Article

pubs.acs.org/IECR

Alternative Approaches for pXylene Production from Starch:

Techno-Economic Analysis
Zhaojia Lin, Vladimiros Nikolakis, and Marianthi Ierapetritou*,

Department of Chemical and Biochemical Engineering, Rutgers - The State University of New Jersey, 98 Brett Road, Piscataway,
New Jersey 08854-8058, United States

Catalysis Center for Energy Innovation & Department of Chemical & Biomolecular Engineering, University of Delaware, Newark,
Delaware 19716, United States
*
S Supporting Information

ABSTRACT: Two alternatives for the production of p-xylene from starch are investigated using detailed techno-economic
analysis. The rst route is based on the patents by GEVO Inc. and uses fermentation followed by the catalytic conversion,
whereas the second one utilizes CCEI-discoveries - solely the catalytic conversion. The minimum p-xylene price for the GEVO
and CCEI processes is found as $4121 and $3637/metric ton, respectively, without accounting for the byproduct values, while
taking into consideration that the byproduct value leads to $3481 and $2885/metric ton. Using sensitivity analysis, the biomass
feedstock cost is determined as a major contributor of the total cost. Other signicant contributors include biocatalyst cost in the
GEVO process and ethylene and solvent (THF) costs in the CCEI process. The impacts of selectivity and conversion on the p-
xylene price for the CCEI process are investigated. High selectivity is favorable compared to high conversion due to recycling
used to improve the performance in the latter case.

INTRODUCTION
Exhausting oil reservoirs, increasing petroleum prices, and
rising environmental concerns have prompted the development
of renewable energy sources including solar, wind, and biomass.
Biomass is among the more promising alternatives since it can
be used to produce both high-volume, low-value fuels, and
high-value chemicals that are currently made from petroleum
resources.1 The U.S. Department of Energy has projected that
biobased fuels and chemicals in the year 2030 will contribute to
20% of US transportation fuel and 25% of the production of US
commodities, respectively, in comparison to 0.5% and 5% in
2001.2 Biobased p-xylene has drawn considerable attention
because it is the principal precursor to polyester polyethylene
terephthalate (PET), a polymer resin broadly used in the
synthesis of bers, lms, and beverage containers.3 Several
companies, such as Coca Cola, Pepsi, and Procter & Gamble,
have launched their projects toward the utilization of biobased
PET.46
A couple of companies (Gevo Inc. and Virent Energy system
Inc.) have announced the pilot scale production of biobased p-
xylene.7 In addition to these two technologies, p-xylene
production based on the discoveries at Catalysis Center of
Energy Innovation (CCEI) is another promising alternative. In
this paper, the work focuses on the comparison of the biobased
p-xylene production routes based on the patents by Gevo Inc.8,9
and discoveries from CCEI1012 using techno-economic
analysis. Unfortunately there are no sucient details in the
literature that will allow the evaluation of the Virent process at
this stage.
The GEVO production route starts with the fermentation of
biomass to form isobutanol. Biobased isobutanol draws a lot of
attention because it can be used as a high-energy-density
biofuel as well as a precursor for the production of commodity
chemicals such as p-xylene.1315 The production of isobutanol
however using fermentation of biomass can only be
accomplished economically by metabolically engineered yeast
or Escherichia coli microorganisms (biocatalyst). No native
microorganism has been identied to produce isobutanol in
large amounts although a small amount of isobutanol has been
produced as a microbial byproduct.13,14 The next step in the
GEVO process is the conversion of isobutanol to p-xylene
following dehydration, dimerization, and dehydrocyclization.8
p-Xylene derived from biobased 5-hydroxymethylfurfural
(HMF) is another alternative. In our previous work,16 the
production of p-xylene from HMF was studied, and the
economics of the process were evaluated. This paper focuses on
the integration of the conversion of biomass to HMF and the
evaluation of the overall production route of p-xylene from
biomass. The conversion of starch to HMF involves rst the
depolymerization of starch into glucose followed by the
isomerization of glucose to fructose and the conversion of
fructose to HMF using a biphasic reaction.17 Dierent solvents
used in the production of HMF from glucose or starch have
been studied. Tetrahydrofuran (THF) is a good extractant for
the removal of HMF from aqueous phase, resulting in the
highest selectivity of HMF.18 Nikolla et al. reported that in the
production of HMF from starch using THF solvent via Sn-Beta
catalyst, 75% conversion of starch, and 69% selectivity to HMF
can be achieved.11 These values are utilized in this work.
Leshkov et al. have reported the possibility of using the same
solvents in the production of HMF from fructose and then 2,5-
Received: July 30, 2013
Revised: September 20, 2013
Accepted: June 11, 2014
Published: June 11, 2014

2014 American Chemical Society 10688 dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
Figure 1. Reaction paths of the (a) GEVO and (b) CCEI processes.
dimethylfuran (DMF) from HMF,12 which is also applied in
this work.
Biobased chemical production routes are still at the early
stage of development, and there is limited knowledge regarding
the optimum reaction conditions, the identication and
utilization of byproducts, and the best purication steps.
10689
Article
Techno-economic analysis is a useful tool to facilitate the
development of the processes, to assess and compare the
feasibility of dierent alternatives, and to determine the most
important parameters that have to be improved to promote
process commercialization. There are just a few published
studies focusing on the economic analysis of the proposed
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
Figure 2. Process diagram of the GEVO process.8,20
alternatives and the evaluation of the productions routes using
detailed owsheets. Among them Kazi et al.19 performed an
economic analysis of HMF production from fructose using n-
butanol and determined the minimum cost of HMF to be
$1070/metric ton (based on 2007 prices for raw materials).
Torres et al.7 studied the same process using 7:3 MIBK:2-
butanol as a solvent that resulted in a minimum cost of HMF of
$1189.44/metric ton. The main dierence between the two
studies is the cost of fructose, which is a signicant factor of
HMF cost. In our previous work, the production of p-xylene
was studied utilizing n-hexanol for the conversion of HMF to
DMF and n-heptane for the production of p-xylene from DMF.
The minimum cost of p-xylene was found to be $3962/metric
ton.16
The aim of this work is to estimate the minimum cost of
biobased p-xylene starting from biomass and to compare the
economics of alternative production routes. Starch is
10690
Article
considered as the biomass feedstock for this study since data
for both processes are readily available in the literature.
Sensitivity analysis is performed to assess the impact of
economic parameters such as the costs of raw materials,
solvents, and catalysts. The rest of the manuscript is structured
as follows. In the next section, the two productions routes (i.e.,
GEVO and CCEI) are introduced followed by a discussion of
the assumptions required to perform the process simulation
and economic analysis. The owsheets of these two production
routes are then described in detail. In the last section of the
manuscript are the economic results for both alternatives, and
the outcomes of the sensitivity analysis are discussed and
analyzed.
METHODOLOGY
Reaction Path. The two reaction routes explored in this
work are shown in Figure 1. In GEVOs route, starch is
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
Figure 3. Process diagram of CCEI production.
depolymerized to glucose using -amylase and glucoamylase
(saccharication), followed by the conversion to isobutanol
using metabolically engineered microorganisms (fermentation)
such as post- or pre-WGD (whole genome duplication) yeast
recombinant microorganisms.20 During the saccharication and
fermentation processes dierent byproducts are produced such
as acetate salt and acetic acid.21 Isobutanol is then dehydrated
to isobutene using a -alumina catalyst at 563 K and 5.15 MPa.
The byproducts of this reaction include n-butene. The
isobutene is then oligomerized to isooctene using HZSM-5
catalyst at 343 K and 5.27 MPa. The byproducts involve trimers
of isobutene. Isooctene is then dehydrocyclizated to produce p-
xylene with 90% selectivity at 823 K and 0.14 MPa. Side
10691
Article
reactions produce o-/m-xylene, isooctane, isobutane, benzene,
toluene, C9+ aromatics, isoolens, and methane.9
In the CCEI process starch is converted to HMF in a
biphasic reactor. This process is integrated with the conversion
of HMF to DMF and its further transformation to p-xylene.16
In particular, starch is depolymerized into glucose in the
presence of HCl at pH equal to 1. Glucose is then isomerized
to fructose and dehydrated to HMF using Sn-Beta as the
isomerization catalyst and HCl as the dehydration catalyst.
These reactions proceed in a biphasic reactor in which the
aqueous phase is saturated with NaCl salt and THF is used as
the organic solvent/extractant. The volumetric ratio of THF to
water is 3:1.11 The byproducts herein contain formic acid,
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research Article

