InsightsintoHeterogeneousCatalystsfortheHMF
SynthesisfromBiomass
vorgelegtvon
M.Sc.
SylviaReiche
ausDahlen
VonderFakultätII‐MathematikundNaturwissenschaften
derTechnischenUniversitätBerlin
zurErlangungdesakademischenGrades
DoktorderNaturwissenschaften
Dr.rer.nat.
genehmigteDissertation
Promotionsausschuss:
Vorsitzender:Prof.Dr.rer.nat.ReinhardSchomäcker
Gutachter:Prof.Dr.rer.nat.RobertSchlögl
Gutachter:Prof.Dr.rer.nat.ArneThomas
Gutachter:Prof.Dr.rer.nat.DangshengSu
TagderwissenschaftlichenAussprache:27.03.2012
Berlin2012
D83
Lagravitationdel’espritnousfaittomberverslehaut.
SimoneWeil
Kurzzusammenfassung
Die säurekatalysierte Dehydratation von Fructose unter Bildung von 5‐
(Hydroxymethyl)furfural (HMF) stellt eine wichtige Modelreaktion für den
notwendigenRohstoffwandelinderchemischenIndustriedar.IndervorgelegtenArbeit
wurdeeineVielzahlheterogenerKatalysatorenfürdieHMF‐Synthesegetestetundmit
ErgebnissenderLiteraturverglichen.Eswurdedeutlich,dassalleMaterialienbereits
nachdererstenAnwendungdeutlichdeaktiviertenundUntersuchungenzurStabilität
der Katalysatoren nur unzureichend dokumentiert sind. Im Rahmen der Dissertation
wurde ein einphasiges Verfahren entwickelt, durch das die Nebenreaktion der
Levulinsäurebildungverhindertwerdenkonnte.DieNutzungvon2‐Butanolalseinziges
LösungsmittelhatdenweiterenVorteil,dassProduktverunreinigungen,z.B.durchdie
notwendigeZugabevonSalzenimzweiphasigenVerfahren,vermiedenwerdenkönnen.
Zudem kann eine Produktakkumulation an der Katalysatoroberfläche graphitischer
Kohlenstoffkatalysatoren verhindert werden, welche andernfalls Folgereaktionen in
FormvonPolymerizationbegünstigenkönnen.
Unter Verwendung des einphasigen Verfahrens in 2‐Butanol wurde die
Desaktivierung saurer Kohlenstoffkatalysatoren systematisch untersucht. Dabei
konnten drei verschiedene Deaktivierungsprozesse unterschieden werden: (1) Das
AuslaugenderfunktionellenGruppenunterReaktionsbedingungen,(2)dieAblagerung
unlöslicher Nebenprodukte, sogenannter Humine, auf der Katalysatoroberfläche und
(3) die Desaktivierung der Katalysatoroberfläche durch die Reaktion mit dem
alkoholischen Lösungsmittel. Das Auslaugen funktioneller Gruppen (Leaching) wurde
fürallepost‐funktionalisiertenKatalysatorendetektiert.DahersindLeaching‐Testsein
essentielles Element vollständiger Katalysatorcharakterisierungundentsprechende
VorbehandlungsschritteeinenotwendigeVoraussetzungfürstabileKatalysatoren.Auch
wenn für ein Material das Auslaugen der Säuregruppen aufgrund entsprechender
Vorbehandlung ausgeschlossen werden konnte, so ist die Stabilität in der Fructose
Dehydratationnichtautomatischgegeben.DieVielzahlmöglicherNebenreaktionführt
zuunlöslichenNebenprodukten,welchewiederumdieaktivenSäurezentrenblockieren
und den Katalysator sukzessive deaktivieren. Aufgrund dessen wurde eine
Referenzreaktion mit monofunktionalem Reaktant, die säurekatalysierte Veresterung
vonEssigsäuremitEthanol,zurvergleichendenBetrachtungherangezogen.Einestabile
KatalysatoraktivitätbeierneutemEinsatz(Recycling)inderVeresterungkonntedabei,
ähnlichwiedieLeaching‐Tests,alsnotwendigeabernichthinreichendeBedingungfür
dieStabilitätdesKatalysatorsinderFructosedehydratationangesehenwerden.Dieses
ErgebnisunterstreichtdieNotwendigkeitzusätzlichermechanistischerKenntnissedes
komplexen Reaktionsnetzwerks der Fructosedehydratation an heterogenen
Katalysatoren. Schließlich konnte mittels in‐situ XPS gezeigt werden, dass eine
Vorbehandlung sehr wohl Einfluss auf die Oberflächenchemie der untersuchten
Kohlenstoffmaterialien haben kann. Speziell die Reaktivitäten gegenüber Wasser
ändertensichfürdieoxidiertenmesoporösenMaterialiennachderVorbehandlung.
Abstract
Theacidcatalyzeddehydrationoffructosein5‐hydroxymethylfurfural(HMF)isan
importantmodelreactionforthenecessaryfeedstockchangeinchemicalindustry.In
the present work, multiple heterogeneous catalysts have been testedfortheHMF
synthesisfromfructoseandtheresultshavebeencomparedtoliteraturedata.Itcould
beshownthatthematerialsdeactivatedstronglyalready after the firstreactionrun.
The comparative studies of the literature showed the leak of stability tests for
heterogeneouscatalystsinfructosedehydration.
Forcomparablestudiesofthecatalysts,aone‐phasedehydrationprocedurein 2‐
butanolhasbeenelaborated.Theuseof2‐butanolastheonlyreactionsolventinhibits
the product accumulation on hydrophobic catalysts (e.g. functionalized carbon
nanotubes),asconfirmedbyadsorptionstudies.Inaddition,therehydrationtotheside
productslevulinicacidandformicacidaresuppressed.Theprocessavoidsimpuritiesin
theHMF‐productthatcanbeproblematicinsubsequentprocessingsteps,suchasS‐and
N‐containingsolventsorsaltresiduesappliedinthebiphasicprocess.Hencetheone‐
phase system in 2‐butanol would require lower purification costs than other
establishedprocesses.
