You are on page 1of 41

Department Editor: Kate Torzewski

Fluid Flow
Laminar PiPe fLow
For steady fow in a pipe (whether laminar or
turbulent), a momentum balance on the fuid gives
the shear stress at any distance from the pipe
centerline.
T T
rx w
r
L
r
R

$&
2

(1)
In Equation (1), = P + gz. The volumetric
fowrate Q can be related to the local shear rate
by doing an integration by parts of Equation (2).
Q r dr
rx
R


P G
2
0


(2)
Newtonian fuid. For a Newtonian fuid,

rx
=

rx
, which gives the following volumetric
fowrate, known as the Hagen-Poiseuille equation.
Q
D
L

P
M
4
128
$&

(3)
It can be written in dimensionless form in Equation
(4) with the two terms defned in Equations (5)
and (6).
f N = 16
Re
/

(4)
f
D
LQ

$& P
R
2 5
2
32


(5)
N
Q
D
Re

R
P M
4


(6)
Power law. A fuid that follows the power law
model obeys the relationship
rx
= (

rx
)
n
. This
gives the following equation.
Q
mR
n
n
R
w
n
n
n
=
+
+


1
3 1
3 1


(7)
Equation (7) can be rearranged into the following
dimensionless form.
f = 16 / N Re, pl

(8)
N
Q
m D
n
n
pl
n n
n n
n
Re,

+ ( )
,

,
]
]
]


2
3 1
7 3 2
2 4 3
p
r

(9)
Bingham plastic. In this case, there is a solid-like
plug fow region from the pipe centerline (where

rx
= 0) to the point where
rx
=
0
(that is, at r
= r
0
= R x
0
/
w
). The result is a fow integral
modifed from Equation (2). For a Bingham plas-
tic,
rx
=
0
+

rx
). Using this expression and
the modifed fow integral, the Buckingham-Reiner
Equation (10) is found.
P T
M
T
T
T
T

d
w
w w
Q
R
3
0 0
4
4
1
4
3
1
3


(10)
The equivalent dimensionless form is given by
Equations (11), (12) and (13).

f
N
N
N
N
f N
He He

d
16
1
1
6
1
3
4
3 7
Re d Re d Re

(11)
N
DV
Red

d
R
M

(12)
N
D
He

d
2
0
2
RT
M

(13)
TurbuLenT PiPe fLow
Since most turbulent fows cannot be analyzed
from a purely theoretical perspective, data and
generalized dimensionless correlations are used.
Newtonian fuid. The friction factor for a
Newtonian fuid in turbulent fow is a function of
both N
Re
and the pipe relative roughness, /D,
which can be read off the Moody diagram [5].
The turbulent part of the Moody diagram (for
N
Re
> 4,000) is accurately represented by the
Colebrook equation (14).
1
4
3 7
1 255
f
D
N f
+
,

,

]

]
]
log
.
.
Re
r

(14)
When N
Re
is very large, the friction factor depends
only on /D. This condition is noted with f
T
as the
fully turbulent friction factor in Equation (15).
1
4
3 7 f
D
T
=

log
.


(15)
The Churchill Equation [2] represents the entire
Moody diagram, from laminar, through transition
fow, to fully turbulent fow. It is presented here as
Equations (16), (17), and (18).
f
N
A B

j
(
,
\
,
(
+
+ ( )
,

,
,
]
]
]
]
2
8 1
12
1 5
1
12
Re
.

(16)
A
D

2 457
1
7 0 27
0 9
16
. ln
Re
.
.
E

(17)
B
N

37 530
16
,
Re

(18)
Power law. For a power-law fuid, the friction fac-
tor depends only upon Equation (9) and the fow
index, as represented by Equations (19)(25) [3].
f f
f f
L
T Tr
= +
+
[ ]

( ) 1
8 8
1
8



(19)
f
N
L
pl

16
Re,

(20)
f
n
N
T
pl
n

0 0682
0 5
1
1 87 2 39
.
[ ]
.
Re,
( . . )

(21)
f N
n
Tr pl
n
s
[


1 79 10
5 24
4
0 414 0 757
. e
. ]
Re,
. .

(22)
A


1
1 4
$

(23)
$ N N
pl plc Re, Re,

(24)
The value of N
Re
where transition from laminar to
turbulent fow occurs (N
Re,plc
) is given by Equa-
tion (25).
N n
plc Re,
, ( ) 2 100 875 1

(25)
Bingham plastic. For the Bingham plastic, f
T
is solely
a function of N
Re
and N
He
, as represented by
Equations (26)(29).
f f f
L
m
T
m m


1

(26)
f
N
T
a

d
10
0 193
Re
.

(27)
a
N
He

; =
s

14 7 1 0 146
2 9 10
5
. . e
.

(28)
m
N
1 7
40 000
.
,
Re

(29)
References
1. Darby, R., Take the Mystery Out of Non-Newtonian
Fluids, Chem. Eng., March 2001, pp. 6673.
2. Churchil, S. W., Friction Factor Equation Spans all Fluid-
Flow Regimes, Chem. Eng., November 1997, p. 91.
3. Darby, R., and Chang, H. D., A Generalized Correla-
tion for Friction Loss in Drag-reducing Polymer Solutions,
AIChE J., 30, p. 274, 1984.
4. Darby, R., and Chang, H. D., A Friction Factor Equation
for Bingham Plastics, Slurries and Suspensions for all
Fluid Flow Regimes, Chem. Eng., December 28, 1981,
pp. 5961.
5. Darby, R., Fluid Mechanics for Chemical Engineers,
Vol. 2, Marcel Dekker, New York, N.Y., 2001.
DEFINITIONS
Newtonian fuid. A fuid is known to be
Newtonian when shear stresses associated with
fow are directly proportional to the shear rate
of the fuid
Power law fuid. A structural fuid has a structure
that forms in the undeformed state, but then
breaks down as shear rate increases. Such a
fuid exhibits power law behavior at intermedi-
ate shear rates
Bingham plastic fuid. A plastic is a material that
exhibits a yield stress, meaning that it behaves
as a solid below the stress level and as a fuid
above the stress level
NOmENclaTurE
a Dimensionless parameter
A Dimensionless parameter
B Dimensionless parameter
D Diameter, m
f Fanning friction factor, dimensionless
f
L
Laminar friction factor, dimensionless
f
T
Fully turbulent friction factor, dimensionless
f
Tr
Transition friction factor, dimensionless
g Gravitational acceleration, m/s
2
L Length of cylinder or pipe, m
m Consistency coefficient, (N)(s)/m
2
n Power law fluid flow index, dimensionless
N
He
Hedstrom number, dimensionless
N
Re
Reynolds Number, dimensionless
N
Re,pl
Power law Reynolds Number, dimensionless
N
Re,plc
Power law Reynolds Number at transition
from laminar to turbulent flow, dimensionless
N
Re

Bingham-plastic Reynolds Number,
dimensionless
P Pressure, Pa
Q Volumetric flowrate, m
3
/s
r Radial position in a pipe or a cylinder, m
R Pipe or cylinder radius, m
V Velocity, m/s
z Vertical elevation above a horizontal refer-
ence plane, m
Dimensionless parameter

rx
Shear rate in tube flow, s
1
Wall roughness, m
Newtonian viscosity, Pas

Bingham Plastic limiting viscosity, Pas


Density, kg/m
3

0
Yield stress, N/m
2

rx
Stress due to force in x direction acting on r
surface, N/m
2

w
Stress exerted by fluid on tube wall, N/m
2
Flow potential, P + gz, Pa
ncrease in flow potential, Pa




















rp seIection or stem trcer Iines



p
p



p
p


1 t 2 tut h 2 25
191
2 o o
t t 88 tulb
72 lbh
lin
2





emperture dierentiI



ety ctor








InstIItion

















con
denste mnioId
shows muItipIe trcer
Iines nd stem trps
boe) he grph
depicts het Ioss o n uninsuIted
pipe he temperture dierentiI is
deried rom the process temper
ture minus mbient design 75 )
Department Editor: Rebekkah Marshall
Heat Transfer
Fluids and Systems
STARTUP
1. Verifycontrolandsafetysystems:Itisvitallyim-
portant to verify all control and safety systems
arecalibratedandreadyforoperationandare
functioningproperly
2.Checkforleakage
3.Remove moisture from the system, using dry,
compressedairorothersuitablemeans.Fillthe
systemwithheattransferfluid
4.Systemfilling
a. Fillthesystemtodesiredlevelavoidingany
unnecessaryaerationofthefluid
b. Openallvalves,thenstartthemaincircula-
tionpumpinaccordancewiththemanufac-
turers recommendations. Allow for thermal
expansion of fluid in determining the cold
chargelevel
c. Circulate the heat transfer fluid through the
systemforabout3to4hourstoeliminateair
pockets, and to assure complete system fill
beforefiringtheheater
5. Starttheheater
a. Bringthesystemuptotemperatureslowlyto
help prevent thermal shock to heater tubes,
tube/heater joints and refractory materials;
and allow operators to check the function-
ing of instruments and controls. The slow
heat-up will also allow moisture trapped in
allsectionsofthesystemtoescapeasvapor.
Inertgasshouldsweeptheexpansiontankto
removenoncondensablesandresidualmois-
turetoasafelocation.Holdthetemperature
stableabove100C(212F)untilnosigns
of moisture remain (knocking or rattling of
pipes,nomoisturefromvents,andsoon)
b. Bring the system to operating temperature,
puttheusersonline,andplacetheexpan-
sion-tankinertingsystemintooperation
c.Thefluidshouldgenerallybeanalyzedwithin
24hofplantstartupandannuallythereafter
d.Checkandcleanstartupstrainersasneeded
The system should be heated and cooled for at
least two cycles with the filter in place since the
resultingexpansionandcontractionwillloosenmill
scale. Reinsulate any surfaces left bare for leak-
checkingpurposes.
OPERATIONS
Heaters: Proper fluid-heater operation will help
ensurelonglifeofthefluid.Commonheaterprob-
lems include flame impingement, excessive heat
flux,controlovershoot,lowfluidflow,andinterlock
malfunctions.
Piping and pumps: Aleak-freesystemwillhelpto
ensuresafeandreliableoperation.Somekeyfea-
turesofaleak-freedesignareasfollows:
Maintainvalvesandpumppackingandseals
Avoid the use of threaded fittings (welded or
flangedconnectionsarepreferred)
Realignpumpsandretorqueflangesoncesys-
temachievesoperatingtemperatureafterinitial
systemstartup
Confirmwithyourfluidsupplierwhattheproper
elastomersare.Notallelastomersarecompat-
iblewithallheattransferfluids
FLUID ANALYSIS
Fluid testing helps detect system malfunction, fluid contamination, moisture, thermal
degradation, as well as other factors that impact system performance (see Table). For
systemsoperatingnearthefluidsmaximumtemperature,annualanalysisissuggested.
Possible actions
1. Filtration: Small diameter particles sus-
pended in heat transfer fluid can be ef-
fectivelyremovedviafiltration.Filterswith
100-mmorlessnominal-particle-removal
ratings should be considered for initial
system treatment. Continuous filtration
through10-mmratedfilterscanmaintain
systemcleanliness
2. Venting: Iflowboilerconcentrationand/
ormoistureisallowedtoreachexcessive
levelsinthefluid,problemssuchaspump
cavitation,increasedsystempressureand
flash-pointdepressioncanoccur.Intermit-
tent,controlledventingtoasafelocation
isacommonsolutiontominimizethepo-
tential for problems caused by excessive
lowboilerormoistureconcentration
3. Inerting: Aneffectivemethodofminimiz-
ing fluid oxidation is to blanket the ex-
pansiontankwithaclean,dry,inertgas,
suchasnitrogen,CO
2
,ornaturalgas
4. Dilution/replacement: Canbeusedtore-
movesomefraction(orall)ofthefluidand
replace with virgin fluid to maintain fluid
propertieswithinnormalranges.Caution
is advised when using reclaimed fluid,
which can return degradation products
and/orcontaminantsintothesystem
5. Cleaning: If a system flush is necessary,
several different methods are available.
Specialty-engineered, heat-transfer flush
fluids may be used to remove sludge or
tar from piping/equipment. Hard car-
bondepositsonheatersurfaces(coke)
generally require the use of mechanical
cleaning techniques like sand or bead
Test result Potential effects Possible cause Possible
actions*
Viscosity
increase
Poor heat-transfer rate, de-
posits, high vapor pressure,
pump cavitation
Contamination
Thermal degradation
Fluid oxidation
4, 5
4, 5
3, 4
Total acid
number
increase
System corrosion,
deposits
Severe oxidation
Acidic contamination
3, 4
4, 5
Moisture
increase
Corrosion, excess system
pressure, pump cavitation,
mechanical knocking
System leaks
Residual moisture in
new or cleaned unit
Unprotected vent or
storage
2
2
2, 3
Insoluble
solids
increase
Poor heat transfer, wear of
pump seals, plugging in
narrow passages
Contamination
Dirt
Corrosion
Oxidation
Thermal stress
1, 4, 5
1, 4
1, 3, 5
1, 3
1, 4
Low- and
high-boiler
increase**
Pump cavitation, poor heat
transfer, excess system
pressure, deposits
Low boilers
High boilers
Contamination
2
4
4, 5
* For detailed guidance on actions, please consult with your fluid engineering specialist.
** For an excellent discussion on low and high boilers, please consult Ref. [4].
blasting,wirebrushing,orhigh-pressure
waterjetting.Forprocesscontamination,
consult with your fluid supplier for sug-
gestedcleaningmethods
SHUTDOWN
Preventoverheatingoffluidduetoresidual
heatintheheater.
1. Shutoffburnercompletelywiththecircu-
lating pump still operating. Continue to
runthepumpatfullcapacitytodissipate
residualheatintheheater
2. Whentheheaterhascooledtothemanu-
facturersrecommendedlowtemperature,
shutoffthecirculatingpumpandswitch
offrequiredheaterelectricalcontrols
3.Caution must be exercised during shut-
downtoensurethatnoareainthesys-
tempipingistotallyandcompletelyiso-
lated. This will prevent a vacuum from
forming,whichcoulddamage(implode)
equipment
4. Operateheattracing,ifneeded
References and further reading
1.Gamble, C.E., Cost Management in Heat
TransferSystems,Chem. Eng. Prog.,July2006
pp.2226.
2. Gamble,C.E.,CleaningOrganicHeatTransfer
FluidSystems,Process Heating,Oct.2002.
3. Beain, others, Properly Clean Out Your Or-
ganicHeatTransferFluidSystem,Chem. Eng.
Prog.,May2001.
4. Spurlin,others,DefiningThermalStability,Pro-
cess Heating,Nov.2000.
5. LiquidPhaseDesignGuide,Pub#7239128C,
Solutia,Inc.,1999.
Facts at Your Fingertips SponSored by
Department Editor: Kate Torzewski
Heat Transfer
System Design
pumps
Centrifugalpumpswithcastorforged
steelcasingsaretypicallyappropriatefor
systemswithlargefowrates
Positivedisplacementpumpsfrequently
canhandlesmallerfowratesoflessthan
100gal/min
Seallesspumps(cannedmotoror
magneticdrive)avoidtheinstallationof
mechanicalseals
Atypicalcentrifugalpumpcanbeexpected
todeliveroperationalheadofabout:

wheregisgravitationalaccelerationinm/
s
2
,Distherotordiameterinmandnisthe
rotorspeedinrpm.
Pumpmanufacturersusuallyspecifythat
above450F(230C),acooled,jack-
etedstuffngboxoracooledmechanical
sealshouldbeused.Secondarysealing
withventanddrainglandsissuggested
tocollectfuidleakageandtoprovide
spaceforinertingtheoutsideoftheseal.
Inertblanketingofthesealwithsteam
ornitrogeneliminatesoxidationdeposit
formation,whichcanleadtosealleakage.
Thissecondarysealingprovidesadditional
safetyinthecaseofsuddensealfailure.
Somefuidleakageatthesealiscommon
inmechanicallysealedpumps.Forthose
applicationswherethesmallleakageis
undesirable,considertheuseofasealless
pump.Whenselectingaseallesspump,the
designermustconsidertheimpactofheat
generatedbythemotorstatorrelativetothe
vaporpressureofthepumpedfuid.
Onpumpswithastuffingbox,atleast
fiveringsofpackingshouldbepro-
vided,suchaslaminargraphiterings.If
expansionloopsareusedinthepump
suctionpiping,theyshouldbeinstalled
horizontally.
insulation
Normalhigh-temperatureinsulation,such
ascalciumsilicate,mineralwoolandcellu-
larglass,canbeusedinheattransferfuid
systems.However,fuid-saturatedinsulation
isapotentialfrehazardatthetempera-
turesoftenencounteredwhileoperating
suchsystems.Organicheattransferfuids
generallycanexhibitaslowoxidation
reactionwithairinthepresenceofporous
insulatingmaterialsattemperaturesabove
500F(260C).
Tominimizethefrehazardpotentialin
insulationsystems:
Usecellularglassinsulation,whichis
resistanttofuidsaturation,inareasofthe
systemwhereleaksarelikelytodevelop
Ifaleakdevelops,removetheinsulation,
andcontainthefuiduntiltheleakcanbe
repaired
Onverticalrunsofpipewhereoccasion-
alleakscandevelopatfanges,install
protectivetight-fttingcapstodivertany
fuidleakageoutsidetheinsulation
Installvalvestemshorizontallyorina
downwardpositionsothatanystemleak-
agedoesnotentertheinsulation
materials of construction
Systemmaterialsofconstructionmustbecom-
patiblewiththeheattransferfuidinuse.
Materialsofconstructionmustalsobe
selectedonthebasisoftheirsuitabilityfor
operationthroughoutthesystemstempera-
turerange(seetablebelow).Theusable
rangesarebasedonthelowtemperature
atwhichthematerialbecomesbrittleand
thehightemperatureatwhichthematerial
beginstolosemechanicalstrength.
Itisadvisabletopracticefre-safe
constructionwhendesigningorganic
heat-transfer-fuidsystems.Fireresistance
addressestheabilityofapipingsystemto
remainfreeofdamagingleakswhenex-
posedtoexternalfre.Apipingcomponent
istypicallyconsideredtobefreresistant
whenitisabletowithstandexposureto
1200F(650C)for30min.
Static seals Typical temperature
range
Fluoro-
elastomers
10F ( 25C)
to 400F (200C)
Flexible
graphite
325F (200C)
to 800F (425C)
Piping
Copper 325F (200C)
to 350F (175C)
Carbon steel 20F ( 30C)
to 800F (425C)
Low alloy steel 150F (100C)
to 800F (425C)
Stainless steel 325F (200C)
to 800F (425C)
piping layout
Thepossibilityofleakagethroughjointsand
fttingsischaracteristicofmostorganicheat-
transferfuidsunlessthefttingsareextremely
tight.Systemdesignshouldminimizethenum-
berofconnectionsinthepipinglayout.The
bestwaytopreventpipingleakageistoweld
allconnections.Useofthreadedfttingsis
stronglydiscouragedduetotheirtendencyto
leak.Whereaccessisnecessary,raised-face
fangeswithweldneckjointsorequivalent
raised-facefangesshouldbeused,asshown
inthetablebelow.
Flange
type
Use when Recommend-
ed gasketing
Class
150
Max. operating
temp. < 350F,
Operating range
< 300F (175C),
Operating range
< 300F (170C)
Metal-
inserted
fexible graph-
ite gaskets
Class
300
Max. operating
temp. > 350F,
Operating range
> 300F (175C),
Operating range
> 300F (170C)
Flexible
graphite-flled
spiral-wound
gaskets
Factorsaffectingtheleakperformanceof
thefangesincludetheabilityofthefange
boltingtoeffectivelyseatthegasketandthe
abilityofthefangetowithstandexternal
momentsinthepipingsystem.
Allpipinglayoutsshouldtakeintoac-
counttheexpansionandcontractionofthe
pipingwithtemperature,accordingtothe
equation ,where isthe
materialscoeffcientofthermalexpan-
sion.Itisvitalthatthestressplacedonthe
systemduringitsexpansionfromambientto
operatingconditionsnotexceedallowable
stresslimits.
filters
Beforestartupofanewsystem,install
awiremeshstrainerofapproximately
120-micronmeshsizeinthepumpsuction.
Pipingsystemsshouldalsobedesigned
withprovisionsfortheinstallationofaside-
streamflter.Filtersthathavegenerallybeen
employedfortheseapplicationsareglass
fberstring-woundcartridgesorcleanable
sintered-metalfltersinthe130-micronrange.
References
1.Wagner,W.,HeatTransferTechniquewithOrganic
Media,Grfelfng-Mnchen:TechnischerVerlagResch
KG,1977.
2.Gamble,C.E.,CostManagementinHeatTransfer
Systems,Chem. Eng. Prog.,July2006pp.22-26.
3.SystemsDesignData,Pub.#7239193ver.C,Solutia
Inc.,2002.
4.SystemDesignandMaintenance,Pub.#TBS10-25
(E),SolutiaInc.,1998.
5.LiquidPhaseDesignGuide,Pub.#7239128C,
SolutiaInc.,1999.
NOTICE:Althoughtheinformationandrecommendationssetforthherein(hereinafterInformation)arepresentedingoodfaithandbelievedtobecorrectasofthedatehereof,SolutiaInc.makesno
representationsorwarrantiesastothecompletenessoraccuracythereof.Informationissuppliedupontheconditionthatthepersonsreceivingsamewillmaketheirowndeterminationastoitssuitability
fortheirpurposespriortouse.InnoeventwillSolutiaberesponsiblefordamagesofanynaturewhatsoeverresultingfromtheuseoforrelianceuponInformationortheproducttowhichInformation
refers.Nothingcontainedhereinistobeconstruedasarecommendationtouseanyproduct,process,equipmentorformulationinconfictwithanypatent,andSolutiamakesnorepresentationorwar-
ranty,expressorimplied,thattheusethereofwillnotinfringeanypatent.NOREPRESENTATIONSORWARRANTIES,EITHEREXPRESSORIMPLIED,OFMERCHANTABILITY,FITNESSFORAPARTICULAR
PURPOSEOROFANYOTHERNATUREAREMADEHEREUNDERWITHRESPECTTOINFORMATIONORTHEPRODUCTTOWHICHINFORMATIONREFERS
Facts at Your Fingertips Sponsored by
By Rebekkah Marshall
Heat
transfer
Basic Equations for HEat transfEr
Conduction:

q
A
k
dT
dx
=

(1)
Convection:

q h A T T
avg s s f
=
( )

(2)
Radiation

q
A
T T
s sur
=
( )

4 4
(3)
HEat transfEr tHrougH
sHEll-and-tuBE HEat ExcHangErs
Heatdutyforexchangertransferringsen-
sibleheat:

q mC T T
p avg o i
=
( )

(4)
Foruseinheat-exchangercalculations,
Equation(2)aboveisoftenwrittenasfollows:

q UA T
s
=

(5)
whereUcanbecalculatedfromthefollowing
relationship;

1 1 1 1 1
U h h D D h h
o i i o w s
= +
( )
+ +
/

(6)
Variousequationsareavailable(seetherefer-
ences)forcalculatingh
i
andh
o
,dependingon
suchfactorsastheReynoldsnumbersforthe
flowingfluidsandwhetherthefluidsundergo
sensibleheattransferor,instead,vaporization
orcondensation.Forinstance,forsensibleheat
transferwithfluidsunderforcedconvection
infullyturbulentflowinsidetubeswithsharp-
edgedentrances,thefollowing,wellestablished
relationshipinvolvingtheNusselt,Reynoldsand
Prandtlnumbersholds:

h D
k
D m
c
k
i i i
p
w
=

0 023
0 8 1 3 0 14
.