levulinic acid, and humins. The production of DMF from HMF byproducts from the fermentation include acetic acid and solid
proceeds in the vapor phase using copperrutheniumcarbon waste.21 At reaction conditions of (3), the only byproduct
(CuRu/C) catalyst at 493 K and 0.69 MPa.12 At this reaction formed is n-butene, while the trimmers of isobutene are
step, the main byproduct is the 2,5-dihydroxymethyltetrahy- considered as byproducts of the reaction (4). The side reactions
drofuran.12 The byproduct formic acid that comes from HMF during the production of p-xylene include o-/m-xylene,
production was reported to decompose to H2 and CO2 at the isooctane, isobutane, n-butane, benzene, toluene, C9+ aromatics,
reaction conditions.22 At the next step, DMF reacts with isoolens, and methane.9
ethylene and forms p-xylene at 573 K and 5.7 MPa using a 2. The byproducts such as isobutane, n-butane, and isooctane
zeolite catalyst in the presence of n-heptane solvent. The are treated as diluents and are recycled to enhance the
production follows a DielsAlder cycloaddition and dehy- conversion of both oligomerization and dehydrocyclization
dration reaction. The competitive side reactions include the reactions.9
hydrolysis of DMF to 2,5-hexanedione and oligomerization or 3. The conversion and selectivity of the CCEI process
polymerization and alkylation of p-xylene.10 reactions are listed in Table 1 (reactions (6)(10)). The
Simulation and Economics. Simulation. The process conversions and selectivity for steps (6)(8) have been
diagrams of the Gevo and CCEI processes are shown in Figures adjusted in order to satisfy the experimentally observed values
2 and 3, respectively, and described in detail later. The for the overall one-pot conversion of starch to HMF (69%
simulation is performed using Aspen Plus V7.3.2. The NRTL HMF selectivity at 75% conversion of starch11). The by-
method is utilized to predict the liquidliquid and liquid products of reactions (6)(8) involve formic acid, levulinic
vapor behavior. In particular, the ELECNRTL method, which is acid, and humins (which also includes all unidentied
the most versatile electrolyte property method that can handle products).11 The reactor pressure is determined by Aspen
an aqueous and mixed solvent systems method, is selected for Plus simulation to satisfy the existence of a liquid-phase.
the units involved salts.17 The PengRobinson method is 4. The reaction parameters of the DMF and p-xylene
exploited for the decanter that is used to separate DMF and production are the same as the ones used in our previous
water because it is found to produce more reliable results study.16 Both the intermediate (2-methyl-5-hydroxymethylfur-
compatible with the solubility properties of DMF, which is an) and the byproduct (2,5-dihydroxymethyltetrahydrofuran)
known to be slightly soluble in water. Most of the components coexist at the reaction (9).24 Since the detailed reaction
involved in the reactions are directly selected from the Aspen mechanism is not known, the eect of solvent on conversion
database, whereas some not included in the database (i.e., and selectivity has not been explicitly considered in this study.
HMF, 2-methyl-5-hydroxymethylfuran) are dened by the In other words, the same conversion and selectivity is assumed
structures. Also there are some components with no specic with using THF as solvent as that with n-hexanol.
molecular structure or complex structure such as humins and 5. The conversion and selectivity of reaction (10) also
trimmers of isobutene for which surrogates are then used. Due include the recent ndings that using H-BEA zeolite catalyst
to the unknown structure of humins, they are modeled as the and n-heptane mixed with a small amount of n-tridecane as
polymer of acetylacetone. Trimmers of isobutene in the GEVO solvent, the conversion and the selectivity for p-xylene
process may contain dierent types of compounds, but they are production from DMF can be increased to 100% and 90%,
represented by 1-dodecene in this work. All the missing respectively.23
parameters are estimated by the molecular structures using the Additionally, it is necessary to make some assumptions to
Unifac Model and thermo data engine (TDE). TDE is a implement the simulation:
thermodynamic data correlation, evaluation, and prediction tool 1. It is assumed that the selectivity to levulinic acid, to formic
developed by the collaboration of Aspen plus and the National acid, and to humins is 7.5%, 18.17%, and 5.33%, respectively.7
Institute of Standard and Technology. 2. At this concentration level, the byproducts (formic and
To enable the process simulation, reaction details and levulinic acid) do not aect the conversion and selectivity of
parameters are selected from the existing literature and reaction (9). The validity of this assumption was veried with a
described below: limited number of experiments carried out at CCEI. Thus, no
1. The conversion and selectivity of the main reactions in the separation steps are considered before reaction (10). It is
GEVO route are listed in Table 1 (i.e., reactions (1)(5)). The possible, however, that this assumption is not valid if the
byproduct concentration is signicantly dierent than what is
Table 1. Conversion and Selectivity of Main Reactions assumed here. In that case additional separation/purication
(GEVO: (1)(5); CCEI: (6)(10)) steps might be required. Based on the literature, formic acid
reaction conversion (%) selectivity (%) decomposes to H2 and CO2 at the conditions of reaction (9).
(1) starch glucose 95 100
To account for this, 75% conversion of formic acid is
(2) glucose isobutanol 99.5 91.46
assumed.22
(3) isobutanol isobutene 98 96.9
3. In a biphasic reactor for the conversion of glucose to
(4) isobutene isooctene 77 97.4 HMF, 93% of HMF and 30% HCl are assumed to exist in the
(5) isooctene p-xylene 47 90 organic phase.25 It is also assumed that the partition coecients
(6) starch glucose 75 100 of formic acid and levulinic acid are the same as HMF.
(7) glucose fructose 100 100 Economics. Aspen Economic Analyzer V7.3.2 is used to
(8) fructose HMF 100 69 perform an economic assessment of the p-xylene production.
(9) HMF DMF 100 78 The discounted cash ow analysis is utilized to perform the
(10) DMF p-xylene 95 (100)a 75.5 (90)a economic evaluation of the dierent processes. The production
cost of p-xylene is used to determine the minimum p-xylene
a
Corresponds to recent conversion and selectivity parameters.23 cost, which is dened as the selling price of the product when
10692 dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
the net present value (NPV) is zero.26 Some additional
assumptions are necessary to perform the economic analysis:
1. All the equipment and operating costs estimated by Aspen
Economic Analyzer V7.3.2 are based on the price of the rst
quarter in 2011.19
2. The capacity of base scenario is assumed 600000 t/year
(75000 kg/h) of dried starch.