Comparative studies on carbon based heterogeneous catalysts revealed three
differentdeactivatingprocesses:(1)theleachingofinstableacidfunctionalgroups,(2)
thesurfacecoverageorsideblockingbyinsolublepolymericbyproducts(humins)and
(3) the surface passivation by the alcoholic solvent. The stabilityofacidfunctional
groupswasinvestigatedbyactivitytestsofthesolventafterpreconditioning of the
catalyst(leachingtests).Itwasfoundthatleachingisacommonproblemforallpost‐
functionalized,e.g.sulfonated,catalystsandappropriatepretreatmentsarerequired,in
ordertostartwithastablefractionofacidfunctionalgroups.Leachingtestsare
essentialforthoroughproofofcatalyststabilityandhencearesuggestedtobeadapted
ascommontestforheterogeneouscatalystsinfructosedehydration.Furthermorewe
establishedthecomparisontothereferencereaction,theesterificationofaceticacidin
ethanol,assuccessfultoolfortheestimationofcatalyststability.Thedeactivationby
leachingorsolventreactionscanbetracedbythereferencereaction. However, the
successfulrecyclingintheesterificationreactiondoesnotguaranteeastablecatalystfor
HMFsynthesis,whichunderlinesthenecessityoffurthermechanisticunderstandingof
thefructosedehydrationonheterogeneouscatalysts.
Finally, it could be shown by in‐situ XPS that the catalyst pretreatment can
significantlyinfluencethesurfacechemistryofthecarbonmaterials.Inparticular,the
reactivity towards water differed for the oxidized mesoporous carbon after the
pretreatmentinalcoholicsolvent.
I
TABLEOFCONTENTS
TableofContents.......................................................................................................................................................I
ListofFigures...........................................................................................................................................................III
ListofTables............................................................................................................................................................VII
IIntroduction.............................................................................................................................1
I.1 TheFeedstockCellulosicBiomass..................................................................................................2
I.2 HMFSynthesisfromBiomass...........................................................................................................4
I.3 MechanismofFructoseDehydrationtoHMF............................................................................9
I.3.1 Overviewschemeofreactionnetwork......................................................................................10
I.3.2 Historicdevelopmentofmechanisticideasonfructosedehydration..........................13
I.3.3 Mechanisticsolventeffects.............................................................................................................19
I.4 References..............................................................................................................................................20
IIOutlineoftheWork.............................................................................................................23
IIIHeterogeneousCatalystsintheDehydrationofFructosetoHMF.....................25
III.1 Introduction..........................................................................................................................................26
III.2 Results.....................................................................................................................................................27
III.3 Discussion..............................................................................................................................................32
III.4 ExperimentalSection........................................................................................................................35
III.4.1 Materials.................................................................................................................................................35
III.4.2 CatalystPerformanceTests............................................................................................................36
III.5 References..............................................................................................................................................37
IVDeactivationPathwaysofCarbonCatalystsforFructoseDehydration...........39
IV.1 Introduction..........................................................................................................................................40
IV.2 Experimental........................................................................................................................................44
IV.2.1 CatalystPreparation..........................................................................................................................44
IV.2.2 CatalystCharacterization................................................................................................................46
IV.2.3 CatalystPerformanceTests............................................................................................................47
IV.3 ResultsandDiscussion.....................................................................................................................49
IV.3.1 CatalystTestingintheDehydrationofGlucosetoHMF.....................................................49
II
IV.3.2 DehydrationofFructosein2‐butanol........................................................................................55
IV.4 SummaryandConclusion................................................................................................................66
IV.5 References..............................................................................................................................................68
VReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy........70
V.1 Introduction..........................................................................................................................................71
V.2 Experimental........................................................................................................................................75
V.2.1 Materialsynthesisandpreliminarycharacterization.........................................................75
V.2.2 Instrumental.........................................................................................................................................76
V.3 ResultsandDiscussion.....................................................................................................................77
V.3.1 Differencesinoriginalcarbonsamples.....................................................................................77
V.3.2 Differencesduringheatinginvacuum.......................................................................................84
V.3.3 Behaviorduringheatinginwateratmosphere......................................................................87
V.3.4 Climpuritiesandtheirevolutionduringheatandvaportreatment.............................91
V.4 Conclusions...........................................................................................................................................93
V.5 References..............................................................................................................................................95
VIFinalDiscussionandOutlook..........................................................................................98
VIISupplementaryInformation................................................................................................i
VII.1 Appreciations..........................................................................................................................................ii
VII.2 DatabaseSampleNumbers...............................................................................................................iii
VII.3 BETIsotherms........................................................................................................................................iv
VII.4 XPS...............................................................................................................................................................v
III
LISTOFFIGURES
IIntroduction......................................................................................................................................1
Figure1:Schematicofsecondarycell‐wallstructureofgrass[3]..............................................3
Figure2:ImportantproductsfromHMF............................................................................................5
Figure3:Overviewofreactionnetworkforthedehydrationoffructose..........................12
Figure4:DehydrationstepsassuggestedbyNefin1910[22].................................................13
Figure6: Comparisonofpossibleconformationsforglucoseandfructose......................15
Figure5:Lobry‐de‐Bruyn‐Alberda‐van‐Ekenstein–rearrangement....................................16
Figure7:Structuralmodelofhydrothermalcarbonparticles[47].........................................18
Figure8:ProposedmechanismforthecatalyticeffectofDMSO[50]......................................19
IIOutlineoftheWork.....................................................................................................................23
IIIHeterogeneousCatalystsintheDehydrationofFructosetoHMF..............................25
Figure1:CatalyticperformanceofdifferentlyfunctionalizedBaytubecatalysts............28
Figure2:ComparisonofBETisothermsofBTsandMC_H2O2.................................................30
Figure3:Comparisonofdifferentoxidicmaterialsinthedehydrationoffructose
toHMF...............................................................................................................................................................31
Figure4: Characteristic color changes of the catalysts after the fructose
