Meunier Sarah PDF

DevelopmentofaPackedbedReactor
ContainingSupportedSolgelImmobilized
LipaseforTransesterification
by
SarahM.Meunier
Athesis
presentedtotheUniversityofWaterloo
infulfillmentofthe
thesisrequirementforthedegreeof
DoctorofPhilosophy
in
ChemicalEngineering
Waterloo,Ontario,Canada,2012
SarahM.Meunier2012
IherebydeclarethatIamthesoleauthorofthisthesis.Thisisatruecopyofthethesis,includingany
requiredfinalrevisions,asacceptedbymyexaminers.
Iunderstandthatmythesismaybemadeelectronicallyavailabletothepublic.
ii
ABSTRACT
Theobjectiveofthisworkwastodevelopanovelenzymeimmobilizationschemeforsupportedlipase
solgelsandtoevaluatethepotentialoftheimmobilizedbiocatalystfortheproductionofbiodieselina
packedbedreactor.Twosourcesoflipase(EC3.1.1.3triacylglycerolhydrolase)wereusedinthisstudy
andthetransesterificationofmethanolandtrioleintoproduceglycerolandmethyloleatewasusedasa
modelreactionofbiodieselproduction.Acommerciallyavailableformofimmobilizedlipase,Novozym
435,wasusedasabasisforcomparisontotheliterature.
Upon establishing a lipase solgel formulation technique, the experimental methodology for the
transesterification reaction using Novozyme 435 was developed. Subsequently, a series of inert
materialswereconsideredbasedontheirsuitabilityassupportsforimmobilizedlipasesolgelsandthe
synthesisofmethyloleate.Thevalueofasupportedlipasesolgelistoimprovetheactivityandstability
oftheenzymeanddevelopanimmobilizedbiocatalystthatispracticalforuseunderpackedbedreactor
conditions. Of the six support materials considered (612 mesh silica gel, Celite R633, Celite R632,
CeliteR647,anionexchangeresin,andQuartzelfelt),thediatomaceousearthsupports(CeliteR633,
R632 and R647) exhibited high enzymatic activity, were thermally stable, and possessed high solgel
adhesion.
From the three types of diatomaceous earth considered, Celite R632 supported lipase solgels were
identified as the most promising supported lipase solgels for methyl oleate production via
transesterification.Uponfurtherevaluation,theCeliteR632lipasesolgelswerefoundtoachievehigh
methyl oleate percent conversions, glycerolwater absorption was only significant at glycerol levels
higherthan75%,andtheimmobilizedlipasehadhighstabilityuponstorageat4Cfor1.5years.
To determine the effects of methanol and glycerol inhibition as well as temperature on the reaction
kinetics, a pingpong bibi kinetic model was developed and validated over a range of methanol
concentrationsandtemperatures.Theoptimalmethanolconcentrationfortheconditionstestedwasin
the range of 1.3 M to 2.0 M, and increased with increasing temperature. The model developed was
consistent with the experimental data and confirmed that glycerol inhibition and the presence of
productshadsignificanteffectsonthereactionkinetics.
iii
ThemethyloleateproductioncapabilitiesoftheCelitesupportedlipasesolgelwereinvestigatedusing
apackedbedreactorandcomparedwithNovozym435undersimilaroperatingconditions.Akinetic
and mass transfer based model was developed for the reactor system using a novel efficiency
correlationtoaccountfortheeffectofglycerolontheenzymaticactivity.Increasingtheflowrate(1.4
mL/min to 20 mL/min) increased the reaction rate, presumably due to the reduction of the glycerol
inhibitioneffectontheimmobilizedbiocatalyst.TheCelitesupportedlipasesolgelwasfoundtohave
superior performance over Novozym 435 both under batch stirred tank reaction conditions and in a
packed bed reactor (83% conversion for Celite solgel vs. 59% conversion for Novozym 435 at 20
mL/mininthepackedbedreactor).
Based on the results obtained, Celite supported lipase solgels exhibited good performance for the
transesterification of triolein with methanol to produce methyl oleate in both batch and packed bed
reactors,andwarrantfurtherexplorationfortheenzymaticproductionofbiodiesel.
iv
ACKNOWLEDGEMENTS
TherearesomanypeoplewhohavemadethisjourneypossibleIdliketothankyouall.Inparticular,I
wouldliketoextendmysinceregratitudetothefollowingpeople:
Tomysupervisor,RaymondLegge,thankyouforgivingmetheopportunitytodomyPhDandforallthe
support,guidance,andadviceyouhaveprovidedalongtheway.
ThankyoutomyPhDexaminationcommitteemembers,Dr.AmarjeetBassi,Dr.ChristineMoresoli,Dr.
EricCroiset,andDr.KesenMa,foryourvaluableinputonmyresearchandthesis.
To Trevor Williams and Amin Rajabzadeh, thank you for collaborating with me and providing valuable
insight on the research completed in Chapters 7 and 8. Trevor helped collect the experimental data
used in Chapters 7 and 8, and Amin contributed to the development of the kinetic and mass transfer
modelsusedinChapters7and8.
Thank you to the Faculty and Staff in the Department of Chemical Engineering at the University of
WaterlooforthemanywaysyouhavehelpedmakemyPhDasuccess.SpecialthankstoRalphDickhout
whoseanalyticalexpertisehassavedmesomuchfrustration,RonNeillwhobuilttheapparatusforthe
packedbed,andBertHabicherwhohelpedmakeallthepiecesfittogether.
ThankyoutotheNaturalSciencesandEngineeringResearchCouncilforthefinancialsupportprovided
byaNSERCDoctoralPostgraduateScholarship;totheUniversityofWaterlooforthefinancialsupport
provided by Presidents Graduate Scholarships, a University of Waterloo Graduate Scholarship, and a
Provost Doctoral Entrance Award for Women; and to the Ontario Student Assistance Program for the
financialsupportprovidebyanOntarioGraduateScholarship.
To all those who have reviewed my thesis partially or completely, including the reviewers of my
publishedwork,thankyouforyourinsight,ithasimprovedmythesistremendously.
ThankyoutomyfriendsandcolleagueswhowerealwaysaroundtogivemetheexpertiseandsupportI
needed.SpecialthankstoRachelCampbellforbeingbymysidethroughallthephasesofthisjourney.
Finally,thankyoutomyfamily,theMeuniersandtheMercers,forhavingunlimitedfaithinme.
DedicatedtoMarc,Flix,ThodoreandNicolas
Youaremyworld.
vi
Contents
AUTHORSDECLARATION_________________________________________________________________ ii
ABSTRACT____________________________________________________________________________iii
ACKNOWLEDGEMENTS___________________________________________________________________ v
DEDICATION __________________________________________________________________________vi
ListofFigures________________________________________________________________________xi
ListofTables_________________________________________________________________________xv
ListofAbbreviations__________________________________________________________________xvii
Nomenclature______________________________________________________________________xviii
Chapter 1 Introduction_______________________________________________________________ 1
1.1
Objectives___________________________________________________________________ 1
1.2
ThesisOrganization ___________________________________________________________ 2
Chapter 2 LiteratureReview___________________________________________________________ 3
2.1
Biodiesel____________________________________________________________________ 3
2.2
Lipase______________________________________________________________________ 5
2.3
Solgelentrapment____________________________________________________________ 6
2.4
Immobilizationofsolgelsonsupportmaterials_____________________________________ 9
2.5
Biodieselproduction _________________________________________________________ 10
2.5.1
Enzymescreening________________________________________________________ 12
2.5.2
Optimization____________________________________________________________ 13
2.5.3
Glycerolremoval ________________________________________________________ 16
2.5.4
Alcoholdeactivation______________________________________________________ 16
2.5.5
Kineticsandreactionmechanism ___________________________________________ 17
2.6
Reactordesign______________________________________________________________ 21
2.6.1
Packedbedreactors______________________________________________________ 21
2.6.2
Modelling______________________________________________________________ 23
Chapter 3 ExperimentalBackground ___________________________________________________ 26

3.1
Lipaseentrapmentinasolgelmatrix____________________________________________ 26
3.2
Thermalpropertiesofsolgelentrappedlipase ____________________________________ 28
vii
3.3
MethyloleatesynthesiswithNovozym435______________________________________ 30
3.4
Solgelmethyloleateformation ________________________________________________ 34
Chapter 4 StudyofSupportMaterialsforSolgelImmobilizedLipase _________________________ 37

4.1
Introduction________________________________________________________________ 38
4.2
Experimental _______________________________________________________________ 39
4.2.1
Materials ______________________________________________________________ 39
4.2.2
Methods_______________________________________________________________ 40
4.3
Resultsanddiscussion________________________________________________________ 42
4.3.1
Surfacecoating__________________________________________________________ 42
4.3.2
Coatingproperties_______________________________________________________ 44
4.3.3
Enzymaticproperties_____________________________________________________ 46
4.4
Conclusions_________________________________________________________________ 48
4.5
Acknowledgements __________________________________________________________ 49
Chapter 5 EvaluationofDiatomaceousEarthasaSupportforSolgelImmobilizedLipasefor
Transesterification___________________________________________________________________ 50
5.1
Introduction________________________________________________________________ 51
5.2
Experimental _______________________________________________________________ 52
5.2.1
Materials ______________________________________________________________ 52
5.2.2
Methods_______________________________________________________________ 53
5.3
5.3.1
Surfacemorphology______________________________________________________ 55
5.3.2
Physicalproperties_______________________________________________________ 57
5.3.3
5.4
Conclusions_________________________________________________________________ 59
5.5
Chapter 6 EvaluationofDiatomaceousEarthSupportedLipaseSolgelsasaMediumforEnzymatic
TransesterificationtoProduceBiodiesel__________________________________________________ 62
6.1
Introduction________________________________________________________________ 63
6.2
Experimental _______________________________________________________________ 64
6.2.1
Materials ______________________________________________________________ 64
6.2.2
Methods_______________________________________________________________ 65
6.3
viii
6.3.1
Physicalproperties_______________________________________________________ 67
6.3.2
6.3.3
Adsorptiveproperties ____________________________________________________ 72
6.4
Conclusions_________________________________________________________________ 75
6.5
Chapter 7 KineticModellingoftheProductionofMethylOleatebyCeliteSupportedLipaseSolgels77
7.1
Introduction________________________________________________________________ 78
7.2
Experimental _______________________________________________________________ 80
7.2.1
Materials ______________________________________________________________ 80
7.2.2
Methods_______________________________________________________________ 80
7.3
7.3.1
Comparisonoflipaseimmobilizationschemes_________________________________ 82
7.3.2
Kineticmodelinitialreactionrate__________________________________________ 83
7.3.3
Kineticmodeltimedependentreactionrate_________________________________ 85
7.3.4
Methyloleateconcentrationprofile_________________________________________ 85
7.3.5
Kineticmodelpredictionsfortransesterification_______________________________ 89
7.4
Conclusions_________________________________________________________________ 90
7.5
7.6
Nomenclature_______________________________________________________________ 90
Chapter 8 KineticandMassTransferModellingofMethylOleateProductioninanImmobilizedLipase
PackedBedReactor__________________________________________________________________ 93
8.1
Introduction________________________________________________________________ 94
8.2
Experimental _______________________________________________________________ 96
8.2.1
Materials ______________________________________________________________ 96
8.2.2
Methods_______________________________________________________________ 96
8.3
Modelling__________________________________________________________________ 99
8.3.1
Kineticmodel___________________________________________________________ 99
8.3.2
Packedbedreactormodel_________________________________________________ 99
8.3.3
Physicalproperties______________________________________________________ 100
8.3.4
Packedbedefficiencycorrelation__________________________________________ 101
8.4
Resultsanddiscussion_______________________________________________________ 102
8.4.1
Batchkinetics__________________________________________________________ 102
ix
8.4.2
Experimentalandmodellingresults ________________________________________ 102
8.4.3
ComparisonofCelitesupportedsolgeltoNovozym435______________________ 105
8.4.4
Modelpredictions ______________________________________________________ 106
8.4.5
Catalystreusability______________________________________________________ 106
8.5
Conclusions________________________________________________________________ 107
8.6
Acknowledgements _________________________________________________________ 108
8.7
Nomenclature______________________________________________________________ 108
Chapter 9 ConclusionsandRecommendations __________________________________________ 111

9.1
Mostsignificantcontributions_________________________________________________ 111
9.1.1
ObjectiveAExperimentalBackground_____________________________________ 111
9.1.2
ObjectiveBStudyofSupportMaterialsforSolgelImmobilizedLipase___________ 111
9.1.3
ObjectiveCEvaluationofDiatomaceousEarthasaSupportforSolgelImmobilized
LipaseforTransesterification______________________________________________________ 112
9.1.4
ObjectiveDEvaluationofDiatomaceousEarthSupportedSolgelsasaMediumfor
EnzymaticTransesterificationtoProduceBiodiesel____________________________________ 112
9.1.5
ObjectiveEKineticModellingoftheProductionofMethylOleatebyCeliteSupported
LipaseSolgels _________________________________________________________________ 113
9.1.6
ObjectiveFKineticandMassTransferModellingofMethylOleateProductioninan
ImmobilizedLipasePackedBedReactor_____________________________________________ 113
9.2
RecommendationsforFutureWork ____________________________________________ 114
References ________________________________________________________________________ 115

AppendixA
ListofPublications______________________________________________________ 129
AppendixB
LipaseSourceChange ___________________________________________________ 130
AppendixC
SampleCalculations_____________________________________________________ 133
ListofFigures
Figure2.1:Transesterificationofatriglyceridewithanalcoholtoproducefattyacidalkylestersandglycerol ____3
Figure2.2:Lipasestructurewiththehelicalloopcircled:(a)closedlidand(b)openlid.AdaptedfromKrebsand
Gerstein(2000). __________________________________________________________________________5
Figure2.3:SolgelformationschemeadaptedfromJinandBrennan(2002)._______________________________8
Figure2.4:Intermediatereactionsforthetransesterificationoftriglycerideswithalcoholtoproducefattyacidalkyl
esters and glycerol: (a) triglycerides to diglycerides, (b) diglycerides to monoglycerides, and (c)
monoglyceridestoglycerol.________________________________________________________________19
Figure 2.5: Schematic representation of the pingpong bibi kinetic mechanism for the transesterification of
triglyceridestoproducebiodiesel.AdaptedfromCheirsilpetal.(2008). ____________________________19
Figure3.1:Linoleicacidactivityofthelipasesolgelbasedon91,104and146daysofstorageat4C.Theerror
barsrepresentthestandarderrorwithn=26. _________________________________________________28
Figure3.2:Relativeactivityofthesolgelentrappedlipaseafterincubationat70C.Theerrorbarsrepresentthe
standarderrorwithn=26._________________________________________________________________29
Figure3.3:Linoleicacidconcentrationprofileasafunctionoftimeforsolgelimmobilizedlipaseat70Cand30C
for12h.Theslopesofthelinesrepresenttheinitialreactionrates.________________________________29
Figure 3.4: Normalized final methyl oleate concentration at various methanol molar ratios for 6 h at 40C
(normalizedto theNovozym 435 batch maximum). The error bars represent the standard deviation with
n=25. _________________________________________________________________________________31
Figure3.5:Percentconversionofmethanoltomethyloleateasafunctionofreactiontemperatureafter6hfor
bothsolventfreeandsolventbased(50%hexane)reactionmediawithamethanoltotrioleinmolarratioof
1:1.____________________________________________________________________________________31
Figure3.6:Percentconversionofmethanoltomethyloleateatamethanoltotrioleinmolarratioof1:1insolvent
freeandsolventbased(50%hexane)reactionmediaat40C.Errorbarsrepresentthestandarderrorwith
n=34. _________________________________________________________________________________33
Figure 3.7: Percent conversion of methanol to methyl oleate at a methanol to triolein molar ratio of 1.5:1 in
solventfreeandsolventbased(50%hexane)reactionmediaat40C. ______________________________33
Figure 3.8: Percent conversion of methanol to methanol oleate at a methanol to triolein molar ratio of 1:1 for
solventfreeandsolventbased(50%hexane)systemswithandwithouttheadditionof3%water. _______35
Figure3.9:MethanoltomethyloleateconversionusingimmobilizedlipasefromNovozym435,80%PTMSsolgel
and20%PTMSsolgelwithamolarratiooftrioleintomethanolof1.5:1at40C.Errorbarsrepresentthe
standarderrorwithn=2.___________________________________________________________________35
xi
Figure 4.1: SEM images at 3000x (a) unsupported lipase solgel; silica gel (b) without and (c) with solgel
immobilizedonthesurface;CeliteR633(d)withoutand(e)withsolgelimmobilizedonthesurface;anion
exchangeresin(f)withoutand(g)withsolgelimmobilizedonthesurface;andQuartzelfelt(h)withoutand
(i)withsolgelimmobilizedonthesurface.____________________________________________________43
Figure4.2:Adhesionofsolgel(mgsolgel/gmaterial)andproteinloading(glipase/gmaterial)foreachsupport
material.Supports:SG(silicagel),R633(CeliteR633),R632(CeliteR632),R647(CeliteR647),AE(anion
exchangeresin),andQF(Quartzelfelt).DuetothelowdensityofQF,dataarepresentedonapercm2basis
ratherthanapergrambasis._______________________________________________________________45
Figure4.3:FinalweightpercentofthesupportswithandwithoutsolgelbasedonTGA.Supports:SG(silicagel),
R633(CeliteR633),R632(CeliteR632),R647(CeliteR647),AE(anionexchangeresin),andQF(Quartzel
felt).Errorbarsrepresenta90%confidenceinterval. ___________________________________________45
Figure4.4:KineticprofileforatypicallipidtransesterificationreactionwithCeliteR632asthesupport._______47
Figure 5.1: SEM images at 3000x magnification of (a) uncoated Celite R633, (b) uncoated Celite R632, (c)
uncoatedCeliteR647,(d)CeliteR633coatedwithlipasesolgel,(e)CeliteR632coatedwithlipasesolgel,
and(f)CeliteR647coatedwithlipasesolgel.Arrowsidentifysomeclustersofsolgelonthesurfacesofthe
Celite. ________________________________________________________________________________55
Figure5.2:PercentsolgelcoverageonthesurfaceofCeliteasdeterminedbySEMimageanalysis.Theerrorbars
representthe95%confidenceintervalsofthesamplemeanbasedonn=45.________________________56
Figure5.3:AdhesionofsolgelonthesupportforeachtypeofCeliteconsidered.Theerrorbarsrepresentthe
95%confidenceintervalsofthesamplemeanbasedonn=3._____________________________________58
Figure5.4:Loadingofproteinpergramforeachsupport:A)theamountofproteinpergramofmaterial,andB)the
amountofproteinpergramofsolgel.Theerrorbarsrepresentthe95%confidenceintervalofthesample
meanbasedonn=3. _____________________________________________________________________58
Figure5.5:Percentmethanolconversionpergramofmaterialafter6hforeachsolgelformulation.Theerrorbars
representthe95%confidenceintervalofthesamplemeanbasedonn=9.__________________________60
Figure 5.6: Initial lipase activity for each solgel formulation measured by the initial amount of methyl oleate
produced per minute per gram of protein. The error bars represent the 95% confidence interval of the
samplemeanbasedonn=9._______________________________________________________________60
Figure 6.1: Protein content and degree of protein immobilization for four solgel formulations. Error bars
representa95%confidenceinterval._________________________________________________________68
Figure 6.2: Percent conversion of methanol to methyl oleate based on a 6 h batch reaction for each solgel
formulation without and with a drying step prior to the enzymatic assay. Error bars represent a 95%
confidenceinterval. ______________________________________________________________________70
xii
Figure6.3:Theenzymaticactivityonapergramofmaterialbasisforeachsolgelformulationasdeterminedfrom
theconversionofmethyloleateandtheproteincontentofthegels.Boththesolgelformationswithoutand
withadryingsteppriortothereactionwereconsidered.Errorsbarsrepresenta95%confidenceinterval. 70
Figure 6.4: Average percent conversion of methanol to methyl oleate for solgels after storage at 4C for
unsupported () and Celite R632 supported solgels (o). The lines represent the lines of best fit for the
unsupported solgel (broken) and the Celite R632 supported solgel (solid). Error bars represent a 95%
confidenceinterval. ______________________________________________________________________71
Figure6.5:TypicalTGAprofileincludingthesampleweight(solid)andthederivativeweight(broken).Thepeak
desorptionpointandthetotalmasslossareindicatedonthegraph.ThesampleshownisCeliteR632sol
gel(CSG4)at50%glycerolequilibratingsolution.______________________________________________73
Figure6.6:PeakdesorptionratesforCeliteR632solgels()andplainCeliteR632(o)basedondifferentlevelsof
glycerolintheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval. _________________73
Figure6.7:TotalpercentagemasslossforCeliteR632solgels()andplainCeliteR632(o)basedondifferent
levelsofglycerolintheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval.__________74
Figure6.8:PeakdesorptiontemperaturesforCeliteR632()solgelsandplainCeliteR632(o)basedondifferent
levelsofglycerolintheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval.__________75
Figure 7.1: Initial reaction rate as a function of methanol concentration at three temperatures (40C, 50C, and
60C). The lines represent the predictions from the model provided by Equation 7.2 and the symbols
representtheexperimentaldata.____________________________________________________________83
Figure 7.2: Methyl oleate concentration at 6 h with respect to the initial methanol concentration at three
temperatures(40C,50C,and60C).Thelinesrepresentthemodelpredictionsandthesymbolsrepresent
theexperimentaldata.ThemainplotrepresentstheresultsfromEquation7.3whiletheinsetplot(upper
left)representstheresultsfromEquation7.2. _________________________________________________84
Figure 7.3: Correlation between the predicted and experimental methyl oleate concentration values using the
kineticmodeldescribedbyEquation7.3.______________________________________________________86
Figure 7.4: Comparison of model predictions (lines) and experimental data (symbols) for four methanol
concentrations(0.3M,0.6M,0.9M,and1.8M)andthreetemperatures(40C,50C,and60C).Plota)T=
40C,b)T=50C,c)T=60C,d)[A]=1.8M. ___________________________________________________88
Figure7.5:Surfaceplotofthekineticmodel(Equation7.3)predictionsforthemethyloleateconcentrationafter6
hatdifferentinitialmethanolconcentrations(04M)andtemperatures(3070C).____________________89
Figure8.1:Schematicofthepackedbedreactorapparatus.___________________________________________98
Figure8.2:Packedbedresultswiththeexperimental(symbols)andmodel(lines)methyloleateconcentrationsasa
functionofcontacttimeinthereactorwithCelitesupportedsolgelsat4flowrates(1.4mL/min,5mL/min,
12.5mL/minand20mL/min).Theconfidenceintervalsshownareestimatedfromapooledvarianceof16
samplesintriplicate._____________________________________________________________________103
xiii
Figure8.3:Packedbedresultswiththeexperimental(symbols)andmodelling(lines)methyloleateconcentrations
asafunctionofcontacttimeinthereactorforNovozym435atflowratesof5mL/minand20mL/min.The
confidenceintervalsshownareestimatedfromapooledvarianceof16samplesintriplicate. __________105
Figure8.4:Contourmapformodelpredictionsformethyloleateconcentrationwithrespecttotheflowrateand
contacttimebasedonkineticsandmasstransfer. _____________________________________________106
Figure8.5:Reusabilityrunsinthepackedbedreactorover5daysatT=40C,substrateratio1:1,andflowrate=5
mL/min. A new substrate mixture (triolein and methanol) was supplied to the existing CSG bed at the
beginningofeachday. ___________________________________________________________________107
xiv
ListofTables
Table2.1:Resultsoflipasescreeningstudiesforenzymaticbiodieselproduction.__________________________11
Table2.2:Enzymaticbiodieselproductionreactionoptimizationliteratureresults._________________________13
Table3.1:ReportedandmeasuredactivitiesforNovozym435,80%PTMSsolgeland20%PTMSsolgel.______36
Table4.1Propertiesofthelipasesolgelsupportmaterials.Porediameterandparticlesizeareasprovidedbythe
suppliers.Initialsolgelcapacityisdefinedastheamountofliquidsolgelrequiredtofullywetthesupport
withoutexcessmaterial.___________________________________________________________________40
Table4.2:Enzymaticpropertiesofthesupportedlipasesolgels:percentconversionofmethanoltomethyloleate
(MO)after6hofreaction,initialreactionrate,andenzymaticactivity.ForQuartzelfelt,theconversionand
reactionratearepresentedintermsofcm2ratherthangduetothelowdensityofthefeltmaterial.Allerror
rangesarebasedonthe90%confidenceintervalofreplicatedexperiments._________________________47
Table5.1:ParticlesizeandporediameterforeachtypeofCelitebasedonthemanufacturersspecifications,and
surface areas of Celite without lipase solgel (support) and with lipase solgel (coated). The confidence
limitsindicatedrepresentthe95%confidenceintervalofthemeanbasedonn6. ___________________56
Table5.2:QuantificationofthesolgelclustersfoundoneachoftheCelitesamples.45imagesforeachtypeof
Celitewereusedforthisdetermination._____________________________________________________56
Table 5.3: Effect of the presence of silica and Celite on methyl oleate production with Novozym 435 and
unsupportedsolgelcontaininglipase.________________________________________________________59
Table 6.1: Description of the lipase preparations used for analysis in terms of the support material, solgel
formulation,andlipasesolutionconcentration. ________________________________________________66
Table7.1:ComparisonofNovozym435(N435),freelipase(Free),unsupportedsolgel(US),andCelitesupported
solgel (CSG) in terms of their final percent conversion, initial reaction rate, and initial activity at 40C and
methanoltotrioleinmolarratioof1.5:1.*ForN435,theinitialactivityisbasedontheweightofN435rather
thantheweightoflipasesincetheamountoflipaseintheenzymepreparationisunknown.____________82
Table7.2:KineticparameterestimationforthemodelspresentedinEquations7.2and7.3._________________84
Table7.3:Comparisonoftheperformanceofthetwoproposedmodels(Equations7.2and7.3)intermsoftheir
SSE,R2,andRMSEforthe6hmethyloleateconcentration._______________________________________86
Table 8.1: Kinetic constants determined by fitting the results of the batch experiments to Equation 8.1 for the
Celite solgel (CSG) and the Novozym 435 (N435) immobilized enzyme preparations. Ethanolysis kinetic
constantsfromtheliteratureusingN435areincludedforcomparison(Chesterfieldetal.,2012). _______101
Table 8.2: Parameters used to model the experimental packed bed reactor results for each of the immobilized
biocatalysts. ___________________________________________________________________________103
xv
Table 8.3: Summary of some of the parameters calculated from the modelling of the packed bed reactor
experiments.___________________________________________________________________________104
Table8.4:Comparisonofsolventfreebiodieselpackedbedreactorperformanceusingimmobilizedlipase. ___104
xvi
ListofAbbreviations
AE
Anionexchangeresin
BCA
Bicinchoninicacid
BET
Brunauer,EmmettandTellermethod
BSA
Bovineserumalbumin
BSTFA
N,OBis(trimethylsilyl)trifluoroacetamide
CSG
Celitesupportedlipasesolgel
DOI
Degreeofimmobilization
FAME
Fattyacidmethylester
Free
Freelipase
GCMS
Gaschromatographymassspectrometry
HDAME
heptadecanoicacidmethylester
HPLC
Highperformanceliquidchromatography
LA
Linoleicacid
MO
Methyloleate
N435
Novozym435
PDA
Pentadecanoicacid
PTMS
Propyltrimethoxysilane
QF
Quartzelfelt
Coefficientofdetermination
R632
CeliteR632
R633
CeliteR633
R647
CeliteR647
RMSE
Rootmeansquarederror
SAA
Surfaceareaanalyzer
SEM
Scanningelectronmicroscopy
SG
Silicagel
SSE
Sumofsquarederrors
TGA
Thermogravimetricanalysis
TMOS
Tetramethoxysilane
TMS
Trimethylsilyl
US
Unsupportedlipasesolgel
UV
Ultraviolet
xvii
Nomenclature
Reactorcrosssectionalarea(m2)
Bulkconcentrationofsubstrate (mol/m3)
Concentrationof atthecatalyticsurface(mol/m3)
Diffusionofsolute insolvent (m2/s)
Diffusionofmethanolintriolein(m2/s)
Particlediameter(m)
Reactordiameter(m)
Temperatureconstant(dimensionless)
Masstransfercoefficient(m/min)
Equilibriumconstant(dimensionless)
Inhibitionconstantforcomponent
Batchreaction(M).Continuousreaction(mol/m3)

Michaelisconstantforcomponent
Molarmassofcomponent (g/mol)
Volumetricflowrate(m3/min)
ReactionratewithrespecttosubstrateA(mol/m3min)
Reynoldsnumberofthereactor(dimensionless)
Reynoldsnumberoftheparticle(dimensionless)
Externalsurfaceareaofthecatalyst(m2/m3)
Schmidtnumber(dimensionless)
Sherwoodnumber(dimensionless)
Catalystsurfacearea(m2/kg)
xviii
Time(min)
Temperature(K)
Superficialmolarvelocity(m/min)
Reactionvelocity(M/min)
Molalvolume (m3/mol)
Initialreactionvelocity(M/min)
Maximuminitialreactionvelocity
Batchreaction(M/min).Continuousreaction(mol/m3min)
Maximumvelocityoftheforwardreaction(M/min)
Maximumvelocityofthereversereaction(M/min)
Proteinweightinthereactor(kg)
Concentrationofcomponent ,where: isalcohol, istriolein, ismethyloleate,
isglyceroland isenzyme
Positionalongthelengthofthereactor(m)
GreekLetters
Efficiencycorrelation(dimensionless)
Constantsfortheefficiencycorrelation( =1,2,3,4,or5)(dimensionless)
Reactorbedvoidfraction(dimensionless)
Dynamicviscosity(Pas)
Kinematicviscosity(cSt)
Densityoftheliquid(g/cm3)
Bulkcatalystdensity(kg/m3)
Densityofcomponent (kg/m3)
Dissociationfactorforthesolvent(dimensionless)
xix
Chapter 1
Introduction
Theresearchdescribedinthisthesisfocusesontheimmobilizationoflipaseinsupportedsolgelsand
the evaluation of the immobilization scheme based on the ability of the supported biocatalyst to
producemethyloleatethroughthetransesterificationoftrioleinandmethanol.Themainfocuswasthe
investigationofthesupportmaterialdevelopedandthepotentialeffectsofthesupportmaterialonthe
productionoffattyacidmethylesters.
1.1
Objectives
Themainobjectiveoftheresearchwastodevelopalipaseimmobilizationschemeforuseasacatalyst
packing material in a packed bed reactor for the production of methyl oleate from triolein and
methanol.Sixsubobjectiveswereidentifiedtoachievethismainobjective:
A. To entrap lipase in solgels and to evaluate the enzymatic activity and stability of the
immobilizedlipase,aswellastoevaluatetheirpotentialtoproducemethyloleateviaenzymatic
transesterification.
B. To evaluate a variety of inert supports for the solgel entrapped lipase for use ultimately in a
packedbedreactor.
C. Tofullycharacterizethemostpromisingsupportmaterialsintermsoftheefficacytoproduce
methyloleateviaenzymatictransesterification.
D. To evaluate the potential of the optimal support material and to determine any evident
shortcomingsfortheenzymaticproductionofmethyloleate.
E. Tomodelthetransesterificationkineticsofthesupportedlipasesolgelinabatchreactorwith
theprimaryvariablesbeingtemperatureandmethanolconcentration.
F. Todevelop,model,andoptimizeapackedbedreactorcontainingsupportedlipasesolgeland
to compare the methyl oleate production capabilities of the supported lipase solgel with
respecttoacommerciallyimmobilizedlipase.
1.2
ThesisOrganization
Thisthesiscontainsfourmainparts:theliteraturereview,theexperimentalbackground,aseriesoffive
chapters addressing the research objectives, and the overall conclusions with recommendations for
futurework.
Theliteraturereview(Chapter2)isacomprehensivereviewoftheworkthathasbeenpublishedwith
respecttolipase,solgelentrapment,supportedsolgels,biodieselproduction,andreactordesignand
modelling.
The experimental background (Chapter 3) lays the foundation for immobilizing lipase in solgels,
assessing their activity and stability, producing methyl oleate via the enzymatic transesterification of
trioleinwith thecommerciallyimmobilizedlipaseNovozym435,anddeterminingwhetherlipasesol
gelshavepotentialfortheproductionofmethyloleate.
Themainresultschaptersencompassaseriesofselfcontainedstudies,eachwithauniqueintroduction,
materialsandmethods,resultsanddiscussion,conclusions,andreferences.Chapter4isanevaluation
of a seriesof lipase solgel support materials to determine the optimal material for the production of
methyl oleate via enzymatic transesterification. Chapter 5 is an evaluation of three types of
diatomaceousearthandwithadetailedcomparisontoelucidatethebestsupportmaterialforlipasesol
gels.Chapter6isaninvestigationofCeliteR632supportedsolgelswithanevaluationoftheeffectof
protein loading and drying on the methyl oleate production capacity. It also evaluates the potential
effectofglycerolandthestabilityoftheimmobilizedenzyme.Chapter7focusesonthedevelopmentof
a kinetic model for the Celite supported solgels with attention on the effects of temperature,
methanol concentration, and glycerol inhibition. Chapter 8 describes the performance of the Celite
supported solgel in a packed bed reactor in comparison to the commercially immobilized lipase
Novozym435throughexperimentaldataandmodelling.
Chapter9presentstheoverallconclusionsofthethesisandrecommendsadditionalareasofresearch
thatwouldcomplementtheresearchdescribed.
Chapter 2
LiteratureReview
Thefollowingsectionsprovideareviewofthecurrentstateofknowledgeintheareasoflipase,solgel
based enzyme entrapment, immobilization of solgels on support materials, enzymatic biodiesel
production,reactordesign,andmodellingasitisrelevanttotheresearchpresentedinthisthesis.
2.1
Biodiesel
Biodiesel is a mixture of fatty acid alkyl esters that can be produced via the transesterification of
triglycerides. Transesterification is a catalytic reaction that converts triglycerides into fatty acid alkyl
estersandglycerolinthepresenceofanalcoholasshowninFigure2.1.Biodieselisanenvironmentally
friendly fuel that emits low exhaust emissions and can be used either on its own or blended with
petroleum diesel for use in unmodified engines (Tan et al., 2010). Since biodiesel can be produced
entirely from vegetable oils or animal fats, it is both renewable and biodegradable (Vasudevan and
Briggs,2008),anditisconsideredanalternativeenergysource(JaegerandEggert,2002).Themajority
of biodiesel is produced using alkaline catalysts (sodium or potassium hydroxide) due to the short
reactiontimerequired,butthisprocessisenergyintensive,glycerolrecoveryisdifficult,thecatalystis
complicatedtoremovefromtheproduct,wastewaterisproducedwhichmustbetreated,andfreefatty
acidsandwaterinhibitthereaction(Fukudaetal.,2001;Meheretal.,2006;Ranganathanetal.,2008;
Vasudevan and Briggs, 2008). Acid catalysts (sulfuric, hydrochloric, or sulfonic acid) are also used to
produce biodiesel, but these reactions typically have a low reaction rates, the acidic environment is a
challengeforreactordesign,andhighalcoholtooilratiosarenecessary(VasudevanandBriggs,2008).
CH2-OOC-R1
CH-OOC-R2
CH2-OH
+
3 ROH
catalyst
CH2-OOC-R3
3 R1-3-COO-R
CH-OH
CH2-OH
Figure2.1:Transesterificationofatriglyceridewithanalcoholtoproducefattyacidalkylestersandglycerol
Therefore, alternative catalysts that overcome the disadvantages of acidic and alkaline catalysts are
desirable.Transesterificationcanalsobecatalyzedbytheenzyme,lipase,sothereissignificantinterest
intheenzymaticproductionofbiodiesel.Someofthebenefitsofenzymaticbiodieselproductionover
chemically catalyzed biodiesel production are that it is less energy intensive and more efficient, it
produces less waste and byproducts, it involves milder operating conditions (temperature and pH),
enzyme and product recovery are less complicated, and there is no soap (fatty acid salt) formation
(Akoh et al., 2007; Fjerbaek et al., 2009; Fukuda et al., 2001; Marchetti et al., 2007; Vasudevan and
Briggs,2008).
Thechallengesoflipasecatalyzedbiodieselproductionincludethepotentialforproductandsubstrate
inactivationoftheenzymeandthehighcostofenzymes(Akohetal.,2007;Fukudaetal.,2001;Ganesan
etal.,2009;Marchettietal.,2007;RoblesMedinaetal.,2009;VasudevanandBriggs,2008).Thereis
currently considerable interest in immobilizing lipase for biodiesel production because immobilization
improves the biocatalyst thermal and chemical stability, and facilitates recovery and reuse of the
enzyme which in turn decreases the costs associated with the enzyme (Akoh et al., 2007; Bajaj et al.,
2010;Fukudaetal.,2001;JaegerandEggert,2002;Macarioetal.,2009;VasudevanandBriggs,2008).
Biodieselproductionusingimmobilizedlipaseisknowntobelesscostlythanthatusingfreelipase,and
thecostsassociatedwithenzymaticbiodieselproductionareexpectedtobecompetitivewiththoseof
alkalinebiodieselproductioniftheimmobilizedlipaseisrecoveredandreusedmorethanfivetimesor
thecostassociatedwiththeenzymeisreducedsignificantly(Jegannathanetal.,2011).
Solgels, as immobilization supports for enzymes, have been shown to improve the thermal and
chemicalstabilityaswellastheactivityoflipase(Aucoinetal.,2004;Pirozzietal.,2009;Reetzetal.,
1996;Reetz,1997).Inaddition,lipaseentrappedsolgelshavethepotentialtobeimmobilizedoninert
supports or carriers which could be used in enzymatic bioreactors. By immobilizing solgel entrapped
lipase on as a thin film on an inert support, the diffusion limitations that are common for solgel
immobilizedsystemscanbeovercome(Brnyiketal.,2000;Oraireetal.,2006;Pogorilyietal.,2007)
whileatthesametimefacilitatingthereuseoftheenzyme(Oraireetal.,2006).