3. The plant operates in a continuous mode for 8000 h per
year. The economic life of the project is assumed to be 20 years.
The internal rate of return (ROR) on investment and the rate
of escalation are assumed as 15% and 3.5%, respectively. 35%
tax is applied to the prots. The simplest depreciation method
-- the straight-line method is applied as the salvage value is 10%
of the original capital cost after 20 years.
4. Corn starch is considered as biomass feedstock. The cost
of corn starch is averaged to 13.29 cents per lb ($293/metric
ton) based on the data of 2012 from USDA.27
5. The market price of THF is estimated as $1500/metric ton
based on the prices reported at www.alibaba.com28 - a
worldwide e-commerce market site originally from China.
The price of chemicals may vary with many factors, such as
season, location, producer, policy, etc. To address this
introduced uncertainty, sensitivity analysis is carried and
presented in the section of sensitivity analysis.
6. The enzyme price is assumed to be $821/metric ton, and
it is estimated that 12000 t/year will be required.29 The
microorganisms used for the fermentation of isobutanol are
assumed at higher cost ($1642/metric ton) since they are
metabolically engineered. The metabolically engineered micro-
organism to starch ratio is assumed to be 10:1, which results in
60000 t/year.9
7. The catalyst costs are taken as the average of the prices
reported in www.alibaba.com.28 The catalysts can be regen-
erated;30 however, complete replacement (except CuRu/C
catalyst) is assumed in this work. The price of -alumina that is
used in the isobutanol dehydration reaction is $1400/metric
ton. The life cycle of alumina catalyst depends on the nature of
the feed, the temperature, and other operating conditions,
which aect the life cycle of alumina catalyst that can vary from
months to years.31 For the base scenario calculation, a median
life cycle of six months is used. A weighted hourly space
velocity (WHSV) of 10 h1 is assumed for the dehydration
reaction which represents the mass ow rate over the catalyst.9
The price of the H-ZSM-5 catalyst used in the isobutene
oligomerization is related to the zeolite topology and the
detailed composition of the catalyst. The median unit price of
ZSM is used that is $5000/metric ton.28 WHSV in the
oligomerization reaction is 100 h1.9 The deactivation rate of
HZM-5 is sensitive to the byproducts and operating conditions.
In this study a one-month catalyst cycle time is applied.32 The
cost of chromium oxide doped alumina catalyst used in the
dehydrocyclization reaction is $2250/metric ton,28 and WHSV
is 1 h1.9 The dehydrocyclization catalyst deactivates rapidly at
the beginning and then remains at a roughly stable rate until 80
h;33 therefore, an 80-h cycle time is applied for the
dehydrocyclization catalyst.
8. The unit cost of a Sn-Beta zeolite catalyst is assumed as
$10000/metric ton. The molar ratio of Sn-Beta assumed for the
reaction is 1:200 of metal:glucose.11
9. The cost of hydrochloric acid is $115/metric ton34 and of
NaCl $190/metric ton.28
10. Other costs include raw materials, solvents, and other
catalysts: the ethylene price is $1202/metric ton and hydrogen
10693
Article
is priced at $631/metric ton.16 The market price of n-heptane is
estimated as $1500/metric ton. The CuRu/C catalyst cost is
$279.66 million as an initial investment plus an additional cost
of $1.05 million/year.19 The consumption of HZM-5 zeolite is
estimated at 1406 metric ton/year with a xed cost at $5000/
metric ton.
11. In the base scenario the value of byproducts is not
considered, even though the impact of considering their value is
discussed.
12. Although wastewater treatment is important in such
processes especially in the fermentation process, the design of
wastewater treatment and the associated cost is not included in
this work.
Even though the assumptions are necessary to perform the
simulation and study the economics, they introduce un-
certainties regarding the results obtained. Therefore, sensitivity
analysis is performed to assess the impacts of uncertainties of
economic parameters such as the costs of raw materials,
solvents, and catalysts and to determine the eects of reaction
parameters such as selectivity and conversion.
PROCESS DIAGRAM AND DESCRIPTION
GEVO Process. The GEVO process development is based
on the patented process of isobutanol fermentation and p-
xylene production from isobutanol.8,20 The detailed process
diagram is shown in Figure 2. (The following description refers
to the stream numbers shown in the gure.) Starch, steam, and
recycled water (Stream 15) containing unreacted components
are fed to a pretreatment reactor (R1) to depolymerize to
glucose by -amylace enzyme. Reactor (R1) operates at 99 C.
The outlet stream 1 is cooled to about 32 C and fed to a
fermentor (R2) with recycled slurries (Stream 6) from the
bottom of the vacuum ash (FL1) containing cells, water,
nutrients, carbohydrates, and about 2 wt % unvaporized
isobutanol. Saccharication and fermentation occur simulta-
neously to produce isobutanol using glucoamylase enzyme and
metabolically engineered yeast at about 32 C and atmospheric
pressure. Stream 2 containing about 4 wt % butanol from
reactor (R2) enters a ash (FL1) that operates at 34 C and
vacuum (about 0.0067 MPa) to form an azeotropic vapor of
isobutanol and water (Stream 5). Stream 5 is then pumped by
the vacuum pump (VP1) to a condenser (CD1) to remove
CO2 (Stream 9). The condensed liquid phase (Stream 10) is
separated into a butanol rich phase (Stream 12) and a water
rich phase (Stream 11) using a decanter (D1). The water rich
heavy phase (Stream 11), as well as Stream 18 from decanter
(D2) and Stream 24 from decanter (D3), are fed to a beer still
column (BC1). A butanol-water azeotropic Stream 13 is
generated at the top of the beer still and fed to another
condenser (CD2) to collect CO2 (Stream 16), followed by
another decanter (D2) to separate into a water rich heavy phase
(Stream 18) and butanol-rich light phase (Stream 19). The
butanol-rich streams (12, 19, and 25) from three decanters
(D1, D2, and D3) connect to a distillation column (DC1) to
remove water. An isobutanol-water azeotrope (Stream 20) from
the top is fed into a third condenser and decanter (CD3 and
D3) to remove CO2 (Stream 22) and produce a water-rich
phase (Stream 24) and a butanol-rich phase (Stream 25).
Stream 21 then enters another distillation column (DC2) to
remove acetic acid and produce puried isobutanol (Stream
26).20
Isobutanol (Stream 26) is fed to a xed-bed catalytic reactor
(R3) and is dehydrated to isobutene (Stream 27) using a -
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
alumina catalyst at 563 K and 5.15 MPa. A decanter (D4) is
used to remove most of the water of stream 27. Isobutene
(Stream 29), as well as recycling stream (37) that contains
unreacted isobutene and diluents such as isobutane and n-
butane, enters the oligomerization reactor (R4) to produce
