. low-income population groups. The
Assessment document available in the Confidential business information,
X. Executive Order 12898: Federal RCRA docket established for today’s Hazardous materials transportation,
Actions To Address Environmental rule discusses our Environmental Hazardous waste, Reporting and
Justice in Minority Populations and Justice analysis. recordkeeping requirements.
Low-Income Populations 40 CFR Part 271
XI. Congressional Review
Executive Order 12898, ‘‘Federal Administrative practice and
Actions to Address Environmental The Congressional Review Act (CRA),
5 U.S.C. 801 et seq., as added by the procedure, Hazardous materials
Justice in Minority Populations and transportation, Hazardous waste,
Low-Income Populations’ (February 11, Small Business Regulatory Enforcement
Fairness Act of 1996, generally provides Intergovernmental relations, Reporting
1994) requires us to complete an and recordkeeping requirements.
analysis of today’s rule with regard to that before a rule may take effect, the
equity considerations. The Order is agency promulgating the rule must Dated: September 14, 2005.
designed to address the environmental submit a rule report, which includes a Stephen L. Johnson,
and human health conditions of copy of the rule, to each House of the Administrator.
minority and low-income populations. Congress and to the Comptroller General ■ For the reasons set out in the
This section briefly discusses potential of the United States. Prior to publication preamble, title 40, chapter I, of the Code
impacts (direct or disproportional) of the final rule in the Federal Register, of Federal Regulations is amended as
today’s rule may have in the area of we will submit all necessary follows:
environmental justice. information to the U.S. Senate, the U.S.
We have recently analyzed House of Representatives, and the PART 9—OMB APPROVALS UNDER
demographic data from the U.S. Census, Comptroller General of the United THE PAPERWORK REDUCTION ACT
and have previously examined data States. Under the CRA, a major rule
cannot take effect until 60 days after it ■ 1. The authority citation for part 9
from two other reports: ‘‘Race, Ethnicity, continues to read as follows:
and Poverty Status of the Populations is published in the Federal Register.
Living Near Cement Plants in the United This action is not a ‘‘major rule’’ as Authority: 7 U.S.C. 135 et seq., 136–136y;
States’’ (EPA, August 1994) and ‘‘Race, defined by 5 U.S.C. 804(2). 15 U.S.C. 2001, 2003, 2005, 2006, 2601–2671;
21 U.S.C. 331j, 346a, 348; 31 U.S.C. 9701; 33
Ethnicity, and Poverty Status of the List of Subjects U.S.C. 1251 et seq., 1311, 1313d, 1314, 1318,
Populations Living Near Hazardous 1321, 1326, 1330, 1342, 1344, 1345 (d) and
Waste Incinerators in the United States’’ 40 CFR Part 9
(e), 1361; E.O. 11735, 38 FR 21243, 3 CFR,
(EPA, October 1994). These reports Environmental protection, Reporting 1971–1975 Comp. p. 973; 42 U.S.C. 241,
examine the number of low-income and and recordkeeping requirements. 242b, 243, 246, 300f, 300g, 300g–1, 300g–2,
300g–3, 300g–4, 300g–5, 300g–6, 300j–1, Emission Standards for Hazardous Air § 9.1 OMB approvals under the Paperwork
300j–2, 300j–3, 300j–4, 300j–9, 1857 et seq., Pollutants for Source Categories’’ by Reduction Act.
6901–6992k, 7401–7671q, 7542, 9601–9657,
adding entry ‘‘63.1200–63.1221’’ in * * * * *
11023, 11048.
numerical order to read as follows:
■ 2. Section 9.1 is amended in the table
under center heading ‘‘National
40 CFR citation OMB control No.
* * * * * * *
National Emission Standards for Hazardous Air Pollutants for Source Categories 3
* * * * * * *
63.1200–63.1221 ...................................................................................... 2050–0171
3 The ICRs referenced in this section of the table encompass the applicable general provisions contained in 40 CFR part 63, subpart A, which
are not independent information collection requirements.
* * * * * Superintendent of Documents, U.S. (2) [Reserved]

Government Printing Office, ■ 3. Section 63.1200 is amended by:
PART 63—NATIONAL EMISSION Washington, DC 20402, (202) 512–1800: ■ a. Revising the introductory text.
STANDARDS FOR HAZARDOUS AIR (1) The following methods as ■ b. Revising paragraph (a)(2).
POLLUTANTS FOR SOURCE published in the test methods ■ c. Adding entry (4) in Table 1 in
CATEGORIES compendium known as ‘‘Test Methods paragraph (b).
for Evaluating Solid Waste, Physical/ The revisions and additions read as
■ 1. The authority citation for part 63 Chemical Methods,’’ EPA Publication
continues to read as follows: follows:
SW–846, Third Edition. A suffix of ‘‘A’’
Authority: 42 U.S.C. 7401 et seq. in the method number indicates § 63.1200 Who is subject to these
■ 2. Section 63.14 is amended by: revision one (the method has been regulations?
■ a. Removing paragraphs (i)(1) and revised once). A suffix of ‘‘B’’ in the The provisions of this subpart apply
(i)(2). method number indicates revision two to all hazardous waste combustors:
■ b. Redesignating paragraph (i)(3) as (the method has been revised twice). hazardous waste incinerators, hazardous
(i)(1). (i) Method 0023A, ‘‘Sampling Method waste cement kilns, hazardous waste
■ c. Adding and reserving new for Polychlorinated Dibenzo-p-Dioxins lightweight aggregate kilns, hazardous
paragraph (i)(2). and Polychlorinated Dibenzofuran waste solid fuel boilers, hazardous
■ d. Revising paragraph (k). Emissions from Stationary Sources,’’ waste liquid fuel boilers, and hazardous
The revisions and additions read as dated December 1996 and in Update III, waste hydrochloric acid production
follows: IBR approved for § 63.1208(b)(1) of furnaces. Hazardous waste combustors
Subpart EEE of this part. are also subject to applicable
§ 63.14 Incorporations by reference. (ii) Method 9071B, ‘‘n-Hexane requirements under parts 260 through
* * * * * Extractable Material (HEM) for Sludge, 270 of this chapter.
(i) * * * Sediment, and Solid Samples,’’ dated (a) * * *
(2) [Reserved] April 1998 and in Update IIIA, IBR (2) Both area sources and major
* * * * * approved for § 63.7824(e) of Subpart sources subject to this subpart, but not
(k) The following materials are FFFFF of this part. previously subject to title V, are
available for purchase from the National (iii) Method 9095A, ‘‘Paint Filter immediately subject to the requirement
Technical Information Service (NTIS), Liquids Test,’’ dated December 1996 to apply for and obtain a title V permit
5285 Port Royal Road, Springfield, VA and in Update III, IBR approved for in all States, and in areas covered by
22161, (703) 605–6000 or (800) 553– §§ 63.7700(b) and 63.7765 of Subpart part 71 of this chapter.
6847; or for purchase from the EEEEE of this part. (b) * * *
TABLE 1 TO § 63.1200.—HAZARDOUS WASTE COMBUSTORS EXEMPT FROM SUBPART EEE

If And If Then
* * * * * * *
(4) You meet the definition of a small quantity burn- ............................................................. You are not subject to the requirements of this
er under § 266.108 of this chapter subpart (Subpart EEE).
* * * * * definitions for ‘‘Btu’’, ‘‘Hazardous waste and ‘‘System removal efficiency’’ in