. / .

(7)
Withtheassumptionthatthe(m/m
w
)termcan
beignored,theimmediatelyaboveequationhas
beenrearrangedasfollows[1,2]tofacilitate
assessingtheeffectsoffluid(andsystem)prop-
ertiesuponh
i
(assumingsensibleheattransfer,
fullturbulence,fluidinsidetubes):

h
m k c
D
i
p
i
= 0 023
0 8 2 3
1 3
0 2
0 47
.

. /
/
.
.

(8)
Forsensibleheattransferwithfluidsunder
forcedconvectionflowingacrosstubebanks
(thus,outsidethetubes),thefollowingrelation-
shiphasbeenpublished[3]

h
cm
a
c
k
D m
o
m
w

/ .
=

2 3 0 14
(9)
Inthisequation,thevaluesofaandmaretobe
asfollows:
Tube
pattern
Reynolds
number
m a
Staggered above200,000 0.300 0.166
Staggered 300to200,000 0.365 0.273
Staggered lessthan300 0.640 1.309
Inline above200,000 0.300 0.124
Inline 300to200,000 0.349 0.211
Inline lessthan300 0.569 0.742
ForanexcellentdiscussionoftheEquation(6)
foulingfactor,h
s
,seeReference[4].
TheappropriateDTdependsontheconfigura-
tionoftheheatexchanger(seereferences).
Forexample,forasimplecountercurrent-
flowexchanger,theappropriatetemperature
(referredtoasthelogmeantemperaturediffer-
ence,DT
LM
),isfoundasfollows;

T
T t T t
T t T t
LM
h h c c
h h c c
=

( )

( )

( )

( )
ln / (10)
Energybalanceforaheatexchanger
Ifanyheatexchangewiththeambientairis
neglected,thefollowingrelationshipisvalid;
mt H H mt H H q
h ha hb c cb ca

( )
=
( )
=
(11)
BatcH HEating
Forheatingabatchoffluidfromtemperature
T
1
toT
2
,bymeansofaninternalcoilofareaA
andanisothermalheatingmediumattempera-
tureT,thefollowingrelationshipholds:

ln
T T
T T
UA
cM

1
2

(12)
stEady-statE HEat flow
By conduction
Forconductionthroughahomogeneousplane
wallofthicknessxandconstant(oraverage)
thermalconductivityk,

q
A
k
T
x
=

(13)
whereDTisthetemperaturedifferencethrough
thewall
Forconductionthroughathree-layerplanewall
(forexample,awallwiththermalinsulationon
eachside),havinglayersofthicknessesx
1
, x
2

andx
3
andthermalconductivitiesk
1
, k
2
andk
3
,

q
T
x
k A
x
k A
x
k A
=
+ +

1
1
2
2
3
3
(14)
whereDTistheoveralltemperaturedifference
acrossallthreelayers
Forconductionthroughthewallofacylinderof
lengthL,whoseinnerandouterradiiarer
inner

andr
outer
,withinnerandouterwallsattem-
peraturesT
s,inner
andT
s,outer
,

q
k L T T
r
r
s inner s outer
outer
inner
=

( )( )
ln
, ,
2
(15)
nOMenCLatUre:
A cross-sectionalareaperpendiculartothe
flowofheat
a parameterinconvection-coefficient
equation
A
s
surfacearea
c,C
p
specificheat;specificheatatconstant
pressure
C
p,avg
specificheatataveragefluidtempera-
ture
D
i
innerdiameterofheat-exchangertube
D
o
outerdiameterofheat-exchangertube
H
ca
,H
ha
enthalpyperunitmassofentering
coldandwarmfluid,respectively
H
cb
,H
hb
enthalpyperunitmassofexiting
coldandhotfluid,respectively
h
avg
averageconvectioncoefficient
h
i
convectioncoefficientforinnertubewall
h
o
convectioncoefficientforoutertubewall
h
s
foulingheat-transfercoefficient
h
w
coefficientofheat-transferradially
throughtubewall;afunctionoftube
thicknessandthermalconductivity
k thermalconductivity
L length
M weightofbatch
m parameterinconvection-coefficient
equation
m

massflowrateoffluid
m

t
c
massflowrateofcoldfluid
m

t
h
massflowrateofhotfluid
q rateofheatflow
T temperature(forradiationcalculations,
useabsolutetemperature)
T
c
inaheatexchanger,theexittempera-
tureforthestreambeingcooled
T
f
temperatureoffluid
T
h
inaheatexchanger,theinlettempera-
tureforthestreambeingcooled
T
i
inlettemperature
T
o
outlettemperature
T
s
temperatureofsurface(forradiation
calculations,useabsolutetemperature)
T
sur
temperatureofsurroundings(forradia-
tioncalculations,useabsolutetempera-
ture)
t
c
inaheatexchangertheinlettempera-
tureforthestreambeingheated
t
h
inaheatexchanger,theoutlettempera-
tureforthestreambeingheated
DT
LM
logmeantemperaturedifference
DT/dx temperaturegradientduringconductive
heatflow
U overallheattransfercoefficient
x distancetheheatflowsduringconduc-
tion
e emissivity
m viscosity;viscosityatbulkfluidtempera-
ture
m
w
viscosityattube-walltemperature
s Stefan-Boltzmannconstant
q timerequiredforbatchheating
References
1. Guffey, G.E., Sizing Up Heat Transfer Fluids
and Heaters, Chem. Eng., pp. 126131,
October1997.
2. McCabe,SmithandHarriott,UnitOperations
ofChemicalEngineering,McGraw-Hill.
3. Chopey, ed., Handbook of Chemical
EngineeringCalculations,,McGraw-Hill.
4. Polley,PutFoulinginItsPlace,Chem.Eng.,pp.
46-49,December2002.
5. Incropera and DeWitt, Introduction to Heat
Transfer,Wiley.
6. Jones, Thermal Design of the Shell-and-Tube,
Chem.Eng.,pp.6065,March2002.






p p




eerences nd urther reding











cknowIedgment












Department Editor: Kate Torzewski
Biodiesel
Production
B
iodiesel can be produced from
vegetable oils by three types
of reactions: base catalyzed
transesterifcation of the oil; direct
acid-catalyzed transesterifcation of
the oil; and conversion of the oil to its
fatty acids, and then to biodiesel.
Biodiesel is typically produced by
a base-catalyzed reaction (Figure 2).
This method of production has several
advantages, including the following:
low temperature (150F) and pressure
(20 psi) reaction that requires only
standard materials of construction;
direct conversion to biodiesel with no
intermediate compounds; and high
conversion (98%) with minimal side
reactions and a low reaction time.
In the chemical reaction for base-
catalyzed biodiesel production,
vegetable oil is reacted with a short
chain alcohol (signifed by ROH)
in the presence of a catalyst to
produce glycerin and biodiesel. The
fatty acid chains associated with
the oil, which are mostly palmitic,
stearic, oleic, and linoleic acids for
naturally occurring oils, are repre-
sented by R', R'' and R''' (Figure 1).
Production stePs
Mixing of alcohol and cata-
lyst. The catalyst is typically sodium
hydroxide (caustic soda) or potassium
hydroxide (potash). It is dissolved in
the alcohol using a standard agitator
or mixer. Methanol or ethanol is com-
monly used as the alcohol.

Reaction. The mixture of alcohol
and catalyst is charged into a closed
reaction vessel, and the oil is added.
The reaction mix is kept just above the
boiling point of the alcohol, 160F,
to speed up the reaction, although it
is sometimes recommend to run the
reaction at room temperature. The
reaction time can vary from 18 h.
Excess alcohol is used to ensure total
conversion of the oil to its esters.
The amount of water and free
fatty acids in the incoming oil must
be monitored, because if either
level is too high, it can inhibit soap
formation and the separation of
glycerin downstream.

Separation. Glycerin and biodiesel
are the two main products of reac-
tion, with each containing an amount
of unreacted alcohol. Since the
glycerin phase is much more dense
than biodiesel phase, the two phases
can be separated by gravity in a
settling vessel, with glycerin simply
drawn off the bottom of the settling
vessel. Alternatively, a centrifuge can
be used to separate the two materials
more quickly.

Glycerin neutralization. The
separated glycerin contains unused
catalyst and soaps, which are neutral-
ized with an acid. Water and alcohol
are removed to produce glycerin
at 8088% purity to sell as crude
glycerin. Alternatively, glycerin can
distilled to 99% purity or higher for
selling to the cosmetic and pharma-
ceutical industries.

Methyl ester wash. After the
biodiesel is separated from glyc-
erin, residual catalyst or soaps can
be removed with a gentle warm
water wash.

Alcohol removal. Unreacted alco-
hol in both the glycerin and biodiesel
phases is removed by fash evapora-
tion or distillation. The recovered alco-
hol is then reused for mixing with the
catalyst. Alcohol removal can occur
after the wash and neutralization, as
shown in Figure 2 to the right, but it
can occur before these steps as well.
Product quality and registra-
tion. Prior to use as a commercial
fuel in the U.S., the fnished biodiesel
must be analyzed to ensure it meets
ASTM specifcations. Additionally, all
biodiesel produced must be regis-
tered with the U.S. Environmental
Protection Agency (Washington, D.C)
under 40 CFR Part 79.
References
1. Biodiesel Production & Quality Standards,
July, 2008. National Biodiesel Board,
www.biodiesel.org/resources/fuelfactsheets
Mixer
Reactor
Settler
Biodiesel
Wash
D
i
s
t
i
l
l
a
t
i
o
n
Neutral-
ization
Glycerin
Biodiesel Glycerin
Soybean oil
D
i
s
t
i
l
l
a
t
i
o
n
Catalyst Methanol
E
x
c
e
s
s

m
e
t
h
a
n
o
l
Mixer
Reactor
Settler
B
i
o
d
i
e
s
e
l
Wash
D
i
s
t
i
l
l
a
t
i
o
n
Neutral-
ization
G
l
y
c
e
r
i
n
Biodiesel Glycerin
Soybean oil
D
i
s
t
i
l
l
a
t
i
o
n
Catalyst Methanol
CH
2
OCOR
CH
2
OCOR + 3 ROH +
CH
2
OCOR
CH
2
OH
CH
2
OH
CH
2
OH
R COOR
R COOR
R COOR
Vegetable
oil
Alcohol Glycerin Biodiesel
Catalyst
Department Editor: Scott Jenkins
Burner Operating
Characteristics*
B
urners are critical for the successful op-
eration of industrial furnaces. Presented
here is a set of equations that can be
used to calculate characteristics of burner
operation, including ame length, ame di-
ameter, ignitability and ameout conditions.
Equations are based on pre-mix burners
operating at atmospheric pressure and r-
ing natural gas only. Premix burners create
short and compact ames compared to raw
gas burners, and are designed to function
with fuel-gas mixtures that have consistent
specic gravity and composition.
Burner requirements
For direct-red heaters to function correctly,
burners must be capable of providing suf-
cient heat liberation from the fuel to meet
heater processing requirements based on
the lower heating value (LHV) of the fuel. A
fuels LHV can be dened as the amount of
heat produced by combusting a specied vol-
ume, and returning the combustion products
to 150C. For the heater to operate at the
design process owrate, the burners need to
provide the heat necessary to maintain pro-
cess uid temperature and meet vaporization
requirements at the heating coil outlet.
The number, size and placement of t
burners must allow each coil to operate
at the same design outlet temperature
Design tube-metal temperature cannot t
be exceeded at any point on the coils
Burner size must allow an outlet veloc- t
ity that does not result in malfunction
over the range of ow conditions
Burner ame length should be less than t
rebox height (for vertical cylindrical
heaters) or less than rebox length (for
end-wall-red heaters)
Excessive ame height and diameter t
should be avoided to prevent ame
impingement on tubes
Burner spacing should be sufcient t
to allow burner-to-burner, as well as
burner-to-tube clearance
The following equations can help establish
optimal burner diameter:
(1)

(2)

(3)

(4)

Burner clearance
Establishing burner-to-burner clear-
ance and burner spacing should be
based on maximum burner ame
diameter. Further, burner ame
diameter should be evaluated at
maximum burner-ame length.
Sufcient burner-to-burner, outside
diameter clearance should take into
account the placement of structural
elements between burners.
Sufcient burner-to-burner clear-
ance prevents interference between
the ame bodies and unburned
fuel cores generated by adjacent
burners, which results in the absence
of unburned fuel within the burner
ame when maximum ame length
is reached. Burner center-to-center
spacing should be at least one fully
combusted ame diameter.
Clearance between the burner-
ame (at maximum diameter) and
the outside diameter of tubular
heating surfaces should be set
such that burner-to-tube ame
impingement is avoided. Doing so
will prevent tube damage due to
overheating and will make best use
of heating surfaces.
Flameout
At high burner velocities, ame
loss can occur if the heat gain due
to burner ignition is less than the
heat loss from the burner ame.
Burner velocities may be pushed
well above that used in normal
heater operation in an effort to
achieve higher heater capacity. Aside from
ame loss while the heater is in operation,
ameout can also be characterized by dif-
culty maintaining a stable ame at startup,
or an inability to ignite the burner. The
following equations can help predict the
circumstances under which amout condi-
tions might occur:
(5)

(6)
(7)
(8)
(9)

Flame velocity
The heat generated by combustion is de-
pendent on the ame propagation velocity.
In a situation with 0% excess air, the ratio
of fuel-to-fuel+air is about 0.1. In that case,
evaluation of the ame propagation velocity
is straightforward. However, at fuel-to-
fuel+air ratios higher or lower than 0.1, it is
more difcult. The following equations can
help predict ame propagation velocity in
those cases:
(10)
(11)
(12)
References
1. Cross, A., Fired-Heater Burner Performance,
Chem. Eng., April 2008, pp. 4447.
NOMENCLATURE
Q
lib heater
= Heater liberation, Btu/h
N
b
= Number of burners
D
b
= Burner diameter, ft
V
b
= Burner exit velocity, ft/s
C
fuel
= Fuel, ft
3
LHV = Lower heating value of fuel, Btu/lb
C
air+fuel
= Volume of air and fuel mixture, ft
3
SV
fuel
= Specific volume of fuel, ft
3
/lb
D
f max
= Maximum flame diameter, ft
L
f
= Flame length, ft
SV
flame
= Specific volume of flame, ft
3
/lb
V
f
= Flame propagation velocity, ft/s
Q
gain
= Burner heat gain, Btu/h
Q
loss
= Burner heat loss, Btu/h
A
s
= Flame front area, ft
2

(HTC)
c
(HTC)
f
, (HTC)
r
= Natural convective, forced-
convective, and radiative heat transfer
coefficients, respectively, Btu/h-ft
2
-F
T
flame
= Flame temperature, R
T
surr
= Surrounding temperature, R
E
g
= Flame emissivity
C
p
= Gas specific heat, Btu/lb-F
A = Frequency factor in the Arrhenius equation
H = Heat of activation, Btu/lb-mol R
R = Gas constant, 1.987 Btu/lb-mol R
T = Gas Temperature, R
dC
m
/dt = Fuel concentration change, mol per ft
3
/s
K = Reaction velocity constant, s
1
W
f
= Fuel, lb/h
*The text was adapted from the article Fired-Heater Burner Performance, by Alan Cross. It appeared in
the April 2008 issue of Chemical Engineering.
















eturns rom eiciency projects
















xmpIe ir prehet









onsidertions









eerence




rude heter remp
nergy eiciency projects economics
cuum tower stem ejectors
ot eed or stem genertion
reenhouse gses price, $metric ton
1
8
6

2 6 8 1
ower recoery
bbId rreIsperstremdy
ir prehet economics remp
5

25
2
15
1
5

5 1 15 2 25
rocess bsorbed duty, miIIion tuh
$1
$9
$8
$7
$6
$5
$
$
$25
$2
$15
$
rious energy eiciency projects with cost o emissions
incIuded
emp heter ir prehet economics with emissions cost
incIuded
Department Editor: Rita L. D'Aquino
AchievabIe combustibIes in ash, %
Unit heat output,10
6
Btu/h
Average fueI cost, $/10
6
Btu
F
u
e
I

s
a
v
i
n
g
s
,

$
1
,
0
0
0
/
y
r
5
10
15
20
50
CombustibIes in ash, %
320
240
160
80
0
6.00 4.00 3.00 2.50 2.00
1.50
1.00
0.50
40 30 20 10
10 25 50 10 100
150
200
300
400
0
B
C
A
FIGURE 1. This nomograph is used to
estimate annual cost savings from reducing
combustible losses due to unburned carbon
This article has been drawn from the work of Wayne Turner and Steve Doty, Boilers and Fired
Systems, Energy Management Handbook, 6th Ed., Ch. 5, The Fairmont Press, Lilburn, Ga., 2006.
Fuel Selection
Considerations
T
he selection and application of fuels to various
combustors are complex. Most existing units have
limited fexibility in their ability to fre alternative
fuels. New units must be carefully planned to assure
the lowest frst costs without jeopardizing the future
capability to switch to a different fuel.
Natural gas
Natural gas has traditionally been the most attrac-
tive fuel type for combustors because of the limited
need for fuel-handling equipment (e.g., pipelines,
metering, a liquid-knockout drum, and appropriate
controls) and the freedom from pollution-control
equipment. Drawbacks include rising fuel costs, in-
adequate gas supplies and lower boiler efficiencies
that result from firing natural gas, particularly when
compared to the firing efficiencies of oil or coal.
Fuel oil
Fuel oils are graded as No. 1, No. 2, No. 4, No.
5 (light), No. 5 (heavy), and No. 6. Distillates are
Nos. 1 and 2, and residual oils are Nos. 4, 5, and
6. Oils are classifed according to their physical
characteristics by the American Society for Testing
and Materials (ASTM) per Standard D-396. No.
2 oil is suitable for industrial use and for home
heating. The primary advantage of using a distillate
oil rather than a residual oil is that it is easier to
handle, requiring no heating to transport and no
temperature control to lower the viscosity for proper
atomization and combustion. However, consider-
able purchase cost penalties exist between residual
and distillate.
Distillates can be divided into two classes:
straight-run, which is produced from crude oil by
heating it and then condensing the vapors; and
cracked, which involves refning at a high tempera-
ture and pressure, or refning with catalysts to pro-
duce the required oil from heavier crudes. Cracked
oils contain substantially more aromatic and olefnic
hydrocarbons, which are more diffcult to burn
than the paraffnic and naphthenic hydrocarbons
obtained from the straight-run process. Sometimes
a cracked distillate, called industrial No. 2, is used
in fuel-burning installations of medium size (small
package boiler or ceramic kilns, for example).
Because of the viscosity range permitted by
ASTM, No. 4 and No. 5 oil can be produced in
a variety of ways: blending of No. 2 and No.
6, mixing refnery by-products, utilization of off-
specifcation products, and so on. Because of the
potential variations in characteristics, it is important
to monitor combustion performance routinely to
obtain optimum results. Burner modifcations may
be required to switch from, say, a No. 4 blend to a
No. 4 distillate.
Light (or cold) and heavy (or hot) No. 5 fuel oil
are distinguished primarily by their viscosity rang-
es: 150 to 300 SUS (Saybolt Universal Seconds) at
100F and 350 to 750 SUS at 100F, respectively.
The (No.) classes normally delineate the need for
preheating for proper atomization.
The No. 6 fuel oil is also referred to as residual,
Bunker C, and reduced- or vacuum bottoms. Because
of its high viscosity, 900 to 9,000 SUS at 100F, it
can only be used in systems designed with heated
storage and a high enough temperature (to achieve
proper viscosity) at the burner for atomization.
Notable fuel oil properties include the following:
1) Viscosity indicates the time required in seconds
for 60 cm
3
of oil to fow through a standard-size
orifce at a specifc temperature. In the U.S., it is
normally determined with a Saybolt viscosimeter,
which comes in Universal and Furol variants. The
differences between them are the orifce size and
the sample temperature. Thus, when stating an oils
viscosity, the type of instrument and temperature
must also be stated. The Universal has the small-
est opening and is used for lighter oils. 2) The
fash point is the temperature at which oil vapors
are ignited by an external fame. As heating
continues above this point, suffcient vapors are
driven off to produce continuous combustion. The
fash point is also an indication of the maximum
temperature for safe handling. Distillate oils have
fash points of 145200F; heavier oils have
fash points up to 250F. 3) The pour point is the
lowest temperature at which an oil fows under
standard conditions, and is roughly 5F above
the solidifcation temperature. Knowledge of the
pour point helps determine the need for heated
storage, the storage temperature, and the need
for heating and pour-point depressant.
Coal
The selection of coal as fuel involves higher
capital investments because of the need for
handling equipment, coal preparation (crush-
ing, conveying, pulverizing, etc.) and storage;
ash handling and storage; pollution-abatement
equipment; and maintenance. The operating
cost savings at current (2007) fuel prices of coal
over oil or gas justifes a great portion, if not
all, of the signifcantly higher capital invest-
ments required for coal.
Coal-fred steam generators and vessels
inherently suffer effciency losses due to a failure
to burn all the available fuel. The unburned fuel
is the remaining carbon in the leftover ash. The
nomograph (Figure 1) may be used to assess
how a reduction in unburned carbon translates
into energy and cost savings. A sample calcula-
tion follows.
Example: The system is a coal-fred steam
generator with a continuous rating of 145,000
lb/h; average (avg.) boiler load = 125,000 lb/h;
existing combustibles in ash = 40% (measured);
obtainable combustibles in ash = 5%; actual oper-
ating time = 8,500 h/yr; design-unit heat output =
150 10
6
Btu/h; avg. heat output = 129 10
6