isooctene (Stream 30) using the HZSM-5 catalyst at 343 K and
5.27 MPa. The byproducts involve trimers of isobutene. Stream
(30) passes through a valve to reduce the pressure and forms
Stream (31). Stream (31) combined with the recycle stream
(18) containing unreacted isooctene, and diluents such as
isooctane, enters the dehydrocyclization reactor (R5) to
produce p-xylene at 823 K and 0.14 MPa. Side reactions
produce o-/m-xylene, isooctane, isobutane, benzene, toluene,
C9+ aromatics, isoolens, and methane. A ash (FL2) is used to
remove the light components, which is connected to another
ash (FL3) to remove hydrogen and methane (Stream 36) and
to recycle other light components (Stream 37). Stream (34)
from the bottom of (FL2) is pumped to a distillation column
(DC3). The light byproducts such as benzene and toluene are
collected from the top (Stream LPROD), whereas the bottom
stream (38) is fed into another distillation column (DC4) to
remove the heavy byproducts that involve C9+ aromatics and
trimmers (Stream HPROD). A distillation column (DC5) is
used to recycle unreacted isooctene and diluent isooctane
(Stream 40). Stream (41) consisting mainly of p-xylene and o-
xylene is separated via a distillation column (DC6) to achieve a
99.5 wt % pure p-xylene.8
CCEI Process. The p-xylene production using the CCEI-
process is based on the laboratory-scale biphasic reaction for
the conversion of starch to HMF in water using THF as the
organic solvent,11 integrated with the HMF to DMF and DMF
to p-xylene routes that have been discussed in our previous
work.16 The detailed process diagram of the process is shown in
Figure 3. (The following description refers to the stream
numbers shown in the gure.)
Starch (Stream STARCH) and the recycle stream (4),
containing mainly water and unreacted components, are fed
into reactor (R1) to depolymerize into glucose (Stream 1)
using HCl at 1.9 MPa and 453 K. Stream (1) mixed with
solvent THF (THF, recycling stream 7 and 13) and NaCl
enters a biphasic reactor (R2) to convert into HMF by using
zeolite Sn-Beta as isomerization catalyst and HCl as
dehydration catalyst at 453 K and 1.9 MPa. The volume ratio
of organic and aqueous phases is 3:1.11 The aqueous phase
(Stream 3), containing mainly water, other components such as
starch, glucose, fructose, HMF, levulinic acid, formic acid, NaCl
(which based on the partition coecients reported is primarily
in the aqueous phase) and the solid byproducts humins, enters
a lter (FI) to remove solids and recycle the other components
(Stream 4) back to reactor (R1). The organic rich phase
(Stream 2), which is composed of THF, water, HMF, formic
acid, and levulinic acid, is fed into a ash (FL1) to recycle the
excess THF back to the reactor and to form a 10 wt % HMF in
THF solution (Stream 6), which also contains some amount of
formic acid, and levulinic acid. The top stream (5) of ash
(FL1) enters a distillation column (DC1) to separate formic
acid, levulinic acid, and HMF and to collect THF from the top.
Stream 6 is evaporated and then enters a uidized bed
reactor (R3) together with the recycled intermediate 2-methyl-
5-hydroxymethylfuran and the unreacted HMF (Stream 17) at
493 K and 0.69 MPa. Compressed hydrogen is also fed to the
reactor (R3) to form DMF (Stream 10). The reaction proceeds
in the vapor phase using the CuRu/C catalyst. The excess
10694
Article
hydrogen (stream 11) is collected from the top of the ash
(FL2) and is recycled back to R3. The bottom stream (Stream
12) of FL2 is connected to a distillation column (DC2) to
separate and recycle THF and water (Stream 13). The rest of
the components from the bottom (Stream 14) enter a second
distillation column (DC3) used to purify DMF (Stream 15)
from the high boiling point intermediates and byproducts
(Stream 16). Stream 16 is separated into the 2-methyl-5-
hydroxymethylfuran (Stream 17) and byproducts such as 2,5-
dihydroxymethyltetrahydrofuran and levulinic acid (Stream
BYRPOD1) via another distillation column (DC4).
Puried DMF (Stream 15) is mixed with n-heptane, recycle
solvent, and unreacted components (Stream 24) and pumped
to a CSTR (R4). Ethylene feed is compressed to the reactor R4
with recycle ethylene (Stream 21). p-Xylene (Stream 20) is
produced at 573 K and 5.7 MPa in R4.10 The main byproducts
consist of 1-ethyl-2,5-dimethylbenzene, 2,5-hexanedione, and
oligomers products, all of which have higher boiling points than
p-xylene. A ash (FL3) is then used to separate the low boiling
point ethylene (Stream 21) from the production stream of R4
(Stream 20), followed by a decanter (D) where most of the
water is removed. The organic stream (23) from the decanter
(D) is fed to another distillation column (DC5) to split the
unreacted raw materials and solvents (Stream 24). Heavy
stream (25) is puried using a distillation column (DC6)
resulting in 99.5 wt % pure p-xylene.
For the case of the reaction (10)* shown in Table 1, the
owsheet is modied by removing the ash drum (FL3).
Stream 20 containing solvents, water, unreacted ethylene, p-
xylene, and byproducts (1-ethyl-2,5-dimethylbenzene, 2,5-
hexanedione, and oligomers products) is fed into a decanter
(D) to remove the water. Then the organic stream 23 from D is
fed to a distillation column (DC5). DC5 has three outlets; the
top and bottom (streams 24 and 25, respectively) of the
column are recycled to the reactor. The top stream contains
unreacted ethylene and n-heptane (Stream 24) while the
bottom is primarily n-tridecane (Stream 25). The side stream
from DC5 containing p-xylene and high-boiling point by-
products is fed to a distillation column (DC6) and essentially
obtains over 99.5 wt % pure p-xylene.
RESULTS AND DISCUSSION
Simulation. The capacity of both GEVO and CCEI
processes is 600000 metric ton of starch per year. The
GEVO process produces 103956 metric ton/year of p-xylene,
while the CCEI process requires other raw materials such as
17600 metric ton/year of hydrogen and 71589 metric ton/year
of ethylene to achieve 196345 metric ton/year of p-xylene. The
detailed composition and ow rates of the raw material streams
and the main product and byproduct streams obtained using
simulation for the GEVO and CCEI processes are provided as
the Supporting Information in Table S1 and S2, correspond-
ingly.
The GEVO process produces several byproducts including
solid waste (stream SW) and other commodity chemicals. Solid
waste, which comprises of unreacted starch, enzyme nutrients,
and solid waste from fermentation such as acetate salts, etc., can
either be sold as animal feed or burnt as fuel. Other commodity
chemicals which are common in the petroleum reneries such
as methane, H2, olens, etc. are also produced. Thus, this
process has a high potential if other options are exploited in the
production moving toward an integrated biorenery. In this
work, the target is to produce biobased p-xylene; therefore, no
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research Article