Figure2.2:Lipasestructurewiththehelicalloopcircled:(a)closedlidand(b)openlid.AdaptedfromKrebsandGerstein
(2000).
2.2
Lipase
Lipase (E.C. 3.1.1.3) is a triacylglycerol hydrolase which catalyzes the hydrolysis of ester bonds of
triglycerides at lipid/water interfaces such as the reaction shown in Figure 2.1. Some of the
biotechnological applications for lipases are for the production of biopolymeric substances and
biodiesel,fortriglycerideandbloodcholesteroldeterminationinenzymecatalyzedbiosensors,forthe
synthesisoftherapeuticstoaiddigestion,andfortheproductionofherbicides,cosmeticlotionsandoils,
andperfumescontaining menthol(Hasanetal.,2006;JaegerandEggert,2002).Inaddition,theyare
also used in the pulp and paper industry for wastepaper deinking, in the textile industry for desizing
(removing size lubricants) and enzyme wash treatments (stone washing), in household and industrial
detergents to remove oils, and in the food processing industry to modify oils, improve flavour, and
removefatsfrommeats(Hasanetal.,2006).
Athydrophobicinterfaces,lipaseundergoesaconformationalchangecalledinterfacialactivation.Prior
to contact with a hydrophobic interface, the enzyme has limited catalytic activity; at hydrophobic
interfaces, the movement of an helical loop (or lid) uncovers the enzyme's active site and the
enzymatic activity is dramatically increased (Sarda and Desnuelle, 1958). The closed and openlid
structuresareshowninFigure2.2where(a)representstheclosedlidconfigurationand(b)represents
theopenlidconfigurationandthehelicalloopiscircled.
Duetotheinterfacialactivationphenomenonexhibitedbylipases,itiscrucialtomaintainahydrophobic
interfaceduringlipasecatalyzedreactions.Thepresenceofwaterduringenzymatictransesterification
andinteresterificationreactionsisnecessaryduetotheformationofaliquidliquidinterfaceinvolving
waterandtheoilysubstratewheretheenzymaticreactionoccurs.Theadditionofasmallamountofa
watermisciblesolvent,suchasethanolormethanol,mayalsoprovidetheinterfacenecessaryforthe
reaction in addition to acting as a reaction substrate. However, as the concentration of the
water/solventisincreasedsufficientlythereisaninhibitoryeffectontheenzymaticactivitysincewater
hinderstheinteractionbetweentheenzymeandsubstrate(AlZuhair,2005;Fukudaetal.,2001;Hsuet
al.,2001a;RoblesMedinaet al.,2009;Samukawa et al.,2000; Shimada etal.,2002;Watanabeetal.,
2001). Therefore, the activity of lipase is strongly influenced by the nature of the interface, the
interfacialproperties,andtheinterfacialarea(Akohetal.,2007).
2.3
Solgelentrapment
Immobilization facilitates enzyme recovery and reuse thereby decreasing the cost associated with the
enzyme, and in some cases also improves the thermal stability, chemical stability and activity of the
enzyme(Akohetal.,2007;Bajajetal.,2010;Fukudaetal.,2001;JaegerandEggert,2002;Macarioet
al., 2009; Vasudevan and Briggs, 2008). There are a variety of methods for enzyme immobilization
which can be classified as immobilization by either binding or physical retention (Hartmeier, 1988).
Immobilizationbybinding,whichissubdividedintobindingtocarriersandcrosslinking,tendstoleadto
amorestableimmobilizedbiocatalystwithloweractivityduetothebondbetweentheenzymeandthe
carrier (Hartmeier, 1988). Alternatively, immobilization by physical retention, which is further
subdivided into matrix entrapment and membrane enclosure, does not involve a bond between the
enzymeandthecarriersotheenzymecanbeimmobilizedwithoutsignificantlyalteringtheenzymeand
thereforemaintainingtheenzymaticactivity(Hartmeier,1988).However,masstransferlimitationsare
oftenaconcernwhenimmobilizationbyphysicalretentionisconsidered(Hartmeier,1988).Therefore,
for any type of enzyme immobilization, a balance between activity and stability is desired. Although
there is a broad spectrum of methods for immobilizing enzymes, including covalent attachment and
adsorption, enzyme entrapment in polymers is the most desirable as it provides a stable enzyme
support with stronger bonds than those that are attainable via adsorption, and it does not require
complexchemicalpreparationasincovalentattachment(Reetzetal.,1996).
The immobilization of enzymes in polymers, such as solgels, diminishes the inhibition of lipase in the
presenceofalcohols(Hsuetal.,2001a).Inadditiontowatermisciblesolventsensitivity,lipasesarealso
sensitive to high temperatures and pressures which can cause irreversible enzyme
denaturation (Noel and Combes, 2003). According to Reetz et al., solgel entrapped lipases are
chemically and thermally stable and have very high activities up to 88fold enhanced esterification
activity(Reetzetal.,1996;Reetz,1997).Inadditiontoimprovingthermalandchemical stability,and
enzymatic activity and selectivity, the immobilization of enzymes is also beneficial due to the ease of
enzyme reuse, continuous process control, product separation, and minimal effluent and materials
handlingchallenges(Gupta,1993).
SomeofthekeychallengesofenzymeentrapmentaccordingtoJinandBrennan(2002)are:
Ensuringthattheproteinitselfandtheprotein'sactivesiteareaccessibletothesubstrate.
Preventing the protein from becoming denatured during the immobilization process by
temperature,pH,ornonaqueousconditions.
Maintaining the polymer matrix pore size so that the substrate can easily diffuse into the
polymer, the hydrolysis products can diffuse out of the matrix, and the enzyme remains
entrappedwithinthepolymer.
Tuning the material properties of the immobilization material to maximize the activity of the
entrappedenzymeandthediffusionofanalytes.
Developingamaterialthatisstraightforwardandreproducibletofabricateinavarietyofforms
basedonthedesiredfunction.
Inaddition,thenatureoftheprotein,precursors,andadditivesusedtoformthesolgelmayaffectthe
homogeneity, orientation, aggregation, and interaction of the enzyme with the solgel. The substrate
and product interactions with the solgel matrix such as electrostatic, hydrogen bonding, and
hydrophobic interactions which affect the accessibility of the enzyme to the substrates may also be
important(JinandBrennan,2002).
Duringthepasttwodecades,solgelentrapmentofenzymeshasseensignificantadvancesandhasbeen
thesubjectofavarietyofreviews(Avniretal.,1994;GillandBallesteros,2000;JinandBrennan,2002;
Livage et al., 2001; Pierre, 2004; Reetz, 1997). Solgels are formed through the hydrolysis and
subsequentcondensationofsilicaprecursorsasshowninFigure2.3.Thefirststep,precursorhydrolysis,
forms a colloidal suspension (sol) which, under the appropriate conditions, can be crosslinked upon
gelation (condensation). Finally, the solgel is strengthened and conditioned with a series of aging,
drying,crushing,andwashingsteps.
Additives
polymers,surfactants,
proteinstabilizers,fillers
Biomolecules
proteins,
DNA,cells
PrecursorSol
Hydrolysis
Condensation
Polycondensation
Gelation
Solgelformationwithentrappedbiomolecules
Aging
Condensationandpolycondensationcontinue
Solventevaporation
Solgelshrinkage
Figure2.3:SolgelformationschemeadaptedfromJinandBrennan(2002).
In order for an enzyme to be entrapped within the polymer matrix, it must be added between the
hydrolysisandcondensationstagessincetheacidicconditionsnecessaryforhydrolysiswilldenaturethe
enzymeandthesolgelmatrixmustdeveloparoundtheenzyme.Someoftheparametersthatcanbe
adjustedtoachievethedesiredentrapmentpropertiesarethewatersilaneratio,solvent,precursorand
catalystproperties,agitationconditions,andenzymeentrapmenttime(Pierre,2004).
Hydrophobicinterfaces,suchasmicellesofsubstratesandimmiscibleorganicsolvents,stimulatethe
helical loop conformation of lipases which permits substrate accessibility to the active site (Bastida et
al., 1998; Galarneau et al., 2006). Therefore, solgel materials with high hydrophobicity provide
excellent supports for lipase immobilization if the lipase is located at the hydrophobic/hydrophilic
interfaceupongelationitispossibletoimmobilizetheenzymeinitsactiveconformation(Reetzetal.,
1996).
A variety of applications have been developed for solgel immobilized enzymes, including selective
coatings for biosensors (optical and electrochemical), affinity chromatography stationary phases,
immunoadsorbent and solidphase extraction materials, solid phase biocatalysts, controlled release
agents (as in drug delivery), biophysical study matrices, fine organic chemical synthesis, and
biocompatibleprostheses(JinandBrennan,2002;Pierre,2004;Soaresetal.,2006).
2.4
Immobilizationofsolgelsonsupportmaterials
Several groups have studied the immobilization of solgels on support materials. Oraire et al.
consideredlipaseimmobilizedinsilicaaerogelsreinforcedwithsilicaquartzfiberfelts(Nassreddineet
al., 2008; Oraire et al., 2006). The gels were most active when composed of 40%
methyltrimethoxysilane and 60% tetramethoxysilane due to the increase in hydrophobic groups. The
enzymaticactivitywascomparabletothecommerciallyimmobilizedlipase,Novozym435,forbiodiesel
production in the presence of an organic solvent; however, in the absence of solvent the diffusion
limitationsinthesilicaaerogelsresultedinalessactivepreparation(Oraireetal.,2006).
Similarly,Lietal.consideredtheimmobilizationofpapainoncottonfabricusingtetraethylorthosilicate
solgels(Lietal.,2007).Inthiscase,thefabricwasimmersedinthesolandenzymesolutionwhichwas
subsequentlydriedontothefabric.Theresidualenzymaticactivityaftersixcontinuoususesforthesol
gel immobilized papain was 30% a significant improvement over the adsorption immobilization
schemewhichhadlessthan20%residualactivityaftertwoconsecutiveuses.Thethermostabilityofthe
enzymewasnotaffectedbytheimmobilizationprocedureandtheimmobilizedlipasewaslesssensitive
topHthanthenativepapain(Lietal.,2007).
Topreparehighlystableandactivebiocatalystsforcommercialuse,Pogorilyietal.developedadouble
immobilization technique where urease was immobilized using the solgel procedure on a silica gel
(Pogorilyi et al., 2007). In this case, a polysiloxane layer containing the enzyme was formed on the
inorganicsupport.Theactivityincreasedwhenhydrophobicgroupsandpolyvinylalcoholwereusedin
the solgel preparation, and the activity also increased as the thickness of the polysiloxane layer
decreased (Pogorilyi et al., 2007). The enzyme was more accessible to the substrate in films and,
therefore,itcanbeinferredthatdiffusionlimitationsinthexerogelsresultedinloweractivity(Pogorilyi
etal.,2007).
Comparingsolgelswithpolyurethanefoamandceramicfoaminpackedbedandfluidizedbedreactors,
tetraethoxysilane solgels with whole cell enzymes performed inferiorly for phenol degradation in
comparisontobothceramicandpolyurethanefoamsduetotheirmasstransferlimitations(Brnyiket
al.,2000).
Enzymeentrappedsolgelshaveenhancedenzymaticpropertieswhendiatomaceousearthisusedasa
supportmaterial(Furukawaetal.,2000;Furukawaetal.,2002a;Furukawaetal.,2002b;Furukawaand
Kawakami,1998;Kawakami,1996;KawakamiandYoshida,1996;Kawakamietal.,2003;Koszelewskiet
al., 2010). In one such study, transaminases entrapped in Celite solgels demonstrated enhanced
activity despite high pH levels and high temperatures which are typically detrimental to enzymatic
activity(Koszelewskietal.,2010).Inaddition,thisimmobilizationschemeallowedthecatalystmaterial
toberecycledeighttimeswithoutcausinganadverseeffectontheachievableconversion(Koszelewski
et al., 2010). When lipase was supported on Celite 545, higher activity, thermal stability, and
reusabilitywereexhibitedincomparisontobothfreelipaseandlipasesupportedonAmberliteIRA938
(Sairolu and Telefoncu, 2004; Sairolu, 2008; Sairolu et al., 2004). Comparing lipase solgels
immobilized on Celite 545 with lipase bound directly to Celite 545, the solgel entrapment scheme
hadimprovedactivityandthermostability(Kawakami,1996;KawakamiandYoshida,1996).
Althoughsolgelimmobilizedenzymesarepromisingforstabilityandreuse,therearemanyhurdlesto
overcomebeforetheyseepracticalapplication.Masstransferlimitationsarecommonoccurrencesin
solgelimmobilizedsystems(Bajajetal.,2010;Brnyiketal.,2000;McDue1991;Oraireetal.,2006;
Pogorilyietal.,2007).Preparingenzymesinsmallparticlesizesoronouterlayersofcarriermatricesare
commonmethodstoimprovereactorperformance(McDuffie,1991).Therefore,immobilizationofsol
gelsasfilmsoninorganicsupportsappearstobeareasonablesolutiontoovercomethischallenge.
2.5
Biodieselproduction
Biodiesel production is a very active area of research with numerous reviews recently published
(Abbaszaadehetal.,2012;Atadashietal.,2012;ChouhanandSarma,2011;Fanetal.,2012;Gogetal.,
2012; Semwal et al., 2011; Shahid and Jamal, 2011; Tan et al., 2010; Yusuf et al., 2011; Zhang et al.,
2012). The production of biodiesel using lipase has advantages over chemically (acid/base) catalyzed
processesbecauseitislessenergyintensive,moreefficient,andhighlyselective;inaddition,enzymatic
biodieselproductioninvolvesmildoperatingconditions,haslittledownstreamprocessing,producesless
wasteandbyproducts,anddoesnotinvolvesoapformation(Akohetal.,2007;Marchettietal.,2007;
Vasudevan and Briggs, 2008). However, some challenges associated with enzymatic biodiesel
production are the potential for enzyme inhibition by the alcohol substrate, inhibition by the glycerol
product, and the high cost of enzymes (Akoh et al., 2007; Fukuda et al., 2001; Vasudevan and Briggs,
2008).Immobilizedenzymesareundergoingsignificantresearchforbiodieselproduction,andhavethe
10
potential to be competitive with chemical catalysts for commercial biodiesel production (Zhang et al.,
2012).
Table2.1:Resultsoflipasescreeningstudiesforenzymaticbiodieselproduction.
LipasesConsidered
Results
Reference
LipozymeTLIMandNovozym435 Novozym435wasthemostrobustcatalyst.
(Hernndez
Martnand
Otero,2008)
Novozym435andLipozymeTM
IM
Novozym435achievedthehighest
conversions,butamixtureof60%Novozym
435and40%LipozymeTMIMwasoptimal.
(Huangetal.,
2010)
C.antarctica,P.cepacia,andT.
lanugunosus(immobilized)
P.cepaciaandT.lanugunosushadthe
highestreactionratesinalkanesolvents
(Gagnonand
Vasudevan,
2011)
P.fluorescens,P.cepacia,M.
javanicus,C.rugosaandR.niveus
(freeandimmobilized)
C.rugosa,P.camembertii,P.
roqueforti,P.cepacia,P.
fluorescens,C.lypolyticaandK.
oxytoca.
Novozym435andLipozymeIM60
T.lanuginosus,P.fluorescens,B.
cepacia,P.camembertii,and
porcinepancreaticlipase
Novozym435,LipozymeRMIM,
PSCandPSD
C.viscosum(freeandimmobilized),
C.rugosaandporcinepancreatic
lipase
R.delemar,A.niger,F.
heterosporum,Novozym435and
LipozymeIM6
Free:R.oryzae,C.rugosa,P.
camembertii,P.cepaciaandP.
fluorescens.Immobilized:
LipozymeRMIM,LipozymeTLIM,
C.antarcticaAandBandR.miehei.
P.fluorescenshadthehighestactivity.
Immobilizedenzymeshadimprovedactivity
overfreeenzymes.
C.rugosa,P.fluorescensandP.cepaciahad
highcatalyticactivity.P.cepaciawasthe
mostmethanolresistantandwasless
dependentontheamountofwaterpresent.
Novozym435wasmostactive.
P.fluorescenshadthehighestactivity.
(Isoetal.,2001)
(Kaiedaetal.,
2001)
(Laietal.,2005)
(Moreiraetal.,
2007)
PSD(immobilizedondiatomaceousearth)
wasthemostactiveforbiodieselproduction.
C.viscosumwastheonlylipasetohave
appreciablebiodieselyield.ImmobilizingC.
viscosumincreasedtheyieldby10%.
(Salisetal.,
2005)
Novozym435wasthemosteffectivelipase
formethanolysis.
(Shimadaetal.,
1999)
ThelipasewiththehighestconversionwasP.
fluorescens.LipozymeRMIMalsoachieved
highconversion.
(Soumanouand
Bornscheuer,
2003a)
Amixtureof1/3Novozym435and2/3
Novozym435,LipozymeTLIM,and
LipozymeRMIM(byweight)hadtheoptimal
LipozymeRMIM
performance.
11
(Shahetal.,
2004)
(Yceland
Demir,2012)
2.5.1
Enzymescreening
One of the challenges that has hampered the development of enzymatic biodiesel production at
industrialscaleisthehighcostoflipase(JaegerandEggert,2002).Toavoidcostconstraints,theactivity
of the lipase must be both enhanced and prolonged, and the lipase must be tolerant to the desired
solventsandsubstrates(Fukudaetal.,2001).Toaddresstheseconcerns,severalstudiesscreenlipases
basedontheirabilitytoproducebiodiesel,andimmobilizelipaseonsupportmaterialssothatitcanbe
moreeasilyreused.Table2.1presentsasummaryofvariouslipasescreeningstudies.
Comparingthesestudies,thebestlipaseforbiodieselproductionisquitespecifictothenatureofthe
reaction being performed, for example the alcohol and oil substrates used and the amount of water
presentinthesystem.Thecommerciallyimmobilizedlipasesconsideredinthesestudiestypicallyhave
high activities in comparison to the free lipases. The demonstrated advantages of immobilized lipase
includeimprovedreusabilityandstability.
Manynoncommercialimmobilizedlipaseshavealsobeenconsidered.Forexample,aporouskaolinite
immobilization medium improves lipase activity (Iso et al., 2001), and cotton membranes are good
immobilization supports (Nie et al., 2006). Kumari et al. compared a variety of free lipases,
commerciallyimmobilizedlipaseonanionexchangeresins,lipaseimmobilizedonamicroporousresin,
crosslinkedenzymeaggregates,andproteincoatedmicrocrystals(2007).Fromthatstudy,theprotein
coated microcrystals had the highest percent conversion, but involved a very complicated formation
procedure.
Finally,severalgroupsconsideredsolgelimmobilizedlipase.Hsuetal.achievedhighconversionsand
consideredprocessoptimizationwithphyllosilicateclaysolgelsfromcetyltrimethylammoniumchloride
and tetramethylorthosilicate (Hsu et al., 2001b; Hsu et al., 2003; Hsu et al., 2004). The immobilized
lipase reacted more slowly, but it achieved higher conversion, was more reusable, more thermally
stable,andnotaffectedbymethanolinhibitionincomparisontothefreelipase(Hsuetal.,2001;Hsuet
al.,2003;Hsuet al.2004).Similarly,lipaseimmobilizedinaTMOS(tetramethylorthosilicate)andiso
BTMS (isobutyltrimethoxysilane) solgel had good methanol resistance, good reusability and more
activity than free lipase (Noureddini et al., 2005). Oraire et al. considered silica aerogels reinforced
with silica quartz fibre felt and dried with supercritical carbon dioxide (Oraire et al., 2006). As in the
other studies, solgel immobilized lipase exhibited improved reusability and higher activity than free
lipase,butathighsubstrateconcentrationsaseverediffusionlimitationwasnotedduetotheplugging
oftheaerogelpores(Oraireetal.,2006).Moreiraetal.(2007)alsousedsolgelstoimmobilizelipase
12
tetraethoxysilane (TEOS) and poly(vinyl alcohol) (PVA) solgels had high activity at elevated
temperatures (50oC) and high ethanol to oil molar ratios (18:1). The solgels were much more active
than free lipase, and the viscosity of the biodiesel produced was comparable to that of commercial
diesel(Moreiraetal.,2007).
2.5.2
Optimization
Aswellasscreeningforthebestsourceandpreparationoflipaseforbiodieselproduction,considerable
current research focuses on optimizing the reaction conditions, such as the substrate molar ratio,
solventpresenceandtype,temperature,watercontent,substrateflowrate,andtypeofacylacceptor,
toachievethehighestpercentconversionandreusabilityoftheenzyme(Akohetal.,2007).Table2.2
provides a review of the major results gathered from some current studies on the optimization of
reactionparametersforbiodieselproduction.
Table2.2:Enzymaticbiodieselproductionreactionoptimizationliteratureresults.
Lipase
Novozym435
Parameters
Oil;substrate
ratio;enzyme
amount;
temperature
Alcohol;enzyme
Calciumalginate
amount;time;
immobilizedR.
temperature;
oryzae
substrateratio
Results
Optimalconditions:methanol:oil=3.8:1,
100%wastefryingoil,15wt%enzyme,
and44.5C.
Reference
(Azcaretal.,
2010)
Optimalconditions:methanol,30C,1:3
(oil:alcohol),24h,and10wt%enzyme:oil. (Balasubramaniam
Pureenzymehadhigherconversionthan
etal.,2012)
thewholecellimmobilizedbiocatalyst.
Lipoprime50T
Solvent;
temperature;
substrateratio;
time
Methanolysisinnhexanewasslowwhile
intbutanolhadhighreactionrates.
Optimalconditions:40C,oil:alcohol=1:6
1:8,and1.5h.
(Bendikieneetal.,
2011)
Immobead150
immobilized
Solvent
Alkanesolventshadhighyieldsandrates.
Isooctaneandnhexanewereoptimal.
(Gagnonand
Vasudevan,2011)
Phyllosilicate
solgel
immobilized
Alcohol;enzyme
amount
Methanol,ethanolandnbutanolhadthe
highestconversion,andtheoptimal
enzymeamountwas150mg/mL.
(Hsuetal.,2001b)
Continuedonnextpage
13
Table2.2continued
Lipase
Parameters
Results
Reference
Phyllosilicate
solgel
immobilized
Temperature;
solvent
Optimalreactiontemperature:freelipase:
40Candimmobilized:40C70C.
(Hsuetal.,2003)
Hexanehadhigheryieldthansolventfree.
Phyllosilicate
solgel
immobilized
Temperature;
flowrates;time
Optimalconditionsusingarecirculating
packedcolumnreactorwere50C,30
mL/minflowrate,and48h.
(Hsuetal.,2004)
Novozym435
andLipozym
TMIMmixture
Enzymeamount;
enzyme;solvent
amount;
substrate
amount;time
Optimalconditions:4wt%enzyme:oil,
49%Novozym435,55vol%tertbutanol
tooil,5.12:1methanol:oil,and20h.
(Huangetal.,2010)
Solvent;
temperature;
watercontent
Methanolandethanolrequiredsolvent,
but1propanoland1butanoldidnot.1
propanolhadthehighestsolventfree
activity.Optimalconditions:50Cand0.3
wt%water.
(Isoetal.,2001)
R.oryzae
Enzymeamount;
watercontent
Reactionrateincreasedwithincreasing
enzymeamountuntil25IU/mL.Without
water,lipasewasinactiveandinsufficient
watercausedirreversibleinactivation.
(Kaiedaetal.,
1999)
Novozym435
Enzymeamount;
substrateratio;
temperature;
time
Optimalconditionsusingrefinedcotton
seedoilandmethanolwere30%enzyme,
1:4oil:alcohol,30C,and7h.
(Kseetal.,2002)
Solgel
immobilized
Substrateratio;
temperature
Optimalconditionsusingpalmoiland
ethanolwere1:18oil:alcoholand58oC.
(Moreiraetal.,
2007)
Solgel
immobilized
Temperature;
substrateratio;
watercontent;
enzymeamount;
alcohol
Optimalconditions:35C,1:15.2
oil:ethanol,1:7.5oil:methanol,0.5gwater
(Noureddinietal.,
(methanol)or0.3gwater(ethanol),and
2005)
475mglipase(methanol).Methanol
reducedthereusabilityoftheenzyme.
Porouskaolinite
immobilized
Continuedonnextpage
14
Table2.2continued
Lipase
Parameters
Results
Reference
Novozym435
Temperature;
enzymeamount;
substrateratio;
wateraddition;
acylacceptor
Optimalconditions:45C,3wt%enzyme,
3:1methanol:oil,andnowater.Methyl
acetateimprovedthelipasehalflifewhile
maintaininghighyield.
(Ognjanovicetal.,
2009)
Commercial
immobilized
Temperature;
water;substrate
ratio;alcohol
Optimalconditions:40C,0.40.6water
activity,1:3and1:6oil:alcohol.Butanol
wasthemostefficientalcohol.
(Salisetal.,2005)
Commercialfree Temperature;
andimmobilized solvent
Optimaltemperature:40C50C,hexane
wasagoodsolvent,andsolventfree
reactionsweresuccessful.
(Soumanouand
Bornscheuer,
2003a)
Commercialfree
Oil;alcohol
andimmobilized
Thehighestconversionoccurredwith
palmoleinandmethanolwasthebest
alcohol.
(Soumanouand
Bornscheuer,
2003b)
LipozymeRM
IM
Temperature;
enzymeamount;
substrateratio
Optimalconditionsinaclosedbatch
reactor:67C,4.5wt%enzyme,and1:2
palmiticacid:ethanol.
(Vieiraetal.,2006)
Commercial
immobilized
lipases
Substrateratio;
temperature;
time
Optimalconditions:40C,10h,and1:12
oil:methylacetate.
(Xuetal.,2003)
Temperature;
Optimalconditionswiththreestepalcohol
LipozymeTLIM enzymeamount; additionof1molareach:40C,10wt%
(Xuetal.,2004)
alcohol
enzyme,andmethanol.
Novozym435,
LipozymeTLIM,
andLipozyme
RMIM
Time;
temperature;
enzymeamount;
substrateratio
Optimalconditionsvarydependingon
lipaseused:3.96.2h;39.555.8C;4.4
21.2%enzyme;1.6:14.1:1oil:alcohol.
15
(YcelandDemir,
2012)
Basedontheseresults,thebestoperatingconditionsaredependentonthetypeoflipase,alcohol,and
oilusedforthereaction.Immobilizedlipaseatapproximately40oCwithanoiltomethanolmolarratio
of1:3(thestoichiometricratio)areconsistentlysuccessfulreactionconditions.
Chang et al. used a statistical approach to determine the optimal reaction parameters using response
surfacemethodologywitha5level5factorcentralcompositerotatabledesign(Changetal.,2005).The
optimal reaction parameters for the reaction of canola oil with methanol using Novozym 435 in a
hexanereactionmediumwere38oC,12.4hourreactiontime,42.3wt%enzyme,1:3.5oiltomethanol
ratio,and7.2wt%water.Thepredictedconversionwas99.4%,whiletheactualachievedconversionfor
thisreactionwas97.9%.Theseresultsarecomparabletothosepreviouslyreportedintheliterature.
2.5.3
Glycerolremoval
Thepresenceofglycerolinthereactionmediummaycauseenzymeinhibition(Samukawaetal.,2000;
Vasudevan and Briggs, 2008). Removal of glycerol by dialysis using an ultrafiltration flat sheet
membrane for continuous methanolysis was very successful, and increasing the removal of glycerol
improved the conversion of the reaction significantly 20% conversion without glycerol compared to
10%conversionwith5.0gglycerolinthereactionmedium(BlafiBaketal.,2002).Usingamembrane
separationtechnique,asopposedtothetypicalglycerolremovalviasettling,ismuchmorepracticalfor
acontinuousbiodieselproductionprocess(BlafiBaketal.,2002).
Thepresenceofsilicabedshasalsobeenusedtoadsorbglycerolinbiodieselstreams(Mazzierietal.,
2008; Samukawa et al., 2000; Yori et al., 2007). One complication of this process is that one of the
reaction substrates, methanol, causes the glycerol saturation capacity of the silica gel to be reduced
(Mazzierietal.,2008)andglyceroltodesorbfromthesilicagel(Yorietal.,2007).Also,silicageladsorbs
methanolwhichcanalsoleadtoareducedbiodieselyield(Wangetal.,2006).
2.5.4
Alcoholdeactivation
To minimize the tendency for lipase to be inhibited by methanol and ethanol substrates, several
procedureshavebeenconsideredincludinglipasepretreatment,alternativeacylacceptorstoreplace
methanol,andstepwiseadditionofmethanol(BlafiBaketal.,2002;ChenandWu,2003;Duetal.,
2004;RuzichandBassi,2010a;RuzichandBassi,2010b;RuzichandBassi,2011;Samukawaetal.,2000;
Shimadaetal.,1999;Shimadaetal.,2002;Watanabeetal.,2000;Watanabeetal.,2001;Watanabeet
al.,2002;Xuetal.,2003;Xuetal.,2004).
16
Lipasepretreatmentbyincubationwithmethyloleateandsoybeanoilfortwelvehourshelpedprevent
deactivation of Novozym 435 by methanol, increased the initial reaction rate, reduced the effect of
wateronthereactionrate,andhelpedpreventactivitylossevenaftertwentyuses(Samukawaetal.,
2000).AnothermethodthathelpedreduceNovozym435inhibitionbymethanolandethanolwasa
pretreatment by immersion with tertbutanol which increased the fatty acid methyl ester yield from
2.5%to24.5%(ChenandWu,2003).Inaddition,periodicregenerationoftheenzymewitha2butanol
ortertbutanolwashallowedcontinuoususeforseventydayswhilemaintainingtheconversionabove
70% (Chen and Wu, 2003). The conclusion of this study was that the alcohol is adsorbed onto the
immobilized enzyme support thereby blocking the oil substrate from reaching the reaction site and
consequentlypreventingthereactionfromprogressing(ChenandWu,2003).
Anotherapproachtopreventingalcoholdeactivationoflipaseinvolvesusingmethylacetateastheacyl
acceptor as opposed to methanol or ethanol. Du et al. (2004) showed that an oil to methyl acetate
molarratioof1:12canbeusedwithNovozym435withoutdeactivatingtheenzymeforbothcrudeand
refinedsoybeanoil,andthereisnoactivitylossina0.5Lbioreactorafter100cycles.Inasimilarstudy,
methylacetateasanacylacceptorgaveahighermethylesteryieldandminimalactivityloss(Xuetal.,
2003).
The most common approach for preventing methanol inactivation of lipase in biodiesel production is
threestep methanolysis. In these studies, methanol deactivation is prevented by adding methanol in
threestepsofonemoleofmethanolpermoleofoileachtoachievethestoichiometricratioofthree
molesofmethanolpermoleofoil.Inonestudy,continuousmethanoladditionhadthebestconversion
of 97% (BlafiBak et al., 2002). Several studies have shown that both threestep methanolysis with
onemoleratioofmethanolineachstep(Shimadaetal.,1999;Watanabeetal.,2000;Watanabeetal.,
2001;Watanabeetal.,2002;Xuetal.,2004)andatwostepprocesswithonemoleratioofmethanolin
thefirststepandtwomoleratiosofmethanolin thesecondstep (Shimadaetal.,2002)bothhelpto
preventmethanolinhibition.
2.5.5
Kineticsandreactionmechanism
Thekineticsofenzymaticbiodieselproductionhasnotbeenwidelystudied.AlZuhair(2005)considered
thekineticsanddevelopedamathematicalmodelbasedonthereactionmechanismwithvegetableoil
as a substrate. The model was compared to the experimental results from an ionexchange resin
immobilizedlipaseandsilicagelimmobilizedlipasewithreasonableagreementwiththeinitialreaction
rate (AlZuhair, 2005). A pingpong mechanism was used with Michaelis Menten kinetics, and, in
17
contrast to many other studies, both the substrates (oil and alcohol) could be studied independently
sinceanorganicsolventwasusedtokeepthebulkvolumeconstant.Theinhibitioneffectsoftheoiland
alcohol were dependent on the immobilization support and as the oil concentration increased, the
inhibitioneffectofthealcoholdecreased(AlZuhair,2005).Asubsequentstudyusingfreelipaserather
thanimmobilizedlipasefoundthatthemodelunderestimatedtheinhibitioneffectsofbothsubstrates,
andthatthereactionwasmoreinhibitedbythealcoholthantheoil(AlZuhairetal.,2007).Comparing
solventfreeandnhexanebasedreactionmediausinglipaseimmobilizedonceramicbeadsusingping
pong bibi kinetics with competitive inhibition by both substrates, a higher yield could be achieved
withoutsolventandthattheratedeterminingstepisthesurfacereactionratherthanmasstransfer(Al
Zuhairetal.,2009).
Based on a kinetics study on biodiesel production with methyl acetate as the acyl acceptor and the
immobilized lipase Novozym 435, three consecutive secondorder reversible reactions were describe
fortheinteresterificationoftriglyceridesandmethylacetateandakineticmodelwithapingpongbibi
mechanismwithsubstratecompetitiveinhibitionwasdeveloped(Xuetal,2005).Thethreereactions
are: triglycerides to diglycerides, diglycerides to monoglycerides, and monoglycerides to
triacetylglycerol.Fromthekineticconstants,thefirstreactionstep,triglyceridestodiglycerides,wasthe
ratelimitingstepfortheoverallinteresterificationreaction(Xuetal.,2005).Similarly,threereversible
reactions can be elucidated for the transesterification of triglycerides to fatty acid alkyl esters and
glycerolusinganalcoholratherthanmethylacetate(Figure2.4).
IndepthstudiesofthekineticmechanismofenzymecatalyzedreactionsbyClelandarecommonlyused
asstartingpointsforthekineticstudiesdescribedintheliteratureincludingaseriesofpingpongbibi
mechanisms with substrate and product inhibition (Cleland, 1963a; Cleland, 1963b; Cleland, 1963c).
Lipasecatalyzedtransesterificationhasbeensuccessfullydescribedbyapingpongbibimechanism(Al
Zuhair2005;AlZuhairetal.,2007;AlZuhairetal.,2009;Cheirsilpetal.,2008;Dossatetal.,2002;Xuet
al.,2005).Insuchakineticmechanism,thefirstsubstrate(ester)bindstotheenzyme,formsanenzyme
intermediateandthefirstproduct(alcohol)isreleasedbeforethesecondsubstrate(alcohol)canbindto
the enzyme to form an intermediate and release the second product (ester) (AlZuhair et al., 2007;
Cheirsilpetal.,2008;Rizzietal.,1992;YadavandDevi2004).Figure2.5showsthisreactionmechanism
whenappliedtotheproductionofbiodieselviatransesterification.
18