Btu/h; avg. fuel cost = $1.50/10
6
Btu.
Analysis: In Figure 1A, the percent of existing
combustibles (measured) are shown on the
horizontal axis. The curves above it represent the
proposed improvement in percent of unburned
carbon in ash. From the coordinates in Figure
1A draw a horizontal line to the curve in Figure
1B that represents the design-unit heat output.
Drop the line to the appropriate fuel-cost curve in
Figure 1C. Extend the line from that point to the
left to obtain the corresponding annual fuel sav-
ings, assuming continuous operation at full boiler
design output. To calculate actual annual fuel
savings, a correction factor (CF) is required that
considers actual boiler load and actual run time:
Actual savings, $ = Savings from chart x CF
where CF = operating avg. heat output/design
heat output [(actual operating h/yr)/(8,760
h/yr)]
Savings for this example =
$210,000/yr [(129 10
6
Btu/h)/(150 10
6

Btu/h)] [(8,500 h/yr)/(8,760 h/yr)]
= $175,200/yr.
Note: If the heat output of the unit or the average
fuel cost exceeds the limit of the fgures, use half
of the particular value and double the savings
obtained from Figure 1C.
It is probable that pulverized-coal-fred instal-
lations suffer from high UCL whenever any of
the following are experienced: a change in the
raw-fuel quality from the original design basis;
deterioration of the fuel burners, burner throats,
or burner swirl plates or impellers; increased
frequency of soot blowing to maintain heat-transfer
surface cleanliness; a noted increase in stack gas
opacity; uneven flame patterns characterized by a
particularly bright spot in one section of the flame
and a notably dark spot in another; CO formation
as determined from a flue-gas analysis; frequent
smoking observed in the combustion zone;
increases in refuse quantities in collection devices;
neglect of critical pulverizer internals and classifier
assemblies; a high incidence of coal hang-up in
the distribution piping to the burners; and frequent
manipulation of the air/coal primary and second-
ary air registers.
Techniques used successfully to reduce high
UCL and/or high-excess-air operation include:
modifying or replacing the pulverizer internals to
increase the coal fneness; installing additional or
new combustion controls to maintain consistent
burner performance; purchasing new coal feeders
that are compatible with and responsive to unit
demand fuctuations; calibrating air fow and
metering devices to ensure correct air/coal mixtures
and velocities at the burner throats; installing
turning vanes or air foils in the secondary air-supply
duct or air plenum to ensure even distribution and
proper air/fuel mixing at each burner; replacing
worn and abraded burner impeller plates; installing
new classifers to ensure that proper coal fnes
reach the burners for combustion; rerouting or
modifying air/coal distribution piping to avoid
coal hang-up; increasing the air/coal mixture
temperature exiting the pulverizers to ensure good
ignition without coking; and cleaning deposits from
burner throats.
Department Editor: Kate Torzewski
Causes of
Overpressurization
T
he failure of a device or of a group of
components can lead to overpressuriza-
tion and subsequent adverse events, such
as fre, explosion, spill or release. The most
common causes of overpressurization are
listed below. Understanding the circumstanc-
es surrounding overpressurization will help
an engineer to avoid these failures.
External fre
According to API RP 520 and 521 standards,
a fre-exposed area is within an area of
2,500 and 5,000 ft
2
, and below a height of
25-ft above the grade. In this scenario, the
exposed vessel is blocked in. Potential vapors
resulting from the fre must be relieved using
a PRV on the vessel, or via a vent path that
remains in a locked-open position between
the vessel and an adjoining vessel.
Blocked outlets
The closure of a block valve on the outlet of a
pressure vessel can cause the vessel's internal
pressure to exceed its maximum allowable
working pressure if the source pressure
exceeds the vessel design pressure. Blocked
outlets can be caused by control valve failure,
inadvertent valve operation, instrument-air
or power failure, and other factors. A PRV
must have suffcient capacity to pass a fuid
fowrate that is equivalent to the difference
between those of the incoming fuids and the
outgoing fuids.
Utility failures
These failures can include the following: gen-
eral power failure, partial power failure, loss
of instrument air, loss of cooling water, loss of
steam, and loss of fuel gas or fuel oil. For these
cases, a fare header should be designed and
sized based on the maximum relief load that
could result from a potential utility failure.
Loss of cooling duty
Cooling-duty losses can include the following:
loss of quench stream, air-cooled exchanger
failure, loss of cold feed and loss of refux.
Relieving capacity should be calculated by
performing a heat balance on the system,
based on the loss of the condensing duty.
Thermal expansion
When liquid is blocked in a vessel or
pipeline, external heat input can cause liquid
temperature, and hence volume, to rise. This
can be caused by the following: liquid that
is blocked in a pipeline and is being heated,
the cold side of a heat exchanger being flled
while the hot side is fowing, or a flled vessel
at ambient temperature that is being heated
by direct solar radiation. PRVs used in these
cases can be easily analyzed and sized.
Abnormal heat input
This failure can be caused by: the supply of
heating medium, such as fuel oil or fuel gas to
a fred heater, being increased; heat transfer
occuring in a new and clean heat exchanger
after revamp; control valve for the fuel supply
failing to fully open; or supply pressure of the
heating steam being changed from normal
range to maximum pressure. As a general
rule, when sizing a PRV, maximum heat-duty
assumed for the abnormal case should be no
more than 125% of normal heat duty.
Abnormal vapor input
Abnormal vapor input can be caused by the
failure of the upstream control valve to fully
open, or upstream-relieving or inadvertent
valve opening. The required relieving
capacity must be equal to or greater than the
amount of the vapor accumulation expected
under the relieving conditions.
Loss of absorbent fow
When gas removal by absorbent is more
than 25% of the total inlet-vapor fow, an
interruption of absorbent fow could cause
pressure to rise in the absorber. The PRV
should be sized base on the net accumulation
of the vapor at the relieving conditions.
Entrance of volatile materials
The entrance of a volatile liquid, such as
water or light hydrocarbons into hot oil dur-
ing a process upset, can cause instantaneous
phase expansion. Instead of relying on PRVs,
processes should be properly designed with
the use of double block valves, the avoid-
ance of water-collecting pockets and use of
steam condensate traps and bleeds on water
connections.
Accumulation of non-condensibles
Accumulation can result from blocking of the
normal non-condensible vent or accumulation
in the pocket of a piping confguration or
equipment. Because this can result in a loss of
cooling duty, PRV analysis should be handled
the same way.
Valve malfunction
Check-valve malfunction results in backfow,
which can be from 5 to 25% of the normal
fowrate. Required relief capacity should be
based on this.
Inadvertent valve operation results in a
valve position that is opposite from normal
operating conditions, which is largely caused
by human error and can be avoided by care-
ful operation.
Control valve failure to open or close is
caused by electronic- or mechanical-signal
failure. This typically will affect just one valve
at a time and should be analyzed on a case-
by-case basis.
Process control failure
This situation refers to the failure of process
controllers, such as programmable logic
controllers and distributed control systems.
The potential impact of failure of every con-
trol loop should be analyzed, as well as the
impact if one loop fails but all others remain
active. As a general rule, the required relief
capacity must be greater than the vapor gen-
erated because of heat buildup in the system.
Exchange tube rupture
When an upstream vessel is relieving by
discharge fuid to a downstream vessel,
the downstream vessel should should be
designed to handle the pressure and volume
of the incoming stream without overpressur-
izing. If the upstream vessel does not have
adequate relief capacity, the downstream
vessel should have a PRV of its own.
When two vessels are connected by an
open path and the frst has its own PRV and
discharges to a fare header, the second
will experience the impact from the relieving
pressure of the frst vessel and should be
analyzed accordingly.
Upstream relieving
Required relief capacity should be greater
than the vapor generated because of heat
buildup in the system.
Runaway chemical reaction
Runaway reactions tend to accelerate with
rising temperature; extremely high volumes
of non-condensibles with high energy can
cause the internal pressure of a vessel or
pipeline to rise rapidly. PRVs may not provide
suffcient relief, so vapor-depressurizing
systems, rupture disks and emergency vents
are preferable.
References
1. Wong, W., Protect Plants Against Overpres-
sure, Chem. Eng. June 2001, pp. 6673.
2. Goodner, H., A New Way of Quantifying
Risks: Part 2, Chem. Eng. November 1993, pp.
140146.
3. Emerson, G., Selecting Pressure Relief Valves,
Chem. Eng. March 18, 1985, pp. 195200.
Spring
Disc
NozzIe
A pressure relief valve (PRV)
is an automatic pressure-
relieving device that is actu-
ated by pressure at the inlet
of the valve. Though safety
valve, or safety relief
valve, is the terminology
for valves relieving gas
or vapor, we will use PRV
to describe all types of
pressure relief valves. A
relief valve, used for
liquid service, generally
opens in proportion to
any increase in pressure
over opening pressure.
A safety valve has char-
acteristics similar to a
relief valve except that
it usually opens rapidly
(pops), and is primarily
used for gas or
vapor service
Source:
Farris
Engineering,
Brecksville,
Ohio
Department Editor: Rebekkah Marshall
Hazardous Area
Classifcation
Guidelines by location
Over the years, hazardous area clas-
sification requirements for the U.S.
have evolved around a single area-clas-
sification system known as the Class/
Division system. Today, the system
addresses establishment of boundaries
of hazardous areas and the equipment
and wiring used in them. Meanwhile,
European countries, as well as some
other countries around the world, have
developed their own area classification
systems to address hazardous locations
safety issues. This independent develop-
ment has resulted in systems for these
countries or groups of countries based
on the International Electrotechnical
Commission (IEC) Zone system, with de-
viations to meet each countrys national
codes. While other countries do accept
and use the Division system (most nota-
bly Canada and Mexico), the majority
of the worlds hazardous locations are
classified using the concepts of the IEC
Zone system. The U.S. National Electri-
cal Code (NEC; NFPA 70) also recog-
nizes the Zone system and allows its
use in the U.S. under article 505 of the
NEC. ATEX requires the use of IEC-type
hazardous area classifications.
defininG hazardous areas
A hazardous area is designated as any
location in which a combustible material
is or may be present in the atmosphere
in sufficient concentration to produce an
ignitable mixture. The North American
method identifies these areas by Class,
Division and Group or optionally by
Class, Zone and Group, while the IEC
and CENELEC designate these areas by
Gas/Dust, Zone and Group. The likeli-
hood that the explosive atmospheres are
present when the equipment is operating
are designated in Tables 1, 2 and 5.
equipment selection
For equipment selection purposes, haz-
ardous area classifications also consider:
The maximum surface temperature of
the equipment under normal operat-
ing conditions (see the Temperature
Code designations in Table 3)
The ignition-related properties of
the explosive atmosphere (see the
Group designations in Table 4)
The protection method(s) used by
the equipment to prevent ignition of
the surrounding atmosphere (see the
Protection Method designations in
Table 6)
Acknowledgement and references
We would like to thank Vladimir Stetsovsky of Chilworth
Technology, Inc. for his contributions to this page
1. National Electrical Code-2005-NFPA 70, National
Fire Protection Association.
2. NFPA 497-2004, Recommended Practice for the
Classification of Flammable Liquids, Gases, or
Vapors and of Hazardous (Classified) Locations for
Electrical Installations in Chemical Process Areas.
3. NFPA 499-2004, Recommended Practice for the
Classification of Combustible Dusts and of Hazard-
ous (Classified) Locations for Electrical Installations
in Chemical Process Areas.
4. IEC 60079-10-2002 Electrical apparatus for explo-
sive gas atmospheres Part 10: Classification of
hazardous areas.
5. IEC 61241-3-2005 Electrical apparatus for use in
the presence of combustible dust Part 3: Classifi-
cation of areas where combustible dusts are or may
be present.
Table 3. Temperature Codes
The Temperature class defines the maximum
surface temperature of the device. Ratings are
given with reference to 40C ambient
T1 450C T3A 180C
T2 300C T3B 165C
T2A 280C T3C 160C
T2B 260C T4 135C
T2C 230C T4A 120C
T2D 215C T5 100C
T3 200C T6 85C
The additional temperature classifications high-
lighted above are for USA and Canada only
Table 4. Gas and Dust Groups
Hazardous locations are grouped according to their ignition
properties
Typical
gas
IEC gas
group
North Ameri-
can group
Minimum ignition
energy
Acetylene IIC A 20J
Hydrogen IIC + H
2
B 20J
Ethylene IIB C 60J
Propane IIA D 100J
*Methane I
Metal dust E
Coal dust F
Grain dust G
Fibers
*Mining application under jurisdiction of U.S. Mine Safety and
Health Administration (MSHA)
Table 2. Relationship Between Divisions and Zones
North America Europe
Division
method
Zone
method IEC standard
Ignitable mixture present
continuously (long periods) Division
1
Zone
0
Zone 0
(Zone 20-Dust)
Ignitable mixture present
intermittently
Zone
1
Zone 1
(Zone 21-Dust)
Ignitable mixture is not
normally present
Division
2
Zone
2
Zone 2
(Zone 22-Dust)
Table 5. Information Required For Establishing
Extent of Hazardous Area
Gas/Vapors Dust
Flash point
Flammability limits
Auto-ignition temperature
Minimum ignition energy,
MIC or MESG for equipment
selection purposes
Gas/Vapor group
Vapor/Gas density
Area ventilation conditions
Location of gas/vapor release
points. Frequency and rate of
release
A/B classification
Minimum explosible dust
concentration
Minimum ignition energy
Minimum ignition tempera-
ture (cloud/layer)
Electrical resistivity
Dust group
Area ventilation conditions
Location of dust release
points. Frequency and rate of
release
Table 6. Types of Protection for Electri-
cal Equipment (IEC/ATEX and NEC)
Technique
IEC
Des-
cription
Permitted
Zone
Permit-
ted Di-
vision
Oil immersion Ex o 1 & 2
Pressurization Ex p 1 & 2 1 & 2
Powder filling Ex q 1 & 2
Flameproof Ex d 1 & 2
Explosion Proof 1 & 2
Increased safety
Intrinsic safety Ex ia 0,1 & 2 1 & 2
Intrinsic safety Ex ib 1 & 2
Encapsulation Ex m 1 & 2
Special protection Ex s 0,1 & 2
Nonincendive 2
Nonsparking Ex nA 2
Enclosed break Ex nC 2
Energy limited Ex nL 2
Simplified
pressurization
Ex nP 2
Restricted
breathing
Ex nR 2
Table 7. Types of Ignition Protection for Mechanical Equipment (ATEX)
Method Description
To ensure that ignition sources cannot arise Construction safety c, Inherent safety g,
To ensure that ignition sources cannot become active Control of ignition sources b
To prevent the explosive atmosphere from reaching
the ignition source
Inert liquid immersion k, Inert gas pressuriza-
tion p, Flow restricting enclosure fr
To contain the explosion and prevent flame propagation Flame proof enclosures d, Flame arresters
Table 1. Hazardous Areas*
North America IEC (Europe)
Class Division Zones
Class I Gas or
vapor
Class II Dust
Class III Fiber
or flying (no group
designation)
Division 1: Pres-
ent or likely to be
present in normal
operation
Division 2: Not
present in normal
operation, could
be present in ab-
normal operation
Gas/Vapor
or Dust
Zone 0 (Gas) /
Zone 20 (Dust)
Zone 1 (Gas) /
Zone 21 (Dust)
Zone 2 (Gas) /
Zone 22 (Dust)
An area in which an explosive atmosphere is
continually present or present for long
periods or frequently
An area in which an explosive atmosphere is
likely to occur in normal operation
An area in which an explosive atmosphere is
not likely to occur in normal operations and,
if it does occur, will exist for only a short time
* This table represents a corrected version from that in the original printing
































































































tep 2 etermine sys
tem chrcteristic cure








percent process ribIe %)



tep 5 eeIop instIIed gin grph

tep 6 Interpret resuIts
























eerences






225
2
175
15
125
1
75
1 5 1, 1,5 2,
Iow, gImin
ximum
ormI
inimum
1 or Ie
2 or Ie
ystem pressure chrcteristic
2,5 , ,5 ,
22
2
18
16
1
12
1
8
6

1 2
Ie treI, %
ximum
gin
inimum
gin
InstIIed gin
5 6 7 8 9 1
,
2,5
2,
1,5
1,
5

1 2
Ie treI, %
ximum
ormI
inimum
5 6 7 8 9 1
InstIIed Iow chrcteristic



Department Editor: Scott Jenkins
Infrared
Temperature
Measurement
C
ontact-based temperature sensors, such as
thermocouples and resistance temperature
detectors (RTDs), have demonstrated ac-
curate and cost-effective operation throughout
the chemical process industries (CPI). However,
there are many applications and settings
where they are simply not practical. In those
cases, engineers can turn to a host of non-
contact temperature measurement devices,
many of which are based on measuring
infrared (IR) radiation. IR thermometers can
routinely perform measurements in situations
where readings with contact thermometers
would be virtually impossible. Situations where
IR-based temperature measurement should be
considered include the following:
* Ropld ond lrequenl lemperolure reodlngs
ore requlred
* Cb|ecls ore movlng
* Cb|ecls ore wllhln slrong eleclromognellc
elds, such as in processes involving induc-
tion or microwave heating
* Ropld lhermol chonges ore lhe norm
* Cb|ecls needlng meosuremenl ore locoled
in process chambers or behind windows
* lhe surloce lemperolure ol on ob|ecl or
equlpmenl requlres meosuremenl
* Locollons ore lnoccesslble lo conlocl
thermometers
* Areos lo be meosured would be domoged
or contaminated by contact measurement
* Vorylng surlocelemperolure dlslrlbullons ore
present
* Cb|ecls ore mode lrom molerlols wllh low
heat capacity and low thermal conductivity
* Molerlols lo be meosured ore goseous, such
as combustion gases and ames
Within the CPI, IR thermometry is most ef-
lecllve ond used mosl lrequenlly ln semlcon-
ductor and wafer processing, cement and
lime processing, rotating kiln shells, waste
incineration, glass processing, sintering and
heat treating, metals processing and drying
applications. While IR thermometers are gen-
erally more expensive than contact thermom-
eters, they usually have longer lifetimes and
requlre less molnlenonce.
A generol knowledge ol lhe key ospecls ol
thermal radiation physics can help users apply
and operate the devices more effectively.
Stefan-Boltzmann law
lhe energy rodloled by on ob|ecl per unll
surface area per unit time is related to its
temperature by the Stefan-Boltzmann law, which
states that irradiance (in J/s/m
2
) is proportional
lo lhe lourlh power ol lhe ob|ecls lemperolure
mulllplled by lls emlsslvlly. A conslonl ol propor-
tionality (the Stefan-Boltzmann constant) is re-
qulred lor colculollons. ln o perlecl [lheorellcol|
blockbody rodlolor, emlsslvlly equols one.
Operation
Mosl lR lhermomelers operole
according to the same basic
operating principles (Figure 1).
A bondwldlh hller ond opllcol
lens are used to focus the IR
energy emllled by on ob|ecl
onto a detector, which converts
the IR radiation into an electri-
col slgnol. Aller compensollng
for emissivity (see below) and
ambient temperature, an analog
output is generated to provide
temperature measurement. The
analog signal can be converted
lo dlgllol when losl ocqulslllon
roles ore requlred.
Emissivity
The emission of thermal radia-
tion is a surface phenomenon
for most materials. The term
emlsslvlly relers lo on ob|ecls
ability to emit thermal radiation.
Emissivity is dened as the ratio between the
energy emllled by on ob|ecl ol o glven lem-
perature and a perfect radiator, or blackbody,
at the same temperature. Emissivity values
lie between zero and one. IR thermometers
generally have the ability to compensate for
the different emissivity values of materials.
Molerlols wllh lhe hlghesl emlsslvlly volues
are the easiest to measure accurately with IR
thermometers, while those with low emissivities
are more difcult. For example, some polished,
shiny metallic surfaces, such as aluminum, are
so reective in the infrared that accurate tem-
perature measurement is not always possible.
Tables listing emissivity values for various
materials have been published, and are avail-
able for reference (Table). Some IR thermom-
eters allow users to change emissivity values
according to the material being measured,
while others have a pre-set emissivity value.
When using IR thermometry, it is important
to consider that materials can have different
emissivity values at different wavelengths. To
determine an emissivity value, you can heat a
molerlol lo o known lemperolure, lhen od|usl
the emissivity value of the instrument until the IR
thermometer matches the known temperature.
Field of view (FOV)
lhe FCV ls lhe ongle ol vlslon ol whlch lhe
instrument operates and is determined by the
optics of the system. The optical system of the
IR thermometer collects the IR energy from a
circular measurement spot, and focuses the
energy on the detector. The optical resolution
of the instrument is determined by the ratio
belween lhe dlslonce lrom lnslrumenl lo ob|ecl
and the size of the spot being measured
(distance-to-spot ratio). Higher ratios mean
better resolution.
Ideally, the target being measured should ll
lhe lnslrumenls FCV. For lhe besl perlor-
monce, lhe lorgel ob|ecl oreo should exceed
lhe FCV by o loclor ol oboul 1. [Flgure 2|.
Calibration
IR thermometers can be calibrated by aiming
at blackbody radiators that are designed
specically for calibration and testing. By vary-
ing the source temperature of the blackbody,
collbrolors con lune lhe lR sensors lnlernol
measurement signal to known temperatures.
Selection questions to consider
When selecting an IR thermometer for a CPI
application, it is important to consider the fol-
lowlng quesllons:
* Vhol ls lhe requlred proxlmlly lo lhe lorgel
* Vhol ls lhe slze ol ob|ecl lo be meosured,
ond wlll ll hl lhe lnslrumenls FCV
* Are lhere ony physlcol obslrucllons belween
lhe ob|ecl lo be meosured ond lhe meosur-
ing device
* Are smoke, dusl or olher porllculole moller
llkely lo be ln lhe meosuremenl oreo
* Vhol ore lhe meosuremenl conlrol lre-
quency requlremenls
* Does lhe ob|ecl hove o shlny surloce
Vhol ore lhe oulpul,lnlerloce requlremenls *
References
1. Young, A., lR lhermomelry Flnds CPl Mlches,
Chem. Eng., Feb. 2002, pp. 0.
2. Cmego Englneerlng lnc. |lnlernel]. c2003
2010. Slomlord [Cl|. |clled Feb. 2011].
Avolloble lrom: hllp:,,www.omego.com,.
EMISSIVITY VALUES OF COMMON MATERIALS*
Material Emissivity
Silver (polished) 0.01
Aluminum (unoxidized) 0.02
Gold (polished) 0.02
Aluminum (heavily oxidized) 0.20
Zinc (bright galvanized) 0.23
Steel (316 polished) 0.28
Soil (plowed eld) 0.38
Iron (liquid) 0.43
Iron (rusted) 0.65
Water 0.67
Sand 0.76
Steel (cold rolled) 0.80
Wood (oak planed) 0.91
Brick (red, rough) 0.93
Carbon (Lampblack) 0.95
Ice 0.98
*Provided for illustrative purposes only.
Filter
Lens assembly
Aperture
Detector
Amplifier
and electronics
Object
Field of view
S
D
Target greater
than spot size
Recommended Incorrect
IR Thermometer
Target equal
to spot size
Target smaller
than spot size Backround
FIGURE 1. The optical system of an IR thermometer collects IR en-
ergy from a circular measurement spot, and focuses it on a detector
FIGURE 2. For accurate IR temperature measurement, the target area
should be greater than the instruments FOV by a factor of about 1.5