■ 4. Section 63.1201 is amended in hydrochloric acid production furnace’’, alphabetical order to read as follows:
paragraph (a) by revising the definitions ‘‘Hazardous waste liquid fuel boiler’’,
of ‘‘Hazardous waste combustor’’, ‘‘New ‘‘Hazardous waste solid fuel boiler’’,
source’’, and ‘‘TEQ’’, and adding
§ 63.1201 Definitions and acronyms used Interim Emissions Standards and on their concentration or mass in the
in this subpart. Operating Limits For Incinerators, hazardous waste feed, considering the
(a) * * * Cement Kilns, and Lightweight results of hazardous waste analyses or
Aggregate Kilns other data and information.
Btu means British Thermal Units.
§ 63.1203 What are the standards for
* * * * *
* * * * * ■ 8. Section 63.1206 is amended by:
hazardous waste incinerators that are
Hazardous waste combustor means a effective until compliance with the ■ a. Revising paragraph (a).
hazardous waste incinerator, hazardous standards under § 63.1219? ■ b. Revising paragraphs (b)(1)(ii), (b)(6)
waste burning cement kiln, hazardous * * * * * introductory text, (b)(7)(i)(A), (b)(7)(ii),
waste burning lightweight aggregate (c) * * * (b)(9)(i) introductory text, (b)(9)(i)(A),
kiln, hazardous waste liquid fuel boiler, (3) * * * (b)(9)(iv)(A), (b)(9)(vi), (b)(9)(vii)
hazardous waste solid fuel boiler, or (ii) You must specify one or more introductory text, (b)(9)(viii)(D),
hazardous waste hydrochloric acid POHCs that are representative of the (b)(9)(ix)(D), (b)(10)(i) introductory text,
production furnace. most difficult to destroy organic (b)(10)(i)(A), (b)(10)(vi), (b)(10)(vii)
* * * * * compounds in your hazardous waste introductory text, (b)(10)(viii)(D),
feedstream. You must base this (b)(10)(ix)(D), (b)(11), (b)(13)(i)
Hazardous waste hydrochloric acid introductory text, (b)(13)(ii), and (b)(14).
production furnace and Hazardous specification on the degree of difficulty
■ c. Adding paragraph (b)(16).
Waste HCl production furnace mean a of incineration of the organic
■ d. Revising paragraphs (c)(1)(i)
halogen acid furnace defined under constituents in the hazardous waste and
introductory text, (c)(3)(iv), (c)(6)(iii)(B)
on their concentration or mass in the
§ 260.10 of this chapter that produces introductory text, (c)(6)(iv) introductory
hazardous waste feed, considering the
aqueous hydrochloric acid (HCl) text, and (c)(7).
results of hazardous waste analyses or ■ e. Adding paragraphs (c)(8) and (c)(9).
product and that burns hazardous waste
other data and information. The revisions and additions read as
at any time.
* * * * * follows:
* * * * *
■ 6. The section heading to § 63.1204
Hazardous waste liquid fuel boiler § 63.1206 When and how must you comply
and paragraph (c)(3)(ii) are revised to with the standards and operating
means a boiler defined under § 260.10 of read as follows:
this chapter that does not burn solid requirements?
fuels and that burns hazardous waste at § 63.1204 What are the standards for (a) Compliance dates. (1) Compliance
any time. Liquid fuel boiler includes hazardous waste burning cement kilns that dates for incinerators, cement kilns, and
boilers that only burn gaseous fuel. are effective until compliance with the lightweight aggregate kilns that burn
standards under § 63.1220? hazardous waste. (i) Compliance date
* * * * * for standards under §§ 63.1203,
* * * * *
Hazardous waste solid fuel boiler (c) * * * 63.1204, and 63.1205. (A) Compliance
means a boiler defined under § 260.10 of (3) * * * dates for existing sources. You must
this chapter that burns a solid fuel and (ii) You must specify one or more comply with the emission standards
that burns hazardous waste at any time. POHCs that are representative of the under §§ 63.1203, 63.1204, and 63.1205
* * * * * most difficult to destroy organic and the other requirements of this
compounds in your hazardous waste subpart no later than the compliance
New source means any affected source
feedstream. You must base this date, September 30, 2003, unless the
the construction or reconstruction of
specification on the degree of difficulty Administrator grants you an extension
which is commenced after the dates
of incineration of the organic of time under § 63.6(i) or § 63.1213.
specified under §§ 63.1206(a)(1)(i)(B), (B) New or reconstructed sources. (1)
constituents in the hazardous waste and
(a)(1)(ii)(B), and (a)(2)(ii). If you commenced construction or
on their concentration or mass in the
* * * * * hazardous waste feed, considering the reconstruction of your hazardous waste
System removal efficiency means [1 ¥ results of hazardous waste analyses or combustor after April 19, 1996, you
Emission Rate (mass/time) / Feedrate other data and information. must comply with the emission
(mass/time)] X 100. * * * * * standards under §§ 63.1203, 63.1204,
and 63.1205 and the other requirements
* * * * * ■ 7. The section heading to § 63.1205
of this subpart by the later of September
TEQ means the international method and paragraph (c)(3)(ii) are revised to
30, 1999 or the date the source starts
of expressing toxicity equivalents for read as follows:
operations, except as provided by
dioxins and furans as defined in U.S. § 63.1205 What are the standards for paragraph (a)(1)(i)(B)(2) of this section.
EPA, Interim Procedures for Estimating hazardous waste burning lightweight The costs of retrofitting and replacement
Risks Associated with Exposures to aggregate kilns that are effective until of equipment that is installed
Mixtures of Chlorinated Dibenzo-p- compliance with the standards under specifically to comply with this subpart,
dioxins and -dibenzofurans (CDDs and § 63.1221? between April 19, 1996 and a source’s
CDFs) and 1989 Update, March 1989. * * * * * compliance date, are not considered to
* * * * * (c) * * * be reconstruction costs.
(3) * * * (2) For a standard under §§ 63.1203,
■ 5. Section 63.1203 is amended by: (ii) You must specify one or more 63.1204, and 63.1205 that is more
■ a. Revising an undesignated center POHCs that are representative of the stringent than the standard proposed on
heading above the section heading. most difficult to destroy organic April 19, 1996, you may achieve
■ b. Revising the section heading.
compounds in your hazardous waste compliance no later than September 30,
feedstream. You must base this 2003 if you comply with the standard
■ c. Revising paragraph (c)(3)(2). specification on the degree of difficulty proposed on April 19, 1996 after
The revisions and additions read as of incineration of the organic September 30, 1999. This exception
follows: constituents in the hazardous waste and does not apply, however, to new or
reconstructed area source hazardous the later of October 12, 2005, or the date monitoring system and documenting
waste combustors that become major the source starts operations, except as compliance with the hydrocarbon
sources after September 30, 1999. As provided by paragraph (a)(2)(ii)(B) of standard during the destruction and
provided by § 63.6(b)(7), such sources this section. The costs of retrofitting and removal efficiency (DRE) performance
must comply with the standards under replacement of equipment that is test or its equivalent.
§§ 63.1203, 63.1204, and 63.1205 at installed specifically to comply with * * * * *
startup. this subpart, between April 20, 2004, (7) * * * (i) * * *
(ii) Compliance date for standards and a source’s compliance date, are not (A) You must document compliance
under §§ 63.1219, 63.1220, and 63.1221. considered to be reconstruction costs. with the Destruction and Removal
(A) Compliance dates for existing (B) For a standard in the subpart that Efficiency (DRE) standard under this
sources. You must comply with the is more stringent than the standard subpart only once provided that you do
emission standards under §§ 63.1219, proposed on April 20, 2004, you may not modify the source after the DRE test
63.1220, and 63.1221 and the other achieve compliance no later than in a manner that could affect the ability
requirements of this subpart no later October 14, 2008, if you comply with of the source to achieve the DRE
than the compliance date, October 14, the standard proposed on April 20, standard.
2008, unless the Administrator grants 2004, after October 12, 2005. This
you an extension of time under § 63.6(i) * * * * *
exception does not apply, however, to
or § 63.1213. new or reconstructed area source (ii) Sources that feed hazardous waste
(B) New or reconstructed sources. (1) hazardous waste combustors that at locations other than the normal flame
If you commenced construction or become major sources after October 14, zone. (A) Except as provided by
reconstruction of your hazardous waste 2008. As provided by § 63.6(b)(7), such paragraph (b)(7)(ii)(B) of this section, if
combustor after April 20, 2004, you sources must comply with this subpart you feed hazardous waste at a location
must comply with the new source at startup. in the combustion system other than the
emission standards under §§ 63.1219, (3) Early compliance. If you choose to normal flame zone, then you must
63.1220, and 63.1221 and the other comply with the emission standards of demonstrate compliance with the DRE
requirements of this subpart by the later this subpart prior to the dates specified standard during each comprehensive
of October 12, 2005 or the date the in paragraphs (a)(1) and (a)(2) of this performance test;
source starts operations, except as section, your compliance date is the (B)(1) A cement kiln that feeds
provided by paragraph (a)(1)(ii)(B)(2) of earlier of the date you postmark the hazardous waste at a location other than
this section. The costs of retrofitting and Notification of Compliance under the normal flame zone need only
replacement of equipment that is § 63.1207(j)(1) or the dates specified in demonstrate compliance with the DRE
installed specifically to comply with paragraphs (a)(1) and (a)(2) of this standard during three consecutive
this subpart, between April 20, 2004, comprehensive performance tests
section.
and a source’s compliance date, are not provided that:
(b) * * *
considered to be reconstruction costs. (i) All three tests achieve the DRE
(2) For a standard under §§ 63.1219, (1) * * *
(ii) When hazardous waste is not in standard in this subpart; and
63.1220, and 63.1221 that is more (ii) The design, operation, and
stringent than the standard proposed on the combustion chamber (i.e., the
hazardous waste feed to the combustor maintenance features of each of the
April 20, 2004, you may achieve three tests are similar;
compliance no later than October 14, has been cut off for a period of time not
less than the hazardous waste residence (iii) The data in lieu restriction of
2008, if you comply with the standard § 63.1207(c)(2)(iv) does not apply when
proposed on April 20, 2004, after time) and you have documented in the
operating record that you are complying complying with the provisions of
October 12, 2005. This exception does paragraph (b)(7)(ii)(B) of this section;
not apply, however, to new or with all otherwise applicable
requirements and standards (2) If at any time you change your
reconstructed area source hazardous
promulgated under authority of sections design, operation, and maintenance
waste combustors that become major
112 (e.g., 40 CFR part 63, subparts LLL, features in a manner that could
sources after October 14, 2008. As
DDDDD, and NNNNN) or 129 of the reasonably be expected to affect your
provided by § 63.6(b)(7), such sources
Clean Air Act in lieu of the emission ability to meet the DRE standard, then
must comply with the standards under
standards under §§ 63.1203, 63.1204, you must comply with the requirements
§§ 63.1219, 63.1220, and 63.1221 at
63.1205, 63.1215, 63.1216, 63.1217, of paragraph (b)(7)(ii)(A) of this section.
startup.
(2) Compliance dates for solid fuel 63.1218, 63.1219, 63.1220, and 63.1221; * * * * *
boilers, liquid fuel boilers, and hydrogen the monitoring and compliance (9) * * * (i) You may petition the
chloride production furnaces that burn standards of this section and §§ 63.1207 Administrator to request alternative
hazardous waste for standards under through 63.1209, except the modes of standards to the mercury or hydrogen
§§ 63.1216, 63.1217, and 63.1218. (i) operation requirements of § 63.1209(q); chloride/chlorine gas emission
Compliance date for existing sources. and the notification, reporting, and standards of this subpart, to the
You must comply with the standards of recordkeeping requirements of semivolatile metals emission standards
this subpart no later than the §§ 63.1210 through 63.1212. under §§ 63.1205, 63.1221(a)(3)(ii), or
compliance date, October 14, 2008, * * * * * 63.1221(b)(3)(ii), or to the low volatile
unless the Administrator grants you an (6) Compliance with the carbon metals emissions standards under
extension of time under § 63.6(i) or monoxide and hydrocarbon emission §§ 63.1205, 63.1221(a)(4)(ii), or
§ 63.1213. standards. This paragraph applies to 63.1221(b)(4)(ii) if:
(ii) New or reconstructed sources. (A) sources that elect to comply with the (A) You cannot achieve one or more
If you commenced construction or carbon monoxide and hydrocarbon of these standards while using
reconstruction of your hazardous waste emissions standards of this subpart by maximum achievable control
combustor after October 12, 2005, you documenting continuous compliance technology (MACT) because of raw
must comply with the new source with the carbon monoxide standard material contributions to emissions of
emission standards of this subpart by using a continuous emissions mercury, semivolatile metals, low
volatile metals, or hydrogen chloride/ volatile metals, or hydrogen chloride/ the following alternative metal emission
chlorine gas; or chlorine gas; or control requirements:
* * * * * * * * * * (ii) Alternative metal emission control
(iv) * * * (A) The alternative (vi) You must include data or requirements for existing incinerators.
standard petition you submit under information with semivolatile metals, (A) You must not discharge or cause
paragraph (b)(9)(i)(A) of this section low volatile metals, and hydrogen combustion gases to be emitted into the
must include data or information chloride/chlorine gas alternative atmosphere that contain cadmium, lead,
documenting that raw material standard petitions that you submit and selenium in excess of 240 µg/dscm,
contributions to emissions prevent you under paragraph (b)(10)(i)(A) of this combined emissions, corrected to 7
from complying with the emission section documenting that emissions of percent oxygen; and,
standard even though the source is the regulated metals and hydrogen (B) You must not discharge or cause
using MACT, as defined under chloride/chlorine gas attributable to the combustion gases to be emitted into the
paragraphs (b)(9)(viii) and (ix) of this hazardous waste only will not exceed atmosphere that contain antimony,
section, for the standard for which you the emission standards in this subpart. arsenic, beryllium, chromium, cobalt,
are seeking relief. (vii) You must not operate pursuant to manganese, and nickel in excess of 97
your recommended alternative µg/dscm, combined emissions,
* * * * * corrected to 7 percent oxygen.
standards in lieu of emission standards
(vi) You must include data or specified in this subpart: (iii) Alternative metal emission
information with semivolatile metals, control requirements for new
* * * * *
low volatile metals, and hydrogen incinerators. (A) You must not discharge
(viii) * * *
chloride/chlorine gas alternative (D) For hydrogen chloride/chlorine or cause combustion gases to be emitted
standard petitions that you submit gas, a hazardous waste chlorine feedrate into the atmosphere that contain
under paragraph (b)(9)(i)(A) of this corresponding to an MTEC of 720,000 cadmium, lead, and selenium in excess
section documenting that semivolatile µg/dscm or less. of 24 µg/dscm, combined emissions,
metals, low volatile metals, and (ix) * * * corrected to 7 percent oxygen; and,
hydrogen chloride/chlorine gas (D) For hydrogen chloride/chlorine (B) You must not discharge or cause
emissions attributable to the hazardous gas, a hazardous waste chlorine feedrate combustion gases to be emitted into the
waste only will not exceed the emission corresponding to an MTEC of 420,000 atmosphere that contain antimony,
standards of this subpart. µg/dscm or less. arsenic, beryllium, chromium, cobalt,
(vii) You must not operate pursuant to (11) Calculation of hazardous waste manganese, and nickel in excess of 97
your recommended alternative residence time. You must calculate the µg/dscm, combined emissions,
standards in lieu of emission standards hazardous waste residence time and corrected to 7 percent oxygen.
specified in this subpart: include the calculation in the (iv) Operating limits. Semivolatile and
* * * * * performance test plan under § 63.1207(f) low volatile metal operating parameter
(viii) * * * and the operating record. You must also limits must be established to ensure
provide the hazardous waste residence compliance with the alternative
(D) For hydrogen chloride/chlorine
time in the Documentation of emission limitations described in
gas, a hazardous waste chlorine feedrate
Compliance under § 63.1211(c) and the paragraphs (e)(2) and (e)(3) of this
corresponding to an MTEC of 2,000,000
Notification of Compliance under section pursuant to § 63.1209(n), except
µg/dscm or less, and use of an air
§§ 63.1207(j) and 63.1210(d). that semivolatile metal feedrate limits
pollution control device with a
* * * * * apply to lead, cadmium, and selenium,
hydrogen chloride/chlorine gas removal
(13) * * * combined, and low volatile metal
efficiency of 85 percent or greater.
(i) Cement kilns that feed hazardous feedrate limits apply to arsenic,
(ix) * * * beryllium, chromium, antimony, cobalt,
waste at a location other than the end
(D) For hydrogen chloride/chlorine where products are normally discharged manganese, and nickel, combined.
gas, a hazardous waste chlorine feedrate and where fuels are normally fired must * * * * *
corresponding to an MTEC of comply with the carbon monoxide and (16) Compliance with subcategory
14,000,000 µg/dscm or less, and use of hydrocarbon standards of this subpart as standards for liquid fuel boilers. You
an air pollution control device with a follows: must comply with the mercury,
hydrogen chloride/chlorine gas removal semivolatile, low volatile metal, and
* * * * *
efficiency of 99.6 percent or greater. total chlorine standards for liquid fuel
(ii) Lightweight aggregate kilns that
(10) * * * (i) You may petition the feed hazardous waste at a location other boilers under § 63.1217 as follows:
Administrator to request alternative than the end where products are (i) You must determine the as-fired
standards to the mercury or hydrogen normally discharged and where fuels heating value of each batch of hazardous
chloride/chlorine gas emission are normally fired must comply with the waste fired by each firing system of the
standards of this subpart, to the hydrocarbon standards of this subpart as boiler so that you know the mass-
semivolatile metals emission standards follows: weighted heating value of the hazardous
under §§ 63.1204, 63.1220(a)(3)(ii), or (A) Existing sources must comply waste fired at all times.
63.1220(b)(3)(ii), or to the low volatile with the 20 parts per million by volume (ii) If the as-fired heating value of the
metals emissions standards under hydrocarbon standard of this subpart; hazardous waste is 10,000 Btu per
§§ 63.1204, 63.1220(a)(4)(ii), or (B) New sources must comply with pound or greater, you are subject to the
63.1220(b)(4)(ii) if: the 20 parts per million by volume thermal emission concentration
(A) You cannot achieve one or more hydrocarbon standard of this subpart. standards (lb/million Btu) under
of these standards while using (14) Alternative to the particulate § 63.1217.
maximum achievable control matter standard for incinerators. (i). (iii) If the as-fired heating value of the
technology (MACT) because of raw General. In lieu of complying with the hazardous waste is less than 10,000 Btu/
material contributions to emissions of particulate matter standards under lb, you are subject to the mass or
mercury, semivolatile metals, low § 63.1203, you may elect to comply with volume emission concentration
standards (µg/dscm or ppmv) under pollution control equipment, that could (F) Following initial adjustment, you
§ 63.1217. affect emissions of regulated hazardous must not adjust the sensitivity or range,
(iv) If the as-fired heating value of air pollutants. averaging period, alarm set points, or
hazardous wastes varies above and (ii) The plan must prescribe how you alarm delay time, except as detailed in
below 10,000 Btu/lb over time, you are will operate and maintain the the operation and maintenance plan
subject to the thermal concentration combustor in a manner consistent with required under paragraph (c)(7) of this
standards when the heating value is good air pollution control practices for section. You must not increase the
10,000 Btu/lb or greater and the mass minimizing emissions at least to the sensitivity by more than 100 percent or
concentration standards when the levels achieved during the decrease the sensitivity by more than 50
heating value is less than 10,000 Btu/lb. comprehensive performance test. percent over a 365 day period unless
You may elect to comply at all times (iii) This plan ensures compliance such adjustment follows a complete
with the more stringent operating with the operation and maintenance baghouse inspection which
requirements that ensure compliance requirements of § 63.6(e) and minimizes demonstrates the baghouse is in good
with both the thermal emission emissions of pollutants, automatic waste operating condition;
concentration standards and the mass or feed cutoffs, and malfunctions. (G) For negative pressure or induced
volume emission concentration (iv) You must record the plan in the air baghouses, and positive pressure
standards. operating record. baghouses that are discharged to the
* * * * * (8) Bag leak detection system atmosphere through a stack, the bag leak
(c) * * * (1) * * * (i) You must requirements. (i) If your combustor is detector shall be installed downstream
operate only under the operating equipped with a baghouse (fabric filter), of the baghouse and upstream of any
requirements specified in the you must continuously operate either: wet acid gas scrubber; and
Documentation of Compliance under (A) A bag leak detection system that (H) Where multiple detectors are
§ 63.1211(c) or the Notification of meets the specifications and required, the system’s instrumentation
Compliance under §§ 63.1207(j) and requirements of paragraph (c)(8)(ii) of and alarm system may be shared among
63.1210(d), except: this section and you must comply with the detectors.
* * * * * the corrective measures and notification (iii) Bag leak detection system
(3) * * * requirements of paragraphs (c)(8)(iii) corrective measures requirements. The
(iv) Failure of the AWFCO system. If and (iv) of this section; or operating and maintenance plan
the AWFCO system fails to (B) A particulate matter detection required by paragraph (c)(7) of this
automatically and immediately cutoff system under paragraph (c)(9) of this section must include a corrective
the flow of hazardous waste upon section. measures plan that specifies the
exceedance of a parameter required to (ii) Bag leak detection system procedures you will follow in the case
be interlocked with the AWFCO system specification and requirements. (A) The of a bag leak detection system alarm.
under paragraph (c)(3)(i) of this section, bag leak detection system must be The corrective measures plan must
you have failed to comply with the certified by the manufacturer to be include, at a minimum, the procedures
AWFCO requirements of paragraph capable of continuously detecting and used to determine and record the time
(c)(3) of this section. If an equipment or recording particulate matter emissions and cause of the alarm as well as the
other failure prevents immediate and at concentrations of 1.0 milligrams per corrective measures taken to correct the
automatic cutoff of the hazardous waste actual cubic meter unless you control device malfunction or minimize
feed, however, you must cease feeding demonstrate, under § 63.1209(g)(1), that emissions as specified below. Failure to
hazardous waste as quickly as possible. a higher detection limit would routinely initiate the corrective measures required
* * * * * detect particulate matter loadings by this paragraph is failure to ensure
(6) * * * during normal operations; compliance with the emission standards
(iii) * * * (B) The bag leak detection system in this subpart.
(B) Be trained under the requirements shall provide output of relative or (A) You must initiate the procedures
of, and certified under, one of the absolute particulate matter loadings; used to determine the cause of the alarm
following American Society of within 30 minutes of the time the alarm
(C) The bag leak detection system
Mechanical Engineers (ASME) first sounds; and
shall be equipped with an alarm system
standards: QHO–1–1994, QHO–1a– (B) You must alleviate the cause of the
that will sound an audible alarm when
1996, or QHO–1–2004 (Standard for the alarm by taking the necessary corrective
an increase in relative particulate
Qualification and Certification of measure(s) which may include, but are
loadings is detected over a preset level;
Hazardous Waste Incinerator Operators). not to be limited to, the following:
(D) The bag leak detection system (1) Inspecting the baghouse for air
If you elect to use the ASME program: shall be installed and operated in a leaks, torn or broken filter elements, or
* * * * * manner consistent with available any other malfunction that may cause
(iv) Control room operators of cement written guidance from the U.S. an increase in emissions;
kilns, lightweight aggregate kilns, solid Environmental Protection Agency or, in (2) Sealing off defective bags or filter
fuel boilers, liquid fuel boilers, and the absence of such written guidance, media;
hydrochloric acid production furnaces the manufacturer’s written (3) Replacing defective bags or filter
must be trained and certified under: specifications and recommendations for media, or otherwise repairing the
* * * * * installation, operation, and adjustment control device;
(7) Operation and maintenance of the system; (4) Sealing off a defective baghouse
plan—(i) You must prepare and at all (E) The initial adjustment of the compartment;
times operate according to an operation system shall, at a minimum, consist of (5) Cleaning the bag leak detection
and maintenance plan that describes in establishing the baseline output by system probe, or otherwise repairing the
detail procedures for operation, adjusting the sensitivity (range) and the bag leak detection system; or
inspection, maintenance, and corrective averaging period of the device, and (6) Shutting down the combustor.
measures for all components of the establishing the alarm set points and the (iv) Excessive exceedances
combustor, including associated alarm delay time; notification. If you operate the
combustor when the detector response particulate loadings is detected over the (A) You may extrapolate the detector
exceeds the alarm set-point more than 5 set-point response up to a particulate matter
percent of the time during any 6-month (D) You must install, operate, and concentration that is 50% of the
block time period, you must submit a maintain the particulate matter particulate matter emission standard or
notification to the Administrator within detection system in a manner consistent 125% of the highest particulate matter
30 days of the end of the 6-month block with the provisions of paragraph (c)(9) concentration used to develop the
time period that describes the causes of of this section and available written correlation under paragraph (c)(9)(iii)(B)
the exceedances and the revisions to the guidance from the U.S. Environmental of this section, whichever is greater. The
design, operation, or maintenance of the Protection Agency or, in the absence of extrapolated emission concentration
combustor or baghouse you are taking to such written guidance, the must not exceed the particulate matter
minimize exceedances. To document manufacturer’s written specifications emission standard.
compliance with this requirement: and recommendations for installation, (B) To establish an approximate
(A) You must keep records of the date, operation, maintenance and quality correlation of the detector response to
time, and duration of each alarm, the assurance of the system; particulate matter emission
time corrective action was initiated and (E) You must include procedures for concentrations, you should use as
completed, and a brief description of the installation, operation, maintenance, guidance Performance Specification-11
cause of the alarm and the corrective and quality assurance of the particulate for PM CEMS (40 CFR Part 60,
action taken; matter detection system in the site- Appendix B), except that you need only
(B) You must record the percent of the specific continuous monitoring system conduct 5 runs to establish the initial
operating time during each 6-month test plan required under § 63.8(e)(3) of correlation under Section 8.6 of
period that the alarm sounds; this chapter. Performance Specification 11.
(C) In calculating the operating time (F) Where multiple detectors are (C) For quality assurance, you should
use as guidance Procedure 2 of
percentage, if inspection of the fabric required to monitor multiple control
Appendix F to Part 60 of this chapter
filter demonstrates that no corrective devices, the system’s instrumentation
and the detector manufacturer’s
action is required, no alarm time is and alarm system may be shared among
recommended procedures for periodic
counted; and the detectors.
quality assurance checks and tests,
(D) If corrective action is required, (G) You must establish the alarm set-
except that:
each alarm shall be counted as a point as provided by either paragraph (1) You must conduct annual Relative
minimum of 1 hour. (c)(9)(ii) or paragraph (c)(9)(iii) of this Response Audits as prescribed by
(9) Particulate matter detection section. Procedure 2 of Appendix F to Part 60 of
system requirements for electrostatic (ii) Establishing the alarm set-point this chapter (Section 10.3(6));
precipitators and ionizing wet without extrapolation. (A) The alarm (2) You need only conduct Relative
scrubbers. If your combustor is set-point is the average of the test run Response Audits on a 3-year interval
equipped with an electrostatic averages of the detector response after passing two sequential annual
precipitator or ionizing wet scrubber, achieved during the comprehensive Relative Response Audits.
and you elect not to establish under performance test demonstrating (D) An exceedance of the particulate
§ 63.1209(m)(1)(iv) site-specific control compliance with the particulate matter matter emission standard by a
device operating parameter limits that emission standard. particulate matter detection system for
are linked to the automatic waste feed (B) During the comprehensive which particulate emission
cutoff system under paragraph (c)(3) of performance test, you may simulate concentrations have been approximately
this section, you must continuously emission concentrations at the upper correlated with the detector response
operate a particulate matter detection end of the range of normal operations by under paragraph (c)(9)(iii) of this section
system that meets the specifications and means including feeding high levels of is not evidence that the standard has
requirements of paragraph (c)(9)(i) ash and detuning the emission control been exceeded. The approximate
through (iii) of this section and you equipment. correlation is used for compliance
must comply with the corrective (C) You must comply with the alarm assurance to determine when corrective
measures and notification requirements set-point on a 6-hour rolling average, measures must be taken rather than for
of paragraphs (c)(9)(iv) through (v) of updated each hour with a one-hour compliance monitoring.
this section. block average that is the average of the (iv) Particulate matter detection
(i) Particulate matter detection system detector responses over each 15-minute system corrective measures
requirements.—(A) The particulate block; requirements. The operating and
matter detection system must be (iii) Establishing the alarm set-point maintenance plan required by paragraph
certified by the manufacturer to be with extrapolation. You may extrapolate (c)(7) of this section must include a
capable of continuously detecting and the average of the test run averages of corrective measures plan that specifies
recording particulate matter emissions the detector response achieved during the procedures you will follow in the
at concentrations of 1.0 milligrams per the comprehensive performance test as case of a particulate matter detection
actual cubic meter unless you provided by paragraph (c)(9)(iii)(A) of system alarm. The corrective measures
demonstrate, under § 63.1209(g)(1), that this section to establish an alarm level plan must include, at a minimum, the
a higher detection limit would routinely after you approximate the correlation of procedures used to determine and
detect particulate matter loadings the detector response to particulate record the time and cause of the alarm
during normal operations; matter concentration as prescribed by as well as the corrective measures taken
(B) The particulate matter detector paragraph (c)(9)(iii)(B) of this section. to correct the control device
shall provide output of relative or You must comply with the extrapolated malfunction or minimize emissions as
absolute particulate matter loadings; alarm set-point on a 6-hour rolling specified below. Failure to initiate the
(C) The particulate matter detection average, updated each hour with a one- corrective measures required by this
system shall be equipped with an alarm hour block average that is the average of paragraph is failure to ensure
system that will sound an audible alarm the detector responses over each 15- compliance with the emission standards
when an increase in relative or absolute minute block. in this subpart.
(A) You must initiate the procedures compliance with the emission standards (2) * * * (iii) The data in lieu test age
used to determine the cause of the alarm provided by this subpart, establish restriction provided in paragraph
within 30 minutes of the time the alarm limits for the operating parameters (c)(2)(i)(A) of this section does not apply
first sounds; and provided by § 63.1209, and demonstrate for the duration of the interim standards
(B) You must alleviate the cause of the compliance with the performance (i.e., the standards published in the
alarm by taking the necessary corrective specifications for continuous Federal Register on February 13, 2002,
measure(s) which may include shutting monitoring systems. 67 FR 6792). See 40 CFR parts 63, 264,
down the combustor. * * * * * 265, 266, 270, and 271 revised as of July
(v) Excessive exceedances (3) One-Time Dioxin/Furan Test for 1, 2002. Paragraph (c)(2)(i)(A) of this
notification. If you operate the Sources Not Subject to a Numerical section does not apply until EPA
combustor when the detector response Dioxin/Furan Standard. For solid fuel promulgates permanent replacement
exceeds the alarm set-point more than 5 boilers and hydrochloric acid standards pursuant to the Settlement
percent of the time during any 6-month production furnaces, for lightweight Agreement noticed in the Federal
block time period, you must submit a aggregate kilns that are not subject to a Register on November 16, 2001 (66 FR
notification to the Administrator within numerical dioxin/furan emission 57715).
30 days of the end of the 6-month block standard under § 63.1221, and liquid * * * * *
time period that describes the causes of fuel boilers that are not subject to a (3) For incinerators, cement kilns, and
the exceedances and the revisions to the numerical dioxin/furan emission lightweight aggregate kilns, you must
design, operation, or maintenance of the standard under § 63.1217, you must commence the initial comprehensive
combustor or emission control device conduct a one-time emission test for performance test to demonstrate
you are taking to minimize exceedances. dioxin/furan under feed and operating compliance with the standards under
To document compliance with this conditions that are most likely to reflect §§ 63.1219, 63.1220, and 63.1221 not
requirement: daily maximum operating variability, later than 12 months after the
(A) You must keep records of the date, similar to a dioxin/furan comprehensive compliance date.
time, and duration of each alarm, the performance test.
time corrective action was initiated and (i) You must conduct the dioxin/furan (d) * * *
completed, and a brief description of the emissions test no later than the deadline (4) * * * (i) Waiver of periodic
cause of the alarm and the corrective for conducting the initial comprehensive performance tests.
action taken; comprehensive performance test. Except as provided in paragraph (c)(2)
(B) You must record the percent of the (ii) You may use dioxin/furan of this section, you must conduct only
operating time during each 6-month emissions data from previous testing to an initial comprehensive performance
period that the alarm sounds; meet this requirement, provided that: test under the interim standards (i.e.,
(C) In calculating the operating time (A) The testing was conducted under the standards published in the Federal
percentage, if inspection of the emission feed and operating conditions that are Register on February 13, 2002); all
control device demonstrates that no most likely to reflect daily maximum subsequent comprehensive performance
corrective action is required, no alarm operating variability, similar to a testing requirements are waived under
time is counted; and dioxin/furan compliance test; the interim standards. The provisions in
(D) If corrective action is required, (B) You have not changed the design the introductory text to paragraph (d)
each alarm shall be counted as a or operation of the source in a manner and in paragraph (d)(1) of this section
minimum of 1 hour. that could significantly affect stack gas do not apply until EPA promulgates
■ 9. Section 63.1207 is amended by:
dioxin/furan emission concentrations; permanent replacement standards
■ a. Revising paragraph (b)(1). and pursuant to the Settlement Agreement
■ b. Adding paragraph (b)(3).
(C) The data meet quality assurance noticed in the Federal Register on
■ c. Revising paragraphs (c)(1) and objectives that may be determined on a November 16, 2001.
(c)(2)(iii). site-specific basis. * * * * *
■ d. Adding paragraph (c)(3).
(iii) You may use dioxin/furan
emissions data from a source to (e) * * *
■ e. Revising paragraph (d)(4)(i).
represent emissions from another on- (2) You must make your site-specific
■ f. Revising paragraphs (e)(2) and
site source in lieu of testing (i.e., data in test plan and CMS performance
(e)(3)(iv). evaluation test plan available to the
■ g. Revising paragraphs (f)(1)(ii)(D), lieu of testing) if the design and
operation, including hazardous waste public for review no later than 60
(f)(1)(x) introductory text, (f)(1)(xiii),
feed and other feedstreams, of the calendar days before initiation of the
(f)(1)(xiv), (f)(1)(xvi), and (f)(1)(xxv).
sources are identical. test. You must issue a public notice to
■ h. Adding paragraph (f)(1)(xv).
(iv) You must include the results of all persons on your facility/public
■ i. Revising paragraph (h)(2)(i).
the one-time dioxin/furan emissions test mailing list (developed pursuant to 40
■ j. Revising paragraph (j)(3).
■ k. Revising paragraph (l)(1) with the results of the initial CFR 70.7(h), 71.11(d)(3)(i)(E) and
introductory text. comprehensive performance test in the 124.10(c)(1)(ix)) announcing the
■ l. Revising paragraph (m)(2) Notification of Compliance. availability of the test plans and the
introductory text. (v) You must repeat the dioxin/furan location where the test plans are
The revisions and additions read as emissions test if you change the design available for review. The test plans must
follows: or operation of the source in a manner be accessible to the public for 60
that may increase dioxin/furan calendar days, beginning on the date
§ 63.1207 What are the performance emissions. that you issue your public notice. The
testing requirements? (c) * * * (1) Test date. Except as location must be unrestricted and
* * * * * provided by paragraphs (c)(2) and (c)(3) provide access to the public during
(b) * * * of this section, you must commence the reasonable hours and provide a means
(1) Comprehensive performance test. initial comprehensive performance test for the public to obtain copies. The
You must conduct comprehensive not later than six months after the notification must include the following
performance tests to demonstrate compliance date. information at a minimum:
(i) The name and telephone number of performance test plan, and provide the (2) You are not required to conduct
the source’s contact person; information required by the emission performance tests to document
(ii) The name and telephone number averaging provision; compliance with the mercury,
of the regulatory agency’s contact (xiv) For preheater or preheater/ semivolatile metals, low volatile metals,
person; precalciner cement kilns with dual or hydrogen chloride/chlorine gas
(iii) The location where the test plans stacks, if you elect to use the emissions emission standards under the
and any necessary supporting averaging provision of this subpart, you conditions specified in this paragraph
documentation can be reviewed and must notify the Administrator of your (m)(2). You are deemed to be in
copied; intent in the initial (and subsequent) compliance with an emission standard
(iv) The time period for which the test comprehensive performance test plan, if the twelve-hour rolling average
plans will be available for public and provide the information required by maximum theoretical emission
review; and the emission averaging provision; concentration (MTEC) does not exceed
(v) An expected time period for (xv) If you request to use Method 23 the emission standard:
commencement and completion of the for dioxin/furan you must provide the * * * * *
performance test and CMS performance information required under ■ 10. Section 63.1208 is amended by
evaluation test. § 63.1208(b)(1)(i)(B); removing and reserving paragraph (a)
(3) * * * (xvi) If you are not required to and revising paragraphs (b)(1)(i) and
(iv) Public notice. At the same time conduct performance testing to (b)(5) to read as follows:
that you submit your petition to the document compliance with the
Administrator, you must notify the mercury, semivolatile metals, low § 63.1208 What are the test methods?
public (e.g., distribute a notice to the volatile metals, or hydrogen chloride/ (a) [Reserved]
facility/public mailing list developed chlorine gas emission standards under (b) * * *
pursuant to 40 CFR 70.7(h), paragraph (m) of this section, you must (1) * * * (i) To determine compliance
71.11(d)(3)(i)(E) and 124.10(c)(1)(ix)) of include with the comprehensive with the emission standard for dioxins
your petition to waive a performance performance test plan documentation of and furans, you must use:
test. The notification must include all of compliance with the provisions of that (A) Method 0023A, Sampling Method
the following information at a section. for Polychlorinated Dibenzo-p-Dioxins
minimum: * * * * * and Polychlorinated Dibenzofurans
(A) The name and telephone number (xxv) If your source is equipped with emissions from Stationary Sources, EPA
of the source’s contact person; a dry scrubber to control hydrogen Publication SW–846 (incorporated by
(B) The name and telephone number chloride and chlorine gas, you must reference— see § 63.14); or
of the regulatory agency’s contact document in the comprehensive (B) Method 23, provided in appendix
person; performance test plan key parameters A, part 60 of this chapter, after approval
(C) The date the source submitted its that affect adsorption, and the limits by the Administrator.
site-specific performance test plan and you establish for those parameters based (1) You may request approval to use
CMS performance evaluation test plans; on the sorbent used during the Method 23 in the performance test plan
and performance test, if you elect not to required under § 63.1207(e)(i) and (ii).
(D) The length of time requested for specify and use the brand and type of (2) In determining whether to grant
the waiver. sorbent used during the comprehensive approval to use Method 23, the
(f) * * * performance test, as required by Administrator may consider factors
(1) * * * § 63.1209(o)(4)(iii)(A); and including whether dioxin/furan were
(ii) * * * detected at levels substantially below
(D) The Administrator may approve * * * * *
the emission standard in previous
on a case-by-case basis a hazardous (h) * * *
testing, and whether previous Method
waste feedstream analysis for organic (2) * * *
0023 analyses detected low levels of
hazardous air pollutants in lieu of the (i) Operations when stack emissions
dioxin/furan in the front half of the
analysis required under paragraph testing for dioxin/furan, mercury,
sampling train.
(f)(1)(ii)(A) of this section if the reduced semivolatile metals, low volatile metals,
(3) Sources that emit carbonaceous
analysis is sufficient to ensure that the particulate matter, or hydrogen
particulate matter, such as coal-fired
POHCs used to demonstrate compliance chloride/chlorine gas is being
boilers, and sources equipped with
with the applicable DRE standards of performed; and
activated carbon injection, will be
this subpart continue to be * * * * * deemed not suitable for use of Method
representative of the most difficult to (j) * * * 23 unless you document that there
destroy organic compounds in your (3) See §§ 63.7(g), 63.9(h), and would not be a significant improvement
hazardous waste feedstreams; 63.1210(d) for additional requirements in quality assurance with Method
* * * * * pertaining to the Notification of 0023A.
(x) If you are requesting to extrapolate Compliance (e.g., you must include
* * * * *
metal feedrate limits from results of performance tests in the (5) Hydrogen chloride and chlorine
comprehensive performance test levels Notification of Compliance). gas—(i) Compliance with MACT
under §§ 63.1209(l)(1)(v) or * * * * * standards. To determine compliance
63.1209(n)(2)(vii): (l) Failure of performance test—(1) with the emission standard for hydrogen
* * * * * Comprehensive performance test. The chloride and chlorine gas (combined),
(xiii) For cement kilns with in-line provisions of this paragraph do not you must use:
raw mills, if you elect to use the apply to the initial comprehensive (A) Method 26/26A as provided in
emissions averaging provision of this performance test if you conduct the test appendix A, part 60 of this chapter; or
subpart, you must notify the prior to your compliance date. (B) Methods 320 or 321 as provided
Administrator of your intent in the * * * * * in appendix A, part 63 of this chapter,
initial (and subsequent) comprehensive (m) * * * or
(C) ASTM D 6735–01, Standard Test this section. ASTM D 6735–01 is (1) A test must include three or more
Method for Measurement of Gaseous available for purchase from at least one runs in which a pair of samples is
Chlorides and Fluorides from Mineral of the following addresses: American obtained simultaneously for each run
Calcining Exhaust Sources—Impinger Society for Testing and Materials according to section 11.2.6 of ASTM
Method to measure emissions of (ASTM), 100 Barr Harbor Drive, Post Method D6735–01.
hydrogen chloride, and Method 26/26A Office Box C700, West Conshohocken, (2) You must calculate the test run