CH2-OOC-R
(a)
(b)
(c)
CH-OOC-R
CH2-OOC-R
+
ROH
R-COO-R
CH-OOC-R
CH2-OOC-R
CH-OH
CH2-OOC-R
CH2-OOC-R
CH-OOC-R
ROH
R-COO-R
CH-OH
CH-OH
CH-OH
CH2-OOC-R
CH2-OH
CH-OH
ROH
R-COO-R
CH-OH
CH-OH
CH2-OH
Figure2.4:Intermediatereactionsforthetransesterificationoftriglycerideswithalcoholtoproducefattyacidalkylesters
andglycerol:(a)triglyceridestodiglycerides,(b)diglyceridestomonoglycerides,and(c)monoglyceridestoglycerol.
Figure2.5:Schematicrepresentationofthepingpongbibikineticmechanismforthetransesterificationoftriglyceridesto
producebiodiesel.AdaptedfromCheirsilpetal.(2008).
19
Themechanisticstepsareasfollows:thefreeenzyme(E)reactswiththetriglyceride(T)toproducean
enzymetriglyceride complex(E.T)fromwhich thesecond complex(E.D.F)releasesthediglyceride (D).
The third complex (E.F) reacts with the alcohol (A) and releases the fatty acid alkyl ester (F). Similar
mechanismsexistforthediglyceride(D)tomonoglyceride(M)reactionandthemonoglyceride(M)to
glycerol (G) reaction as shown in Figure 2.5. This is consistent with the results obtained using TLC
analysisofthereactionintermediatesbyKaiedaetal.(1999)supportingthenotionthateachesterbond
of the triglyceride undergoes a two step mechanism: first the ester bond is hydrolyzed to produce
partialglyceridesandfreefattyacids,followedbyesterificationofthefreefattyacidswiththealcohol
toproducethefattyacidalkylester.
The initial rate equation with inhibition of one substrate (Equation 2.1) is often used to model the
kineticsoflipasecatalyzedtransesterificationfortheproductionofbiodiesel(AlZuhair,2005;Dossatet
al.,2002;Xuetal.,2005).
where istheinitialreactionvelocity,
concentration,
the triglyceride,
isthemaximuminitialreactionvelocity,
isthetriglycerideconcentration,
(2.1)
isthealcohol
istheapparentMichaelisMentenconstantfor
is the inhibition constant for the alcohol,
is the apparent Michaelis Menten
constantforthealcohol.
AlZuhairetal.consideredtheinitialinhibitioneffectsofbothsubstratesbyinvolvingasolventsothat
the initial oil and alcohol concentrations could be independently varied and studied (AlZuhair et al.,
2007; AlZuhair et al., 2009). The equation used for the kinetic modeling was similar to Equation 2.1
withtheadditionofa
termtotakeintoaccounttheeffectoftriglycerideinhibition.Thesestudies
show comparable inhibition constants for the triglyceride and alcohol which were quite large
(approximately 30004500) indicating that the effect of inhibition was low (the inhibition constant
represents the dissociation of the inhibitor from the enzymeinhibitor complex so smaller numbers
indicatehigherinhibitoryeffects).
Alternatively,acompleterateequation(Equation2.2)forapingpongbibimechanismwithinhibitionof
bothsubstratesandbothproductshasbeendevelopedandcanbeusedasamoresophisticatedmodel
fortheentirereactionratherthansolelytheinitialconditions(Rizzietal.,1992;YadavandDevi,2004).
20
(2.2)
with:
1
1
The reaction for this model is: A+BP+Q; is the reaction velocity;
maximum velocities of the reverse and forward reactions;
concentrationsofcomponentsA,B,P,andQ;
and
, and
istheequilibriumconstant;
aretheMichaelisconstantsforA,B,PandQ;and
and
and the
represent the
,
,and
aretheinhibitionconstantsfor
A,B,PandQ.
2.6
Reactordesign
Some common immobilized biocatalyst continuous reactors are stirred tank reactors, fluidized bed
reactors, and packed bed reactors (Hartmeier, 1988; Messing, 1975). Stirred reactors are simple and
inexpensive,butaremorecommonforaerobicfermentationsthanimmobilizedbiocatalysisbecausethe
intensive stirring introduces unnecessary shearing forces on the immobilized biocatalysts (Hartmeier,
1988).Fluidizedbedreactorshavebedsthatarelooselyfilledwithcatalystparticlesandthesubstrateis
forced upwards through the bed. Although fluidized beds are advantageous for immobilized
biocatalysts,theretentionofbiocatalystparticlesischallengingiftheviscosityofthesubstrateishigh
(Hartmeier,1988).
2.6.1
Packedbedreactors
Packed bed reactors can accommodate the highest density of catalyst particles, and, therefore, the
highestpossiblesubstrateconversionisattainable(Hartmeier,1988).Inaddition,packedbedreactors
21
helpminimizediffusionlimitationsofimmobilizedenzymesbecausethereactorensurespropermixing
between the immobilized catalyst and the reaction medium thereby improving external mass transfer
(Messing, 1975). Packed bed reactors have the simplest reactor design to achieve a high degree of
contactbetweenthesolidcatalystparticlesandtheliquidsubstrates(Thoenes,1994).
Using a packed bed reactor, immobilized enzymes can easily be reused and continuous processing is
feasible.However,onechallengeofapackedbedreactorisbasedonthelifetimeofthebiocatalystit
isimpracticaltoshutdownthereactortochangethecatalystparticlesfrequently(Thoenes,1994).Care
must be taken to ensure the catalyst has a long lifetime within the reactor so it can run continuously
withlittlemaintenance.
Forpackedbedreactors,particlediametersaretypicallybetween3and10mm,andtheheightofthe
reactor is typically between 10 and a few hundred times the particle diameter (Thoenes, 1994).
Continuousenzymaticmethanolysisinpackedcolumnreactorscommonlyutilizescolumnswithaheight
ofapproximately5.5timestheinternaldiameter(Hsuetal.,2003;Nieetal.,2006;Shimadaetal.,1999;
Watanabeetal.,2000).
Several packed bed reactor studies have been completed using the commercially immobilized
Novozym 435. Hama et al. developed a bench scale solventfree packed bed reactor with a glycerol
separatingtankandachievedfinalfattyacidmethylestercontentsabove96%usingeither10passeson
a single reactor bed (Hama et al., 2011a) or 550 h continuous production using five reactors in series
(Hamaetal.,2011b).Similarly,Shimadaetal.andWatanabeetal.developedacontinuous,threestep,
fixedbedreactorwithNovozym435thatachieved98%(Shimadaetal.,1999)and96%(Watanabeet
al., 2000; Watanabe et al., 2001) conversion with good enzyme reusability and without enzyme
deactivation.OtherpackedbedreactorstudiesusingNovozym435alsohavepromisingresults:76%
molar conversion with no activity loss after 7 days (Chang et al., 2009), 83% conversion and 30 days
continuousproductionwithoutconversiondecrease(Chenetal.,2011),and75.2%conversionusinga
tertbutanolcosolvent(Shawetal.,2008).
Considering alternative immobilized lipases, Hsu et al. (2004) successfully developed a recirculating
packedcolumnreactorwithlipaseimmobilizedontenpackagesofcommercialpapercoffeefilters.Nie
etal.(2006)developedacontinuousreactorthatachieved92%conversionbyimmobilizinglipaseona
cottonmembraneforuseinathreestepmethanolysiscontinuousfixedbedreactorwithninecolumns
packedwithimmobilizedlipase.Wangetal.(2001)developedalipaseFe3O4immobilizedoncottonfor
22
useinbothasinglebedreactorandfourpackedbedsinseries.Atextileclothimmobilizedlipasewas
used in a three step packed bed reactor with a hexane solvent and gravity driven glycerol separation
achievinga91%fattyacidmethylesterproduct(Chenetal.,2009).Afterthreehoursinasinglepacked
bed reactor using an immobilized lipase of mixed sources and stepwise methanol addition, 98%
conversionwasachieveddroppingto90%after108hduetoglycerolaccumulationinthereactor(Leeet
al.,2010).
A study using Lipozyme IM20 compared the performance of batch reactors and tubular reactors for
biodiesel production. The tubular reactor required the addition of glass beads to increase the void
space in the reactor, and had higher reactor rates, caused less stress on the lipase, and was more
flexibleintermsofrecyclerates(Mukeshetal.,1993).
2.6.2
Modelling
Few studies exist in the literature dealing with modelling enzymatic packed bed reactors for biodiesel
production. Most optimization studies are purely empirical using response surface methodology to
modelandoptimizetheexperimentaldata(Chenetal.,2011;Ognjanovicetal.,2009;Shawetal.,2008).
Halimetal.(2009)usedacombinationofresponsesurfacemethodologyandmasstransferstudiesto
optimize an immobilized lipase packed bed reactor for biodiesel production. First, the experimental
datawasmodellingusingresponsesurfacemethodologyandplotsofareactioncontrolledmodelwith
thatofamasstransfercontrolledmodelweresubsequentlycomparedtodeterminewhichfitthedata
better.Theauthorsfoundthatthemasstransfercontrolledmodelhadbetterfit(R2=0.9441)thanthe
reaction controlled model (R2=0.7184) indicating the importance of considering the mass transfer in
additiontothereactionkinetics.
AccordingtoFogler(1999),thereactionandflowinapackedbedcanbedescribedbythedifferential
equationshownasEquation2.3.
where isthesuperficialmolarvelocity,
area of the catalyst,
(2.3)
isthemasstransfercoefficient,
is the bulk concentration of substrate ,
istheexternalsurface
is the concentration of at the
catalytic surface, and refers position along the length of the reactor. Equations 2.4 and 2.5 can be
used to solve Equation 2.3 for a packed bed reactor where
23
is the Sherwood number,
is the
diffusion coefficient,
is the particle diameter, is the volumetric flow rate, is the reactor cross
sectionalareaand isthereactorvoidfraction.
(2.4)
(2.5)
The Sherwood number is a function of the Reynolds number and the Schmidt number, but the exact
relation between the dimensionless numbers is typically determined experimentally. Considering a
seriesof16packedandfluidizedbedstudies,DwivediandUpadhyay(1977)developedanequationfor
theSherwoodnumber(
)tofitawiderangeofdataasshowninEquation2.6forReynoldsnumbers
) and the Schmidt number (
less than 10 where the Reynolds number of the particle (
) are
described by Equations 2.7 and 2.8 respectively, is the liquid dynamic viscosity, and is the liquid
density(DwivediandUpadhyay,1977;Seguinetal.,1996).
1.1
ReP. Sc
Sh
(2.6)
(2.7)
(2.8)
The diffusion coefficient (
) used in these equations can be estimated by the Wilke and Chang
equationasshowninEquation2.9where istheassociationfactorforthesolvent(1.0forunassociated
solvents),
isthemolalvolumeforthesolute(
),and
0.5
isthemolarmassforcomponent
(Treybal,1980).
117.3
18
10
0.6
(2.9)
The void fraction of the packed bed can be estimated for uniformly sized spheres and small reactor
diametersbyEquation2.10where
isthereactordiameter(Rase,1990).
0.4 1
0.42
24
(2.10)
Finally, the differential term in Equation 2.3 (
) can be determined from the kinetics of the
reaction.CombiningEquations2.3through2.10withthereactionkineticsofEquation2.2willprovidea
completemodelofboththekineticsandmasstransferinapackedbedreactorforenzymaticbiodiesel
production.
Santacearia et al. (2007) employ a similar set of equations for an immobilized enzyme packed bed
reactor for free fatty acid esterification. The reaction kinetics obtained from batch experiments was
combinedwithexternalmasstransferequationstomodeltheoleicacidconversioninthepackedbed
reactor (Santacearia et al., 2007; Tesser et al., 2005). In the system described, a unique immobilized
catalyst was developed that incorporated springs to help prevent problems arising from catalyst
swelling. Since the authors did not have a uniform spherical catalyst, an alternative Sherwood
correlationwasdevelopedfortheircatalyst.
Considering an immobilized enzyme packed bed reactor, Rahman et al. (2011) used an ordered bibi
reactionmechanismandexternalmasstransferequationstodescribeesterificationtoproducefarnesyl
laurateusingLipozymeRMIM.Comparingareactionlimitedmodeltoamasstransferlimitedmodel,
thereactorwasfoundtobemasstransferlimited(Rahmanetal.,2011).Halimetal.(2009)performed
asimilaranalysiscomparingreactionandmasstransferlimitedmodelsforaNovozym435packedbed
reactorforbiodieselproductionandalsofoundthatthereactionwasmasstransferlimited.
ConsideringsolelytheexternalmasstransferinaLipozymeTLIMpackedbedreactorforthehydrolysis
ofpalmolein,Chewetal.(2008)usedafirstorderapparentreactionratealongwiththeexternalmass
transfer correlations. The authors presented only the initial reaction rate data since the apparent
reaction rate assumption was only valid for the initial conditions and determined that significant
externalmasstransfereffectsexistedthatcouldnotbeneglectedespeciallyasthereactorisscaledup
(Chewetal.,2008).
25
Chapter 3
ExperimentalBackground
The following sections present preliminary results that demonstrate the feasibility of the enzyme
immobilizationprotocolalongwiththemethodologydevelopment.Lipasewassuccessfullyentrapped
in solgels with good activity and enhanced thermal stability. A commercially available immobilized
lipase,Novozym435,wasusedtoproducemethyloleateandsomeofthechallengesassociatedwith
itsapplicationwereidentified.
3.1 Lipaseentrapmentinasolgelmatrix
Lipase was entrapped in an 80% propyltrimethoxysilane (PTMS) and 20% tetramethoxysilane (TMOS)
solgelfollowingaprocedureadaptedfromthatofCliffordandLegge(2005).Theprecursors(PTMSand
TMOS)weresonicatedinthepresenceofwater(H2O:Si=12)andhydrochloricacid(HCl:Si=0.0002)for
approximatelytwohours.Thehydrolysisofthismixtureformedanaqueoussolwhichwaskeptonice
topreventuntimelygelation.Asolutionoflipase(NovozymesN44035)and50mMphosphatebuffer
at pH 7 (enzyme:buffer = 1:2 by volume) was subsequently added to the precursor solution. The
enzymeandprecursorsolutiongelledslowlywhilecoveredat4Cfor24hours.Upongelation,thesol
gelwasallowedtodryfurtheruntilthedryingratewaslessthan1mg/hr.Finally,thegelwascrushed
intoafinepowderandwashedtwicewith50mMphosphatebufferatpH7,thenwithisopropylalcohol,
andfinallywithhexane.Thesolgelpowderwassizedusingamicrosieve(number200mesh)andonly
particles less than 80 m in diameter were used for the analysis. Approximately 12 g of solgel were
producedusingthisprocedure.
Todeterminethedegreeoflipaseimmobilizationinthesolgel,asampleoftheenzymebuffersolution
loaded onto the solgel and the first phosphate buffer wash were analyzed for total protein content
usingaWatersHPLCsystemwithanAgilentZorbaxBioSeriesGF250column.TheHPLCwascalibrated
usingBSAstandardsasperaBCAproteinassaykitfromPierceBiotechnology.Themobilephasewas1
mL/min of 200 mM phosphate buffer with pH 7, the sample injection volume was 20 L, and the UV
detectionwavelengthwas280nm.Basedonthedifferencebetweentheamountofproteinloadedinto
26
thegelsandtheamountofproteininthewash,thedegreeofimmobilizationwasdetermined.Itwas
determinedthat44.3%ofthelipaseloadedontothesolgelwasentrapped(1.89mglipase/gsolgel).
Thesolgelswereanalyzedforactivityusingthehydrolysisoflinoleicacidethylestertolinoleicacidas
the basis for activity determination. In this reaction, 5 mL of 500 mM linoleic acid ethyl ester in
disopropyletherwascombinedwith50mgofsolgeland50Lof50mMphosphatebuffer(pH7).The
reaction was conducted at 30C with agitation at 500 rpm. Each hour, a 10 L sample was removed
fromthereactionvialanddilutedin990Lhexanecontaining50Lof5mMpentadecanoicacidasan
internalstandardand5LofthederivitizingagentBSTFA.Priortoanalysisofthereactionproducts,the
solution washeld at 60C for one hour to allow the TMS derivitization to occur. The formation of the
reactionproduct,linoleicacid,wasanalyzedusingaVarianCP3800gaschromatographequippedwitha
VarianFactorFourVF5msfusedsilicacapillarycolumnwithdimensions0.25mmx0.25mx30mandan
oventemperatureof220C.
Theinitialactivityofthesolgelisdefinedasthemolesoflinoleicacid(LA)producedperunitoftimeper
mass of lipase. The initial activity of the solgel was 12.00 mol LA/(min mglipase) with a relative
standarddeviationof3.68%.Inaddition,thepresenceofnonenzymaticactivitywasdetermined.Two
systems were considered: one with blank solgel (solgel without lipase) and one without solgel. In
bothsystemsnomeasurablelinoleicacidwasproducedwhichconfirmedthatthereactionobservedin
thelipaseentrappedsolgelreactionmediumwasinherentlyenzymatic.
Finally, the stability of the solgel activity was considered with respect to linoleic acid under storage
conditionsof4C.AsshowninFigure3.1,thesolgelentrappedlipasemaintainedastablelinoleicacid
activity up to 146 days of storage at 4C. Noureddini et al. (2003) studied solgel immobilized lipase
stability and activity and found that after 120 h of incubation at 40C solgels retained 95% of their
activity,butfreelipaselost67%activityafter24hofincubationat40Candhadalmostnoactivityafter
96hofincubation.
27
14
Activity(mol/min mgLipase)
12
10
0
91
104
StorageTime(days)
146
Figure3.1:Linoleicacidactivityofthelipasesolgelbasedon91,104and146daysofstorageat4C.Theerrorbars
representthestandarderrorwithn=26.
3.2 Thermalpropertiesofsolgelentrappedlipase
Immobilization of lipase in solgels was shown to improve the thermostability of the enzyme
(Noureddini et al., 2002; Noureddini et al., 2003). To analyze the thermostability of the solgel
entrappedlipase,thesolgelswereincubatedinaPierceReactiThermHeatingModuleat70Cforthe
desiredperiodoftime.Sampleswereperiodicallytakenandassayedforactivityat30Cinthepresence
oflinoleicacidethylesterandbufferaspreviouslydescribed.Adeclineinactivitywasobservedwith
increasingincubationtimefrom100%attimezerotoapproximately50%afterthreedaysofincubation
at 70C (Figure 3.2). Noureddini et al. (2002) found solgel entrapped lipase to be stable up to 70C,
whilefreelipasehadsevereactivitylossat40C.
To investigate the activity of the solgel at increased temperatures, the linoleic acid assay was
performed at 30C and at 70C with neither of the solgel samples incubated before the assay. From
Figure3.3,thereactionat70Chasahigherinitialreactionratethanthatat30C(28.7mMLA/hvs.16.0
mMLA/h),butthe12hlinoleicconcentrationforbothtemperatureswasthesame(approximately100
mM).
28
y=0.0063x+0.9198
R=0.8874
RelativeActivity
100%
75%
50%
25%
0%
0
10
20
30
40
Incubation Time(h)
50
60
70
Figure3.2:Relativeactivityofthesolgelentrappedlipaseafterincubationat70C.Theerrorbarsrepresentthestandard
errorwithn=26.
100
[LA]mM
80
60
y=28.7x
40
y=16.0x
20
70C
30C
0
0
6
Time(h)
10
12
Figure3.3:Linoleicacidconcentrationprofileasafunctionoftimeforsolgelimmobilizedlipaseat70Cand30Cfor12h.
Theslopesofthelinesrepresenttheinitialreactionrates.
29
3.3 MethyloleatesynthesiswithNovozym435
To determine the feasibility of producing methyl oleate, a biodiesel component, using solgels, a
commerciallyimmobilizedlipase(Novozym435,NovozymesNorthAmericaInc.,Washington,DC)was
investigated.Novozym435isalipasefromCandidaantarcticaimmobilizedonamacroporousacrylic
resin (SigmaAldrich Canada Ltd., Oakville, ON). As a case study, this commercially available lipase is
usefulsincethereisabroadspectrumofliteratureavailableontheuseofNovozym435forbiodiesel
production(Changetal.,2005;ChenandWu,2003;Duetal.,2004;HernndezMartnandOtero,2008;
Kseetal.,2002;Laietal.,2005;Oraireetal.,2006;Salisetal.2005;Samukawaetal.,2000;Shimada
etal.,1999;Watanabeetal.,2000;Watanabeetal.,2001;Xuetal.,2003).
An assay for methyl oleate production was employed using 5.94 g of triolein (60% purity), 50 mg of
Novozym435,and162Lofmethanol(1:1methanol:trioleinunlessnotedotherwise).Sampleswere
combinedandagitatedinavialmaintainedat40C.Thecourseofthetransesterificationreactionwas
followedwithtime:10Laliquotsperiodicallytakenanddilutedwith990Lofhexane.Thereactions
werefollowedfor6hunlessotherwisenoted.
Since alcohols are known to inhibit enzymes, it is expected that as the concentration of methanol is
increasedbeyondacertaintolerancetherewillbeanegativeimpactontheobservedactivityandlevel
ofconversion(KargiandShuler,2001).Therequiredstoichiometricratioofmethanoltotrioleinforthis
transesterificationreactionis3:1.AsshowninFigure3.4,theoptimalmethanolmolarratiorequiredto
maximize the concentration of methyl oleate produced after 6 h was 0.5 to 1.5. Ratios in this range
likely minimize alcoholmediated inhibition of the enzyme while providing adequate substrate for the
reaction.Asthisisabatchreaction,theamountofmethyloleateproducedislimitedbytheamountof
alcoholthatcanbepresentinthesystemwithoutdeactivatingtheenzyme.
Enzymaticreactionsareverytemperaturesensitivewherelowtemperatureslimitthekineticenergyof
the system and high temperatures cause enzyme denaturation, both of which result in lower rates of
reaction(KargiandShuler,2001).Sincethereactionmediumfortheformationofmethyloleateisvery
viscous, the potential to decrease the viscosity by increasing the temperature or by introducing non
viscoussolventsisofgreatinterest.Therefore,theeffectoftemperatureontheconversionoftriolein
andmethanoltomethyloleatewasinvestigatedforbothsolventbasedandsolventfreesystems.
30
Normalized6hMethylOleateConcentration
1.0
0.8
0.6
0.4
0.2
0.0
0
6
Methanol MolarRatio
10
12
Figure3.4:Normalizedfinalmethyloleateconcentrationatvariousmethanolmolarratiosfor6hat40C(normalizedtothe
Novozym435batchmaximum).Theerrorbarsrepresentthestandarddeviationwithn=25.
100
Solventfree
Solventbased
FinalConversion(%)
80
60
40
20
0
25
35
45
Temperature(C)
55
65
Figure3.5:Percentconversionofmethanoltomethyloleateasafunctionofreactiontemperatureafter6hforbothsolvent
freeandsolventbased(50%hexane)reactionmediawithamethanoltotrioleinmolarratioof1:1.
31
AsseeninFigure3.5,anoptimumtemperatureforproductformationwasobservedat60C,bothinthe
absenceandpresenceoftheorganicsolvent(50%hexane).Inthesolventbasedsystemtheconversion
was more than double that of the solventfree system. This can likely be attributed to the improved
mixing in the solventbased system as the hexane reduces the viscosity thereby allowing for an
improvementinthemixing.
Figure3.6comparesthemethyloleateformationat40Cwitha1:1molarratioofmethanoltotriolein
with and without solvent. The solventfree reaction medium consisted of triolein, methanol and
immobilizedlipase,whilethesolventbasedmediumconsistedoftriolein,methanol,immobilizedlipase,
and50%hexanebyvolume.Thesolventbasedmediumandthesolventfreemediuminitiallyshowed
the same extent of conversion which increased linearly for the first 12 h. However, as the reaction
progressed beyond 12 h, the solventbased system continued to increase in methyl oleate production
whereas the solventfree system plateaued. As discussed previously, this may be attributed to the
mixinglimitationsinthesolventfreemedium.After24h,thesolventbasedsystemachievedapercent
conversion that was approximately three times that of the solventfree system (79% conversion for
solventbasedand26%conversionforsolventfree).
In a similar study except that the methanol to triolein molar ratio was increased to 1.5:1 (Figure 3.7),
similarobservationsweremade.Atthebeginningofthereactionperiodtheconversionofmethanolto
methyl oleate was comparable between the solventbased and solventfree systems, but the solvent
freesystemreachedaplateauafterabout23h,whilethesolventbasedsystemcontinuedtoshowan
increaseinconversionupto30hwhentheconversionleveledoff.Unlikeatlowermethanoltotriolein
ratios,theinitialconversionofthesolventfreesystemwasslightlyhigherthanthatofthesolventbased
system. Also, the point at which the solventfree system leveled off was much later (about 23 h as
comparedto12hinthepreviousstudy).Whentheratioofmethanolisincreaseditcanbeassumed
that more reactant is available to proceed further without deactivating the lipase despite the poor
mixing.Thereactioninthesolventfreemediumwasstilllimitedbypoormixingsincethesolventbased
system surpassed the solventfree system conversion before leveling off (50% conversion for solvent
and30%conversionforsolventfree).
32
100
1:1Methanol Solventfree
1:1Methanol Solventbased
Conversion(%)
80
60
40
20
0
0
10
15
20
Time(h)
Figure3.6:Percentconversionofmethanoltomethyloleateatamethanoltotrioleinmolarratioof1:1insolventfreeand
solventbased(50%hexane)reactionmediaat40C.Errorbarsrepresentthestandarderrorwithn=34.
100
1.5:1Methanol Solventfree
1.5:1Methanol Solventbased
Conversion(%)
80
60
40
20
0
0
10
15
20
25
Time(h)
30
35
40
45
50
Figure3.7:Percentconversionofmethanoltomethyloleateatamethanoltotrioleinmolarratioof1.5:1insolventfreeand
solventbased(50%hexane)reactionmediaat40C.
33
Lipase undergoes interfacial activation at oilwater interfaces and the lipase activity is typically
proportionaltotheinterfacialarea,butexcessivewatercancausethecompetinghydrolysisreactionto
proceed as opposed to the transesterification reaction (Nouredinni et al., 2005). Therefore, to
determinewhetherasmallamountofwaterwouldbebeneficialforthesereactionconditions,asetof
experiments was performed with and without 3% water added to the solventfree and solventbased
(50%hexane)reactionsatamethanoltotrioleinmolarratioof1:1at40C.Forboththesolventfree
andsolventbasedsystems,addingthissmallamountofwaterreducedtheachievableconversionduring
thesixhourtimeframebyabout75%(Figure3.8).
3.4 Solgelmethyloleateformation
To demonstrate that the solgel immobilized lipase had methyl oleate activity, an experiment was
conducted with solgels of different compositions (Figure 3.9). The 6 h methanol to methyl oleate
conversionobservedforthe80%PTMSsolgelwascomparabletothatofNovozym435,butthe20%
PTMS solgel was considerably lower. Although comparable results are observed between Novozym
435andthe80%PTMSsolgel,itwillbenotedthattoachievethislevelofconversion350mgofsolgel
(approximately0.66mglipase)wereusedincomparisonto50mgofNovozym435.
The suppliers reported activities for the each of the enzyme preparations (Table 3.1) show that
Novozym 435 has half the activity compared to Novozyme N44035 which is the lipase used in the
preparationofthesolgel.Basedonthemeasuredmethyloleateactivities,thelipaseimmobilizedinthe
solgelhadapproximately6and70timesmoreactivitythanNovozym435forthe20%and80%PTMS
solgels, respectively. The results also indicate that the more hydrophobic solgel (80% PTMS) was
significantly more active in relation to the more hydrophilic preparation. This is expected as the
interfacialactivationoflipaseathydrophobicinterfacesiswelldocumented(KrebsandGerstein,2000;
SardaandDesnuelle,1958).Theseresultsindicatethatthe80%PTMSsolgelisagoodcandidateforthe
immobilization of solgels on silica supports for use in a packed bed enzymatic reactor to produce
methyloleate.
34
20
1:1Methanol Solventfree NoWater

1:1Methanol Solventfree 3%water
1:1Methanol Solventbased NoWater
1:1Methanol Solventbased 3%water
Conversion(%)
15
10
0
0
3
Time(h)
Figure3.8:Percentconversionofmethanoltomethanololeateatamethanoltotrioleinmolarratioof1:1forsolventfree
andsolventbased(50%hexane)systemswithandwithouttheadditionof3%water.
20
Conversion(%)
15
10
0
Novozym435
80%PTMSSolgel
20%PTMSSolgel
Figure3.9:MethanoltomethyloleateconversionusingimmobilizedlipasefromNovozym435,80%PTMSsolgeland20%
PTMSsolgelwithamolarratiooftrioleintomethanolof1.5:1at40C.Errorbarsrepresentthestandarderrorwithn=2.
35
Table3.1:ReportedandmeasuredactivitiesforNovozym435,80%PTMSsolgeland20%PTMSsolgel.
Enzymepreparation
Reportedactivity
Novozym435(N435)
10000PLU/gN435*
80%PTMSsolgel
20000PLU/glipase*
20%PTMSsolgel
20000PLU/glipase*
*
PLU=1molpropyllaurateformedperminute
U=1molmethyloleateformedperminute
36
Measuredactivity
53.46U/gN435
3707U/glipase
331U/glipase
Chapter 4
Study of Support Materials for Solgel Immobilized

Lipase*
Overview
A variety of support materials for solgel immobilized lipase were considered based on their ability to
providesuperiorsolgeladhesion,loadprotein,andsynthesizemethyloleate.Astandardapproachwas
developed to formulate the supported lipase solgels and to allow comparison of the resulting hybrid
materials.Thesesupportedsolgelsareproposedasanalternativeimmobilizationregimetoovercome
somechallengesassociatedwithenzymaticbiodieselproductionsuchasenzymestabilityandcost.The
supportmaterialsconsideredwere612meshsilicagel,CeliteR633,CeliteR632,CeliteR647,anion
exchange resin, and Quartzel felt. Each support material exhibited unique properties that would be
beneficial for this application including: Quartzel felt had the highest initial solgel capacity; silica gel
had the most uniform coating of deposited solgel; the anion exchange resin had the highest protein
loading (1060 g lipase/g) and reaction rate (1.25 mmol/min g); the Celite support series were the
mostthermallystableandhadthelowestwatercontent;CeliteR632hadthehighestsolgeladhesion
(275 mg solgel/g material), 6 h biodiesel conversion per gram of supported material (68%), and
enzymaticactivity(9.4mmol/minglipase).Thethreesupportswiththehighestenzymaticproperties
(conversion,activityandreactionrate)wereCeliteR632,anionexchangeresin,andQuartzel.These
supports had superior performance in comparison to the unsupported lipase solgels. Based on this
study the lipase solgel support material with the most potential for biodiesel production is Celite
R632.
Keywords
Enzymeimmobilization;solgelentrapment;lipidtransesterification;lipase;supportedsolgel;biodiesel
ThischapterwassubmittedforpublicationinBiocatalysisandBiotransformationandiscurrentlyunderreview
StudyofSupportMaterialsforSolgelImmobilizedLipase
S.M. Meunier and R.L. Legge, Biocatal. Biotransform., submitted July 2012, Manuscript ID GBAB12
404.R1.
37
4.1 Introduction
Growing environmental concerns have resurfaced interest in alternative fuels that are sustainable,
renewableandbiodegradable.Biodieselisonesuchenvironmentallyfriendlyfuelthatcanbeproduced
from natural oils and fats, can be used on its own or blended with petroleumbased diesel fuels, and
emitslowerexhaustemissionsandgreenhousegases(Tanetal.,2010).Biodieselisamixtureoffatty
acid alkyl esters that can be produced via transesterification of triglycerides in the presence of an
alcoholwithaglycerolbyproduct.
Industrially,biodieselisproducedusingalkalinecatalystswhichhaveveryhighconversionsandreaction
rates, but are burdened with high energy requirements and extensive downstream recovery and
purification steps (Meher et al., 2006). Enzymebased transesterification can overcome some of the
challenges associated with alkaline transesterification requiring milder operating conditions, minimal
downstreamprocessingandproduceaveryhighpurityproductwithlowerwasteproduction(Fukudaet
al.,2001;Akohetal.,2007;Marchettietal.,2007;VasudevanandBriggs,2008).
There are limitations that accompany enzymatic biodiesel production including the potential for
inactivation of the enzyme and the high cost of enzymes (RoblesMedina et al., 2009). Immobilizing
enzymes can help overcome these limitations as immobilization renders them more chemically and
thermally stable, and they are more easily recovered and reused (Macario et al., 2009; Bajaj et al.,
2010). Lipases immobilized in solgels also benefit from enhanced lipase activity (Aucoin et al., 2004;
Reetz,1997).Inarecentreviewofbiodieselproductionviaimmobilizedenzymes,Zhangetal.(2012)
concluded that immobilized enzymes have the potential in the near future to be competitive with
conventionalchemicalcatalysisfortheproductionofbiodiesel.
Sinceenzymeimmobilizationinsolgelscancauseinternalmasstransferlimitations(Bajajetal.,2010),a
doubleimmobilizationprocedureisproposedinwhichlipaseisimmobilizedinasolgelandsupported
onaninert material.Thebenefitsofthissystemincludeboth theadvantagesofsolgelsandenzyme
immobilization in addition to improving internal mass transfer and the practicality of supported
materials.Researchershaveimmobilizedsolgels onsupport materialssuchascottonfabric(Li et al.,
2007), silica quartz fiber felt (Nassreddine et al., 2008; Oraire et al., 2006), diatomaceous earth
(Furukawaetal.,2002a;Furukawaetal.,2002b;FurukawaandKawakami,1998;KawakamiandYoshida,
1996; Kawakami et al., 2003; Koszelewski et al., 2010; Meunier and Legge, 2010; Meunier and Legge,
2012),andsilicagel(Pogorilyietal.,2007).
38
In this study, four materials were chosen for development as supports that are analogous in
compositiontosolgel.Silicagelhasbeenusedsuccessfullyforsupportingureasesolgels(Pogorilyiet
al., 2007). Diatomaceous earth is a common enzyme immobilization material; Celite 545 has been
usedsuccessfullyfortransaminase,proteaseandlipasesolgels(Furukawaetal.,2002a;Furukawaet
al., 2002b; Furukawa and Kawakami, 1998; Kawakami and Yoshida, 1996; Kawakami et al., 2003;
Koszelewskietal.,2010),andCeliteR632,R633,andR647supportedlipasesolgelshavebeenusedfor
methyloleateproduction(MeunierandLegge,2010;MeunierandLegge,2012).Quartz fiberfeltisa
usefulsupportmaterialforlipasesolgels(Nassreddineetal.,2008;Oraireetal.,2006).Finally,anion
exchangeresinwaschosenforitsvaluableseparationandimmobilizationproperties(Wangetal.,2010).
This study encompasses a variety of support materials, applies a consistent solgel immobilization
technique and compares each of the materials formed on the basis of their surface, coating, and
enzymaticcharacteristics.ThesurfacecharacteristicsofthematerialsarecomparedviaSEMimagingto
determinethepresenceofavisiblesolgelcoating.Thecoatingpropertiesconsideredaretheadhesion
of solgel tothe support surface, the protein loading of the supported solgel, and the water capacity
and thermal stability of the coated and uncoated support materials. The enzymatic properties
consideredaremethyloleateproductioncapacity,enzymaticactivity,andreactionrate.Basedonthese
results,thesupportmaterialswiththegreatestpotentialareidentifiedforfurtherstudy.
4.2 Experimental
4.2.1
Materials
Novozym435wasagiftfromNovozymesNorthAmericaInc.(Franklinton,NC)itsreportedactivityis
10 000 PLU/g. Lipase (NS44035, Novozymes North America Inc., Franklinton, NC) with a reported
activity of 20 000 PLU/g was immobilized in 80% PTMS trimethoxypropylsilane and 20% TMOS
tetramethyl orthosilicate (SigmaAldrich Canada Ltd, Oakville, ON) solgels. The lipase solgels were
supportedondesiccantactivatedsilicagel612mesh(EagleChemicalCo.,Mobile,AL),Celite(World
Minerals, Santa Barbara, CA), analytical grade anion exchange resin AG3X4 (BioRad Laboratories,
Richmond,CA),andQuartzelfelt(SaintGobainQuartzUSA,Louisville,KY).Ultrapurewaterwasfroma
MilliQwaterpurificationsystem(Millipore,Billerica,MA).Allotherchemicalsusedwerereagentgrade
fromlocalsuppliers.
39
4.2.2
4.2.2.1
Methods
Supportlipasesolgelimmobilization
ThesupportmaterialschosenforconsiderationandtheirknownpropertiesareprovidedinTable4.1.
The lipase was immobilized using a method adapted from previous work (Meunier and Legge, 2010)
using80%PTMSand20%TMOSwhichresultedinasolgelwithhighhydrophobicityforenhancedlipase
activity(CliffordandLegge,2005)andsuperiorgelationproperties.Aprecursormixturecontaining0.08
mol PTMS, 0.02 mol TMOS, 1 mol ultrapure water and 20 L HCl (0.1 M) was hydrolyzed for 1 h in a
sonicating water bath at room temperature. The hydrolyzed precursor mixture was then evaporated
undervacuumfor30minat40C.Subsequently,14mLoflipaseandphosphatebuffer(50mM,pH7.0)
was added to the precursor mixture. This mixture was then added to the desired support material
based on a predetermined solgel capacity to maximize the amount of solgel on the support without
allowingexcesssolgeltogelseparatelyfromthesupportedmaterial(Table4.1).Afterthoroughmixing,
the support and solgel was deposited in a Petri dish, sealed and aged for 24 h at 4C, then dried
unsealedat4Cuntilthedryingrateslowedtolessthan1mg/h.Oncedry,thegelswerewashedwith
phosphatebuffer(5mL/gsolgel)twice,isopropylalcohol(3mL/gsolgel),andhexane(5mL/gsolgel).
Afterwashing,theexcesssolventwasevaporatedfromthegelsovernightandthegelswerethenstored
inasealedcontainerat4Cuntiluse.
Unsupported lipase solgels were prepared in a similar manner omitting the mixing step with the
support.SincetheunsupportedsolgelsadheretothePetridishupongelation,theywerecrushedinto
apowderuponremovalandthewashingsolutionsremovedviacentrifugationatapproximately1250xg
for10min.
Table4.1Propertiesofthelipasesolgelsupportmaterials.Porediameterandparticlesizeareasprovidedbythesuppliers.
Initialsolgelcapacityisdefinedastheamountofliquidsolgelrequiredtofullywetthesupportwithoutexcessmaterial.
Support
SilicaGel
CeliteR633
CeliteR632
CeliteR647
AnionExchangeResin
QuartzelFelt
Porediameter
(m)
0.005
6.5
7.0
0.007
large
Particlesize
(m)
16803360
297595
5951410
5951410
74149
40
Initialsolgelcapacity
(mL/g)
0.3
1
1
1
1.5
62.5
4.2.2.2
ThesurfacecoatingofthesupportedlipasesolgelswasstudiedqualitativelyusingaHitachiS570SEM
(HitachiHighTechnologies,Berkshire,England).Thesampleswerecoatedwithgoldinpreparationfor
microscopy and an electron beam energy of 15 kV was used. Images were taken at a variety of
magnificationsforcomparison,butonlythosetakenat3000xarepresented.
4.2.2.3
Proteinmeasurement
The methodology for protein determination of the supported solgels has been previously described
(MeunierandLegge,2010).Samplesfromtheoriginalenzymesolutionand thetwophosphate buffer
washeswereanalyzedforproteincontentusingaVarianHPLC(AgilentTechnologies,Mississauga,ON)
equipped with an Agilent Zorbax Bio Series GF250 column calibrated using a BSA protein standard
(PierceBiotechnologyInc.,Rockford,IL).TheHPLCmobilephasewas200mMphosphatebuffer(pH7.0)
andtheabsorbancewavelengthwas280nm.Basedonaproteinmassbalance,theamountofprotein
remaininginthesupportedsolgelswasdetermined.
4.2.2.4
Adhesionofsolgel
Theadhesionofsolgeltothesupport(mgsolgelpergsupportmaterial)wasdeterminedbasedonthe
changeinmassofthesupportbeforeandafterthesolgelimmobilizationandwashingprocedures.The
ratio of lipase to solgel (mg lipase per g solgel) was also measured and calculated from the ratio of
proteinloadingtosolgeladhesion.
4.2.2.5
Thermogravimetricanalysis
The water retention and thermal stability of the supported solgels and the neat supports was
determined basedonthermogravimetricanalysisusingaQ500TGA (TAInstruments,NewCastle,DE).
The TGA analysis method used a temperature ramp from 30C to 400C at 10C/min under N2 at 50
mL/minandaircoolingfor10min.
4.2.2.6
Enzymatictransesterificationoflipids
GCMS analysis was used to determine the enzymatic activity of the supported lipase solgels as
previouslypublished(MeunierandLegge,2010).Areactionvialcontaining1gofsupportedlipasesol
gel, 4 mmol triolein, and 4 mmol methanol was heated to 40C and agitated for 6 h. Hourly, 10 L
aliquots were sampled from the reaction vial and diluted with 990 L hexane and 100 L of
heptadecanoic acid methyl ester (internal standard) and subsequently analyzed for methyl oleate
content using a Varian GCMS (CP3800 gas chromatograph, Saturn 2000 mass spectrometer/mass
41
spectrometer)equippedwithaCPWax52CBfusedsilicacolumn(CP8513,VarianInc.,Mississauga,ON).
TheGCinjectortemperaturewas250Cwithasplitratioof50andaheliumgasflowof1mL/min.The
column oven temperature was held at 170C for 10 min, ramped at 10C/min to 250C, and held at
250Cfor2min.Themethyloleatedetectionlimitwasapproximately0.7mM.
4.3 Resultsanddiscussion
4.3.1
Surfacecoating
In most cases, the SEM images collected confirm a visible presence of solgel on the surface of the
supportmaterialthatisconsistentwiththenatureoftheunsupportedsolgel(Figure4.1a).However,it
isevidentthateachofthesupportmaterialsiscoatedwithsolgelinadifferentmanner.Thesilicagel
exhibits an ideal solgel surface coverage without solgel the silica gel has a rough surface (Figure
4.1b),butwhencoatedwithsolgelthesurfaceissmoothandappearsuniform(Figure4.1c).Incontrast
withthesilicagelsurfacecoverage,allthreediatomaceousearthsupportsexhibitcoatingthatisnon
uniform (Figure 4.1d and e for Celite R633) there are visible sections of solgel coating on the
diatomaceousearth.Thesolgeltendstoformpreferentiallyonitselfratherthanonthediatomaceous
earthindicatingstrongercohesiveforcesthanadhesiveforces.Thisisconsistentwithobservationsof
Celitesolgelcoatingspreviouslyreported(MeunierandLegge,2010).Theanionexchangeresindoes
notshowanyevidenceofsolgelcoatingonthesurfaceat3000xmagnification(Figure4.1fandg).Since
the coating and enzymatic properties considered show evidence that solgel is indeed immobilized on
the anion exchange resin, the lack of visible solgel in the SEM images is likely due to the similarities
between the surface properties of the anion exchange resin and the solgel. The SEM images of the
Quartzel felt show solgel immobilized on the individual fibers (Figure 4.1h and i) and also bonding
individualfiberstoeachother(notshown).Unlikethesolgelformedontheothersurfaces,thesolgel
formedontheQuartzelfibersisrougherthan the uncoatedfibers.Thiscouldbeduetomechanical
issueshinderingthesolgelformationonthecylindricalsurface.
42