further purication of other byproducts were considered. In Table 2. Summary of the Capital and Operating Cost of the
addition, as mentioned in our previous study,16 many of the GEVO Process
components for the GEVO process have very similar boiling
cost cost
points, resulting in loss of intermediate materials and products item (million $) item (million $/yr)
during the separation process. purchased equipment 170.7 raw materials 180.0
In addition, in the GEVO process benzene, toluene, and (fermentation)
other xylene isomers are also formed. These chemicals are also purchased equipment 10.4 enzyme/biocatalyst 108.4
made in the existing petroleum-based p-xylene production via (isobutanol to costs
p-xylene)
catalytic reforming of naphtha.8,35 Such byproducts have close
equipment setting 0.37 catalysts costs 13.2
boiling points, making the purication of p-xylene a dicult
piping 6.7 operating labor and 17.7
process. Based on the normal boiling and melting points of p-, maintenance cost
m-, and o-xylene (listed in Table S3), it is dicult to separate civil 1.3 utilities 43.2
xylene isomer mixtures especially to separate p-xylene from m- steel 1.2 plant overhead 8.9
xylene. Considering that the existing byproduct is mainly o- instrumentation 3.5 general and 17.9
xylene, we assume no existence of m-xylene. Therefore, one administrative
cost
distillation column can separate p-xylene and o-xylene, though a
electrical 1.2 total operating cost 389.3
large one with high reux ratio is used to achieve 99.5 wt % of
insulation 3.8
p-xylene of nal product. However, if m-xylene is necessarily to paint 0.28
be included, then other technologies are needed to be other 34.9
considered for xylene isomer purication. Adsorption using general and 6.9
the UOP Parex process and the Toray Aromax process is two administrative
common ones,36 while crystallization is another potential overheads
alternative for p-xylene recovery. In the latter case, high capital contract fee 5.6
and utility cost are expected due to the really low operating contingencies 44.6
temperature and low recovery of p-xylene (6070%).37 total project capital cost 291.6
However, the integration with adsorption increases the
concentration of p-xylene in the feed stream of crystallization Table 3. Summary of the Capital and Operating Cost of the
and reduces both utility and capital cost.37,38 Both crystal- CCEI Process
lization and adsorption require the use of multiple units. cost cost
In comparison to the GEVO process, one advantage of the item (million $) item (million $/yr)
CCEI route is that only p-xylene is produced. Therefore, the catalyst cost 279.7 catalysts 165.7
CCEI process requires fewer and simpler separation steps purchased equipment 36.2 starch 197.2
thereby reducing the related cost. However, most of the CCEI equipment setting 0.51 H2 12.2
process byproducts such as 2,5-dihydroxymethyltetrahydrofur- piping 6.2 ethylene 86.0
an, 2,5-hexanedione, oligomers, and 1-ethyl-2,5-dimethylben- civil 1.8 solvents 95.0
zene have not been used for other purposes, so more studies steel 0.53 operating labor and 3.9
maintenance
are required to improve the selectivity of p-xylene or to identify
instrumentation 2.8 utilities 94.0
commercial value for these products.
electrical 1.7 plant overhead 1.9
The main drawback of the CCEI process is the need for a
insulation 2.5 general and 32.2
large amount of solvent. A 10% starch aqueous solution is used administrative
initially based on the existing literature data. Higher overheads
concentrations of starch can be achieved, but since there are paint 0.18 total operating cost 688.1
no data regarding the impact of initial biomass concentration other 17.8
on the conversion and selectivity, 10% is used in this work. A general and 1.9
administrative
3:1 volumetric ratio of organic to aqueous phase is used in the overheads
experiment11 to achieve better phase partition and high contract fee 2.1
selectivity of HMF, which results in a large amount of solvent. contingencies 13.4
However, the large amount of solvent causes diculties in the total project capital 367.2
separation process and results in increasing utility costs. cost
Economics. The detailed capital and operating cost of both
GEVO and CCEI processes are listed in Tables 2 and 3, component is for equipment purchasing, accounting for 10.2%
correspondingly. The fraction of the capital and operating cost of the total capital cost.
of the GEVO and CCEI processes are also shown in Figures S1 In terms of the operating cost, the three main contributors
and S2, respectively. The capital cost of the GEVO process is for the GEVO process are raw materials 46.2%, enzyme
estimated at $291.6 million while that of the CCEI process is (including biocatalyst) 27.8%, and utilities 11.1%. In compar-
$367.2 million. The main component of the capital cost of the ison, the operating cost of the CCEI process breaks down into
GEVO process comes from the equipment purchase required 43.0% from raw materials, 24.1% from catalysts, 13.8% from
for the fermentation process, which accounts for 58.6% of the solvents and 13.7% from utilities. In particular, the fraction of
capital cost. The cost of the remaining equipment purchase raw materials cost can be further decomposed to 28.7% from
corresponds to only 3.6%. The main component of the CCEI starch, 12.5% from ethylene, and 1.8% from H2. The catalyst
capital cost is due to the CuRu/C catalyst cost which is costs herein contain the annual replacement or regeneration of
around 79.0% of the total capital cost. The second largest all kinds of catalysts. Among them, the dominant cost comes
10695 dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research Article