mittingcthode ioniztion gges











eIecting Is


eerences


Department Editor: Scott Jenkins
Pressure
measurement
considerations
P
ressure measurement in the chemical
process industries (CPI) is crucial to
many unit operations, and selecting the
most effective pressure sensors for a given
situation can be complicated by a range of
factors. An initial key to selection is establish-
ing an accurate understanding of exactly
what is meant when the term pressure is
used, since there are different types. Other
critical considerations include the following:
media compatibility, environment, process
control, electrical isolation and output signal.
Pressure types
Pressure measurements can be affected by
what type of pressure sensing equipment is
used, and understanding the different types
of pressure is a prerequisite for selecting sen-
sors or gages for your application. Accuracy
can suffer if pressure types are misunder-
stood. Differences in pressure types have
everything to do with the reference point for
a given pressure measurement. Here are
denitions for ve common pressure types:
Gage pressure Gage pressure, the type
that most people rst imagine when thinking
of measuring pressure, covers a positive
pressure range. Its zero (reference) point is set
at ambient pressure, and it is unaffected by
changes in barometric pressure because the
sensor is open to the atmosphere. This allows
the current atmospheric pressure to be the ref-
erence against which all subsequent changes
in pressure are measured. Gage pressure ef-
fectively can measure pressures below 1 psi,
as well as pressures up to 200,000 psi.
Vacuum pressure Like gage pressure,
vacuum pressures zero point is ambient pres-
sure, and sensors measuring it are vented
and therefore unaffected by barometric
change. Since vacuum pressure refers to
a negative pressure range, the distinction
between vacuum and gage pressure is really
a function of direction and magnitude. Sen-
sors measuring this type are commonly used
in vacuum pump systems and applications
where suction is required.
Compound gage pressure This pres-
sure type is the combination of gage and
vacuum pressure in that it involves both
positive and negative pressure changes.
Its zero is therefore set at atmospheric
pressure, and it is vented. The value of a
compound gage is seen when used in ap-
plications where the pressure uctuates from
positive to negative and vice-versa. Sensors
measuring this pressure type typically do
not exceed 100 psi in range.
Sealed pressure Sealed pressure refers
to a situation where the pressure sensor is
not vented. This is primarily done to protect
the sensor, by avoiding the introduction of
moisture or dust into the sensor housing.
The sensor is sealed with a pressure equal
to the atmospheric pressure at the time of
sealing. This pressure then becomes the ref-
erence pressure against which all pressure
changes are measured.
Because it is sealed, unvented pressure
sensors are unavoidably affected by baro-
metric pressure changes. It is not typically
used in low-pressure applications because
the barometric shift of a few psi would affect
measurement accuracy signicantly. How-
ever, at 1,000 psi and above, the relatively
small shift would go unnoticed and can be
smaller than the error band of the sensor.
In one real-world case, a sealed pressure
type sensor was calibrated at a manufac-
turing facility in Utah and then shipped to
Indiana. The atmospheric pressure differ-
ences between the locations caused the unit
to fail in Indiana, while it worked properly
in Utah.
Absolute pressure Absolute pressure
is used when the zero point must be set to
absolute zero. To achieve this, the sensor is
also sealed, but under a vacuum condition, so
that air molecules are removed from the enclo-
sure. This then becomes the reference point
and allows measurements to be made with
reference to absolute zero. By denition and
design, this is sensitive to barometric changes.
Unlike sealed pressure, absolute pressure
is often used in low-pressure applications
measuring atmospheric conditions, such as in
weather stations, aircraft and laboratories.
Notes
Material on pressure types was contributed by
Elden Tolman, product design engineer at Automa-
tion Products Group Inc. (APG; Logan, Utah;
www.apgsensors.com).
MAIN FACTORS TO CONSIDER FOR CHEMICAL PRESSURE MEASUREMENTS
Factor Reason Solution
Media
compatibility
A pressure-sensing element will
come in contact with varying
concentrations of chemicals,
temperatures and pressure ranges
depending on the industry sector
in which the application appears,
including petrochemicals, food,
pharmaceuticals, water, refrigera-
tion, alternative energy or power
generation
Pressure sensors constructed from
one-piece 316L stainless steel,
nickel and cobalt-based superal-
loys are free from internal welds,
O-rings and very thin isolation
diaphragms offer excellent media
compatibility for most chemicals.
The one-piece design ensures
outside media do not permeate
the sensor body
Process
control
New processes for heavy oil,
alternative energy and water puri-
cation systems demand extreme
operating conditions, such as low
ambient temperatures (50C),
high media temperature (150C),
as well as complex and volatile
gas-liquid mixtures
Pressure sensors with new tech-
nologies and wetted materials are
needed. High-temperature, oil-free,
bulk silicon piezoresistive sensors
are ideal for these emerging mar-
kets. Superalloys, such as Inconel,
Hastelloy and Waspalloy, with thick
sensing diaphragms, offer the
best solutions without the need for
complex sensor packaging and
expensive secondary seals
Environment Rain, ice, dust and pressure
washers can cause water to seep
into sensor housings and cause
electronics to shortcircuit
Absolute and sealed-gage refer-
ence pressure sensors protect
electronics from these conditions.
If venting is required to maintain
accuracy at low pressures, provi-
sions must be made for dry, non-
corrosive environments for sensors
to breathe
Electrical
isolation
Improper grounding and lightning
strikes can cause electrical
failures of pressure sensors, as a
result of isolation failure
Pressure sensors with custom
electronics and a sensing element
able to withstand 500 V d.c. isola-
tion can work in extreme electri-
cal conditions
Output signal Depending on distance and en-
vironment, certain output signals
can experience signal loss or
generate noisy signals
A 420-mA output signal is recom-
mended for transmission lengths
greater than 15 ft in environments
with electrical noise
Table content submitted by Karmjit Sidhu, vice president of business development at American Sensor
Technologies (Mount Olive, N.J.; www.astsensors.com)
Reference
Positive
pressure
Ambient
Negative
pressure
Gage Vacuum Compound Sealed Absolute
Source: APG
Department Editor: Kate Torzewski
Valves
gate
Gate valves are designed to
operate fully open or fully closed;
When fully opened, there is very
little pressure drop across the
valve, and when fully closed there
is good sealing against pressure.
With the proper mating of a disk
to the seat ring, very little or no
leakage occurs across the disk
when the gate valve is closed.
Gate valves open or close slowly,
which prevents fuid hammer and
subsequent damage to the piping
system. Gate valves are usually
classifed by the type of disk used,
and a variety of disk types are
available, such as solid wedge,
split wedge or parallel disk.
globe
The basic principle of globe valve
operation is the perpendicular
movement of the disk toward, or
away from, the seat. This causes
the annular space between the
disk and seat ring to gradually
close as the valve is closed. It is
this characteristic that gives the
globe valve good throttling ability.
When the valve is closed, there
is no blocked-in volume, as occurs
in a gate valve, so a globe valve
has much less leakage around
the seat. Also, the disk-to-seat-ring
contact is much closer to forming
right angles, so the force of clos-
ing tightly seats the disk.
pinch
This valve consists of a fexible
tube that is mechanically pinched
from the outside of the valve
body. The principal advantages of
this type of valve are that the fow
passage is straight without any
crevices, and there are no internal
moving parts. The soft valve body
has the ability to seal around
trapped solids, so pinch valves
are suitable for handling slurries
and solids, which would clog in
the obstructed fow passages of
other valve types. They are also
used for the sanitary handling of
foodstuffs and pharmaceuticals
because the media are isolated
from the working parts.
diaphragm
The fow passage in diaphragm
valves is free of crevices and
is unobstructed by moving
parts, making them suitable for
applications where cleanliness,
bubble-tight shutoff and chemical
compatibility are important. The
diaphragm valve is considered
to be the valve least likely to
cause contamination. For this
reason, it is popular in high-
purity applications.
It is available in two general
designs, weir and straightway.
The weir-style diaphragm valve is
utilized for higher-pressure applica-
tions. The straightway diaphragm
valve, having no fow path obstruc-
tions, is well suited for higher-fow
and slurry applications.
ball
This rotational-motion valve uses
a ball-shaped disk with a hole
bored through to stop or start fuid
fow. When the valve handle is
turned to the open position, the
ball is rotated so that the hole
lines up with the valve bodys
inlet and outlet. When the ball is
rotated so the hole is perpendicu-
lar to the fow, the valve is closed.
Because the ball moves across
the seats with a wiping motion,
ball valves can handle fuids with
suspended solids.
Ball valves are available in
Venturi, reduced and full-port
patterns. The full-port pattern has
a ball with a bore equal to the
inside diameter of pipe. Most ball
valves instead have a reduced
bore with a Venturi shaped fow
passage of about three quarters
the nominal valve size.
plug
Plug valves have a cylindrical or
tapered plug with a hole bored
through. As with ball valves, fuid
fows when the hole in the plug is
aligned with the pipe, and a quar-
ter turn of the plug stops the fow.
Plug valves often have fuoro-
carbon seating materials and in
some cases are fully lined with
fuorocarbons, which provides
excellent protection for corrosive
applications that require bubble-
tight shutoff. There are several
different types of plug valves
commonly used in the CPI, includ-
ing lubricated, nonlubricated and
eccentric types.
butterfly
Butterfy valves get their name from
the winglike action of the fow-con-
trolling disk that opens and closes
at right angles to the fow path.
Butterfy valves were introduced to
counteract the problems associated
with linear-valve designs (espe-
cially gate valves), such as the
relatively large size and weight,
the high operating force required,
and the tendency to leak. Instead
of a long stroke, the butterfy valve
requires a quarter turn to cycle
from a fully open to fully closed
position. Butterfy valves can be
used for both on/off and throttling
applications.
check
The purpose of a check (or non-
return) valve is to allow fuid fow
in one preferred direction and
to prevent back fow, or fow in
the opposite direction. Ideally, a
check valve will begin to close as
the pressure drops in a pipeline
and the fuid momentum slows.
When the fow direction reverses,
the check valve should close
completely. Check valves can be
of the following types: swing, lift
and tilting disk.
References
1.Sahoo, T., Pick the Right Valve,
Chem. Eng., August 2004, pp.
3439.
Gate
valve
Diaphragm
valve
Butterfly
valve
Ball
valve
Check
valve
Plug
valve
Globe
valve
Pinch
valve
Source for figures: Valve Manufacturers
Association of America, Washington, D.C.















f p



Department Editor: Rita L. D'Aquino


Materials of
Construction
Low-temperature
appLications [1]
One key engineering consideration is the
choice of materials of construction for frigid
applications. Nickel-chromium (Ni-Cr) type
stainless steels are notably versatile at low or
cryogenic temperatures. They offer a combi-
nation of high impact strength (IS) and corro-
sion resistance. In the austenitic phase, with
face-centered-cubic crystals, the combination
of Cr and Ni in the material improves IS and
toughness down to temperatures as low as
250C. For good IS at temperatures down
to 45C, C-Mn-Si steels are recommended.
The most preferred grades are fine-grained
steels of pressure-vessel quality, such as ASTM
A 516 and ASTM A 537 (in all grades). For
temperatures between 45 and 100C (for
example, for liquid-ethylene storage), steels
containing 2.59% Ni are useful. Between
150 and 250C, the Ni-Cr austenitic steels
(300 series, of 18/8 varieties), are highly
recommended. In the nonferrous category,
Al has excellent properties for temperatures
as low as 250C. Also attractive are Cu and
some of its alloys, which can withstand tem-
peratures down to 195C.
chemicaL resistance
CPVC [2]. Many nonmetals do not have the
tensile strength to meet the pressure require-
ments of various process applications, espe-
cially at elevated temperatures. But years of
testing and actual field performance prove
that chlorinated polyvinyl chloride (CPVC)
systems can be pressure rated for operation
as high as 200F. CPVCs high heat-distor-
tion temperature and resistance to corrosion
make it suitable for applications such as
metal processing, pulp and paper, and in-
dustrial wastewater treatment, where harsh
and corrosive chemicals are commonly used
(see Figure 1). Another advantage of CPVC
is that it is lighter than metal, and therefore
less expensive to install, from both a mate-
rial cost and labor perspective. CPVC is not
recommended where aromatic solvents and
esters are present in high concentrations.
FRP pipe [3]. Composite fiberglass-rein-
forced plastic (FRP) pipe has been replac-
ing conventional pipe material, such as
steel and concrete, in numerous applica-
tions because of its corrosion resistance,
low design weight (25% of concrete pipe
and 10% of steel pipe), high fatigue en-
durance, and adaptability to numerous
composite blends (Table 5, Ref. 3) and
manufacturing methods. FRP pipe may
be divided into two broad categories:
gravity pipe (dia. from 8 to 144 in.) and
pressure pipe (dia. from 1 to 16 in.). It is
not unusual to see FRP pressure pipe han-
dling pressures as high as 2,0005,000
psi during chemical processing, with the
higher-pressure pipe at the lower end of
the diameter scale.
heat transfer
properties [4]
Metals, including specialty materials, are
the best choice in terms of good heat trans-
fer. In the lined category, glass is used ex-
tensively for process equipment where good
heat transfer is required. Lined materials,
however, often have the problem of uneven
thermal expansion, which may weaken the
bonding of the lining in due course. While
fluoropolymers have excellent compatibility
with various chemicals and special surface
and physical chemistries, they are gener-
ally not used for reaction vessels because
of their poor heat-transfer properties. Ther-
mal conductivities for various materials are
listed in the Table, and typical applications
are shown in Figure 2.
THERMAL CONDUCTIVITY OF VARIOUS
MATERIALS OF CONSTRUCTION [4]
Material Thermal conduc-
tivity, W/(m)(K)
Carbon Steel (CS) 60.59
SS 304 40.71
SS 316 14.23
SS 316 L 14.23
Hastelloy B
2
9.12
Hastelloy C
2
10.21
Tantalum
2
57.5
Titanium
2
21.67
Zirconium
2
20.77
Graphite 121.15
Hexoloy 125.65
Glass
1
1.00
Lead 35.30
Inconel
2
12.00
CPVC 0.14
PTFE (Polytetra-
fluoroethylene)
1
0.25
PFA (Perfluoro-
alkoxy resin)
1
0.19
ETFE (Ethylene
tetrafluoroethylene)
1
0.24
PVDF (Polyvinylidene
fluoride)
1
0.23
ECTFE (Ethylene chlo-
rotrifluoroethylene)
1
0.16
1. Common choice for lining material
2. Exotic metals
Excellent
Good
Fair
Poor
Weak acids
Weak bases
Salts
Strong acids
Aliphatics
Strong bases
Strong oxidants
Halogens
Aromatic solvents
Esters and ketones
Figure 1. CPVC offers resistance to a
variety of harsh chemicals
Exotic 300
250
200
150
100
50
0
-50
-100
T
e
m
p
e
r
a
t
u
r
e
,