dehydrationreaction..................................................................................................................................32
IVDeactivationPathwaysofCarbonCatalystsforFructoseDehydration....................39
Figure1:Overviewofcarbonmaterialbasis...................................................................................41
Figure2:Comparisonofthedehydrationoffructoseandthereferencereactionthe
esterificationofaceticacidinethanol.................................................................................................43
Figure3:Catalytictestingfacilities.....................................................................................................48
Figure4: Optimal reaction temperature for the glucose dehydration in aqueous
phase.................................................................................................................................................................50
Figure5:G&Fconversion, HMFselectivity,andHMF yield fordifferentreaction
timesat40barN2for10wt%glucosesolutionsinwater.........................................................50
Figure6: Comparison of differently functionalized Baytube® catalysts in the
dehydrationofglucoseinaqueousmedia.........................................................................................52
IV
Figure7:AdsorptionisothermsofglucoseandHMFonBTsindifferentsolvents
after20hat30°C..........................................................................................................................................53
Figure8:HMFyieldoverreactiontimeforthedehydrationofglucoseandfructose
inaqueousphaseincomparisontotheone‐phasesystemin2‐butanol..............................54
Figure9:Schematicreactionassemblyofonephasesystemin2‐butanol.........................56
Figure10:Catalytictestingmodes......................................................................................................57
Figure11:BTscatalystinthedehydrationoffructosein2‐butanol.....................................58
Figure12:BTscatalystintheesterificationofaceticacid........................................................59
Figure13:CatalyticperformanceofBS‐CNFinthedehydrationoffructose.....................60
Figure14:Left:Amberlyst®15(Amb)inthefructosedehydration.Right:Nafion®
(Naf) in the fructose dehydration. Comparison of first run (RUN1),leachingtest
(LEACH)andreuseofthecatalystafterRUN1inarecyclingrun(RECYCL)......................61
Figure15:Left:ComparisonoffirstrunactivitiesofAGPsandBTs.Right:Number
ofacidfunctionalgroupsdeterminedbytitration.........................................................................62
Figure16:Upper:CatalyticperformanceofAGPsafterfourpreconditioningstep
(AGPs_PC4) in dehydration of fructose in 2‐butanol. Lower: Recyclability of
AGPs_PC4intheesterificationreaction..............................................................................................63
Figure17: Left: OMCs in the dehydration of fructose in 2‐butanol after
preconditioning(PC),andrecyclingofthepreconditionedOMCs(RECYCL).Right:
LeachingactivityofOMCsintheesterificationofaceticacid(LEACH).................................64
Figure18:MesoporouscarboncatalystTDP0.2inthedehydrationoffructosein2‐
butanol..............................................................................................................................................................65
Figure19:MesoporousTDP0.2‐catalystintheesterificationofaceticacid......................65
VReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy................70
Figure1: Possible reaction mechanism for theformationoftheresorcinol‐
formaldehydepolymerastheprecursorforOMCsynthesis[5,9]..............................................72
Figure2:Fructosedehydrationinto5‐hydroxymethylfurfuralunderliberationof
3moleculesofwater...................................................................................................................................73
Figure3: Catalytic performance of mesoporous carbon materials in the
dehydrationoffructosetoHMF.............................................................................................................74
Figure4:SchematicdrawingofthehighpressureXPSsystematBESSYII[11]..................77
Figure5:Schemeofpossiblecarbonstructure..............................................................................79
Figure6:O1sfitsofMC_0,MC_1andMC_2ofex‐situXPSmeasurements.........................80
Figure7:C1sfitsofMC_0,MC_1andMC_2ofex‐situXPSmeasurements..........................83
V
Figure8:TG‐MSexperimentofMC_1.................................................................................................85
Figure9:EvolutionofC1scomponentsduringheatinginvacuumandsubsequent
additionofwater(0.1mbar)at130°C................................................................................................87
Figure10:Reactivityofdifferentoxygenfunctionalgroupstowardswater[28]and
consequentialchangesinthesolvationchemistry.........................................................................89
Figure11:EvolutionofC1scomponentsduringheatinginvacuumandsubsequent
additionofwater(0.1mbar)at130°C................................................................................................90
Figure12:InsituXPSspectraofsampleMC1(left)andMC2(right)....................................94
VIFinalDiscussionandOutlook..................................................................................................98
Figure1:Mechanismoffructosedehydrationcatalyzedbyhomogeneousacids..........100
VIISupplementaryInformation........................................................................................................i
Figure1:O1sfits(leftpanel)andC1sfits(rightpanel)ofMC_0,MC_1andMC_2
obtainedbyex‐situXPS................................................................................................................................v
Figure2:FitsforO1sspectraofMC_1andMC_2duringheatinginvacuum.......................vi
Figure3:FitsforC1sspectraofMC_1andMC_2duringheatinginvacuum........................vi
Figure4:FitsforO1sspectraofMC_1andMC_2duringheatingin0.1mbarvapor.......vii
Figure5:FitsforC1sspectraofMC_1andMC_2duringheatingin0.1mbarvapor.......vii
Figure6:Cl2pspectraforMC_1andMC_2duringheatinginvacuum...............................viii
Figure7:Cl2pspectraforMC_1andMC_2duringheatinginvapor....................................viii
VII
LISTOFTABLES
IIntroduction......................................................................................................................................1
Table1:Monosaccharidebuildingblocksofbiomass....................................................................7
IIOutlineoftheWork.....................................................................................................................23
IIIHeterogeneousCatalystsintheDehydrationofFructosetoHMF..............................25
Table1:Overviewofcatalyticperformanceoftestedmaterialsandcomparisonto
literaturereferences...................................................................................................................................29
IVDeactivationPathwaysofCarbonCatalystsforFructoseDehydration....................39
Table1:Overviewofcatalysts...............................................................................................................46
Table2:BETsurfaceareabeforeandafterreaction....................................................................52
VReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy................70
Table1:SummaryofmaterialbasisincludingBETareasandtitrationresults................76
Table2:O1speakassignmentaccordingtotheliterature........................................................78
Table3:QuantificationofoxygenspeciesofMC_0,MC_1andMC_2byex‐situXPS.......81
Table4:C1speakassignmentaccordingtoliterature................................................................82
Table5:QuantificationofcarbonspeciesintheC1spectraofMC_0,MC_1 and
MC_2obtainedbyex‐situXPS.................................................................................................................84
Table6: Quantification of oxygen species of MC_1 and MC_2 during heating in
vacuumandsubsequentadditionofwater(0.1mbar)at130°C.............................................85
Table7: Quantification of oxygen species for MC_1 and MC_2 during heating in
0.1mbarvapor..............................................................................................................................................88
Table8:QuantificationofClspeciesofMC_1andMC_2duringin‐situXPS.......................92
VIFinalDiscussionandOutlook..................................................................................................98
VIISupplementaryInformation........................................................................................................i
Table1:BETisothermsofmesoporouscarbonsamplesinvestigatedbyXPS.....................v
VIII
Table2:QuantificationofcarbonspeciesintheO1speakduringheating in
vacuumandsubsequentadditionofwaterat130°C.......................................................................x
Table3:QuantificationofcarbonspeciesintheC1speakduringheatinginvacuum
andsubsequentadditionofwaterat130°C.........................................................................................x
Table4:QuantificationofcarbonspeciesintheO1speakduringheating in 0.1
mbarvaporpressure....................................................................................................................................xi
Table5:QuantificationofcarbonspeciesintheC1speakduringheating in 0.1
mbarvaporpressure....................................................................................................................................xi
1
I INTRODUCTION
Thediminishingreservesoffossilfuelscoupledwiththesteadygrowthin
theirconsumptionnecessitatetheexplorationofalternativeresources.Inorder
to secure the future supply of fuelsandchemicals,arenewable feedstock is
required to provide the sustainable foundation of prospective processes.