from the replacement of the Sn-Beta catalyst, accounting for

78.5% of the catalyst costs. Notably, the catalyst cost for the
CCEI process is an expensive part for both capital and
operating cost because of the high price and easy deactivation.
Identifying new cheap catalysts with a long life cycle or
operating conditions that prolong the lifetime of the currently
used catalysts will be useful to lower the initial costs and further
reduce the annual operating costs. Around 90% of the annual
solvent cost is due to the large amounts of THF used. No heat
integration is included in this work. If heat integration is
considered, it is expected to result in higher reduction of utility
cost of the GEVO process compared to the CCEI process due
to large dierences of temperatures used in the GEVO process.
In the base case scenario, where no byproducts value is
considered, the minimum price of p-xylene from the GEVO
and CCEI processes, which is dened as the price when the
NPV equal to zero, is $4121/metric ton and $3637/metric ton,
respectively.
If the byproduct value is considered, then the economics are
substantially improved. Solid waste from the fermentation is
considered as animal feed at $40/metric ton.21 Humins from
starch conversion to HMF is considered as waste of no value.
No further purication of byproducts is included in the
owsheet diagrams studied, and the value of the byproducts
streams are considered based on the price of these chemicals
produced from petroleum. For example, the value of acetic acid
is $607/metric ton and that of o-xylene is $1405/metric ton.34
For the unpuried byproducts or those byproducts with
unknown usage at the current stage, they are accounted as
bioethanol fuel, at the price of $2.2/gal ($736/metric ton).39
Stream 36 containing mainly hydrogen and methane is
considered to have the same value as natural gas which is
around $124/thousand cubic meters40 (i.e., $155/metric ton
assuming the density of natural gas is 0.8 kg/m3). Considering Figure 4. Variation of PX minimum price of the (a) GEVO and (b)
the byproduct value results in lower cost for p-xylene, which is CCEI processes.
calculated to be $3481/metric ton, and $2885/metric ton for
the GEVO and CCEI processes (for reaction (10)), Table 4. Price of Biobased Ethylene from Dierent
respectively. Further investigation of separation steps to purify Alternatives42
selected byproducts will further improve the economics of the
processes. In particular, isooctane can be used as an oil additive; source price ($/metric ton)
benzene and toluene are important intermediates at a cost of sugarcane 1200
$1400/metric ton;34 and levulinic acid is precursor of biofuel sweet sorghum 1700
and pharmaceuticals with a price of $6000 to 9000/metric corn 2000
ton.41 sugar beet 2600
Sensitivity Analysis. Sensitivity analysis is used to assess lignocellulose 19002000 (estimation)
the impact of variability of cost of raw material, catalysts
(including enzyme and byproducts), and solvent (Figure 4). Biocatalyst is the second signicant contributor to the GEVO
50% to 150% variation of the cost of the base scenario has been process cost. 50% to 150% biocatalyst cost variation results in
studied. Biomass feedstock cost is the most signicant factor of the minimum price of p-xylene ranging from $3667 to $4575/
p-xylene price. Biomass feedstock price varies signicantly metric ton. Since we use the biocatalyst amount based on the
based on the choice of biomass and its source. It is found that a batch process, the amount needed may be much smaller than
50% increase of biomass feedstock cost raises the minimum what is estimated in this study. In addition, the biocatalyst cost
cost of p-xylene by 21.8% in the GEVO process and by 12.5% may drop signicantly with the future advances in the
in the CCEI process, while a 50% increase of ethylene increases fermentation process, which is similar to the development of
the minimum cost of p-xylene by 5.9%. In the base scenario, a the ethanol fermentation. Therefore, a scenario of 10% of the
petroleum-based ethylene price is used in the economic analysis base scenario is considered that leads to a minimum price of p-
of the CCEI process. Biobased ethylene can be produced via xylene of $3462/metric ton.
dehydration of ethanol that is obtained from fermentation of Another important factor of the CCEI process is the solvent
biomass. In order to investigate the use of biobased ethylene in used (especially THF solvent). A 50% reduction of THF cost
the CCEI process dierent source alternatives are considered as decreases the minimum cost of p-xylene to $3429/metric ton;
given in Table 4. Using these dierent prices of bioethylene, the while a 50% increase results in the p-xylene price of $3863/
p-xylene minimum price may vary from $3636 to $4162 per metric ton. Solvent eects have been studied, and the results
metric ton (for reaction (10)). showed that HMF selectivity depends on its partition
10696 dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
Figure 5. Sensitivity analysis on the reaction parameters of the CCEI process (* corresponds to the overall conversion and selectivity from starch to
HMF (reactions (6)(8))).
coecient.18 A high volumetric ratio is often used to achieve
high conversion and selectivity but also results in a large
amount of solvent used, increasing the utility cost. Given that
there is a trade-o between high yield and high utility cost, an
economic optimal volumetric ratio can be found. However,
more information, regarding the relation between yield and
solvent amount for all the reactions involved using THF
solvent, is needed to achieve this. In addition, it is assumed that
there are no eects on the conversion of HMF to DMF
reaction, which might not be the case. To study these eects
sensitivity analysis is performed on the selectivity and
conversion of all the reactions. However, more studies on the
production of DMF using THF as a solvent are required to
understand the eects of the solvent and be able to design a
better separation process. The impacts of other costs such as
the H2 cost for the CCEI process, the enzyme cost for GEVO,
and the catalysts cost for both process are found to be
negligible.
Given the early stage of the experimental study of the CCEI
process, the impacts of the selectivity and conversion of the
reactions involved in this process have been investigated and
shown in Figure 5. The variability considered for the
conversion and selectivity of all the reactions is 5% to ensure
the same separation method can be used. For the reactions
characterized by high conversion rates (i.e., HMF to DMF and
DMF to p-xylene), only a 5% conversion reduction is
considered.
It should be pointed out that the results of the sensitivity
analysis illustrate that selectivity is favored compared to
conversion in terms of the cost reduction that can be achieved.
For example just a 5% increase of selectivity of the overall
reaction from starch to HMF results in a decrease of p-xylene
minimum cost by 3.6%, while a 5% decrease results in an
increase of p-xylene minimum cost by 6.6%. It is found that the
decrease of selectivity leads to more variation on p-xylene price
than the increase, due to the raw materials or intermediates loss
in the separation process. A 5% increase of the conversion of
the overall reaction from starch to HMF (reactions (6)(8))
results in very small change in p-xylene cost due to the recycling
of the raw materials. In comparison, a 5% decrease of the
overall conversion from starch to HMF leads to an increase of
1.9% of the p-xylene minimum price. This is due to the loss of
starch in the separation of humins. A 5% decrease in the
conversion of the HMF to DMF reaction results in a 5.5%
increase of the p-xylene cost. This change is higher compared to
10697
Article
the result reported in our previous work.16 This is because in
the current study the integration of upstream unit operations
results in an increasing loss of HMF due to the transformation
of the HMF that is present in the recycle stream to humins. A
5% decrease of conversion of the reaction of DMF to p-xylene
raises the p-xylene minimum price only 1.3%.
Finally, if the selectivity and conversion reported in a recently
published work23 is used (reaction (10*)), then an economic
analysis of the new process as described in the subsection
Process diagram and description (CCEI process) leads to a p-