to measure emissions of chlorine gas, PA 19428–2959; or ProQuest, 300 North standard deviation of each set of paired
provided that you follow the provisions Zeeb Road, Ann Arbor, MI 48106. samples to quantify data precision,
in paragraphs (b)(5)(C)(1) through (6) of according to Equation 1 of this section:
 C1 − C2a 
RSD a = (100 ) AbsoluteValue  a  (Eq. 1)
 C1a + C2a 
Where: (A) For cement kilns and sources §§ 63.1204(a)(7) and (b)(7) at each point
RSDa = The test run relative standard equipped with a dry acid gas scrubber, where emissions are vented from these
deviation of sample pair a, percent. you must use Methods 320 or 321 as affected sources including the bypass
C1a and C2a = The HCl concentrations, provided in appendix A, part 63 of this stack of a preheater or preheater/
milligram/dry standard cubic meter chapter, or ASTM D 6735–01 to measure precalciner kiln with dual stacks.
(mg/dscm), from the paired hydrogen chloride, and the back-half, (B) Cement kilns under § 63.1220—
samples. caustic impingers of Method 26/26A as Except as provided by paragraphs
(3) You must calculate the test average provided in appendix A, part 60 of this (a)(1)(iv) and (a)(1)(v) of the section and
relative standard deviation according to chapter to measure chlorine gas; and unless your source is equipped with a
Equation 2 of this section: (B) For incinerators, boilers, and bag leak detection system under
lightweight aggregate kilns, you must § 63.1206(c)(8) or a particulate matter
p use Methods 320 or 321 as provided in detection system under § 63.1206(c)(9),
∑ RSD a appendix A, part 63 of this chapter, or you must use a COMS to demonstrate
RSDTA = a =1
( Eq. 2 ) ASTM D 6735–01 to measure hydrogen
chloride, and Method 26/26A as
and monitor compliance with the
p opacity standard under §§ 63.1220(a)(7)
provided in appendix A, part 60 of this and (b)(7) at each point where emissions
Where:
chapter to measure total chlorine, and are vented from these affected sources
RSDTA = The test average relative calculate chlorine gas by difference if: including the bypass stack of a
standard deviation, percent. (1) The bromine/chlorine ratio in preheater or preheater/precalciner kiln
RSDa = The test run relative standard feedstreams is greater than 5 percent; or with dual stacks.
deviation for sample pair a. (2) The sulfur/chlorine ratio in (C) You must maintain and operate
p = The number of test runs, ≥3. feedstreams is greater than 50 percent. each COMS in accordance with the
(4) If RSDTA is greater than 20 requirements of § 63.8(c) except for the
* * * * *
percent, the data are invalid and the test requirements under § 63.8(c)(3). The
must be repeated. ■ 11. Section 63.1209 is amended by:
■ a. Revising paragraphs (a)(1)(ii), requirements of § 63.1211(c) shall be
(5) The post-test analyte spike
(a)(1)(iv)(D), (a)(1)(v)(D), and (a)(5). complied with instead of § 63.8(c)(3);
procedure of section 11.2.7 of ASTM
■ b. Revising paragraph (b)(2)(ii). and
Method D6735–01 is conducted, and the (D) Compliance is based on a six-
■ c. Revising the heading of paragraph
percent recovery is calculated according minute block average.
to section 12.6 of ASTM Method (g)(1) introductory text and paragraph
D6735–01. (g)(1)(i). * * * * *
■ d. Adding paragraph (g)(1)(iv). (iv) * * *
(6) If the percent recovery is between
■ e. Revising paragraphs (k)(1)(i) and (D) To remain in compliance, all six-
70 percent and 130 percent, inclusive,
the test is valid. If the percent recovery (k)(2)(i). minute block averages must not exceed
■ f. Revising paragraph (l)(1). the opacity standard.
is outside of this range, the data are
■ g. Revising paragraphs (m)(1)(iv) (v) * * *
considered invalid, and the test must be
introductory text and (m)(3). (D) To remain in compliance, all six-
repeated.
■ h. Revising paragraph (n)(2). minute block averages must not exceed
(ii) Compliance with risk-based limits
■ i. Revising the heading of paragraph the opacity standard.
under § 63.1215. To demonstrate
(o) introductory text and paragraph * * * * *
compliance with emission limits
(o)(1). (5) Petitions to use CEMS for other
established under § 63.1215, you must
■ j. Adding paragraph (r).
use Method 26/26A as provided in standards. You may petition the
The revisions read as follows: Administrator to use CEMS for
appendix A, part 60 of this chapter,
Method 320 as provided in appendix A, § 63.1209 What are the monitoring compliance monitoring for particulate
part 63 of this chapter, Method 321 as requirements? matter, mercury, semivolatile metals,
provided in appendix A, part 63 of this (a) * * * low volatile metals, and hydrogen
chapter, or ASTM D 6735–01, Standard (1) * * * chloride and chlorine gas under § 63.8(f)
ER12OC05.002</MATH>
Test Method for Measurement of (ii) (A) Cement kilns under in lieu of compliance with the
Gaseous Chlorides and Fluorides from § 63.1204—Except as provided by corresponding operating parameter
Mineral Calcining Exhaust Sources— paragraphs (a)(1)(iv) and (a)(1)(v) of the limits under this section.
Impinger Method (following the section, you must use a COMS to * * * * *
provisions of paragraphs (b)(5)(C)(1) demonstrate and monitor compliance (b) * * *
ER12OC05.001</MATH>
through (6) of this section), except: with the opacity standard under (2) * * *
(ii) Accuracy and calibration of establish the hourly rolling average limit thermal concentration that is updated
weight measurement devices for as the average of the test run averages. each hour.
activated carbon injection systems. If * * * * * (5) If you select an averaging period
you operate a carbon injection system, (2) * * * (i) For sources other than for the feedrate limit that is greater than
the accuracy of the weight measurement cement kilns, you must measure the a 12-hour rolling average, you must
device must be ± 1 percent of the weight temperature of each combustion calculate the initial rolling average as
being measured. The calibration of the chamber at a location that best though you had selected a 12-hour
device must be verified at least once represents, as practicable, the bulk gas rolling average, as provided by
each calendar quarter at a frequency of temperature in the combustion zone. paragraph (b)(5)(i) of this section. You
approximately 120 days. You must document the temperature must calculate rolling averages
* * * * * measurement location in the test plan thereafter as the average of the available
(g) * * * you submit under §§ 63.1207(e) and (f); one-minute values until enough one-
(1) Requests to use alternatives to minute values are available to calculate
* * * * *
operating parameter monitoring (l) Mercury. * * * the rolling average period you select. At
requirements. (i) You may submit an (1) Feedrate of mercury. (i) For that time and thereafter, you update the
application to the Administrator under incinerators and solid fuel boilers, when rolling average feedrate each hour with
this paragraph for approval of complying with the mercury emission a 60-minute average feedrate.
alternative operating parameter (C) If you burn hazardous waste with
standards under §§ 63.1203, 63.1216
monitoring requirements to document a heating value of less than 10,000 Btu/
and 63.1219, you must establish a 12-
compliance with the emission standards lb, you must calculate the mercury
hour rolling average limit for the total
of this subpart. For requests to use feedrate limit as follows:
feedrate of mercury in all feedstreams as (1) You must calculate the mercury
additional CEMS, however, you must the average of the test run averages.
use paragraph (a)(5) of this section and feedrate limit as the mercury emission
(ii) For liquid fuel boilers, when standard divided by [1 ¥ System
§ 63.8(f). Alternative requests to complying with the mercury emission
operating parameter monitoring Removal Efficiency].
standards of § 63.1217, you must (2) The feedrate limit is expressed as
requirements that include unproven establish a rolling average limit for the
monitoring methods may not be made a mass concentration per unit volume of
mercury feedrate as follows on an stack gas (µg/dscm) and is converted to
under this paragraph and must be made averaging period not to exceed an
under § 63.8(f). a mass feedrate (lb/hr) by multiplying it
annual rolling average: by the average stack gas flowrate of the
* * * * * (A) You must calculate a mercury test run averages.
(iv) Dual Standards that incorporate system removal efficiency for each test (3) You must comply with the
the Interim Standards for HAP metals. run and calculate the average system feedrate limit by determining the
(A) Semivolatile and Low Volatile removal efficiency of the test run mercury feedrate (lb/hr) at least once a
Metals. You may petition the averages. If emissions exceed the minute to calculate a 60-minute average
Administrator to waive a feedrate mercury emission standard during the feedrate.
operating parameter limit under comprehensive performance test, it is (4) You must update the rolling
paragraph (n)(2) of this section for either not a violation because the averaging average feedrate each hour with this 60-
the emission standards expressed in a period for the mercury emission minute feedrate measurement.
thermal emissions format or the interim standard is (not-to-exceed) one year and (5) If you select an averaging period
standards based on documentation that compliance is based on compliance for the feedrate limit that is greater than
the feedrate operating parameter limit is with the mercury feedrate limit with an a 12-hour rolling average, you must
not needed to ensure compliance with averaging period not-to-exceed one year. calculate the initial rolling average as
the relevant standard on a continuous (B) If you burn hazardous waste with though you had selected a 12-hour
basis. a heating value of 10,000 Btu/lb or rolling average, as provided by
(B) Mercury. You may petition the greater, you must calculate the mercury paragraph (b)(5)(i) of this section. You
Administrator to waive a feedrate feedrate limit as follows: must calculate rolling averages
operating parameter limit under (1) The mercury feedrate limit is the thereafter as the average of the available
paragraph (l)(1) of this section for either emission standard divided by [1 ¥ one-minute values until enough one-
the feed concentration standard under system removal efficiency]. minute values are available to calculate
§§ 63.1220(a)(2)(i) and (b)(2)(i) or the (2) The mercury feedrate limit is a the rolling average period you select. At
interim standards based on hazardous waste thermal concentration that time and thereafter, you update the
documentation that the feedrate limit expressed as pounds of mercury in rolling average feedrate each hour with
operating parameter limit is not needed hazardous waste feedstreams per a 60-minute average feedrate.
to ensure compliance with the relevant million Btu of hazardous waste fired. (D) If your boiler is equipped with a
standard on a continuous basis. (3) You must comply with the wet scrubber, you must comply with the
* * * * * hazardous waste mercury thermal following unless you document in the
(k) * * * concentration limit by determining the performance test plan that you do not
(1) * * * (i) For sources other than a feedrate of mercury in all hazardous feed chlorine at rates that may
lightweight aggregate kiln, if the waste feedstreams (lb/hr) at least once a substantially affect the system removal
combustor is equipped with an minute and the hazardous waste efficiency of mercury for purposes of
electrostatic precipitator, baghouse thermal feedrate (MM Btu/hr) at least establishing a mercury feedrate limit
(fabric filter), or other dry emissions once a minute to calculate a 60-minute based on the system removal efficiency
control device where particulate matter average thermal emission concentration during the test:
is suspended in contact with as [hazardous waste mercury feedrate (1) Scrubber blowdown must be
combustion gas, you must establish a (lb/hr) / hazardous waste thermal minimized during a pretest conditioning
limit on the maximum temperature of feedrate (MM Btu/hr)]. period and during the performance test:
the gas at the inlet to the device on an (4) You must calculate a rolling (2) Scrubber water must be
hourly rolling average. You must average hazardous waste mercury preconditioned so that mercury in the
water is at equilibrium with stack gas at mercury from all hazardous waste limits as specified in paragraphs (l)(1)(i)
the mercury feedrate level of the feedstreams that ensures the MTEC through (iv) of this section, you may
performance test; and calculated in paragraph (l)(1)(iii)(B)(4) request as part of the performance test
(3) You must establish an operating of this section is below the operating plan under §§ 63.7(b) and (c) and
limit on minimum pH of scrubber water requirement under paragraphs §§ 63.1207 (e) and (f) to use the mercury
as the average of the test run averages §§ 63.1220(a)(2)(iii) and (b)(2)(iii); and feedrates and associated emission rates
and comply with the limit on an hourly (2) Initiate an automatic waste feed during the comprehensive performance
rolling average. cutoff that immediately and test to extrapolate to higher allowable
(iii) For cement kilns: automatically cuts off the hazardous feedrate limits and emission rates. The
(A) When complying with the waste feed when either the gas flowrate extrapolation methodology will be
emission standards under or mercury feedrate exceeds the limits reviewed and approved, as warranted,
§§ 63.1220(a)(2)(i) and (b)(2)(i), you identified in paragraph (l)(1)(iv)(D)(1) of by the Administrator. The review will
must: this section. consider in particular whether:
(1) Comply with the mercury (iv) For lightweight aggregate kilns: (A) Performance test metal feedrates
hazardous waste feed concentration (A) When complying with the are appropriate (i.e., whether feedrates
operating requirement on a twelve-hour emission standards under §§ 63.1205, are at least at normal levels; depending
rolling average; 63.1221(a)(2)(i) and (b)(2)(i), you must on the heterogeneity of the waste,
(2) Monitor and record in the establish a 12-hour rolling average limit whether some level of spiking would be
operating record the as-fired mercury for the total feedrate of mercury in all appropriate; and whether the physical
concentration in the hazardous waste feedstreams as the average of the test form and species of spiked material is
(or the weighted-average mercury run averages; appropriate); and
concentration for multiple hazardous (B) Except as provided by paragraph (B) Whether the extrapolated feedrates
waste feedstreams); (l)(1)(iv)(C) of this section, when you request are warranted considering
(3) Initiate an automatic waste feed complying with the hazardous waste historical metal feedrate data.
cutoff that immediately and feedrate corresponding to a maximum * * * * *
automatically cuts off the hazardous theoretical emission concentration (m) * * *
waste feed when the as-fired mercury (MTEC) under §§ 63.1221(a)(2)(ii) and (1) * * *
concentration operating requirement is (b)(2)(ii), you must: (iv) Other particulate matter control
exceeded; (1) Comply with the MTEC operating devices. For each particulate matter
(B) When complying with the requirement on a twelve-hour rolling control device that is not a fabric filter
emission standards under §§ 63.1204, average; or high energy wet scrubber, or is not an
63.1220(a)(2)(ii) and (b)(2)(ii), you must (2) Monitor and record the feedrate of
electrostatic precipitator or ionizing wet
establish a 12-hour rolling average limit mercury for each hazardous waste
scrubber for which you elect to monitor
for the total feedrate of mercury in all feedstream according to § 63.1209(c);
particulate matter loadings under
feedstreams as the average of the test (3) Monitor with a CMS and record in
§ 63.1206(c)(9) of this chapter for
run averages; the operating record the gas flowrate
process control, you must ensure that
(C) Except as provided by paragraph (either directly or by monitoring a
the control device is properly operated
(l)(1)(iii)(D) of this section, when surrogate parameter that you have
and maintained as required by
complying with the hazardous waste correlated to gas flowrate);
(4) Continuously calculate and record § 63.1206(c)(7) and by monitoring the
feedrate corresponding to a maximum
in the operating record a MTEC operation of the control device as
theoretical emission concentration
assuming mercury from all hazardous follows:
(MTEC) under §§ 63.1220(a)(2)(iii) and
(b)(2)(iii), you must: waste feedstreams is emitted; * * * * *
(1) Comply with the MTEC operating (5) Initiate an automatic waste feed (3) Maximum ash feedrate. Owners
requirement on a twelve-hour rolling cutoff that immediately and and operators of hazardous waste
average; automatically cuts off the hazardous incinerators, solid fuel boilers, and
(2) Monitor and record the feedrate of waste feed when the MTEC operating liquid fuel boilers must establish a
mercury for each hazardous waste requirement is exceeded; maximum ash feedrate limit as a 12-
feedstream according to § 63.1209(c); (C) In lieu of complying with hour rolling average based on the
(3) Monitor with a CMS and record in paragraph (l)(1)(iv)(B) of this section, average of the test run averages. This
the operating record the gas flowrate you may: requirement is waived, however, if you
(either directly or by monitoring a (1) Identify in the Notification of comply with the particulate matter
surrogate parameter that you have Compliance a minimum gas flowrate detection system requirements under
correlated to gas flowrate); limit and a maximum feedrate limit of § 63.1206(c)(9).
(4) Continuously calculate and record mercury from all hazardous waste (n) * * *
in the operating record a MTEC feedstreams that ensures the MTEC (2) Maximum feedrate of semivolatile
assuming mercury from all hazardous calculated in paragraph (l)(1)(iv)(B)(4) of and low volatile metals. (i) General. You
waste feedstreams is emitted; this section is below the operating must establish feedrate limits for
(5) Initiate an automatic waste feed requirement under paragraphs semivolatile metals (cadmium and lead)
cutoff that immediately and §§ 63.1221(a)(2)(ii) and (b)(2)(ii); and and low volatile metals (arsenic,
automatically cuts off the hazardous (2) Initiate an automatic waste feed beryllium, and chromium) as follows,
waste feed when the MTEC operating cutoff that immediately and except as provided by paragraph
requirement is exceeded; automatically cuts off the hazardous (n)(2)(vii) of this section.
(D) In lieu of complying with waste feed when either the gas flowrate (ii) For incinerators, cement kilns, and
paragraph (l)(1)(iii)(C) of this section, or mercury feedrate exceeds the limits lightweight aggregate kilns, when
you may: identified in paragraph (l)(1)(iv)(C)(1) of complying with the emission standards
(1) Identify in the Notification of this section. under §§ 63.1203, 63.1204, 63.1205, and
Compliance a minimum gas flowrate (v) Extrapolation of feedrate levels. In 63.1219, and for solid fuel boilers when
limit and a maximum feedrate limit of lieu of establishing mercury feedrate complying with the emission standards
under § 63.1216, you must establish 12- test, it is not a violation because the least once a minute to calculate a 60-
hour rolling average limits for the total averaging period for the semivolatile minute average feedrate.
feedrate of semivolatile and low volatile metal emission standard is one year and (iv) You must update the rolling
metals in all feedstreams as the average compliance is based on compliance average feedrate each hour with this 60-
of the test run averages. with the semivolatile metal feedrate minute feedrate measurement.
(iii) Cement kilns under § 63.1220— limit that has an averaging period not to (v) If you select an averaging period
(A) When complying with the emission exceed an annual rolling average. for the feedrate limit that is greater than
standards under §§ 63.1220(a)(3)(i), (2) Boilers that feed hazardous waste a 12-hour rolling average, you must
(a)(4)(i), (b)(3)(i), and (b)(4)(i), you must with a heating value of 10,000 Btu/lb or calculate the initial rolling average as
establish 12-hour rolling average greater. You must calculate the though you had selected a 12-hour
feedrate limits for semivolatile and low semivolatile metal feedrate limit as the rolling average, as provided by
volatile metals as the thermal semivolatile metal emission standard paragraph (b)(5)(i) of this section. You
concentration of semivolatile metals or divided by [1 ¥ System Removal must calculate rolling averages
low volatile metals in all hazardous Efficiency]. thereafter as the average of the available
waste feedstreams. You must calculate (i) The feedrate limit is a hazardous one-minute values until enough one-
hazardous waste thermal concentrations waste thermal concentration limit minute values are available to calculate
for semivolatile metals and low volatile expressed as pounds of semivolatile the rolling average period you select. At
metals for each run as the total mass metals in all hazardous waste that time and thereafter, you update the
feedrate of semivolatile metals or low feedstreams per million Btu of rolling average feedrate each hour with
volatile metals for all hazardous waste hazardous waste fed to the boiler. a 60-minute average feedrate.
feedstreams divided by the total heat (B) Chromium. (1) Boilers that feed
(ii) You must comply with the
input rate for all hazardous waste hazardous waste with a heating value of
hazardous waste semivolatile metal
feedstreams. The 12-hour rolling 10,000 Btu/lb or greater. (i) The feedrate
thermal concentration limit by
average feedrate limits for semivolatile limit is a hazardous waste thermal
determining the feedrate of semivolatile
metals and low volatile metals are the concentration limit expressed as pounds
metal in all hazardous waste of chromium in all hazardous waste
average of the hazardous waste thermal feedstreams (lb/hr) and the hazardous
concentrations for the runs. feedstreams per million Btu of
waste thermal feedrate (MM Btu/hr) at hazardous waste fed to the boiler.
(B) When complying with the
least once a minute to calculate a 60- (ii) You must comply with the
emission standards under
minute average thermal emission hazardous waste chromium thermal
§§ 63.1220(a)(3)(ii), (a)(4)(ii), (b)(3)(ii),
concentration as [hazardous waste concentration limit by determining the
and (b)(4)(ii), you must establish 12-
semivolatile metal feedrate (lb/hr) / feedrate of chromium in all hazardous
hour rolling average limits for the total
hazardous waste thermal feedrate (MM waste feedstreams (lb/hr) and the
feedrate of semivolatile and low volatile
metals in all feedstreams as the average Btu/hr)]. hazardous waste thermal feedrate (MM
of the test run averages. (iii) You must calculate a rolling Btu/hr) at least once a minute to
(iv) Lightweight aggregate kilns under average hazardous waste semivolatile calculate a 60-minute average thermal
§ 63.1221—(A) When complying with metal thermal concentration that is emission concentration as [hazardous
the emission standards under updated each hour. waste chromium feedrate (lb/hr) /
§§ 63.1221(a)(3)(i), (a)(4)(i), (b)(3)(i), and (iv) If you select an averaging period hazardous waste thermal feedrate (MM
(b)(4)(i), you must establish 12-hour for the feedrate limit that is greater than Btu/hr)]. You must update the rolling
rolling average feedrate limits for a 12-hour rolling average, you must average feedrate each hour with this 60-
semivolatile and low volatile metals as calculate the initial rolling average as minute average feedrate measurement.
the thermal concentration of though you had selected a 12-hour (2) Boilers that feed hazardous waste
semivolatile metals or low volatile rolling average, as provided by with a heating value less than 10,000
metals in all hazardous waste paragraph (b)(5)(i) of this section. You Btu/lb. You must establish a 12-hour
feedstreams as specified in paragraphs must calculate rolling averages rolling average limit for the total
(n)(2)(iii)(A) of this section. thereafter as the average of the available feedrate (lb/hr) of chromium in all
(B) When complying with the one-minute values until enough one- feedstreams as the average of the test
emission standards under minute values are available to calculate run averages. You must update the
§§ 63.1221(a)(3)(ii), (a)(4)(ii), (b)(3)(ii), the rolling average period you select. At rolling average feedrate each hour with
and (b)(4)(ii), you must establish 12- that time and thereafter, you update the a 60-minute average feedrate
hour rolling average limits for the total rolling average feedrate each hour with measurement.
feedrate of semivolatile and low volatile a 60-minute average feedrate. (vi) LVM limits for pumpable wastes.
metals in all feedstreams as the average (3) Boilers that feed hazardous waste You must establish separate feedrate
of the test run averages. with a heating value less than 10,000 limits for low volatile metals in
(v) Liquid fuel boilers under Btu/lb. (i) You must calculate the pumpable feedstreams using the
§ 63.1217. (A) Semivolatile metals. You semivolatile metal feedrate limit as the procedures prescribed above for total
must establish a rolling average limit for semivolatile metal emission standard low volatile metals. Dual feedrate limits
the semivolatile metal feedrate as divided by [1 ¥ System Removal for both pumpable and total feedstreams
follows on an averaging period not to Efficiency]. are not required, however, if you base
exceed an annual rolling average. (ii) The feedrate limit is expressed as the total feedrate limit solely on the
(1) System removal efficiency. You a mass concentration per unit volume of feedrate of pumpable feedstreams.
must calculate a semivolatile metal stack gas (µg/dscm) and is converted to (vii) Extrapolation of feedrate levels.
system removal efficiency for each test a mass feedrate (lb/hr) by multiplying it In lieu of establishing feedrate limits as
run and calculate the average system by the average stack gas flowrate (dscm/ specified in paragraphs (l)(1)(i) through
removal efficiency of the test run hr) of the test run averages. (iii) of this section, you may request as
averages. If emissions exceed the (iii) You must comply with the part of the performance test plan under
semivolatile metal emission standard feedrate limit by determining the §§ 63.7(b) and (c) and §§ 63.1207(e) and
during the comprehensive performance semivolatile metal feedrate (lb/hr) at (f) to use the semivolatile metal and low
volatile metal feedrates and associated inorganic) in all feedstreams as the Btu/lb. You must establish a 12-hour
emission rates during the average of the test run averages. rolling average limit for the total
comprehensive performance test to (ii) Liquid fuel boilers. (A) Boilers that feedrate of chlorine (organic and
extrapolate to higher allowable feedrate feed hazardous waste with a heating inorganic) in all feedstreams as the
limits and emission rates. The value not less than 10,000 Btu/lb. (1) average of the test run averages. You
extrapolation methodology will be The feedrate limit is a hazardous waste must update the rolling average feedrate
reviewed and approved, as warranted, thermal concentration limit expressed each hour with a 60-minute average
by the Administrator. The review will as pounds of chlorine (organic and feedrate measurement.
consider in particular whether: inorganic) in all hazardous waste * * * * *
(A) Performance test metal feedrates feedstreams per million Btu of (r) Averaging periods. The averaging
are appropriate (i.e., whether feedrates hazardous waste fed to the boiler. periods specified in this section for
are at least at normal levels; depending (2) You must establish a 12-hour operating parameters are not-to-exceed
on the heterogeneity of the waste, rolling average feedrate limit as the averaging periods. You may elect to use
whether some level of spiking would be average of the test run averages. shorter averaging periods. For example,
appropriate; and whether the physical (3) You must comply with the you may elect to use a 1-hour rolling
form and species of spiked material is feedrate limit by determining the mass average rather than the 12-hour rolling
appropriate); and feedrate of hazardous waste feedstreams average specified in paragraph (l)(1)(i) of
(B) Whether the extrapolated feedrates (lb/hr) at least once a minute and by this section for mercury.
you request are warranted considering knowing the chlorine (organic and ■ 12. Section 63.1210 is amended by:
historical metal feedrate data. inorganic) content and heating value ■ a. Revising the table in paragraph
(million Btu/lb) of hazardous waste (a)(1) and the table in paragraph (a)(2).
* * * * * feedstreams at all times to calculate a ■ b. Redesignating paragraph (b) as (d).
(o) Hydrogen chloride and chlorine 60-minute average feedrate ■ c. Adding new paragraph (b).
gas. * * * measurement as [hazardous waste ■ d. Adding new paragraph (c).
(1) Feedrate of total chlorine and chlorine feedrate (lb/hr) / hazardous The revisions and additions read as
chloride. (i) Incinerators, cement kilns, waste thermal feedrate (million Btu/hr)]. follows:
lightweight aggregate kilns, solid fuel You must update the rolling average
boilers, and hydrochloric acid feedrate each hour with this 60-minute § 63.1210 What are the notification
production furnaces. You must establish average feedrate measurement. requirements?
a 12-hour rolling average limit for the (B) Boilers that feed hazardous waste (a) * * *
total feedrate of chlorine (organic and with a heating value less than 10,000 (1) * * *
Reference Notification
63.9(b) ................................................................. Initial notifications that you are subject to Subpart EEE of this Part.
63.9(d) ................................................................. Notification that you are subject to special compliance requirements.
63.9(j) .................................................................. Notification and documentation of any change in information already provided under § 63.9.
63.1206(b)(5)(i) ................................................... Notification of changes in design, operation, or maintenance.
63.1206(c)(7)(ii)(C) ............................................. Notification of excessive bag leak detection system exceedances.
63.1207(e), 63.9(e) 63.9(g)(1) and (3) ............... Notification of performance test and continuous monitoring system evaluation, including the
performance test plan and CMS performance evaluation plan.1
63.1210(b) ........................................................... Notification of intent to comply.
63.1210(d), 63.1207(j), 63.1207(k), 63.1207(l), Notification of compliance, including results of performance tests and continuous monitoring
63.9(h), 63.10(d)(2), 63.10(e)(2). system performance evaluations.
1 You may also be required on a case-by-case basis to submit a feedstream analysis plan under § 63.1209(c)(3).
(2) * * *
Reference Notification, request, petition, or application 6
63.9(i) .................................................................. You may request an adjustment to time periods or postmark deadlines for submittal and re-
view of required information.
63.10(e)(3)(ii) ...................................................... You may request to reduce the frequency of excess emissions and CMS performance reports.
63.10(f) ................................................................ You may request to waive recordkeeping or reporting requirements.
63.1204(d)(2)(iii), 63.1220(d)(2)(iii) ..................... Notification that you elect to comply with the emission averaging requirements for cement kilns
with in-line raw mills.
63.1204(e)(2)(iii), 63.1220(e)(2)(iii) ..................... Notification that you elect to comply with the emission averaging requirements for preheater or
preheater/precalciner kilns with dual stacks.
63.1206(b)(4), 63.1213, 63.6(i), 63.9(c) ............. You may request an extension of the compliance date for up to one year.
63.1206(b)(5)(i)(C) .............................................. You may request to burn hazardous waste for more than 720 hours and for purposes other
than testing or pretesting after making a change in the design or operation that could affect
compliance with emission standards and prior to submitting a revised Notification of Compli-
ance.
63.1206(b)(8)(iii)(B) ............................................. If you elect to conduct particulate matter CEMS correlation testing and wish to have federal
particulate matter and opacity standards and associated operating limits waived during the
testing, you must notify the Administrator by submitting the correlation test plan for review
and approval.
63.1206(b)(8)(v) .................................................. You may request approval to have the particulate matter and opacity standards and associ-
ated operating limits and conditions waived for more than 96 hours for a correlation test.
Reference Notification, request, petition, or application 6
63.1206(b)(9) ...................................................... Owners and operators of lightweight aggregate kilns may request approval of alternative emis-
sion standards for mercury, semivolatile metal, low volatile metal, and hydrogen chloride/
chlorine gas under certain conditions.
63.1206(b)(10) .................................................... Owners and operators of cement kilns may request approval of alternative emission standards
for mercury, semivolatile metal, low volatile metal, and hydrogen chloride/chlorine gas under
certain conditions.
63.1206(b)(14) .................................................... Owners and operators of incinerators may elect to comply with an alternative to the particulate
matter standard.
63.1206(b)(15) .................................................... Owners and operators of cement and lightweight aggregate kilns may request to comply with
the alternative to the interim standards for mercury.
63.1206(c)(2)(ii)(C) ............................................. You may request to make changes to the startup, shutdown, and malfunction plan.
63.1206(c)(5)(i)(C) .............................................. You may request an alternative means of control to provide control of combustion system
leaks.
63.1206(c)(5)(i)(D) .............................................. You may request other techniques to prevent fugitive emissions without use of instantaneous
pressure limits.
63.1207(c)(2) ...................................................... You may request to base initial compliance on data in lieu of a comprehensive performance
test.
63.1207(d)(3) ...................................................... You may request more than 60 days to complete a performance test if additional time is need-
ed for reasons beyond your control.
63.1207(e)(3), 63.7(h) ......................................... You may request a time extension if the Administrator fails to approve or deny your test plan.
63.1207(h)(2) ...................................................... You may request to waive current operating parameter limits during pretesting for more than
720 hours.
63.1207(f)(1)(ii)(D) .............................................. You may request a reduced hazardous waste feedstream analysis for organic hazardous air
pollutants if the reduced analysis continues to be representative of organic hazardous air
pollutants in your hazardous waste feedstreams.
63.1207(g)(2)(v) .................................................. You may request to operate under a wider operating range for a parameter during confirm-
atory performance testing.
63.1207(i) ............................................................ You may request up to a one-year time extension for conducting a performance test (other
than the initial comprehensive performance test) to consolidate testing with other state or
federally-required testing.
63.1207(j)(4) ....................................................... You may request more than 90 days to submit a Notification of Compliance after completing a
performance test if additional time is needed for reasons beyond your control.
63.1207(l)(3) ....................................................... After failure of a performance test, you may request to burn hazardous waste for more than
720 hours and for purposes other than testing or pretesting.
63.1209(a)(5), 63.8(f) .......................................... You may request: (1) Approval of alternative monitoring methods for compliance with stand-
ards that are monitored with a CEMS; and (2) approval to use a CEMS in lieu of operating
parameter limits.
63.1209(g)(1) ...................................................... You may request approval of: (1) Alternatives to operating parameter monitoring requirements,
except for standards that you must monitor with a continuous emission monitoring system
(CEMS) and except for requests to use a CEMS in lieu of operating parameter limits; or (2)
a waiver of an operating parameter limit.
63.1209(l)(1) ....................................................... You may request to extrapolate mercury feedrate limits.
63.1209(n)(2) ...................................................... You may request to extrapolate semivolatile and low volatile metal feedrate limits.
63.1211(d) ........................................................... You may request to use data compression techniques to record data on a less frequent basis
than required by § 63.1209.
(b) Notification of intent to comply (E) Waste minimization and emission resources for installing emission control
(NIC). These procedures apply to control technique(s) effectiveness; systems or making process changes for
sources that have not previously (F) A description of the evaluation emission control, or the date by which
complied with the requirements of criteria used or to be used to select you will issue orders for the purchase of
paragraph (b) of this section, and to waste minimization and/or emission component parts to accomplish
sources that previously complied with control technique(s); and emission control or process changes.
the NIC requirements of § 63.1210, (G) A general description of how you (C) The date by which you anticipate
which were in effect prior to October 11, intend to comply with the emission you will submit construction
2000, that must make a technology standards of this subpart. applications;
change requiring a Class 1 permit (ii) As applicable to each source, (D) The date by which you anticipate
modification to meet the standards of information on key activities and you will initiate on-site construction,
§§ 63.1219, 63.1220, and 63.1221. estimated dates for these activities that installation of emission control
(1) You must prepare a Notification of will bring the source into compliance equipment, or process change;
Intent to Comply that includes all of the with emission control requirements of (E) The date by which you anticipate
following information: this subpart. You must include all of the
(i) General information: you will complete on-site construction,
(A) The name and address of the following key activities and dates in installation of emission control
owner/operator and the source; your NIC: equipment, or process change; and
(B) Whether the source is a major or (A) The dates by which you anticipate (F) The date by which you anticipate
an area source; you will develop engineering designs you will achieve final compliance. The
(C) Waste minimization and emission for emission control systems or process individual dates and milestones listed
control technique(s) being considered; changes for emissions; in paragraphs (b)(1)(ii)(A) through (F) of
(D) Emission monitoring technique(s) (B) The date by which you anticipate this section as part of the NIC are not
you are considering; you will commit internal or external requirements and therefore are not
enforceable deadlines; the requirements standards of this subpart. You must post once on at least one local radio station
of paragraphs (b)(1)(ii)(A) through (F) of a sign-in sheet or otherwise provide a or television station.
this section must be included as part of voluntary opportunity for attendees to (iv) Notice to the facility mailing list.
the NIC only to inform the public of provide their names and addresses; You must provide a copy of the notice
how you intend to comply with the (2) You must submit a summary of the to the facility mailing list in accordance
emission standards of this subpart. meeting, along with the list of attendees with § 124.10(c)(1)(ix) of this chapter.
(iii) A summary of the public meeting and their addresses developed under (4) You must include all of the
required under paragraph (c) of this paragraph (b)(1) of this section, and following in the notices required under
section; copies of any written comments or paragraph (c)(3) of this section:
(iv) If you intend to cease burning materials submitted at the meeting, to (i) The date, time, and location of the
hazardous waste prior to or on the the Administrator as part of the final meeting;
compliance date, the requirements of NIC, in accordance with paragraph
paragraphs (b)(1)(ii) and (b)(1)(iii) of this (ii) A brief description of the purpose
(b)(1)(iii) of this section; of the meeting;
section do not apply. You must include (3) You must provide public notice of
in your NIC a schedule of key dates for (iii) A brief description of the source
the NIC meeting at least 30 days prior and proposed operations, including the
the steps to be taken to stop hazardous to the meeting and you must maintain,
waste activity at your combustion unit. address or a map (e.g., a sketched or
and provide to the Administrator upon copied street map) of the source
Key dates include the date for submittal request, documentation of the notice.
of RCRA closure documents required location;
You must provide public notice in all of
under subpart G, part 264 or subpart G, (iv) A statement encouraging people
the following forms:
part 265 of this chapter. to contact the source at least 72 hours
(2) You must make a draft of the NIC (i) Newspaper advertisement. You before the meeting if they need special
available for public review no later than must publish a notice in a newspaper of access to participate in the meeting;
30 days prior to the public meeting general circulation in the county or
(v) A statement describing how the
required under paragraph (c)(1) of this equivalent jurisdiction of your facility.
draft NIC (and final NIC, if requested)
section or no later than 9 months after In addition, you must publish the notice
can be obtained; and
the effective date of the rule if you in newspapers of general circulation in
adjacent counties or equivalent (vi) The name, address, and telephone
intend to cease burning hazardous waste number of a contact person for the NIC.
prior to or on the compliance date. jurisdiction where such publication
would be necessary to inform the (5) The requirements of this paragraph
(3) You must submit the final NIC to
affected public. You must publish the do not apply to sources that intend to
the Administrator no later than one year
notice as a display advertisement. cease burning hazardous waste prior to
following the effective date of the
(ii) Visible and accessible sign. You or on the compliance date.
emission standards of this subpart.
(c) NIC public meeting and notice. (1) must post a notice on a clearly marked ■ 13. Section 63.1211 is amended by:
Prior to the submission of the NIC to the sign at or near the source. If you place ■ a. Revising the table in paragraph (b).
permitting agency, and no later than 10 the sign on the site of the hazardous ■ b. Revising paragraph (c)(1).
months after the effective date of the waste combustor, the sign must be large The revisions read as follows:
emission standards of this subpart, you enough to be readable from the nearest
must hold at least one informal meeting spot where the public would pass by the § 63.1211 What are the recordkeeping and
with the public to discuss anticipated site. reporting requirements?
activities described in the draft NIC for (iii) Broadcast media announcement. * * * * *
achieving compliance with the emission You must broadcast a notice at least (b) * * *
Reference Document, Data, or Information
63.1200, 63.10(b) and (c) ................................... General. Information required to document and maintain compliance with the regulations of
Subpart EEE, including data recorded by continuous monitoring systems (CMS), and copies
of all notifications, reports, plans, and other documents submitted to the Administrator.
63.1204(d)(1)(ii), 63.1220(d)(1)(ii) ...................... Documentation of mode of operation changes for cement kilns with in-line raw mills.
63.1204(d)(2)(ii), 63.1220(d)(2)(ii) ...................... Documentation of compliance with the emission averaging requirements for cement kilns with
in-line raw mills.
63.1204(e)(2)(ii), 63.1220(e)(2)(ii) ...................... Documentation of compliance with the emission averaging requirements for preheater or pre-
heater/precalciner kilns with dual stacks.
63.1206(b)(1)(ii) .................................................. If you elect to comply with all applicable requirements and standards promulgated under au-
thority of the Clean Air Act, including Sections 112 and 129, in lieu of the requirements of
Subpart EEE when not burning hazardous waste, you must document in the operating
record that you are in compliance with those requirements.
63.1206(b)(5)(ii) .................................................. Documentation that a change will not adversely affect compliance with the emission standards
or operating requirements.
63.1206(b)(11) .................................................... Calculation of hazardous waste residence time.
63.1206(c)(2) ...................................................... Startup, shutdown, and malfunction plan.
63.1206(c)(2)(v)(A) ............................................. Documentation of your investigation and evaluation of excessive exceedances during malfunc-
tions.
63.1206(c)(3)(v) .................................................. Corrective measures for any automatic waste feed cutoff that results in an exceedance of an
emission standard or operating parameter limit.
63.1206(c)(3)(vii) ................................................. Documentation and results of the automatic waste feed cutoff operability testing.
63.1206(c)(4)(ii) .................................................. Emergency safety vent operating plan.
63.1206(c)(4)(iii) .................................................. Corrective measures for any emergency safety vent opening.
63.1206(c)(5)(ii) .................................................. Method used for control of combustion system leaks.
63.1206(c)(6) ...................................................... Operator training and certification program.
63.1206(c)(7)(i)(D) .............................................. Operation and maintenance plan.
63.1209(c)(2) ...................................................... Feedstream analysis plan.
Reference Document, Data, or Information
63.1209(k)(6)(iii), 63.1209(k)(7)(ii), Documentation that a substitute activated carbon, dioxin/furan formation reaction inhibitor, or
63.1209(k)(9)(ii), 63.1209(o)(4)(iii). dry scrubber sorbent will provide the same level of control as the original material.
63.1209(k)(7)(i)(C) .............................................. Results of carbon bed performance monitoring.
63.1209(q) ........................................................... Documentation of changes in modes of operation.
63.1211(c) ........................................................... Documentation of compliance.
(c) * * * (4) Hold an informal public meeting § 63.1214 Implementation and