Biomass, as steadily produced by photosynthesis, is not only a renewable
feedstock, it also offers a CO2‐neutral bases for the production of essential
chemicalsandtransportationfuels.However,thechangeinfeedstockrequires
the complete reconstruction of today’s infrastructure and technology. All
establishedprocessesfortherefiningofalkanebasedcrudeoilarefoundedon
theintroductionoffunctionalgroups,mainlybyselectiveoxidation.Theexisting
functionalizationtechnologyneedstobetransferredintostrategiesofselective
oxygenremovalforthehighlyfunctionalizedbiomassfeedstocks.
Starting from the choice of proper raw materials, over economicand
ecologicalcultivation,downtothedevelopmentofnewrefiningstrategies,there
areseveralcomplexchallengesthatrequiremultidisciplinarysolutions.Itisin
theresponsibilityofscientificresearchtocontributetotheimplementationof
necessaryfeedstockchanges,startingfromin‐depthanalysisandthesuccessive
compilationoffundamentalunderstanding.
Introduction
2
I.1 TheFeedstockCellulosicBiomass
The choice of a suitable biomass feedstock for future chemical and fuel
productionhastobetakenwithcare.Politicalcriteria,aswellasethicalissues
andtechnicalfeasibilityneedtobetakenintoaccount.Politicalcriteriacanbe
global dependencies and requirements in development and industrialization.
The ethical arguments concern price competition to food sourcesorthe
questionofsustainabilityofbiomassproductionanduse.Lastbutnotleast,the
questionoftechnicalfeasibilityisimportantforlargescale,industrialprocesses.
Herein, thenumber andcosts of necessaryreactionandpurificationsteps,as
wellasthefinancialinvestmentsforthedevelopmentofnewproductionlines,
comparedtothepriceofthefinalproductplayarole.Sinceesculentplantsare
generally easier to digest and existing technology of food processing can be
used; they are preferred feedstocks according to the argument of technical
applicability. For this reason, the first generationbiofuels are made of sugar,
starchandvegetableoil[1].
Incontrast,secondgenerationbiofuels are produced from sustainable
sources, i. e. available feedstocks preventing impact on greenhouse gas
emission,biodiversityandlanduse.Infact,themostabundantclassofbiomass
raw materials is carbohydrates, contributing 95% of the 200 billion tones of
annualbiomassproductionbyphotosynthesis
[2]. Considering the other
sustainabilitycriteriaofbiodiversityandlandusechange,possible second
generation feedstocks are harvest waste (sugar cane begasse, corn stover),
forestindustrysidestreamsorfastgrowingenergycrops.Themaincomponent
of those materials is lignocellulose which consists of a complex network of
benzylic lignin, hemicellulose and cellulose (compare grass structure
Figure1)[3].
Introduction
3
Figure1:Schematicofsecondarycell‐wallstructureofgrass.Themaincomponentsare
cellulose hemicelluloses and lignin, interconnected by p‐coumaric acid (pCA), ferulic
acid(FA),p‐hydroxybenzoicacid(BA),sinapicacid(SA)andcinnamicacid(CA)[3].
Thecarbohydratefractionbasedonhemicelluloseandcellulosecomprises
hexose‐andpentose‐unitsthatarepolymerizedbytheformationofverystable
glycosidicbonds.Incellulose,β‐(1,4)‐linkedD‐glucoseunitsformlong‐chained
polymers.Throughinterconnectionbyhydrogenbonds,thosecellulosechains
resultinfiberswithpartlycrystallineareas.Hemicellulose,bycontrast,isofa
lessorderedstructure,duetothevarietyofsugarunitsinthepolymerwhich
arecross‐linkedtoformanetworkratherthanlinearstructures.Typicalsugar
Introduction
4
units of hemicellulose are the pentoses xylose and arabinose, as well as the
hexoses glucose, mannose and galactose. Lignin is incorporated into the
carbohydratenetworkofhemicelluloseandcellulose,andleadstothetypical
wooden appearance of plants. All fragments are interconnected by non‐core
lignincomponents,suchasp‐coumaricacid,ferulicacid,p‐hydroxybenzoicacid,
sinapicacidorcinnamicacid.
Innature,thoserobuststructuresgivetheplantstructuralstability.Onlyfew
organisms,suchasruminantanimals,areabletobreakthestructure,hydrolyze
glycosidicbondsandmetabolizethemonosaccharidesfromcellulosicbiomass.