xylene price of $3040/metric using the cost of catalyst and
solvent shown in Table 5.
Table 5. Catalyst and Solvent Cost of Reaction (10*)
cost ($/metric ton)
n-tridecane 3250
H-BEA catalyst 5000
SUMMARY
With the increasing attention on biobased p-xylene, the techno-
economic analysis of the alternatives production routes is of
great importance. This work focuses on the analysis of the
GEVO and CCEI processes. For the base case scenario, in
which no byproduct value is considered, it is found that the
minimum p-xylene price of the GEVO and CCEI processes is
$4121/metric ton and $3637 metric ton, respectively. These
prices are larger than the price of petroleum based p-xylene
which is $1630/metric ton.34 However, once the value of the
byproducts is considered and better integration is achieved as it
is done for petroleum reneries, the economics of biobased p-
xylene will be substantially improved. For example it is shown
that the consideration of some of the byproducts result in a
drop of the p-xylene minimum price to $3481/metric ton and
$2885/metric ton for the GEVO and CCEI processes,
respectively.
It was also found based on sensitivity analysis performed in
this study that the cost of biomass is one of the most signicant
factors on the price of p-xylene independent of the process
used, and this can vary a lot depending on the source of
biomass.
In addition, the economics of the GEVO process can be
improved with the development of biocatalyst used for the
fermentation of isobutanol. However, the maximum yield of
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research
isobutanol is 95% of the theoretical yield, namely 39 g
isobutanol/100 g glucose,43 which limits the economics of the
whole p-xylene production. The economics of the CCEI
process also depends on other material needed such as
ethylene. The replacement of petro-based ethylene with
bioethylene may increase the minimum p-xylene price since
most biobased ethylene is much more expensive than petro-
based ethylene except that produced from sugarcane feedstocks.
Sensitivity analysis is performed on the conversion and
selectivity of the reactions involved in the CCEI process to
assess their impacts on the economics. However, more studies
are necessary in order to further understand the eects of
dierent solvent choices on the reactions involved and the
separation process.
ASSOCIATED CONTENT
*
S Supporting Information
Tables S1S3 and Figures S1 and S2. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Phone: 732-445-2971. Fax: 732-445-2581. E-mail: marianth@
soemail.rutgers.edu.
Notes
The authors declare no competing nancial interest.
ACKNOWLEDGMENTS
This work was nancially supported from the Catalysis Center
for Energy Innovation, an Energy Frontier Research Center
funded by the U.S. Department of Energy, Oce of Science,
and Oce of Basic Energy Sciences under Award Number DE-
SC0001004.
REFERENCES
(1) Bozell, J. J. Feedstocks for the Future Biorefinery Production of
Chemicals from Renewable Carbon. Clean: Soil, Air, Water 2008, 36
(8), 641647.
(2) Perlack, R. D. et al. Biomass as Feedstock for a Bioenergy and
Bioproducts Industry: The technical Feasibility of a Billion-Ton Annual
Supply; United States Department of Energy: 2005.
(3) Chadwick, S. S. Ullmanns Encyclopedia of Industrial Chemistry;
1988(A21); pp 235237.
(4) Coca-Cola. The Coca-Cola Company Announces Partnerships to
Develop Commercial Solutions for Plastic Bottles Made Entirely From
Plants. Available from: http://www.thecoca-colacompany.com/
dynamic/press_center/2011/12/plantbottle-partnerships.html (ac-
cessed May 31, 2014).
(5) P&G. P&G Helps Push Toward Plant-Based Plastics. Available
from: http://news.pg.com/blog/pg-corporate-brand/pg-helps-push-
toward-plant-based-plastics (accessed May 31, 2014).
(6) Pepsi. PepsiCo Develops Worlds First 100% Plant-Based,
Renewably Sourced PET Bottle. Available from: http://www.pepsico.
com/PressRelease/PepsiCo-Develops-Worlds-First-100-Percent-
Plant-Based-Renewably-Sourced-PET-Bott03152011.html (accessed
May 31, 2014).
(7) Torres, A. I.; Daoutidis, P.; Tsapatsis, M. Continuous production
of 5-hydroxymethylfurfural from fructose: a design case study. Energy
Environ. Sci. 2010, 3 (10), 15601572.
(8) Peters, M. W. T., Joshua D.; Jenni, M.; Manzer, L. E.; Henton, D.
E. Integrated process to selectively convert renewable isobutanol to p-
xylene. 2011; US20110087000A1.
(9) Peters, M. W.; Taylor, J. D.; Jenni, M.; Manzer, L. E.; Henton, D.
E. Integrated process to selectively convert renewable isobutanol to p-xylene;
Gevo Inc.: 2011.
10698
Article
(10) Williams, C. L.; et al. Cycloaddition of Biomass-Derived Furans
for Catalytic Production of Renewable p-Xylene. ACS Catal. 2012,
935939.
(11) Nikolla, E.; et al. One-Pot Synthesis of 5-(Hydroxymethyl)-
furfural from Carbohydrates using Tin-Beta Zeolite. ACS Catal. 2011,
1 (4), 408410.
(12) Roman-Leshkov, Y.; et al. Production of dimethylfuran for
liquid fuels from biomass-derived carbohydrates. Nature 2007, 447
(7147), 982985.
(13) Atsumi, S.; Hanai, T.; Liao, J. C. Non-fermentative pathways for
synthesis of branched-chain higher alcohols as biofuels. Nature 2008,
451 (7174), 8689.
(14) Feldman, R. M. R.; Gunawardena, U.; Urano, J.; Meinhold, P.;
Aristidou, A.; Dundon, C. A.; Smith, C. Fermentative isobutanol
production by metabolically engineered yeasts with pyruvate decarboxylase
gene knockouts; Gevo, Inc.: USA, 2011.
(15) Peters, M. W., et al. Manufacture of xylene from C4 and C5
molecules obtained from fermentation of biomass; Gevo, Inc.: USA, 2012;
36pp.
(16) Lin, Z.; Ierapetritou, M.; Nikolakis, V. Aromatics from
Lignocellulosic Biomass: Economic Analysis of the Production of p-
Xylene from 5-Hydroxymethylfurfural. AIChE J. 2013, 59 (6), 2079
2087.
(17) Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Phase
Modifiers Promote Efficient Production of Hydroxymethylfurfural
from Fructose. Science 2006, 312 (5782), 19331937.
(18) Roman-Leshkov, Y.; Dumesic, J. Solvent Effects on Fructose
Dehydration to 5-Hydroxymethylfurfural in Biphasic Systems
Saturated with Inorganic Salts. Top. Catal. 2009, 52 (3), 297303.
(19) Kazi, F. K.; et al. Techno-economic analysis of dimethylfuran
(DMF) and hydroxymethylfurfural (HMF) production from pure
fructose in catalytic processes. Chem. Eng. J. 2011, 169 (13), 329
338.
(20) Evanko, W. A.; Eyal, A. M.; Glassner, D. A; Aristidou, A. A.;
Evans, K.; Gruber, P. R.; Hawkins, A. C. Recovery of higher alcohols from
dilute aqueous solutions; Gevo Inc.: 2009.
(21) Hawkins, A. C.; Glassner, D. A.; Buelter, T.; Wade, J.; Meinhold,
P.; Peters, M. W.; Gruber, P. R.; Evanko, W. A.; Aristidou, A. A.
Methods for the Economical Production of Biofuel Precursor That Is Also a
Biofuel from Biomass; Gevo, Inc.: 2009.
(22) Fellay, C.; Dyson, P. J.; Laurenczy, G. A Viable Hydrogen-
Storage System Based On Selective Formic Acid Decomposition with
a Ruthenium Catalyst. Angew. Chem. 2008, 120 (21), 40304032.
(23) Chang, C.-C.; et al. Ultra-selective cycloaddition of
dimethylfuran for renewable p-xylene with H-BEA. Green Chem.
2014, 16, 585588.
(24) Williams, C. L.; Chang, C.-C.; Phuong, D.; Nikbin, N.;
Caratzoulas, S.; Vlachos, D. G.; Lobo, R. F.; Fan, W.; Dauenhauer, P. J.
Cycloaddition of biomass-derived furans for catalytic production of
renewable p-xylene. ACS Catal. 2012, 2 (6), 935939.
(25) Gallo, J. M. R.; et al. Production and upgrading of 5-
hydroxymethylfurfural using heterogeneous catalysts and biomass-
derived solvents. Green Chem. 2013, 15 (1), 8590.
(26) Hirshleifer, J. On the Theory of Optimal Investment Decision. J.
Political Econ. 1958, 66 (4), 329352.
(27) Net cost of corn starch to U.S. wet millers. Available from: www.
ers.usda.gov (accessed May, 2013).
(28) Alibaba pricing. Available from: http://price.alibaba.com/
(accessed May 2013).
(29) Humbird, D.; Davis, R.; Tao, L.; Kinchin, C.; Hsu, D.; Aden, A.;
Schoen, P.; Lukas, J.; Olthof, B.; Worley, M.; Sexton, D.; Dudgeon, D.
Process Design and Economics for Biochemical Conversion of
Lignocellulosic Biomass to Ethanol; National Renewable Energy
Laboratory: 2011.
(30) Corma, A.; Martinez, C.; Doskocil, E. J.; Yaluris, G. Alkene
oligomerization process; BP Corporation North America Inc.; BP Oil
International Limited: June 3, 2009.
dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699
Industrial & Engineering Chemistry Research Article