C
Exotic Exotic
Exotic
Fluoropolymer,
glass lined,
exotic
Fluoropolymer,
glass lined,
exotic
Glass lined,*
exotic
Glass lined,*
exotic
Exotic
Exotic Exotic
Exotic
Exotic
Exotic
Application:
Typical
equipment:
Tanks,
vessels
Pipelines,
valves,
owmeters
Mixers
Reactors
Storage Transport Agitation (Agitation +
heat transfer)
Exotic Exotic Exotic
Figure 2. When looking beyond steel for materials of construction, it is impor-
tant to consider the intended application and temperature range. Exotic (specialty)
metals (see Table) are shown here to serve well in all applications. Another mate-
rial, equally suited to a specific requirement, however, may be chosen as the more
cost-effective option
References
1. Nalli, K., Materials of Construction For Low-Tem-
perature and Cryogenic Processes, Chem. Eng.
July 2006, pp. 4447.
2. Newby, R. and Knight, M., Specifying CPVC In
Chemical Process Environments, Chem. Eng., Oc-
tober 2006, pp. 3438.
3. Beckwith, S., and Greenwood, M., Dont Over-
look Composite FRP Pipe, Chem. Eng., May
2006, pp. 42-48.
4. Robert, J., Selecting Materials of Construction,
Chem. Eng., September 2005, pp. 6062.
Department Editor: Kate Torzewski
Pristine
Processing
Equipment
P
rocesses in the pharmaceutical, biotechnology, food and semiconductor industries must
meet a high set of standards to ensure high product purity. Equipment criteria specifc to
high-purity processes are established to minimize contamination and maintain prod-
uct integrity. In designing a pristine process, material and equipment style are of upmost
importance. Bacteria is the main cause of contamination and is prone to growing in the dead
cavities of equipment created by sharp corners, crevices, seams and rough surfaces. Another
source of contamination is leaking, which allows undesirable chemicals to compromise the
quality of the process ingredients, by causing contamination, rusting and particle generation.
MATERIALS OF CONSTRUCTION
Many factors must be taken into consideration when selecting materials of construction for
use in pristine process applications where high-purity and sanitation are paramount. All sur-
faces should be constructed of a smooth material that will not corrode, generate particles
or harbor dead cavities. These criteria can be met with three standard materials: 316L
stainless steel (SS), polyvinylidene fuoride (PVDF) and polytetrafuoroethylene (PTFE). The
advantages and disadvantages of these materials are summarized below to facilitate the
material selection process for a given application with consideration of chemical compat-
ibility, cost, and temperature stability.
Valves
Valves should not harbor contaminants and
must be easy to clean. By these criteria,
diaphragm and pinch valves are excellent
choices for ultrapure processes, as they
have smooth, gently curved surfaces that
will not harbor contaminants. Ball check,
full-port plug and full-port ball valves are
good choices as well, while butterfy, spring
check, gate and swing check valves are all
unacceptable, since contamination can col-
lect in the corners that are essential to their
design. Though several valves are appropri-
ate for pristine processes, certain valves
are better suited for particular applications.
Diaphragm valves are the most widely used
in high-purity systems for their resistance
to contamination and ability to be used as
a control valve. Ball and plug valves, on
the other hand, are less costly and are not
limited by temperature and pressure. Also,
in applications using sterile steam and
freeze-drying, ball valves are preferred over
diaphragm valves because they eliminate
the risk of catastrophic seat failure.
Seals
As with all pristine processing equipment,
high-purity seals should not have any
cavities where contaminants can breed. By
choosing a seal with gland rings that do not
need to be threaded or ported, the areas
where bacteria can breed are minimized. In
choosing a seal material, it is important to
fnd a compound that will not swell, crack,
pit or fake, thus reducing seal failure and
contamination. To ensure the success of
seals, furoelastomers are a top choice in
pristine processing applications for their ex-
cellent thermal stability, chemical resistance
and mechanical durability.
Piping
The surface of piping, as well as any wet-
ted equipment parts, should have a very
smooth surface. When 316 SS is being
used, electropolishing is a good method
for achieving an ultra-smooth fnish. Joining
methods should minimize crevices and dead
cavities, and all materials should be free of
biological degradable substances, leachable
substances, and glues and solvents that may
migrate into the product stream.
References
1. Smith, B., What Makes a Pump for High-Purity
Fluids?, Chem. Eng., pp. 8789, April 2002.
2. Schmidt, M., Selecting Clean Valves, Chem.
Eng., pp. 107111, June 2001.
3. Wulf, B., Pristine Processing: Designing Sanitary
Systems, Chem. Eng., pp. 7679, Nov. 1996.
4. Weeks, D. T. and Bennett, T., Specifying Equip-
ment for High-Purity Process Flow, Chem. Eng.,
pp. 2730, Aug. 2006.
EQUIPMENT STYLE SELECTION
Critical factors in high-purity equipment selection include cleanability, cost, fow capabili-
ties and product compatibility. With these considerations in mind, criteria useful for choos-
ing pumps, valves, seals and piping are described in this section.
Pumps
A fundamental requirement of pristine processing pumps is the ability to clean a pump in
place without disassembly. Pump seals, gaskets and internal surfaces should eliminate the
buildup of material and should clean out easily during wash cycles. The most common pump
styles for high-purity processes are centrifugal, lobe-style and peristaltic pumps, which are
outlined below.
MATERIALS OF CONSTRUCTION
Material Advantages Disadvantages
Stainless
Steel
Mechanical strength
Functions at 121C
(steam-sterilization temperature)
Vulnerability to corrosion by
certain chemicals, which
increases with temperature
PVDF
Chemically inert
Resistant to corrosion and leaching
Durable and long-lasting
Retains circumferential strength
Functions only intermittently
at 121C
PTFE
The most chemically inert plastic
Resistant to corrosion and leaching
Avoids leaching
Suitable for coating equipment
Complex shapes are
diffcult to form
PUMPS
Pump
style
Advantages Disadvantages
Applications best
suited for this style
Centrifugal
Low cost
Easy cleanability
Effciency and fow de-
crease with increasing
pressure and volume
Handling low-
viscosity products
Handling high fowrates
(401,500 gal/min)
Peristaltics
Low cost
Easy cleanability
No mechanical
seals
Non-damaging to
delicate products
The need for hoses may
cause issues in elastomeric
compatibility, temperature
and pressure limitations,
and a need to change
hose regularly
Small, batch-oriented
applications
Laboratory or pilot-
scale plants
Rotary
Lobes
Higher pressure
and fow
capabilities
Unaffected by
pressure variations
High cost
Large, continuous
duty applications
Steaming and high
pressure applications
Department Editor: Scott Jenkins
Controlling
Membrane Fouling
T
he deposition and accumulation of suspended and dissolved
particles on membrane surfaces leads to performance loss.
Fouling can dramatically reduce the efciency and econom-
ic benets of a membrane process. The type of fouling and how
strongly it appears depends on several parameters, including
the following:
* Molure ol solules ond solvenls
* Membrone process
* Poreslze dlslrlbullon
* Membrone surloce choroclerlsllcs ond molerlol ol conslrucllon
* Hydrodynomlcs ol lhe membrone module
* Process condlllons
COMMON FOULING MODES
Membrone loullng con be o conloundlng problem ln woler lreol-
ment systems. Several of the most common types of fouling are
shown in the following table.
REDUCING FOULING
Infuence of the bulk solution
Properlles ol lhe bulk solullon con ollecl membrone loullng, bul
whether these properties can be manipulated depends on the
actual process conditions.
Ad|usllng pH, vorylng lemperolure ond chonglng porllcle slze
are possible parameter changes that may be manipulated to
lnNuence lhe loullng behovlor ol o bulk solullon. For exomple,
operollon ol low pH con help prevenl scollng ol colclum sullole
by changing the solubility of the salts. This can signicantly
reduce the precipitation of calcium sulfate on the membrane.
Concentration polarization
Concenlrollon polorlzollon ls lhe occumulollon ol re|ecled por-
ticles, especially during microltration and ultraltration, to the
exlenl lhol lronsporl lo lhe membrone surloce becomes llmlled.
Hlgh Nux lhrough lhe membrone con couse re|ecled porllcles
to accumulate on the surface of the membrane. Concentration
polorlzollon reduces lhe permeoblllly ol lhe solvenl ond con
leod lo o llmlllng Nux, where on lncreose ln pressure does nol
correspond lo o rlse ln Nux.
Reduclng concenlrollon polorlzollon leods lo hlgher llmlllng
Nux ond lower loullng lendency. Conlrolllng concenlrollon polor-
lzollon ls essenllol lor o process lo be economlcolly benehclol.
Concenlrollon polorlzollon con be conlrolled by lncreoslng
crossow velocity or by increasing turbulence on the membrane
surloce, whlch lncreoses lhe bock lronsporl ol lhe porllcle owoy
from the membrane. The challenge becomes balancing high
Nuxes ond low loullng wllh low lnveslmenl ond operollng cosls.
Membrane properties
As an essential part of the membrane process, the membrane
itself has a strong inuence on fouling. Typically, hydrophilic
membrones ore speclhed becouse lhey exhlbll on olhnlly lor
water, which is one of the main tools used to reduce the ad-
sorption of foulants onto the membrane surface. A hydrophilic
membrone ls surrounded by woler molecules, whlch work os o
protective layer. The hydrophilicity and hydrophobicity of some
polymeric membrane materials are shown in Table 2.
Because hydrophilic membranes have lower chemical resistance
than hydrophobic ones, their chemical stability and cleanability
have to be evaluated as part of the selection process. It should
be noted that most membranes are polymer blends.
Mosl polyelhersullone [PES| membrones conloln some poly-
vlnylpyrolldone [PVP| lo lncreose hydrophlllclly. lhe problem ls
lhol PVP ls nol sloble ogolnsl oxldlzlng ogenls, whlch moy leod
to changes in membrane porosity if not closely monitored.
Porosity
Poroslly con reduce loullng durlng mlcrohllrollon ond ullrohllro-
llon. lhe slrongesl loullng ls coused by lhe blocklng ol membrone
pores. lherelore, lhelr pore slze should be smoller lhon lhe
overoge porllcle slze, ond o membrone wllh o norrow poreslze
dlslrlbullon ls prelerred lo ovold lhe blocklng ol blgger pores.
Infuence of the permeate fux
Crlllcol Nux ls onolher loclor lhol con be lnNuenced lo reduce
fouling in microltration, ultraltration and nanoltration
processes. ll ls dehned os lhe Nux below whlch o decllne ln
Nux wllh llme does nol occur, whlle obove lhls Nux, loullng
slorls. Al crlllcol Nux, lhe number ol porllcles lronsporled lo lhe
membrane is similar to the number of particles that diffuse away
lrom lhe membrone. ln operollon below crlllcol Nux, lhe Nuxes
ore reverslble, whlch meons lhol os long os Nux ls below lhe
crlllcol Nux, membrone permeoblllly ls nol chonged by loullng.
lhe odvonloges ol crlllcol Nux ore lhol conslonl Nuxes ond
membrane properties can be sustained for longer periods. On
lhe olher hond, lhe condlllons ol crlllcol Nux requlre lower pres-
sure, hence o hlgher membrone oreo ls requlred. lherelore,
the tradeoff between higher investment costs and lower fouling
tendencies has to be assessed for each process.
lhe molerlol ln lhls column ls odoped lrom lhe lollowlng orllcle: Llpnlzkl, ).,
Slrolegles lor Conlrolllng Membrone Foullng, Chem. Eng., September 2007,
pp. 6264.
POLYMERIC MEMBRANE PROPERTIES
Property Polymer
Hydrophobic Polytetrauoroethylene (PTFE, Teon)
Polyvinylideneuoride (PVDF)
Polypropylene (PP)
Polyethylene (PE)
Hydrophilic Regenerated cellulose
Cellulose ester
Polycarbonate (PC)
Polysulfone/polyethersulfone (PS/PES)
Polyimide/polyetherimide (PI/PEI)
(aliphatic) Polyamide (PA)
Polyetheretherketone (PEEK)
Cellulose triacetate (CTA)
FOULING EXAMPLES
Foulants Fouling mode
Large
suspended
particles
Particles present in the original feed or
developed in the process can block module
channels
Small colloi-
dal particles
Colloidal particles can raise a fouling layer or
block the porous structure of the membrane
Macro-
molecules
Gel-like cake formation on top of the mem-
brane or macromolecular fouling within the
structure of porous membranes
Small
molecules
Molecules such as substituted aromatics can
adsorb onto the membrane structure and
reduce the water content of the membrane,
which lowers permeability
Scalants Depending on the pH, salt may precipitate on
the membrane. This reduces the membrane
area and may reduce the water content in the
membrane
Biological
material
Growth of bacteria on the membrane surface,
which leads to a gel-like cake on the membrane
Department Editor: Kate Torzewski
Membranes
S
eparation by a membrane is achieved by
creating a boundary between different
bulk gas or liquid mixtures. As different
solvents and solutes fow through a membrane
at different rates, separation is achieved.
Here, we will focus on three fltration
techniques: microfltration (MF), ultrafltration
(UF) and nanofltration (NF). These processes
are characterized by the size of the particle
that can be separated by the membrane, as
illustrated in the fgure. Each membrane type
is best suited for unique applications and is
designed with the module and material that
will allow the best separation.
Flow through a membrane is characterized
as either tangential fow fltration (TFF), where
the feed stream fows at a velocity vector nor-
mal to the membrane surface, or normal fow
fltration (NFF), where the stream fows tangent
to the membrane surface. The fow pattern is
dependent on the type of module utilized. NFF
modules include: cartridges, stacked disks
and fat sheets. TFF modules include: plate-
and-frame (cassettes), hollow fbers, tubes,
monoliths, spirals and vortex fow.
microfiltration
MF separates particles from true solutions. This
technique is able to separate particles from
about 0.1 to as high as 10 m. As can be
seen from the fgure, large, soluble macromol-
ecules, bacteria and other microorganisms can
be retained by MF membranes.
Membrane materials
MF membranes have the largest pore open-
ings of any other membrane. Typically, they
can be classifed as having tortuous or capil-
lary pores.
From solids. When membranes are made
by sintering or agglomeration of micropar-
ticles, pores are formed by the interstices
between solid particles. Common materi-
als include: metal, metal oxide, graphite,
ceramic and polymer.
Ceramic. These membranes are typically
created by the sol-gel process, which is the
successive deposition of smaller ceramic
precursor spheres, followed by fring to form
multitube monoliths.
Track etched. A polymer flm is exposed to
a collimated beam of radiation that breaks
chemical bonds in the polymer chains. The
flm is then etched in a bath that selectively at-
tacks the damaged polymer, a technique that
produces a flm with photogenic pores.
Chemical phase inversion. A solution of a
concentrated polymer in solvent is spread into a
thin flm, then precipitated through the slow ad-
dition of a nonsolvent to produce tortuous pores.
Thermal phase inversion. A solution of
polymer in a poor solvent is prepared at an
elevated temperature. After being formed into
its fnal shaped, the temperature is dropped
and the polymer precipitates, and the solvent
is washed out.
Streched polymers. Semicrystalline polymers,
which are stretched perpendicular to the axis
of crystallite orientation, fracture in such a way
that reproducible microchannels are made.
Membrane modules
Many conventional designs are used in MF,
including cartridge-flter housing, plate-and-
frame-type devices, capillary bundles, tubular
membranes, spiral-wound modules and belt
flters. Ceramic MF membranes are available
as fat sheet, single tubes, disc, and other forms,
primarily for lab use. Finally, cassettes are two
different cross-fow membrane devices.
ultrafiltration
UF membranes, with pore sizes ranging from
about 1 to 100 nm in diameter, employ pres-
sure driving forces of 0.21.0 MPa. This tech-
nique drives liquid solvents and small solutes
through the membrane, while retaining larger
particles, like large dissolved molecules, col-
loids and suspended solids.
Membrane materials
UF membranes are typically made of polymeric
structures, such as polyethersulfone, regener-
ated cellulose, polysulfone, polyamide, poly-
acrylonitrile or various furopolymers. They are
formed by immersion casting on a web or as a
composite on an MF membrane. Membrane se-
lection is based on molecular-weight rating for
high yields, chemical and mechanical robust-
ness during product processing and Clean In
Place, and process fux for sizing and costing.
Membrane modules
Modules include cassettes, spirals, hollow
fbers, tubes, fat sheets, and inorganic mono-
liths. These primarily operate in TFF to increase
fux by reducing plugging. For virus removal
and water treatment, however, NFF operation
is run with cartridge and hollow fber modules.
nanofiltration
NF, sometimes referred to as loose RO
(reverse osmosis), utilizes a driving force
of 0.3 to 10.5 MPa to drive liquid solvents
through the membrane while retaining small
solutes of about 10 to 100 nm in diameter.
NF membranes are different from the mem-
branes previously discussed, because they are
usually charged, utilizing ion repulsion as a
major method of charged-species rejection.
They have 2080% NaCl retention and retain
> 2001,000 Daltons of neutral organics,
with a low retention of dissolved gases.
Neutral or undissociated solutes have a lower
retention than charged or dissociated solutes.
Membrane materials
Cellulose polymers. These are formed by
immersion casting of 3040% polymer lac-
quers, which can include cellulose acetate,
triacetate and acetate-butyrate, on a web
immersed in water.
Thin flm composites. Formed by interfacial
polymerization, TFCs involve coating a
microporous membrane substrate with an
aqueous prepolymer solution, then immers-
ing it in a water-immiscible solvent contain-
ing a reactant.
Crosslinked polyetherurea. Some of these
membranes feature NaCl retention and
water permeability.
Membrane modules
NF membrane modules are available in
spiral, hollow fber, tubular, and plate-and-
frame formats. Spirals are most common,
as they have low feed-side pressure props,
are less prone to clogging, are easily
cleaned, are mechanically robust, and are
most economical.
Aqueous salts
Colloidal silica Blue indigo dye
Asbestos fibers
Virus
Tobacco smoke Milled flour
Gelatin Coal dust
Latex emulsion
Albumin protein
Paint pigment
Bacteria
A.C. fine test dust
Yeast cells
Carbon black
Synthetic
dye
Red
blood
cells
Sugar
Pesticide
Ultrafiltration
Nanofiltration Microfiltration
Molecular Macromolecular Microparticle
Scanning electron microscope Optical microscope
0.001 Micrometers,
log scale
Typical
particles
Particle size
range
Process for
filtration
0.01 0.1 1.0 10
Facts at your Fingertips sponsored by:
References
1. Perrys Chemical Engineers Handbook, 8th ed. McGraw Hill, New York, 2008.
2. Seidel, A., ed. in chief, Separation Technology, second edition, John Wiley and Sons, Inc., New
Jersey, 2008.
Department Editor: Rita L. D'Aquino
T
he formation of crystals requires the birth of
new particles, also called nucleation, and
the growth of these particles to the fnal
product size. The driving force for both rates is
the degree of supersaturation, or the numerical
difference between the concentration of solute in
the supersaturated solution in which nucleation
and growth occurs vs. concentration of solute in
a solution that is theoretically in equilibrium with
the crystals.
In a batch crystallizer, the crystal size
distribution (CSD) is controlled by frst seeding
the initially supersaturated batch with a known
number and size distribution of crystals, and
then controlling the rate of evaporation or
cooling (i.e., rate of energy transfer) so as to
achieve a level of supersaturation that supports
adequate crystal growth and an acceptable
rate of nucleation. The relationship between
supersaturation and growth is linear, but that
between nucleation and growth is raised to a
power that is usually greater than one, making
it diffcult to grow large crystals when nucleation
is occurring. The following procedure describes
how to achieve the optimal growth rate:
1. Screen the seeds at the beginning of the ex-
periment to determine the cumulative number
of crystals that are greater than a given size
N. Estimate N
Li,
the number of crystals of a
given size (L
av
) obtained from the screening:

(1)

N
W
L k
L
i
av v c
i
=

3



The parameters are defned in the table
of nomenclature. To convert from m to ft,
multiply by 3.28 x 10
6
.
2. Continue to measure the number and size
of crystals as the cooling or evaporation
program is in progress. Prepare an inverse
cumulative plot of the number of crystals
greater than a given size vs. size of the
crystal (Figure 1). The crystal growth rate de-
pends on the energy transfer rate, so modify
the rate of energy transfer until a desirable
product is obtained.
3. Repeat the frst two steps at intervals
throughout the batch cycle and plot the
results as shown in Figure 1. The family of
curves resulting from data plotted under the
selected conditions indicates that the number
of crystals is not increasing with time. Thus,
no additional nucleation is occurring yet.
4. Proceed to collect crystal samples, an-
ticipating the onset of nucleation. Figure
2 indicates that the number of crystals is
signifcantly increasing with time. In this
fgure, t
1
(not to be confused with t
1
in Figure
1) represents the start of this new set of batch
dynamics. It is safe to assume that signifcant
nucleation is now occurring and that the rate
of energy transfer is too high.
5. By taking the slope of the curve represent-
ing the estimated number of nuclei present
at the measured point in time (N
ti
) vs. time
(t
i
), one can determine the nucleation rate.
Using your representation of Figure 3, create
a dashed, horizontal line across the lower
portion of the graph depicting the selected,
cumulative number of crystals (N
i
), and their
sizes (L
1
L
4
) over time (t
1
t
4
).
6. For a selected cumulative number of crystals
(N
i

), plot the crystal size (L) vs. time (t), as


demonstrated in Figure 3. The slopes repre-
sent the crystal growth rate (G). If the level of
supersaturation changes during the run, the
growth rate also changes. Non-parallel lines
would indicate that the larger crystals are
growing at a faster rate, due to a reduced
diffusional resistance [layer] at the crystal
surface. With larger particles, the resistance
layer may be smaller, allowing the solute to
more readily reach the crystal surface and
incorporate itself into the lattice. These fac-
tors collectively contribute to the accelerated
growth rate of the larger particles. Parallel
lines indicate that the growth rate is not
dependent on crystal size.
7. Increase the rate of cooling or evaporation
until additional nucleation occurs, upon
which you can safely assume that the growth
rate is too high.
8. Develop a seeding and evaporation profle
that will yield a growth rate that is lower than
the value found in Step 6.
When determining the growth rate, keep in
mind the difference in mixing characteristics
between a laboratory-scale vessel and a com-
mercial confguration. A small tank generally
offers a higher relative pumping capacity,
shorter blend time, and higher average shear
rates within a narrower range.
UsefUl observations
Most processors will agree that when it
comes to crystals, the larger, the better. Large
crystals are easier to handle in downstream
operations, such as washing, centrifugation
and drying.
As previously mentioned, it is desirable for
the seeds size distribution to refect a nar-
row cut of particles. In this cut, the weight
of crystals with sizes fner than Ls should be
minimal because these tiny particles add
enormously to the number of crystals that
compete for supersaturation and growth.
Studies show that milled seeds may not grow
as well as unmilled seeds. Furthermore, not
all crystals of a given size grow at the same
constant rate. This is sometimes attributed to
the differences in the surface characteristics
of particles that have equal dimensions.
Fines destruction in a batch system can
greatly reduce the effects of secondary nucle-
ation on the CSD, and signifcantly increase
crystal size while narrowing the CSD.
In practice, not all additional nucleation can
be suppressed. Crystallizations carried out at
low levels of supersaturation near the meta-
stable zone (i.e., the conditions under which
crystals grow, but do not typically nucleate)
will display some secondary nucleation, due
to crystal-crystal interactions and contact
between the crystals and the impeller. Nev-
ertheless, the mean crystal size, shape and
distribution are dramatically improved when
seeding is followed by a programmed rate of
energy transfer.
Reference: Genck, W., Better Growth in Batch Crystal-
lizers, Chem. Eng., Vol. 106, No. 8, pp. 9095, Aug.
2000. E-mail: genckintl@aol.com
/
h
n
u
m
b
e
r

o
f

c
r
y
s
t
a
I
s

I
a
r
g
e
r

t
h
a
n
-
Size, - (Mm)
-h = smallest measurable size
/h
1
U
1
-
1
-
2
-
3
U
2
U
3
U
4
-h
/
h

n
u
m
b
e
r

o
f

c
r
y
s
t
a
l
s

l
a
r
g
e
r

t
h
a
n

-
-h = smallest measurable size
/U
1
/U
2
/U
3
/U
4
U
1
U
2
U
3
U
4
-h
4
J
[
F

-
( =
/h
1
/h
2
/h
3
U
1
U
2
U
3
U
4
Time, U (min)
E-
EU
Controlling
Crystal Growth
NomeNclature
A Crystalsurfacearea,ft
2
B Nucleationrate,(numberofnuclei)/
ft
3
/s
G Crystalgrowthrate,m/s
k
v
Crystal-volumeshapefactor,
dimensionless
L Crystalsize,m
L Smallest-measurablesize,m
L
av
Sizeofcrystalfraction,m
L
f
Finalsizeofcrystal,m
L
s
Seedsize,m
N Numberofseeds
N
i
Constant,cumulativenumber
ofcrystalsincrystallizer
N
Li
Numberofcrystalsofagivensize,
L
av
N
ti
Numberofcrystalnucleiatanytime
S Rateofsupersaturation
S* Maximumallowablesupersatura-
tion,lb/ft
3
solvent
t, t
i
, t
f
Time,h
W
i
Weightofcrystalsonscreen

c
Crystaldensity,lb/ft
3
FiGure 1. FiGure 2. FiGure 3.
Department Editor: Kate Torzewski
Crystallization
C
rystallization is a method of solid-
fuid separation in which pure
chemical crystals are formed.
Crystallization kinetics consists of
three major phenomena: nucleation
(the birth of a crystal), transfer of the
solute from the supersaturated solution
to the crystal surface, and a reaction
during which the solute becomes incor-
porated into the crystal.
supersaturation
In order to drive the process of crystal-
lization, the solution must be supersatu-
rated with solute.
The solubility of the solution, related
to crystal size, is defned by the Kelvin
equation:
Supersaturation can be quantifed by
the ratio of the mass solute concen-
tration in the bulk solution to the
concentration in the solution at the
point of saturation. Alternatively, it
can be described by relative super-
saturation, which is calculated by the
following equation:
nucleation
Nucleation occurs at the point that a
crystal begins to form. The relative
rates of nucleation and growth are
critical to crystallization kinetics, as
they determine both crystal size and
size distribution. Nucleation is catego-
rized as primary if the supersaturated
solution is free of crystals, or second-
ary if the solution already contains
crystals. Primary nucleation requires a
higher level of supersaturation and is
the principal mechanism occurring in
precipitation. Secondary nucleation oc-
curs in commercial crystallizers, where
crystalline surfaces are present in order
to produce large crystals.
Primary nucleation
Primary nucleation, when occurring
homogeneously, takes place in a
supersaturated solution that is free of
foreign matter. First, molecules in the
solution will associate into a micro-
scopic cluster, which will either dis-
sociate or continue to grow. When the
cluster develops until it forms a lattice
structure, it is then called an embryo. A
stable crystalline nucleus is established
when the crystal size exceeds D
p
given
by the Kelvin equation for the solutions
specifc supersaturation ratio.
Combining the Kelvin equation with
laws of chemical kinetics gives the rate
of homogeneous nucleation, which is
described as follows:
Secondary nucleation
Industrial crystallizers typically rely on
secondary nucleation, which is caused
by the presence of existing crystals in
the supersaturated solution. This can
occur by one of three mechanisms in
which nuclei are removed from a crys-
tal surface, including the following:
fuid shear
collision of crystals with each other
collision of crystals with metal
surfaces
The phenomena of secondary nucle-
ation is too complex to derive a simple
kinetic theory, so an empirical power-
law function has been developed to
describe this process, which is based
on experimentally derived constants for
a particular system:
crystal growth
Based on data proving that a solution
in contact with a crystal is supersatu-
rated, there is an accepted two-step
theory of crystal growth, referred to as
the diffusion-reaction theory. In the frst
step, mass transfer of solute from the
solution to the crystal-solution interface
occurs. In the second step, the kinetic
step, a frst-order reaction occurs at the
surface of the crystal, during which sol-
ute molecules from the solution become
incorporated into the crystal-lattice
structure. By combining these steps, the
rate of mass transfer in crystal growth
is expressed as follows:
Through a series of calculations, this
expression can be used to defne the
particle size of a crystal at any time,
assuming that D
p0
<< D
p
:
References
1. Seader, J. D. and Henley, E. J., Separation
Process Principles, 2nd ed., John Wiley and
Sons, Inc., New Jersey, 2006.
Nomenclature
A Frequency factor
A
s
Surface area of crystal
B
0
Rate of homogeneous pri-
mary nucleation
c Mass solute concentration
in the bulk supersaturated
solution
c
s
Mass solute concentration
in the solution at satura-
tion
c/c
s
Supersaturation ratio
dm/dt Rate of mass deposited on
the crystal surface
D
p
Crystal diameter
k
c
Mass transfer coefficient
k
i
Kinetic coefficient
k
N
, b, j, r Constants determined ex-
perimentally
M
T
Mass of crystals per vol-
ume of magma
N Agitation rate
N
a
Avogadros number
R Gas Constant
s Relative supersaturation
t Time
T Temperature