This indicates the difficulty of controlled cleavage of lignocellulose into well
definedbuildingblocks.Sincelignocelluloseistheonlybiomassfeedstockthat
canbeexploitedinsustainablemannerinlargescale,themainchallengeofits
industrial use is the development of an adequate refining technology. This
impliestheimprovementofenzymaticandchemicalhydrolysisstrategiesonthe
onehand,andprocessesfortheselectiveconversionoflignocellulosesderived
buildingblocksontheotherhand.
I.2 HMFSynthesisfromBiomass
Themaincarbohydrateunitsoflignocellulose,aswellasoffirstgeneration
feedstocks, are summarized in Table1. They provide a pool of potential
buildingblocksforthesynthesisofchemicalsandfuels.Oneimportantproperty
ofthematerialsisthehighoxygencontent.Theconsequenceofthehighoxygen
contentisthelowerheatingvalue,whichisdisadvantageousfortheapplication
asfuels.Additionallythehighdensityoffunctionalgroupsresultsinahighand
Introduction
5
unspecificreactivityofmonosaccharides.Thusthemajorchallenge is the
selective,chemicalremovalofoxygenwhichisefficientlydoneinacidcatalyzed
dehydrationreactions.
One important dehydration product is 5‐hydroxymethyl furfural (HMF).
HMFcanbefurtherconvertedtodimethylfuran(DMF)andusedasafuelorfuel
additive[4].Inaddition,HMFisanimportantintermediatefortheproductionof
polymers[5],e.g.informofthefurtheroxidizedderivatefurandicarboxylicacid
(FCDA)whichisseenaspotentialalternativetoterephthalicacidderivedfrom
crudeoilprocessing.Anotherprocessalreadyappliedinindustrialscaleisthe
use of the HMF‐derivate 3,5‐dihydroxy methylfuran by Oaker Oatsforthe
production of polyurethane foams[6]. Furthermore, essential future platform
chemicals can be produced ofHMF,suchaslevulinicacid
[7]orpyrrols
[8]
(Figure2).
Figure2:ImportantproductsfromHMF
Inadditiontothedirectindustrialrelevanceofthereactionstudied, the
dehydrationoffructoseisonegeneralpathwayofeffectiveoxygenremovalin
biomassconversionchemistry[9].HMFhasconsequentlybeenlistedas oneof
the “key substance between carbohydrate chemistry and mineral oil based
chemistry”[10].Thereforethisreactionwasalsostudiedasanimportantmodel
Introduction
6
reactionintheextensionandtheimprovementofalreadyexistingconceptsof
biorefinery[11].
Duetothereasonsmentioned,thenumberofpublicationsonHMFsynthesis
increasedexponentiallywithinthelastdecade.Greateffortshavebeenmadeto
studyavarietyofpossiblefeedstocks.Thespectrumrangesfrom
monosaccharide and polysaccharide, over cellulose and hemicellose up to
biomassrawmaterials,suchasawoodchipsorstraw.Thelatterrequirespecial
digestion or extraction mechanisms. One approach is to apply ball milling to
mechanocatalytically breaknaturalcellulosesources
[12] prior to further
dehydration steps. However, even by addition of a homogeneous acid
(p‐toluenesulfonic acid, p‐TSA) the mechanically assisted depolymerization
processyieldsonlyintolessthat0.3%furfuralproducts[13].Sofarballmilling
studies have aimed the production of soluble, short‐chained polysaccharides,
andtheremightbemorepotentialforthismethodtoassistdirectproductionof
HMFfromnaturallignocellulosesources.
Themostpromisingpossibilitytoefficientlydissolvelignocellulosicbiomass
and convert it directly to HMF or HMF‐derivatives is the use of ionic liquids
(ILs).BinderandRainesreportedtheuseofN,N‐dimethylacetamide(DMA)in
combination with 1‐ethyl‐3‐methylimidazolium chloride ([EMIM]Cl), lithium
chloride(LiCl)andCrCl3forthedirectproductionofHMFfromcornstoverina
singlereactionstep[14].Inapplyingmicrowaveheating(400Wfor3min),Zhang
etal.wereabletoobtain45‐52%HMFyieldsfromcornstalk,ricestrawand
pinewood[15].Thesolventusedinthatworkwas1‐butyl‐3‐methylimidazolium
chloride([BMIM]Cl),againtogetherwithCrCl3.Thusthereactionmixturemight
containtoxicchromiumspecies.Inaddition,therearethegeneralproblemsof
productseparationandpurification,aswellasthehighsolventcosts.Onthis
account,thedirectconversionofrawbiomassintoHMFcanbestillseenasina
“proof‐of‐principle”‐state.
Introduction
7
Table1:Monosaccharidebuildingblocksofbiomass
Introduction
8
In principal, there are similar procedures applied while using purified
cellulose as feedstock for HMF synthesis. In combination with column
chromatographyonsilicagelisolatedHMFyieldsof61%wereobtainedfrom
cellulose[16]. Here 10 wt% CrCl3 in [BMIM]Cl were irradiated by microwave
heatingat 400 W for2min.Other groups, such asValenteetal.[17],reported
loweryieldsforHMFproductionfromcelluloseinsimilarcomplex systems.
TheydidnotdetectHMFformationafter4hat100°C(oilbathheating),while
using a 2‐phase mixture of [BMIM]Cl and methyl isobutyl ketone (MIBK)
together with CrCl3 as catalyst. The addition of 1‐ethyl‐3‐methylimidazolium
hydrogen sulfate ([EMIM][HSO4])leadsto9%HMFyieldafter4hat100°C.
BinderandRainescouldincreasetheHMFyieldto53%bytheadditionofHCl
usingoilbathheatingat140°C for1 h[14]. Although the reaction systems are
similar,adirectcomparisonofthoseresultsisdifficultduetothedifferencesin
reaction conditions applied. In general, it can be concluded that microwave
heatingandhighioncontentsfavorhighHMFyieldsfromcellulose[16].