Figure4.1:SEMimagesat3000x(a)unsupportedlipasesolgel;silicagel(b)withoutand(c)withsolgelimmobilizedonthe
surface;CeliteR633(d)withoutand(e)withsolgelimmobilizedonthesurface;anionexchangeresin(f)withoutand(g)
withsolgelimmobilizedonthesurface;andQuartzelfelt(h)withoutand(i)withsolgelimmobilizedonthesurface.
43
4.3.2
Coatingproperties
Theadhesionofsolgelonthesupportmaterialandtheloadingofproteininthesupportedsolgelwere
compared to determine the presence of preferential solgel or lipase loading for each of the support
materialstested(Figure4.2).Silicagel(SG),CeliteR632(R632),andanionexchangeresin(AE)allhave
similarratiosoflipasetosolgel(3.7,3.1,and4.3glipase/mgsolgelrespectively).CeliteR633(R633)
andCeliteR647(R647)bothhaveanaffinitytoloadlipaseoversolgel(9.1and14.6glipase/mgsol
gel respectively). A possible explanation for the difference between the different Celite supports is
that the R632 has a much lower water adsorption capacity than the other two supports which could
causeR632topreferentiallyloadsolgelwithoutentrappingthelipaseinitspores.Quartzelfelt(QF)
hasastrongaffinityforsolgeloverthelipase(0.86glipase/mgsolgel)possiblyduetothecylindrical
natural of the solgel formation which does not provide a welldefined pore structure to effectively
entrap the protein. Based on these results, the highest protein loadings and adhesion levels are
attainedbyAEandR632(proteinloadingsabove800glipase/gmaterialandadhesionlevelsabove200
mgsolgel/gmaterial).Coatingonglassbeadswasalsoexplored,butthesolgeldidnotadheretothe
surfaceofthebeadswithoutsurfacemodification.
TGAwasconductedtoascertainthewatercapacityandthermalstabilityofthematerialsstudied.The
weightlossoftheneatsupportsandthesupportedsolgels(Figure4.3)showthataddingsolgeltothe
supportmaterialcausesnochangeoranincreaseinthewatercontentofthematerialsgiventhatthe
weight change is primarily due to the presence of water. One notable exception to this is AE, which
exhibits less of a weight loss with solgel, but a substantially larger weight loss than any of the other
supportmaterials.Basedonthethermograms(notshown),itisevidentthatAEisnotthermallystable
attemperaturesaboveapproximately250C,thusathightemperatures,theweightchangeisattributed
todegradationoftheanionexchangeresin.Therefore,theadditionofthesolgelincreasestheoverall
stabilityofthemateriallikelyduetothepresenceofthehighlystablesilicon.NeatQFhasaverylow
densityandthesampleweightisveryclosetothedetectionlimitsoftheTGAsothethermogramsfor
the neat QF are highly variable. Overall, the Celite series of support materials displays the highest
thermalstabilityandlowestwatercontent.
44
Adhesion(mgsolgel/gmaterial)
1000
ProteinLoading(glipase/gmaterial)
800
600
400
200
0
SG
R633
R632
R647
AE
QF
Figure4.2:Adhesionofsolgel(mgsolgel/gmaterial)andproteinloading(glipase/gmaterial)foreachsupportmaterial.
Supports:SG(silicagel),R633(CeliteR633),R632(CeliteR632),R647(CeliteR647),AE(anionexchangeresin),andQF
(Quartzelfelt).DuetothelowdensityofQF,dataarepresentedonapercm2basisratherthanapergrambasis.
100
FinalWeight(%)
90
80
70
60
50
US
SG
R633
R632
NeatSupport
Solgel
R647
AE
QF
Figure4.3:FinalweightpercentofthesupportswithandwithoutsolgelbasedonTGA.Supports:SG(silicagel),R633
(CeliteR633),R632(CeliteR632),R647(CeliteR647),AE(anionexchangeresin),andQF(Quartzelfelt).Errorbars
representa90%confidenceinterval.
45
4.3.3
Enzymaticproperties
Basedonthemethyloleateconcentrationprofilesforthebatchreactionsasafunctionoftime(Figure
4.4),theconversionofmethanoltomethyloleate(%)after6hr,theinitialreactionrate(mM/ming),
andtheinitialenzymaticactivity(mmol/minglipase)werecalculated(Table4.2).The6hconversionis
calculatedbasedonthefinalmethyloleateconcentrationwithrespecttothestartingconcentrationof
methanolforthereaction;theinitialreactionrateisbasedoninitialslopefortheconcentrationprofile;
and the initial enzymatic activity is calculated from the initial reaction rate based on the amount of
lipasepresentinthereaction(Figure4.4).
Foreachofthesupportedlipasesolgels,withtheexceptionofSGandQF,thepercent conversionof
methanol to methyl oleate was comparable to or beyond that of the unsupported solgel (Table 4.2).
Althoughthesilicagelappearedtobeapromisingsupportmaterialbasedonthecoatingproperties,it
did not produce any detectable methyl oleate. This is likely due to protein inactivation caused by an
interactionbetweenthesilicagelandthelipaseduringtheimmobilizationprocess.Althoughtheseries
of Celite supports all have similar solgel coatings (Figure 4.1), Celite R632 exhibits a substantially
higherpercentconversionthanCeliteR633andCeliteR647,indicatingthattheenzymaticproperties
ofthesolgelcoatingweremoredependentontheabilityofthesupportmaterialtoachievegoodsol
gelandproteinloadingsthanconsistentsupportcoverage.Althoughtheanionexchangeresindidnot
haveasolgelcoatingthatwasvisiblebySEM,itshighpercentconversionconfirmsthepresenceofvery
activelipase.QFexhibitedaverysmallpercentconversion,butduetoitslowdensityitisreportedona
percm2basisratherthanapergbasisliketheothersupportmaterials,sothevaluesarenotdirectly
comparable. R632 and AE achieved enhanced percent conversion in comparison to that of US, with
R632 having the highest conversion of 68%. Therefore, the Celite R632 is considered the most
promisingsupporttoreducemasstransferlimitationsthatinhibitreactantsandproductsfromdiffusing
toandfromtheentrappedlipase.
46
ConcentrationofMethylOleate(mM)
400
300
6hMethylOleate
Concentration
InitialReactionRate
200
y=74.939x
R=0.9993
100
0
0
3
Time(h)
Figure4.4:KineticprofileforatypicallipidtransesterificationreactionwithCeliteR632asthesupport.
Table4.2:Enzymaticpropertiesofthesupportedlipasesolgels:percentconversionofmethanoltomethyloleate(MO)after
6hofreaction,initialreactionrate,andenzymaticactivity.ForQuartzelfelt,theconversionandreactionrateare
presentedintermsofcm2ratherthangduetothelowdensityofthefeltmaterial.Allerrorrangesarebasedonthe90%
confidenceintervalofreplicatedexperiments.
Support
Unsupported
Silicagel
CeliteR633
CeliteR632
CeliteR647
Anionexchangeresin
Quartzelfelt
6hMOconversion
(%/gmaterial)
36.012.1
0
31.95.3
68.13.4
38.12.6
50.81.6
8.84(percm2)
Reactionrate
(mM/mingmaterial)
0.6170.239
0
0.5290.253
1.2020.212
0.6160.142
1.2490.011
0.181(percm2)
Enzymaticactivity
(mmol/minglipase)
1.010.39
0
4.822.31
9.361.64
5.611.29
7.800.69
8.29
Similar conclusions can be drawn from the initial reaction rate and enzymatic activity (Table 4.2).
AlthoughthereismuchmorevariabilityintheR632datafortheseparametersmakingthereactionrate
ofR632notsignificantlyhigherthanofAE,andtheactivitynotsignificantlyhigherthanthatofAEand
QF.Intermsofreactionrate,twodistinctgroupswereobserved:ahighreactionrategroupofR632and
AE(approximately1.2mmol/ming)andalowreactionrategroupofUS,R633,andR647(approximately
47
0.6mmol/ming).QFdoesnotfollowthistrendasitispresentedonapercm2basisratherthanperg
duetoitslowdensityincomparisontotheothersupports.WiththeexceptionofSG,allthesupports
demonstratedenhancedlipaseactivityincomparisontotheunsupportedsolgelformulationduetothe
improved mass transfer expected for the supported solgel system. However, QF does have a high
enzymaticactivitysinceactivityiscalculatedonthebasisoftheamountoflipaseinthereactionmedia
andQFhasaverylowlipaseloadingincomparisontotheothersupportmaterials(Figure4.2).Basedon
the enzymatic properties presented (conversion, reaction rate, and activity) R632, AE and QF are the
mostpromisingsupportmaterials.
Forcomparison,theactivityofthecommerciallyimmobilizedlipase(Novozym435)andthefreelipase
(NS44035)werealsoassayedunderthesamereactionconditions.TheenzymaticactivityofNovozym
435(8.81.1mmol/ming)wasnotsignificantlydifferentthanthatofR632,AE,andQF,andNS44035
didnotproducemethyloleateatdetectablelevels.
4.4 Conclusions
Astudyofpotentiallipasesolgelsupportmaterials(silicagel,CeliteR633,CeliteR632,CeliteR647,
anionexchangeresin,andQuartzelfelt)wasconductedwithacomparisonoftheirsolgelcoatingand
enzymatic capacities to that of unsupported lipase solgels demonstrating their potential for biodiesel
production. Based on SEM imaging, the silica gel possessed a smooth and uniform coating on the
surface; however, it did not exhibit any enzymatic activity. Thermal analysis demonstrated that the
anionexchangeresinwasnotstableathightemperaturesliketheothersupportmaterials,andthatthe
Celite series of supports exhibited the highest stability and lowest water content. Based on solgel
adhesion,proteinloadingandenzymaticproperties,thesuperiorsupportmaterialswereCeliteR632,
anionexchange,andQuartzelfelt.Thesesupportsshowedimprovedenzymaticabilitiesincomparison
to the unsupported solgels likely due to the decrease in mass transfer resistance. Based on this
analysis, Celite R632 is the most promising support material, and supported lipase solgels warrant
furtherstudyforenzymaticbiodieselproduction.
48
4.5 Acknowledgements
ThisworkwassupportedbytheNaturalSciencesandEngineeringResearchCouncil(NSERC)intheform
of a Discovery grant to RLL and an NSERC Postgraduate Scholarship to SMM. We thank Novozymes
NorthAmericaforsamplesofthelipaseformulation,WorldMineralsforsamplesoftheCelitesupport
materials,andSaintGobainQuartzfortheQuartzelsamples.
49
Chapter 5
EvaluationofDiatomaceousEarthasaSupportfor
SolgelImmobilizedLipaseforTransesterification
Overview
Enzymatic production of biodiesel by triglyceride transesterification is a promising alternative to
chemicallycatalyzedbiodieselproductiondespitethechallengesinvolvedwithusingenzymes.Celite
supportedlipasesolgelswereinvestigatedasanoptionforsolvingsomeofthechallengesassociated
with the use of enzymes for biodiesel production addressing such problems as activity, stability and
reusability of the enzyme. Three types of Celite were considered (R633, R632, and R647) and
compared tounsupportedlipasesolgels.Variousfactorswereconsideredwithregardstocomparing
thesupportmaterials.Theyincludedsurfacemorphologycharacterizedusingsurfaceareaanalysisand
scanning electron microscopy, physical properties including adhesion of the solgel to the Celite and
the protein loading on the Celite, and finally enzymatic properties based on the conversion of
methanoltomethyloleateandtheenzymaticactivityoflipase.Allthesolgelsshowedgoodconversion
andinitiallipaseactivity,andalltheCelitesupportshadsimilarsolgeladhesionandproteinloading.
Solgel immobilized lipase supported on Celite R632 had an average 6 h percent conversion of
approximately60%,andanaverageinitiallipaseactivitycomparabletothatoftheunsupportedsolgel
formulation.
Keywords
Enzymeimmobilization,solgelentrapment,Celite,transesterification,lipase
ThischapterispublishedintheJournalofMolecularCatalysisB:Enzymatic
EvaluationofDiatomaceousEarthasaSupportforSolgelImmobilizedLipaseforTransesterification
S.M.MeunierandR.L.Legge,J.Mol.Catal.B:Enzym.,vol.62,no.1,pp.5357,2010.
50
5.1 Introduction
Biodieselisanalternativetofossilfuelsthatisbecomingincreasinglyimportantduetoenvironmental
concerns, rising fossil fuel prices, demands for renewable fuel sources, and market requirements for
agricultural surpluses (RoblesMedina et al., 2009). Industrially, most biodiesel is produced using
alkalinecatalystsbecausethisprocessiscosteffectiveandefficient;however,itisenergyintensiveand
requires significant downstream processing steps such as washing, separation and purification
(Ranganathanetal.,2008).Alternatively,transesterificationoftriglyceridestoproducebiodieselcanbe
accomplished enzymatically using lipase. The enzymatic process consumes less energy, produces less
wasteandbyproducts,andinvolvesmilderoperatingconditions(Akohetal.,2008;Fukudaetal.,2001;
Marchettietal.,2007;VasudevanandBriggs,2008).
Industrially,enzymaticbiodieselproductionpossessesseveralchallengesincludingaslowreactionrate,
the risk of enzyme inactivation by the alcohol substrate and glycerol byproduct, and the high cost of
enzymes (RoblesMedina et al., 2009). By immobilizing enzymes, their reusability can be improved
thereby reducing their associated costs, and simplifying downstream processing (Akoh et al., 2008).
Further, lipase immobilized in solgel matrices has improved chemical and thermal stability and has
higheractivitycomparedtootherimmobilizationprocedures(Reetz,1997;Reetzetal.,1996).
Solgelimmobilizedlipasecanbesupportedoninertmaterialstoimprovethediusionofsubstratesand
productstoandfromtheenzymeandthusimprovethereactionrate(Brnyiketal.,2000;Oraireetal.,
2006; Pogorilyi et al, 2007). Studies have shown that lipase supported on Celite 545 had greater
activity, higher thermal stability, and improved reusability in comparison to lipase supported on
Amberlite IRA938 and free lipase (Sairolu 2008; Sairolu et al., 2004; Sairolu and Telefoncu,
2004).StudieshavealsoshownthatsolgelentrappedlipaseimmobilizedonCelite545hasimproved
activityandthermostabilitycomparedtolipasedepositeddirectlyonCelite545withoutusingthesol
gelmethod(Kawakami,1996;KawakamiandYoshida,1996).
Deactivation of lipase by methanol is a common challenge for the production of biodiesel via
methanolysis. Although the stoichiometric ratio of oil to methanol is 1:3, studies have shown that
enzyme deactivation can be significantly reduced by adding methanol in 3 steps of 1 mole ratio each
(Shimadaetal.,1999;Watanabeetal.,2000;Watanabeetal.,2001;Watanabeetal.,2002;Xuetal.,
2004).OurownlaboratoryworkusingNovozym435supportsthisfindingwithanoptimalmethanolto
triolein mole ratio of 0.5 1.5. As an alternative to stepwise methanolysis, one study shows that
51
continuousadditionofmethanolwasfoundtoachievethehighestbiodieselconversionof97%(Blafi
Baketal.,2002).
The presence of glycerol is also inhibitory to the enzymatic production of biodiesel (Samukawa et al.,
2000). Studies have shown that silica beds are successful for the adsorption of glycerol in biodiesel
streams,butmethanolisdetrimentaltothisprocess(Yorietal.,2007;Mazzierietal.,2008).Methanol
desorbsglycerolfromthesilica(Yorietal.,2007),andaddingmethanolreducesthesaturationcapacity
of the silica gel by half (Mazzieri et al., 2008). In addition, when silica gel was used to control the
amount of water in the biodiesel system, excess silica reduced the biodiesel yield due to methanol
adsorption(Wangetal.,2006).
Theobjectiveofthisstudywastobettercharacterizelipasesolgelssupportedondiatomaceousearth
based supports and to evaluate their efficacy in terms of the production of biodiesel via enzymatic
transesterification.ThesurfacemorphologyoftheCelitesupportedlipasesolgelswascharacterized
using surface area analysis and scanning electron microscopy. The physical properties investigated
includedtheadhesionofsolgeltotheCeliteandtheloadingofproteinontheCelitesolgel.Finally,
the conversion and enzymatic activity of the hybrid materials were considered. For each property
studied, three different types of Celite were compared and, where applicable, the Celite supported
solgel was compared to unsupported solgel material. This information is valuable in evaluating an
optimal support system for solgel immobilized lipase and for understanding the interaction between
thesupport,solgel,andlipase.
5.2 Experimental
5.2.1
Materials
CelitesampleswereagiftfromWorldMinerals(SantaBarbara,CA).Lipase(NS44035)andNovozym
435weregiftsfromNovozymesNorthAmericaInc.(Franklinton,NC).ThebiologicalsourceofNS44035
wasnotprovidedbythesupplier;thecommercialactivityofNS44035is20000PLU/g.Thebiological
source of Novozym 435 is C. antarctica and its commercial activity is 10 000 PLU/g. Tetramethyl
orthosilicate (TMOS), trimethoxypropylsilane (PTMS), triolein, and methyl heptadecanoate (HDAME)
were obtained from SigmaAldrich Canada Ltd. (Oakville, ON). Sodium phosphate was obtained from
MallinckrodtBaker(Phillipsburg,NJ).IsopropylalcoholwasobtainedfromEMDChemicals(Gibbstown,
NJ).HexanewasobtainedfromFisherScientificCompany(Ottawa,ON).Thesilicagel(612mesh)was
52
obtained from Eagle Chemical Co., Mobile, AL. Ultrapure water was produced using a MilliQ water
purification system from Millipore (Billerica, MA). All other chemicals were obtained from local
suppliersandwereofreagentgrade.
5.2.2
5.2.2.1
Methods
Immobilizationoflipase
ToimmobilizelipaseonCelite,0.08molPTMSand0.02molTMOSwerehydrolyzedinthepresenceof
1molultrapurewaterand200LHCl(0.1M).Themixturewassonicatedinawaterbathsonicatorfor1
h. The precursor solution was then rotary evaporated in a heated water bath at 40C for 30 min to
remove water and alcohol. A solution of lipase and phosphate buffer (50 mM, pH 7.0) with an
approximate protein concentration of 4000 g/mL was prepared, and 14 mL was added to the
hydrolyzedprecursorsolution.NinemLoftheresultantmixturewascombinedwith6gofthedesired
Celitesupport,thoroughlymixed,anddepositedinaPetridish.ThePetridishwassealedandagedat
4Cfor24h.Thegelwasthendriedunsealedat4Cuntilthedryingratewaslessthan1mg/h.Once
dry,thegelwasremovedfromthePetridishandwashedtoremoveanyproteinnotentrappedwiththe
followingwashsequence:twicewithphosphatebuffer(50mM,pH7.0,5mLper3mLsolgelmixture),
isopropylalcohol(2.5mLper3mLsolgelmixture),andhexane(5mLper3mLsolgelmixture).Excess
solventwasevaporatedfromthegel atroomtemperatureovernightbeforethegelswerestoredina
sealedcontainerat4C.
Unsupportedgelswerepreparedinasimilarmannerexceptthattheevaporatedprecursorandenzyme
mixturewasdepositeddirectlyintoaPetridishforaginganddrying,thedriedsolgelwascrushedina
mortarfollowingremovalfromthePetridish,andthewashingsolutionswereseparatedfromthesolgel
bycentrifugationat4000rpmfor10min.
5.2.2.2
Surfaceareaanalysis
The nitrogen physisorption experiments were carried out in a Micrometrics Gemini 2375 surface area
analyzer(MicrometricsInstrumentCorporation,Norcross,GA).Thesamplesweredegassedat120Cfor
15hundernitrogenflowpriortoeachmeasurement.TheCeliteR647samplesweredegassedforan
additional3hat300C.StarDriversoftwarewasusedtofittheadsorptioncurveandevaluatetheBET
surfacearea.
53
5.2.2.3
The surface morphology of the gels was evaluated using a Hitachi S570 scanning electron microscope
(Hitachi HighTechnologies, Berkshire, England). The samples were coated with gold prior to analysis.
Anelectronbeamenergyof15kVwasusedforanalysis.Solgelclusterswereidentifiedvisuallyfrom
theSEMimages,andthe SEMimages wereused todeterminethepercentcoverageofsolgelonthe
surface of the Celite particles. For this analysis, 45 random images of each type of Celite were
collectedandthepercentcoveragewasdeterminedforeachimage.
5.2.2.4
Proteindetermination
Thetotalproteincontentofthegelswasdeterminedfromtheproteincontentoftheenzymesolution
loadedintothesolgelsandtheproteincontentinthetwobufferwashes.Theamountofproteinwas
quantifiedusingaVarian HPLCsystem(VarianInc.,Mississauga,ON)equippedwithanAgilentZorbax
Bio Series GF250 column (Agilent Technologies, Mississauga, ON) and calibrated using a BCA protein
assaykit(PierceBiotechnologyInc.,Rockford,IL).ThemobilephasefortheHPLCanalysiswas200mM
phosphatebuffer(pH7.0),anddetectionwasatanabsorbancewavelengthof280nm.
5.2.2.5
Enzymaticlipidtransesterification
TheenzymaticactivityoftheCelitesupportedlipasesolgelswasdeterminedbyGCMSanalysis.The
reactions were carried out at 40C with agitation for 6 h. The reaction vial consisted initially of
approximately1gofthesupportedlipasesolgel,4mmoloftriolein,and4mmolofmethanol.At1h
intervalsa10Lsamplewasremovedfromthereactionvialanddilutedin990Lhexaneand100Lof
the internal standard, HDAME. The formation of methyl oleate, the reaction product, was followed
using a Varian GCMS system (CP3800 gas chromatograph, Saturn 2000 mass spectrometer/mass
spectrometer)equippedwithaCPWax52CBfusedsilicacolumn(CP8513,VarianInc.,Mississauga,ON).
OneLsamplesofthedilutedreactionmixturewereinjectedintotheGCataninjectortemperatureof
250Candasplitratioof50.Heliumwasusedasthecarriergaswithacolumnflowof1mL/min.The
GCoventemperaturewasinitiallysetto170Cfor10min,rampedat10C/minto250C,andheldat
250Cfor2min.
54
5.3.1
Surfacemorphology
ComparisonofthetexturalcharacteristicsofthethreetypesofCeliteconsideredforthisstudy,R633,
R632, and R647 will be found in Table 5.1. The three types of Celite were chosen based on their
particlesizesandporediameters.Surfaceareaanalysiswascompletedfor eachofthesupportswith
andwithoutalipasesolgelcoating.ThecoatedR633hadasignificantincreaseinsurfaceareaoverthe
uncoatedR633,buttheothertwosupports,R632andR647,hadnosignificantchangeinsurfacearea
upon coating (Table 5.1). With R633, some solgel formation occurred between particles promoting
particle agglomeration which could increase the surface area. In addition, the presence of solgel
clustersontheCeliteconfirmedthatcohesiveforceswereperhapsstrongerthantheadhesiveforces
duringsolgelformation(Figure5.1).SolgelclusterswerevisibleforallthreetypesofCelite(Figure
5.1).
Figure5.1:SEMimagesat3000xmagnificationof(a)uncoatedCeliteR633,(b)uncoatedCeliteR632,(c)uncoatedCelite
R647,(d)CeliteR633coatedwithlipasesolgel,(e)CeliteR632coatedwithlipasesolgel,and(f)CeliteR647coatedwith
lipasesolgel.ArrowsidentifysomeclustersofsolgelonthesurfacesoftheCelite.
55
Table5.1:ParticlesizeandporediameterforeachtypeofCelitebasedonthemanufacturersspecifications,andsurface
areasofCelitewithoutlipasesolgel(support)andwithlipasesolgel(coated).Theconfidencelimitsindicatedrepresent
the95%confidenceintervalofthemeanbasedonn6.
Particlesize
(um)
300600
6001400
6001400
Celite
R633
R632
R647
Porediameter
(um)
6.5
7
0.07
Supportsurfacearea
(m2/g)
0.900.12
1.520.17
58.920.77
Coatedsurfacearea
(m2/g)
1.370.33
1.490.23
64.394.50
Table5.2:QuantificationofthesolgelclustersfoundoneachoftheCelitesamples.45imagesforeachtypeofCelite
wereusedforthisdetermination.
Surfaceareaimaged(m2)
Numberofsolgelclusters
Averageclustersize(m2)
95%confidenceintervalforclustersize
R633
477000
38
1270
540
R632
477000
48
1230
480
R647
477000
2
3670
6130
18%
16%
Surfaceareacoverage(%)
14%
12%
10%
8%
6%
4%
2%
0%
R633
R632
R647
Figure5.2:PercentsolgelcoverageonthesurfaceofCeliteasdeterminedbySEMimageanalysis.Theerrorbarsrepresent
the95%confidenceintervalsofthesamplemeanbasedonn=45.
56
As seen in Figure 5.2, R647 has significantly less percent solgel on the surface of the Celite in
comparison with R633 and R632. This indicates that the lipase solgel did not adhere as well to the
Celite R647 support. Celite R647 has a much larger surface area (Table 5.1) than R633 and R632
whichmayfavoursolgelcohesionoveradhesiontotheCelitesurface.
5.3.2
Physicalproperties
BoththeCeliteR633andR632haveacomparablenumberofsolgelclustersandaverageclustersize
basedonthe45imagesrecordedforeachsample;however,CeliteR647hadveryfewclusters(Table
5.2). There is no significant difference between the levels of solgel adhesion on any of the support
materials (Figure 5.3). For R647, although the same mass of solgel is immobilized on the support
(Figure5.3),thesurfacecoverageissignificantlylower(Figure5.2).ThiswouldsuggestthattheR633
andR632favouradhesionofsolgelasathinnerlayeronthesurfaceratherthancohesionofthickersol
gelclustersaswithR647.
Figure5.4showstheloadingofproteinascalculatedbasedonHPLCanalysis.Theunsupportedsolgel
has a high protein loading per gram of material (Figure 5.4 A) because there is no additional mass
contributedbytheCelitewhichisthecaseforsupportedsolgels.Whentheloadingisconsideredon
the basis of solgel only (Figure 5.4 B), the unsupported formulation is significantly lower than the
CeliteR632andR647formulations.Forboththematerialbasisandsolgelbasis(Figure5.4AandB),
R633 has significantly lower protein loading than R632 and R647. Based on the manufacturers
specifications for the Celite, R633 has a much higher water adsorption capacity than R632 and R647
(R633 240%, R632 84%, and R647 163%). R633 may preferentially absorb the lipasebuffer solution
from the solgel mixture, whereas the other types of Celite absorb less water which may promote
gelationofthelipaseinthesolgelpolymermatrix.Inthelattercasethelipasewouldnotbewashedoff
thesupportasintheformercaseresultinginmoreproteinimmobilizedontheR632andR647supports.
5.3.3
Enzymaticproperties
InitialstudieswereperformedtodeterminetheeffectofsilicaandCeliteontheenzymaticactivityof
the solgel. Adding 1 g of silica gel to either the 50 mg Novozym 435 or 200 mg solgel reaction
systems reduced the amount of methyl oleate produced after 6 h (Table 5.3). Further, adding 1 g of
CeliteR632tothe200mgsolgelsystemcausednomethyloleatetobeproducedin6h.Therefore,
anyimprovedactivityinthesupportedsolgelappearstobeduetotheCelitesolgelcomplexrather
thantheadsorptionofglycerolorwaterinthepresenceofsilica.
57
0.40
Adhesion(gsolgel/gmaterial)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
R633
R632
R647
Figure5.3:AdhesionofsolgelonthesupportforeachtypeofCeliteconsidered.Theerrorbarsrepresentthe95%
confidenceintervalsofthesamplemeanbasedonn=3.
5.0
A
4.5
B
Proteinloading (mgprotein/g)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
R633
R632
R647
Unsupported
Figure5.4:Loadingofproteinpergramforeachsupport:A)theamountofproteinpergramofmaterial,andB)theamount
ofproteinpergramofsolgel.Theerrorbarsrepresentthe95%confidenceintervalofthesamplemeanbasedonn=3.
58
Table5.3:EffectofthepresenceofsilicaandCeliteonmethyloleateproductionwithNovozym435andunsupportedsol
gelcontaininglipase.
Enzyme
50mgNovozym435
200mgunsupportedsolgel
200mgunsupportedsolgel
Silica
1gsilicagel
1gsilicagel
1gCeliteR632
Effectonmethyloleateproduction
Reduction
Reduction
Elimination
Figure5.5showsthepercentconversionofmethanoltomethyloleatepergramofmaterialbasedona
6hreactionperiod.R632,R647,andtheunsupportedsolgelsallshowahigherconversionthanR633.
Since R633 has a lower protein loading than the other gels, a lower conversion per mass is expected.
Figure5.6showsthatCeliteR633exhibitsahighinitiallipaseactivity;however,itslowproteinloading
andlowconversionmakeitapoorsupportmaterialfortransesterification.
Figure5.6showsthattheinitiallipaseactivityoftheCeliteR647issignificantlylowerthanthatofboth
R633andR632.ThiswouldindicatethattheR633andR632supportsprovideafavourableenvironment
fortheimmobilizationofactivelipaseincomparisontoR647.Thehighpercentsurfacecoverage(Figure
5.2)onR632whilemaintainingcomparablesolgeladhesion(Figure5.3)indicatesthatthinnerlayersof
solgelareformedonthesurfaceofthissupport,thereforethehighlipaseactivityfortheR632lipase
solgelmaybeduetoimprovedinternalmasstransfer.
5.4 Conclusions
Lipase immobilized solgels supported on diatomaceous earth are promising for biodiesel production.
All three Celite supports considered (R633, R632, and R647) as well as the unsupported solgel
exhibitedlipaseactivityandshowedconversionrangesfromapproximately25%to65%in6h.Similar
solgeladhesionlevelsweremeasuredforeachofthesupportmaterials.Basedonthethreesupport
materialsconsidered,CeliteR632exhibitedgoodinitialactivityandachievedanaverageconversionof
approximately60%pergramofmaterialafter6hofreaction.Thehighsurfaceareacoverageofthesol
gel on R632 indicates a thinner coating that would lead to improved internal mass transfer. Thus,
comparing the three supports, Celite R633, R632, and R647, and unsupported lipase solgel, Celite
R632couldbeconsideredapromisingsupportforlipaseimmobilizedsolgelstoproducebiodieselvia
enzymatictransesterification.
59
100
90
Conversionat6h(%/gmaterial)
80
70
60
50
40
30
20
10
0
R633
R632
R647
Unsupported
Figure5.5:Percentmethanolconversionpergramofmaterialafter6hforeachsolgelformulation.Theerrorbarsrepresent
the95%confidenceintervalofthesamplemeanbasedonn=9.
20
18
Initialactivity(mmol/min g)
16
14
12
10
8
6
4
2
0
R633
R632
R647
Unsupported
Figure5.6:Initiallipaseactivityforeachsolgelformulationmeasuredbytheinitialamountofmethyloleateproducedper
minutepergramofprotein.Theerrorbarsrepresentthe95%confidenceintervalofthesamplemeanbasedonn=9.
60
of a Discovery grant to RLL and an NSERC Postgraduate Scholarship to SMM. We thank Novozymes
North America for supplying samples of the lipase formulation and World Minerals for supplying
samplesoftheCelitesupportmaterials.
61
Chapter 6
EvaluationofDiatomaceousEarthSupported
LipaseSolgelsasaMediumforEnzymatic
TransesterificationtoProduceBiodiesel
Overview
Immobilized lipase has the potential to be the catalyst of choice for biodiesel production since it is
efficient, effective, and environmentally friendly; however, the stability and activity of lipase must be
addressed before enzymatic biodiesel production processes can be industrially accepted. This study
investigates an enzyme immobilization procedure that immobilizes lipase in a solgel supported on
diatomaceous earth (Celite R632), and determines its potential for biodiesel production in terms of
achievable conversion and apparent stability. Four immobilized materials (lipase solgels with and
without Celite at two protein loading levels) were compared in terms of their immobilized protein
content,conversionofmethanoltomethyloleate,lipaseactivity,longtermstability,andglycerolwater
adsorption.TheCeliteR632solgelwithhighproteinloadingachievedthemaximumconversioninthe
6hreactionperiod(80%).Adryingstepwasfoundtobeadvantageouspriortothereaction,andthe
absorption of glycerolwater on the Celite was only found to be significant at high levels of glycerol.
Thematerialwasfoundtobeverystableuponstorageat4Cforupto1.5years,losingonlyabout15%
of its percent conversion capacity per year. Based on this study, the supported immobilization
techniqueshowssignificantpotentialasanovelcatalystforbiodieselproduction.
Keywords
enzymeimmobilization,biodiesel,solgel,transesterification
ThischapterispublishedintheJournalofMolecularCatalysisB:Enzymatic
Evaluation of Diatomaceous Earth Supported Lipase Solgels as a Medium for Enzymatic
TransesterificationofBiodiesel
S.M.MeunierandR.L.Legge,J.Mol.Catal.B:Enzym.,vol.77,pp.9297,2012.
62
6.1 Introduction
Lipaseisatriacylglycerolhydrolaseandisresponsibleforthehydrolysisofesterbondsintriglycerides.
Athydrophobicinterfaces,lipasesareknowntoundergointerfacialactivationwhichcausesasurgein
enzymatic activity (Sarda and Desnuelle, 1958). Enzymatic reactions typically have many advantages
such as enhanced specificity and efficiency; however, decreased activity and stability are common
elementsofconcernwhenworkingwithenzymes.
Onewaytoextendtheoperationalstability,andthusdecreasetheeffectivecostoftheenzyme,isby
immobilization. Although biodiesel production with alkali catalysts is known to be the least costly
option, processes using immobilized lipase are significantly less costly than those using free lipase
(Jegannathanetal.,2011).Jegannathanetal.(2011)alsoshowthatifimmobilizedlipaseisreusedmore
than five times or the production cost of the lipase is reduced significantly, enzymatic biodiesel
productioncouldbecompetitivewiththealkalibiodieselprocess.
Enzyme entrapment in polymers such as solgels provides a stable enzyme support with very strong
bonds while not requiring complex chemical preparation (Reetz et al., 1996). Both the thermal and
chemical stability and activity of enzymes can be improved via solgel immobilization (Pirozzi et al.,
2009; Reetz, 1997; Reetz et al., 1996). Further, transaminases encapsulated via a supported
immobilization procedure with Celite solgels exhibit enhanced activity even at extreme conditions
suchashighpHandtemperature,andcanberecycleduptoeighttimeswithoutdramaticallyreducing
the achievable conversion (Koszelewski et al., 2010). Other studies comparing Celite solgels show
similarresultsusinglipaseastheenzymeofinterest(Kawakami,1996;KawakamiandYoshida,1996).
Due to environmental and economical concerns, biodiesel has become an invaluable alternative to
standarddieselfuels(RoblesMedinaetal.,2009).Biodieseliscomprisedoffattyacidalkylestersandis
producedviathetransesterificationofoils.Conventionally,biodieselisproducedusingeitheranalkali
or acid catalyst, but using lipase as the catalyst for transesterification could provide significant
advantagesoverthesetraditionalprocesses.Enzymaticprocessesselfadapttochangesinrawmaterial
quality,producebiodieselinfewprocessingsteps,uselittleenergy,producelittlewastewater,andyield
high quality byproducts (Fjerbaek et al., 2009). The challenges associated with enzymatic
transesterification include low reaction rates, high enzyme cost, and the potential for enzyme
deactivation(Ganesanetal.,2009).
63
Since methanol, glycerol and water are all known to have inhibitive effects on lipase each of these
componentsmustbecloselymonitoredinthebiodieselproductionprocess.Methanolisasubstrateas
wellasalipaseinhibitorinthebiodieselproductionprocess;therefore,lowerthanstoichiometricratios
ofmethanolarecommonlyusedforbiodieselproductiontoavoidenzymedeactivation(Shimadaetal.,
1999;Watanabeetal.,2000;Watanabeetal.,2001;Watanabeetal.,2002;Xuetal.,2004).Glycerolis
a byproduct of enzymatic biodiesel production that is inhibitory to the lipase, but glycerol produced
during transesterification can be absorbed by silica beds (Samukawa et al., 2000; Yori et al., 2007;
Mazzierietal.,2008).Intermsofwaterabsorption,accordingtheCelitesupplier,WorldMinerals,the
water absorption capacity of Celite is up to 500% by weight, depending on the type of Celite
considered. This is a crucial parameter since, although water is necessary for lipase activation in
biodieselproduction,excesswaterlevelswillinhibitthelipasebyoccupyingtheenzymesupportpore
spaceandlimitingcontactbetweentheenzymeandsubstrates(RoblesMedinaetal.,2009).
This study considers the enzymatic activity of lipase immobilized in Celite R632 solgels based on a
supportedimmobilizationscheme.Acomparisonismadebetweentheachievableenzymaticconversion
ofmethanoltomethyloleateforunsupportedsolgelsandsolgelssupportedonCeliteR632atboth
high and low protein content levels. Further, the glycerolwater absorption is considered for both
Celite R632 and solgel Celite R632 using thermogravimetric analysis in an attempt to elucidate the
possibleeffectsofglycerolandwaterinabiodieselprocessforlipaseimmobilizedinCeliteR632sol
gels. This information is valuable in evaluating the potential of Celite R632 solgels as a supported
immobilizationmediumforenzymaticbiodieselproduction.
6.2 Experimental
6.2.1
Materials
CelitesampleswereagiftfromWorldMinerals(SantaBarbara,CA).Lipase(NS44035)wasagiftfrom
NovozymesNorthAmericaInc.(Franklinton,NC).ThebiologicalsourceofNS44035cannotbedisclosed
by the supplier; the activity of NS44035 is 20 000 PLU/g. Tetramethyl orthosilicate (TMOS),
trimethoxypropylsilane(PTMS),triolein,glycerol,methyloleateandmethylheptadecanoate(HDAME)
were obtained from SigmaAldrich Canada Ltd. (Oakville, ON). Acetonitrile was obtained from EMD
Chemicals(Gibbstown,NJ).SodiumphosphatewasobtainedfromMallinckrodtBaker(Phillipsburg,NJ).
Hexane and hydrochloric acid were obtained from Fisher Scientific Company (Ottawa, ON). Ultrapure
64
waterwasproducedusingaMilliQwaterpurificationsystemfromMillipore(Billerica,MA).Allother
chemicalswereobtainedfromlocalsuppliers.
6.2.2
6.2.2.1
Methods
FourdifferentlipasesolgelswereproducedforanalysiswithandwithoutCeliteR632asasupport
materialandwithhighandlowconcentrationsoflipase(Table6.1).
To immobilize lipase in diatomaceous earth solgels, 0.08 mol PTMS and 0.02 mol TMOS were
hydrolyzedinthepresenceof1molultrapurewaterand200LHCl(0.1M).Themixturewassonicated
for one hour to allow for complete hydrolization of the precursors. The precursor solution was then
rotary evaporated in a heated water bath at 40C for 30 min to remove excess water and alcohol. A
solutionoflipaseandphosphatebuffer(50mM,pH7.0)withanapproximateproteinconcentrationof4
mg/mLwasprepared,and14mLwasaddedtothehydrolyzedprecursorsolution.Forthehighlipase
concentration Celite solgels, the lipase solution was used undiluted at a concentration of
approximately12mg/mL.Theresultantmixturewasaddedtothesupportmaterialwithapproximately
3 mL solgel mixture for every 2 g of Celite R632. After thorough mixing, the solgel Celite was
deposited in a Petri dish, sealed and aged at 4C for 24 hours. The Celite solgel was then dried
uncovered at 4C until the drying rate was less than 1 mg/h. Finally, the supported gel was removed
fromthePetridishandwashedtwicewithphosphatebuffer(50mM,pH7.0,5mLbufferpergramof
solgel for each wash) to remove any protein that was not completely immobilized within the gel.
Excesssolventwasevaporatedfromthegelatroomtemperatureovernightpriortostoringthegelsina
sealedcontainerat4C.
Unsupportedgelswerepreparedinasimilarmannerexceptthattheevaporatedprecursorandenzyme
mixturewasdepositeddirectlyintothePetridishforaginganddrying,thedriedsolgelwascrushedina
mortaruponremovalfromthePetridish,andthewashingsolutionswereseparatedfromthesolgelby
centrifugation.
65
Table6.1:Descriptionofthelipasepreparationsusedforanalysisintermsofthesupportmaterial,solgelformulation,and
lipasesolutionconcentration.
Lipasepreparation
CSG4
CSG12
USG4
USG12
Free
6.2.2.2
Supportmaterial
CeliteR632
CeliteR632
Unsupported
Unsupported
Unsupported
Solgel
80%PTMS/20%TMOS
80%PTMS/20%TMOS
80%PTMS/20%TMOS
80%PTMS/20%TMOS
NoSolgel
Lipasesolution
4mg/mL
12mg/mL
4mg/mL
12mg/mL
4mg/mL
Proteinmeasurement
Thetotalproteincontentofthegelswascalculatedusingamassbalanceoftheproteincontentofthe
enzyme solution loaded to the solgels and content in the two buffer wash solutions. The degree of
immobilization was calculated as the percentage of protein in the solgel compared to the amount of
protein desired in the supported solgel. The amount of protein was quantified using a Varian HPLC
system(VarianInc.,Mississauga,ON)equippedwithanAgilentZorbaxBioSeriesGF250column(Agilent
Technologies,Mississauga,ON)andcalibratedusingaBCAproteinassaykit(PierceBiotechnologyInc.,
Rockford, IL). The mobile phase for the HPLC analysis was 200 mM phosphate buffer (pH 7.0) and
detectionwasatawavelengthof280nm.
6.2.2.3
Enzymaticlipidtransesterification
The enzymatic activity of the supported lipase solgels was determined by GCMS analysis. The
reactions were carried out at 40C with agitation for 6 h. The reaction mixture consisted initially of
approximately 1 g of the supported lipase solgel, 4 mmol of triolein and 4 mmol of methanol (total
reactionvolume3.89mL).Eachhour,a10Lsamplewasremovedfromthereactionvialanddilutedin
990 L hexane with 100 L of the internal standard, HDAME. The formation of methyl oleate, the
reactionproduct,wasfollowedusingaVarianGCMSsystem(CP3800gaschromatograph,Saturn2000
mass spectrometer/mass spectrometer) equipped with a CPWax 52 CB fused silica column (CP8513,
VarianInc.,Mississauga,ON).1LsamplesofthedilutedreactionmixturewereinjectedintotheGCat
an injector temperature of 250C and a split ratio of 50. Helium was used as the carrier gas with a
columnflowof1mL/min.The GCoventemperaturewasinitiallysetto170Cfor10min,rampedat
10C/minto250C,andheldat250Cfor2min.
The enzymatic activity of the dried solgel formulations was carried out in the same manner as the
originalgelswiththeexceptionthatpriortotheenzymaticassaythegelsweredriedovernightina60C
oven. At this point, no further change in mass was observed from the drying process and thus any
remainingwaterwasassumedtobecompletelyremovedfromthesolgelformulation.
66
UsingtheGCMSmethoddescribed,acalibrationwascompletedbasedonamethyloleatestandard(40
mM800mM)andusedasthebasisforallmethyloleateconcentrationsprovided.FromtheGCMS
chromatograms,methyloleatewastheonlyvisiblepeakindicatingthatnosidereactionsoccurred.The
enzymaticactivityofthelipasewasdeterminedfromtheslopeofthemethyloleateconcentrationtime
profileovertheweightofimmobilizedenzymematerialused.
6.2.2.4
Desorptionofglycerolandwater
The desorption of glycerol and water from Celite R632 upon equilibration was measured using a
thermogravimetricanalyzer(Q500TGA,TAInstruments,NewCastle,DE).Priortoanalysis,eachsample
wasimmersedinthedesiredsolutioncontainingglycerolandwateratvariousconcentrations(0%,10%,
25%, 50%, 75% and 100% by volume glycerol) and equilibrated for 24 hours. The sample was then
washedwithwatertoremoveexcessglycerolandairdriedovernight.TheTGAanalysismethodramped
thetemperaturefrom30Cto400Cat10C/minfollowedbyaircoolingfor10minunder50mL/minN2
gas.Theonsetofdesorption(temperatureatwhich0.025%masslossisachieved),thepeakdesorption
rate (maximum mass loss rate achieved), the peak temperature (temperature at the peak desorption
rate),andthetotalmassloss(percentmasslostattheendoftheanalysis)weredeterminedfromthe
weightlosscurveandthederivativeweightlosscurve.
6.3.1
Physicalproperties
Comparing the protein content and the degree of immobilization of the four different lipase solgels
formed,severalphenomenawereobserved(Figure6.1).First,forboththelowandhighproteinlevel
solgels,theunsupportedsolgelshavehigherproteincontentsincomparisontotheCelitesupported
solgels.SincethereisadditionalmaterialaddedtotheCelitesolgels,lowerproteincontentpergram
isexpectedforthesegels.
Theproteinlevelimmobilizedinthesolgelwasnotproportionaltotargetloadforthesolgelduringthe
formationprocedure(Figure6.1).FortheCelitesolgelstheproteincontentwasapproximatelydouble
when triple the lipase was loaded, but for the unsupported solgels the protein content was
approximatelytripleasexpected.OnepossibleexplanationforthelackofproportionalityfortheCelite
solgelsisthattheexcessoflipaseinthetripleunsupportedformulationcausedproteinaggregationand
subsequent immobilization. This effect might not be as pronounced for the lower levels of lipase
67
becausethereisnotenoughlipaseintheformulationandforthetriplelipaseCelitesystembecause
thepresenceofCelitereducestheapparentproteinconcentration.
The observed degree of protein immobilization for the Celite solgels is comparable to that for the
unsupportedsolgels(Figure6.1).Whentheamountoflipaseistripled,thedegreeofimmobilization
decreasesslightlyforboththeCeliteandunsupportedsolgels.Sincethereisexcesslipaseinthetriple
lipasesolgels,thereismoreopportunityforlipasethatisincompletelyentrappedtobeeasilyremoved
duringthewashingsteps.
Thehighproteincontentunsupportedsolgel(USG12)didexhibitmuchmorevariabilitythantheother
immobilizationregimes (Figure6.1). Thisislikelycausedby thepotentialforaggregation,incomplete
immobilizationandlipasedeactivationduringthesolgelformationprocedure.Withsuchhighlevelsof
protein and without the Celite support material, the lipase is much more exposed and thus more
sensitive to environmental changes. This enhanced potential for enzyme deactivation and poor
immobilization can conceivably greatly reduce the reproducibility of the lipase solgel immobilization
procedure.
25
100
20
80
70
15
60
50
10
40
30
DegreeofImmobilization (%)
ProteinContent(mgprotein/ggel)
90
20
10
0
CSG4
CSG12
ProteinContent
USG4
DegreeofImmobilization
USG12
Figure6.1:Proteincontentanddegreeofproteinimmobilizationforfoursolgelformulations.Errorbarsrepresenta95%
confidenceinterval.
68
6.3.2
Enzymaticproperties
Free lipase at a comparable level to that found in the immobilized lipase preparation CSG4 was
assayed to determine a base methyl oleate conversion level for comparison to the immobilized
preparations. Based on GCMS analysis, no measurable methyl oleate was detected during the 6 h
reaction.
Basedonthepercentconversionafter6hformethyloleateproduction(Figure6.2),theCeliteR632
solgelsexhibitedanincreasedconversionwhentheamountoflipaseloadedontothegelswastripled.
In addition, the Celite supported solgels achieved slightly more conversion when dried due to an
excessofwaterabsorbedonthesupportmaterialsinceCelite isabsorbentandlipaseisinhibitedby
excesswatercontent.
The unsupported solgels demonstrate the opposite trend the solgel preparation with triple lipase
achievedalowerpercentconversionthantheregularlipaseloadingandthedriedgelpreparationshave
lowerconversionthanthosethatarenotdried.Asdiscussedwithregardstotheproteincontentofthe
gels, this formulation does have much more variability than the other formulations which could be
caused by aggregation, incomplete immobilization, and deactivation of the protein due to the excess
quantitiesofprotein.SincethereisnoCelitepresentintheunsupportedsolgelstoremovethewater,
dryingtheenzymepreparationmayresultininactivationofthelipaseratherthanremovingtheexcess
waterandtherebycausingthereductioninproductconcentration.
In comparison to the Celite supported solgel, adding neat Celite to the unsupported solgel (SG +
Celite)greatlyreducesthepercentconversionlikelyduetoreactant,productandwaterabsorptionby
the Celite thereby preventing the reaction from progressing. Comparing the supported and
unsupported solgel formulations, the addition of the Celite as a support material for the solgel
increasesthepercentconversiontomethyloleateforbothsolgelformulations.
Comparingtheactivityofthedifferentsolgelpreparationsonamassbasis(Figure6.3)revealsthatthe
dried unsupported and supported solgels all have comparable activities. This demonstrates the
beneficial effects of using Celite and providing a support since the Celite solgels contain much less
enzymepergramofmaterialincomparisontotheunsupportedsolgels(Figure6.1).Theunsupported
solgel preparation with the higher level of lipase (USG12) does have a slightly lower activity in
comparison with the other materials when dried which is likely due to the excess lipase in the
preparationwhichiseitherinactiveand/orinaccessibletothesubstrates.
69
6hPerentConversion ofMethanol toMethylOleate
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
CSG4
CSG12
SG+Celite
Original
USG4
USG12
Dried
Figure6.2:Percentconversionofmethanoltomethyloleatebasedona6hbatchreactionforeachsolgelformulation
withoutandwithadryingsteppriortotheenzymaticassay.Errorbarsrepresenta95%confidenceinterval.
35
30
Activity(U/g)
25
20
15
10
0
CSG4
CSG12
USG4
Original
Dried
USG12
Figure6.3:Theenzymaticactivityonapergramofmaterialbasisforeachsolgelformulationasdeterminedfromthe
conversionofmethyloleateandtheproteincontentofthegels.Boththesolgelformationswithoutandwithadryingstep
priortothereactionwereconsidered.Errorsbarsrepresenta95%confidenceinterval.
70
TheunsupportedsolgelshaveamuchhigheractivitythantheCelitesolgelswhenassayedwithouta
dryingstep.TheadsorptivepropertiesoftheCeliteandtheinhibitionoflipasecausedbyexcesswater
aretheexpectedcausesofthisphenomenon.Despitetheincreasedlipasecontent(Figure6.1)inthe
highconcentrationlipaseunsupportedsolgel(USG12),theactivityiscomparabletotheregularlipase
levelunsupportedsolgel(USG4).Therefore,theexcesslipaseinthissystemmaynotbeaccessibleto
thesubstratesorhasbeenrenderedinactive.
TheCelitesupportedsolgelsandtheunsupportedsolgelswerealsocomparedbasedontheirstability
overaperiodofapproximately1.5years(Figure 6.4).TheunsupportedandCelitesupportedsolgel
formulations show almost identical trends with a very gradual decrease in product output over time.
The unsupported solgel (USG4) has a 0.05% decrease in product concentration per day while the
Celitesupportedsolgel(CSG4)hasa0.04%decreaseperday.Thisisaboutan18%(USG4)and15%
(CSG4) decrease in product concentration per year. This indicates that a very stable enzyme
formulationhasbeendevelopedwithalongshelflifewhenstoredat4C.
Normalized6hMethylOleateConcentration (%)
120%
y=0.0005x+1.0156
R=0.9885
100%
80%
y=0.0004x+0.9831
R=0.8561
60%
40%
20%
0%
0
100
200
300
400
Time(days)
CSG4
USG4
Linear(CSG4)
Linear(USG4)
500
Figure6.4:Averagepercentconversionofmethanoltomethyloleateforsolgelsafterstorageat4Cforunsupported()and
CeliteR632supportedsolgels(o).Thelinesrepresentthelinesofbestfitfortheunsupportedsolgel(broken)andthe
CeliteR632supportedsolgel(solid).Errorbarsrepresenta95%confidenceinterval.
71
6.3.3
Adsorptiveproperties
ATGAspectrum(Figure6.5)showstypicalweighttimeandderivativeweighttimeprofilesforasample
ofCeliteR632solgel(CSG4).Thepeakdesorptionandtotalmasslossparametersareindicatedon
the graph. These profiles are used to determine the onset of desorption, peak desorption rate, peak
temperature,andtotalmassloss.
Based on the weight change with respect to temperature data obtained from the TGA analysis, the
temperatureatwhichdesorptionwasestablishedforeachcase(10%,25%,50%,75%,and100%glycerol
forbothneatCeliteandtheCelitesolgelCSG4,datanotshown)wascompared.Theaverageonset
ofdesorptionforallcaseswas113.31.2C.Thisindicatesthattheobserveddesorptioniscausedbya
combinationofthewaterandtheglycerolratherthanthecomponentsseparately.
ThepeakdesorptionrateobtainedfromtheTGAanalysisisameasureofthelevelofadhesionofsolvent
totheCeliteorCelitesolgelwithrespecttothecohesionofsolventmoleculestoeachother.High
peakdesorptionratesindicatethatthesolventcoheresmorestronglytoitselfandlowpeakdesorption
ratessignifythatthesolventmoleculesadhereverywelltothesupportmaterial.Thetotalmassloss
indicatestheabsorptivecapacityoftheCeliteorCelitesolgelwithrespecttotheapplicableglycerol
water solution. At the end of the TGA analysis, any glycerolwater absorbed and retained by the
materialisdesorbedandthusquantifiedbythetotalmassloss.
Consideringboththepeakdesorptionrate(Figure6.6)andthe totalmassloss(Figure6.7)whenneat
Celite and Celite solgels are incubated in glycerolwater solutions similar trends are evident. Both
thepeakdesorptionratesandthetotalmasslossarehigherineachcasefortheneatCelitethanfor
theCelitesolgel.Itislikelythatthesolgelprovidesaprotectivebarrierthatreducesabsorptionof
theglycerolwatersolution.
Additionally, the amount of solution absorbed does generally increase with increasing percentage of
glycerolintheequilibratingsolution.OnenotableexceptionisthedecreasefortheneatCelitefrom
75%glycerolto100%glycerol.Thisshowsthatthewaterisnecessaryforabsorptionofglycerolonto
theCelite.However,thisisnotthecasefortheCelitesolgel.Theincreasingadsorptiontrendofthe
supportedsolgelmaterialisnotaffectedbyeliminatingthewaterfromtheequilibratingsolution.
72
100
0.4
PeakDesorption
98
0.35
96
Weight(%)
TotalMassLoss
92
0.25
90
0.2
88
0.15
86
DerivativeWeight(%/oC)
0.3
94
0.1
84
0.05
82
80
0
50
100
150
200
250
Temperature(oC)
Weight
300
350
400
DerivativeWeight
Figure6.5:TypicalTGAprofileincludingthesampleweight(solid)andthederivativeweight(broken).Thepeakdesorption
pointandthetotalmasslossareindicatedonthegraph.ThesampleshownisCeliteR632solgel(CSG4)at50%glycerol
equilibratingsolution.
0.8
PeakDesorption Rate(%/oC)
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
20
30
40
50
60
70
80
Glycerol/WaterEquilibrating Solution (v%glycerol)
CSG4
NeatCelite
90
100
Figure6.6:PeakdesorptionratesforCeliteR632solgels()andplainCeliteR632(o)basedondifferentlevelsofglycerol
intheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval.
73
30
TotalMass Loss(%)
25
20
15
10
0
0
10
20
30
40
50
60
70
80
CSG4
NeatCelite
90
100
Figure6.7:TotalpercentagemasslossforCeliteR632solgels()andplainCeliteR632(o)basedondifferentlevelsof
glycerolintheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval.
Whentheequilibratingsolutionconsistedsolelyofwater,therewasnoevidenceofdesorptionfromthe
neat Celite or solgel Celite (Figures 6.6 and 6.7). This is evidence that the water is not readily
retainedbytheCelitematerialdespiteitsabsorptiveproperties.
ThepeakdesorptiontemperatureasdeterminedbytheTGAisameasureofthestrengthofcohesionof
theglycerolwatersolutiontotheCeliteorCelitesolgelmaterial.Thepeakdesorptiontemperatures
foreachsolutionweredifferent(Figure6.8),rangingfrom132Cto187C,andshowedsimilartrendsto
those observed for the peak desorption rate (Figure 6.6) and the total mass loss (Figure 6.7). One
phenomena observed is the similarity between the values for the 100% glycerol solution absorption
when comparing the regular Celite and the solgel supported Celite (Figure 6.8), indicating that the
differencecausedbythesolgelismostprominentwhenwaterispresentintheequilibratingsolution.
AccordingtoCliffordandLegge(2005),bothcontactanglemeasurementsandTGAindicatethatPTMS
TMOSsolgelsarehydrophobicandthatincreasingtheproportionofPTMSincreasesthehydrophobicity
ofthesolgels.Therefore,thehydrophobicityofthesolgelappearstohaveastrongereffectwhenpure
waterisusedastheequilibratingsolutionratherthanwhenglycerolwatersolutionsareused.
74
200
PeakDesorption Temperature(oC)
190
180
170
160
150
140
130
0
10
20
30
40
50
60
70
80
CSG4
NeatCelite
90
100
Figure6.8:PeakdesorptiontemperaturesforCeliteR632()solgelsandplainCeliteR632(o)basedondifferentlevelsof
glycerolintheequilibratingsolution.Errorbarsrepresenta95%confidenceinterval.
Thesephenomenaareimportantinabiodieselproductionprocessessinceglycerolisabyproductofthe
reaction.ThehighlyabsorptivepropertiesoftheCelitesolgelwillinhibitthelipaseandpreventthe
reactionfromproceeding.Caremustbetakentoensurethatglycerolisremovedwhenthereactionis
runonacontinuousbasis.However,atlowlevelsofglycerol(i.e.10%byvolume),verylittleglycerolis
absorbed in the Celite solgel material (approximately 2.5% by mass), so this phenomena becomes
moreofaconcernastheconcentrationofglycerolincreases.
6.4 Conclusions
AproceduretoimmobilizelipaseinsolgelssupportedonCeliteR632wasdevelopedandshowntobe
valuable for the transesterification of triolein to produce methyl oleate. Three main properties were
consideredforcomparison:proteinloading,enzymaticactivity,andglycerolwateradsorption,forfour
differentsolgelpreparations:unsupportedsolgellowproteinlevel(USG4),Celitesolgellowprotein
level(CSG4),unsupportedsolgelhighproteinlevel(USG12),andCelitesolgelhighproteinlevel(C
SG12).
75
Based on the amount of protein retained with respect to that loaded onto the solgel, the
immobilizationprocedureismoreeffective(i.e.incompleteproteinimmobilizationisminimized)atthe
lowproteinlevels.ThewaterabsorbedbytheCeliteduringsolgelpreparationmustberemovedasit
inhibitsthelipasethemaximumproductionofmethyloleate(80%in6h)wasachievedwiththehigh
protein content Celite solgel (CSG12) after a predrying step. Over a 1.5 yr time frame, both the
unsupportedandCelitesupportedsolgelsexhibitedhighstabilitywhenstoredat4C.Thesolgelwas
found to provide a protective barrier that inhibits the absorption of the glycerolwater solution.
Therefore,theremovalofglyceroliskeyinpreventingtheinhibitionofthelipase,butisonlyaconcern
atveryhighlevelsofglycerol(approximately75%glycerolbyvolume).
Based on this study, a Celite supported solgel immobilized lipase was developed that achieves high
productionofmethyloleateinashortreactiontime.Althoughtheimmobilizationprocedureisstraight
forward,somesubtletiesofthenewmaterialexistincludingadesirablepredryingsteptoremovewater
absorbedduringimmobilization,andthepotentialforglycerolwaterabsorptionwhenhighquantitiesof
glycerolarepresent.Inadditiontothehighconversion,theimmobilizationregimeishighlystableover
1.5yearswithoutsophisticatedstoragerequirements.
of a Postgraduate Scholarship to SMM and a Discovery Grant to RLL. We thank Novozymes North
America for supplying samples of the lipase formulation and World Minerals for supplying samples of
theCelitesupportmaterials.
76
Chapter 7
KineticModellingoftheProductionofMethylOleate
byCeliteSupportedLipaseSolgels
Overview
This study illustrates the benefits of Celite supported lipase solgels for the transesterification of
trioleintoproducemethyloleate.Differentmethanolconcentrationsandtemperaturesweretestedto
identify any effects due to methanol or temperature. A pingpong bibi kinetic model was developed
and validated taking into account the inhibition effects of methanol and glycerol as well as the
temperatureeffectontheenzyme.Althoughinitialreactionratekineticmodelsareusefulforpredicting
thekineticswhennoproductsarepresentinthereactionmedium,acompletekineticmodelbeyondthe
initial conditions that considers glycerol inhibition is important. The model developed was consistent
with the experimental data (R2=0.95) showing an increase in methyl oleate production capacity with
increasing methanol concentration up to an optimal range of 1.3 M to 2.0 M depending on the
temperature.Increasingthetemperatureincreasedtheinitialreactionrate,butnothermalinactivation
of the enzyme was observed over the temperature range examined. Based on the kinetic constants
obtained,themaximumvelocityofthereversereactionisabout25%slowerthanthatoftheforward
reaction and glycerol inhibition has a more significant effect on the reaction kinetics than methanol
inhibition.Themodelthatwasdevelopedisoffundamentalimportanceforunderstandingtheeffects
of methanol and glycerol inhibition as well as the effect of temperature on the production of methyl
oleateusinglipasemediatedenzymatictransesterification.
Keywords
enzymeimmobilization,biodiesel,solgel,transesterification,kineticmodelling,Celite
77
7.1
Introduction
Biodieselisacleanburningfuelthatconsistsofamixtureoffattyacidalkylestersandcanbeproduced
via the transesterification of triglycerides. Biodiesel is typically produced using an alkaline or acid
catalyst chemical catalysts produce very high yields, but are there are issues with high energy
consumption, significant downstream processing, the necessity for wastewater treatment, and low
selectivityleadingtoundesirablesidereactions(AlZuhair,2005;Bajajetal.,2010;Fukudaetal.,2001;
Shah et al., 2004; Soumanou and Bornscheuer, 2003). Alternatively, lipase can be used to catalyze
transesterification reactions without the associated drawbacks of chemical catalysts. Enzymes do,
however, have their own shortcomings high cost, potential for deactivation, and low reaction rates
(Akohetal.,2007;Fukudaetal.,2001;Ganesanetal.,2009;Marchettietal.,2007;RoblesMedinaet
al.,2009;VasudevanandBriggs,2008).
Economically,itisessentialtoimmobilizelipaseforreuseifanenzymatictransesterificationprocessisto
be competitive with the alkaline process (AlZuhair, 2007; Jegannathan et al., 2011). Immobilized
enzymes,specificallythoseimmobilizedinsolgels,areknowntohaveincreasedactivity,selectivity,and
stabilityaswellasbeingmoreeasilyseparatedandreused(Aucoinetal.,2004;Pirozzietal.,2009;Reetz
et al., 1996; Reetz, 1997). Combined with the challenges of chemical catalysis, these benefits make
immobilizedenzymesespeciallyappealingforfattyacidalkylesterproduction.
Inanattempttofurtherimproveenzymaticactivityandstability,reduceany complicationscaused by
mass transfer limitations, and facilitate separation of the enzyme from the reaction media for reuse,
there has been a significant amount of research on immobilizing solgels on a variety of inorganic
supportmaterials.Immobilizinglipasesolgelsondiatomaceousearthsupportshasproventoproducea
verystableandactivebiocatalystfortheenzymaticproductionoffattyacidmethylesters(Meunierand
Legge,2010;MeunierandLegge,2012).
Enzymatic transesterification is complicated by the potential for lipase inactivation by the alcohol
substrate, glycerol product, and temperature. The required stoichiometric ratio of triglyceride to
alcohol is 1:3; however, the alcohol concentration is typically limited by lipase inhibition. In several
studies, a methanol concentration less than the stoichiometric amount was necessary to prevent
adverse effects of alcohol inhibition during transesterification (Shimada et al., 1999; Watanabe et al.,
2000;Watanabeetal.,2001;Watanabeetal.,2002;Xuetal.,2004).
78
When glycerol inhibition affects the enzymatic production of fatty acid methyl esters, the removal of
glycerolfromthereactionmediaviadialysis(BlafiBaketal.,2002),silicabedadsorption(Mazzieriet
al.,2008;Samukawaetal.,2000;Yorietal.,2007),orgravitationalsettling(Chenetal.,2009;Hamaet
al.,2011a;Hamaetal.,2011b)hasbeenexplored.
Enzymatic transesterification is advantageous due to low temperature requirements and lower
operatingcostsassociatedwiththedecreasedenergyrequirements.However,workingwithenzymes
that are very sensitive to temperature can add undesirable constraints to the operating conditions.
Immobilizationoflipaseinsolgelshasbeenshowntoreducethetemperatureandalcoholsensitivityof
theenzyme(Hsuetal.,2001a;Hsuetal.,2001b;Reetzetal.,1996;Reetz,1997).Immobilizedenzymes
show improved thermal stability in comparison to free enzymes, but there is still the limitation that
although higher temperatures increase the enzymatic activity they also cause thermal deactivation of
theenzyme(Balcoetal.,1996).
Pingpongbibimechanismshavebeenwidelyusedforthekineticmodellingofbiodieselproductionvia
enzymatictransesterification(AlZuhair,2005;AlZuhairetal.,2007;AlZuhairetal.,2009;Cheirsilpet
al.,2008;Dossatetal.,2002;Xuetal.,2005).Comprehensivekineticequationsforenzymecatalyzed
reactions, including pingpong bibi mechanisms with competitive inhibition by both substrates and
products, have been developed and are widely used as a starting point for enzymatic kinetic studies
(Cleland,1963a;Cleland,1963b;Cleland,1963c).Kineticstudiesofenzymaticbiodieselproductionare
often based on the initial rate version of the pingpong bibi kinetic equation, and, therefore, only
accountforsubstrateinhibitionsinceglycerolisnotpresentinthesysteminitially(AlZuhair,2005;Al
Zuhairetal.,2007;AlZuhairetal.,2009;Dossatetal.,2002;Xuetal.,2005).Equation7.1isoftenused
forinitialreactionratestudiesbasedonthereaction
where isthetriglyceride, is
thealcohol, istheglycerol,and isthemethyloleate:
where: istheinitialreactionrate,
isthemaximuminitialreactionvelocity,
alcoholandtriglycerideconcentrationsrespectively,
forthealcoholandtriglyceriderespectively,and
and
(7.1)
and
arethe
aretheapparentMichaelisconstants
isthealcoholinhibitionconstant.
Theobjectiveofthisstudywastomodelthekineticsofthetransesterificationoftrioleinandmethanol
for the production of methyl oleate and glycerol using Celite supported lipase solgels. A series of
79
experimentswerecompletedwithvarioustemperaturesandmethanolconcentrationstodeterminethe
effect of these parameters. The initial rate kinetic equation was originally considered, but a more
comprehensivemodelwassubsequentlydevelopedandvalidatedwhichincludestheeffectofglycerol
inhibition. The kinetic study presented provides a fundamental understanding the effects of glycerol
andmethanolinhibitionontheenzymeaswellasthebenefitsofCelitesupportedsolgelscontaining
immobilizedlipase.
7.2
Experimental
7.2.1
Materials
Novozym435wasagiftfromNovozymesNorthAmericaInc.(Franklinton,NC)containinglipasefrom
Candida antarctica with a reported activity of 10 000 PLU/g. Lipase PS Amano SD (Amano Enzyme
USACo.,Elgin,IL)withaactivityof25800U/gfromBurkholderiacepaciawasimmobilizedinsolgels
composedof80%trimethoxypropylsilaneand20%tetramethylorthosilicate(SigmaAldrichCanadaLtd,
Oakville,ON).The lipase solgelwassupportedon Celite R632 fromWorld Minerals (SantaBarbara,
CA). Ultrapure water was from a MilliQ water purification system from Millipore (Billerica, MA). All
otherchemicalsusedwerereagentgrade.
7.2.2
Methods
7.2.2.1
CeliteR632supportedlipasesolgelswereproducedaspreviouslyreported(MeunierandLegge,2010).
Trimethoxypropylsilane(14mL)andtetramethylorthosilicate(2.95mL)werecombinedwithultrapure
water(17.4mL)and0.1MHCl(200L)andsonicatedfor1htoallowhydrolysisoftheprecursors.The
solutionwasrotaryevaporatedat40Cfor30mintoremoveanywateroralcohol.6.5gofLipasePS
AmanoSDwasdissolvedin18mLof50mMphosphatebufferatpH=7.0,and14mLwasaddedto
therotaryevaporatedprecursorsolution.Thelipasesolgelsolutionwasaddedtothesupportmaterial
inaratioof1.5gmixture/gCeliteR632andmixedthoroughly.Themixturewasthendepositedintoa
Petridishandsealedfor24hat4C.Finally,theCelitesolgelwasdrieduncoveredat4Cuntilthe
dryingrateslowedtolessthan1mg/hr.AftertheCelitesolgelsweredried,thesupportedsolgelwas
washedwith50mMpH=7.0phosphatebuffertwice(5mLbuffer/gsolgelperwash)toremoveanyfree
protein. The resulting gels were dried overnight at room temperature and then stored in a sealed
containerat4Cuntiluse.
80
Unsupported lipase solgels were prepared in a similar fashion but without the support and solgel
mixing step. The resulting solgel was crushed with a mortar and pestle and the wash solutions
separatedfromthesolgelbycentrifugationat1250xgfor10min.
7.2.2.2
Totalproteinquantification
ThetotalproteincontentofthesupportedsolgelswasdeterminedusingaBCAproteinassaykit(Pierce
BiotechnologyInc.,Rockford,IL)usingamicroplatemethodwithabovineserumalbumin(BSA)standard
from 252000 g/mL. The absorbance was measured at 562 nm using a BioTek Synergy 4 Microplate
Reader(BioTekUS,Winooski,VT).Samplesofthestockenzymesolutionwerecomparedwithsamples
fromthewashsolutionsandamassbalanceperformedtodeterminetheamountofproteinremaining
inthesolgel.
7.2.2.3
Enzymatictransesterification
Todeterminetheenzymatictransesterificationcapacitiesofthesupportedsolgels,Celitesolgel(1g),
triolein (6.5 mmol), and methanol (4 mmol unless otherwise stated) were combined in a reaction vial
and agitated at a controlled temperature for 6 h as previously described (Meunier and Legge, 2010).
TenLsamplesweretakeneveryhourfromthereactionvialanddilutedwith990Lhexaneand100L
internal standard (heptadecanoic acid methyl ester). The diluted samples were analyzed for methyl
oleate content using a GCMS (CP3800 gas chromatograph, Saturn 2000 mass spectrometer/mass
spectrometer,VarianInc.,Mississauga,ON)equippedwithaCPWax52CBfusedsilicacolumn(CP8513,
VarianInc.,Mississauga,ON).Theinjectorwassetat250Cwithasplitratioof50;theovenwasheldat
170Cfor10min,rampedfrom170Cto250Cat10C/minandheldat250Cfor2min;heliumwasthe
carriergasataflowrateof1mL/min;andtheinjectionvolumewas1L.
Only the methyl oleate concentration could be measured directly using this GCMS method, so the
concentrationsoftheothercomponents(alcohol,triolein,andglycerol)weredeterminedusingamass
balance based on the reaction stoichiometry. No other components were observed in the GCMS
chromatograms.
7.2.2.4
Kineticmodelling
MATLAB software (Version 7) from Mathworks, Natick, Massachusetts was employed for parameter
estimation. The fitting coefficients were taken as positive values and optimized using builtin Matlab
functionsfornonlinearleastsquaresfittingwiththeTrustregionalgorithm.Theordinarydifferential
equation (ODE) of methyl oleate concentration was solved using a builtin Matlab function, ode45,
81
which is based on an explicit fourth order RungeKutta formula. The relative error tolerance and
absoluteerrortoleranceweresetat104and106,respectively.
The 5 kinetic constants for the initial rate equation (section 7.3.2) were obtained using the initial
reactionratedatafromall13reactionconditionstoensureagoodmodelfit.Toconsidertheentire6h
reaction (section 7.3.3), 5 additional kinetic constants were necessary to be used in combination with
the5constantsfromtheinitialrateequation(totalof10kineticconstants).Toobtainthe5additional
kineticconstants,oneexperimentalconditionwasused(T=40Cand[A]=0.6M).Theremaining12
reactionconditions(T=40,50,and60Cand[A]=0.3,0.6,0.9and1.8M)wereusedtoverifythefitof
thekineticmodelandtotestthemodelflexibilitybeyondtheprimaryreactionconditions.Themodel
developed was validated using 83 data points at different temperatures, alcohol concentrations, and
times.
7.3
Resultsanddiscussion
7.3.1
Comparisonoflipaseimmobilizationschemes
Comparing the performance of Novozym 435 (N435), free lipase (Free), unsupported lipase solgels
(US), and Celite supported lipase solgels (CSG) at 40C and a molar ratio of methanol to triolein of
1.5:1, CSG had the highest conversion after 6 h, initial reaction rate and initial activity compared to
N435, US and Free (Table 7.1). Celite supported lipase solgels have enhanced transesterification
abilities giving them a unique advantage over both unsupported lipase solgels, free lipase, and the
commerciallyimmobilizedNovozym435.
Table7.1:ComparisonofNovozym435(N435),freelipase(Free),unsupportedsolgel(US),andCelitesupportedsolgel
(CSG)intermsoftheirfinalpercentconversion,initialreactionrate,andinitialactivityat40Candmethanoltotriolein
molarratioof1.5:1.*ForN435,theinitialactivityisbasedontheweightofN435ratherthantheweightoflipasesincethe
amountoflipaseintheenzymepreparationisunknown.
Sample
N435
Free
US
CSG
Finalconversion
(%)
8.2
3.5
38.2
62.2
Initialreactionrate
(M/min)
333
257
1105
3740
82
Initialactivity
(mol/minglipase)
43*
398
5712
9750
7.3.2
Kineticmodelinitialreactionrate
Toevaluatethevalidityoftheinitialreactionratemodelatdifferenttemperaturesandmethanolratios,
Equation 7.1 was adapted to include a temperature dependent exponential term (T in degrees K) as
isthetemperatureconstant.
showninEquation7.2where
(7.2)
Thiskinetic modeldemonstratesgoodagreementbetween theinitialexperimentaldatacollectedand