(1) By the compliance date, you must 30 days following notice of NIC public enforcement.
develop and include in the operating meeting and notice of the pre- * * * * *
record a Documentation of Compliance. application meeting or notice of the (c) * * *
You are not subject to this requirement, permit modification request. (1) Approval of alternatives to
however, if you submit a Notification of (c) Information Repository specific to requirements in §§ 63.1200, 63.1203,
Compliance under § 63.1207(j) prior to new combustion units. (1) Any source 63.1204, 63.1205, 63.1206(a), 63.1215,
the compliance date. Upon inclusion of 63.1216, 63.1217, 63.1218, 63.1219,
that files a RCRA permit application or
the Documentation of Compliance in the 63.1220, and 63.1221.
modification request for construction of
operating record, hazardous waste (2) Approval of major alternatives to
a new hazardous waste combustion unit
burning incinerators, cement kilns, and test methods under §§ 63.7(e)(2)(ii) and
after October 12, 2005 may be required
lightweight aggregate kilns regulated (f), 63.1208(b), and 63.1209(a)(1), as
to establish an information repository if
under the interim standards of defined under § 63.90, and as required
deemed appropriate.
§§ 63.1203, 63.1204, and 63.1205 are no in this subpart.
longer subject to compliance with the (2) The Administrator may assess the
need, on a case-by-case basis for an (3) Approval of major alternatives to
previously applicable Notification of monitoring under §§ 63.8(f) and
Compliance. information repository. When assessing
the need for a repository, the 63.1209(a)(5), as defined under § 63.90,
* * * * * and as required in this subpart.
Administrator shall consider the level of
■ 14. Section 63.1212 is added to (4) Approval of major alternatives to
public interest, the presence of an
subpart EEE to read as follows: recordkeeping and reporting under
existing repository, and any information
§ 63.1212 What are the other requirements available via the New Source Review §§ 63.10(f) and 63.1211(a) through (c), as
pertaining to the NIC? and Title V permit processes. If the defined under § 63.90, and as required
(a) Certification of intent to comply. Administrator determines a need for a in this subpart.
The Notice of Intent to Comply (NIC) repository, then the Administrator shall ■ 16. Section § 63.1215 is added to
must contain the following certification notify the facility that it must establish subpart EEE to read as follows:
signed and dated by a responsible and maintain an information repository.
§ 63.1215 What are the health-based
official as defined under § 63.2 of this (3) The information repository shall compliance alternatives for total chlorine?
chapter: I certify under penalty of law contain all documents, reports, data,
that I have personally examined and am (a) General. (1) Overview. You may
and information deemed necessary by establish and comply with health-based
familiar with the information submitted the Administrator. The Administrator
in this document and all attachments compliance alternatives for total
shall have the discretion to limit the chlorine under the procedures
and that, based on my inquiry of those contents of the repository.
individuals immediately responsible for prescribed in this section for your
(4) The information repository shall hazardous waste combustors other than
obtaining the information, I believe that
be located and maintained at a site hydrochloric acid production furnaces.
the information is true, accurate, and
chosen by the source. If the You may comply with these health-
complete. I am aware that there are
Administrator finds the site unsuitable based compliance alternatives in lieu of
significant penalties for submitting false
for the purposes and persons for which the emission standards for total chlorine
information, including the possibility of
it was established, due to problems with provided under §§ 63.1216, 63.1217,
fine and imprisonment.
(b) New units. Any source that files a location, hours of availability, access, or 63.1219, 63.1220, and 63.1221. To
RCRA permit application or permit other relevant considerations, then the identify and comply with the limits, you
modification request for construction of Administrator shall specify a more must:
a hazardous waste combustion unit after appropriate site. (i) Identify a total chlorine emission
October 12, 2005 must: (5) The Administrator shall require concentration (ppmv) expressed as
(1) Prepare a draft NIC according to the source to provide a written notice chloride (Cl(-)) equivalent for each on-
§ 63.1210(b) and make it available to the about the information repository to all site hazardous waste combustor. You
public upon issuance of the notice of individuals on the source mailing list. may select total chlorine emission
NIC public meeting per § 63.1210(c)(3); concentrations as you choose to
(6) The source shall be responsible for
(2) Prepare a draft comprehensive demonstrate eligibility for the risk-based
maintaining and updating the repository
performance test plan pursuant to the limits under this section, except as
with appropriate information
requirements of § 63.1207 and make it provided by paragraph (b)(4) of this
throughout a period specified by the
available for public review upon section;
Administrator. The Administrator may
issuance of the notice of NIC public (ii) Apportion the total chlorine
close the repository at his or her
meeting; emission concentration between HCl
discretion based on the considerations
(3) Provide notice to the public of a and Cl2 according to paragraph (b)(6)(i)
in paragraph (c)(2) of this section.
pre-application meeting pursuant to of this section, and calculate HCl and
§ 124.30 or notice to the public of a ■ 15. Section 63.1214 is amended by Cl2 emission rates (lb/hr) using the gas
permit modification request pursuant to revising paragraphs (c)(1), (c)(2), (c)(3), flowrate and other parameters from the
§ 270.42 and; and (c)(4) to read as follows: most recent regulatory compliance test.
(iii) Calculate the annual average HCl- equating the toxicity of chlorine to HCl ERCl2 is the emission rate of chlorine in
equivalent emission rate as prescribed using RfCs as the health risk metric for lbs/hr
in paragraph (b)(2) of this section. long-term exposure. RfCHCl is the reference concentration of
(iv) Perform an eligibility Annual Average HCl-Equivalent HCl
demonstration to determine if your HCl- Emission Rate Limit means the HCl- RfCCl2 is the reference concentration of
equivalent emission rate meets the equivalent emission rate (lb/hr) chlorine
national exposure standard and thus is determined by equating the toxicity of (3) 1-hour average rates. You must
below the annual average HCl- chlorine to HCl using RfCs as the health calculate 1-hour average toxicity-
equivalent emission rate limit, as risk metric for long-term exposure and weighted HCl-equivalent emission rates
prescribed by paragraph (c) of this which ensures that maximum annual for each combustor as follows:
section; average ambient concentrations of HCl ERSTtw = ERHCl + ERCl2 × (aRELHCl/
(v) Submit your eligibility equivalents do not exceed a Hazard aRELCl2)
demonstration for review and approval, Index of 1.0, rounded to the nearest Where:
as prescribed by paragraph (e) of this tenths decimal place (0.1), at an off-site
section, which must include receptor location. ERSTtw is the 1-hour average HCl
information to ensure that the 1-hour Hazard Index (HI) means the sum of toxicity-weighted emission rate
average HCl-equivalent emission rate more than one Hazard Quotient for (HCl-equivalent emission rate)
limit is not exceeded, as prescribed by multiple substances and/or multiple considering 1-hour (short-term)
paragraph (d) of this section; exposure pathways. In this section, the exposures, lb/hr
(vi) Demonstrate compliance with the Hazard Index is the sum of the Hazard ERHCl is the emission rate of HCl in lbs/
annual average HCl-equivalent emission Quotients for HCl and chlorine. hr
rate limit during the comprehensive Hazard Quotient (HQ) means the ratio ERCl2 is the emission rate of chlorine in
performance test, as prescribed by the of the predicted media concentration of lbs/hr
testing and monitoring requirements a pollutant to the media concentration aRELHCl is the 1-hour Reference
under paragraph (e) of this section; at which no adverse effects are Exposure Level of HCl
(vii) Comply with compliance expected. For chronic inhalation aRELCl2 is the 1-hour Reference
monitoring requirements, including exposures, the HQ is calculated under Exposure Level of chlorine
establishing feedrate limits on total this section as the air concentration (4) You must use the RfC values for
chlorine and chloride, and operating divided by the RfC. For acute inhalation hydrogen chloride and chlorine found at
parameter limits on emission control exposures, the HQ is calculated under http://epa.gov/ttn/atw/toxsource/
equipment, as prescribed by paragraph this section as the air concentration summary.html.
(f) of this section; and divided by the aREL. (5) You must use the aREL values for
(viii) Comply with the requirements Look-up table analysis means a risk hydrogen chloride and chlorine found at
for changes, as prescribed by paragraph screening analysis based on comparing http://www.oehha.ca.gov/air/
(h) of this section. the HCl-equivalent emission rate from acute_rels/acuterel.html.
(2) Definitions. In addition to the the affected source to the appropriate (6) Cl2HCl ratios—(i) Ratio for
definitions under § 63.1201, the HCl-equivalent emission rate limit calculating annual average HCl-
following definitions apply to this specified in Tables 1 through 4 of this equivalent emission rates. (A) To
section: section. calculate the annual average HCl-
1-Hour Average HCl-Equivalent Reference Concentration (RfC) means equivalent emission rate (lb/hr) for each
Emission Rate means the HCl-equivalent an estimate (with uncertainty spanning combustor, you must apportion the total
emission rate (lb/hr) determined by perhaps an order of magnitude) of a chlorine emission concentration (ppmv
equating the toxicity of chlorine to HCl continuous inhalation exposure to the chloride (Cl(-)) equivalent) between HCl
using 1-hour RELs as the health risk human population (including sensitive and chlorine according to the historical
metric for acute exposure. subgroups) that is likely to be without average Cl2/HCl volumetric ratio for all
1-Hour Average HCl-Equivalent an appreciable risk of deleterious effects regulatory compliance tests.
Emission Rate Limit means the HCl- during a lifetime. It can be derived from (B) You must calculate HCl and Cl2
equivalent emission rate (lb/hr) various types of human or animal data, emission rates (lb/hr) using the
determined by equating the toxicity of with uncertainty factors generally apportioned emission concentrations
chlorine to HCl using 1-hour RELs as applied to reflect limitations of the data and the gas flowrate and other
the health risk metric for acute exposure used. parameters from the most recent
and which ensures that maximum 1- (b) HCl-equivalent emission rates. (1) regulatory compliance test.
hour average ambient concentrations of You must express total chlorine (C) You must calculate the annual
HCl-equivalents do not exceed a Hazard emission rates for each hazardous waste average HCl-equivalent emission rate
Index of 1.0, rounded to the nearest combustor as HCl-equivalent emission using these HCl and Cl2 emission rates
tenths decimal place (0.1), at an off-site rates. and the equation in paragraph (b)(2) of
receptor location. (2) Annual average rates. You must this section.
Acute Reference Exposure Level calculate annual average toxicity- (ii) Ratio for calculating 1-hour
(aREL) means health thresholds below weighted HCl-equivalent emission rates average HCl-equivalent emission rates.
which there would be no adverse health for each combustor as follows: (A) To calculate the 1-hour average HCl-
effects for greater than once in a lifetime ERtw = ERHCl + ERCl2 × (RfCHCl/RfCCl2) equivalent emission rate for each
exposures of one hour. ARELs are Where: combustor as a criterion for you to
developed by the California Office of ERLTtw is the annual average HCl determine under paragraph (d) of this
Health Hazard Assessment and are toxicity-weighted emission rate section if an hourly rolling average
available at http://www.oehha.ca.gov/ (HCl-equivalent emission rate) feedrate limit on total chlorine and
air/acute_rels/acuterel.html. considering long-term exposures, chloride may be waived, you must
Annual Average HCl-Equivalent lb/hr apportion the total chlorine emission
Emission Rate means the HCl-equivalent ERHCl is the emission rate of HCl in concentration (ppmv chloride (Cl(-))
emission rate (lb/hr) determined by lbs/hr equivalent) between HCl and chlorine
according to the historical highest Cl2/ ratio, the permitting authority may prescribed by paragraph (c)(4) of this
HCl volumetric ratio for all regulatory require you to conduct periodic testing section.
compliance tests. to establish representative ratios. (ii) You must also determine in your
(B) You must calculate HCl and Cl2 (v) Updating Cl2/HCl ratios. You must eligibility demonstration whether each
emission rates (lb/hr) using the include the Cl2/HCl volumetric ratio combustor may exceed the 1-hour HCl-
apportioned emission concentrations demonstrated during each performance equivalent emission rate limit absent an
and the gas flowrate and other test in your data base of historical Cl2/ hourly rolling average limit on the
parameters from the most recent HCl ratios to update the ratios you feedrate of total chlorine and chloride,
regulatory compliance test. establish under paragraphs (b)(6)(i) and as provided by paragraph (d) of this
(C) You must calculate the 1-hour (ii) of this section for subsequent section.
average HCl-equivalent emission rate calculations of the annual average and (2) Definition of eligibility. (i)
using the se HCl and Cl2 emission rates 1-hour average HCl-equivalent emission Eligibility for the risk-based total
and the equation in paragraph (b)(3) of rates. chlorine standard is determined by
this section. (7) Emission rates are capped. The comparing the annual average HCl-
(iii) Ratios for new sources. (A) You hydrogen chloride and chlorine equivalent emission rate for the total
must use engineering information to emission rates you use to calculate the chlorine emission rate you select for
estimate the Cl2/HCl volumetric ratio for HCl-equivalent emission rate limit for each combustor to the annual average
a new source for the initial eligibility incinerators, cement kilns, and HCl-equivalent emission rate limit.
demonstration. lightweight aggregate kilns must not (ii) The annual average HCl-
(B) You must use the Cl2/HCl result in total chlorine emission equivalent emission rate limit ensures
volumetric ratio demonstrated during concentrations exceeding: that the Hazard Index for chronic
the initial comprehensive performance (i) For incinerators that were existing exposure from HCl and chlorine
test to demonstrate in the Notification of sources on April 19, 1996: 77 parts per emissions from all on-site hazardous
Compliance that your HCl-equivalent million by volume, combined waste combustors is less than or equal
emission rate does not exceed your HCl- emissions, expressed as chloride (Cl(-)) to 1.0, rounded to the nearest tenths
equivalent emission rate limit. equivalent, dry basis and corrected to 7 decimal place (0.1), for the actual
(C) When approving the test plan for individual most exposed to the facility’s
percent oxygen;
the initial comprehensive performance emissions, considering off-site locations
(ii) For incinerators that are new or
test, the permitting authority will where people reside and where people
reconstructed sources after April 19,
establish a periodic testing requirement, congregate for work, school, or
1996: 21 parts per million by volume,
such as every 3 months for 1 year, to recreation.
combined emissions, expressed as
establish a record of representative Cl2/ (iii) Your facility is eligible for the
chloride (Cl(-)) equivalent, dry basis and
HCl volumetric ratios. health-based compliance alternative for
(1) You must revise your HCl- corrected to 7 percent oxygen;
(iii) For cement kilns that were total chlorine if either:
equivalent emission rates and HCl- (A) The annual average HCl-
equivalent emission rate limits after existing sources on April 19, 1996: 130
equivalent emission rate for each on-site
each such test using the procedures parts per million by volume, combined
hazardous waste combustor is below the
prescribed in paragraphs (b)(6)(i) and emissions, expressed as chloride (Cl(-))
appropriate value in the look-up table
(ii) of this section. equivalent, dry basis and corrected to 7
determined under paragraph (c)(3) of
(2) If you no longer are eligible for the percent oxygen;
this section; or
health-based compliance alternative, (iv) For cement kilns that are new or (B) The annual average HCl-
you must notify the permitting authority reconstructed sources after April 19, equivalent emission rate for each on-site
immediately and either: 1996: 86 parts per million by volume, hazardous waste combustor is below the
(i) Submit a revised eligibility combined emissions, expressed as annual average HCl-equivalent emission
demonstration requesting lower HCl- chloride (Cl(-)) equivalent, dry basis and rate limit you calculate based on a site-
equivalent emission rate limits, corrected to 7 percent oxygen; specific compliance demonstration
establishing lower HCl-equivalent (v) For lightweight aggregate kilns that under paragraph (c)(4) of this section.
emission rates, and establishing by were existing sources on April 19, 1996: (3) Look-up table analysis. Look-up
downward extrapolation lower feedrate 600 parts per million by volume, tables for the eligibility demonstration
limits for total chlorine and chloride; or combined emissions, expressed as are provided as Tables 1 and 2 to this
(ii) Request a compliance schedule of chloride (Cl(-)) equivalent, dry basis and section.
up to three years to demonstrate corrected to 7 percent oxygen; (i) Table 1 presents annual average
compliance with the emission standards (vi) For lightweight aggregate kilns HCl-equivalent emission rate limits for
under §§ 63.1216, 63.1217, 63.1219, that are new or reconstructed sources sources located in flat terrain. For
63.1220, and 63.1221. after April 19, 1996: 600 parts per purposes of this analysis, flat terrain is
(iv) Unrepresentative or inadequate million by volume, combined terrain that rises to a level not exceeding
historical Cl2/HCl volumetric ratios. (A) emissions, expressed as chloride (Cl(-)) one half the stack height within a
If you believe that the Cl2/HCl equivalent, dry basis and corrected to 7 distance of 50 stack heights.
volumetric ratio for one or more percent oxygen. (ii) Table 2 presents annual average
historical regulatory compliance tests is (c) Eligibility demonstration—(1) HCl-equivalent emission rate limits for
not representative of the current ratio, General. (i) You must perform an sources located in simple elevated
you may request that the permitting eligibility demonstration to determine terrain. For purposes of this analysis,
authority allow you to screen those whether the total chlorine emission simple elevated terrain is terrain that
ratios from the analysis of historical rates you select for each on-site rises to a level exceeding one half the
ratios. hazardous waste combustor meet the stack height, but that does not exceed
(B) If the permitting authority believes national exposure standards using either the stack height, within a distance of 50
that too few historical ratios are a look-up table analysis prescribed by stack heights.
available to calculate a representative paragraph (c)(3) of this section, or a site- (iii) To determine the annual average
average ratio or establish a maximum specific compliance demonstration HCl-equivalent emission rate limit for a
source from the look-up table, you must nearest property boundary do not match one hazardous waste combustor on site,
use the stack height and stack diameter the exact values in the look-up table, the sum across all hazardous waste
for your hazardous waste combustors you must use the next lowest table combustors of the ratio of the adjusted
and the distance between the stack and value. HCl-equivalent emission rate limit to
the property boundary. (v) Adjusted HCl-equivalent emission the HCl-equivalent emission rate limit
(iv) If any of these values for stack rate limit for multiple on-site provided by Tables 1 or 2 cannot exceed
height, stack diameter, and distance to combustors. (A) If you have more than 1.0, according to the following equation:
n
HC1-Equivalent Emission Rate Limit Adjusted i
∑
i =1 HCI-Equivalent Emission Rate Limit Tablei
≤ 1.0
Where: (C) Use site-specific, quality-assured emission rate limit for each affected
i = number of on-site hazardous waste data wherever possible; source using either a look-up table
combustors; (D) Use health-protective default analysis or site-specific analysis:
HCl-Equivalent Emission Rate Limit assumptions wherever site-specific data (i) Look-up table analysis. Look-up
Adjustedi means the apportioned, are not available, and: tables are provided for 1-hour average
allowable HCl-equivalent emission (E) Contain adequate documentation HCl-equivalent emission rate limits as
rate limit for combustor i, and of the data and methods used for the Table 3 and Table 4 to this section.
HCl-Equivalent Emission Rate Limit assessment so that it is transparent and Table 3 provides limits for facilities
Tablei means the HCl-equivalent can be reproduced by an experienced located in flat terrain. Table 4 provides
emission rate limit from Table 1 or risk assessor and emissions limits for facilities located in simple
2 to § 63.1215 for combustor i. measurement expert. elevated terrain. You must use the
(B) The adjusted HCl-equivalent (iv) Your site-specific compliance Tables to establish 1-hour average HCl-
emission rate limit becomes the HCl- demonstration need not: equivalent emission rate limits as
equivalent emission rate limit. (A) Assume any attenuation of prescribed in paragraphs (c)(3)(iii)
(4) Site-specific compliance exposure concentrations due to the through (c)(3)(v) of this section for
demonstration. (i) You may use any penetration of outdoor pollutants into annual average HCl-equivalent emission
scientifically-accepted peer-reviewed indoor exposure areas; rate limits.
risk assessment methodology for your (B) Assume any reaction or deposition (ii) Site-specific analysis. The 1-hour
site-specific compliance demonstration of the emitted pollutants during average HCl-equivalent emission rate
to calculate an annual average HCl- transport from the emission point to the limit is the HCl-equivalent emission rate
equivalent emission rate limit for each point of exposure. that ensures that the Hazard Index
on-site hazardous waste combustor. An (d) Assurance that the 1-hour HCl- associated with maximum 1-hour
example of one approach for performing equivalent emission rate limit will not average exposures is not greater than 1.0
the demonstration for air toxics can be be exceeded. To ensure that the 1-hour rounded to the nearest tenths decimal
found in the EPA’s ‘‘Air Toxics Risk HCl-equivalent emission rate limit will place (0.1). You must follow the risk
Assessment Reference Library, Volume not be exceeded when complying with assessment procedures under paragraph
2, Site-Specific Risk Assessment the annual average HCl-equivalent (c)(4) of this section to estimate short-
Technical Resource Document,’’ which emission rate limit, you must establish term inhalation exposures through the
may be obtained through the EPA’s Air a 1-hour average HCl-equivalent estimation of maximum 1-hour average
Toxics Web site at http://www.epa.gov/ emission rate for each combustor, ambient concentrations.
ttn/fera/risk_atra_main.html. establish a 1-hour average HCl- (3) Criteria for determining whether
(ii) The annual average HCl- equivalent emission rate limit for each the 1-hour HCl-equivalent emission rate
equivalent emission rate limit is the combustor, and consider site-specific may be exceeded absent an hourly
HCl-equivalent emission rate that factors including prescribed criteria to rolling average limit on the feedrate of
ensures that the Hazard Index determine if the 1-hour average HCl- total chlorine and chloride. An hourly
associated with maximum annual equivalent emission rate limit may be rolling average feedrate limit on total
average exposures is not greater than 1.0 exceeded absent an hourly rolling chlorine and chloride is waived if you
rounded to the nearest tenths decimal average limit on the feedrate of total determine considering the criteria listed
place (0.1). chlorine and chloride. If the 1-hour below that the long-term feedrate limit
(iii) To determine the annual average average HCl-equivalent emission rate (and averaging period) established
HCl-equivalent emission rate limit, your limit may be exceeded, you must under paragraph (c)(4)(i) of this section
site-specific compliance demonstration establish an hourly rolling average will also ensure that the 1-hour average
must, at a minimum: feedrate limit on total chlorine as HCl-equivalent emission rate will not
(A) Estimate long-term inhalation provided by paragraph (f)(3) of this exceed the 1-hour average HCl-
exposures through the estimation of section. equivalent emission rate limit you
annual or multi-year average ambient (1) 1-hour average HCl-equivalent calculate for each combustor.
concentrations; emission rate. You must calculate the 1- (i) The ratio of the 1-hour average
(B) Estimate the inhalation exposure hour average HCl-equivalent emission HCl-equivalent emission rate based on
for the actual individual most exposed rate from the total chlorine emission the total chlorine emission rate you
to the facility’s emissions from concentration you select for each source select for each hazardous waste
hazardous waste combustors, as prescribed in paragraph (b)(6)(ii)(C) combustor to the 1-hour average HCl-
considering off-site locations where of this section. equivalent emission rate limit for the
people reside and where people (2) 1-hour average HCl-equivalent combustor; and
congregate for work, school, or emission rate limit. You must establish (ii) The potential for the source to
ER12OC05.003</GPH>
recreation; the 1-hour average HCl-equivalent vary total chlorine and chloride
feedrates substantially over the (C) The annual average HCl- support your determination of the
averaging period for the feedrate limit equivalent emission rate limit for each annual average HCl-equivalent emission
established under paragraph (c)(4)(i) of hazardous waste combustor, and the rate limit for each combustor:
this section. limits on operating parameters required (A) Identification of the risk
(e) Review and approval of eligibility under paragraph (g)(1) of this section; assessment methodology used;
demonstrations—(1) Content of the (D) Determination of the long-term (B) Documentation of the fate and
eligibility demonstration—(i) General. chlorine feedrate limit, including the transport model used;
The eligibility demonstration must total chlorine system removal efficiency (C) Documentation of the fate and
include the following information, at a for sources that establish an (up to) transport model inputs, including the
minimum: annual rolling average feedrate limit stack parameters listed in paragraph
(A) Identification of each hazardous under paragraph (g)(2)(ii) of this section; (d)(1)(i)(C) of this section converted to
waste combustor combustion gas (iii) Additional content to implement the dimensions required for the model;
emission point (e.g., generally, the flue the 1-hour average HCl-equivalent (D) As applicable:
gas stack); emission rate limit. You must include (1) Meteorological data;
(B) The maximum and average the following in your eligibility (2) Building, land use, and terrain
capacity at which each combustor will demonstration to implement the 1-hour data;
operate, and the maximum rated average HCl-equivalent emission rate (3) Receptor locations and population
capacity for each combustor, using the limit: data, including areas where people
metric of stack gas volume (under both (A) Determination of whether the congregate for work, school, or
actual and standard conditions) emitted combustor may exceed the 1-hour HCl- recreation; and
per unit of time, as well as any other equivalent emission rate limit absent an (4) Other facility-specific parameters
metric that is appropriate for the hourly rolling average chlorine feedrate input into the model;
combustor (e.g., million Btu/hr heat limit, including: (E) Documentation of the fate and
input for boilers; tons of dry raw (1) Determination of the 1-hour transport model outputs; and
material feed/hour for cement kilns); average HCl-equivalent emission rate
(F) Documentation of any exposure
(C) Stack parameters for each from the total chlorine emission rate
assessment and risk characterization
combustor, including, but not limited to you select for the combustor;
calculations.
stack height, stack diameter, stack gas (2) Determination of the 1-hour
(2) Review and approval—(i) Existing
temperature, and stack gas exit velocity; average HCl-equivalent emission rate
sources. (A) If you operate an existing
limit using either look-up Tables 3 and
(D) Plot plan showing all stack source, you must submit the eligibility
4 to this section or site-specific risk
emission points, nearby residences and demonstration to your permitting
analysis;
property boundary line; authority for review and approval not
(3) Determination of the ratio of the 1-
(E) Identification of any stack gas later than 12 months prior to the
hour average HCl-equivalent emission
control devices used to reduce compliance date. You must also submit
rate to the 1-hour average HCl-
emissions from each combustor; a separate copy of the eligibility
equivalent emission rate limit for the
(F) Identification of the RfC values combustor; and demonstration to: U.S. EPA, Risk and
used to calculate annual average HCl- (4) The potential for the source to vary Exposure Assessment Group, Emission
equivalent emission rates and the aREL total chlorine and chloride feedrates Standards Division (C404–01), Attn:
values used to calculate 1-hour average substantially over the averaging period Group Leader, Research Triangle Park,
HCl-equivalent emission rates; for the long-term feedrate limit North Carolina 27711, electronic mail
(G) Calculations used to determine the established under paragraphs (g)(2)(i) address REAG@epa.gov.
annual average and 1-hour average HCl- and (g)(2)(ii) of this section; and (B) Your permitting authority should
equivalent emission rates and rate (B) Determination of the hourly notify you of approval or intent to
limits, including calculation of the Cl2/ rolling average chlorine feedrate limit, disapprove your eligibility
HCl ratios as prescribed by paragraph including the total chlorine system demonstration within 6 months after
(b)(6) of this section; removal efficiency. receipt of the original demonstration,
(ii) Additional content to implement (iv) Additional content of a look-up and within 3 months after receipt of any
the annual average HCl-equivalent table demonstration. If you use the look- supplemental information that you
emission rate limit. You must include up table analysis to establish HCl- submit. A notice of intent to disapprove
the following in your eligibility equivalent emission rate limits, your your eligibility demonstration, whether
demonstration to implement the annual eligibility demonstration must also before or after the compliance date, will
average HCl-equivalent emission rate contain, at a minimum, the following: identify incomplete or inaccurate
limit: (A) Documentation that the facility is information or noncompliance with
(A) For incinerators, cement kilns, located in either flat or simple elevated prescribed procedures and specify how
and lightweight aggregate kilns, terrain; and much time you will have to submit
calculations to confirm that the annual (B) For facilities with more than one additional information or to achieve the
average HCl-equivalent emission rate on-site hazardous waste combustor, MACT standards for total chlorine
that you calculate from the total documentation that the sum of the ratios under §§ 63.1216, 63.1217, 63.1219,
chlorine emission rate you select for for all such combustors of the HCl- 63.1220, and 63.1221. If your eligibility
each combustor does not exceed the equivalent emission rate to the HCl- demonstration is disapproved, the
limits provided by paragraph (b)(7) of equivalent emission rate limit does not permitting authority may extend the
this section; exceed 1.0. compliance date of the total chlorine
(B) Comparison of the annual average (v) Additional content of a site- standards to allow you to make changes
HCl-equivalent emission rate limit for specific compliance demonstration. If to the design or operation of the
each combustor to the annual average you use a site-specific compliance combustor or related systems as quickly
HCl-equivalent emission rate for the demonstration, your eligibility as practicable to enable you to achieve
total chlorine emission rate you select demonstration must also contain, at a compliance with the MACT total
for each combustor; minimum, the following information to chlorine standards.
(C) If your permitting authority has source that increases its emissions or its comprehensive performance test at
not approved your eligibility potential to emit such that it becomes a levels specified under paragraph
demonstration by the compliance date, major source of HAP on or after April (e)(2)(ii)(B) of this section, you must use
and has not issued a notice of intent to 12, 2007, you must either: EPA Method 320/321 or ASTM D 6735–
disapprove your demonstration, you (1) Comply with the final total 01, or an equivalent method, to measure
may nonetheless begin complying, on chlorine emission standards under hydrogen chloride, and Method 26/26A,
the compliance date, with the HCl- §§ 63.1216, 63.1217, 63.1219, 63.1220, or an equivalent method, to measure
equivalent emission rate limits you and 63.1221 upon startup. If the final chlorine and hydrogen chloride, and
present in your eligibility standard is more stringent than the determine your chlorine emissions as
demonstration. standard proposed for your source on follows:
(D) If your permitting authority issues April 20, 2004, however, and if you start (A) You must determine you chlorine
a notice of intent to disapprove your operations before October 14, 2008, you emissions to be the higher of the value
eligibility demonstration after the may comply with the proposed standard measured by Method 26/26A, or an
compliance date, the authority will until October 14, 2008, after which you equivalent method, or the value
identify the basis for that notice and must comply with the final standard; or calculated by difference between the
specify how much time you will have to (2) Submit an eligibility combined hydrogen chloride and
submit additional information or to demonstration for review and approval chlorine levels measured by Method 26/
comply with the MACT standards for under this section 12 months prior to 26a, or an equivalent method, and the
total chlorine under §§ 63.1216, startup. hydrogen chloride measurement from
63.1217, 63.1219, 63.1220, and 63.1221. (f) Testing requirements—(1) General. EPA Method 320/321 or ASTM D 6735–
The permitting authority may extend You must comply with the requirements 01, or an equivalent method.
the compliance date of the total chlorine for comprehensive performance testing (B) The procedures under paragraph
standards to allow you to make changes under § 63.1207. (f)(2)(ii) of this section for determining
to the design or operation of the (2) System removal efficiency. (i) You hydrogen chloride and chlorine
combustor or related systems as quickly must calculate the total chlorine emissions apply if you feed bromine or
as practicable to enable you to achieve removal efficiency of the combustor sulfur during the performance test at the
compliance with the MACT standards during each run of the comprehensive levels specified in this paragraph
for total chlorine. performance test. (f)(5)(ii)(B):
(ii) New or reconstructed sources. (A) (ii) You must calculate the average (1) If the bromine/chlorine ratio in
General. The procedures for review and system removal efficiency as the average feedstreams is greater than 5 percent by
approval of eligibility demonstrations of the test run averages. mass; or
applicable to existing sources under (iii) If your source does not control (2) If the sulfur/chlorine ratio in
paragraph (e)(2)(i) of this section also emissions of total chlorine, you must feedstreams is greater than 50 percent
apply to new or reconstructed sources, assume zero system removal efficiency. by mass.
except that the date you must submit (3) Annual average HCl-equivalent (g) Monitoring requirements. (1)
the eligibility demonstration is as emission rate limit. If emissions during General. You must establish and comply
prescribed in this paragraph (e)(2)(ii). the comprehensive performance test with limits on the same operating
(B) If you operate a new or exceed the annual average HCl- parameters that apply to sources
reconstructed source that starts up equivalent emission rate limit, complying with the MACT standard for
before April 12, 2007, or a solid fuel eligibility for emission limits under this total chlorine under § 63.1209(o), except
boiler or liquid fuel boiler that is an area section is not affected. This emission that feedrate limits on total chlorine and
source that increases its emissions or its rate limit is an annual average limit chloride must be established according
potential to emit such that it becomes a even though compliance is based on a to paragraphs (g)(2) and (g)(3) of this
major source of HAP before April 12, 12-hour or (up to) an annual rolling section:
2007, you must either: average feedrate limit on total chlorine (2) Feedrate limit to ensure
(1) Comply with the final total and chloride because the feedrate limit compliance with the annual average
chlorine emission standards under is also used for compliance assurance HCl-equivalent emission rate limit. (i)
§§ 63.1216, 63.1217, 63.1219, 63.1220, for the semivolatile metal emission For sources subject to the feedrate limit
and 63.1221, by October 12, 2005, or standard for total chlorine and chloride under
upon startup, whichever is later, except (4) 1-hour average HCl-equivalent § 63.1209(n)(4) to ensure compliance
for a standard that is more stringent emission rate limit. Total chlorine with the semivolatile metals standard:
than the standard proposed on April 20, emissions during each run of the (A) The feedrate limit (and averaging
2004 for your source. If a final standard comprehensive performance test cannot period) for total chlorine and chloride to
is more stringent than the proposed exceed the 1-hour average HCl- ensure compliance with the annual
standard, you may comply with the equivalent emission rate limit. average HCl-equivalent emission rate
proposed standard until October 14, (5) Test methods. (i) If you operate a limit is the same as required by
2008, after which you must comply with cement kiln or a combustor equipped § 63.1209(n)(4), except as provided by
the final standard; or with a dry acid gas scrubber, you must paragraph (g)(2)(i)(B) of this section.
(2) Submit an eligibility use EPA Method 320/321 or ASTM D (B) The numerical value of the total
demonstration for review and approval 6735–01, or an equivalent method, to chlorine and chloride feedrate limit (i.e.,
under this section by April 12, 2006, measure hydrogen chloride, and the not considering the averaging period)
and comply with the HCl-equivalent back-half (caustic impingers) of Method you establish under § 63.1209(n)(4)
emission rate limits and operating 26/26A, or an equivalent method, to must not exceed the value you calculate
requirements you establish in the measure chlorine gas. as the annual average HCl-equivalent
eligibility demonstration. (ii) Bromine and sulfur emission rate limit (lb/hr) divided by [1
(C) If you operate a new or considerations. If you operate an ¥ system removal efficiency], where the
reconstructed source that starts up on or incinerator, boiler, or lightweight system removal efficiency is calculated
after April 12, 2007, or a solid fuel aggregate kiln and your feedstreams as prescribed by paragraph (f)(2) of this
boiler or liquid fuel boiler that is an area contain bromine or sulfur during the section.
(ii) For sources exempt from the prescribed by paragraph (f)(2)(ii) of this conducting a performance test to
feedrate limit for total chlorine and section. reestablish the combustor’s system
chloride under § 63.1209(n)(4) because (h) Changes—(1) Changes over which removal efficiency. You must also
they comply with § 63.1207(m)(2), the you have control. (i) Changes that would submit to the permitting authority a
feedrate limit for total chlorine and affect the HCl-equivalent emission rate revised eligibility demonstration
chloride to ensure compliance with the limit. (A) If you plan to change the documenting the higher system removal