s,L
Interfacial tension

s
Molar volume of crystals
Number of ions/molecule
of solute
Solute diffusivity
Solution density
Department Editor: Kate Torzewski
Humidity
Control
applications
Dehumidifcation by cooling or dessication
has a variety of applications, including:
Preventing moisture regain. Nearly all
materials have some affnity for moisture
based on surface characteristics and
the amount of surface exposed to humid
air. Moisture regain occurs when moist
particles stick together.
Preventing condensation. Air holds water
vapor in proportion to its temperature. Cold
surfaces of pipes, vessels, valves and heat
exchangers condense moisture unless the
air around them is dried to a dewpoint
below the temperature of the cold surface.
Preventing corrosion. The exposure of metal
surfaces to atmospheric corrosion can be
reduced by surrounding the surfaces with
dry air. Dehumidifers also keep humidity
low in process control rooms, preventing
the corrosion of electrical contacts and
sensitive electrical components.
Drying heat-sensitive products. Typically,
drying time is reduced by heating a prod-
uct. If the product is susceptible to damage
by heat, drying time can be reduced by
using dehumidifed air, which reduces the
vapor pressure of air above the wet surface.
cooling
A common method for dehumidifcation is
the use of air conditioning. Figure 1 shows
a typical vapor-compression cooling-based
dehumidifcation process. Air to be dried
passes through a cooling coil, which lowers
the temperature of the airstream below its
dewpoint. As the air cools, it loses its capac-
ity to hold water vapor. The water condenses
on the cooling coil surface, and falls to the
drain pan as liquid. The air is then drier in
absolute terms, but it also has a relative hu-
midity close to 100%. If low relative humidity
is needed in addition to a lower absolute
amount of moisture, the air can be heated
after it leaves the cooling coil.
For industrial purposes, cooling-based
dehumidifcation units are optimized for re-
moving moisture rather than removing heat.
These units provide deep cooling of small
amounts of air rather than slight cooling
of large amounts of air, condensing more
moisture. Standard refrigeration equipment
can produce dewpoints of +40F (4C) on a
reliable basis.
dessication
In a desiccant system, the process airstream
passes through a desiccant medium. The
desiccant adsorbs moisture directly from
the airstream. Desiccant dehumidifers can
produce dewpoints below 0F (18C)
a fvefold reduction in the air moisture
beyond what can be achieved with a stan-
dard-grade air conditioning system.
This equipment uses differences in vapor
pressure to remove moisture from air by
chemical attraction. The surface of dry
desiccant has a very low vapor pressure,
compared with the much higher vapor pres-
sure of humid air.
Water vapor moves out of the humid air
onto the desiccant surface to eliminate the
vapor pressure difference, as shown in
Figure 2. Eventually, the desiccant surface
collects enough water vapor to equal the
vapor pressure of the humid air. Then the
desiccant must be dried (reactivated) by ap-
plying heat before it is recycled to remove
more moisture from the air stream.
cooling vs. dessication
In most chemical process applications, both
technologies work best together. Cooling-
based dehumidifcation handles the mois-
ture load occurring at high dewpoints, and
desiccant-based dehumidifcation removes
the moisture load at low dewpoints. The op-
timal mix of the two technologies depends
on the characteristics of the application.
Factors to consider include the following:
Dewpoint control level. When the required
moisture-control level is relatively high
(above a 50F dewpoint), cooling-based
dehumidifcation is economical in terms of
both operating cost and initial equipment
cost. Below 50F, precautions need to be
taken to avoid freezing the condensed
water on the cooling coil. Consequently,
desiccants are more economical than cool-
ing-based systems at lower dewpoints.
Relative humidity sensitivity. When a pro-
cess needs a low moisture level in absolute
terms, but can tolerate a high relative
humidity, cooling-based dehumidifcation
without desiccants is cost effective. By con-
trast, in processes that demand a low rela-
tive humidity in addition to a low dewpoint,
desiccant systems are used for humidity
control, with supplementary cooling systems
to keep temperature within acceptable lim-
its. When a product is sensitive to relative
humidity but not to temperature, a desiccant
dehumidifer is used without a cooling unit
to maintain a constant relative humidity.
Temperature tolerance. If the application
can tolerate a wide temperature range,
then dehumidifcation alone may suffce. In
most cases, both temperature and moisture
must be maintained within set limits, so both
cooling and desiccant equipment are used
in a combination to maintain control.
References
1. Harriman, L., Dont Sweat It, Dehumidify,
Chem. Eng., August 1997, pp. 8087.
2. Soleyn, K., Humidity Control: Preventing
Moisture Contamination, Chem. Eng., October
2003, pp. 5051.
100%
relative humidity
75%
50%
25%
Air temperature, oF
S
p
e
c
i
f
i
c

h
u
m
i
d
i
t
y
,

g
r
a
i
n
s
/
l
b
10 20 30 50 60 70 80
Leave
Enter
Cooling
coil
160
120
80
10
40
Condenser
Compressor
Receiver
Fan
Heater
Condensate pan
100%
relative humidity
75%
50%
25%
Air temperature, oF
10 20 30 50 60 70 80
Leave
Enter
Process air Dry air
S
p
e
c
i
f
i
c

h
u
m
i
d
i
t
y
,

g
r
a
i
n
s
/
l
b
160
120
80
10
40
Drive motor
Desiccant
wheel
Heater
Figure 1. Air through a cooling
dehumidifier
Figure 2. Air through a desiccant
dehumidifier
Department Editor: Rebekkah Marshall
Preventing
Runaway
Reactions
general considerations [1]
A process is considered to be thermally
safe only if the reactions can easily be
controlled, and if the raw material, the
products, the intermediates and the re-
action masses are thermally stable under
the considered process conditions. Check
into the process equipment, its design, its
sequence of operation and the control
strategies. In addition to the engineer-
ing aspects, get detailed information on
thermodynamic and kinetic properties
of the substances involved, such as the
reaction rates or heat-release rates as
a function of process conditions. Deter-
mine the physical and chemical proper-
ties, as well.
Understanding of thermal-hazard po-
tential requires knowledge of various
skills and disciplines [3]. These include:
Operating mode: The mode of opera-
tion is an important factor. For instance,
a batch reaction, where all the reactants
are charged initially, is more difficult to
control than a semi-batch operation in
which one of the reactants is charged
progressively as the reaction proceeds
(for more, see Design Options).
Engineering: Design and layout of the
plant and equipment and its built-in con-
trols impact the entire process. The ca-
pacity of the heating or cooling system
is important in this context. Process en-
gineering is used to understand the con-
trol of the chemical processes on a plant
scale. It determines which equipment
should be used and how the chemical
processes should be performed. In ad-
dition, take into account technical failure
of equipment, human errors (deviations
from operating instructions), unclear
operating instructions, interruption of
energy supply, and external influences,
such as frost or rain (for more, see De-
sign Options).
Chemistry: The nature of the process and
the behavior of products must be known,
not only under reaction conditions, but
also in case of unexpected deviations
(for example, side reactions, instability
of intermediates). Chemistry is used to
gain information regarding the reaction
pathways that the materials in question
follow.
Physical chemistry and reaction kinetics:
The thermophysical properties of the reac-
tion masses and the kinetics of the chemi-
cal reaction are of primary importance.
Physical chemistry is used to describe the
reaction pathways quantitatively.
data collection
The following data are especially rel-
evant in avoiding runaway reactions:
Physical and chemical properties, ig-
nition and burning behavior, electro-
static properties, explosion behavior
and properties, and drying, milling,
and toxicological properties
Interactions among the chemicals
Interactions between the chemicals
and the materials of construction
Thermal data for reactions and de-
composition reactions
Cooling-failure scenarios
design options [2]
If a reaction is has the potential for
runaway, the following design changes
should be considered:
Batch to continuous. Batch reactors
require a larger inventory of reac-
tants than continuous reactors do, so
the potential for runaway in continu-
ous systems is less by comparison
Batch to semi-batch. In a semi-batch
reaction, one or more of the reactants
is added over a period of time. There-
fore, in the event of a temperature or
pressure excursion, the feed can be
switched off, thereby minimizing the
chemical energy stored up for a sub-
sequent exothermic release
Continuous, well-mixed reactors to
plug flow designs. Plug-flow reactors
require comparatively smaller volumes
and therefore smaller (less dangerous)
inventories for the same conversion
Reduction of reaction inventory via
increased temperature or pressure,
changing catalyst or better mix-
ing. A very small reactor operating
at a high temperature and pressure
may be inherently safer than one
operating as less extreme conditions
because it contains a much lower in-
ventory [3]. Note that while extreme
conditions often result in improved
reaction rates, they also present their
own safety challenges. Meanwhile, a
compromise solution employing mod-
erate pressure and temperature and
medium inventory may combine the
worst features of the extremes [3].
Less-hazardous solvent
Externally heated or cooled to inter-
nally heated or cooled
thermal stability
criteria [1, 4]
As a guideline, three levels are sufficient
to characterize the severity and prob-
ability of a runaway reaction, as shown
in the Table.
Defining high, meDium
anD low risk [1]
Severity Probability
High T
ad
>

200K TMR
ad
< 8 h
Medium 50K < T
ad
<

200K 8 h < TMR
ad
<
24 h
Low T
ad
<

50K and
the boiling point
cannot be sur-
passed
TMR
ad
> 24 h
adiabatic temperature rise
The adiabatic temperature rise is calculated
by dividing the energy of reaction by
the specific heat capacity as shown in
Equation (1).
T
ad
= 1,000Q
r
/C
p
(1)
where:
T
ad
= adiabatic temperature rise, K
Q
r
= energy of reaction, kJ/kg
C
p
= heat capacity, J/(kg)(K)
time to maximum rate (tmr)
TMR
ad
(the time to maximum rate, adiabatic) is
a semiquantitative indicator of the probability
of a runaway reaction. Equation (2), defining
TMR
ad
in hours, is derived for zero-order
reaction kinetics:
TMR
ad
= C
p
RT
o
2
/3,600q
o
E
a
(2)
where:
R = gas constant, 8.314 J/molK
T
o
= absolute initial temperature, K
q
o
= specific heat output at To, W/kg
E
a
= activation energy, J/mol
The TMR value provides operating personnel
with a measure of response time. Knowledge
of the TMR allows decisions to be based on
an understanding of the time-frame available
for corrective measures in case heat transfer
is lost during processing.
References
1. Venugopal, Bob, Avoiding Runaway Reac-
tions, Chem. Eng., June 2002, pp. 5458.
2. Smith, Robin, Chemical Process Design,
McGraw-Hill, New York, 1995.
3. Kletz, T. A., Cheaper, Safer Plants,
IChemE Hazard Workshop, 2d., IChemE,
Rugby, U.K., 1984.
4. Gygax, R., Reaction Engineering Safety,
Chem. Eng. Sci., 43, 8, pp. 175971, Au-
gust 1998.
Department Editor: Kate Torzewski
Sedimentation
Centrifuging
C
entrifugation is the method of choice
in the chemical process industries (CPI)
for separating solids from liquids. It
relies on the G-forces generated by high-
speed rotation to recover solids or liquids
from slurries, as well as clarify liquids or
classify solids.
Centrifuges can be categorized as either
sedimentation or fltration units. Sedimenta-
tion centrifugation relies on a difference in
density between the solid and liquid being
separated. Filtering is performed with a
rotating basket ftted with a flter medium,
where the centrifugal force of rotation
expels the liquid through the flter.
basics of sedimentation
centrifuging
The mechanics of sedimentation centrifug-
ing make it ideal for two-phase systems
with a high-density differential. As an
incoming slurry spins in a sedimentation
centrifuge, it forms an annulus adjacent to
the bowl wall. The centrifugal force causes
the denser material to move outwardly to-
ward the wall of the centrifuge bowl, while
the liquid overfows from the bowl or is
picked up by a skimmer. Periodically, the
solid must be removed from the centrifuge
manually or with a cutter knife. Alterna-
tively, it can be removed continuously with
a screw conveyor. Sedimentation centrifu-
gation allows material to be separated
hundreds or thousands of times faster than
simple sedimentation by gravity alone.
stoKes' law
According to Stokes Law, the terminal
velocity of a particle is determined by the
centrifugal gravity (
2
r) created by the
centrifuge with particle-balancing buoy-
ancy and viscous drag taken into account.
This terminal (settling) velocity is deter-
mined by the equation below:
V r d
s s L

1
18
2 2
M
R R 7 ( )
where
V
s
= Settling velocity, m/s
= viscosity, kg/ms
= angular speed of rotation, rev/min