One example for an ionic‐liquid‐free method for the direct conversion of
cellulose to HMF was reported by Chareonlimkun et al.[18]. They used hot
compressedwater(HCW)andaheterogeneousZrO2‐TiO2catalyst.Astainless
steelreactorwasfilledby0.1gcelluloseand1mlwater,andheatedto250°C
for5min.DuetoconstantN2‐pressureof34.5MPawaterisinthestateofliquid
phase during the experiment. Although this method avoids the used of toxic
additivesand expensive ionic liquids, theobtained HMFselectivity of 13%at
70%conversionisratherlow.
Theselectedexamplesontheuseofrawbiomassorcelluloseshouldshow
that,despitethepartiallypromisingHMFyields,thecurrentlyrequiredreaction
conditions do not allow a sustainable synthesis. Highly complexproduction
routes, including toxic additives,expensivesolventsandimpractical product
separation, are required. Additionally, the systematic variation of reaction
Introduction
9
conditions,inordertovalidatetheinfluenceoftheindividualparametersand
provideacomparablebasisofreactivityresultsremainstobecarriedout.
Duetothereasonsmentioned,allfollowingconsiderationswillfocusonthe
useofthemonosaccharidesglucoseandfructoseasinitialproduct for the
synthesisofHMF.Inparallelcomprehensiveresearcheffortsareobservablein
theselectivebreakageofrawbiomassintomonosaccharidebyenzymaticand
chemicalhydrolysis,aswellasmechanocatalyticalballmillingprocess.
I.3 MechanismofFructoseDehydrationtoHMF
As described above, sugars are complex structures containing reactive
carbonylandhydroxylfunctionalgroupsinhighdensity.Inbiologicalsystems
sugarscanbeselectivelyconvertedbyperfectlydesignedenzymesundermild
conditions.Theformationofbyproductsissuppressedbytheconstrainedlocal
environmentoftheactivesite.Chemicalsugarconversionhowever,suchasan
acidic treatment of aqueous fructose solutions, leads to a complex reaction
networkofmanypossiblereactionroutesandbyproducts.Drivenbydifferent
motives, various scientists triedtomapthisreactionnetworkover the last
century.Inthissection,asummaryoftheirfindingsiscomprisedinanoverall
reaction scheme. Secondly, the literature basis for the currentmechanistic
understanding is summarized, in order to distinguish between the scientific
levelofknowledge,ongoingdebates,andremainingqueriesdowntothepresent
day.
Introduction
10
I.3.1 Overviewschemeofreactionnetwork
Anoverviewofthereactionnetworkoftheacidcatalyzeddehydration of
fructoseintoHMFisdepictedinFigure2.Inthecenteroftheschemestands
the overall reaction, indicated bytheboldarrows.Fructosereacts over
intermediates to the dehydration product HMF. In aqueous media HMF can
further react via re‐hydration levulinic acid and formic acid, again passing
certain intermediate steps. The explanatory boxes (dotted) for the
intermediatescontainthemechanismsdiscussedintheliterature. For the
dehydration of fructose to HMF two general pathways are reported (upper
dottedbox).Theupperone,comprisingtheintermediates1‐5, maintains the
five‐memberedringstructureoffructofuranose(FαforFβf).Inthefirststep,
fructoseisprotonatedintheC‐2–OHposition.TheformedC‐2–OH2+splitsoff
water as a good leaving group. Thusthe first dehydration step results in the
fructofuranosyl cation (2).Atthesametime,allinitialstereoisomeric
informationinC‐2positionislost,asindicatedbywavylines.Bythefollowing
deprotonation,theenolintermediate3evolveswhichisinequilibriumwiththe
corresponding keto form 4. Subsequently, either the C‐4–OH or the C‐3–OH
position can be dehydrated by water abstraction, resulting in 5aor5b,
respectively. Finally the last dehydration step, which is the only irreversible
stepinthemechanism,resultsintheformationofaromaticHMF.
Alternatively,thedehydrationcanfollowanopen‐chainmechanismoverthe
acyclic keto‐form of fructose (6). Thereby 6isinequilibriumtotheenol‐
tautomer7andcanbedehydratedattheC‐3–OHposition.A3‐deoxyhexosulose
intermediate is formed, labeled as 8. The subsequent dehydration at C‐4–OH
positionresultsintheunsaturatedosone9.Finally,theabstractionofthethird
watermoleculeleadstotheformationofthefive‐ringstructureofHMF.
Introduction
11
InaqueousmediathedehydrationproductHMFcanfurtherreactwithwater
over the intermediates 11‐15 (lower explanatory box). The “rehydration”
reactionisaswellcatalyzedbytheacidabundantinthereactionmixture.The
formaladditionofonewatermoleculeresultsinthecleavageofformicacid,and
furtherwateradditiontointermediate15resultsintheformationoflevulinic
acid.BothlevulinicacidandformicacidleadtoaloweringofthetotalpHofthe
solutionandautocatalyzethedecompositionoffructose.
Amajorcontributiontotheoverallreactionnetworkcomesfromunwanted
byproducts. They can be classified into soluble and insoluble polymerization
products. Soluble polymers can be formed by intermolecular condensationof
fructose.Thereactionoftwofructosemoleculesunderwaterabstractionresults
inthedimerdi‐fructose‐di‐anhydride.Theformationofinsoluble byproducts,
so‐calledhumins,canproceedoveroligomerizationoffructosewithitselfand
with HMF. Furthermore side products of levulinic acid formation(17) can
undergounwantedpolymerizationreactions.Generally,allintermediatescarry
multiplefunctionalgroupspredestinatedfortheformationofsideproductsand
humins, in particular in acidic environment. Thus the process of humin
formationishighlycomplexandathroughoutmechanisticunderstandingofthis
processisstillmissing.
Introduction
12
Figure3:Overviewofreactionnetworkforthedehydrationoffructoseunderacidic
conditionsinaqueousmedia.Theoverallreaction( )appearsinthemiddleofthe
scheme.TheexplanatoryboxesshowthemechanismofHMFformation(upperdotted
box)andthere‐hydrationofHMFtolevulinicacidandformicacid(lowerdottedbox).