annual average HCl-equivalent emission design, operation, or maintenance of the efficiency and the increased feedrate
rate must be established as follows: facility in a manner than would limits on total chlorine and chloride.
(A) You must establish an average decrease the annual average or 1-hour (2) Changes over which you do not
period for the feedrate limit that does average HCl-equivalent emission rate have control that may decrease the HCl-
not exceed an annual rolling average; limit, you must submit to the permitting equivalent emission rate limits. These
(B) The numerical value of the total authority prior to the change a revised requirements apply if you use a site-
chlorine and chloride feedrate limit (i.e., eligibility demonstration documenting specific risk assessment under
not considering the averaging period) the lower emission rate limits and paragraph (c)(4) of this section to
must not exceed the value you calculate calculations of reduced total chlorine demonstrate eligibility for the health-
as the annual average HCl-equivalent and chloride feedrate limits. based limits.
emission rate limit (lb/hr) divided by [1 (B) If you plan to change the design, (i) Proactive review. You must review
¥ system removal efficiency], where the operation, or maintenance of the facility the documentation you use in your
system removal efficiency is calculated in a manner than would increase the eligibility demonstration every five
as prescribed by paragraph (f)(2) of this annual average or 1-hour average HCl- years from the date of the
section. equivalent emission rate limit, and you comprehensive performance test and
(C) You must calculate the initial elect to increase your total chlorine and submit for review and approval with the
rolling average as though you had chloride feedrate limits. You must also comprehensive performance test plan
selected a 12-hour rolling average, as submit to the permitting authority prior either a certification that the
provided by paragraph (b)(5)(i) of this to the change a revised eligibility information used in your eligibility
section. You must calculate rolling demonstration documenting the demonstration has not changed in a
averages thereafter as the average of the increased emission rate limits and manner that would decrease the annual
available one-minute values until calculations of the increased feedrate average or 1-hour average HCl-
enough one-minute values are available limits prior to the change. equivalent emission rate limit, or a
to calculate the rolling average period (ii) Changes that could affect system revised eligibility demonstration.
you select. At that time and thereafter, removal efficiency. (A) If you plan to
(ii) Reactive review. If in the interim
you update the rolling average feedrate change the design, operation, or
between your comprehensive
each hour with a 60-minute average maintenance of the combustor in a
performance tests you have reason to
feedrate. manner than could decrease the system
know of changes that would decrease
(3) Feedrate limit to ensure removal efficiency, you are subject to
the annual average or 1-hour average
compliance with the 1-hour average the requirements of § 63.1206(b)(5) for
HCl-equivalent emission rate limit, you
HCl-equivalent emission rate limit. (i) conducting a performance test to
must submit a revised eligibility
You must establish an hourly rolling reestablish the combustor’s system
demonstration as soon as practicable but
average feedrate limit on total chlorine removal efficiency and you must submit
not more frequently than annually.
and chloride to ensure compliance with a revised eligibility demonstration
the 1-hour average HCl-equivalent documenting the lower system removal (iii) Compliance schedule. If you
emission rate limit unless you efficiency and the reduced feedrate determine that you cannot demonstrate
determine that the hourly rolling limits on total chlorine and chloride. compliance with a lower annual average
average feedrate limit is waived under (B) If you plan to change the design, HCl-equivalent emission rate limit
paragraph (d) of this section. operation, or maintenance of the during the comprehensive performance
(ii) You must calculate the hourly combustor in a manner than could test because you need additional time to
rolling average feedrate limit for total increase the system removal efficiency, complete changes to the design or
chlorine and chloride as the 1-hour and you elect to document the increased operation of the source, you may request
average HCl-equivalent emission rate system removal efficiency to establish that the permitting authority grant you
limit (lb/hr) divided by [1 ¥ system higher feedrate limits on total chlorine additional time to make those changes
removal efficiency], where the system and chloride, you are subject to the as quickly as practicable.
removal efficiency is calculated as requirements of § 63.1206(b)(5) for BILLING CODE 6560–50–P
ER12OC05.006</GPH>
ER12OC05.007</GPH>
ER12OC05.008</GPH>
■ 17. Section 63.1216 and an cause combustion gases to be emitted (4) For arsenic, beryllium, and
undesignated center heading are added into the atmosphere that contain: chromium combined, except for an area
to subpart EEE to read as follows: (1) For dioxins and furans, either source as defined under § 63.2,
carbon monoxide or hydrocarbon emissions in excess of 380 µg/dscm,
Emissions Standards and Operating corrected to 7 percent oxygen;
emissions in excess of the limits
Limits for Solid Fuel Boilers, Liquid
provided by paragraph (a)(5) of this (5) For carbon monoxide and
Fuel Boilers, and Hydrochloric Acid
section; hydrocarbons, either:
Production Furnaces
(2) Mercury in excess of 11 µg/dscm (i) Carbon monoxide in excess of 100
§ 63.1216 What are the standards for solid corrected to 7 percent oxygen; parts per million by volume, over an
fuel boilers that burn hazardous waste? (3) For cadmium and lead combined, hourly rolling average (monitored
(a) Emission limits for existing except for an area source as defined continuously with a continuous
sources. You must not discharge or under § 63.2, emissions in excess of 180 emissions monitoring system), dry basis
µg/dscm, corrected to 7 percent oxygen;
ER12OC05.009</GPH>
and corrected to 7 percent oxygen. If
you elect to comply with this carbon § 63.1206(b)(7), hydrocarbons do not (ii) You must specify one or more
monoxide standard rather than the exceed 10 parts per million by volume POHCs that are representative of the
hydrocarbon standard under paragraph during those runs, over an hourly most difficult to destroy organic
(a)(5)(ii) of this section, you must also rolling average (monitored continuously compounds in your hazardous waste
document that, during the destruction with a continuous emissions monitoring feedstream. You must base this
and removal efficiency (DRE) test runs system), dry basis, corrected to 7 specification on the degree of difficulty
or their equivalent as provided by percent oxygen, and reported as of incineration of the organic
§ 63.1206(b)(7), hydrocarbons do not propane; or constituents in the hazardous waste and
exceed 10 parts per million by volume (ii) Hydrocarbons in excess of 10 parts on their concentration or mass in the
during those runs, over an hourly per million by volume, over an hourly hazardous waste feed, considering the
rolling average (monitored continuously rolling average (monitored continuously results of hazardous waste analyses or
with a continuous emissions monitoring with a continuous emissions monitoring other data and information.
system), dry basis, corrected to 7 system), dry basis, corrected to 7 (d) Significant figures. The emission
percent oxygen, and reported as percent oxygen, and reported as limits provided by paragraphs (a) and
propane; or propane; (b) of this section are presented with
(ii) Hydrocarbons in excess of 10 parts (6) For hydrogen chloride and two significant figures. Although you
per million by volume, over an hourly chlorine combined, except for an area must perform intermediate calculations
rolling average (monitored continuously source as defined under § 63.2, using at least three significant figures,
with a continuous emissions monitoring emissions in excess of 73 parts per you may round the resultant emission
system), dry basis, corrected to 7 million by volume, expressed as a levels to two significant figures to
percent oxygen, and reported as chloride (Cl(-)) equivalent, dry basis and document compliance.
propane; corrected to 7 percent oxygen; and (e) Alternative to the particulate
(6) For hydrogen chloride and (7) For particulate matter, except for matter standard. (1) General. In lieu of
chlorine combined, except for an area an area source as defined under § 63.2 complying with the particulate matter
source as defined under § 63.2, or as provided by paragraph (e) of this standards of this section, you may elect
emissions in excess of 440 parts per section, emissions in excess of 34 mg/ to comply with the following alternative
million by volume, expressed as a dscm corrected to 7 percent oxygen. metal emission control requirement:
chloride (Cl(-)) equivalent, dry basis and (c) Destruction and removal efficiency (2) Alternative metal emission control
corrected to 7 percent oxygen; and (DRE) standard. (1) 99.99% DRE. Except requirements for existing solid fuel
(7) For particulate matter, except for as provided in paragraph (c)(2) of this boilers. (i) You must not discharge or
an area source as defined under § 63.2 section, you must achieve a DRE of cause combustion gases to be emitted
or as provided by paragraph (e) of this 99.99% for each principle organic into the atmosphere that contain
section, emissions in excess of 68 mg/ hazardous constituent (POHC) cadmium, lead, and selenium in excess
dscm corrected to 7 percent oxygen. designated under paragraph (c)(3) of this of 180 µg/dscm, combined emissions,
(b) Emission limits for new sources. section. You must calculate DRE for corrected to 7 percent oxygen; and,
You must not discharge or cause each POHC from the following equation: (ii) You must not discharge or cause
combustion gases to be emitted into the DRE = [1 ¥ (Wout ÷ Win)] × 100% combustion gases to be emitted into the
atmosphere that contain: atmosphere that contain antimony,
(1) For dioxins and furans, either Where:
Win = mass feedrate of one POHC in arsenic, beryllium, chromium, cobalt,
carbon monoxide or hydrocarbon
a waste feedstream; and manganese, and nickel in excess of 380
emissions in excess of the limits
Wout = mass emission rate of the same µg/dscm, combined emissions,
provided by paragraph (b)(5) of this
POHC present in exhaust emissions corrected to 7 percent oxygen.
section;
(2) Mercury in excess of 11 µg/dscm prior to release to the atmosphere. (3) Alternative metal emission control
corrected to 7 percent oxygen; (2) 99.9999% DRE. If you burn the requirements for new solid fuel boilers.
(3) For cadmium and lead combined, dioxin-listed hazardous wastes F020, (i) You must not discharge or cause
except for an area source as defined F021, F022, F023, F026, or F027 (see combustion gases to be emitted into the
under § 63.2, emissions in excess of 180 § 261.31 of this chapter), you must atmosphere that contain cadmium, lead,
µg/dscm, corrected to 7 percent oxygen; achieve a DRE of 99.9999% for each and selenium in excess of 180 µg/dscm,
(4) For arsenic, beryllium, and POHC that you designate under combined emissions, corrected to 7
chromium combined, except for an area paragraph (c)(3) of this section. You percent oxygen; and,
source as defined under § 63.2, must demonstrate this DRE performance (ii) You must not discharge or cause
emissions in excess of 190 µg/dscm, on POHCs that are more difficult to combustion gases to be emitted into the
corrected to 7 percent oxygen; incinerate than tetra-, penta-, and atmosphere that contain antimony,
(5) For carbon monoxide and hexachlorodibenzo-p-dioxins and arsenic, beryllium, chromium, cobalt,
hydrocarbons, either: dibenzofurans. You must use the manganese, and nickel in excess of 190
(i) Carbon monoxide in excess of 100 equation in paragraph (c)(1) of this µg/dscm, combined emissions,
parts per million by volume, over an section to calculate DRE for each POHC. corrected to 7 percent oxygen.
hourly rolling average (monitored In addition, you must notify the (4) Operating limits. Semivolatile and
continuously with a continuous Administrator of your intent to low volatile metal operating parameter
emissions monitoring system), dry basis incinerate hazardous wastes F020, F021, limits must be established to ensure
and corrected to 7 percent oxygen. If F022, F023, F026, or F027. compliance with the alternative
you elect to comply with this carbon (3) Principal organic hazardous emission limitations described in
monoxide standard rather than the constituents (POHCs). (i) You must treat paragraphs (e)(2) and (e)(3) of this
hydrocarbon standard under paragraph the POHCs in the waste feed that you section pursuant to § 63.1209(n), except
(b)(5)(ii) of this section, you must also specify under paragraph (c)(3)(ii) of this that semivolatile metal feedrate limits
document that, during the destruction section to the extent required by apply to lead, cadmium, and selenium,
and removal efficiency (DRE) test runs paragraphs (c)(1) and (c)(2) of this combined, and low volatile metal
or their equivalent as provided by section. feedrate limits apply to arsenic,
beryllium, chromium, antimony, cobalt, (3) For cadmium and lead combined, equivalent, dry basis and corrected to 7
manganese, and nickel, combined. except for an area source as defined percent oxygen;
(f) Elective standards for area sources. under § 63.2, (ii) When you burn hazardous waste
Area sources as defined under § 63.2 are (i) When you burn hazardous waste with an as-fired heating value of 10,000
subject to the standards for cadmium with an as-fired heating value less than Btu/lb or greater, emissions in excess of
and lead, the standards for arsenic, 10,000 Btu/lb, emissions in excess of 5.08 × 10¥2 lbs combined emissions of
beryllium, and chromium, the standards 150 µg/dscm, corrected to 7 percent hydrogen chloride and chlorine gas
for hydrogen chloride and chlorine, and oxygen, on an (not-to-exceed) annual attributable to the hazardous waste per
the standards for particulate matter averaging period; million Btu heat input from the
under this section if they elect under (ii) When you burn hazardous waste hazardous waste;
§ 266.100(b)(3) of this chapter to comply with an as-fired heating value of 10,000 (7) For particulate matter, except for
with those standards in lieu of the Btu/lb or greater, emissions in excess of an area source as defined under § 63.2
standards under 40 CFR 266.105, 8.2 × 10¥5 lbs combined cadmium and or as provided by paragraph (e) of this
266.106, and 266.107 to control those lead emissions attributable to the section, emissions in excess of 80 mg/
pollutants. hazardous waste per million Btu heat dscm corrected to 7 percent oxygen.
input from the hazardous waste on an (b) Emission limits for new sources.
■ 18. Section 63.1217 is added to read
(not-to-exceed) annual averaging period; You must not discharge or cause
as follows:
(4) For chromium, except for an area combustion gases to be emitted into the
§ 63.1217 What are the standards for liquid source as defined under § 63.2: atmosphere that contain:
fuel boilers that burn hazardous waste? (i) When you burn hazardous waste (1)(i) Dioxins and furans in excess of
(a) Emission limits for existing with an as-fired heating value less than 0.40 ng TEQ/dscm, corrected to 7
sources. You must not discharge or 10,000 Btu/lb, emissions in excess of percent oxygen, for liquid fuel boilers
cause combustion gases to be emitted 370 µg/dscm, corrected to 7 percent equipped with a dry air pollution
into the atmosphere that contain: oxygen; control system; or
(1)(i) Dioxins and furans in excess of (ii) When you burn hazardous waste (ii) Either carbon monoxide or
0.40 ng TEQ/dscm, corrected to 7 with an as-fired heating value of 10,000 hydrocarbon emissions in excess of the
percent oxygen, for liquid fuel boilers Btu/lb or greater, emissions in excess of limits provided by paragraph (b)(5) of
equipped with a dry air pollution 1.3 × 10¥4 lbs chromium emissions this section for sources not equipped
control system; or attributable to the hazardous waste per with a dry air pollution control system;
million Btu heat input from the (iii) A source equipped with a wet air
(ii) Either carbon monoxide or
hazardous waste; pollution control system followed by a
hydrocarbon emissions in excess of the
(5) For carbon monoxide and dry air pollution control system is not
limits provided by paragraph (a)(5) of
hydrocarbons, either: considered to be a dry air pollution
this section for sources not equipped (i) Carbon monoxide in excess of 100 control system, and a source equipped
with a dry air pollution control system; parts per million by volume, over an with a dry air pollution control system
(iii) A source equipped with a wet air hourly rolling average (monitored followed by a wet air pollution control
pollution control system followed by a continuously with a continuous system is considered to be a dry air
dry air pollution control system is not emissions monitoring system), dry basis pollution control system for purposes of
considered to be a dry air pollution and corrected to 7 percent oxygen. If this emission limit;
control system, and a source equipped you elect to comply with this carbon (2) For mercury:
with a dry air pollution control system monoxide standard rather than the (i) When you burn hazardous waste
followed by a wet air pollution control hydrocarbon standard under paragraph with an as-fired heating value less than
system is considered to be a dry air (a)(5)(ii) of this section, you must also 10,000 Btu/lb, emissions in excess of 6.8
pollution control system for purposes of document that, during the destruction µg/dscm, corrected to 7 percent oxygen,
this emission limit; and removal efficiency (DRE) test runs on an (not-to-exceed) annual averaging
(2) For mercury, except as provided or their equivalent as provided by period;
for in paragraph (a)(2)(iii) of this § 63.1206(b)(7), hydrocarbons do not (ii) When you burn hazardous waste
section: exceed 10 parts per million by volume with an as-fired heating value of 10,000
(i) When you burn hazardous waste during those runs, over an hourly Btu/lb or greater, emissions in excess of
with an as-fired heating value less than rolling average (monitored continuously 1.2 × 10¥6 lbs mercury emissions
10,000 Btu/lb, emissions in excess of 19 with a continuous emissions monitoring attributable to the hazardous waste per
µg/dscm, corrected to 7 percent oxygen, system), dry basis, corrected to 7 million Btu heat input from the
on an (not-to-exceed) annual averaging percent oxygen, and reported as hazardous waste on an (not-to-exceed)
period; propane; or annual averaging period;
(ii) When you burn hazardous waste (ii) Hydrocarbons in excess of 10 parts (3) For cadmium and lead combined,
with an as-fired heating value 10,000 per million by volume, over an hourly except for an area source as defined
Btu/lb or greater, emissions in excess of rolling average (monitored continuously under § 63.2:
4.2 × 10¥5 lbs mercury attributable to with a continuous emissions monitoring (i) When you burn hazardous waste
the hazardous waste per million Btu system), dry basis, corrected to 7 with an as-fired heating value less than
heat input from the hazardous waste on percent oxygen, and reported as 10,000 Btu/lb, emissions in excess of 78
an (not-to-exceed) annual averaging propane; µg/dscm, corrected to 7 percent oxygen,
period; (6) For hydrogen chloride and on an (not-to-exceed) annual averaging
(iii) The boiler operated by Diversified chlorine, except for an area source as period;
Scientific Services, Inc. with EPA defined under § 63.2: (ii) When you burn hazardous waste
identification number TND982109142, (i) When you burn hazardous waste with an as-fired heating value greater
and which burns radioactive waste with an as-fired heating value less than than or equal to 10,000 Btu/lb,
mixed with hazardous waste, must 10,000 Btu/lb, emissions in excess of 31 emissions in excess of 6.2 × 10¥6 lbs
comply with the mercury emission parts per million by volume, combined combined cadmium and lead emissions
standard under § 63.1219(a)(2); emissions, expressed as a chloride (Cl(-)) attributable to the hazardous waste per
million Btu heat input from the (c) Destruction and removal efficiency (2) Alternative metal emission control
hazardous waste on an (not-to-exceed) (DRE) standard. (1) 99.99% DRE. Except requirements for existing liquid fuel
annual averaging period; as provided in paragraph (c)(2) of this boilers. (i) When you burn hazardous
(4) For chromium, except for an area section, you must achieve a DRE of waste with a heating value less than
source as defined under § 63.2: 99.99% for each principle organic 10,000 Btu/lb:
(i) When you burn hazardous waste hazardous constituent (POHC) (A) You must not discharge or cause
with an as-fired heating value less than designated under paragraph (c)(3) of this combustion gases to be emitted into the
10,000 Btu/lb, emissions in excess of 12 section. You must calculate DRE for atmosphere that contain cadmium, lead,
µg/dscm, corrected to 7 percent oxygen; each POHC from the following equation: and selenium, combined, in excess of
(ii) When you burn hazardous waste DRE = [1 ¥ (Wout ÷ Win)] × 100% 150 µg/dscm, corrected to 7 percent
with an as-fired heating value of 10,000 oxygen; and
Where:
Btu/lb or greater, emissions in excess of (B) You must not discharge or cause
1.4 × 10¥5 lbs chromium emissions Win = mass feedrate of one POHC in a combustion gases to be emitted into the
attributable to the hazardous waste per waste feedstream; and atmosphere that contain antimony,
million Btu heat input from the Wout = mass emission rate of the same
arsenic, beryllium, chromium, cobalt,
hazardous waste; POHC present in exhaust emissions
manganese, and nickel, combined, in
prior to release to the atmosphere.
(5) For carbon monoxide and excess of 370 µg/dscm, corrected to
hydrocarbons, either: (2) 99.9999% DRE. If you burn the 7 percent oxygen;
(i) Carbon monoxide in excess of 100 dioxin-listed hazardous wastes F020, (ii) When you burn hazardous waste
parts per million by volume, over an F021, F022, F023, F026, or F027 (see with a heating value of 10,000 Btu/lb or
hourly rolling average (monitored § 261.31 of this chapter), you must greater:
continuously with a continuous achieve a DRE of 99.9999% for each (A) You must not discharge or cause
emissions monitoring system), dry basis POHC that you designate under combustion gases to be emitted into the
and corrected to 7 percent oxygen. If paragraph (c)(3) of this section. You atmosphere that contain in excess of
you elect to comply with this carbon must demonstrate this DRE performance 8.2 × 10¥5 lbs combined emissions of
monoxide standard rather than the on POHCs that are more difficult to cadmium, lead, and selenium
hydrocarbon standard under paragraph incinerate than tetra-, penta-, and attributable to the hazardous waste per
(b)(5)(ii) of this section, you must also hexachlorodibenzo-p-dioxins and million Btu heat input from the
document that, during the destruction dibenzofurans. You must use the hazardous waste; and
and removal efficiency (DRE) test runs equation in paragraph (c)(1) of this (B) You must not discharge or cause
or their equivalent as provided by section to calculate DRE for each POHC. combustion gases to be emitted into the
§ 63.1206(b)(7), hydrocarbons do not In addition, you must notify the atmosphere that contain either in excess
exceed 10 parts per million by volume Administrator of your intent to of 1.3 × 10¥4 lbs combined emissions of
during those runs, over an hourly incinerate hazardous wastes F020, F021, antimony, arsenic, beryllium,
rolling average (monitored continuously F022, F023, F026, or F027. chromium, cobalt, manganese, and
with a continuous emissions monitoring (3) Principal organic hazardous nickel attributable to the hazardous
system), dry basis, corrected to 7 constituents (POHCs). (i) You must treat waste per million Btu heat input from
percent oxygen, and reported as the POHCs in the waste feed that you the hazardous waste;
propane; or specify under paragraph (c)(3)(ii) of this (3) Alternative metal emission control
(ii) Hydrocarbons in excess of 10 parts section to the extent required by requirements for new liquid fuel boilers.
per million by volume, over an hourly paragraphs (c)(1) and (c)(2) of this (i) When you burn hazardous waste
rolling average (monitored continuously section. with a heating value less than 10,000
with a continuous emissions monitoring (ii) You must specify one or more Btu/lb:
system), dry basis, corrected to 7 POHCs that are representative of the (A) You must not discharge or cause
percent oxygen, and reported as most difficult to destroy organic combustion gases to be emitted into the
propane; compounds in your hazardous waste atmosphere that contain cadmium, lead,
(6) For hydrogen chloride and feedstream. You must base this and selenium, combined, in excess of 78
chlorine, except for an area source as specification on the degree of difficulty µg/dscm, corrected to 7 percent oxygen;
defined under § 63.2: of incineration of the organic and
(i) When you burn hazardous waste constituents in the hazardous waste and (B) You must not discharge or cause
with an as-fired heating value less than on their concentration or mass in the combustion gases to be emitted into the
10,000 Btu/lb, emissions in excess of 31 hazardous waste feed, considering the atmosphere that contain antimony,
parts per million by volume, combined results of hazardous waste analyses or arsenic, beryllium, chromium, cobalt,
emissions, expressed as a chloride (Cl(-)) other data and information. manganese, and nickel, combined, in
equivalent, dry basis and corrected to 7 (d) Significant figures. The emission excess of 12 µg/dscm, corrected to
percent oxygen; limits provided by paragraphs (a) and 7 percent oxygen;
(ii) When you burn hazardous waste (b) of this section are presented with (ii) When you burn hazardous waste
with an as-fired heating value of 10,000 two significant figures. Although you with a heating value greater than or
Btu/lb or greater, emissions in excess of must perform intermediate calculations equal to 10,000 Btu/lb:
5.08 × 10¥2 lbs combined emissions of using at least three significant figures, (A) You must not discharge or cause
hydrogen chloride and chlorine gas you may round the resultant emission combustion gases to be emitted into the
attributable to the hazardous waste per levels to two significant figures to atmosphere that contain in excess of
million Btu heat input from the document compliance. 6.2 × 10¥6 lbs combined emissions of
hazardous waste; (e) Alternative to the particulate cadmium, lead, and selenium
(7) For particulate matter, except for matter standard. (1) General. In lieu of attributable to the hazardous waste per
an area source as defined under § 63.2 complying with the particulate matter million Btu heat input from the
or as provided by paragraph (e) of this standards of this section, you may elect hazardous waste; and
section, emissions in excess of 20 mg/ to comply with the following alternative (B) You must not discharge or cause
dscm corrected to 7 percent oxygen. metal emission control requirement: combustion gases to be emitted into the
atmosphere that contain either in excess emissions monitoring system), dry basis (3) For lead and cadmium, except for
of 1.4 × 10¥5 lbs combined emissions of and corrected to an area source as defined under § 63.2,
antimony, arsenic, beryllium, 7 percent oxygen. If you elect to comply hydrogen chloride and chlorine gas
chromium, cobalt, manganese, and with this carbon monoxide standard emissions in excess of the levels
nickel attributable to the hazardous rather than the hydrocarbon standard provided by paragraph (b)(6) of this
waste per million Btu heat input from under paragraph (a)(5)(ii) of this section, section;
the hazardous waste; you must also document that, during the (4) For arsenic, beryllium, and
(4) Operating limits. Semivolatile and destruction and removal efficiency chromium, except for an area source as
low volatile metal operating parameter (DRE) test runs or their equivalent as defined under § 63.2, hydrogen chloride
limits must be established to ensure provided by § 63.1206(b)(7), and chlorine gas emissions in excess of
compliance with the alternative hydrocarbons do not exceed 10 parts per the levels provided by paragraph (b)(6)
emission limitations described in million by volume during those runs, of this section;
paragraphs (e)(2) and (e)(3) of this over an hourly rolling average (5) For carbon monoxide and
section pursuant to § 63.1209(n), except (monitored continuously with a hydrocarbons, either:
that semivolatile metal feedrate limits continuous emissions monitoring (i) Carbon monoxide in excess of 100
apply to lead, cadmium, and selenium, system), dry basis, corrected to 7 parts per million by volume, over an
combined, and low volatile metal percent oxygen, and reported as hourly rolling average (monitored
feedrate limits apply to arsenic, propane; or continuously with a continuous
beryllium, chromium, antimony, cobalt, (ii) Hydrocarbons in excess of 10 parts emissions monitoring system), dry basis
manganese, and nickel, combined. per million by volume, over an hourly and corrected to 7 percent oxygen. If
(f) Elective standards for area sources. rolling average (monitored continuously you elect to comply with this carbon
Area sources as defined under § 63.2 are with a continuous emissions monitoring monoxide standard rather than the
subject to the standards for cadmium system), dry basis, corrected to 7 hydrocarbon standard under paragraph
and lead, the standards for chromium, percent oxygen, and reported as (b)(5)(ii) of this section, you must also
the standards for hydrogen chloride and propane; document that, during the destruction
chlorine, and the standards for (6) For hydrogen chloride and and removal efficiency (DRE) test runs
particulate matter under this section if chlorine gas, either: or their equivalent as provided by
they elect under § 266.100(b)(3) of this (i) Emission in excess of 150 parts per § 63.1206(b)(7), hydrocarbons do not
chapter to comply with those standards million by volume, combined exceed 10 parts per million by volume
in lieu of the standards under 40 CFR emissions, expressed as a chloride (Cl(¥) during those runs, over an hourly
266.105, 266.106, and 266.107 to control equivalent, dry basis and corrected to 7 rolling average (monitored continuously
those pollutants. percent oxygen; or with a continuous emissions monitoring
■ 19. Section 63.1218 is added to read (ii) Emissions greater than the levels system), dry basis, corrected to 7
as follows: that would be emitted if the source is percent oxygen, and reported as
achieving a system removal efficiency propane; or
§ 63.1218 What are the standards for (ii) Hydrocarbons in excess of 10 parts
hydrochloric acid production furnaces that
(SRE) of less than 99.923 percent for
total chlorine and chloride fed to the per million by volume, over an hourly
burn hazardous waste? rolling average (monitored continuously
(a) Emission limits for existing combustor. You must calculate SRE
from the following equation: with a continuous emissions monitoring
sources. You must not discharge or system), dry basis, corrected to 7
cause combustion gases to be emitted SRE = [1 ¥ (Cl out / Cl in)] × 100% percent oxygen, and reported as
into the atmosphere that contain: Where: propane;
(1) For dioxins and furans, either Cl in = mass feedrate of total chlorine (6) For hydrogen chloride and
carbon monoxide or hydrocarbon or chloride in all feedstreams, chlorine gas, either:
emissions in excess of the limits reported as chloride; and (i) Emission in excess of 25 parts per
provided by paragraph (a)(5) of this Cl out = mass emission rate of hydrogen million by volume, combined
section; chloride and chlorine gas, reported emissions, expressed as a chloride (Cl(¥)
(2) For mercury, hydrogen chloride equivalent, dry basis and corrected to 7
as chloride, in exhaust emissions
and chlorine gas emissions in excess of percent oxygen; or
prior to release to the atmosphere.
the levels provided by paragraph (a)(6) (ii) Emissions greater than the levels
of this section; (7) For particulate matter, except for
that would be emitted if the source is
(3) For lead and cadmium, except for an area source as defined under § 63.2,
achieving a system removal efficiency
an area source as defined under § 63.2, hydrogen chloride and chlorine gas
(SRE) of less than 99.987 percent for
hydrogen chloride and chlorine gas emissions in excess of the levels
total chlorine and chloride fed to the
emissions in excess of the levels provided by paragraph (a)(6) of this
combustor. You must calculate SRE
provided by paragraph (a)(6) of this section.
from the following equation:
section; (b) Emission limits for new sources.
You must not discharge or cause SRE = [1 ¥ (Cl out / Cl in)] × 100%
(4) For arsenic, beryllium, and
chromium, except for an area source as combustion gases to be emitted into the Where:
defined under § 63.2, hydrogen chloride atmosphere that contain: Cl in = mass feedrate of total chlorine
and chlorine gas emissions in excess of (1) For dioxins and furans, either or chloride in all feedstreams,
the levels provided by paragraph (a)(6) carbon monoxide or hydrocarbon reported as chloride; and
of this section; emissions in excess of the limits Cl out = mass emission rate of hydrogen
(5) For carbon monoxide and provided by paragraph (b)(5) of this chloride and chlorine gas, reported
hydrocarbons, either: section; as chloride, in exhaust emissions
(i) Carbon monoxide in excess of 100 (2) For mercury, hydrogen chloride prior to release to the atmosphere.
parts per million by volume, over an and chlorine gas emissions in excess of (7) For particulate matter, except for
hourly rolling average (monitored the levels provided by paragraph (b)(6) an area source as defined under § 63.2,
continuously with a continuous of this section; hydrogen chloride and chlorine gas
emissions in excess of the levels cadmium and lead, the standards for (i) Carbon monoxide in excess of 100
provided by paragraph (b)(6) of this arsenic, beryllium, and chromium, the parts per million by volume, over an
section. standards for hydrogen chloride and hourly rolling average (monitored
(c) Destruction and removal efficiency chlorine, and the standards for continuously with a continuous
(DRE) standard. (1) 99.99% DRE. Except particulate matter under this section if emissions monitoring system), dry basis
as provided in paragraph (c)(2) of this they elect under § 266.100(b)(3) of this and corrected to 7 percent oxygen. If
section, you must achieve a DRE of chapter to comply with those standards you elect to comply with this carbon
99.99% for each principle organic in lieu of the standards under 40 CFR monoxide standard rather than the
hazardous constituent (POHC) 266.105, 266.106, and 266.107 to control hydrocarbon standard under paragraph
designated under paragraph (c)(3) of this those pollutants. (a)(5)(ii) of this section, you must also
section. You must calculate DRE for ■ 20. Section 63.1219 and a new document that, during the destruction
each POHC from the following equation: undesignated center heading are added and removal efficiency (DRE) test runs
DRE = [1 ¥ (W out / W in)] × 100% to subpart EEE to read as follows: or their equivalent as provided by
Where: Replacement Emissions Standards and § 63.1206(b)(7), hydrocarbons do not
Win = mass feedrate of one POHC in a Operating Limits for Incinerators, exceed 10 parts per million by volume
waste feedstream; and Cement Kilns, and Lightweight during those runs, over an hourly
Wout = mass emission rate of the same Aggregate Kilns rolling average (monitored continuously
POHC present in exhaust emissions with a continuous emissions monitoring
prior to release to the atmosphere. § 63.1219 What are the replacement system), dry basis, corrected to 7
(2) 99.9999% DRE. If you burn the standards for hazardous waste percent oxygen, and reported as
incinerators? propane; or
dioxin-listed hazardous wastes F020,
F021, F022, F023, F026, or F027 (see (a) Emission limits for existing (ii) Hydrocarbons in excess of 10 parts
§ 261.31 of this chapter), you must sources. You must not discharge or per million by volume, over an hourly
achieve a DRE of 99.9999% for each cause combustion gases to be emitted rolling average (monitored continuously
POHC that you designate under into the atmosphere that contain: with a continuous emissions monitoring
(1) For dioxins and furans: system), dry basis, corrected to 7
paragraph (c)(3) of this section. You (i) For incinerators equipped with
must demonstrate this DRE performance percent oxygen, and reported as
either a waste heat boiler or dry air propane;
on POHCs that are more difficult to pollution control system, either: (6) Hydrogen chloride and chlorine
incinerate than tetra-, penta-, and (A) Emissions in excess of 0.20 ng gas (total chlorine) in excess of 32 parts
hexachlorodibenzo-p-dioxins and TEQ/dscm, corrected to 7 percent per million by volume, combined
dibenzofurans. You must use the oxygen; or emissions, expressed as a chloride
equation in paragraph (c)(1) of this (B) Emissions in excess of 0.40 ng (Cl(¥)) equivalent, dry basis and
section to calculate DRE for each POHC. TEQ/dscm, corrected to 7 percent corrected to 7 percent oxygen; and
In addition, you must notify the oxygen, provided that the combustion (7) Except as provided by paragraph
Administrator of your intent to gas temperature at the inlet to the initial (e) of this section, particulate matter in
incinerate hazardous wastes F020, F021, particulate matter control device is excess of 0.013 gr/dscf corrected to 7
F022, F023, F026, or F027. 400°F or lower based on the average of percent oxygen.
(3) Principal organic hazardous the test run average temperatures. (For (b) Emission limits for new sources.
constituents (POHCs). (i) You must treat purposes of compliance, operation of a You must not discharge or cause
the POHCs in the waste feed that you wet particulate matter control device is combustion gases to be emitted into the
specify under paragraph (c)(3)(ii) of this presumed to meet the 400°F or lower atmosphere that contain:
section to the extent required by requirement); (1)(i) Dioxins and furans in excess of
paragraphs (c)(1) and (c)(2) of this (ii) Emissions in excess of 0.40 ng 0.11 ng TEQ/dscm corrected to 7
section. TEQ/dscm, corrected to 7 percent percent oxygen for incinerators
(ii) You must specify one or more oxygen, for incinerators not equipped equipped with either a waste heat boiler
POHCs that are representative of the with either a waste heat boiler or dry air or dry air pollution control system; or
most difficult to destroy organic pollution control system; (ii) Dioxins and furans in excess of
compounds in your hazardous waste (iii) A source equipped with a wet air 0.20 ng TEQ/dscm corrected to 7
feedstream. You must base this pollution control system followed by a percent oxygen for sources not equipped
specification on the degree of difficulty dry air pollution control system is not with either a waste heat boiler or dry air
of incineration of the organic considered to be a dry air pollution pollution control system;
constituents in the hazardous waste and control system, and a source equipped (iii) A source equipped with a wet air
on their concentration or mass in the with a dry air pollution control system pollution control system followed by a
hazardous waste feed, considering the followed by a wet air pollution control dry air pollution control system is not
results of hazardous waste analyses or system is considered to be a dry air considered to be a dry air pollution
other data and information. pollution control system for purposes of control system, and a source equipped
(d) Significant figures. The emission this standard; with a dry air pollution control system
limits provided by paragraphs (a) and (2) Mercury in excess of 130 µg/dscm, followed by a wet air pollution control
(b) of this section are presented with corrected to 7 percent oxygen; system is considered to be a dry air
two significant figures. Although you (3) Cadmium and lead in excess of pollution control system for purposes of
must perform intermediate calculations 230 µg/dscm, combined emissions, this standard;
using at least three significant figures, corrected to 7 percent oxygen; (2) Mercury in excess of 8.1 µg/dscm,
you may round the resultant emission (4) Arsenic, beryllium, and chromium corrected to 7 percent oxygen;
levels to two significant figures to in excess of 92 µg/dscm, combined (3) Cadmium and lead in excess of 10
document compliance. emissions, corrected to 7 percent µg/dscm, combined emissions,
(e) Elective standards for area oxygen; corrected to 7 percent oxygen;
sources. Area sources as defined under (5) For carbon monoxide and (4) Arsenic, beryllium, and chromium
§ 63.2 are subject to the standards for hydrocarbons, either: in excess of 23 µg/dscm, combined
emissions, corrected to 7 percent hexachlorodibenzo-p-dioxins and manganese, and nickel in excess of 23