s
= density of solid, kg/m
3

L
= density of liquid, kg/m
3
d = particle diameter, m
r = centrifuge radius of curvature, m
Stokes Law tells us that settling velocity
can be maximized with a high centrifugal
speed, large particle size, large density
difference between solids and liquid, large
separation radius and low liquid viscosity.
applications
Centrifuge selection is heavily dependent
on characteristics of the incoming slurry, in-
cluding particle size, solids concentration,
liquid viscosity and density differential. Oth-
er factors that come into play are the need
to remove solids periodically or continuous-
ly and the degree of purity required of the
separated products. Table 1 summarizes
the mechanics and suitable applications of
common sedimentation centrifuges.
References
1. Scroder, T. Selecting The Right Centrifuge, Chem.
Eng. September 1998, pp. 8288.
2. Moir, D. N. Sedimentation Centrifuges: Know What
You Need, Chem. Eng. March 1988, pp. 4251.
3. Bershad, B. C., Chaffotte, R. M., Leung, W. F.
Making Centrifugation Work For You, Chem.
Eng. August 1990, pp. 8489.
4. Perrys Chemical Engineers Handbook, 8th ed.
New York: McGraw Hill, 2008.
TABLE 1. types of sedimentation centrifuges
type mechanics Best suited for
Tubular
Bowl
A vertical cylinder with the feed
slurry introduced in the bottom of
the bowl
The use of a distributor and baffle
assembly accelerates the slurry to
the speed of rotation
The heavier phase becomes concen-
trated against the wall, while the lighter
phase floats on top
The two phases are separated by a baffle
Liquid discharges over the top of the bowl,
while solid buildup is removed manually
Purification of lubricating
and industrial oils
Food, biochemical and
pharmaceutical applications
Solids should be less than 1%
in volume of the slurry
Multi-
chamber
Constructed of a series of tubular
sections arranged concentrically
The slurry feed enters in the
smallest tube and continues
through the outer tubes as they
increase in size
Larger solid particles settle in the small
tubes, and particles of smaller sizes settle
in subsequent tubes
Up to six chambers are typical with a
maximum holding capacity of 0.064 m
3
Clarifying fruit juices, wort
and beer
Skimmer
pipe /
knife
discharge
Feed enters the hub end and
is accelerated to speed before
entering the separation pool
Solids settle on the bowl wall while
liquid overflows the ring weir
When a thick solid layer begins to form
on the bowl wall, supernatent liquid is
removed with a skimmer, and solids are
knifed out with centrifugal filters
Heavy-duty applications,
such as coal dewatering
Disc Feed enters through the top axis
of the bowl and is accelerated by
a radial-vane assembly
The unit is constructed of a stack
of typically 50 to 150 closely
spaced conical discs arranged at
an angle between 40 and 50 deg
Solids settle under the disc and move
downward to be released at the bottom of
the bowl wall
Liquids travel up the conical channel,
and their upward movement in the
centrifuge is facilitated by holes across
each disc
Self-cleaning types:
purification of beverages,
mineral oils, and edible oils
Disc nozzle: corn wet
milling (starch separation,
gluten thickening), clay
classification, acid crystal
washing, lube oil dewaxing
Decanter Constructed of a solid external
bowl and an internal screw
conveyor mounted horizontally
Solids are removed from the conical
discharge end (the beach)
Bowl and conveyor rotate in the same
direction, but at different speeds, creating
a speed differential that controls the
speed of solid removal
Applications that require
continuous removal of solids,
where feed solids are high
and volume reduction is
important
Screen-
bowl
A solid bowl decanter with a cylindrical screen added to the conical end Improved cake dryness and
highest product purity
Department Editor: Rita L. D'Aquino
Solvent Selection
Methodology
A STEPWISE ProcEdurE
Organic solvents have been used in many industries for centu-
ries, but the methods and tools to select optimal solvents while
minimizing their adverse environmental, health, safety and op-
erational concerns are still evolving. The appropriate selection
of solvents depends to a large extent on the application more
specifcally on what needs to be dissolved, and under what
conditions. This article presents a four-step approach to solvent
selection based upon Ref. 1*, where the reader will fnd a list of
additional resources on this topic.
Identify the challenge and solvent characteristics.
The frst two steps are: 1) identifying the actual problem and
technology or unit operation required to solve it; and 2) defning
the requirements that must be met by the solvent, using criteria
related to its physical and chemical properties (e.g., pure-solvent
properties, such as normal boiling point, the Hildebrand solubility
parameter at 300 K, the Hansen solubility parameters; solvent-
solute properties, such as the solubility of the solute as a function
of the composition of the mixture; and functional constraints, such
as solute loss in solute).
Obtain reliable values of solvent properties and
narrow down selection. There are several alternatives for this
third step. For example, one can measure the required properties,
use a database of properties of chemicals (or solvents), or, use
property models to estimate them. For solvent-selection problems
not involving chemical reactions, the pattern of the desired
solvent is established through analysis of the solute, application
type, and other constraints.
Once this is established, a database of known solvents can
be used to identify the solvents that match the necessary pattern
(Table 1). On the other hand, when chemical reactions are
involved, the approach is based on transition-state theory and
requires consideration of the solvation energies of the reactants,
products and transition states, and thus, knowledge of the reac-
tion mechanism.
When the crucial values have been found, the solvent search
could be such that frst, solvent-pure properties are used, followed
by solvent-EHS, then solvent-solute, and fnally solvent-function.
Narrow down the list by removing the compounds that do not
match desired properties.
A protocol derived by Britest Ltd. (www.britest.co.uk) seeks to
use mechanistic principles to guide solvent selection (Figure). The
objective is to follow the arrows according to the problem defni-
tion and a search criterion until an end-point is reached, thereby
obtaining the characteristics of the candidate solvents. These
characteristics are used to identify the group to which the solvents
belong using solvents database (see Table 2). The corresponding
group-types are evaluated and a fnal selection is made.
Verify selection. The fourth step is to verify that the solvent
works as expected by performing a computational validation
by simulation. Experimental validation of a solvent candidate is
required at all stages of process development.
Table 2. Well-known solvents together with their related properties
Solvent Name Molecule type Group type Charge NBP NMP Sol. Par.
(K) (K)
1-Methyl-2-pyrrolidinone Amide 1 NE/EPD 475.15 249.15 23.16
Acetonitrile Nitrile 1 E/NPG 354.75 229.35 24.05
Dimethyl sulphoxide S-oxide 1 E/NPG 462.15 291.65 26.75
Dimethyl formamide Amide 1 NE/NPG 426.15 212.75 23.95
Dimethylacetamide Amide 1 NE/NPG 438.15 253.15 22.35
Diisopropyl ether Ether 2 NE/EPD 341.65 181.35 14.45
Dimethyl ether Ether 2 NE/EPD 248.35 131.65 15.12
Methyl tertbutyl ether Ether 2 NE/EPD 328.35 164.55 15.07
Tetrahydrofuran Ether 2 NE/EPD 338.15 164.85 18.97
Chlorobenzene Chloride 3 NE/P 632.35 404.9 19.35
m-xylene (also o-; p-) Aromatic HC 3 NE/P 412.27 225.3 18.05
Toluene Aromatic HC 3 NE/P 383.95 178.25 18.32
Acetic acid Acid 4 PG 391.05 289.81 19.01
Propionic acid Acid 4 E/PG 414.25 252.45 19.41
Sulfuric acid Acid 4 E/PG 610
ordered
283.46 28.41
Propanol Alcohol 5 E/N 370.35 147.05 24.45
Ethanol Alcohol 5 E/N 351.35 159.05 26.13
Butanol Alcohol 5 E/N 390.81 183.85 23.35
Ethylene glycol Alcohol 5 E/N 470.45 260.15 33.7
Dichloromethane Chloride 6 NE/EPD 313.15 178.05 20.37
Heptane Alkane 7 NE/I 371.65 182.55 15.2
Hexane Alkane 7 NE/I 341.85 177.85 14.9
Pentane Alkane 7 NE/I 309.22 143.42 14.4
Methanol Alcohol 4, 5 E/N 337.85 175.47 29.59
Water Aqueous 4, 5 E/N 373.15 273.15 47.81
NE = non-electrolytic solvent; E = electrolytic solvent; P = polarizable; EPD = electron-pair
donor; I = inert; PG = protogenic (proton donor); N= neutral (donor & acceptor); NPG =
non-protogenic (proton acceptor); NBP = normal boiling point; NMP = normal melting point;
Sol. Par. = Hildebrand solubility parameter at 300 K (MPa
1/2
)
Table 1. Some well-known databases and solvent selection tools
Databases Address and comments
ChemFinder Searchable data and hyperlink index: http://chemfnder.cambridgesoft.com
Solvents Databases Solvent substitution data systems at http://es.epa.gov/ssds/ssds.html;
Handbook of Solvents from www.chemtec.org/cd/ct_23.html; and
SOLVDB at http://solvdb.ncms.org/index.html
NIST Webbook
DIPPR and
TAPP
Source of physical and chemical data at http://webbook.nist.gov
www.aiche.org/TechnicalSocieties/DIPPR/About/Mission.aspx; and
www.chempute.com/tapp.htm
CAPEC Database Pure as well as mixture properties data, including solvent-solute database:
www.capec.kt.dtu.dk/Software/ICAS-and-its-Tools
Selection Tools Address and comments
SMSwin A specialized software for property estimation and solvent classifcation:
www.capec.kt.dtu.dk/documents/software/SMSWIN.htm
NRTL-SAC and
eNRTL-SAC
Activity coeffcient method based on segment contributions. Predictive based
on a small set of solubility data. Useful for crystallization solvent selection
and extends to LLE and VLE: www.aspentech.com
Stability, solubility of reactants, products
Single phase or solid-liquid
Homogeous catalysis
by Pt group complexes
Moderate polarity
DPA ethers,
aromatics
Dipolar
aprotic
Fast, low temp,
but recovery
difficult
Slow, high-
temperature,
easy recovery
Aromatic
hydrocarbon
(xylene)
High polarity
Water, carboxylic
acids, inorganic
acids, lower
alcohols
Condensation SN1/E1 SN2/E2
Substrate/product hydroxyl sensitive
No
Yes
Water,
alcohols
Dipolar aprotic
ethers
Consider
solvation
Two-phase or liquid-liquid
(polar phase is water)
Water, immiscible solvent
Choose polarity
based on substrate
and reagent
solubility. May need
phase-transfer catalyst
Group 7 Group 5 Group 2 Group 4 Group 3 Group 1 Group 3
Group 6 Group 1 Group 2
Group 3 Group 1
*Reference: 1. Gani, R., et al., A Modern Approach to Solvent Selection, Chem. Eng., Vol. 113, No. 3, pp. 3043, Mar. 2006. Author E-mail: rag@kt.dtu.dk
Department Editor: Kate Torzewski
Avoiding
Seal Failure
S
eals are assemblies of elements that prevent
the passage of a solid, liquid, gas or vapor
from one system to another. When a seal
allows leakage of material, failure has occurred.
This guide provides an overview of common seal
types and a discussion of seal failure to aid in
choosing the most effective seal and avoiding
future failure.
seal types
Seals types can be classifed within two broad
categories: static and dynamic. Static seals have
no relative motion between mating surfaces, while
dynamic seals do have relative motion between
a moving surface and a stationary surface. Seals
do not have to ft into one category or the other;
rather, seal types can fall anywhere on a spectrum
between static and dynamic, and few seals are
strictly one type or the other. Table 1 describes the
applications and requirements of several common
seal types.
seal failure
Seal failure is caused by a wide variety of circumstances, including improper in-
stallation and environmental factors such as temperature, pressure, fuid incompat-
ibilities, time and human factors.
Most causes of failure can be described as mechanical diffculties or system
operations problems. Examples of mechanical diffculties include strain on the
seal face caused by improper installation and vibration caused by improper net
positive suction head. Meanwhile, system operating problems can include condi-
tions that are outside of a pumps best performance envelope, such as upsets, dry
running, and pressure or temperature fuctuations. Changes in the fuid being pro-
cessed can cause problems as well, especially with fuids that fash or carbonize.
Common visual indicators of failure include short cuts, V-shaped notches in the
seal, skinned surface in localized areas, or thin, peeled-away area on the seal.
Table 2 describes causes of some of the most prevalent types of seal failure with
recommended methods of action.
In some cases, the cause of failure may be diffcult to determine due to the com-
plexity of the seal construction. These unique failure modes can result in faking or
peeling of the seal face, corrosion, faking or pitting of the carbon faces, degrada-
tion of the elastomer energizer seals, and spring or bellows breakage. It is likely
that these rapid degradations are a result of contamination, which can be avoided
with careful installation or using pre-assembled, cartridge-type mechanical seals.
References
1. Ashby, D. M. Diagnos-
ing Common Causes of
Sealing Failure, Chem.
Eng. June 2005, pp.
4145.
2. Netzel, J., Volden,
D., Crane, J. Suitable
Seals Lower the Cost
of Ownership, Chem.
Eng. December 1998,
pp. 9296.
TABLE 1. COMPARISON OF COMMON SEAL TYPES
Type Applications Periodic
Adjustment
Required?
Moving
friction
Tolerances
required (mov-
ing seals)
Gland
adapters
required?
Space
require-
ments
Static Dynamic
O-ring X X No Medium Close No Small
T-seal X X No Medium Fairly close No Small
U-packing X No Low Close No Small
V-packing X Yes Medium Fairly close Yes Small
Cup-type pack-
ing
X No Medium Close Yes Medium
Flat gasket X Yes No Large
Compression or
jam packing
X X Yes High Fairly close Yes Large
TABLE 2. SOLUTIONS TO COMMON CAUSES OF SEAL FAILURE
Failure type Definition Causes Solutions
Compression
set
A lost of resiliency caused by the
failure of a seal to rebound after it has
been deformed for some period of
time. The seal will exhibit a fattened
surface corresponding to the contours
of the mating hardware
Exposure to excessive tempera-
ture or incompatible fuids
Excessive deformation of the
elastomer at installation
An incompletely vulcanized seal
Choose proper defection for the seal
Choose appropriate elastomer mate-
rial for the application in terms of
thermal stability and compression set
resistance
Nibbling and
extrusion
A seal starts to appear to be torn
away in little pieces until it loses its
overall shape and fows into whatever
void area is available
Excessive clearance gaps
Improper seal material
Excessive volume-to-void ratio
Inconsistent clearance gaps
Increase bulk hardness of the sealing
element
Decrease clearance gaps
Redesign volume-to-void ratio
Add anti-extrusion devices
Spiral failure A seal rolls within its gland, resulting
in cuts or marks that spiral around the
circumference of the seal
Applications where a seal is
used in a slow, reciprocating
fashion
Irregular surface over the mating
parts causing the seal to grip to
certain contact points
Use an elastomer with a higher bulk
hardness
For male-type installation, increase the
installed stretch on the seal
Specify a smoother, more uniform fn-
ish on mating hardware
Change the type of seal to a lip-type
confguration
Explosive
decompres-
sion
Seal exhibits blisters, fssure, pock
marks or pits, both externally and
internally
Gas entrapment within the
elastomer during high-pres-
sure cycling, followed by rapid
depressurization
Use an elastomer material that is more
resilient to explosive decompression
Use polymeric or metal seals if
0possible
Wear Smooth burnishing of a sealing
surface
Relative motion of the seal
against the mating surface
Use a harder material
Use a polymeric solution
Department Editor: Scott Jenkins
Flow Prole for
Reciprocating
Pumps
R
eciprocating pumps are often used in
the chemical process industries (CPI)
because of their ability to generate high
pressures at low velocities. A subcategory of
positive-displacement pumps, reciprocating
pumps act through the recipricating motion
of a piston, plunger or diaphragm. Such
pumps work by way of a connecting-rod-
and-crank mechanism with a piston.
By nature, reciprocating pumps generate
pulsing ow, which, when plotted as a func-
tion of time, or of crank angle, produces a
curve that resembles a sine wave to a rst
approximation. For example, manufacturers
of pulsation dampeners and surge suppres-
sors often use sinusoidal curves for piston
pumps and compressors in their product
literature and sizing formulas. However, a
closer examination of the ow prole for
a piston-and-crank pump or compressor
reveals the curve to be a signicantly dis-
torted sine wave because of the interaction
between the crank and the connecting rod.
Calculating fowrate
In graphical form, the crank and crankshaft
of a reciprocating pump can be visual-
ized by placing the crankshaft center at
the 90-deg mark of a 180-deg x-axis, and
placing the crank bearing at the origin (see
gure). A connecting rod links the crank to
the piston.
Determining the position of the piston at
any crank angle can be accomplished by
measuring on a piston pump, compressor,
or piston engine, or it can be calculated
using trigonometric relationships.
The degree to which the actual ow pro-
le curve deviates from the sinusoidal curve
is determined by the ratio of the connecting
rod length to the crankshaft length. Smaller
values of the ratio translate into greater
levels of distortion. As the connecting rod
becomes very long, the ow prole would
approach the sine curve.
To calculate the owrate at a given crank
angle, use the following procedure and
denitions:
Crank length = OC
Piston rod length = CP
For any angle a, Line AC = OC sin (a)
Line SA = OC OC cos (a)
Line AP = (CP
2
AC
2
)
0.5
Line SP = AP + SA
1. Calculate the piston position for two
crank angles, perhaps 2 deg apart.
2. The difference in piston positions equals
piston displacement over the time interval
between the two crank angles. The value
is an average over the span of the two
readings, not an instantaneous read-
ing. As the step size approaches zero,
displacement nears the true velocity.
3. This value can be converted into ow-
rates (gal/min or other units) if the piston
diameter and speed (revolutions per
minute, rpm) are known.
Observations of the plot
In an illustrative example, plots of piston
velocity versus crank angle are shown (see
graph). The ratios of the connecting rod
length to crank shaft length are 1.05 to
1 (blue line), 2 to 1 (red line) and 5 to 1
(green line). The following observations
can be made:
1. At the beginning of the discharge stroke,
owrate approaches zero asympotically,
rather than as a sinusoidal curve
2. Peak owrates do not occur at the
90-deg point, but rather at 95120 deg,
depending on the ratio of rod length to
crank length
3. Peak owrates are higher than would be
predicted with a pure sine curve
4. From 180 to 360 deg (the suction por-
tion of the pump cycle), the curve mirrors
the 0-to-180-deg portion
5. Flowrates during the suction portion
of the curve are also higher and occur
earlier than the 270-deg point
Effects of distorted sine curve
Within the areas of uid ow and mechani-
cal pump design, there are a number of
aspects that are affected by the deviation
of ow prole from a perfect sine curve for
pumps and compressors. The effects include
the following:
* Check volves ond possoges wlll hove
higher-than-predicted peak owrates
and pressure drop will be higher, by the
square of owrate
* lhe hlgher Nowroles ond pressure drops
will affect net positive suction head
(NPSH) and possibly induce vaporization
* Moxlmum cronk revolullons per mlnule wlll
be lower than what would be allowed by
the pure (non-distorted) sinusoidal curve
* Loods experlenced by beorlngs wlll
increase somewhat, especially in high-
speed compressors
* Slress onolysls ol lhe connecllng rods wlll
be affected
* Surge dompeners musl hondle lhe shorper
peak of a bell curve, rather than a
smoother sine curve
* Mulllplslon pumps ond compressors
would have less smoothing effect than
would be predicted because the bell-
shaped curve has a sharper peak
References
1. McGulre, ).l., Pumps lor Chemlcol Process-
lng," Morcel Dekkor, Mew York, 1PP0.
2. Henshaw, T.E., Reciprocating Pumps, Van
Moslrond Relnhold Co., Mew York, 1P8Z.
3. Krugler, A., Piston Pumps and Compressors:
Exploring the Flow Prole, Self-published,
2010.
Note. Molerlol lor lhls edlllon ol Focls ol Your
Fingertips was contributed by Arthur Krugler,
P.E., Krugler Englneerlng Group lnc., Vhllller,
Calif. (www.kruglerengineeringgroup.com).
Crank
Connecting rod
C
S
a
A
O
0 deg
180 deg
P Piston
Cylinder

90 deg
Piston velocity versus crank angle
0.0000
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00 200.00
Crank angle (rotation), deg
Crank length = 2
Connecting
rod length =
2.05 (blue line)
4 (red line)
10 (green line)
P
i
s
t
o
n

v
e
l
o
c
i
t
y
,

i
n
.
/
d
e
g

X

1
0



















tep 1 ht dimeter
(


















tep 2 diI cIernce



tep xiI cIernce




tep eI mounting

















tep 5 ounting surce



InternI ger pumps

















xternI ger pumps











obe pumps








ne umps










eerences








25
2
15
1
5

5 1 15
pcity, gImin
erormnce
1
8
6

11
1
9
8
7
6
5

8
7
6
5

25 5 75 1,
iscosity, ct
1 2 5
iscosity, ct
55 8 15
eet o hed
entriugI
entriugI
ositie
ositie
iscosity
ed Iowrte
entriugI
ositie
ositie
entriugI
otor
ush rod ne
otor
IdIer
rescent
Department Editor: Kate Torzewski
Pump Selection
and Specifcation
PUMP SELECTION
I
n choosing a pump, it is important to match a pumps capabili-
ties with system requirements and the characteristics of the liquid
being processed. These factors include the inlet conditions,
required fowrate, differential pressure and liquid characteristics.
Generally, the quality of the liquid should remain unchanged after
passage through a pump. Therefore, material compatibility, viscos-
ity, shear sensitivity and the presence of particulate matter in a
liquid are important considerations in pump selection.
Most engineering applications employ either centrifugal or
positive displacement (PD) pumps for fuid handling. These pumps
function in very different ways, so pump selection should be based
on the unique conditions of a process.
Centrifugal pumps
The most widely used pump in the chemical process industries for
liquid transfer is the centrifugal pump. Available in a wide range of
sizes and capacities, these pumps are suitable for a wide range of
applications. Advantages of the centrifugal style include: simplic-
ity, low initial cost, uniform fow, small footprint, low maintenance
expense and quiet operation.
Positive displacement pumps
Though engineers may be frst inclined to install centrifugal pumps,
many applications dictate the need for PD pumps. Because of their
mechanical design and ability to create fow from a pressure input,
PD pumps provide a high effciency under most conditions, thus
reducing energy use and operation costs.
Choosing centrifugal versus positive displacement
These two main pump styles respond very differently to various
operating conditions, so it is essential to evaluate the requirements
of a process prior to choosing an appropriate pump. Table 1 il-
lustrates the mechanical differences between these pumps, as well
as the effects of pressure, viscosity and inlet conditions on fowrate
and pump effciency.
Range of operation
Pump styles range far beyond simply PD and centrifugal pumps.
PD pumps encompass many specifc styles, including a variety of
reciprocating, rotary and blow-cover pumps. Likewise, centrifugal
pumps encompass radial, mixed, and axial fow styles, which all
belong to a greater category of kinetic pumps.
A simple way to narrow down pump styles is to determine the
required capacity that your pump must handle. Based upon a
required capacity in gal/min. and a pressure in lb
f
/in.
2
, the pump
coverage chart below can help engineers focus their selection to a
just a few pump styles.
PUMP SPECIfICaTIONS
Based on the application in which a pump will be used, the pump
type, and service and operating conditions, the specifcations of a
pump can be determined.
Casting connection: Volute casing effciently converts velocity en-
ergy impacted to the liquid from the impeller into pressure energy.
A casing with guide vanes reduce loses and improve effciency
over a wide range of capacities, and are best for multistage high-
head pumps
Impeller details: Closed-type impellers are most effcient. Open-
type impellers are best for viscous liquids, liquids containing solid
matter, and general purposes
Sealings: Rotating shafts must have proper sealing methods to
prevent leakage without affecting process effciency negatively.
Seals can be grouped into the categories of noncontacting seals
and mechanical face seals. Noncontacting seals are often used
for gas service in high-speed rotating equipment. Mechanical face
seals provide excellent sealing for high leakage protection
Bearings: Factors to take into consideration while choosing a
bearing type include shaft-speed range, maximum tolerable shaft
misalignment, critical-speed analysis, loading of compressor
impellers, and more. Bearing styles include: cylindrical bore; cy-
lindrical bore with dammed groove; lemon bore; three lobe; offset
halves; tilting pad; plain washer; and taper land
Materials: Pump material is often stainless steel. Material should
be chosen to reduce costs and maintain personnel safety while
avoiding materials that will react with the process liquid to create
corrosion, erosion or liquid contamination
References
1. Perrys Chemical Engineers Handbook, 7th ed. New York: McGraw
Hill, 1997.
2. Petersen, J. and Jacoby, Rodger. Selecting a Positive Displacement Pump,
Chem. Eng. August 2007, pp. 4246.
PumP ComParison Chart
Centrifugal
Pump
Positive displacement
pump
Mechanics
The pump imparts a
velocity to the liquid,
resulting in a pressure
at the outlet.
Pressure is created
and flow results
The pump captures
confined amounts of
liquid and transfers
them from the suction
to discharge port.
Flow is created and
pressure results
Performance
Flow varies with
changing pressure
Flow is constant with
changing pressure
Viscosity
Efficiency decreases
with increasing viscosity
Efficiency increases with
increasing viscosity
Efficiency
Efficiency peaks at the
best-of-efficiency point.
At higher or
lower pressures,
efficiency decreases
Efficiency increases
with increasing pressure
Inlet
conditions
Liquid must be in the
pump to create a pres-
sure differential. A dry
pump will not prime on
its own
Negative pressure is
created at the inlet port.
A dry pump will prime
on its own
100
5
2
2 5 10 1 2 5 100
Capacity, gaI/min
2 5 1,000 2
10
100
1,000
10,000
510,000
1,000
5
2
H
e
a
d
,

f
t

o
f

I
i
q
u
i
d
P
r
e
s
s
u
r
e
,

I
b
f
/
i
n
2
10,000
5
2
10
Metering - pIunger
Metering - diaphragm
Screw
Regenerative
CentrifugaI -singIe stage, singIe suction
CentrifugaI -muItistage
CentrifugaI -doubIe suction
Gear Direct-acting steam
MuIticyIinder AxiaI fIow
Adapted from Perrys Chemical Engineers Handbook
Department Editor: Kate Torzewski
Vacuum
Pumps
V
acuum is any system of reduced
pressure, relative to local
(typically atmospheric) pressure.
Achieved with a pump, vacuum
systems are commonly used to:
Removeexcessairand
its constituents
Removeexcessreactantsor
unwanted byproducts
Reducetheboilingpoint
Drysolutematerial
Createapressuredifferentialfor
initiatingtransportofmaterial
Liquid-ringanddrypumpsofferthe
mostadvantagesforthechemical
processindustries(CPI).Bothof
thesepumptypeshavebearings
sealedofffromthepumpingcham-
ber and do not require any internal
lubrication because the rotors do
notcontactthehousing.Both,when
employingacoolantsystem,prevent
thecoolantfromcontactingthepro-
cessfuidandcausingcontamina-
tion, and both use mechanical shaft
seals for containment.
liquid-ring pumps
Inthecylindricalbodyofthepump,
asealantfuidundercentrifugal
forceformsaringagainsttheinside
ofthecasing(Figure1).
The source of that force is a
multi-bladed impeller whose shaft
is mounted so as to be eccentric to
theringofliquid.Becauseofthis
eccentricity, the pockets bounded
by adjacent impeller blades (also
calledbuckets)andtheringincrease
in size on the inlet side of the pump,
andtheresultingsuctioncontinually
drawsgasoutofthevesselbeing
evacuated. As the blades rotate
towardthedischargesideofthe
pump, the pockets decrease in
size,andtheevacuatedgasiscom-
pressed,enablingitsdischarge.
Theringofliquidnotonlyactsas
a seal; it also absorbs the heat of
compression, friction and condensa-
tion.Popularliquidchoicesinclude
water,ethyleneglycol,mineraloil
andorganicsolvents.
dry pumps
Rotary-claw,rotary-lobeand
rotary-screw pumps dominate as
drypumpsintheCPI,particularlyin
larger-sizepumpapplications.
Rotary Claw. Thegeometricshape
ofthispumpallowsforagreater
compression ratio to be taken
acrosstherotorsathigherpressures
(Figure2).Twoclawrotorsrotate
in opposite directions of rotation
withouttouching,usingtiminggears
tosynchronizetherotation.Thegas
entersthroughaninletportafterit
has been uncovered and flls the
void space between the rotors and
pumphousing.Onthenextrotation,
thatsametrappedsampleofgasis
compressedanddischargedasthe
dischargeportopens.
Aminimumofthreestagesin
series is required to achieve pres-
sures comparable to those of an
oil-sealed mechanical pump. Some
drydesignsusetwotechnologiesin
combination;forexample,arotarly
lobe as a booster for a claw pump.
Rotary Lobe. The rotary-lobe pump
(Figure3)istypicallyusedasame-
chanicalboosteroperatinginseries
with an oil-sealed piston or vane
pumptoboostpumpingcapacityat
low pressures.
This pump consists of two sym-
metrical two-lobe rotors mounted on
separate shafts in parallel, which
rotate in opposite directions to each
otherathighspeeds.Timinggears
are used to synchronize the rotation
of the lobes to provide constant
clearance between the two.
Rotary Screw. Twolonghelical
rotors in parallel rotate in opposite
directionswithouttouching,syn-
chronizedbyhelicaltiminggears
(Figure4).Gasfowmovesaxially
alongthescrewwithoutanyinternal
compression from suction to dis-
charge.Pocketsofgasaretrapped
within the convolutions of the rotors
andthecasing,andtransportedto
thedischarge.Compressionoccurs
atthedischargeport,wherethe
trappedgasmustbedischarged
againstatmosphericpressure.Each
convolution of the rotor acts similarly
toastageinserieswiththeonebe-
hindit;atleastthreeconvolutedgas
pockets in the rotor are required to
achieve acceptable vacuum levels.
References
1. Vilbert,P.,MechanicalPumpsfor
VacuumProcessing,Chem. Eng.
October2004,pp.4451.
2. Aliasso,J.,ChoosetheRight
VacuumPump,Chem. Eng.March
1999,pp.96100.
Advantages Disadvantages
L
i
q
u
i
d
-
R
i
n
g

V
a
c
u
u
m

P
u
m
p
s
Simpler design; employs only one rotating assembly
Can be fabricated from any castable metal
Minimal noise and vibration
Little increase in the temperature of the discharged gas
No damage from liquid or small particulates in the process fuid
Maintenance and rebuilding are simple
Slow rotational speed (1,800 rpm or less), maximizing operating life
Can use any type of liquid for the sealant fuid in situations where min-
gling with the process vapor is permissible
No lubricating liquid in the vacuum chamber to be contaminated
Accommodation of both condensable vapors and noncondensables,
while operating as both a vacuum pump and condenser
Mixing of the evacuated gas with
the sealing fuid
Risk of cavitation requires a portion
of process load to be noncondens-
able under operating conditions
High power requirement to form and
maintain the liquid ring, resulting in
large motors
Achievable vacuum is limited by the
vapor pressure of sealant fuid at
the operating temperature
Power consumption
D
r
y

V
a
c
u
u
m

P
u
m
p
s
Rugged rotor design, constructed of sturdy cast or ductile iron without
any fimsy rotating components
Noncontact design facilitated by timing gears
High rotational speed reduces the ratio of gas slip to displacement,
increases net pumping capacity and reduces ultimate pressure
Multiple staging provides inlet pressures below 1-mm Hg absolute while
discharging to atmosphere
No contamination of evacuated gas
Due to lack of condensation, pump can be fabricated of standard,
inexpensive cast iron
Cannot handle particulate matter,
nor large slugs of liquid
May require a silencer
May discharge gases at high
temperatures
Most diffcult to repair or rebuild
May require a gas purge for cooling,
or to protect the bearings and seals
from the process gas
Due to high operating temperatures,
some process gases may polymerize