Introduction
13
I.3.2 Historicdevelopmentofmechanisticideasonfructosedehydration
Studiesonacidcatalyzedsugardegradationhavebeenintheliteraturefor
morethan100years.Theformationoflevulinicacidandformicacid,aswellas
afurfuralproducthasbeenreportedbyDüllandKiermayer,alreadyin1895[19‐
21].Theyellowfurfuralproductwasdescribedbyanaromaticodorofoverripe
apples.Thefunctionalgroups were identified byFehling’ssolutionandSchiff
test(aldehyde),aswellasbyreactionwithanilineinaceticacid(furfural).The
reactionwithphenylhydrazineresultsinahydrazonewithm.p.=138°C(vs.96°C
forfurfuralhydrazone).Theelementalanalysisofthehydrazone gave a
composition of C6H6O3forthefurfuralproduct,whichconsequentlycanbe
formedfromfructosebytheabstractionof3moleculesofwater[21].Kiermayer
assignedtheproductto3‐hydroxy‐5‐methylfurfural[19].
Figure4:DehydrationstepsassuggestedbyNefin1910[22].
ThecorrectidentificationofthestructureofthefurfuralproductasHMFwas
given by Blanksma and van Ekenstein[23, 24] by analogy observations to
Introduction
14
dehydration products of chitose (5‐(hydroxymethyl)‐tetrahydrofuran‐2,3,4‐
triol) in combination with further specific color reactions. Based on the
identificationofHMFasthedehydrationproduct,Nefsuggestedamechanismin
1910(Figure4)[22].Nefalreadyproposedcyclicintermediatessimilarto5a,
compare Figure2 upper reaction mechanism. One general argument for a
dehydrationmechanismovercyclicintermediates(compare1‐5,Figure2) is
thehigheractivityoffructoseincomparisontoglucose.Blanksma and van
Ekensteinfoundonly1%yieldofHMFfromglucosewhereasfructosegave20‐
25%yieldunderthesamereactionconditions[24].Kiermayer[19],aswellaslater
Haworth and Jones[25] found that in acidic treatments of sucrose only the
fructose half reacts to form HMF, the glucose half remains unreacted in the
reaction mixture. As already summarized inthe previous chapter the highest
HMFyieldshavebeenachievedbyeitherusingfructoseasadirectfeedstockor
facilitatingisomerizationtofructosepriortotheactualdehydrationstep[26,27].
Since the five‐membered ring structure of fructofuranose is close to the
suggestedfructofuranosylcation(2),themeasurableadvantageoffructosein
comparison to glucose supports the cyclic mechanisms. Although fructose
occursinthe5‐ringstateonlytoanextentof24%at20°C,thefuranosecontent
increaseswithtemperature(at80°C42%furanose[28]).
Inthecaseofglucosenoneofthepossiblecyclicconformersprovides a
suitablestructuralelementforadirectHMFformation(Figure6).Consequently
anepimerizationisnecessarydescribedastheLobry‐de‐Bruyn‐Alberda‐van‐
Ekenstein–rearrangement(Figure5)[29].Theepimerizationbetweenthealdose
glucoseandtheketosefructoseistypicallybasecatalyzed,andpassestheopen
chain conformation of glucose (0.002% at 31°C). The enediol intermediate is
alsodiscussedasfirstpossibleintermediate7(Figure2)intheacyclicreaction
mechanism.
Introduction
15
Figure6: Comparison of possible conformations for glucose and fructosewiththeir
abundanceinaqueoussolution[30]:a)31°C,glucose[31,32];b)44°C,glucose[31,32];c)27°C,
fructose[32,33];d)80°C,fructose[28]
Introduction
16
Figure5:Lobry‐de‐Bruyn‐Alberda‐van‐Ekenstein–rearrangement
Theacyclicreactionmechanism,i.e.anotherpossiblemechanismviaopen
chain indermediates (compound 6‐8, Figure2), has first been discussed by
HurdandIsenhourin1932[34].Itisbasedonthewellknown,base‐catalyzedβ‐
eliminationofthehydroxylgroupviaenediolintermediates(7),asdiscussedfor
the epimerization. Since fructose forms a more stable open chain conformer
than glucose (Figure5),thehigherabundanceofthesamecanbean
explanationforthehigherdehydrationactivityoffructose.Thephenylosazone
derivateof9hasbeenisolatedbyWolfrometal.[35]. The same group further
proceededUVabsorptionstudiestoaddfurtherevidencetotheabundanceof
acyclicintermediates[36].However,thosestudiesdidcompareabsorptionbands
whicharenotpartofthe UV spectrumof 3‐deoxyhexosulose(8)as critically
discussed later[37, 38]. Anet et al. also supported the idea of a dehydration
mechanismoveracyclicintermediates[39, 40].Theywereabletoisolateosazone
derivativesof3‐deoxyhexulose(8)and3,4‐dideoxyhexuloseintermediates(9)
bypapercolumnchromatographyandidentifiedthesamebyNMR
[40].Later
isotopeexchangestudieswereperformedinD2O[41, 42]showingthattheHMF
product does not contain any deuterium incorporation. Since theacyclic
mechanism includes several equilibria with enolic tautomers (7, 9), carbon‐
linked deuterium is expected in the final product. Thus the isotope exchange
experimentsinD2Osupporttheideaofthecyclicmechanism[38].Thequestion
remained how measurable amounts of deuterium free HMF can be obtained
Introduction
17
from glucose, since here an epimerization is the first requirement for HMF
formation and in basic media deuterium incorporation was detected.[42]The
answerwasgivenbyHarrisandFeather[43]whoobservedanintramolecularC‐2
C‐1hydrogentransferinthedehydrationofC‐2tritiumlabeled glucose
(glucose‐2‐3H).However,anintramolecularhydrogentransferC‐3C‐2,which
couldproveanacyclicmechanismintheabsenceofdeuteriumincorporation,
has not been presented down to the present day. In 2008, Yaylayan et al.
reportedarelativeefficiencyinHMFformationforglucose,3‐desoxyglucosone
(9)andfructoseof0.16:1:2.4[44].Hence9cannotbethemainprecursorof
HMFfromfructose;otherwiseitwouldhaveleadtomoreHMFthaninthecase
relativetoglucose.