oxygen; dibenzofurans. You must use the µg/dscm, combined emissions,
(5) For carbon monoxide and equation in paragraph (c)(1) of this corrected to 7 percent oxygen.
hydrocarbons, either: section to calculate DRE for each POHC. (4) Operating limits. Semivolatile and
(i) Carbon monoxide in excess of 100 In addition, you must notify the low volatile metal operating parameter
parts per million by volume, over an Administrator of your intent to limits must be established to ensure
hourly rolling average (monitored incinerate hazardous wastes F020, F021, compliance with the alternative
continuously with a continuous F022, F023, F026, or F027. emission limitations described in
emissions monitoring system), dry basis (3) Principal organic hazardous paragraphs (e)(2) and (e)(3) of this
and corrected to 7 percent oxygen. If constituent (POHC). (i) You must treat section pursuant to § 63.1209(n), except
you elect to comply with this carbon each POHC in the waste feed that you that semivolatile metal feedrate limits
monoxide standard rather than the specify under paragraph (c)(3)(ii) of this apply to lead, cadmium, and selenium,
hydrocarbon standard under paragraph section to the extent required by combined, and low volatile metal
(b)(5)(ii) of this section, you must also paragraphs (c)(1) and (c)(2) of this feedrate limits apply to arsenic,
document that, during the destruction section. beryllium, chromium, antimony, cobalt,
and removal efficiency (DRE) test runs (ii) You must specify one or more manganese, and nickel, combined.
or their equivalent as provided by POHCs that are representative of the ■ 21. Section 63.1220 is added to
§ 63.1206(b)(7), hydrocarbons do not most difficult to destroy organic subpart EEE to read as follows:
exceed 10 parts per million by volume compounds in your hazardous waste
during those runs, over an hourly feedstream. You must base this § 63.1220 What are the replacement
rolling average (monitored continuously specification on the degree of difficulty standards for hazardous waste burning
with a continuous emissions monitoring of incineration of the organic cement kilns?
system), dry basis, corrected to 7 constituents in the hazardous waste and (a) Emission and hazardous waste
percent oxygen, and reported as on their concentration or mass in the feed limits for existing sources. You
propane; or hazardous waste feed, considering the must not discharge or cause combustion
(ii) Hydrocarbons in excess of 10 parts results of hazardous waste analyses or gases to be emitted into the atmosphere
per million by volume, over an hourly other data and information. or feed hazardous waste that contain:
rolling average (monitored continuously (d) Significant figures. The emission (1) For dioxins and furans, either:
with a continuous emissions monitoring limits provided by paragraphs (a) and (i) Emissions in excess of 0.20 ng
system), dry basis, corrected to 7 (b) of this section are presented with TEQ/dscm corrected to 7 percent
percent oxygen, and reported as two significant figures. Although you oxygen; or
propane; must perform intermediate calculations (ii) Emissions in excess of 0.40 ng
(6) Hydrogen chloride and chlorine using at least three significant figures, TEQ/dscm corrected to 7 percent
gas in excess of 21 parts per million by you may round the resultant emission oxygen provided that the combustion
volume, combined emissions, expressed levels to two significant figures to gas temperature at the inlet to the initial
as a chloride (Cl(-)) equivalent, dry basis document compliance. dry particulate matter control device is
and corrected to 7 percent oxygen; and (e) Alternative to the particulate 400 °F or lower based on the average of
(7) Except as provided by paragraph matter standard. (1). General. In lieu of the test run average temperatures;
(e) of this section, particulate matter in complying with the particulate matter (2) For mercury, both:
excess of 0.0015 gr/dscf, corrected to 7 standards of this section, you may elect (i) An average as-fired concentration
percent oxygen. to comply with the following alternative of mercury in all hazardous waste
(c) Destruction and removal efficiency metal emission control requirement: feedstreams in excess of 3.0 parts per
(DRE) standard. (1) 99.99% DRE. Except (2) Alternative metal emission control million by weight; and
as provided in paragraph (c)(2) of this requirements for existing incinerators. (ii) Emissions in excess of 120 µg/
section, you must achieve a destruction (i) You must not discharge or cause dscm, corrected to 7 percent oxygen; or
and removal efficiency (DRE) of 99.99% combustion gases to be emitted into the (iii) A hazardous waste feedrate
for each principle organic hazardous atmosphere that contain cadmium, lead, corresponding to a maximum theoretical
constituent (POHC) designated under and selenium in excess of 230 µg/dscm, emission concentration (MTEC) in
paragraph (c)(3) of this section. You combined emissions, corrected to 7 excess of 120 µg/dscm;
must calculate DRE for each POHC from percent oxygen; and, (3) For cadmium and lead, both:
the following equation: (ii) You must not discharge or cause (i) Emissions in excess of 7.6 × 10¥4
combustion gases to be emitted into the lbs combined emissions of cadmium
DRE = [1 ¥ (Wout / Win)] × 100%
atmosphere that contain antimony, and lead attributable to the hazardous
Where: arsenic, beryllium, chromium, cobalt, waste per million Btu heat input from
Win = mass feedrate of one POHC in a manganese, and nickel in excess of 92 the hazardous waste; and
waste feedstream; and µg/dscm, combined emissions, (ii) Emissions in excess of 330 µg/
Wout = mass emission rate of the same corrected to 7 percent oxygen. dscm, combined emissions, corrected to
POHC present in exhaust emissions (3) Alternative metal emission control 7 percent oxygen;
prior to release to the atmosphere. requirements for new incinerators. (i) (4) For arsenic, beryllium, and
(2) 99.9999% DRE. If you burn the You must not discharge or cause chromium, both:
dioxin-listed hazardous wastes F020, combustion gases to be emitted into the (i) Emissions in excess of 2.1 × 10¥5
F021, F022, F023, F026, or F027 (see atmosphere that contain cadmium, lead, lbs combined emissions of arsenic,
§ 261.31 of this chapter), you must and selenium in excess of 10 µg/dscm, beryllium, and chromium attributable to
achieve a DRE of 99.9999% for each combined emissions, corrected to 7 the hazardous waste per million Btu
POHC that you designate under percent oxygen; and, heat input from the hazardous waste;
paragraph (c)(3) of this section. You (ii) You must not discharge or cause and
must demonstrate this DRE performance combustion gases to be emitted into the (ii) Emissions in excess of 56 µg/
on POHCs that are more difficult to atmosphere that contain antimony, dscm, combined emissions, corrected to
incinerate than tetra-, penta-, and arsenic, beryllium, chromium, cobalt, 7 percent oxygen;
(5) Carbon monoxide and system), dry basis, corrected to 7 sampling system and the main stack as
hydrocarbons. (i) For kilns equipped percent oxygen, and reported as follows:
with a by-pass duct or midkiln gas propane. (A) Emissions in the by-pass or
sampling system, either: (6) Hydrogen chloride and chlorine midkiln gas sampling system are limited
(A) Carbon monoxide in the by-pass gas in excess of 120 parts per million by to either:
duct or mid-kiln gas sampling system in volume, combined emissions, expressed (1) Carbon monoxide in excess of 100
excess of 100 parts per million by as a chloride (Cl(-)) equivalent, dry basis, parts per million by volume, over an
volume, over an hourly rolling average corrected to 7 percent oxygen; and hourly rolling average (monitored
(monitored continuously with a (7) For particulate matter, both: continuously with a continuous
continuous emissions monitoring (i) Emissions in excess of 0.028 gr/ emissions monitoring system), dry basis
system), dry basis and corrected to 7 dscf corrected to 7 percent oxygen; and and corrected to 7 percent oxygen. If
percent oxygen. If you elect to comply (ii) Opacity greater than 20 percent, you elect to comply with this carbon
with this carbon monoxide standard unless your source is equipped with a monoxide standard rather than the
rather than the hydrocarbon standard bag leak detection system under hydrocarbon standard under paragraph
under paragraph (a)(5)(i)(B) of this § 63.1206(c)(8) or a particulate matter (b)(5)(i)(A)(2) of this section, you also
section, you must also document that, detection system under § 63.1206(c)(9). must document that, during the
during the destruction and removal (b) Emission and hazardous waste destruction and removal efficiency
efficiency (DRE) test runs or their feed limits for new sources. You must (DRE) test runs or their equivalent as
equivalent as provided by not discharge or cause combustion gases provided by § 63.1206(b)(7),
§ 63.1206(b)(7), hydrocarbons in the by- to be emitted into the atmosphere or hydrocarbons do not exceed 10 parts per
pass duct or mid-kiln gas sampling feed hazardous waste that contain: million by volume during those runs,
system do not exceed 10 parts per (1) For dioxins and furans, either: over an hourly rolling average
million by volume during those runs, (i) Emissions in excess of 0.20 ng (monitored continuously with a
over an hourly rolling average TEQ/dscm corrected to 7 percent continuous emissions monitoring
(monitored continuously with a oxygen; or system), dry basis, corrected to 7
continuous emissions monitoring (ii) Emissions in excess of 0.40 ng percent oxygen, and reported as
system), dry basis, corrected to 7 TEQ/dscm corrected to 7 percent propane; or
percent oxygen, and reported as oxygen provided that the combustion (2) Hydrocarbons in the by-pass duct
propane; or gas temperature at the inlet to the initial or midkiln gas sampling system in
(B) Hydrocarbons in the by-pass duct dry particulate matter control device is excess of 10 parts per million by
or midkiln gas sampling system in 400 °F or lower based on the average of volume, over an hourly rolling average
excess of 10 parts per million by the test run average temperatures; (monitored continuously with a
volume, over an hourly rolling average (2) For mercury, both: continuous emissions monitoring
(monitored continuously with a (i) An average as-fired concentration system), dry basis, corrected to 7
continuous emissions monitoring of mercury in all hazardous waste percent oxygen, and reported as
system), dry basis, corrected to 7 feedstreams in excess of 1.9 parts per propane; and
percent oxygen, and reported as million by weight; and (B) Hydrocarbons in the main stack
propane; (ii) Emissions in excess of 120 µg/ are limited, if construction of the kiln
(ii) For kilns not equipped with a by- dscm, corrected to 7 percent oxygen; or commenced after April 19, 1996 at a
pass duct or midkiln gas sampling (iii) A hazardous waste feedrate plant site where a cement kiln (whether
system, either: corresponding to a maximum theoretical burning hazardous waste or not) did not
(A) Hydrocarbons in the main stack in emission concentration (MTEC) in previously exist, to 50 parts per million
excess of 20 parts per million by excess of 120 µg/dscm; by volume, over a 30-day block average
volume, over an hourly rolling average (3) For cadmium and lead, both: (monitored continuously with a
(monitored continuously with a (i) Emissions in excess of 6.2 x 10-5 lbs continuous monitoring system), dry
continuous emissions monitoring combined emissions of cadmium and basis, corrected to 7 percent oxygen, and
system), dry basis, corrected to 7 lead attributable to the hazardous waste reported as propane.
percent oxygen, and reported as per million Btu heat input from the (ii) For kilns not equipped with a by-
propane; or hazardous waste; and pass duct or midkiln gas sampling
(B) Carbon monoxide in the main (ii) Emissions in excess of 180 µg/ system, hydrocarbons and carbon
stack in excess of 100 parts per million dscm, combined emissions, corrected to monoxide are limited in the main stack
by volume, over an hourly rolling 7 percent oxygen; to either:
average (monitored continuously with a (4) For arsenic, beryllium, and (A) Hydrocarbons not exceeding 20
continuous emissions monitoring chromium, both: parts per million by volume, over an
system), dry basis and corrected to 7 (i) Emissions in excess of 1.5 x 10-5 lbs hourly rolling average (monitored
percent oxygen. If you elect to comply combined emissions of arsenic, continuously with a continuous
with this carbon monoxide standard beryllium, and chromium attributable to emissions monitoring system), dry basis,
rather than the hydrocarbon standard the hazardous waste per million Btu corrected to 7 percent oxygen, and
under paragraph (a)(5)(ii)(A) of this heat input from the hazardous waste; reported as propane; or
section, you also must document that, and (B)(1) Carbon monoxide not exceeding
during the destruction and removal (ii) Emissions in excess of 54 µg/ 100 parts per million by volume, over
efficiency (DRE) test runs or their dscm, combined emissions, corrected to an hourly rolling average (monitored
equivalent as provided by 7 percent oxygen; continuously with a continuous
§ 63.1206(b)(7), hydrocarbons in the (5) Carbon monoxide and emissions monitoring system), dry basis,
main stack do not exceed 20 parts per hydrocarbons. (i) For kilns equipped corrected to 7 percent oxygen; and
million by volume during those runs, with a by-pass duct or midkiln gas (2) Hydrocarbons not exceeding 20
over an hourly rolling average sampling system, carbon monoxide and parts per million by volume, over an
(monitored continuously with a hydrocarbons emissions are limited in hourly rolling average (monitored
continuous emissions monitoring both the bypass duct or midkiln gas continuously with a continuous
monitoring system), dry basis, corrected specify under paragraph (c)(3)(ii) of this Ctotal = {Cmill-off × (Tmill-off / (Tmill-off +
to 7 percent oxygen, and reported as section to the extent required by Tmill-on))} + {Cmill-on × (Tmill-on /
propane at any time during the paragraphs (c)(1) and (c)(2) of this (Tmill-off + Tmill-on))}
destruction and removal efficiency section. Where:
(DRE) test runs or their equivalent as (ii) You must specify one or more Ctotal = time-weighted average
provided by § 63.1206(b)(7); and POHCs that are representative of the concentration of a regulated constituent
(3) If construction of the kiln most difficult to destroy organic considering both raw mill on time and
commenced after April 19, 1996 at a compounds in your hazardous waste off time;
plant site where a cement kiln (whether feedstream. You must base this Cmill-off = average performance test
burning hazardous waste or not) did not specification on the degree of difficulty concentration of regulated constituent
previously exist, hydrocarbons are of incineration of the organic with the raw mill off-line;
limited to 50 parts per million by constituents in the hazardous waste and Cmill-on = average performance test
volume, over a 30-day block average on their concentration or mass in the concentration of regulated
(monitored continuously with a hazardous waste feed, considering the constituent with the raw mill on-
continuous monitoring system), dry results of hazardous waste analyses or line;
basis, corrected to 7 percent oxygen, and other data and information. Tmill-off = time when kiln gases are not
reported as propane. (d) Cement kilns with in-line kiln raw routed through the raw mill; and
(6) Hydrogen chloride and chlorine mills. (1) General. (i) You must conduct Tmill-on = time when kiln gases are
gas in excess of 86 parts per million by performance testing when the raw mill routed through the raw mill.
volume, combined emissions, expressed is on-line and when the mill is off-line (ii) Compliance. (A) If you use this
as a chloride (Cl(-)) equivalent, dry basis to demonstrate compliance with the emission averaging provision, you must
and corrected to 7 percent oxygen; and emission standards, and you must document in the operating record
(7) For particulate matter, both: establish separate operating parameter
(i) Emissions in excess of 0.0023 gr/ compliance with the emission standards
limits under § 63.1209 for each mode of on an annual basis by using the
dscf corrected to 7 percent oxygen; and operation, except as provided by
(ii) Opacity greater than 20 percent, equation provided by paragraph (d)(2) of
paragraphs (d)(1)(iv) and (d)(1)(v) of this this section.
unless your source is equipped with a section.
bag leak detection system under (B) Compliance is based on one-year
(ii) You must document in the block averages beginning on the day you
§ 63.1206(c)(8) or a particulate matter operating record each time you change
detection system under § 63.1206(c)(9). submit the initial notification of
from one mode of operation to the compliance.
(c) Destruction and removal efficiency alternate mode and begin complying
(DRE) standard. (1) 99.99% DRE. Except (iii) Notification. (A) If you elect to
with the operating parameter limits for document compliance with one or more
as provided in paragraph (c)(2) of this that alternate mode of operation.
section, you must achieve a destruction emission standards using this emission
(iii) You must calculate rolling averaging provision, you must notify the
and removal efficiency (DRE) of 99.99% averages for operating parameter limits
for each principle organic hazardous Administrator in the initial
as provided by § 63.1209(q)(2). comprehensive performance test plan
constituent (POHC) designated under (iv) If your in-line kiln raw mill has
paragraph (c)(3) of this section. You submitted under § 63.1207(e).
dual stacks, you may assume that the (B) You must include historical raw
must calculate DRE for each POHC from dioxin/furan emission levels in the by-
the following equation: mill operation data in the performance
pass stack and the operating parameter test plan to estimate future raw mill
DRE = [1 ¥ (Wout / Win)] × 100% limits determined during performance down-time and document in the
Where: testing of the by-pass stack when the performance test plan that estimated
Win = mass feedrate of one POHC in a raw mill is off-line are the same as when emissions and estimated raw mill down-
waste feedstream; and the mill is on-line. time will not result in an exceedance of
Wout = mass emission rate of the same (v) In lieu of conducting a an emission standard on an annual
POHC present in exhaust emissions performance test to demonstrate basis.
prior to release to the atmosphere. compliance with the dioxin/furan (C) You must document in the
(2) 99.9999% DRE. If you burn the emission standards for the mode of notification of compliance submitted
dioxin-listed hazardous wastes F020, operation when the raw mill is on-line, under § 63.1207(j) that an emission
F021, F022, F023, F026, or F027 (see you may specify in the performance test standard will not be exceeded based on
§ 261.31 of this chapter), you must workplan and Notification of the documented emissions from the
achieve a DRE of 99.9999% for each Compliance the same operating performance test and predicted raw mill
POHC that you designate under parameter limits required under down-time.
paragraph (c)(3) of this section. You § 63.1209(k) for the mode of operation (e) Preheater or preheater/precalciner
must demonstrate this DRE performance when the raw mill is on-line as you kilns with dual stacks. (1) General. You
on POHCs that are more difficult to establish during performance testing for must conduct performance testing on
incinerate than tetra-, penta-, and the mode of operation when the raw each stack to demonstrate compliance
hexachlorodibenzo-p-dioxins and mill is off-line. with the emission standards, and you
dibenzofurans. You must use the (2) Emissions averaging. You may must establish operating parameter
equation in paragraph (c)(1) of this comply with the mercury, semivolatile limits under § 63.1209 for each stack,
section to calculate DRE for each POHC. metal, low volatile metal, and hydrogen except as provided by paragraph
In addition, you must notify the chloride/chlorine gas emission (d)(1)(iv) of this section for dioxin/furan
Administrator of your intent to standards on a time-weighted average emissions testing and operating
incinerate hazardous wastes F020, F021, basis under the following procedures: parameter limits for the by-pass stack of
F022, F023, F026, or F027. (i) Averaging methodology. You must in-line raw mills.
(3) Principal organic hazardous calculate the time-weighted average (2) Emissions averaging. You may
constituent (POHC). (i) You must treat emission concentration with the comply with the mercury, semivolatile
each POHC in the waste feed that you following equation: metal, low volatile metal, and hydrogen
chloride/chlorine gas emission (g) [Reserved]. under paragraph (a)(5)(ii) of this section,
standards specified in this section on a (h) When you comply with the you also must document that, during the
gas flowrate-weighted average basis particulate matter requirements of destruction and removal efficiency
under the following procedures: paragraphs (a)(7) or (b)(7) of this section, (DRE) test runs or their equivalent as
(i) Averaging methodology. You must you are exempt from the New Source provided by § 63.1206(b)(7),
calculate the gas flowrate-weighted Performance Standard for particulate hydrocarbons do not exceed 20 parts per
average emission concentration using matter and opacity under § 60.60 of this million by volume during those runs,
the following equation: chapter. over an hourly rolling average
Ctot = {Cmain × (Qmain / (Qmain + Qbypass))} ■ 22. Section 63.1221 is added to (monitored continuously with a
+ {Cbypass × (Qbypass / (Qmain + subpart EEE to read as follows: continuous emissions monitoring
Qbypass))} system), dry basis, corrected to 7
Where: § 63.1221 What are the replacement percent oxygen, and reported as
standards for hazardous waste burning propane; or
Ctot = gas flowrate-weighted average lightweight aggregate kilns?
concentration of the regulated (ii) Hydrocarbons in excess of 20 parts
(a) Emission and hazardous waste per million by volume, over an hourly
constituent;
Cmain = average performance test feed limits for existing sources. You rolling average, dry basis, corrected to 7
concentration demonstrated in the must not discharge or cause combustion percent oxygen, and reported as
main stack; gases to be emitted into the atmosphere propane;
Cbypass = average performance test or feed hazardous waste that contain: (6) Hydrogen chloride and chlorine
concentration demonstrated in the (1) For dioxins and furans, either: gas in excess of 600 parts per million by
bypass stack; (i) Emissions in excess of 0.20 ng volume, combined emissions, expressed
Qmain = volumetric flowrate of main TEQ/dscm corrected to 7 percent as a chloride (Cl(-)) equivalent, dry basis
stack effluent gas; and oxygen; or and corrected to 7 percent oxygen; and
Qbypass = volumetric flowrate of bypass (ii) Rapid quench of the combustion
(7) Particulate matter emissions in
effluent gas. gas temperature at the exit of the (last)
excess of 0.025 gr/dscf, corrected to 7
(ii) Compliance. (A) You must combustion chamber (or exit of any
percent oxygen.
demonstrate compliance with the waste heat recovery system that
(b) Emission and hazardous waste
emission standard(s) using the emission immediately follows the last
feed limits for new sources. You must
concentrations determined from the combustion chamber) to 400°F or lower
not discharge or cause combustion gases
performance tests and the equation based on the average of the test run
to be emitted into the atmosphere or
provided by paragraph (e)(1) of this average temperatures. You must also
feed hazardous waste that contain:
section; and notify in writing the RCRA authority
(1) For dioxins and furans, either:
(B) You must develop operating that you are complying with this option;
(2) For mercury, either: (i) Emissions in excess of 0.20 ng
parameter limits for bypass stack and TEQ/dscm corrected to 7 percent
main stack flowrates that ensure the (i) Emissions in excess of 120 µg/
dscm, corrected to 7 percent oxygen; or oxygen; or
emission concentrations calculated with (ii) Rapid quench of the combustion
the equation in paragraph (e)(1) of this (ii) A hazardous waste feedrate
corresponding to a maximum theoretical gas temperature at the exit of the (last)
section do not exceed the emission combustion chamber (or exit of any
standards on a 12-hour rolling average emission concentration (MTEC) in
excess of 120 µg/dscm; waste heat recovery system that
basis. You must include these flowrate immediately follows the last
limits in the Notification of Compliance. (3) For cadmium and lead, both:
(i) Emissions in excess of 3.0 × 10¥4 combustion chamber) to 400°F or lower
(iii) Notification. If you elect to
lbs combined emissions of cadmium based on the average of the test run
document compliance under this
and lead attributable to the hazardous average temperatures. You must also
emissions averaging provision, you
waste per million Btu heat input from notify in writing the RCRA authority
must:
(A) Notify the Administrator in the the hazardous waste; and that you are complying with this option;
initial comprehensive performance test (ii) Emissions in excess of 250 µg/ (2) For mercury, either:
plan submitted under § 63.1207(e). The dscm, combined emissions, corrected to (i) Emissions in excess of 120 µg/
performance test plan must include, at 7 percent oxygen; dscm, corrected to 7 percent oxygen; or
a minimum, information describing the (4) For arsenic, beryllium, and (ii) A hazardous waste feedrate
flowrate limits established under chromium, both: corresponding to a maximum theoretical
paragraph (e)(2)(ii)(B) of this section; (i) In excess of 9.5 × 10¥5 lbs emission concentration (MTEC) in
and combined emissions of arsenic, excess of 120 µg/dscm;
(B) Document in the Notification of beryllium, and chromium attributable to (3) For cadmium and lead, both:
Compliance submitted under the hazardous waste per million Btu (i) Emissions in excess of 3.7 × 10¥5
§ 63.1207(j) the demonstrated gas heat input from the hazardous waste; lbs combined emissions of cadmium
flowrate-weighted average emissions (ii) Emissions in excess of 110 µg/ and lead attributable to the hazardous
that you calculate with the equation dscm, combined emissions, corrected to waste per million Btu heat input from
provided by paragraph (e)(2) of this 7 percent oxygen; the hazardous waste; and
section. (5) Carbon monoxide and (ii) Emissions in excess of 43 µg/
(f) Significant figures. The emission hydrocarbons. (i) Carbon monoxide in dscm, combined emissions, corrected to
limits provided by paragraphs (a) and excess of 100 parts per million by 7 percent oxygen;
(b) of this section are presented with volume, over an hourly rolling average (4) For arsenic, beryllium, and
two significant figures. Although you (monitored continuously with a chromium, both:
must perform intermediate calculations continuous emissions monitoring (i) In excess of 3.3 × 10¥5 lbs
using at least three significant figures, system), dry basis and corrected to 7 combined emissions of arsenic,
you may round the resultant emission percent oxygen. If you elect to comply beryllium, and chromium attributable to
levels to two significant figures to with this carbon monoxide standard the hazardous waste per million Btu
document compliance. rather than the hydrocarbon standard heat input from the hazardous waste;
(ii) Emissions in excess of 110 µg/ dibenzofurans. You must use the PART 264—STANDARDS FOR
dscm, combined emissions, corrected to equation in paragraph (c)(1) of this OWNERS AND OPERATORS OF
7 percent oxygen; section to calculate DRE for each POHC. HAZARDOUS WASTE TREATMENT,
(5) Carbon monoxide and In addition, you must notify the STORAGE, AND DISPOSAL
hydrocarbons. (i) Carbon monoxide in Administrator of your intent to burn FACILITIES
excess of 100 parts per million by hazardous wastes F020, F021, F022,
volume, over an hourly rolling average F023, F026, or F027. ■ 1. The authority citation for part 264
(monitored continuously with a continues to read as follows:
(3) Principal organic hazardous
continuous emissions monitoring constituents (POHCs). (i) You must treat Authority: 42 U.S.C. 6905, 6912(a), 6924,
system), dry basis and corrected to 7 each POHC in the waste feed that you 6925, 6927, 6928(h), and 6974.
percent oxygen. If you elect to comply specify under paragraph (c)(3)(ii) of this
with this carbon monoxide standard ■ 2. Section 264.340 is amended by
section to the extent required by revising the first sentence of paragraph
rather than the hydrocarbon standard paragraphs (c)(1) and (c)(2) of this
under paragraph (b)(5)(ii) of this section, (b)(1) and adding paragraph (b)(5) to
section. read as follows:
you also must document that, during the
destruction and removal efficiency (ii) You must specify one or more
§ 264.340 Applicability.
(DRE) test runs or their equivalent as POHCs that are representative of the
most difficult to destroy organic * * * * *
provided by § 63.1206(b)(7), (b) * * * (1) Except as provided by
hydrocarbons do not exceed 20 parts per compounds in your hazardous waste
feedstream. You must base this paragraphs (b)(2) through (b)(5) of this
million by volume during those runs, section, the standards of this part do not
over an hourly rolling average specification on the degree of difficulty
of incineration of the organic apply to a new hazardous waste
(monitored continuously with a incineration unit that becomes subject
continuous emissions monitoring constituents in the hazardous waste and
on their concentration or mass in the to RCRA permit requirements after
system), dry basis, corrected to 7 October 12, 2005; or no longer apply
percent oxygen, and reported as hazardous waste feed, considering the
results of hazardous waste analyses or when an owner or operator of an
propane; or existing hazardous waste incineration
(ii) Hydrocarbons in excess of 20 parts other data and information.
unit demonstrates compliance with the
per million by volume, over an hourly (d) Significant figures. The emission maximum achievable control
rolling average, dry basis, corrected to 7 limits provided by paragraphs (a) and technology (MACT) requirements of part
percent oxygen, and reported as (b) of this section are presented with 63, subpart EEE, of this chapter by
propane; two significant figures. Although you conducting a comprehensive
(6) Hydrogen chloride and chlorine must perform intermediate calculations performance test and submitting to the
gas in excess of 600 parts per million by using at least three significant figures, Administrator a Notification of
volume, combined emissions, expressed you may round the resultant emission Compliance under §§ 63.1207(j) and
as a chloride (Cl(-)) equivalent, dry basis levels to two significant figures to 63.1210(d) of this chapter documenting
and corrected to 7 percent oxygen; and document compliance.
(7) Particulate matter emissions in compliance with the requirements of
excess of 0.0098 gr/dscf corrected to 7 part 63, subpart EEE, of this chapter.
PART 260—HAZARDOUS WASTE
percent oxygen. * * *
MANAGEMENT SYSTEM: GENERAL
(c) Destruction and removal efficiency * * * * *
(DRE) standard. (1) 99.99% DRE. Except (5) The particulate matter standard of
■ 1. The authority citation for part 260
as provided in paragraph (c)(2) of this § 264.343(c) remains in effect for
continues to read as follows:
section, you must achieve a destruction incinerators that elect to comply with
Authority: 42 U.S.C. 6905, 6912(a), 6921– the alternative to the particulate matter
and removal efficiency (DRE) of 99.99%
6927, 6930, 6934, 6935, 6937, 6938, 6939, standard of §§ 63.1206(b)(14) and
for each principal organic hazardous and 6974.
constituent (POHC) designated under 63.1219(e) of this chapter.
paragraph (c)(3) of this section. You ■ 2. Section 260.11 is amended by * * * * *
must calculate DRE for each POHC from
■ a. Revising the first sentence in PART 265—INTERIM STATUS
the following equation:
paragraph (a). STANDARDS FOR OWNERS AND
DRE = [1 ¥ (Wout / Win)] × 100%
■ b. Revising paragraph (c)(1). OPERATORS OF HAZARDOUS WASTE
Where:
TREATMENT, STORAGE, AND
Win = mass feedrate of one POHC in The revisions and additions read as
DISPOSAL FACILITIES
a waste feedstream; and follows:
Wout = mass emission rate of the same ■ 1. The authority citation for part 265
§ 260.11 References.
POHC present in exhaust emissions continues to read as follows:
prior to release to the atmosphere. (a) When used in parts 260 through Authority: 42 U.S.C. 6905, 6906, 6912,
(2) 99.9999% DRE. If you burn the 268 of this chapter, the following 6922, 6923, 6924, 6925, 6935, 6936, and
dioxin-listed hazardous wastes F020, publications are incorporated by 6937.
F021, F022, F023, F026, or F027 (see reference. * * * ■ 2. Section 265.340 is amended by
§ 261.31 of this chapter), you must * * * * * revising paragraph (b)(1) to read as
achieve a destruction and removal follows:
(c) * * *
efficiency (DRE) of 99.9999% for each
POHC that you designate under (1) ‘‘APTI Course 415: Control of § 265.340 Applicability.
paragraph (c)(3) of this section. You Gaseous Emissions,’’ EPA Publication * * * * *
must demonstrate this DRE performance EPA–450/2–81–005, December 1981, (b) * * * (1) Except as provided by
on POHCs that are more difficult to IBR approved for §§ 264.1035 and paragraphs (b)(2) and (b)(3) of this
incinerate than tetra-, penta-, and 265.1035. section, the standards of this part no
hexachlorodibenzo-dioxins and * * * * * longer apply when an owner or operator
demonstrates compliance with the (ii) Section 266.107—Standards to compliance with the standards of 40
maximum achievable control control hydrogen chloride and chlorine CFR part 63, subpart EEE alone may not
technology (MACT) requirements of part gas. be protective of human health or the
63, subpart EEE, of this chapter by (4) The particulate matter standard of environment, the Director shall require
conducting a comprehensive § 266.105 remains in effect for boilers the additional information or
performance test and submitting to the that elect to comply with the alternative assessment(s) necessary to determine
Administrator a Notification of to the particulate matter standard under whether additional controls are
Compliance under §§ 63.1207(j) and §§ 63.1216(e) and 63.1217(e) of this necessary to ensure protection of human
63.1210(d) of this chapter documenting chapter. health and the environment. This
compliance with the requirements of * * * * * includes information necessary to
part 63, subpart EEE, of this chapter. evaluate the potential risk to human
* * * * * PART 270—EPA ADMINISTERED health and/or the environment resulting
PERMIT PROGRAMS: THE from both direct and indirect exposure
PART 266—STANDARDS FOR THE HAZARDOUS WASTE PERMIT pathways. The Director may also require
MANAGEMENT OF SPECIFIC PROGRAM a permittee or applicant to provide
HAZARDOUS WASTES AND SPECIFIC information necessary to determine
TYPES OF HAZARDOUS WASTE ■ 1. The authority citation for part 270 whether such an assessment(s) should
MANAGEMENT FACILITIES continues to read as follows: be required.
Authority: 42 U.S.C. 6905, 6912, 6924, (1) The Director shall base the
■ 1. The authority citation for part 266 6925, 6927, 6939, and 6974. evaluation of whether compliance with
continues to read as follows: ■ 2. Section 270.6 is revised to read as the standards of 40 CFR part 63, subpart
Authority: 42 U.S.C. 1006, 2002(a), 3001– follows: EEE alone is protective of human health
3009, 3014, 6905, 6906, 6912, 6921, 6922, or the environment on factors relevant
6924–6927, 6934, and 6937. § 270.6 References.
to the potential risk from a hazardous
■ 2. Section 266.100 is amended by (a) When used in part 270 of this waste combustion unit, including, as
revising the first sentence of paragraph chapter, the following publications are appropriate, any of the following
(b)(1) and adding paragraphs (b)(3) and incorporated by reference. These factors:
(b)(4) to read as follows: incorporations by reference were (i) Particular site-specific
approved by the Director of the Federal considerations such as proximity to
§ 266.100 Applicability. Register pursuant to 5 U.S.C. 552(a) and receptors (such as schools, hospitals,
* * * * * 1 CFR part 51. These materials are nursing homes, day care centers, parks,
(b) * * * (1) Except as provided by incorporated as they exist on the date of community activity centers, or other
paragraphs (b)(2), (b)(3), and (b)(4) of approval and a notice of any change in potentially sensitive receptors), unique
this section, the standards of this part these materials will be published in the dispersion patterns, etc.;
do not apply to a new hazardous waste Federal Register. Copies may be (ii) Identities and quantities of
boiler or industrial furnace unit that inspected at the Library, U.S. emissions of persistent, bioaccumulative
becomes subject to RCRA permit Environmental Protection Agency, 1200 or toxic pollutants considering
requirements after October 12, 2005; or Pennsylvania Ave., NW., (3403T), enforceable controls in place to limit
no longer apply when an owner or Washington, DC 20460, those pollutants;
operator of an existing hazardous waste libraryhq@epa.gov; or at the National (iii) Identities and quantities of
boiler or industrial furnace unit Archives and Records Administration nondioxin products of incomplete
demonstrates compliance with the (NARA). For information on the combustion most likely to be emitted
maximum achievable control availability of this material at NARA, and to pose significant risk based on
technology (MACT) requirements of part call 202–741–6030, or go to: http:// known toxicities (confirmation of which
63, subpart EEE, of this chapter by www.archives.gov/federal_register/ should be made through emissions
conducting a comprehensive code_of_federal_regulations/ testing);
performance test and submitting to the ibr_locations.html. (iv) Identities and quantities of other
Administrator a Notification of (b) The following materials are off-site sources of pollutants in
Compliance under §§ 63.1207(j) and available for purchase from the National proximity of the facility that
63.1210(d) of this chapter documenting Technical Information Service (NTIS), significantly influence interpretation of
compliance with the requirements of 5285 Port Royal Road, Springfield, VA a facility-specific risk assessment;
part 63, subpart EEE, of this chapter. 22161, (703) 605–6000 or (800) 553– (v) Presence of significant ecological
* * * 6847; or for purchase from the considerations, such as the proximity of
* * * * * Superintendent of Documents, U.S. a particularly sensitive ecological area;
(3) If you own or operate a boiler or Government Printing Office, (vi) Volume and types of wastes, for
hydrochloric acid production furnace Washington, DC 20402, (202) 512–1800: example wastes containing highly toxic
that is an area source under § 63.2 of (1) ‘‘APTI Course 415: Control of constituents;
this chapter and you elect not to comply Gaseous Emissions,’’ EPA Publication (vii) Other on-site sources of
with the emission standards under EPA–450/2–81–005, December 1981, hazardous air pollutants that
§§ 63.1216, 63.1217, and 63.1218 of this IBR approved for §§ 270.24 and 270.25. significantly influence interpretation of
chapter for particulate matter, (2) [Reserved]. the risk posed by the operation of the
semivolatile and low volatile metals, ■ 3. Section 270.10 is amended by source in question;
and total chlorine, you also remain adding paragraph (l) to read as follows: (viii) Adequacy of any previously
subject to: conducted risk assessment, given any
(i) Section 266.105—Standards to § 270.10 General application requirements. subsequent changes in conditions likely
control particulate matter; * * * * * to affect risk; and
(ii) Section 266.106—Standards to (l) If the Director concludes, based on (ix) Such other factors as may be
control metals emissions, except for one or more of the factors listed in appropriate.
mercury; and paragraph (l)(1) of this section that (2) [Reserved]
■ 4. Section 270.19 is amended by necessary to ensure compliance with § 270.32 Establishing permit conditions.
revising paragraph (e) to reads as §§ 266.102(e)(1) and 266.102(e)(2)(iii) of * * * * *
follows: this chapter if you elect to comply with (b) * * *
§ 270.235(a)(1)(i) to minimize emissions (3) If, as the result of an assessment(s)
§ 270.19 Specific part B information or other information, the Administrator
of toxic compounds from startup,
requirements for incinerators.
shutdown, and malfunction events; or if or Director determines that conditions
* * * * * you are an area source and elect to are necessary in addition to those
(e) When an owner or operator of a comply with the §§ 266.105, 266.106, required under 40 CFR parts 63, subpart
hazardous waste incineration unit and 266.107 standards and associated EEE, 264 or 266 to ensure protection of
becomes subject to RCRA permit requirements for particulate matter, human health and the environment, he
requirements after October 12, 2005, or hydrogen chloride and chlorine gas, and shall include those terms and
when an owner or operator of an non-mercury metals; or the Director conditions in a RCRA permit for a
existing hazardous waste incineration determines certain provisions apply, on hazardous waste combustion unit.
unit demonstrates compliance with the a case-by-case basis, for purposes of
air emission standards and limitations * * * * *
information collection in accordance ■ 9. Section 270.42 is amended by:
in part 63, subpart EEE, of this chapter with §§ 270.10(k), 270.10(l), ■ a. Revising paragraph (j)(1).
(i.e., by conducting a comprehensive 270.32(b)(2), and 270.32(b)(3). ■ b. Redesignating paragraph (j)(2) as
performance test and submitting a
* * * * * (j)(3).
Notification of Compliance under ■ c. Adding new paragraph (j)(2).
§§ 63.1207(j) and 63.1210(d) of this ■ 6. Section 270.24 is amended by
■ d. Adding new paragraph (k); and
chapter documenting compliance with revising paragraph (d)(3) to read as ■ e. Adding a new entry 10 in numerical
all applicable requirements of part 63, follows: order in the table under section L of
subpart EEE, of this chapter), the Appendix I.
§ 270.24 Specific part B information
requirements of this section do not requirements for process vents. The revisions and additions read as
apply, except those provisions the follows:
Director determines are necessary to * * * * *
ensure compliance with §§ 264.345(a) (d) * * * § 270.42 Permit modification at the request
and 264.345(c) of this chapter if you (3) A design analysis, specifications, of the permittee.
elect to comply with § 270.235(a)(1)(i) to drawings, schematics, and piping and * * * * *
minimize emissions of toxic compounds instrumentation diagrams based on the (j) * * *
from startup, shutdown, and appropriate sections of ‘‘APTI Course (1) Facility owners or operators must
malfunction events. Nevertheless, the 415: Control of Gaseous Emissions’’ have complied with the Notification of
Director may apply the provisions of (incorporated by reference as specified Intent to Comply (NIC) requirements of
this section, on a case-by-case basis, for in § 270.6) or other engineering texts 40 CFR 63.1210 that were in effect prior
purposes of information collection in acceptable to the Regional to October 11, 2000, (See 40 CFR part
accordance with §§ 270.10(k), 270.10(l), Administrator that present basic control 63 §§ 63.1200–63.1499 revised as of July
270.32(b)(2), and 270.32(b)(3). device information. The design analysis 1, 2000) in order to request a permit
■ 5. Section 270.22 is amended by shall address the vent stream modification under this section for the
revising the introductory text to read as characteristics and control device purpose of technology changes needed
follows: operation parameters as specified in to meet the standards under 40 CFR
§ 264.1035(b)(4)(iii). 63.1203, 63.1204, and 63.1205.
§ 270.22 Specific part B information * * * * * (2) Facility owners or operators must
requirements for boilers and industrial comply with the Notification of Intent to
furnaces burning hazardous waste. ■ 7. Section 270.25 is amended by
Comply (NIC) requirements of 40 CFR
When an owner or operator of a revising paragraph (e)(3) to read as 63.1210(b) and 63.1212(a) before a
cement kiln, lightweight aggregate kiln, follows: permit modification can be requested
solid fuel boiler, liquid fuel boiler, or § 270.25 Specific part B information under this section for the purpose of
hydrochloric acid production furnace requirements for equipment. technology changes needed to meet the
becomes subject to RCRA permit * * * * * 40 CFR 63.1215, 63.1216, 63.1217,
requirements after October 12, 2005, or (e) * * * 63.1218, 63.1219, 63.1220, and 63.1221
when an owner or operator of an standards promulgated on October 12,
(3) A design analysis, specifications,
existing cement kiln, lightweight 2005.
drawings, schematics, and piping and
aggregate kiln, solid fuel boiler, liquid * * * * *
instrumentation diagrams based on the
fuel boiler, or hydrochloric acid (k) Waiver of RCRA permit conditions
appropriate sections of ‘‘APTI Course
production furnace demonstrates in support of transition to the part 63
415: Control of Gaseous Emissions’’
compliance with the air emission MACT standards. (1) You may request
(incorporated by reference as specified
standards and limitations in part 63, to have specific RCRA operating and
in § 270.6) or other engineering texts
subpart EEE, of this chapter (i.e., by emissions limits waived by submitting a
acceptable to the Regional
conducting a comprehensive Class 1 permit modification request
Administrator that present basic control
performance test and submitting a under Appendix I of this section,
device information. The design analysis
Notification of Compliance under section L(10). You must:
shall address the vent stream
§§ 63.1207(j) and 63.1210(d) of this (i) Identify the specific RCRA permit
characteristics and control device
chapter documenting compliance with operating and emissions limits which
operation parameters as specified in
all applicable requirements of part 63, you are requesting to waive;
§ 264.1035(b)(4)(iii).
subpart EEE, of this chapter), the (ii) Provide an explanation of why the
requirements of this section do not * * * * * changes are necessary in order to
apply. The requirements of this section ■ 8. Section 270.32 is amended by minimize or eliminate conflicts between
do apply, however, if the Director adding paragraph (b)(3) to read as the RCRA permit and MACT
determines certain provisions are follows: compliance; and
(iii) Discuss how the revised (2) To request this modification in (i) Submit your modification request
provisions will be sufficiently conjunction with MACT performance to the Director at the same time you
protective. testing where permit limits may only be submit your test plans to the
(iv) The Director shall approve or waived during actual test events and Administrator; and
deny the request within 30 days of pretesting, as defined under 40 CFR (ii) The Director may elect to approve
63.1207(h)(2)(i) and (ii), for an aggregate or deny the request continent upon
receipt of the request. The Director may,
time not to exceed 720 hours of approval of the test plans.
as his or her discretion, extend this 30
day deadline one time for up to 30 days operation (renewable at the discretion of * * * * *
by notifying the facility owner or the Administrator) you must: Appendix I to § 270.42—Classification
operator. of Permit Modification
Modifications Class
* * * * * * *
L. * * * ................................................................................................................................................................................................................ 11
10. Changes to RCRA permit provisions needed to support transition to 40 CFR part 63 (Subpart EEE—National Emission Standards for
Hazardous Air Pollutants From Hazardous Waste Combustors), provided the procedures of § 270.42(k) are followed..
* * * * * * *
1 Class 1 modifications requiring prior Agency approval.
* * * * * solid fuel boiler, liquid fuel boiler, or ■ b. Revising paragraphs (b)(1)