Figure 1. Liquid-Ring Pump
Source: Medical Gas Info
Figure 2. Rotary-Claw Pump
Source: Gardner Denver
Hanover, Inc.
Figure 3.
Rotary-
Lobe
Pump
Source: Kurt J. Lesker Co.
Figure 4. Rotary-Screw Pump
Source: Kurt
J. Lesker Co.









































Department Editor: Kate Torzewski
Acid Storage
M
ost common acids can be stored in
horizontal or vertical ASME-type
tanks, as shown in the fgures to the
right, or vertical API-type tanks. Horizontal,
carbon-steel ASME-type tanks of 10,000
40,000 gal capacity should have a plate
thickness of 3/8 in. with dished heads of
the same thickness. The thickness includes
a corrosion allowance of 1/4 in., which
provides a tank life of 1520 years.
Sulfuric acid (H
2
SO
4
)
Storage. This acid is prone to enter into
reactions that generate hydrogen, so in
addition to keeping the vessel vented
adequately, exclude potential sources of
ignition from the vicinity.
Materials of construction. Carbon
steel is satisfactory for concentrated
technical grades of sulfuric acid at
normal atmospheric temperature. H
2
SO
4

solutions that are more dilute corrode
carbon steel severely. To avoid inadver-
tent dilution of concentrated acid, keep
acid away from contact with moist air.
Polyvinyl chloride pipe is recommended
for ordinary sulfuric acid, but for oleum,
Type 316 stainless steel or carbon-steel
lined with a fuorocarbon is best.
Phosphoric acid (H
3
PO
4
)
Storage. The tank bottom should be rolled
to a height of 3 in. (upward). This allows
welds to freely expand or contract. Corner
welds should be avoided, as undue stresses
can occur and aggravate corrosion [2].
Depending on the acid grade, the freezing
point varies and may necessitate heating to
avoid freeze-up in storage. In any case, to
avoid corrosion, high-pressure steam should
not be used; steam coils located several
inches below the bottom of the tank are
recommended. The space below the tank
bottom should be enclosed to permit heat-
ing of the air to 50C, and the tank walls
should be insulated.
Materials of construction. Tanks can
be fabricated of Type 316 extra-low-carbon
stainless steel, rubber--lined carbon steel or
fberglass-reinforced plastic. Carbon steel
should not be used, as it will corrode.
Hydrochloric acid (HCl)
Storage. HCl of all strengths should be
stored in tanks similar to those men-
tioned above. Containment areas should
be provided around tanks, and storage
facilities should include a pressure- and
vacuum-relief service, primary and
redundant level indicators, a high-level
alarm, an overfow line, an emergency
block valve at the tank outlet nozzle and
a vent-fume scrubber.
Materials of construction. These
storage tanks should be fabricated of rub-
ber-lined carbon steel, glass-lined carbon
steel or fber-reinforced polymer (FRP). Soft
natural-rubber compounds are used as lin-
ers for concentrated acid storage tanks at
temperatures up to 60C with a minimum
lining thickness of 3/16 in. Semi-hard
rubber is used for lining equipment and
piping for acid up to 70C with FRP tanks
of vinyl-ester resin.
Nitric acid (HNO
3
)
Storage. Storage tanks for HNO
3
of less
than 95 wt.% concentration should be
designed for at least a slight pressure and
vacuum, permitting the venting of nitrogen-
oxide fumes to collection and disposal
equipment, such as a scrubber or a fare.
When locating the tank vent and overfow
pipe, consider that escaping vapors and
liquid can corrode exterior welds as the
acid is diluted with atmospheric moisture.
Materials of construction. For concen-
trations up to 95 wt.% at ambient tempera-
ture, storage units should be fabricated of
Type 304L stainless steel. For concentra-
tions of 95 wt.% and above, Type 3003
aluminum alloy should be used. Acid in
the range of 5255 wt.% should be stored
in tanks of Type 347 stainless steel using
No. 12 gage sheet. Above 90 wt.%, cor-
rosion allowance in the tank-wall thickness
may be necessary. Glass-lined carbon
steel tanks are satisfactory for all acid
grades.
Hydrofuoric acid (HF)
Storage. Because of anhydrous HFs high
vapor pressure, tanks are designed for a
minimum pressure of 60 psig and have
X-rayed and stress-relieved welds. Tanks
holding 70 wt.% HF are also designed per
ASME code, or for lower pressure, as its
vapor pressure is much lower than that of an-
hydrous HF. These tanks should be equipped
with a relief device, and discharge piping
should be routed to a scrubber. Aqueous HF
tanks should have a vent, with the vent line
also going to the scrubber.
Materials of construction. Carbon-steel
storage tanks can be used for anhydrous HF
at temperatures up to 66C and 70 wt.%
HF. Acid of concentrations greater than 60
wt.% may be handled in steel up to a tem-
perature of 38C. In steel tanks, hydrogen
blistering may be caused by the accumula-
tion of H
2
, so periodic tank inspections are
required to evaluate blistering.
References
1. Grossel, S., Safe Effcient Handling of Acids,
Chem. Eng. December 1998, pp. 104112.
2. Anon., Phosphoric Acid, Rhone-Poulene Basic
Chemicals Co., Shelton, Conn. (1992).
2A*
Weld lines
* Minimum width of the shell course at each saddle. Width may be increased if more economical for shell course layout.
Knuckle radius line
2A*
DImension all nozzle
locations from this line
Standard flanged
and dished heads
Manhole Vent
A A
Top plate of saddle
welded to shell
6 in. typical
Nominal capacity level
Bottom
outlet
baffle
Channel
legs
Submerged
fill pipe
Grounding
clip
Name
plate
R

=

O
D

(
a
p
p
r
o
x
)
1
5

i
n
.
OD
Department Editor: Scott Jenkins
Distillation
Tray Design
I
n a distillation column tray, vapor
passes upward through liquid that is
owing across a horizontal perforated
plate. Vapor passing through the perfo-
rated plate forms a two-phase mixture
with the liquid and enables mass transfer
contacting. This mixture is typically quite
turbulent. Tray design must allow the
turbulent liquid to fall away from the
rising vapor in the space above the tray,
while also enabling the vapor bubbles
to rise out of the falling liquid in the
downcomer. The downcomer is usually
a vertical plate that enables the already
contacted froth to travel down to the next
tray without remixing with the up-owing
vapor from the tray below.
Generally, designing a column tray en-
tails determining the minimum downcom-
er area that still allows vapor bubbles
to rise through the liquid, selecting the
number of downcomers, determining
the active area, and checking the ow
path length to see if a person can pass
through a tray manway. These factors
are the primary drivers for determining
overall tower size.
Downcomer area is determined by the
maximum recommended downcomer ve-
locity. Divide the volumetric ow of liquid
by the downcomer velocity to obtain the
downcomer top area. Typically a curve
of maximum downcomer velocity versus
the density difference between liquid and
vapor is consulted during this process.
A downcomer is generally straight
unless its area exceeds 8% of the tower
area. In that case, the downcomer is
sloped such that its bottom area is 60%
of its top area.
Active area
The active area of a distillation tower is
where the vapor contacts the liquid to
effect mass transfer. Above the active
area, where the liquid falls away from
the rising vapor, is the volume where the
vapor can expand. Typically, the active
area is calculated to be the tower cross-
sectional area minus the downcomer top
and downcomer bottom area.
The minimum active area (ft
2
) for nor-
mal valve trays can be determined from
the following relationship, which is a
modication of a commonly used correla-
tion [1] taken at 82% of jet ood:
Active area = V-Load / [TS
0.5
(0.0762
0.00092(SV
2
)) 0.011W
L
]
Where,
V-Load = CFS
V
(SV / (SL SV))
0.5

TS = Tray spacing, in.
SV = Vapor density, lb/ft
3

W
L
= Weir loading, gal/min per in.
CFS
V
= Vapor volumetric ow, ft
3
/s
The required active area is dependant
on the vapor density and weir loading.
Note that the weir loading need not be
known at this point. Assume a weir load-
ing value of 5 gal/min per in. intially.
Typical tray spacings are 24 in.
Tower area and diameter
Based on the above areas, the overall
tower area and diameter can be deter-
mined by the following:
A
T
= A
Dtop
+ A
Dbottom
+ A
A
D = 2(A
T
/ )
0.5

Where,
A
T
= Tower area, ft
2
A
Dtop
= Downcomer area at top, ft
2
A
Dbottom
= Downcomer area at bottom, ft
2
A
A
= Active Area, ft
2
D = Tower inner dia., ft
Number of downcomers
Once the tower diameter is determined,
then the number of downcomers can be
chosen. As a starting point, an initial
design should use a single downcomer.
The resulting weir length is calculated
from a standard chord-length calcula-
tion, which is iterative for a given
downcomer area.
B
W
= {[(D
2
/360) cos
1
(2Z/D)] 2A
D
}/Z
Where,
Z = [(D
2
/4) B
W
2
]
0.5

B
W
= Weir length of one downcomer, ft
A good place to start the iterative
process is with a weir length 0.8 times
the tower diameter. If the resulting weir
loading is greater than 12 gal/min per
in., then increase the number of tray
passes to two. Recalculate the outlet weir
length for each of the side downcomers
of the column by using half the downcom-
er area. Check the weir loading again
(for the tray with side downcomers). If
the weir loading continues to exceed 12
gal/min per in., increase the number of
tray passes to four. It is assumed that the
two-pass tray with side downcomers has
the shortest weir length.
The simplest approach to designing
4-pass trays is to assume equal bubbling
area and make the side downcomers one-
quarter of the total downcomer area, and
make the center (and off-center) downcom-
ers one-half of the total downcomer area.
Maintaining the resulting downcomer
widths at 6 in. or more will allow a person
to reach into the downcomer for installa-
tion. In addition, make sure the resulting
tray-ow path-length is 16 in. or greater
to enable a person to physically pass
through the trays. These minimum size
criteria may increase the column diameter
to above the previously calculated value.
Other considerations
Other criteria that need to be consid-
ered are; downcomer backup, spray
uidization, and entrainment. In addition,
minimum load conditions need to be
determined. The criteria for determining
the low-end vapor and liquid range are
weeping, tray stability and dry-tray pres-
sure drop.
Reference
1. Glitsch Inc. Ballast Tray Design Manual; Bul-
letin No. 4900. 3rd Ed. Glitsch Inc., Dallas,
Tex.,1974.
Note: Material for the June Facts at Your Finger-
tips was supplied by Dan Summers, tray technol-
ogy manager, Sulzer Chemtech USA Inc.
Side view of a simple tray arrangement
Downcomers
Vapor flow
Liquid flow
M
a
x
.

d
o
w
n
c
o
m
e
r

e
n
t
r
a
n
c
e

v
e
l
o
c
i
t
y
,

f
t
/
s
Delta-density (R
L
R
V
), lb/ft
3
Maximum downcomer velocity guideline
0.45
0.4
0.35
0.3
0.25
0.2
10 20 30 40 50 60
0.15
D
Z
B
W
A
D
Tray geometry parameters
Department Editor: Kate Torzewski
Random Tower
Packing
Packed columns [1]
A packed column is a vertical,
cylindrical pressure vessel con-
taining one or more sections of a
packing material over whose sur-
face the liquid fows downward
by gravity, as a flm or as drop-
lets, between packing elements.
Vapor fows upward through the
wetted packing, contacting the
liquid and facilitating absorption
of the vapor into the liquid.
Packings are offered in either ran-
dom or structured designs. Here,
we will focus on random pack-
ings, which are separate pieces
of packing that have a uniform
geometric shape. Instead of be-
ing arranged in a structured way,
they are dumped or randomly
packed into the column shell.
design
considerations [2]
Size. Random packings are typically avail-
able in diameters of 13.5 in. Generally,
as packing size increases, mass-transfer
effciency and pressure drop decrease.
By this correlation, for a given column
diameter, an optimal packing size can be
determined that represents a compromise
between achieving low pressure drop and
high mass-transfer rates. A rule of thumb
that must also be taken into account is to
choose a packing diameter that is less than
one-eighth of the column diameter, which
minimizes liquid channeling.
Material. Metal packings are usually
preferred because of their superior strength
and good wettability. Ceramic packings,
on the other hand, have superior wettability
but inferior strength, and are used only in
situations at elevated temperatures where
corrosion resistance is needed and plastics
would fail. Plastic packings, usually made
of polypropylene, are inexpensive and
have suffcient strength; however, they may
experience poor wettability, especially at
low liquid fowrates.
Packing objectives [1]
1. Maximize the specifc surface area. In-
creasing the surface area per unit volume
maximizes the vapor-liquid contact area,
and, therefore, effciency. Effciency gener-
ally increases as the random packing size
is decreased.
2. Spread the surface area uniformly. This
improves vapor-liquid contact, and there-
fore, effciency. For instance, a Raschig
ring and a Pall ring of identical size have
identical surface areas per unit volume,
but the Pall ring has a superior spread of
surface area and therefore gives much
better effciency.
3. Maximize the void space per unit
column volume. This minimizes resistance
to gas upfow, thereby enhancing packing
capacity. Capacity increases with random
packing size. This poses a trade off,
however, in that the ideal size of packing
is a compromise between maximizing ef-
fciency and maximizing capacity.
4. Minimize friction. An open shape
minimizes friction, providing good aerody-
namic characteristics.
5. Minimize costs. Packing costs, as well
as the requirements for packing supports
and column foundations, generally increase
with the weight per unit volume of packing.
Packings generally become cheaper as the
size of random packing increases.
Packing structures [3,2]
Raschig Rings are hollow cylinders with
a height that is equal to the ring diam-
eter. This structure is the oldest form of
random packing.
The original saddle-shaped packings, Berl
Saddles, have a smaller free-gas design
than Raschig Rings. However, they are of-
ten a more preferable choice, as they offer
a lower pressure drop and higher capacity.
The invention of the Intalox Saddle marked
the start of the second generation of
random packings. When packed
together, they prevent signifcant
portions of wetting liquid from
being blocked off, thus avoiding
pools of liquid, trapped gas and
violent directional changes of
gas. They offer higher capacity,
higher effciency and lower pres-
sure drop than Berl Saddles.
The Intalox Saddle was further
improved into the Super Intalox
Tower Packing, which has scal-
loped edges and holes in the
material. This allows further
liquid drainage, the elimina-
tion of stagnant pockets, and
more open area for vapor rise,
thus providing higher capacity
and efficiency.
Pall rings are modifed Raschig
Rings that have windows cut and
bent inward. This lowers friction
while improving packing area
distribution, wetting and liquid
distribution. This design allows
higher capacity and effciency than all
previously developed packings.
The next generation of packings features
through-fow structures of a lattice-work
design. The Metal Intalox IMTP offers the
best features of packings that preceded it,
combines the high void fraction and the
well-distributed surface area of the Pall
ring with the low aerodynamic drag of the
saddle shape.
Similar in structure to the Pall Ring is the
Cascade Mini-Ring, which has a height
to diameter ratio of 1:3 compared to 1:1
in the Pall Ring. This allows the individual
packing components to be oriented with
their open side facing vapor fow, thus re-
ducing friction and exposing more surface
to mass transfer.
The latest generation of random packings
features a very open, smooth and wave-like
geometry that promotes wetting, but still
promotes recurrent turbulence. This allows
a decreased pressure drop while sustain-
ing mass-transfer effciency that may be
independent of column diameter, and may
allow a greater depth of packing without a
liquid redistributor.
References
1. Perrys Chemical Engineers Handbook, 8th
ed. McGraw Hill, New York, 2008.
2. Seader, J. D. and Henley, E. J., Separation
Process Principles, 2nd ed., John Wiley and
Sons, Inc., New Jersey, 2006.
3. Schweitzer, P., Handbook of Separation
Techniques for Chemical Engineers, 3rd ed.,
McGraw Hill, New York, 1997.
Facts at your Fingertips sponsored by:
Source: Separation Process Principles, 2nd ed., John Wiley and Sons, Inc.
Department Editor: Scott Jenkins
Hansen Solubility
Parameters (HSP)
I
n recent years, a number of factors have
further emphasized the importance of
solvent selection in the chemical process
industries (CPI). These include the increas-
ing regulatory scrutiny leveled on industrial
solvents, and an effort by CPI companies to
reduce the environmental, health and safety
(EHS) impact of their solvent use. In addi-
tion, specialized, high-performance solvents
are available for some applications [1].
Key tools for helping determine which sol-
vents are suitable for particular applications
have been predicated on the concept that
thermodynamic afnity between solvent and
solute can help predict solution behavior.
Conceptually, when assessing whether
or not a molecule will dissolve in a solvent,
engineers consider the difference between
the energy cost of disrupting the intermo-
lecular attractions of the solvent and solute
molecules, compared to the energy gained
through solute-solvent interactions. As
Hansen and Abbott explain [2], to dissolve
something, you are essentially making a
hole in the solvent, and that takes energy.
If the solute-solvent interactions are greater
than the sum of the energy losses from the
two substances, then solute will dissolve.
THE FOUNDATION OF HSP
Hildebrand [3] developed a thermody-
namics-based, one-dimensional solubility
parameter in 1936 as a tool for eliminating
certain solvents from consideration in pro-
cess development. The Hildebrand solubility
parameter (represented by lower-case delta)
for a pure liquid substance is dened as the
square root of the cohesive energy density
(the amount of energy needed to remove a
volume of molecules from their neighbors
to an innite distance). The same energy
needs to be overcome for molecules to
separate from each other and be surround-
ed by solvent. Equation (1) represents the
Hildebrand solubility parameter.
(1)
Where H
v
is the heat of vaporization, V
m

is the molar volume and RT is the ideal gas
term. The units are (MPa)
1/2
.
The Hildebrand method formed the founda-
tion for a later multidimensional solubility
parameter system that has been rened over
time by Charles Hansen [2]. The resulting
Hansen Solubility Parameters (HSP), have
become a useful tool in selecting solvents for
chemical processes.
THE HANSEN MODEL
HSP are a set of three values that describe
the thermodynamics of dissolving one
substance into another. Solvents with similar
Hansen solubilities generally are miscible
in most proportions, and Hansen values
that are different mean limited solubility. As
Hansen puts it, The HSP describe whether
things prefer to be near each other or not.
HSP attempt to characterize the energies
and assign values to them as a way to
predict solution behavior. Hansens model
essentially splits the Hildebrand solubility
parameter into three components:
* Dlsperslon lorces [Von der Vools lnleroc-
tions) are the attractions experienced by
atoms placed at close distances
* Polor [dlpole| lnlerocllons ore eleclronlc
forces due to the positive-negative attrac-
tion of molecular dipoles
* Hydrogen bondlng ls o porllculor lype
of polar interaction, but can also be
considered a form of electron exchange,
according to Hansen and Abbott [1]
The three components of the HSP are respre-
sented by Equation (2).

(2)
The three terms (in units of pressure)
comprising the HSP are calculated for
each substance through empirical solubility
experiments of various types, depending
on the material being measured. The HSP
values are known for a large number of
solutes. Also, molecular dynamics methods
exist that can estimate HSP without the need
for experimental data [4].
Once determined, the values are then
treated as coordinate points in a three-di-
mensional space, called Hansen space, (see
Figure 1), with axes corresponding to each
of the three HSP components. The nearer
to each other that two molecules are in this
three-dimensional space, the more likely they
are to dissolve into each other. To calculate
the distance between the HSP of two sub-
stances (R
a
), the following equation is used:
(3)
Equation (3) was developed by plotting
experimental data, and represents solubility
data as a sphere, where solvents inside the
sphere are effective at dissolving a given
substance, and those outside are progres-
sively less so.
Polymer solubilities
For solubilities of polymers, Hansen
introduced a concept called the interaction
radius (R
0
). Solubility data determined ex-
perimentally on a given polymer for a range
of solvents allow engineers to estimate a
volume of solubility in the three-dimension-
al Hansen space. For most polymers, this
volume of solubility is spherical in shape if
the scale of the dispersion axis relative to
the other axes is doubled. Solvents whose
data values lie within the sphere will gener-
ally dissolve the polymer, while those lying
outside the sphere will not.
Once an R
0
value is established, the ratio
of R
a
to R
0
can be determined. This ratio
has been called the relative energy distance
[RED| number.
* Syslems wllh RED volues less lhon one ore
sufciently alike and will dissolve
* RED volues equol lo one meon lhe syslem
will partially dissolve
* RED greoler lhon one meons lhe syslem
will not dissolve
Applications of HSP
HSP are useful in selecting process solvents,
additives in formulations, for the blending
of polymers, and for the control of kinetics
and monomer-sequence distributions in
co-polymers. HSP have been used as the
basis for solvent selection models and
computerized methods, but they can also be
used in manual calculations. The following
represent some of the applications in which
HSP have been used:
* Predlcllng ellecllve solvenls lor polymers
* Sludylng dlsperslon ol plgmenls
* Delermlnlng solublllly properlles ol corbon
nanotubes
* Flndlng solvenl blends lhol ore cheoper,
perform better or that have smaller nega-
tive environmental impact
Limitations of HSP
HSP are often best used for screening pur-
poses, with experimental data then used to
validate (or not) the predictions. Limitations
of the HSP that are discussed by Hansen
include the following:
* lhe poromelers con be dllhcull lo meosure
* Moleculor slze ploys o slgnlhconl role ln
whether a substance dissolves within a
given time period
* lhe HSP ore on opproxlmollon
* Poromelers vory wllh lemperolure
Ongoing work by Hansen, discussed in [2]
has begun to address HSP limitations. O
References
1. Vypych, G., lmporlonl Delermlnonls ol Solvenl
Selection, Chem. Eng., June 2006, pp. 5460.
2. Hansen, C. and Abbott, S., Hansen Solubility
Parameters in Practice, eBook, www.hansen-
solubility.com, 2009.
3. Hildebrand, J.H., The Solubility of Non-electro-
lytes, Reinhold, New York, 1936.
4. Belmares, M., Blanco, W.M., Goddard III, A.
and others. Hildebrand and Hansen solubility
parameters from molecular dynamics with ap-
plications to electronic nose polymer sensors.
Journal of Computational Chemistry: 25 (15),
pp. 1,8141,826, 2004.
25
D H
P
20
20
25 27.5
25
22.5
20
15.5
15
15
15
10
10
5
5
0
FIGURE 1. HSP values for dispersion (D),
polar (P) and hydrogen bonding (H) factors
are plotted within a 3D space

You might also like