the model predictions using the kinetic constants in Table 7.2 (Figure 7.1). Both the model and
experimental data show an increase in the initial rate with increasing temperature and methanol
concentrationuptoapproximately1 Matwhich pointthealcoholinhibitionresultedinadecreasein
theinitialreactionrate.
InitialRate(M/min)
0.004
0.003
0.002
0.001
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Methanol Concentration (M)
1.6
T=40CExp
T=50CExp
T=60CExp
T=40CModel
T=50CModel
T=60CModel
1.8
Figure7.1:Initialreactionrateasafunctionofmethanolconcentrationatthreetemperatures(40C,50C,and60C).The
linesrepresentthepredictionsfromthemodelprovidedbyEquation7.2andthesymbolsrepresenttheexperimentaldata.
83
Table7.2:KineticparameterestimationforthemodelspresentedinEquations7.2and7.3.
Coefficient
(mol/Lmin)
(mol/Lmin)
(mol/L)
(mol/L)
(mol/L)
(mol/L)
(mol/L)
(mol/L)
Equation7.2
1.136
5.5x104
1.2x104
0.137
0.0188
Equation7.3
0.8236
1.136
0.3709
5.5x104
1.2x104
0.7362
0.6133
0.137
0.0394
0.0188
AlthoughEquation7.2agreeswith the experimentaldatafortheinitialreactionrate,significantover