■ 10. Section 270.62 is amended by hydrochloric acid production furnace introductory text and (b)(2).
revising the introductory text to read as becomes subject to RCRA permit ■ c. Adding new paragraph (c).
follows: requirements after October 12, 2005 or
when an owner or operator of an The revisions read as follows:
§ 270.62 Hazardous waste incinerator existing cement kiln, lightweight * * * * *
permits.
aggregate kiln, solid fuel boiler, liquid
When an owner or operator of a fuel boiler, or hydrochloric acid § 270.235 Options for incinerators, cement
hazardous waste incineration unit kilns, lightweight aggregate kilns, solid fuel
production furnace demonstrates boilers, liquid fuel boilers and hydrochloric
becomes subject to RCRA permit compliance with the air emission
requirements after October 12, 2005, or acid production furnaces to minimize
standards and limitations in part 63, emissions from startup, shutdown, and
when an owner or operator of an subpart EEE, of this chapter (i.e., by malfunction events.
existing hazardous waste incineration conducting a comprehensive
unit demonstrates compliance with the performance test and submitting a (a) * * * (1) Revisions to permit
air emission standards and limitations Notification of Compliance under conditions after documenting
in part 63, subpart EEE, of this chapter §§ 63.1207(j) and 63.1210(d) of this compliance with MACT. The owner or
(i.e., by conducting a comprehensive chapter documenting compliance with operator of a RCRA-permitted
performance test and submitting a all applicable requirements of part 63, incinerator, cement kiln, lightweight
Notification of Compliance under subpart EEE, of this chapter), the aggregate kiln, solid fuel boiler, liquid
§§ 63.1207(j) and 63.1210(d) of this requirements of this section do not fuel boiler, or hydrochloric acid
chapter documenting compliance with apply. The requirements of this section production furnace may request that the
all applicable requirements of part 63, do apply, however, if the Director Director address permit conditions that
subpart EEE, of this chapter), the minimize emissions from startup,
determines certain provisions are
requirements of this section do not shutdown, and malfunction events
necessary to ensure compliance with
apply, except those provisions the under any of the following options
§§ 266.102(e)(1) and 266.102(e)(2)(iii) of
Director determines are necessary to when requesting removal of permit
this chapter if you elect to comply with
ensure compliance with §§ 264.345(a) conditions that are no longer applicable
§ 270.235(a)(1)(i) to minimize emissions
and 264.345(c) of this chapter if you according to §§ 264.340(b) and
of toxic compounds from startup,
elect to comply with § 270.235(a)(1)(i) to 266.100(b) of this chapter:
shutdown, and malfunction events; or if
minimize emissions of toxic compounds
you are an area source and elect to * * * * *
from startup, shutdown, and
comply with the §§ 266.105, 266.106, (2) Addressing permit conditions
malfunction events. Nevertheless, the
Director may apply the provisions of and 266.107 standards and associated upon permit reissuance. The owner or
this section, on a case-by-case basis, for requirements for particulate matter, operator of an incinerator, cement kiln,
purposes of information collection in hydrogen chloride and chlorine gas, and lightweight aggregate kiln, solid fuel
accordance with §§ 270.10(k), 270.10(l), non-mercury metals; or the Director boiler, liquid fuel boiler, or
270.32(b)(2), and 270.32(b)(3). determines certain provisions apply, on hydrochloric acid production furnace
a case-by-case basis, for purposes of that has conducted a comprehensive
* * * * * information collection in accordance
■ 11. Section 270.66 is amended by
performance test and submitted to the
with §§ 270.10(k), 270.10(l), Administrator a Notification of
revising the introductory text to read as 270.32(b)(2), and 270.32(b)(3).
follows: Compliance documenting compliance
* * * * * with the standards of part 63, subpart
§ 270.66 Permits for boilers and industrial ■ 12. Section 270.235 is amended by: EEE, of this chapter may request in the
furnaces burning hazardous waste. ■ a. Revising the section heading and application to reissue the permit for the
When an owner or operator of a paragraphs (a)(1) introductory text and combustion unit that the Director
cement kiln, lightweight aggregate kiln, (a)(2) introductory text. control emissions from startup,
shutdown, and malfunction events boiler, liquid fuel boiler, or minimized from startup, shutdown, and
under any of the following options: hydrochloric acid production furnace malfunction events, including releases
* * * * * that is operating under the interim from emergency safety vents, based on
(b) * * * (1) Interim status status standards of parts 265 or 266 of review of information including the
operations. In compliance with this chapter submits a RCRA permit source’s startup, shutdown, and
§§ 265.340 and 266.100(b), the owner or application, the owner or operator may malfunction plan and design. The
operator of an incinerator, cement kiln, request that the Director control director will specify that these permit
lightweight aggregate kiln, solid fuel emissions from startup, shutdown, and conditions apply only when the facility
boiler, liquid fuel boiler, or malfunction events under any of the is operating under its startup,
hydrochloric acid production furnace options provided by paragraphs (a)(2)(i), shutdown, and malfunction plan.
that is operating under the interim (a)(2)(ii), or (a)(2)(iii) of this section.
(c) New units. Hazardous waste PART 271—REQUIREMENTS FOR
status standards of part 265 or 266 of AUTHORIZATION OF STATE
incinerator, cement kiln, lightweight
this chapter may control emissions of HAZARDOUS WASTE PROGRAMS
aggregate kiln, solid fuel boiler, liquid
toxic compounds during startup,
fuel boiler, or hydrochloric acid
shutdown, and malfunction events ■ 1. The authority citation for part 271
production furnace units that become
under either of the following options continues to read as follows:
subject to RCRA permit requirements
after conducting a comprehensive
after October 12, 2005 must control Authority: 42 U.S.C. 6905, 6912(a), and
performance test and submitting to the 6926.
emissions of toxic compounds during
Administrator a Notification of
startup, shutdown, and malfunction ■ 2. Section 271.1(j) is amended by
Compliance documenting compliance
events under either of the following adding the following entries to Table 1
with the standards of part 63, subpart
options: in chronological order by date of
EEE, of this chapter.
(1) Comply with the requirements publication in the Federal Register, to
* * * * * specified in § 63.1206(c)(2) of this read as follows:
(2) Operations under a subsequent chapter; or
RCRA permit. When an owner or (2) Request to include in the RCRA § 271.1 Purpose and scope.
operator of an incinerator, cement kiln, permit, conditions that ensure * * * * *
lightweight aggregate kiln, solid fuel emissions of toxic compounds are (j) * * *
TABLE 1.—REGULATIONS IMPLEMENTING THE HAZARDOUS AND SOLID WASTE AMENDMENTS OF 1984
Promulgation date Title of Reglation Federal Register reference Effective date
* * * * * * *
Oct. 12, 2005 ................................ Standards for Hazardous Air Pol- [Insert FR page numbers] ............. Oct. 12, 2005.
lutants for Hazardous Waste
Combustors.
[FR Doc. 05–18824 Filed 10–11–05; 8:45 am]