(31) de Klerk, A. Key catalyst types for the ecient rening of

FischerTropsch syncrude: alumina and phosphoric acid. In Catalysis;
The Royal Society of Chemistry: 2011; Vol. 23, pp 149.
(32) de Klerk, A. Oligomerization of FischerTropsch Olefins to
Distillates over Amorphous SilicaAlumina. Energy Fuels 2006, 20 (5),
17991805.
(33) Fung, S. C.; Querini, C. A.; McCoy, C. J. Coke and Product
Proles Along Catalyst Bed in Reforming Reactions. In Studies in
Surface Science and Catalysis; Calvin, H. B., John, B. B., Eds.; Elsevier:
1991; pp 135142.
(34) ICIS pricing. Available from: www.icispricing.com (accessed
May 18, 2013).
(35) Cortright, R. D.; Blommel, P. G. Synthesis of liquid fuels and
chemicals from oxygenated hydrocarbons; Virent Energy Systems, Inc.:
USA, 2008; 63 pp.
(36) Chauvel, A.; Lefebvre, G. Petrochemical Processes Technical and
Economic Characteristics 1: Synthesis-gas Derivatives and Major Hydro-
carbons; ed. OPHRYS. 1989.
(37) Exxon, PxMax with Crystallization -- An integrated process for
high purity paraxylene production, 2009.
(38) Zhong, L.; Xiao, J.; Lu, Y.; Guo, Y.; Kong, D. Process for the
Crystallization Separation of p-Xylene; China Petroleum & Chemical
Corporation: Beijing, China, and Shanghai Research Institute of
Petrochemical Technology Sinopec: Shanghai, China, 2009.
(39) ICIS, http://www.icis.com/energy/ethanol/us/ (accessed June
28, 2012).
(40) Mundi, 2013. Available from: www.indexmundi.com/
commodities/?commodity=natural-gas (accessed May 2013).
(41) van Groenestijn, J.; Enzing, C.; van Dongen, M.; Bossenbroek, J.
Biobased economy Exploring the opportunities for the Netherlands:
Overview of high value added applications of biomass and biorefenery;
TNO Report.
(42) Production of Bio-ethylene. 2012. Available from: http://iea-
etsap.org/web/HIGHLIGHTS%20PDF/I13_HL_Bioethylene_
Broeren_Mar2012_FINAL9_GSOK.pdf (accessed March 2012).
(43) Hawkins, A. C. G.; David, A.; Buelter, T.; Wade, J.; Meinhold,
P.; Peters, M. W.; Gruber, P. R.; Evanko, W. A.; Aristidou, A. A.
Methods for the Economical Production of Biofuel Precursor That Is Also a
Biofuel from Biomass; Gevo Inc.: USA, 2009.

10699 dx.doi.org/10.1021/ie402469j | Ind. Eng. Chem. Res. 2014, 53, 1068810699