predictionforeverydatapointatthe6hmethyloleateconcentrationwasfound(Figure7.2inset).This
indicates that the glycerol inhibition is an important consideration for kinetic modelling of the entire
concentrationprofile(BlafiBaketal.,2002;Dossatetal.,1999;Hong etal.,2011;Watanabeet al.,
2000).
1.2
1.0
[M](M)
6hMethylOleateConcentration (M)
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.8
0.5
1.5
[A](M)
0.6
0.4
0.2
0.0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
T=40CExp
T=50CExp
T=60CExp
T=40CModel
T=50CModel
T=60CModel
Figure7.2:Methyloleateconcentrationat6hwithrespecttotheinitialmethanolconcentrationatthreetemperatures
(40C,50C,and60C).Thelinesrepresentthemodelpredictionsandthesymbolsrepresenttheexperimentaldata.The
mainplotrepresentstheresultsfromEquation7.3whiletheinsetplot(upperleft)representstheresultsfromEquation7.2.
84
7.3.3
Kineticmodeltimedependentreactionrate
A more comprehensive model (Equation 7.3) was developed by modifying the classic pingpong bibi
kineticmodeltoincludeonlythecompetitiveinhibitiontermsforthealcoholsubstrateandtheglycerol
product(Rizzietal.,1992).
(7.3)
T
with:
where: isthereactionrate;
reactions,respectively;
andglycerol,respectively;
and
,
,and
andthemaximumvelocitiesofthereverseandforward
aretheconcentrationsofalcohol,triolein,methyloleate
istheequilibriumconstant;
constantsforalcohol,triolein,methyloleateandglycerol,respectively;
,and
aretheMichaelis
and
aretheconstants
for dissociation of the inhibitor from the enzymeinhibitor complex for alcohol and glycerol,
respectively;and
isthetemperatureconstant.
The performance of each model was validated based on the final methyl oleate concentration
demonstrating the improvement in prediction when glycerol inhibition is considered (Table 7.3 and
Figure7.2). Considering onlythefinalobservedmethyloleate concentration,Equation7.2providesa
verypoorfitwithaR2=0.201whileEquation7.3providesagoodfitwithaR2=0.910.
7.3.4
Methyloleateconcentrationprofile
Usingonlyonedatasetformodelfitting(T=40Cand[A]=0.6M)andtheremainingdatatovalidate
the model, the prediction of the entire methyl oleate concentration profile based on methanol and
glycerol inhibition kinetics (Equation 7.3) demonstrates good agreement with the experimental values
(Figure7.3).Thismodelprovidesanexcellentfitfortheentiresetofexperimentalresults(thefitting
parameterspresentedinTable7.3consideronlythefinalconcentration),withtheaccuracyofthemodel
determinedasfollows:SSE=0.20,R2=0.95,andRMSE=0.049.
85
Table7.3:Comparisonoftheperformanceofthetwoproposedmodels(Equations7.2and7.3)intermsoftheirSSE,R2,and
RMSEforthe6hmethyloleateconcentration.
Model(Equation)
InitialRate(7.2)
WithGlycerolInhibition(7.3)
Sumofsquared
errors(SSE)
0.447
0.050
Coefficientof
determination(R2)
0.201
0.910
Rootmeansquared
error(RMSE)
0.193
0.064
The kinetic parameters used to fit the model (Table 7.2, Equation 7.3) indicate that the maximum
velocity of the reverse reaction is approximately 25% slower than that of the forward reaction and
glycerolinhibitionhasamuchstrongereffectonthereactionratethanmethanolinhibitionsincehigher
inhibitionconstantscorrespondtolessofaninhibitioneffectastheyarepresentinthedenominatorof
the kinetic model. These kinetic constants demonstrate the importance of glycerol inhibition and the
reversereactionwhenconsideringthekineticsofenzymatictransesterification.Inimmobilizedenzyme
reactors, glycerol is known to adsorb onto the enzymatic support and cause an activity decrease by
creatingahydrophiliclayeraroundtheenzymethatpreventsthehydrophobicsubstrate(triolein)from
accessingtheenzyme(BlafiBaketal.,2002;Dossatetal.,1999;Hongetal.,2011;Watanabeetal.,
2000).
0.8
0.7
PredictedMethylOleate(M)
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.1
0.2
0.3
0.4
0.5
ExperimentalMethylOleate(M)
0.6
0.7
0.8
Figure7.3:Correlationbetweenthepredictedandexperimentalmethyloleateconcentrationvaluesusingthekineticmodel
describedbyEquation7.3.
86
Comparing the concentration profiles at 3 temperatures (40C, 50C, and 60C) and 4 methanol
concentrations(0.3M,0.6M,0.9M,and1.8M),verygoodagreementbetweenthemodelpredictions
and the experimental data are observed (Figure 7.4). The model does overpredict the methyl oleate
concentrations at the low methanol concentration (0.3 M), which is likely due to a mass transfer
inhibitioncausedbythe relativelysmallamountofmethanolpresentinthereactionmediumandthe
insolubilityofmethanolintrioleinsincethisisasolventfreesystem(Ognjanovicetal.,2009).Atevery
temperature considered, the concentration of methyl oleate produced during the reaction increased
with increasing initial methanol concentration indicating that at these levels methanol inhibition does
not have a detrimental effect on the performance of the enzyme at concentrations up to 1 M
(approximately 1:1 molar ratio of methanol:triolein). When the methanol concentration is increased
beyond 1 M no further increase in methyl oleate concentration was observed, demonstrating the
commencement of methanol inhibition. This is in agreement with the literature where methanol
inhibitionisdocumentedindicatingthatlessthanstoichiometricamountsofmethanolarenecessaryto
preventinhibitionandthata1:1molarratiooftriolein:methanolisoptimal(Leeetal.,2011;Shimadaet
al.,2002;Xuetal.,2004).
Atahighmethanolconcentration(1.8M,Figure7.4d),thereisaminimaleffectoftemperaturebetween
40C and 50C which is expected since the optimal temperature for lipase is within this temperature
range. These results indicate that the Celite supported lipase solgels are quite temperature and
methanolresistantovertherangeoftemperaturesandconcentrationsconsidered(Figure7.4).Other
studiesexaminingtheoptimaltemperaturerangeofimmobilizedlipasefortheproductionofbiodiesel
indicatealowertemperatureoptimumof3040Cwithcommerciallyimmobilizedenzymes(Kseetal.,
2002;Salisetal.,2005)comparedtosolgelimmobilizedlipaseswhichhavereportedoptimabetween
40and70C(Hsuetal.,2003;Hsuetal.,2004;Moreiraetal.,2007).Inarecentstudy,Chesterfieldetal.
(2012)foundthatthermaldenaturationofNovozym435beganat47C.
Consideringtheinitialreactionratedata(Figure7.1)andthefinalmethyloleateconcentration(Figure
7.2),itisevidentthatincreasingthetemperatureincreasestheinitialreactionrate,buthasaminimal
effectonthefinalmethyloleateconcentration.Thisindicatesthattheincreasedperformancefromthe
temperature effect causes the final product concentration to be attained more quickly, but does not
cause the final product concentration to increase. It is argued that the inhibitory effect of glycerol
supersedesthetemperatureeffectseeninitially.
87
a)T=40C
MethylOleate(M)
0.8
0.7
[A]=0.9MModel
0.6
[A]=0.9MExp
0.5
[A]=0.6MModel
0.4
0.3
[A]=0.6MExp
0.2
[A]=0.3MModel
0.1
[A]=0.3MExp
0
0
100
200
300
Time(min)
b)T=50C
0.8
MethylOleate(M)
0.7
0.6
[A]=0.9MModel
0.5
[A]=0.9MExp
0.4
[A]=0.6MModel
0.3
[A]=0.6MExp
0.2
0.1
0
0
100
200
300
Time(min)
c)T=60C
MethylOleate(M)
0.8
0.7
[A]=0.9MModel
0.6
[A]=0.9MExp
0.5
[A]=0.6MModel
0.4
0.3
[A]=0.6MExp
0.2
[A]=0.3MModel
0.1
[A]=0.3MExp
0
0
100
200
300
Time(min)
d)[A]=1.8M
0.8
MethylOleate(M)
0.7
0.6
T=40CModel
0.5
T=40CExp
0.4
T=50CModel
0.3
T=50CExp
0.2
0.1
0
0
100
200
300
Time(min)
Figure7.4:Comparisonofmodelpredictions(lines)andexperimentaldata(symbols)forfourmethanolconcentrations(0.3
M,0.6M,0.9M,and1.8M)andthreetemperatures(40C,50C,and60C).Plota)T=40C,b)T=50C,c)T=60C,d)
[A]=1.8M.
88
7.3.5
Kineticmodelpredictionsfortransesterification
Basedonthekineticmodel,theconditionswereextendedbeyondthosetestedexperimentallytoassess
the potential of the model developed (Figure 7.5). This model predicts an optimal methanol
concentrationrangeofapproximately1.3Mto2.0Mwhichisinagreementwithexperimentsthatshow
that increasing the methanol to triolein molar ratio from 1.8 to 3 decreases the 6 h methyl oleate
concentration by about 60% (data not presented). From the model predictions, increasing the
temperature from 30C from 70C shifted the optimal methanol concentration from 1.3 M to 2.0 M
indicating that there is a positive interaction between temperature and methanol (Figure 7.5). The
optimalalcoholconcentrationincreasedlinearlywithtemperatureatarateof0.017M/K.Additionally,
at higher temperatures the impact of increasing the methyl oleate concentration was more profound
than at lower temperatures demonstrating that at higher temperatures higher methyl oleate
concentrations can be achieved, but small changes in the initial methanol concentration have a more
intenseeffectonthefinalproductconcentration.Sincediffusivityinliquidsisknowntoincreasewith
increasing temperature, the mass transfer of glycerol away from the enzyme will improve at higher
temperaturesresultinginanincreaseinthereactionrate(Treybal,1980).
Methyl Oleate Concentration at 6-hr (M)
0.8
0.7
1
0.6
0.8
0.5
0.6
0.4
0.4
0.2
0.3
0
4
3
2
1
300
305
310
315
320
325
330
335
340
345
0.2
0.1
Temperature (K)
Figure7.5:Surfaceplotofthekineticmodel(Equation7.3)predictionsforthemethyloleateconcentrationafter6hat
differentinitialmethanolconcentrations(04M)andtemperatures(3070C).
89
7.4
Conclusions
Celitesupportedlipasesolgelshavesuperiorenzymaticperformanceoverunsupportedlipasesolgels,
free lipase, and Novozym 435. The typical kinetic model for predicting enzymatic transesterification
performance was shown to be useful only when considering the initial conditions of the reaction.
However, as the reaction progresses, the presence of glycerol in the reaction medium becomes more
importantanditisessentialforakineticmodeltotakeintoaccounttheenzymaticinhibitioncausedby
this product. A kinetic model based on a pingpong bibi mechanism was developed that takes into
accountbothmethanolandglycerolinhibitioneffectsontheproductionofmethyloleateviaenzymatic
transesterification using Celite supported lipase solgels. The model developed shows excellent
agreement with the experimental data at varying temperatures and initial methanol concentrations
(R2=0.95).Basedonthekineticparameters,glycerolhasastrongerinhibitioneffectthanmethanoland
the maximum velocity of the reverse reaction is approximately 25% slower than that of the forward
reaction. The experimental data show that increasing the concentration of methanol in the reaction
mediumincreasestheproductionofmethyloleateindicatingthatalthoughmethanolisinhibitory,itis
not enough to decrease the achievable product concentration within the range of concentrations
studied. There is also a positive temperature effect on the enzyme that increases the initial reaction
rate,butnotthefinalmethyloleateconcentrationachieved.Theresultsfromthemodelindicatethat
there an optimal methanol concentration range of 1.3 M to 2.0 M that increases with increasing
temperature,butnotemperaturedeactivationoftheenzymewasobserved.
7.5
Acknowledgements
ofaPostgraduateScholarshiptoSMMandaDiscoveryGranttoRLL.WethankAmanoEnzymeUSAfor
supplying samples of the lipase formulation and World Minerals for supplying samples of the Celite
supportmaterials.
7.6
Nomenclature
Acronyms
CSG
90
Free
Freelipase
N435 Novozym435
R2
RMSE Rootmeansquarederror
SSE
Sumofsquarederrors
US
Unsupportedlipasesolgel
Parameters
Alcoholconcentration(M)
Glycerolconcentration(M)
Methyloleateconcentration(M)
Triglycerideconcentration(M)
Temperatureconstant(dimensionless)
Equilibriumconstant(dimensionless)
Alcoholinhibitionconstant(M)
Glycerolinhibitionconstant(M)
ApparentMichaelisconstantforalcohol(M)
ApparentMichaelisconstantforglycerol(M)
ApparentMichaelisconstantformethyloleate(M)
ApparentMichaelisconstantfortriglyceride(M)
Temperature(K)
Reactionvelocity(M/min)
Initialreactionvelocity(M/min)
91
Maximumvelocityoftheforwardreaction(M/min)
Maximumvelocityofthereversereaction(M/min)
92
Chapter 8
KineticandMassTransferModellingofMethylOleate
ProductioninanImmobilizedLipasePackedBed
Reactor
Overview
ThisstudyinvestigatesaCelitesupportedsolgelimmobilizedlipasepackedbedreactoralongwiththe
developmentofakineticsandmasstransfermodelfortheproductionofmethyloleate.Todetermine
the reaction kinetics, an initial rate equation accounting for alcohol inhibition was used for both the
CelitesupportedlipasesolgelandNovozym435.ItwasfoundthattheCelitesupportedsolgelhad
ahighmaximumreactionvelocityandwaslessinhibitedbymethanolincomparisontoNovozym435,
supportingthepracticalpotentialofCelitesupportedsolgelsforuseinenzymebasedreactorsforthe
production of biodiesel. To describe the continuous packed bed reactor performance, an efficiency
correlationwasdevelopedtoaccountfortheeffectoftheflowrateandglycerolconcentrationonthe
enzymatic activity. These reaction kinetics, mass transfer, and efficiency equations were used to
describetheeffectofflowrateinthepackedbedreactor;themodeldevelopedwasingoodagreement
with the experimental data. Increasing the flow rate was found to increase the reactor performance
presumablybyaidingintheremovalofaglycerollayeraroundthebiocatalyst.Additionally,amaximum
achievable methyl oleate concentration was observed due to presence of glycerol. Comparing the
Celite solgel packed bed reactor performance to that of the Novozym 435, a higher product
conversion was achievable using the Celite solgel in a shorter period of time indicating the superior
performance of the supported solgel for enzymatic transesterification. The Celite solgel did not
exhibitanyactivitylossoverafivedaytimespansupportingthepotentialforthiscatalystinapacked
bedreactorfortransesterification.
Keywords
Packedbedreactor,lipaseimmobilization,transesterification,solgel,biodiesel,Celite
93
8.1 Introduction
Theindustrialproductionofbiodiesel(fattyacidalkylesters)fromtriglyceridesiscatalyzedchemically
byanacidicoralkalineprocess.Theuseofthesechemicalcatalystsiscomplicatedbytherequirement
of higher than ambient operating temperatures (5580C), undesirable sidereactions caused by raw
material impurities (such as free fatty acids and water), difficult recovery and purification of the
products as well as the catalyst, and the need for substantial wastewater treatment (Ganesan et al.,
2009). Alternatively, enzymatic transesterification of triglycerides using lipase can overcome these
challenges by lowering the operating temperature and providing a more selective reaction with less
processingstepsandwithouttheneedforextensivewastewatertreatment(Fjerbaeketal.,2009).The
majordrawbacksofenzymaticallycatalyzedbiodieselproductionarethehighcostoftheenzyme,low
reactionrate,andthepotentialforenzymeinhibitionanddeactivationovertime(RoblesMedinaetal.,
2009).
Basedonaneconomicstudy,biodieselproductionusingimmobilizedlipasewasfoundtobemorecost
effectivethanusingsolublelipase,buttheimmobilizedlipaseprocesswasmorecostlythanthealkali
process and the enzyme must be reused more than five times to be industrially competitive
(Jegannathan et al., 2011). The benefits of immobilized enzymes over soluble enzymes include the
enhancedabilityfortheirrecoveryandreusewhichdecreasestheeffectivecostoftheenzyme,andthe
potentialforimprovedactivityandstabilitydependingonthenatureoftheimmobilizationscheme(Fan,
2012).
Byimmobilizinglipaseinasolgel,thethermalandchemicalstabilityisknowntobeenhancedaswellas
theenzymaticactivity(Aucoinetal.,2004;Pirozzietal.,2009;Reetzetal.,1996;Reetz,1997).Many
studieshavefocusedonasecondarysupportmaterialfortheimmobilizationoflipasesolgelstohelp
overcome the potential for diusion limitations, to make the immobilized lipase more practical for
enzymaticreactors,andtofurthersimplifytherecoveryandreuseoftheenzyme(Brnyiketal.,2000;
MeunierandLegge,2010;MeunierandLegge,2012;Oraireetal.,2006;Pogorilyietal.,2007).
Research on enzymatic biodiesel production can be classified based on whether the process is
conducted using a solventbased or solventfree reaction media. Although solventfree
transesterificationiscomplicatedbyhighviscosityandanenhancedpotentialforinhibition,theuseof
organic solvents is undesirable due to the increased costs, environmental concerns, and the need for
furtherdownstreamseparation(Dossatetal.,2002;SelmiandThomas,1998;SzczesnaAntczak etal.,
94
2009;Tongboriboonetal.,2010).Amajorconcernforsolventfreebiodieselproductionistheincrease
inenzymeinhibitionbythealcoholwhichexistsasdropletsintheorganicphaseduetoitsinsolubilityin
theoilphase(Ognjanovicetal.,2009).
Most enzymatically catalyzed biodiesel production plants operate with stirred tank batch reactors
(Fjerbaek et al., 2009). However, the agitation required for this type of reactor is one of the primary
causes of enzyme deactivation, and by implementing a packed bed reactor without agitation, higher
conversions can be achieved (Ognjanovic et al., 2009). In addition, packed bed reactors have more
industrial flexibility than batch stirred tank reactors. The commercially immobilized lipase, Novozym
435,hasbeenthesubjectofmanypackedbedreactorstudiesandissufficientlystableandactivefor
useforthecontinuousproductionofbiodiesel(Changetal.,2009;Chenetal.,2011;Halimetal.,2009;
Hamaetal.,2011a;Hamaetal.,2011b;Ognjanovicetal.,2009;Royonetal.,2007;Shawetal.,2008).
Studies have also been conducted with other immobilized lipases (Lee et al., 2010; Nie et al., 2006;
Wangetal.,2011).
Enzymatic activity is known to be affected by a variety of chemicals as well as operating conditions.
Lipase inhibition by the alcohol substrate in biodiesel production has been an extensive area for
research with some alternatives for maintaining enzymatic activity including the pretreatment of the
lipase with organic solvents or the biodiesel reaction substrates or products (Chen and Wu, 2003;
Samukawaetal.,2000);theuseofalternateacylacceptorstoalleviatetheneedforalcoholsubstrates
(Du et al., 2004; Ruzich and Bassi, 2011; Xu et al., 2003); and the use of smaller than the required
stoichiometricamountsofalcoholinthereactionmedium(Shimadaetal.,1999;Shimadaetal.,2002;
Watanabeetal.,2000;Watanabeetal.,2001;Watanabeetal.,2002;Xuetal.,2004).
Inadditiontoinhibitionbythealcoholsubstrate,thepresenceoftheglycerolproductisknowntohave
anadverseeffectontheenzymaticactivity.Topreventundesirablereductionsinenzymaticactivityin
continuous biodiesel production, glycerol is typically removed from the product stream via dialysis
(BlafiBaketal.,2002),adsorptiononsilicabeds(Mazzierietal.,2008;Samukawaetal.,2000;Yoriet
al.,2007),orgravitydrivensettling(Chenetal.,2009;Hamaetal.,2011a;Hamaetal.,2011b).Froman
indepthstudyoftheinfluenceofglycerolonimmobilizedlipasetransesterification,Dossatetal.(1999)
determined that the phenomena causing the decrease in the enzymatic activity involves the
accumulationofahydrophiliclayerofglycerolaroundtheimmobilizedenzymewhichpreventsaccessof
thehydrophobicsubstrate(triglyceride)totheenzyme.
95
In this work, the initial reaction kinetics for Celite supported lipase solgels and Novozym 435 were
determined usingabatchstirredtankreactoratdifferentmethanolconcentrations.Usingthekinetic
data from these studies along with the appropriate mass transfer equations, the concentration of the
methyloleateproductwithrespecttotimeinapackedbedreactorwasmodelled,andtheeffectofflow
rate on the conversion analyzed. The performance of the batch and continuous reactorswith Celite
solgelimmobilizedlipaseandNovozym435weresubsequentlycompared.Finally,thereusabilityof
theCelitesolgelcatalystbedwasevaluated.
8.2 Experimental
8.2.1
Materials
Two lipases were used in this study: Novozym 435 (Novozymes North America Inc., Franklinton, NC)
andLipasePSAmanoSD(AmanoEnzymeUSACo.,Elgin,IL).Thebiologicalsourceoflipaseusedfor
Novozym435isCandidaantarctica,theactivitywas10000PLU/g,anditisreportedtobeimmobilized
on a macroporous acrylic resin. The biological source of the lipase for Lipase PS Amano SD is
Burkholderiacepaciaandthereportedactivitywas25800U/g.LipasePSAmanoSDwasimmobilized
inasolgelcomposedof80%trimethoxypropylsilaneand20%tetramethylorthosilicate(SigmaAldrich
Canada Ltd, Oakville, ON) and supported on Celite R632 (World Minerals, Santa Barbara, CA).
Ultrapure water was obtained from a MilliQ water purification system (Millipore, Billerica, MA). All
otherchemicalsusedwereofreagentgrade.
8.2.2
8.2.2.1
Methods
Enzymeimmobilization
CeliteR632supportedlipasesolgelswereproducedaspreviouslyreported(MeunierandLegge,2010).
Trimethoxypropylsilane(14mL),tetramethylorthosilicate(2.95mL),ultrapurewater(17.4mL),and0.1
MHCl(200L)werecombinedandsonicatedfor1htoallowhydrolysisoftheprecursors.Thesolution
wasrotaryevaporatedat40Cfor30mintoremoveanywateroralcohol.6.5goflipasePSAmanoSD
was dissolved in 18 mL of 50 mM phosphate buffer at pH 7.0, and 14 mL was added to the rotary
evaporatedprecursorsolution.Thelipasesolgelsolution(27mL)wasthenmixedthoroughlywiththe
CeliteR632support(18g)anddepositedintoaPetridishandsealedfor24hat4C.Finally,theCelite
solgelwasdrieduncoveredat4Cuntilthedryingrateslowedtolessthan1mg/hr.AftertheCelite
solgelsweredried,theywerewashedtwicewith50mMpH=7.0phosphatebuffer(45mLeachwash)to
96
removeanyfreeprotein.Theresultinggelsweredriedovernightatroomtemperatureandstoredina
sealedcontainerat4C.Thisprocedureproducedapproximately25gofCelitesupportedsolgel.
8.2.2.2
Proteincontentdetermination
Thetotalproteincontentinthesupportedsolgelswasdeterminedbasedonamassbalanceapproach.
A BCA protein assay (Pierce Biotechnology Inc., Rockford, IL) with bovine serum albumin (BSA) as the
standardoveraconcentrationrangeof25to2000g/mLwasused.Theabsorbancewasmeasuredat
562 nm using a BioTek Synergy 4 Microplate Reader (BioTek US, Winooski, VT) according to the
recommendations from Pierce Biotechnology Inc. for the microplate method. Samples of the stock
enzymesolutionwerecomparedtosamplesofthewashsolutionsandamassbalancewasperformedto
determinetheamountofproteinremaininginthesupportedsolgel.
8.2.2.3
Batchstirredtankreactoroperation
Todeterminethekineticconstantsfortransesterification,theimmobilizedlipase(1gCelitesolgelor
50mgNovozym435)and6.5mLtrioleinwerecombinedwithvariousconcentrationsofmethanolinan
agitated(500rpm)batchreactorat40Cfor6haspreviouslydescribed(MeunierandLegge,2010).A10
LsamplewastakeneveryhourfromthereactionvialforGCMSquantificationofthemethyloleate.
8.2.2.4
Packedbedreactoroperation
Figure 8.1 is a schematic of the packed bed reactor apparatus which was fabricated inhouse and
consisted of a glass double walled reaction column of 1 cm diameter packed to approximately 17 cm
highwithimmobilizedlipase(6.00gfortheNovozym435and7.50gfortheCelitesupportedlipase
solgel). According to Thoenes (1994), the recommended particle diameter for packing materials is
typically310mmandthereactorheightshouldbe10toafewhundredtimestheparticlediameter.In
thisstudy, theparticlediameterwas1 mmforCelitesolgeland0.6mmforNovozym435andthe
reactorheightwas170timestheparticlediameter.Thedirectionoftheflowwasdownwardduetothe
highsubstrateviscosityandtheneedtoavoidanychallengesassociatedwithcatalystparticleretention.
The reactor was maintained at a constant temperature (40C) using a circulating water bath (NESLAB
RTE111,NeslabInstrumentsInc.,Portsmouth,NH,USA).Thereactionmixture(1.95mLmethanoland
49.5mL97%triolein)wasvigorouslymixedanddeliveredtothetopofthecolumnataconstantflow
rate using a peristaltic pump (Cole Parmer Masterflex Easyload L/S 751800, Montreal, QC, Canada).
Thereactionmixturewasrecycledthroughthepackedbeduntilnofurtherincreaseinconversionwas
observed.Aftereachpass,a10Lsamplewastakenfromtheliquidphaseformethyloleateanalysis.
97

Figure8.1:Schematicofthepackedbedreactorapparatus.
8.2.2.5
Quantificationoffattyacidmethylesters
The10Lsamplesfromthebatchstirredtankandpackedbedreactorsweredilutedin990Lhexane
and 100 L of 320 M heptadecanoic acid methyl ester (internal standard), and analyzed for methyl
oleate content using a GCMS (CP3800 gas chromatograph, Saturn 2000 mass spectrometer/mass
spectrometer,VarianInc.,Mississauga,ON)equippedwithaCPWax52CBfusedsilicacolumn(CP8513,
Varian Inc., Mississauga, ON). The injector was set at 250C with a split ratio of 50; the oven
temperature was held at 170C for 10 min, ramped 10C/min to 250C and held at 250C for 2 min;
helium was the carrier gas at a flow rate of 1 mL/min; and the injection volume was 1 L. Only the
methyl oleate concentration was measured using this method the methanol and triolein
concentrationsweredeterminedusingamassbalancebasedonthereactionstoichiometry.Noother
componentswereobservedintheGCMSchromatograms.
98
8.3 Modelling
8.3.1
Kineticmodel
Theinitialreactionrateforenzymaticbiodieselproductionwithinhibitionbythealcoholsubstratecan
be described by Equation 8.1 for the reaction
where is the triglyceride, is the
alcohol, istheglycerol,and isthemethyloleate(AlZuhair,2005;AlZuhairetal.,2007;AlZuhairet

al.,2009;Dossatetal.,2002;Xuetal.,2005).
where
is the maximum initial reaction velocity,
concentrations respectively,
triglyceriderespectively,and
,and
and
(8.1)
and
are the alcohol and triglyceride
are the apparent Michaelis constants for the alcohol and
isthealcoholinhibitionconstant.The4kineticconstants(
)forCelitelipasesolgel(CSG)andNovozym435(N435)catalystswereoptimizedusing
nonlinear least squares fitting with the Trustregion algorithm conducted with MATLAB software
(Version7,Mathworks)basedon13datapointsforCSGand17datapointsforN435.
8.3.2
Packedbedreactormodel
Consideringtheflowandreactioninapackedbedreactor,Equation8.2presentsadifferentialequation
to describe the substrate concentration with respect to the position along the height of the reactor
(Fogler,1999).
(8.2)
where isthefluidlinearvelocity,
istheexternalmasstransfercoefficient(
is the external surface area of the catalyst,
is the bulk concentration of the substrate,
),
is the
concentration of the substrate at the catalyst surface, and refers to the position along the reactor
height. The steady state operation of an enzymatic packed bed reactor can also be described by
Equation8.3(Dunn,2003).
1
99
(8.3)
where referstothereactionratewithrespecttothesubstrate.Consideringaseriesofpackedbed
andfluidizedbedstudies,Equation8.4presentsageneralequationfortheSherwoodnumber(
can be used to determine the mass transfer coefficient (
)that
) in Equation 8.2 (Dwivedi and Upadhyay,
1977;Seguinetal.,1996).
1.1
where is the void fraction in the reactor,
(8.4)
is the Reynolds number for the particle and
is the
Schmidtnumber.
8.3.3
Physicalproperties
TheWilkeandChangequation(Equation8.5)wasusedtoestimatethediffusivity,
,ofsolute ina
verydilutesolutioninsolvent where referstothemethanoland tothetriolein(Treybal,1980).