BILLING CODE 6560–50–P

Rule: Air Pollutants, Hazardous National Emission Standards: Hazardous Waste Combustors

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rule: Air Pollutants, Hazardous National Emission Standards: Hazardous Waste Combustors

Uploaded by

Copyright:

Available Formats

Wednesday,

October 12, 2005

Any industry that combusts hazardous waste as defined in

EPA Environmental Protection Agency F. Testing Requirements IV. Use of Surrogates

TABLE 2.—SUMMARY OF EMISSION LIMITS FOR NEW OR RECONSTRUCTED SOURCES

A. Compliance Dates midway between comprehensive 3. You must install an automatic

Standard Proposed limit Final limit

The changes in the incinerator

Particulate Matter (gr/dscf) .................................................................................................................................................. 0.0007 0.0015

Hazardous Waste Burning Cement Kilns

Standard Proposed limit Final limit

The changes in the standards for new

Standard Proposed limit Final limit

C. Hazardous Waste Burning

Standard Proposed limit Final limit

The changes in the standards for new

Standard Proposed limit Final limit

Standard Proposed limit Final limit

Particulate matter ............................................... 0.0099 gr/dscf .................................................. 0.0098 gr/dscf.

D. Solid Fuel Boilers

Mercury (µg/dscm) ................................................................................................................................................................... 10 11

The changes in the solid fuel boiler

Mercury (µg/dscm) ................................................................................................................................................................... 10 11

Mercury (lb/MM Btu) ........................................................................................................................... 3.7E–6 ......... 19 µg/dscm 4.2E–5

The changes in the liquid fuel boiler

F. Hydrochloric Acid Production

Standard Proposed limit Final limit

The changes in the hydrochloric acid

Standard Proposed limit Final limit

achieve a standard whenever it is tested (as the

standards consequently are not work nature of carbon monoxide as an

of whom would violate a 99 + percent standard based on their own operating

LaFarge (OH) ................................................................................................... 3 99.1 99.4 99.3

firing nozzle, you must determine the mass-

combust/toolkit/factshts.htm to learn § 63.10(f). www.epa.gov/ttn/emc/staffdir.html.

been promulgated for mercury CEMS, there has not

40 CFR citation OMB control No.

* * * * * Superintendent of Documents, U.S. (2) [Reserved]

TABLE 1 TO § 63.1200.—HAZARDOUS WASTE COMBUSTORS EXEMPT FROM SUBPART EEE

* * * * * definitions for ‘‘Btu’’, ‘‘Hazardous waste and ‘‘System removal efficiency’’ in

Reference Notification, request, petition, or application 6

Reference Notification, request, petition, or application 6

Reference Document, Data, or Information

Reference Document, Data, or Information

(c) * * * (4) Hold an informal public meeting § 63.1214 Implementation and

and corrected to 7 percent oxygen. If

emissions, corrected to 7 percent hexachlorodibenzo-p-dioxins and manganese, and nickel in excess of 23

* * * * * solid fuel boiler, liquid fuel boiler, or ■ b. Revising paragraphs (b)(1)

[FR Doc. 05–18824 Filed 10–11–05; 8:45 am]

You might also like