117.3
10
where is the dissociation factor for the triolein,

temperature, isthetrioleinviscosity,and
(8.5)
is the molecular mass of the triolein, T is the
isthemolarvolumeofthemethanol.Assuminguniform
sphericalparticles,thevoidfraction( )canbecalculatedbasedontheparticlediameter(
reactordiameter(
)andthe
)asshowninEquation8.6(Rase,1990).
0.4 1
0.42
(8.6)
Sincetrioleinconsistedofapproximately97.6%(v/v)oftheliquidmixture,thedensity( )anddynamic
viscosity ( ) of pure triolein were used for the modelling. Equation 8.7 was used to determine the
kinematicviscosity(
)withrespecttotemperaturebasedonpuretrioleinwhere iscSt, iscP,
isg/cm3and isK(ValeriandMeirelles,1997).
ln
1.5554
2182.0021
888879.2445
(8.7)
2
The surface area and particle diameter for the Celite R632 solgel were 1.49 m /g and 1 mm,
respectively(MeunierandLegge,2010).ForNovozym435thesurfaceareaandparticlediameterwere
89m2/gand0.6mm,respectively(Wiemannetal.,2009).
100
8.3.4
Packedbedefficiencycorrelation
Despite mass transfer effects, the solventfree packed bed reactor with Celite supported solgel
containinglipaseachievedhigherconversionsmorequicklythanthestirredtankbatchreactorthatwas
usedtodeveloptheinitialkineticequations.Basedontheeffectofglycerolforcontinuousenzymatic
transesterification described by Dossat et al. (1999), the presence of glycerol in the liquid phase is
assumed to cause a diffusion limitation preventing the organic substrate from accessing the enzyme.
Additionally, Dossat et al.(1999) found that increasing the amount of enzyme in the system caused a
delayinthepointatwhichtheenzymaticactivitybeginstodecrease.Toaccommodatethiseffect,an
efficiencycorrelation, ,wasdevelopedthatdependsonboththeReynoldsnumberforthereactor(
andtheconcentrationofglycerolwithrespecttotheweightofenzymeinthepackedbed(
)
)as
describedbyEquation8.8.
(8.8)
Fromthiscorrelation,astheReynoldsnumberincreasestheglycerollayerisremovedtherebyreducing
the resistance layer resulting in an increase in the observed enzymatic activity. Conversely, as the
glycerol concentration increases, so does the glycerol layer resulting in an increase in resistance
resulting in a decrease in the observed activity. The constants of the model (
, and
were obtained using an inverse parameter estimation approach. The experimental methyl oleate
concentrationsobtainedat1.4,5and20mL/min(29datapoints)wereusedtocalibratethemodelby
estimatingtheparametersoftheefficiencyequation(Equation8.8).Ageneticalgorithmcodeavailable
inthe MATLABsoftwarewasusedforparameter estimationbyminimizing thesumofsquarederror
(SSE)betweenexperimentalandmodelestimationofthemethyloleateconcentration.Themodelwas
validatedagainsttheexperimentaldataobtainedat12.5mL/minandasecondsetofdataobtainedat
20mL/min(30additionaldatapoints).
Table8.1:KineticconstantsdeterminedbyfittingtheresultsofthebatchexperimentstoEquation8.1fortheCelitesolgel
(CSG)andtheNovozym435(N435)immobilizedenzymepreparations.Ethanolysiskineticconstantsfromtheliterature
usingN435areincludedforcomparison(Chesterfieldetal.,2012).
Kineticcoefficient
(mol/Lmin)
(mol/L)
(mol/L)
(mol/L)
CSG
408.26
1.2x104
5.5x104
0.137
N435
1.022
0.132
830
4.2x105
101
Chesterfieldetal.(2012)
3.26
0.029
0.948
0.083
8.4.1
Batchkinetics
ThekineticconstantsdeterminedforCelitesupportedsolgelsandNovozym435(Table8.1)indicate
that a faster reaction was observed for Celite solgel (
inhibitionwasgreaterforNovozym435(
) and that alcohol
).TheseresultsshowthattheCelitesol
gel has better kinetic performance in comparison to the commercially immobilized Novozym 435.
However, the Celite solgels exhibit lower apparent affinities for both substrates in comparison
Novozym435(
and
).Thekineticmodelsdevelopedforboth
enzyme preparations provide good fit with respect the experimental data (SSECSG=0.219, R2CSG=0.888,
SSEN435=3.9x109,R2N435=0.895).
ArecentstudybyChesterfieldetal.(2012)consideredthereactionkineticsforNovozym435biodiesel
productionandfound
and
valuesbetweenthosefoundforCSGandN435inthisstudy(Table
8.1). A similar catalyst loading was used, however ethanol was used as the acyl acceptor, which is
knowntobelessinhibitorytolipase(HernndezMartnandOtero,2008;Salisetal.,2005).Thiseffect
is substantiated by the lower alcohol inhibition constant and higher apparent Michaelis constant
observedinthisstudyincomparisontotheobservationsofChesterfieldetal.(2012).
8.4.2
Experimentalandmodellingresults
Todeterminetheconstantsforthecorrelationcoefficient(Equation8.8),Equations8.18.8weresolved
at a varietyof flow rates to achieve the best fit for all conditions tested. Table 8.2 provides the data
used for the immobilized enzyme including the catalyst density, surface area, particle size, porosity,
efficiency correlation constants, diffusivity coefficient and viscosity while Table 8.3 provides the
additionaldatanecessaryforthepackedbedreactormodelincludingtheReynoldsnumber,Sherwood
number,andmasstransfercoefficientasdescribedinSection8.3.2.
The model developed showed very good agreement with the experimental data for all the flow rates
examined (Figure 8.2). Increasing the flow rate from 1.4 mL/min to 5 mL/min and from 5 mL/min to
12.5mL/mininfluencedthemethyloleateconcentrationprofile,butafurtherincreaseintheflowrate
from12.5mL/minto20mL/minhadaminimaleffectontheproductconcentrationprofile.Asimilar
effectwasobservedbyHamaetal.(2011)whoobservedthattheglycerolremovalefficiencyincreased
withincreasingflowrateuntiltheoptimalflowforglycerolremovalwasachieved.Atthetwohighest
flow rates (12.5 mL/min and 20 mL/min), the maximum conversion (90%) is reached within
102
approximately 10 minutes of contact time between the substrates and the enzyme. The inability to
achieve100%conversionisattributedtothehighglycerolconcentrationwhichresultsinahydrophilic
layeraroundtheenzymeandsupportpreventingthetriglyceridefromreachingtheenzymesactivesite
(Dossatetal.,1999).
Table8.2:Parametersusedtomodeltheexperimentalpackedbedreactorresultsforeachoftheimmobilizedbiocatalysts.
Parameter
(kg/m3)
(m2/kg)
(m2/m3)
(m)
2
(m
/s)
(Pas)
CSG
562
1490
8.37x105
0.001
0.42
23
6
14
1x104
4
N435
449
82000
3.69x107
0.0006
0.41
8
4
14
5x102
4
2.11x1010
0.0355
1000000
[Methyloleate](M)
800000
600000
400000
200000
0
0
10
15
Contacttime(min)
20
25
CSG 1.4mL/min
CSG 5mL/min
CSG 12.5mL/min
CSG 20mL/min
Model 1.4mL/min
Model 5mL/min
Model 12.5mL/min
Model 20mL/min
Figure8.2:Packedbedresultswiththeexperimental(symbols)andmodel(lines)methyloleateconcentrationsasafunction
ofcontacttimeinthereactorwithCelitesupportedsolgelsat4flowrates(1.4mL/min,5mL/min,12.5mL/minand20
mL/min).Theconfidenceintervalsshownareestimatedfromapooledvarianceof16samplesintriplicate.
103
Table8.3:Summaryofsomeoftheparameterscalculatedfromthemodellingofthepackedbedreactorexperiments.
Enzyme
(mL/min)
(cm/min)
CSG
1.4
0.0075
38.77
0.477
CSG
5
0.0267
55.37
0.681
CSG
12.5
0.0669
71.56
0.881
CSG
20
0.1070
81.62
1.004
N435
5
0.0160
48.74
1.002
N435
20
0.0640
71.86
1.477
Table8.4:Comparisonofsolventfreebiodieselpackedbedreactorperformanceusingimmobilizedlipase.
20
20
Reaction
time
(min)
60
60
56
0.1
560
76
290
1020
255
75
Solgel
67
155
30
180
30
N435
10.68
110
16.6
60
25
NS8801
17
120
0.8
150
16
CSG
N435
Reactor
volume
(cm3)
13.35
13.35
Substrate
volume
(mL)
52
52
N435
4.15
N435
Reference
Catalyst
Thiswork
Thiswork
Changetal.
(2009)
Hamaetal.
(2011b)
Hsuetal.
(2004)
Ognjanovicetal.
(2009)
Xuetal.
(2012)
Flowrate
(mL/min)
Conversion
(%)
84
63
Theobservedeffectofflowrateisconsistentwithwhatisdescribedintheliterature(Changetal.,2009,
Hama et al., 2011b; Lee et al., 2010). Hama et al. (2011b) observed an optimal flow rate of
approximately 9 mL/min based on the glycerol removal efficiency and the fatty acid methyl ester
contentoftheproductstreamforasimilarlysizedreactor(15.7mminternaldiameterincomparisonto
the 10 mm internal diameter for this study). Table 8.4 presents a review of the literature for
immobilizedenzymepackedbedbiodieselreactorsinsolventfreereactionmedia.Thereactiontimeis
presentedinthistable,differentfromthecontacttimeinthefiguresinthatthereactiontimeisthetime
requiredfortheliquidtoflowthroughthereactor(substratevolume/flowratexnumberofpasses)as
opposedtothecontacttimebetweentheliquidandthecatalystwhichaccountsforthevoidfractionin
thereactorandthereactordimensions.Thepackedbedreactorresultsachievedinthisstudysurpass
theconversionsreportedintheliterature.
104
8.4.3
ComparisonofCelitesupportedsolgeltoNovozym435
Asabasisforcomparison,acommerciallyavailableimmobilizedlipase,Novozym435,wastestedinthe
same packed bed reactor at flow rates of 5 mL/min and 20 mL/min. The same model (with different
efficiencycorrelationconstants,Table8.2)wasusedtopredictthemethyloleateconcentrationprofile
andwasfoundtobeingoodagreementwiththeexperimentaldata(Figure8.3).Therewasasignificant
improvementofthepackedbedreactorperformanceusingCSGincomparisontoN435(Figures8.2and
8.3).
MuchhighermethyloleateconcentrationswereobservedforthepackedbedreactorwithCSGthanfor
N435. After 5 min of contact time at 5 mL/min the CSG had a conversion of 58% while N435 had a
conversionof42%,andat20mL/mintheCSGhadaconversionof83%whileN435had59%conversion.
SincethemaximumreactionrateishigherandthemethanolinhibitioneffectislowerforCSGthanfor
N435,animprovementintheperformanceofthebatchstirredtankandpackedbedreactorswithCSGis
expected.AninternaldiffusionlimitationformanybiocatalystsincludingN435hasbeenreportedinthe
literature(Almeidaetal.,1998;JungandBauer,1992;Xiaoetal.,2012;XuandChuang,1997),whichisa
likelycauseofthedecreaseinreactorperformanceusingN435.
1000000
[Methyloleate](M)
800000
600000
400000
200000
0
0
N435 5mL/min
4
Contacttime(min)
N435 20mL/min
Model 5mL/min
8
Model 20mL/min
Figure8.3:Packedbedresultswiththeexperimental(symbols)andmodelling(lines)methyloleateconcentrationsasa
functionofcontacttimeinthereactorforNovozym435atflowratesof5mL/minand20mL/min.Theconfidenceintervals
shownareestimatedfromapooledvarianceof16samplesintriplicate.
105

Figure8.4:Contourmapformodelpredictionsformethyloleateconcentrationwithrespecttotheflowrateandcontact
timebasedonkineticsandmasstransfer.
8.4.4
Modelpredictions
Thepredictionsusingthemodeldevelopedsupporttheexperimentalobservationsthatincreasingthe
flowrateandthecontacttimecauseanincreaseinthemethyloleateconcentration,butbothvariables
have thresholds for which further increases in the flow rate do not substantially increase the final
achievableconcentrationofmethyloleate(Figure8.4).Forexample,acontacttimeof5minataflow
rateof0.5mL/minresultsinamethyloleateconcentrationof0.35M,aflowrateof16mL/minresults
in a methyl oleate concentration of 0.79 M, and a flow rate of 50 mL/min results in a methyl oleate
concentration of 0.81 M. A similar effect is observed with contact time. This is consistent with the
literaturewhichindicatesthatahydrophilicglycerollayerhinderstheenzymaticactivitybylimitingthe
access of the triglyceride to the enzyme increasing the flow rate will remove the glycerol, but
increasingthecontacttime(conversion)willincreasetheglycerolconcentration(BlafiBaketal.,2002;
Dossatetal.,1999;Hongetal.,2011).
8.4.5
Catalystreusability
The reusability of the Celite solgel was considered over a five day period where each day new
substratesweresuppliedtotheexistingcatalystbedinthereactor.Fromtheexperimentaldata(Figure
8.5),nosignificantactivitylosswasobservedoverthefivedaytimespansupportingtheclaimthatthe
Celitesolgelisapromisingimmobilizedlipasesuitablefortransesterificationinapackedbedreactor.
106
MethylOleateConcentration (M)
1000000
750000
500000
250000
DAY1
0
0
50
DAY 2DAY3DAY4DAY5
100
150
200
250
300
Time(min)
350
400
450
500
Figure8.5:Reusabilityrunsinthepackedbedreactorover5daysatT=40C,substrateratio1:1,andflowrate=5mL/min.
Anewsubstratemixture(trioleinandmethanol)wassuppliedtotheexistingCSGbedatthebeginningofeachday.
8.5 Conclusions
A packed bed reactor containing Celite supported lipase solgel was developed and its performance
was analyzed using a combination of reaction kinetics and mass transfer with the addition of a novel
efficiency correlation to account for the effect of glycerol on the observed enzymatic activity. The
reaction kinetics for Celite supported solgels and Novozym 435 were determined using an initial
reactionratemodelwithalcoholinhibition.TheCSGhadamuchfastermaximumreactionvelocityand
alcohol was found to be less inhibitory. The reaction kinetics from the stirred batch reactor in
combination with mass transfer correlations were insufficient to describe the performance of the
packedbedreactorwhichisattributedtothepresenceofaglycerollayerwhichcreatesanadditional
mass transfer resistance. An efficiency correlation dependent on the Reynolds number and the
concentration of glycerol with respect to the amount of enzyme in the packedbed reactor was
developed and used in the model. The efficiency correlation describes the relationship between the
reductionoftheglycerollayerwithincreasingflowrateandtheincreasingglycerollayerwithincreasing
glycerol concentration. Combining the reaction kinetics, mass transfer, and efficiency correlation, the
resultingmodelexhibitedgoodagreementwiththeexperimentaldataoverarangeofflowratesfrom
107
1.4mL/minto20mL/min.Boththeexperimentalresultsandthemodelpredictionsdemonstratethe
strong effect of flow rate on the methyl oleate concentration at low flow rates, but as the flow rate
increases beyond 12.5 mL/min the activity enhancing effect is minimized. The results for the Celite
supported solgel system were compared to a commercially available immobilized lipase, N435, and
showedthatCSGsystemresultedinsignificantlyhigherfinalmethyloleatepercentconversions(83%at
20 mL/min) in comparison to N435 (59% at 20 mL/min) supporting the inherent advantages of this
enzymebased system for the production of fatty acid methyl esters. Additionally, over five days no
enzymaticactivitylosswasobservedfortheCSGfurtherindicatingitspotentialforuseinapackedbed
reactorconfiguration.
ofaPostgraduateScholarshiptoSMMandaDiscoveryGranttoRLL.WethankAmanoEnzymeUSAfor
supplying samples of the lipase formulation, World Minerals for the samples of the Celite support
materialsandNovozymesNorthAmericaInc.forthesamplesoftheNovozym435.
8.7 Nomenclature
Alcoholconcentration(mol/m3)
Concentrationofsubstrate (mol/m3)
Concentrationofsubstrate atthecatalystsurface(mol/m3)
CSG
Diffusionofmethanolintriolein(m2/s)
Catalystparticlediameter(m)
Reactordiameter(m)
Glycerolconcentration(mol/m3)
Masstransfercoefficient(m/min)
108
Alcoholinhibitionconstant(mol/m3)
ApparentMichaelisconstantforalcohol(mol/m3)
ApparentMichaelisconstantfortriglyceride(mol/m3)
Molecularmassofthetriolein(g/mol)
N435 Novozym435
Flowrate(m3/min)
R2
Reactionratewithrespectsubstrate (mol/m3min)
Reynoldsnumberforthereactor
10
10
Reynoldsnumberfortheparticle
Externalsurfaceareaofthecatalyst(m2/m3)
Schmidtnumber(dimensionless)
Sherwoodnumber(dimensionless)
SSE
Catalystsurfacearea(m2/kg)
Time(min)
Temperature(K)
60
Sumofsquarederrors
60
Triglycerideconcentration(mol/m3)
Linearvelocity(m/min)
Molarvolumeofthemethanol(m3/mol)
Apparentmaximuminitialreactionvelocity(mol/m3min)
Proteinweightinthereactor(kg)
109
(dimensionless)
(dimensionless)
Positionalongthereactorheight(m)
Efficiencycorrelation(dimensionless)
Constantsfortheefficiencycorrelation( =1,2,3,4or5)
Voidfractioninthereactor(dimensionless)
Oildynamicviscosity(Pas)
Kinematicviscosity(cSt)
Density(g/cm3)
Bulkcatalystdensity(kg/m3)
Dissociationfactorfortheoil(dimensionless)
110
Chapter 9
ConclusionsandRecommendations
9.1
Mostsignificantcontributions
The research conducted for this thesis focused on the development and characterization of a novel
immobilized enzyme, Celite supported lipase solgel, and the elucidation of its potential for the
productionofbiodieselinanenzymaticpackedbedreactor.Aseriesoflipasesolgelsupportmaterials
were evaluated based on their fatty acid methyl ester production capabilities in a stirred tank batch
reactor;thesupportmaterialwiththemostpotential,diatomaceousearthCeliteR632,waschosenand
further characterized to predict any challenges that may be associated with enzymatic
transesterification;thereactionkineticsofthesupportedsolgelforfattyacidmethylesterproduction
was ascertained; and the performance of the immobilized enzyme was assessed using a packed bed
reactor with kinetic and mass transfer modelling. The following sections describe the specific
conclusionsdrawnfromeachofthechapterspresentedinthisthesis.
9.1.1
ObjectiveAExperimentalBackground
Chapter 3 presents the preliminary results collected to determine the feasibility of formulating and
evaluating lipase solgels for biodiesel production. Based on the research presented, lipase solgels
werefoundtobeveryactiveenzymaticallywithhighthermalstabilityandgoodpotentialforfattyacid
methyl ester production when high hydrophobicity solgels (80% PTMS) are used. Additionally, using
thecommerciallyimmobilizedlipase,Novozym435,thepresenceofmethanolinthereactionmixture
wasfoundtohaveastrongeffectontheachievableconversion,theeffectsofthepresenceofahexane
solventandtemperaturewereonlyevidentathightemperatures(above60C),andwaterhadaminimal
effectonconversion.Thisinformationwasusedasabasisforthedevelopingasupportedlipasesolgel
fortheproductionoffattyacidmethylesters.
9.1.2
ObjectiveBStudyofSupportMaterialsforSolgelImmobilizedLipase
InChapter4,aseriesofsupportmaterialswereexaminedandusedtoformulatesupportedlipasesol
gelswhichwereevaluatedbasedontheirabilitytoretainsolgel,loadprotein,andremaincatalytically
activefortheproductionofmethyloleate.Thegoalinselectingasupportmaterialwasitspotentialfor
useasabiocatalystpreparationpracticalforuseinapackedbedreactorforbiodieselproduction.Six
111
supportmaterialswereconsidered:silicagel,threetypesofdiatomaceousearth(CeliteR633,Celite
R632, and Celite R647), an anion exchange resin, and Quartzel felt. Each supported solgel had
unique advantages for this application, but Celite R632 was chosen as the support material with the
most potential for biodiesel production due to its observed high transesterification conversion ability,
activityandreactionrate,aswellasitshighthermalstabilityandsolgeladhesionlevelsincomparison
totheothersupportsconsidered.
9.1.3
Objective C Evaluation of Diatomaceous Earth as a Support for Solgel Immobilized

LipaseforTransesterification
Chapter5isastudyofthediatomaceoussupportedsolgelsinmoredepthbylookingattheirsurface
morphology, physical properties, and enzymatic properties. The main differences between the three
Celitematerialsusediswithrespecttotheirparticlesize,porediameter,andsurfacearea:R633hasa
smallparticlesizeandlargeporediameter,R632hasalargeparticlesizeandlargeporediameter,and
R647haslargeparticlesize,smallporediameterandverylargesurfacearea.Thesolgeladhesionand
proteinloadingwerecomparableforallthreetypesofCeliteconsidered.CeliteR632exhibitedgood
initial activity and methyl oleate conversion in comparison to the other supported solgels and the
unsupported solgel validating the potential of Celite R632 supported lipase solgels for biodiesel
production.
9.1.4
Objective D Evaluation of Diatomaceous Earth Supported Solgels as a Medium for

EnzymaticTransesterificationtoProduceBiodiesel
In Chapter 6, the Celite R632 supported lipase solgels were further investigated for biodiesel
production potential based on their protein content, methyl oleate conversion, enzymatic activity,
stability, and glycerolwater adsorption. Unsupported and Celite supported lipase solgels with
different protein loading levels were compared. Based on the results obtained, the Celite R632
supportedsolgelwiththehighestproteincontentachievedthebestconversionoverthe6hreaction
period(80%),adryingsteppriortoreactionwasfoundtoimprovetheconversion,andtheadsorptionof
glycerol was only found to be significant at high concentrations of glycerol (75% by volume). The
unsupportedandCelitesupportedlipasesolgelswerefoundtobecatalyticallystableuponstorageat
4Covera1.5yrperiod. TheseresultsdemonstratethatCeliteR632supportedlipase solgelsshow
promiseasanovelbiocatalystforbiodieselproduction.
112
9.1.5
ObjectiveEKineticModellingoftheProductionofMethylOleatebyCeliteSupported
LipaseSolgels
Chapter7presentsakineticstudybasedonapingpongbibimechanismformethyloleateproduction
using Celite R632 supported lipase solgels. A kinetic model was developed to include inhibition by
methanol and glycerol, the effect of temperature, the presence of products, and the reverse reaction
andwasshowntobeconsistentwiththeexperimentaldata.Anoptimalmethanolconcentrationrange
(1.32.0M)thatwasincreasedwithincreasingreactiontemperaturewasidentified.Glycerolinhibition
was found to have a more significant effect on the kinetics than methanol inhibition, indicating the
importanceofconsideringthereactionkineticsbeyondtheinitialconditions.Themodeldevelopedin
this study can be considered valuable for understanding and identifying the combined effects of
methanol inhibition, glycerol inhibition, and temperature on the production of methyl oleate by
enzymatictransesterification.
9.1.6
Objective F Kinetic and Mass Transfer Modelling of Methyl Oleate Production in an

ImmobilizedLipasePackedBedReactor
Chapter 8 provides the results for a modeling study based on kinetics and mass transfer for a packed
bedreactorwithCelitesupportedlipasesolgelsandNovozym435.Comparingthemethyloleate
production capabilities of the Celite supported solgel with Novozym 435 under both batch and
packed bed reactor conditions, the Celite supported solgel had superior performance indicating the
potential of this novel biocatalyst for transesterification. To model the reactor performance, reaction
kinetics and mass transfer equations were used in combination with a novel efficiency correlation to
accountfortheeffectofglycerol.Themodeldevelopedwasconsistentwiththeexperimentaldataover
arangeofflowratesfrom1.4mL/minto20mL/min.Thepresenceofglycerolhadadominanteffecton
themethyloleateconcentrationprofileandincreasingtheflowrate,presumablyreducingtheglycerol
layersurroundingthebiocatalyst,resultedinimprovedapparentenzymaticactivity.
113
9.2
RecommendationsforFutureWork
Basedontheresearchconductedinthisstudy,severalrecommendationsforfurtherworkareprovided:
Fundamentals: Quantification of glycerol and the reaction intermediates (diglycerides and

monoglycerides) in the product stream would be useful for gaining a better understanding of
the reaction mechanism. This could be accomplished using HPLC with a refractive index
detectororaGCwithFIDdetectionandatemperatureprogrammableinjector.
Processoptimization:Thereareseveralvariablesthatshouldbestudiedtocompletetheprocess
optimization for the Celite supported packed bed reactor such as the lipase source, solgel
composition,acylacceptor,watercontent,andtemperature.Eachofthesevariablesiscrucial
fortheperformanceandstabilityofthereactor.Ideally,thesevariablesshouldbeconsidered
simultaneouslytoevaluateanyinteractioneffects.
Applicability for biodiesel production: To determine the industrial applicability of this Celite
supportedpackedbedreactorforbiodieselproduction,thereactorshouldbeconsideredwith
alternativefeedstockssuchaswaste,vegetable,andnonedibleoils.
Operational stability: The long term conversion capabilities and mechanical stability of the
immobilized enzyme should be studied as well as the catalyst lifetime. This information is
essential in evaluating the practicality of large scale biodiesel production using the proposed
enzymaticreactor.
Glycerol separation: Given the effects of glycerol revealed in Chapters 6, 7, and 8, the
continuous removal of glycerol from the packed bed operation could significantly benefit the
achievable conversion. Although there are a variety of options for glycerol removal,
implementationofapackedbedreactorwithmembranebasedseparationforglycerolmaybe
ofvalueforindustrialapplications.
114
References
AbbaszaadehA,GhobadianB,OmidkhahMR,NajafiG.2012.Currentbiodieselproductiontechnologies:
Acomparativereview.EnergyConversManage63:138148.
AkohCC,ChangSW,LeeGC,ShawJF.2007.Enzymaticapproachtobiodieselproduction.JAgricFood
Chem55:89959005.
AlmeidaMC,RuivoR,MaiaC,FreireL,CorraDeSampaioT,BarreirosS.1998.Novozym435activityin
compressedgases.Wateractivityandtemperatureeffects.EnzymeMicrobTechnol22:494499.
AlZuhairS.2007.Productionofbiodiesel:Possibilitiesandchallenges.BiofuelBioprodBior1:5766.
AlZuhair S. 2005. Production of biodiesel by lipasecatalyzed transesterification of vegetable oils: A
kineticsstudy.BiotechnolProgr21:14421448.
AlZuhairS,AlmenhaliA,HamadI,AlshehhiM,AlsuwaidiN,MohamedS.2011.Enzymaticproductionof
biodieselfromused/wastevegetableoils:Designofapilotplant.RenewableEnergy36:26052614.
AlZuhair S, Dowaidar A, Kamal H. 2009. Dynamic modeling of biodiesel production from simulated
wastecookingoilusingimmobilizedlipase.BiochemEngJ44:256262.
AlZuhairS,LingFW,JunLS.2007.Proposedkineticmechanismoftheproductionofbiodieselfrompalm
oilusinglipase.ProcessBiochem42:951960.
Atadashi IM, Aroua MK, Abdul Aziz AR, Sulaiman NMN. 2012. The effects of catalysts in biodiesel
production:Areview.JIndEngChemInpress.
Aucoin MG, Erhardt FA, Legge RL. 2004. Hyperactivation of Rhizomucor miehei lipase by hydrophobic
xerogels.BiotechnolBioeng85:647655.
AvnirD,BraunS,LevO,OttolenghiM.1994.Enzymesandotherproteinsentrappedinsolgelmaterials.
ChemMater6:16051614.
AzcarL,CiudadG,HeipieperHJ,MuozR,NaviaR.2010.Improvingfattyacidmethylesterproduction
yieldinalipasecatalyzedprocessusingwastefryingoilsasfeedstock.JBiosciBioeng109:609614.
115
Bajaj A, Lohan P, Jha PN, Mehrotra R. 2010. Biodiesel production through lipase catalyzed
transesterification:Anoverview.JMolCatalB:Enzym62:914.
BalasubramaniamB,SudalaiyadumPerumalA,JayaramanJ,ManiJ,RamanujamP.2012.Comparative
analysis for the production of fatty acid alkyl esterase using whole cell biocatalyst and purified
enzymefromRhizopusoryzaeonwastecookingoil(sunfloweroil).WasteManage32:15391547.
Bastida A, Sabuquillo P, Armisen P, FernndezLafuente R, Huguet J, Guisn JM. 1998. A single step
purification, immobilization, and hyperactivation of lipases via interfacial adsorption on strongly
hydrophobicsupports.BiotechnolBioeng58:486493.
BlafiBakK,KovcsF,GubiczaL,HancskJ.2002.Enzymaticbiodieselproductionfromsunfloweroil
byCandidaantarcticalipaseinasolventfreesystem.BiocatalBiotransform20:437439.
Bendikiene V, Kiriliauskaite V, Juodka B. 2011. Production of environmentally friendly biodiesel by
enzymaticoiltransesterification.JEnvironEngLandsc19:123129.
BrnyikT,KuncovG,PcaJ.2000.Theuseofsilicagelpreparedbysolgelmethodandpolyurethane
foam as microbial carriers in the continuous degradation of phenol. Appl Microbiol Biotechnol
54:168172.
ChangC,ChenJH,Chang CmJ,WuTT,ShiehCJ.2009.Optimizationoflipasecatalyzedbiodieselby
isopropanolysis in a continuous packedbed reactor using response surface methodology. New
Biotechnol26:187192.
Chang HM, Liao HF, Lee CC, Shieh CJ. 2005. Optimized synthesis of lipasecatalyzed biodiesel by
Novozym435.JChemTechnolBiotechnol80:307312.
Cheirsilp B, HKittikun A, Limkatanyu S. 2008. Impact of transesterification mechanisms on the kinetic
modelingofbiodieselproductionbyimmobilizedlipase.BiochemEngJ42:261269.
ChenHC,JuHY,WuTT,LiuYC,LeeCC,ChangC,ChungYL,ShiehCJ.2011.Continuousproduction
of lipasecatalyzed biodiesel in a packedbed reactor: Optimization and enzyme reuse study. J
BiomedBiotechnol2011ArticleID9507256pagesdoi:10.1155/2011/950725
ChenJW,WuWT.2003.RegenerationofimmobilizedCandidaantarcticalipasefortransesterification.
JBiosciBioeng95:466469.
116
ChenY,XiaoB,ChangJ,FuY,LvP,WangX.2009.Synthesisofbiodieselfromwastecookingoilusing
immobilizedlipaseinfixedbedreactor.EnergyConversManage50:668673.
Chesterfield DM, Rogers PL, AlZaini EO, Adesina AA. 2012. Production of biodiesel via ethanolysis of
wastecookingoilusingimmobilisedlipase.ChemEngJ207208:701710.
ChewYH,LeeCT,SarmidiMR,AzizRA,RazaliF.2008.Externalmasstransfermodelforthehydrolysisof
palmoleinusingimmobilizedlipase.FoodBioprodProcess86:276282.
Chouhan APS, Sarma AK. 2011. Modern heterogeneous catalysts for biodiesel production: A
comprehensivereview.RenewableSustainableEnergyRev15:43784399.
Cleland WW. 1963a. The kinetics of enzymecatalyzed reactions with two or more substrates or
products.I.Nomenclatureandrateequations.BiochimBiophysActa67:104137.
Cleland WW. 1963b. The kinetics of enzymecatalyzed reactions with two or more substrates or
products.II.Inhibition:Nomenclatureandtheory.BiochimBiophysActa67:173187.
Cleland WW. 1963c. The kinetics of enzymecatalyzed reactions with two or more substrates or
products. III. Prediction of initial velocity and inhibition patterns by inspection. Biochim Biophys
Acta67:188196.
Clifford JS, Legge RL. 2005. Use of water to evaluate hydrophobicity of organicallymodified xerogel
enzymesupports.BiotechnolBioeng92:231237.
Dossat V, Combes D, Marty A. 2002. Lipasecatalysed transesterification of high oleic sunflower oil.
EnzymeMicrobTechnol30:9094.
DossatV,CombesD,MartyA.1999.Continuousenzymatictransesterificationofhigholeicsunfloweroil
inapackedbedreactor:Influenceoftheglycerolproduction.EnzymeMicrobTechnol25:194200.
DuW,XuY,LiuD,ZengJ.2004.Comparativestudyonlipasecatalyzedtransformationofsoybeanoilfor
biodieselproductionwithdifferentacylacceptors.JMolCatalB:Enzym30:125129.
Dunn IJ. 2003. Biological reaction engineering: dynamic modelling fundamentals with simulation
examples.Wienheim,Germany:VCHp272.
117
DwivediPN,UpadhyaySN.1977.Particlefluidmasstransferinfixedandfluidizedbeds.IndEngChem
ProcDD16:157165.
FanX.2012.Enzymaticbiodieselproductionthewayofthefuture.LipidTechnol24:3132.
Fan X, Niehus X, Sandoval G. 2012. Lipases as biocatalyst for biodiesel production. Methods Mol Biol
861:471483.
FjerbaekL,ChristensenKV,NorddahlB.2009.Areviewofthecurrentstateofbiodieselproductionusing
enzymatictransesterification.BiotechnolBioeng102:12981315.
FoglerHS.Elementsofchemicalreactionengineering3rded.UpperSaddleRiver,N.J.:PrenticeHallPTR
c1999.
FukudaH,KondoA,NodaH.2001.Biodieselfuelproductionbytransesterificationofoils.JBiosciBioeng
92:405416.
Furukawa SY, Ono T, Ijima H, Kawakami K. 2002a. Activation of protease by solgel entrapment into
organicallymodifiedhybridsilicates.BiotechnolLett24:1316.
FurukawaSY,OnoT,IjimaH,KawakamiK.2002b.Effectofimprintingsolgelimmobilizedlipasewith
chiral template substrates in esterification of (R)(+) and (S)()glycidol. J Mol Catal B: Enzym
17:2328.
Furukawa SY, Ono T, Ijima H, Kawakami K. 2000. Catalytic capabilities of lipase immobilized into
organicallymodifiedsilicatesbysolgelmethod.KagakuKogakuRonbunshu26:885886.
FurukawaSYA,KawakamiK.1998.CharacterizationofCandidarugosalipaseentrappedintoorganically
modifiedsilicatesinesterificationofmentholwithbutyricacid.JFermentBioeng85:240242.
GagnonMD,VasudevanPT.2011.Effectsofsolventandenzymesourceontransesterificationactivity.
EnergyFuels25:46694674.
Galarneau A, Mureseanu M, Atger S, Renard G, Fajula F. 2006. Immobilization of lipase on silicas.
Relevanceoftexturalandinterfacialpropertiesonactivityandselectivity.NewJChem30:562571.
118
Ganesan D, Rajendran A, Thangavelu V. 2009. An overview on the recent advances in the

transesterification of vegetable oils for biodiesel production using chemical and biocatalysts. Rev
EnvironSciBiotechnol8:367394.
Gill I, Ballesteros A. 2000. Bioencapsulation within synthetic polymers (Part 1): Solgel encapsulated
biologicals.TrendsBiotechnol18:282296.
Gog A, Roman M, Toa M, Paizs C, Irimie FD. 2012. Biodiesel production using enzymatic
transesterificationCurrentstateandperspectives.RenewableEnergy39:1016.
GuptaMN.ThermostabilityofEnzymes.Berlin:SpringerVerlagc1993
Halim SFA, Kamaruddin AH, Fernando WJN. 2009. Continuous biosynthesis of biodiesel from waste
cookingpalmoilinapackedbedreactor:Optimizationusingresponsesurfacemethodology(RSM)
andmasstransferstudies.BioresourTechnol100:710716.
Hama S, Tamalampudi S, Yoshida A, Tamadani N, Kuratani N, Noda H, Fukuda H, Kondo A. 2011a.
Enzymatic packedbed reactor integrated with glycerolseparating system for solventfree
productionofbiodieselfuel.BiochemEngJ55:6671.
Hama S, Tamalampudi S, Yoshida A, Tamadani N, Kuratani N, Noda H, Fukuda H, Kondo A. 2011b.
Processengineeringandoptimizationofglycerolseparationinapackedbedreactorforenzymatic
biodieselproduction.BioresourTechnol102:1041910424.
HartmeierW.Immobilizedbiocatalysts:anintroduction.Berlin:SpringerVerlagc1988.
HasanF,ShahAA,HameedA.2006.Industrialapplicationsofmicrobiallipases.EnzymeMicrobTechnol
39:235251.
HernndezMartn E, Otero C. 2008. Different enzyme requirements for the synthesis of biodiesel:
Novozym435andLipozymeTLIM.BioresourTechnol99:277286.
HongWP,ParkJY,MinK,KoMJ,ParkK,YooYJ.2011.Kineticsofglyceroleffectonbiodieselproduction
foroptimalfeedingofmethanol.KoreanJChemEng28:19081912.
Hsu AF, Foglia TA, Jones K, Shen S. 2001a. Phyllosilicate solgel immobilized enzymes: Their use as
packedbedcolumnbioreactors.ACSSympSer776:155164.
119
HsuAF,JonesK,MarmerWN,FogliaTA.2001b.Productionofalkylestersfromtallowandgreaseusing
lipaseimmobilizedinaphyllosilicatesolgel.JAmOilChemSoc78:585588.
HsuAF,JonesKC,FogliaTA,MarmerWN.2004.Continuousproductionofethylestersofgreaseusing
animmobilizedlipase.JAmOilChemSoc81:749752.
Hsu AF, Jones KC, Foglia TA, Marmer WN. 2003. Optimization of alkyl ester production from grease
usingaphyllosilicatesolgelimmobilizedlipase.BiotechnolLett25:17131716.
HuangY,ZhengH,YanY.2010.Optimizationoflipasecatalyzedtransesterificationoflardforbiodiesel
productionusingresponsesurfacemethodology.ApplBiochemBiotechnol160:504515.
IsoM,ChenB,EguchiM,KudoT,ShresthaS.2001.Productionofbiodieselfuelfromtriglyceridesand
alcoholusingimmobilizedlipase.JMolCatalB:Enzym16:5358.
JaegerKE,EggertT.2002.Lipasesforbiotechnology.CurrOpinBiotechnol13:390397.
Jegannathan KR, EngSeng C, Ravindra P. 2011. Economic assessment of biodiesel production:
Comparisonofalkaliandbiocatalystprocesses.RenewableSustainableEnergyRev15:745751.
Jin W, Brennan JD. 2002. Properties and applications of proteins encapsulated within solgel derived
materials.AnalChimicaActa461:136.
Jung HJ, Bauer W. 1992. Determination of process parameters and modelling of lipasecatalyzed
transesterificationinafixedbedreactor.ChemEngTechnol15:341348.
KaiedaM,SamukawaT,KondoA,FukudaH.2001.Effectofmethanolandwatercontentsonproduction
ofbiodieselfuelfromplantoilcatalyzedbyvariouslipasesinasolventfreesystem.JBiosciBioeng
91:1215.
Kaieda M, Samukawa T, Matsumoto T, Ban K, Kondo A, Shimada Y, Noda H, Nomoto F, Ohtsuka K,
IzumotoE,FukudaH.1999.BiodieselfuelproductionfromplantoilcatalyzedbyRhizopusoryzae
lipaseinawatercontainingsystemwithoutanorganicsolvent.JBiosciBioeng88:627631.
Kargi F, Shuler ML. 2001. Bioprocess Engineering: Basic Concepts. United States: Prentice Hall PTR. p.
576.
120
Kawakami K. 1996. Enhancement of thermostability of lipase by the solgel entrapment into methyl
substitutedorganicsilicatesformedondiatomaceousearth.BiotechnolTech10:491494.
KawakamiK,MatsuiY,OnoT,IjimaH.2003.Enhancementofproteaseactivityintransesterificationof
glycidolwithvinylnbutyratebyentrapmentintoalkylsubstitutedsilicatesandpretreatmentwith
asubstrate.BiocatalBiotransform21:4952.
Kawakami K, Yoshida S. 1996. Thermal stabilization of lipase by solgel entrapment in organically
modifiedsilicatesformedonKieselguhr.JFermentBioeng82:239245.
KseO,TterM,AksoyHA.2002.ImmobilizedCandidaantarcticalipasecatalyzedalcoholysisofcotton
seedoilinasolventfreemedium.BioresourTechnol83:125129.
Koszelewski D, Muller N, Schrittwieser JH, Faber K, Kroutil W. 2010. Immobilization of omega
transaminasesbyencapsulationinasolgel/celitematrix.JMolCatalB:Enzym63:3944.
Krebs WG, Gerstein M. 2000. The morph server: A standardized system for analyzing and visualizing
macromolecularmotionsinadatabaseframework.NucleicAcidsRes28:16651675.
Kumari V, Shah S, Gupta MN. 2007. Preparation of biodiesel by lipasecatalyzed transesterification of
highfreefattyacidcontainingoilfromMadhucaindica.EnergyFuels21:368372.
LaiCC,ZullaikahS,ValiSR,JuYH.2005.Lipasecatalyzedproductionofbiodieselfromricebranoil.J
ChemTechnolBiotechnol80:331337.
LeeJH,KimSB,ParkC,TaeB,HanSO,KimSW.2010.Developmentofbatchandcontinuousprocesses
onbiodieselproductioninapackedbedreactorbyamixtureofimmobilizedCandidarugosaand
Rhizopusoryzaelipases.ApplBiochemBiotechnol161:365371.
Lee M, Lee J, Lee D, Cho J, Kim S, Park C. 2011. Improvement of enzymatic biodiesel production by
controlledsubstratefeedingusingsilicagelinsolventfreesystem.EnzymeMicrobTechnol49:402
406.
Li FY, Xing YJ, Ding X. 2007. Immobilization of papain on cotton fabric by solgel method. Enzyme
MicrobTechnol40:16921697.
Livage J, Coradin T, Roux C. 2001. Encapsulation of biomolecules in silica gels. J Phys Condens Matter
13:R673R691.
121
MacarioA,GiordanoG,MolinerM,CormaA.2009.Increasingstabilityandproductivityoflipaseenzyme
byencapsulationinaporousorganicinorganicsystem.MicroporousMesoporousMater118:334
340.
Marchetti JM, Miguel VU, Errazu AF. 2007. Possible methods for biodiesel production. Renewable
SustainableEnergyRev11:13001311.
MazzieriVA,VeraCR,YoriJC.2008.Adsorptivepropertiesofsilicagelforbiodieselrefining.EnergyFuels
22:42814284.
McDuffieNG.Bioreactordesignfundamentals. Massachusetts:ButterworthHeinemannc1991.
MeherLC,SagarDV,NaikSN.2006.TechnicalaspectsofbiodieselproductionbytransesterificationA
review.RenewableSustainableEnergyRev10:248268.
MessingRA,ed.Immobilizedenzymesforindustrialreactors.NewYork:AcademicPress.1975.
MeunierSM,LeggeRL.2012.Evaluationofdiatomaceousearthsupportedlipasesolgelsasamedium
forenzymatictransesterificationofbiodiesel.JMolCatalB:Enzym77:9297.
Meunier SM, Legge RL. 2010. Evaluation of diatomaceous earth as a support for solgel immobilized
lipasefortransesterification.JMolCatalB:Enzym62:5357.
Moreira ABR, Perez VH, Zanin GM, de Castro HF. 2007. Biodiesel synthesis by enzymatic
transesterificationofpalmoilwithethanolusinglipasesfromseveralsourcesimmobilizedonsilica
PVAcomposite.EnergyFuels21:36893694.
Mukesh D, Banerji AA, Newadkar R, Bevinakatti HS. 1993. Lipase catalysed transesterification of
vegetableoilsAcomparativestudyinbatchandtubularreactors.BiotechnolLett15:7782.
NassreddineS,KaroutA,ChristML,PierreAC.2008.Transesterificationofavegetaloilwithmethanol
catalyzedbyasilicafibrereinforcedaerogelencapsulatedlipase.ApplCatal,A344:7077.
NieK,XieF,WangF,TanT.2006.Lipasecatalyzedmethanolysistoproducebiodiesel:Optimizationof
thebiodieselproduction.JMolCatalB:Enzym43:142147.
122
Noel M, Combes D. 2003. Rhizomucor miehei lipase: Differential scanning calorimetry and
pressure/temperature stability studies in presence of soluble additives. Enzyme Microb Technol
33:299308.
NoureddiniH,GaoX,JoshiS.2003.ImmobilizationofCandidarugosaLipasebySolGelEntrapmentand
ItsApplicationintheHydrolysisofSoybeanOil.JAmOilChemSoc80:10771083.
NoureddiniH,GaoX,JoshiS,WagnerPR.2002.ImmobilizationofPseudomonascepacialipasebysolgel
entrapmentanditsapplicationinthehydrolysisofsoybeanoil.JAmOilChemSoc79:3340.
Noureddini H, Gao X, Philkana RS. 2005. Immobilized Pseudomonas cepacia lipase for biodiesel fuel
productionfromsoybeanoil.BioresourTechnol96:769777.
OgnjanovicN,BezbradicaD,KnezevicJugovicZ.2009.Enzymaticconversionofsunfloweroiltobiodiesel
in a solventfree system: Process optimization and the immobilized system stability. Bioresour
Technol100:51465154.
OraireO,BuissonP,PierreAC.2006.Applicationofsilicaaerogelencapsulatedlipasesinthesynthesis
ofbiodieselbytransesterificationreactions.JMolCatalB:Enzym42:106113.
PierreAC.2004.Thesolgelencapsulationofenzymes.BiocatalBiotransform22:145170.
PirozziD,FanelliE,AronneA,PerniceP,MingioneA.2009.Lipaseentrapmentinazirconiamatrix:Sol
gelsynthesisandcatalyticproperties.JMolCatalB:Enzym59:116120.
PogorilyiRP,SiletskayaEY,GoncharikVP,KozharaLI,ZubYL.2007.Immobilizationofureaseonthesilica
gelsurfacebysolgelmethod.RussJApplChem80:330334.
RahmanNK,KamaruddinAH,UzirMH.2011.Enzymaticsynthesisoffarnesyllaurateinorganicsolvent:
Initial water activity, kinetics mechanism, optimization of continuous operation using packed bed
reactorandmasstransferstudies.BioprocessBiosystEng34:687699.
Ranganathan SV, Narasimhan SL, Muthukumar K. 2008. An overview of enzymatic production of
biodiesel.BioresourTechnol99:39753981.
RaseHF.Fixedbedreactordesignanddiagnostics:gasphasereactions.Boston:Butterworthsc1990.
123
Reetz MT. 1997. Entrapment of biocatalysts in hydrophobic solgel materials for use in organic
chemistry.AdvMater9:943954.
Reetz MT, Zonta A, Simpelkamp J. 1996. Efficient immobilization of lipases by entrapment in
hydrophobicsolgelmaterials.BiotechnolBioeng49:527534.
RizziM,StylosP,RiekA,ReussM.1992.Akineticstudyofimmobilizedlipasecatalysingthesynthesisof
isoamylacetatebytransesterificationinnhexane.EnzymeMicrobTechnol14:709714.
RoblesMedinaA,GonzlezMorenoPA,EstebanCerdnL,MolinaGrimaE.2009.Biocatalysis:Towards
evergreenerbiodieselproduction.BiotechnolAdv27:398408.
RoyonD,DazM,EllenriederG,LocatelliS.2007.Enzymaticproductionofbiodieselfromcottonseedoil
usingtbutanolasasolvent.BioresourTechnol98:648653.
RuzichNI,BassiAS.2011.Proposedkineticmechanismofbiodieselproductionthroughlipasecatalysed
interesterificationwithamethylacetateacylacceptorandionicliquid[BMIM][PF6]cosolvent.Can
JChemEng89:166170.
Ruzich NI, Bassi AS. 2010a. Investigation of enzymatic biodiesel production using ionic liquid as a co
solvent.CanJChemEng88:277282.
Ruzich NI, Bassi AS. 2010b. Investigation of lipasecatalyzed biodiesel production using ionic liquid
[BMIM][PF6] as a cosolvent in 500 mL jacketed conical and shake flask reactors using triolein or
wastecanolaoilassubstrates.EnergyFuels24:32143222.
Sairolu A. 2008. Conversion of sunflower oil to biodiesel by alcoholysis using immobilized lipase.
ArtificialCells,BloodSubstitutes,andBiotechnology36:138149.
SairoluA,KilinA,TelefoncuA.2004.Preparationandpropertiesoflipasesimmobilizedondifferent
supports.ArtifCellsBloodSubstituesImmobilizationBiotechnol32:625636.
SairoluA,TelefoncuA.2004.Immobilizationoflipasesondifferentcarriersandtheiruseinsynthesis
ofpentylisovalerates.PrepBiochemBiotechnol34:169178.
Salis A, Pinna M, Monduzzi M, Solinas V. 2005. Biodiesel production from triolein and short chain
alcoholsthroughbiocatalysis.JBiotechnol119:291299.
124
Samukawa T, Kaieda M, Matsumoto T, Ban K, Kondo A, Shimada Y, Noda H, Fukuda H. 2000.

PretreatmentofimmobilizedCandidaantarcticalipaseforbiodieselfuelproductionfromplantoil.
JBiosciBioeng90:180183.
Santacesaria E, Tesser R, Di Serio M, Guida M, Gaetano D, Garcia Agreda A. 2007. Kinetics and mass
transfer of free fatty acids esterification with methanol in a tubular packed bed reactor: A key
pretreatmentinbiodieselproduction.IndEngChemRes46:51135121.
SardaL,DesnuelleP.1958.Actiondelalipasepancratiquesurlesestersenmulsion.BiochimBiophys
Acta30:513521.
SeguinD,MontilletA,BrunjailD,ComitiJ.1996.Liquidsolidmasstransferinpackedbedsofvariously
shapedparticlesatlowReynoldsnumbers:Experimentsandmodel.ChemEngJ63:19.
Selmi B, Thomas D. 1998. Immobilized lipasecatalyzed ethanolysis of sunflower oil in a solventfree
medium.JAmOilChemSoc75:691695.
Semwal S, Arora AK, Badoni RP, Tuli DK. 2011. Biodiesel production using heterogeneous catalysts.
BioresourTechnol102:21512161.
Shah S, Sharma S, Gupta MN. 2004. Biodiesel preparation by lipasecatalyzed transesterification of
Jatrophaoil.EnergyFuels18:154159.
Shahid EM, Jamal Y. 2011. Production of biodiesel: A technical review. Renewable Sustainable Energy
Rev15:47324745.
Shaw JF, Chang SW, Lin SC, Wu TT, Ju HY, Akoh CC, Chang RH, Shieh CJ. 2008. Continuous
enzymaticsynthesisofbiodieselwithNovozym435.EnergyFuels22:840844.
ShimadaY,WatanabeY,SamukawaT,SugiharaA,NodaH,FukudaH,TominagaY.1999.Conversionof
vegetableoiltobiodieselusingimmobilizedCandidaantarcticalipase.JAmOilChemSoc76:789
793.
Shimada Y, Watanabe Y, Sugihara A, Tominaga Y. 2002. Enzymatic alcoholysis for biodiesel fuel
productionandapplicationofthereactiontooilprocessing.JMolCatalB:Enzym17:133142.
Soares CMF, dos Santos OA, de Castro HF, de Moraes FF, Zanin GM. 2006. Characterization of solgel
encapsulatedlipaseusingtetraethoxysilaneasprecursor.JMolCatalB:Enzym39:6976.
125
SoumanouMM,BornscheuerUT.2003a.Improvementinlipasecatalyzedsynthesisoffattyacidmethyl
estersfromsunfloweroil.EnzymeMicrobTechnol33:97103.
Soumanou MM, Bornscheuer UT. 2003b. Lipasecatalyzed alcoholysis of vegetable oils. Eur J Lipid Sci
Technol105:656660.
Szczesna Antczak M, Kubiak A, Antczak T, Bielecki S. 2009. Enzymatic biodiesel synthesis Key factors
affectingefficiencyoftheprocess.RenewableEnergy34:11851194.
Tan T, Lu J, Nie K, Deng L, Wang F. 2010. Biodiesel production with immobilized lipase: A review.
BiotechnolAdv28:628634.
TesserR,DiSerioM,GuidaM,NastasiM,SantacesariaE.2005.Kineticsofoleicacidesterificationwith
methanolinthepresenceoftriglycerides.IndEngChemRes44:79787982.
Thoenes D. Chemical reactor development: from laboratory synthesis to industrial production.
Dordrecht:KluwerAcademicPublishersc1994.
Tongboriboon K, Cheirsilp B, HKittikun A. 2010. Mixed lipases for efficient enzymatic synthesis of
biodieselfromusedpalmoilandethanolinasolventfreesystem.JMolCatalB:Enzym67:5259.
TreybalRE.Masstransferoperations.NewYork:McGrawHillc1980.
ValeriD,MeirellesAJA.1997.Viscositiesoffattyacids,triglycerides,andtheirbinarymixtures.JAmOil
ChemSoc74:12211226.
Vasudevan PT, Briggs M. 2008. Biodiesel productioncurrent state of the art and challenges. J Ind
MicrobiolBiotechnol35:421430.
Vieira APDA, da Silva MAP, Langone MAP. 2006. Biodiesel production via esterification reactions
catalyzedbylipase.LatAmApplRes36:283288.
Wang L, Du W, Liu D, Li L, Dai N. 2006. Lipasecatalyzed biodiesel production from soybean oil
deodorizerdistillatewithabsorbentpresentintertbutanolsystem.JMolCatalB:Enzym43:2932.
Wang X, Liu X, Zhao C, Ding Y, Xu P. 2011. Biodiesel production in packedbed reactors using lipase
nanoparticlebiocomposite.BioresourTechnol102:63526355.
126
Wang YD, Shen XY, Li ZL, Li X, Wang F, Nie XA, Jiang JC. 2010. Immobilized recombinant Rhizopus
oryzaelipasefortheproductionofbiodieselinsolventfreesystem.JMolCatalB:Enzym67:4551.
Watanabe Y, Shimada Y, Sugihara A, Noda H, Fukuda H, Tominaga Y. 2000. Continuous production of
biodiesel fuel from vegetable oil using immobilized Candida antarctica lipase. J Am Oil Chem Soc
77:355360.
Watanabe Y, Shimada Y, Sugihara A, Tominaga Y. 2002. Conversion of degummed soybean oil to
biodieselfuelwithimmobilizedCandidaantarcticalipase.JMolCatalB:Enzym17:151155.
Watanabe Y, Shimada Y, Sugihara A, Tominaga Y. 2001. Enzymatic conversion of waste edible oil to
biodieselfuelinafixedbedbioreactor.JAmOilChemSoc78:703707.
Wiemann L, Nieguth R, Eckstein M, Naumann M, Thum O, AnsorgeSchumacher M. 2009. Composite
particlesofNovozyme435andsilicone:Advancingtechnicalapplicabilityofmacroporousenzyme
carriers.ChemCatChem1:455462.
Xiao Y, Gao L, Xiao G, Fu B, Niu L. 2012. Experimental and modeling study of continuous catalytic
transesterification to biodiesel in a benchscale fixedbed reactor. Ind Eng Chem Res 51:11860
11865.
Xu Y, Du W, Liu D. 2005. Study on the kinetics of enzymatic interesterification of triglycerides for
biodieselproductionwithmethylacetateastheacylacceptor.JMolCatalB:Enzym32:241245.
XuY,DuW,LiuD,ZengJ.2003.Anovelenzymaticrouteforbiodieselproductionfromrenewableoilsin
asolventfreemedium.BiotechnolLett25:12391241.
XuY,DuW,ZengJ,LiuD.2004.ConversionofsoybeanoiltobiodieselfuelusinglipozymeTLIMin a
solventfreemedium.BiocatalBiotransform22:4548.
XuY,NordbladM,WoodleyJM.2012.Atwostageenzymaticethanolbasedbiodieselproductionina
packedbedreactor.JBiotechnol.ArticleinPress:
http://dx.doi.org.proxy.lib.uwaterloo.ca/10.1016/j.bbr.2011.03.031
Xu ZP, Chuang KT. 1997. Effect of internal diffusion on heterogeneous catalytic esterification of acetic
acid.ChemEngSci52:30113017.
127
YadavGD,DeviKM.2004.Immobilizedlipasecatalysedesterificationandtransesterificationreactionsin
nonaqueous media for the synthesis of tetrahydrofurfuryl butyrate: Comparison and kinetic
modeling.ChemEngSci59:373383.
Yori JC, D'Ippolito SA, Pieck CL. 2007. Deglycerolization of biodiesel streams by adsorption over silica
beds.EnergyFuels21:347353.
YcelY,DemirC.2012.Theoptimizationofimmobilizedlipasecatalyzedtransesterificationofcanolaoil
byresponsesurfacemethodologyandmixturedesign.EnergySourcesPartA34:20312040.
Yusuf NNAN, Kamarudin SK, Yaakub Z. 2011. Overview on the current trends in biodiesel production.
EnergyConversManage52:27412751.
ZhangB,WengY,Xu H,MaoZ.2012.Enzymeimmobilization forbiodieselproduction. Appl Microbiol
Biotechnol93:6170.
128
AppendixA
ListofPublications
PeerReviewedPublications
Study of Support Materials for Solgel Immobilized Lipase, S.M. Meunier and R.L. Legge, Biocatal.
Biotransform.,submittedJuly2012.ManuscriptIDGBAB12404.R1.
Evaluation of Diatomaceous Earth Supported Lipase Solgels as a Medium for Enzymatic
TransesterificationofBiodiesel,S.M.MeunierandR.L.Legge,J.Mol.Catal.B:Enzym.,vol.77,pp.
9297,2012.
EvaluationofDiatomaceousEarthasaSupportforSolgelImmobilizedLipaseforTransesterification,
S.M.MeunierandR.L.Legge,J.Mol.Catal.B:Enzym.,vol.62,no.1,pp.5357,2010.
ConferencePresentations
BiodieselProductionPotentialviaSupportedLipaseSolgels,S.M.MeunierandR.L.Legge,14thAnnual
CSChEOntarioQubecBiotechnologyMeeting,May2012.
Biodiesel Production Catalyzed by Celite Supported Lipase Solgels, S.M. Meunier and R.L. Legge,
CurrentResearchinEngineering,Science,&Technology(CREST)Meeting,March2012.
Solgel Immobilized Lipase Supported on Diatomaceous Earth for Lipid Transesterification, S.M.
MeunierandR.L.Legge,11thAnnualCSChEOntarioQuebecBiotechnologyMeeting,June2009.
Supported Solgel Immobilized Enzymes for Lipid Transesterification, S. M. Meunier and R. L. Legge,
WISEInitiative2009InternationalWomensDayConference,March2009.
129
AppendixB
LipaseSourceChange
TheexperimentalworkdescribedinChapters3through6wascompletedusinglipaseNS44035provided
byNovozymesNorthAmericaInc.Thislipasewasusedsincethesolgelimmobilizationprocedureand
linoleic acid assay methodologies were well established based on earlier experiments. Since all the
initial experiments were conducted using the Novozymes lipase, the supported solgels studied for
methyloleateproductionwereproducedusingthesamemethodology.
One of the concerns identified regarding the results obtained was that the Novozymes lipase is not a
commonlipaseandthesupplierwillnotdisclosethesourceoftheenzyme.Areviewoftheliteratureon
enzymaticbiodieselproductionrevealedthatLipasePSAmanoSDfromAmanoEnzymeUSACo.wasa
very active lipase for this application. The enzyme source for the Celite supported lipase solgels
formulatedforChapters7and8isthereforefromAmanoratherthanNovozymes.
The basic information regarding the two types of lipase used in this thesis is provided in Table B.1.
Experimentally,themaindifferencebetweenthetwolipasesisthattheNovozymeslipasewassupplied
as a liquid and Amano lipase as a powder, resulting in a slight change in the solgel formulation
proceduretoensurethattheAmanolipasewaswelldissolvedinthephosphatebufferpriortouse.The
proteincontentoftheAmanolipasewasunknown,sotheamountoflipaseusedwaschosensothatthe
activityoftheproteinloadedwouldbecomparablebetweenthetwoenzymesources.
TableB.1:InformationfromNovozymesNorthAmericaInc.andAmanoEnzymeUSACo.forthetwotypesoflipaseusedin
theexperimentsperformedforthisresearch.
Lipase(Supplier)
NS44035
(Novozymes)
LipasePSAmanoSD
(AmanoEnzyme)
Source
Unknown
1,3specificlipase
Burkholderiacepacia
Form
Activity(U/g)
Liquid
20000
Powder
23000
130
Optimalconditions
pH=7.5
Temperature=40C
pH=7
Temperature=50C
A series of experiments were conducted with the Amano lipase to ensure that similar levels of
performance could be obtained using the new source of lipase (Table B.2). Based on the results
obtained, the unsupported solgel and the Celite supported solgel with Amano lipase had slightly
higherproteincontentcomparedtotheNovozymeslipase.However,thedegreeofimmobilizationthat
couldbeachievedusingtheNovozymeslipasesolgelswasmuchhigherforboththeunsupportedand
CelitesupportedsolgelsthanthatwiththeAmanolipase.Thiscouldbeduetotheneedtodissolve
theAmanolipasepowderinphosphatebufferpriortoitsadditiontothesolgelprecursorsolution.For
the free lipase, unsupported solgel, and Celite supported solgel, the Amano lipase exhibited higher
methyloleateconversionafterthe6hreactionthantheNovozymeslipase.Ineachcase,the6hmethyl
oleate concentration for the Amano lipase was close to double that of the Novozymes lipase.
Consideringtheinitialactivityofthelipase,boththeunsupportedsolgelandtheCelitesupportedsol
gelhadapproximatelydoubletheinitialactivityusingtheAmanoenzyme,butthefreelipaseexhibited
onlyhalftheinitialactivityoftheNovozymeslipase.Therefore,immobilizingtheAmanolipaseinthe
solgel (unsupported or supported) dramatically increased the enzymatic activity for methyl oleate
production.Basedontheseresults,theAmanolipasewasfoundtobeagoodchoiceoflipaseforthe
subsequentCelitesupportedlipasesolgelmethyloleateproductionstudies.
ForthesolgelsproducedusingtheAmanolipase,theHPLCprocedureforproteincontentidentification
wasfoundtobeunreliable.Although,thecompositionfortheAmanolipasefromthesupplierindicated
thatthepowderprovidedwas90%dextrinand10%lipase,HPLCanalysisindicatedmanypeaks(10or
more)whichislikelyduetothemixtureofcarbohydratesinthedextrinthatwasused.Theresolutionof
thepeakswasverypoor,andidentificationofalipasepeakwasnotpossible.Therefore,aBCAassay
wasusedtodeterminetheproteincontentoftheAmanolipasesupportedsolgels.
TableB.2:Proteincontent,degreeofimmobilization(DOI),methyloleateconcentrationafter6h(6h[MO])andinitial
activityfortheNovozymesandAmanolipasesforfreelipase,solgelimmobilizedlipase,andCelitesupportedlipasesolgel.
Support
Free
Solgel
CeliteSolgel
Lipase
Proteincontent
(mg/g)
DOI
(%)
6h[MO]
(M/g)
Initialactivity
(U/g)
Novozymes
Amano
Novozymes
Amano
Novozymes
Amano
NA
NA
4.86
5.60
1.45
2.21
NA
NA
80.11
26.54
85.88
72.52
1.05
1.65
178480
396382
203513
440473
59.78
29.74
6077
10128
4676
10503
131
TheunsupportedsolgelandtheCelitesupportedsolgelcontainingbothtypesoflipasewerestudied
using SEM to determine if there were any structural differences in the nature of the solgel material
(FigureB.1).Themostnotabledifferenceobservedwaswiththeunsupportedlipasesolgelswherethe
Novozymesunsupportedsolgelresultedinanirregularlyshapedsurface,buttheAmanounsupported
solgel produced defined particles possessing a globular structure, even after solgel immobilization.
This is likely due to the differences in other components in the Amano lipase preparation. For the
CelitesolgelwiththeAmanolipase,thesolgelontheCelitesupportwasverysphericalwithnoair
bubbles,butfortheCelitesolgelwiththeNovozymeslipase,thesolgelwasnonsphericalwithmany
airbubbles.
NovozymesLipase
AmanoLipase
NovozymesLipase
AmanoLipase
UnsupportedSolgel
UnsupportedSolgel
CeliteSolgel
CeliteSolgel
FigureB.1:SEMimagesat3000xmagnificationoftheunsupportedlipasesolgelandtheCelitelipasesolgelwiththe
NovozymesandAmanolipases.ArrowsontheCelitesolgelsimagesindicatethepresenceofsolgelontheCelite.
132
AppendixC
SampleCalculations
GCMSEnzymaticProperties
The GCMS method used to follow the enzymatic production of methyl oleate (MO) with time is
described in Sections 4.2.2.6, 5.2.2.5, 6.2.2.3, 7.2.2.3, and 8.2.2.5. Heptadecanoic acid methyl ester
(HDAME)wasusedasaninternalstandardandaplotofthepeakareaofMO/HDAMEvs.theknown
concentrationofMO/HDAMEisacalibrationcurveusedtodeterminetheMOconcentrationfromthe
peakareasobtainedusingtheGCMSVarianProStarsoftware(FigureC.1).
14
y=0.5300x
R=0.9990
PeakAreaMO/HDAME
12
10
0
0
10
15
[MO]/[HDAME]
20
25
FigureC.2:SamplecalibrationcurvefortheGCMScalibrationusedtodeterminetheconcentrationofmethyloleate,[MO],
basedonthepeakareas.Theinternalstandardisheptadecanoicacidmethylester(HDAME).
133
Standard:
4.113minpeakHDAME
6.136minpeakMO
Sample:
4.114minpeakHDAME
6.142minpeakMO
FigureC.2:SamplechromatogramsobtainedfromGCMSwherethetopchromatogramisastandardsolutionandthebottom
chromatogramisasamplesolution.Thepeakat4.11minisHDAMEandat6.14minisMO.
134
Tocalculatetheenzymaticpropertiesforanygivenrun,aplotofMOconcentrationvs.time(FigureC.3)
was created based on the chromatogram and calibration curve obtained from the GCMS. The initial
reaction rate, initial activity, and 6 h percent conversion were calculated based on the equations
provided.Furtherdescriptionoftheseparametersandcalculationscanbefoundinsection4.3.3.
100
ConcentrationofMethylOleate(mM)
400
300
6hrMethylOleate
Concentration
InitialReactionRate
200
y=74.939x
R=0.9993
100
0
0
3
Time(hours)
FigureC.3:Samplemethyloleateconcentrationvs.timeplotusedtodeterminetheenzymaticproperties(initialreaction
rateand6hmethyloleateconcentration)basedona6hbatchreaction.
135
HPLCProteinContent
To determine the protein content of the solgels, the HPLC method described in sections 3.1, 4.2.2.3,
5.2.2.4,and6.2.2.2wasusedandacalibrationcurveforaBSAstandardwasdeveloped(FigureC.4).
To determine the protein content of the solgels, a mass balance was used based on the amount of
proteinloadedtothesolgelfromtheimmobilizationprocedureandtheamountofproteinpresentin
the solgel phosphate buffer washes. Chromatograms were obtained from the HPLC method (Figure
C.5)wheretheupperchromatogramrepresentsthelipasesolutionloadedtothesolgelandthelower
chromatogram represents the phosphate buffer wash. Based on activity assays, the lipase peak is
known to elute after approximately 10 minutes. The 15 minute peak is an additive in the protein
solutionobtainedfromthesupplier.Thepeakheightwasusedintheanalysis.
Other data necessary for the calculations were obtained from the solgel preparation, such as the
volumeoflipasesolutionloadedtothesolgels,thevolumeoflipasesolutioninthebufferwashes,the
mass of support material used in the supported solgels, and the total mass of material produced for
boththesupportedandunsupportedsolgelformulations.
250000
BSAPeakHeight
200000
150000
y=113.4x
R=0.9995
100000
50000
0
0
200
400
600
800
1000 1200 1400
ActualConcentration (g/mL)
1600
1800
2000
FigureC.4:SamplecalibrationcurvefortheHPLCcalibrationbasedonaBSAstandardandusedtodeterminethetotal
proteincontentofthesolgels.
136
Basedonthisdata,theproteinloading,degreeofimmobilization,andsolgeladhesionofthesupported
andunsupportedsolgelswereobtainedfromthefollowingequations.
100
Overlaid chromatogram with the upper curve

representingthelipasestocksolutionandthelower
curverepresentingabufferwash.
Thepeakatapproximately10ministhelipasepeak
and the 15 min peak is an additive in the lipase
solutionobtainedfromthesupplier.
FigureC.5:SamplechromatogramsfortheHPLCproteincontentdeterminationwheretheuppercurveistheproteinsolution
loadedtothesolgelsandthelowercurveisabufferwash.Thepeakat10minutesisthelipasepeak.
137
SEMCoatingAnalysis
To quantify the amount of solgel coating present on Celite solgels as described in sections 5.2.2.3,
5.3.1,and5.3.2aseriesof45SEMimagesweretakenforeachsupportedsolgelandthesolgelclusters
onthesurfaceoftheCelitewereidentifiedvisually.FigureC.6 isanexampleoftwoSEMimagesto
demonstrate this process where a) is a sample SEM image and b) shows 6 solgel clusters that were
identifiedontheoriginalimage.
AMatLabprogramwasdevelopedtocalculatetheareaofeachsolgelclusteridentifiedintheimages.
Basedonthecollectionofimages,thepercentcoverageofsolgelonthesurfaceoftheCeliteandthe
average cluster size were determined and compared between the three different types of Celite
consideredassolgelsupportmaterials.
a)
b)
FigureC.6:SEMimagesdemonstratingthesurfaceareacoverageproceduredevelopedwherea)isasampleimageandb)
showsthesameimagewith6solgelclustersidentifiedforanalysis.
138
TGAWaterContentandThermostability
A variety of properties of the supported solgels were determined based on plots of the weight vs.
temperatureandthederivativeweightvs.temperatureobtainedviaTGA(FigureC.7).
As described in section 4.2.2.6, the water content and thermal stability of the supported solgels was
determinedviaTGA.Forthisprocedure,thesupportedsolgelswereanalyzedwithoutpretreatment,
andthetotalmassloss,whichoccurredatapproximately100C,wasassumedtobesolelyduetothe
presenceofwater.Inaddition,fromthesethermogramsthethermalstabilityofthesupportedsolgel
was evaluated based on whether the weight stabilized after the water was removed (stable) or the
weightcontinuedtodecreaseasthetemperatureincreased(unstable).
Section6.2.2.4describesthemethodologyforanalternativeuseofTGAtomeasureglycerolandwater
desorptionfromtheCelitesolgels.Forthisanalysis,thesolgelswereimmersedinsolutionsofvarying
concentrations of glycerol for 24 hrs, rinsed to remove excess glycerol, and air dried overnight.
Subsequent to this treatment, the Celite solgels were analyzed via TGA and the thermograms were
usedtodeterminethepeakdesorptionrate,peakdesorptiontemperature,totalmassloss,andonsetof
desorption.
100
0.4
PeakDesorption
98
0.35
96
Weight(%)
TotalMassLoss
92
0.25
0.2
90
88
0.15
86
DerivativeWeight(%/oC)
0.3
94
0.1
84
0.05
82
80
0
0
50
100
150
200
250
Temperature(oC)
Weight
300
DerivativeWeight
350
400
FigureC.7:Samplethermogramshowingtheweightpercentandderivativeweightpercentasafunctionoftemperature.
139

Meunier Sarah PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Meunier Sarah PDF

Uploaded by

Copyright:

Available Formats

DevelopmentofaPackedbedReactor

Chapter 3 ExperimentalBackground ___________________________________________________ 26

Chapter 4 StudyofSupportMaterialsforSolgelImmobilizedLipase _________________________ 37

Experimentalandmodellingresults ________________________________________ 102

Modelpredictions ______________________________________________________ 106

Acknowledgements _________________________________________________________ 108

Chapter 9 ConclusionsandRecommendations __________________________________________ 111

RecommendationsforFutureWork ____________________________________________ 114

References ________________________________________________________________________ 115

LipaseSourceChange ___________________________________________________ 130

Diffusionofsolute insolvent (m2/s)

is the inhibition constant for the alcohol,

is the apparent Michaelis Menten

is the bulk concentration of substrate ,

is the concentration of at the

is the Sherwood number,

less than 10 where the Reynolds number of the particle (

The diffusion coefficient (

) used in these equations can be estimated by the Wilke and Chang

Finally, the differential term in Equation 2.3 (

) can be determined from the kinetics of the

1:1Methanol Solventfree NoWater

Study of Support Materials for Solgel Immobilized

6hPerentConversion ofMethanol toMethylOleate

thealcohol, istheglycerol,and isthemethyloleate:

Thiskinetic modeldemonstratesgoodagreementbetween theinitialexperimentaldatacollectedand

AlthoughEquation7.2agreeswith the experimentaldatafortheinitialreactionrate,significantover

Methyl Oleate Concentration at 6-hr (M)

where is the triglyceride, is the

alcohol, istheglycerol,and isthemethyloleate(AlZuhair,2005;AlZuhairetal.,2007;AlZuhairet

is the maximum initial reaction velocity,

are the alcohol and triglyceride

are the apparent Michaelis constants for the alcohol and

is the external surface area of the catalyst,

is the bulk concentration of the substrate,

) in Equation 8.2 (Dwivedi and Upadhyay,

where is the void fraction in the reactor,

is the Reynolds number for the particle and

verydilutesolutioninsolvent where referstothemethanoland tothetriolein(Treybal,1980).

where is the dissociation factor for the triolein,

is the molecular mass of the triolein, T is the

)withrespecttotemperaturebasedonpuretrioleinwhere iscSt, iscP,

) and that alcohol

Objective C Evaluation of Diatomaceous Earth as a Support for Solgel Immobilized

Objective D Evaluation of Diatomaceous Earth Supported Solgels as a Medium for

Objective F Kinetic and Mass Transfer Modelling of Methyl Oleate Production in an

Fundamentals: Quantification of glycerol and the reaction intermediates (diglycerides and

Ganesan D, Rajendran A, Thangavelu V. 2009. An overview on the recent advances in the

Samukawa T, Kaieda M, Matsumoto T, Ban K, Kondo A, Shimada Y, Noda H, Fukuda H. 2000.

Overlaid chromatogram with the upper curve

You might also like