The present compilation of literatureonsugardehydrationstrongly
supportsthecyclicmechanismoverthefuranosylcation.Ifatall,theopen‐chain
mechanismonlyplaysaminorroleintheformationofHMFfromfructose.
Despitethesignificantinfluenceofinsolublepolymerizationproductsonthe
HMF selectivity, very little is known on the formation mechanism of humins.
One possible reaction, which can lead to the formation of humins, is the
intermolecular condensation of fructose[45, 46]. Furthermore, side reactions of
fructose with the product HMF are reported[46], such as the acetalization
betweenthealdehydefunctionalgroupofHMFandhydroxylfunctionalgroups
ofthesugar.Moregeneral,Kusteretal.mentionedthatsidereactionsbetween
all intermediates contributetothehuminformation
[26]. More recently, the
structural analysis of hydrothermally synthesized carbon materials from
glucosegavefurtherinsightintheconstitutionofhumins(Figure7)[47]. The
resultsofthe13Csolid‐stateMASNMRexperimentscanbedirectlyappliedon
thestructuredeterminationofhumins,sincethehydrothermalcarbonsynthesis
fromglucoseisperformedunderverysimilarconditionsascommon
dehydrationreactions,i.e.10wt%glucosesolutioninwaterat180°Cfor24h.
Introduction
18
Onlythechoiceoflongerreactiontimesprovokesthepreferentialformationof
solidcarbonmaterials,insteadofmaximizingHMFyields.Themainstructural
motifdetectedforthehydrothermalcarbonwaswithapproximately65%the
furanring,originatedfromHMF.About23%wasassignedtosp3carbonandthe
remainingfractionattributedtoC=Oandresidualglucose.Consequently, the
NMR results support the idea of a complex polymerization procedure with
major contribution of HMF. However, the question of the origin of the
interconnectingcarbonchainscontaining23%sp3carbonandabout13%C=O
remainsunsolved.Hencetheelucidationofthehuminformationmechanismis
stillopentoprospectiveresearchactivities.
Figure7:Structuralmodelofhydrothermalcarbonparticlesaccordingto[47]
Introduction
19
I.3.3 Mechanisticsolventeffects
The(ring)structureandconformationoftheinitialsugarapparentlyplaya
majorroleinthefacilenessofthedehydrationreaction.Sincesolventshavethe
property to influence the structural appearance of sugars, there have been
several mechanistic considerations on solvent effects. The mostprominent
example is DMSO. Dais and Perlin[48] observed a higher content of furanose
conformersoffructoseinDMSOcomparedtowateralreadyat20°C,with20%
α-furanose(Fαf),55%β‐furanose(Fβf)andonly26%β‐pyranose(Fβp).Itis
known that intramolecular hydrogen bonds are stronger in DMSO than in
water[49]. However a direct connection between stronger intramolecular
hydrogenbondsandconformericpreferencescannotbeconcludedbecausethe
influenceofhydrogenbondsissimilarinpyranoseandfuranoseconformers.An
explanation for the distribution of fructose conformers in DMSO is still
outstanding. Amarasekara et al.[50]foundafurtherdecreaseofFβpdownto
16%at150°Candproposedamechanismforthecatalyticeffectof DMSO
beyond the furanose stabilization (Figure8).Additionally they couldidentify
thecyclicintermediate5b(compareFigure3)bycombinationof1Hand13C
NMRdataandtherewithprovidefurtherevidenceforthecylicmechanism.
Figure8:ProposedmechanismforthecatalyticeffectofDMSOonthedehydrationof
fructose[50]
Introduction
20
I.4 References
[1] European Biofuels Technology Platform (EBTP), BiofuelProduction,
http://www.biofuelstp.eu/fuelproduction.html,17thNovember2011.
[2] N. E. Bassam, International Research Center for Renewable Energy (IFEED),
2009.
[3] J.Bidlack,M.Malone,R.Benson,ProceedingsoftheOklahomaAcademyofScience
1992,72,51‐56.
[4] Y.Roman‐Leshkov,C.J.Barrett,Z.Y.Liu,J.A.Dumesic,Nature2007,447,982‐
U985.
[5] A.Gandini,M.N.Belgacem,ProgressinPolymerScience1997,22,1203‐1379;C.
Moreau,M.N.Belgacem,A.Gandini,TopicsinCatalysis2004,27,11‐30.
[6] W.J.Pentz,inPatent,C08G1814ed.(Ed.:Q.O.Company),QuakerOatsCompany,
US,1984.
[7] A.Corma,S.Iborra,A.Velty,ChemicalReviews2007, 107, 2411‐2502; G.
Dautzenberg,M.Gerhardt,B.Kamm,Holzforschung2011,65,439‐451.
[8] F.W.Lichtenthaler,A.Brust,E.Cuny,GreenChemistry2001,3,201‐209.
[9] R.Rinaldi,F.Schuth,Energy&EnvironmentalScience2009,2,610‐626.
[10] H.Schiweck,M.Munir,K.M.Rapp,B.Schneider,M.Vogel,Zuckerindustrie1990,
115,555‐565.
[11] B.Kamm,P.Gruber,M.Kamm,Biorefineries‐IndustrialProcessesandProducts,
Vol.2,Wiley‐VCH,Weinheim,2005.
[12] S.M.Hick,C.Griebel,D.T.Restrepo,J.H.Truitt,E.J.Buker,C.Bylda,R.G.Blair,
GreenChemistry2010,12,468‐474;S.VandeVyver,J.Geboers,P.A.Jacobs,B.F.
Sels,Chemcatchem2011,3,82‐94.
[13] N. Meine, J. Hilgert, R. Rinaldi, F. Schüth, in XV.InternationalSymposiumon
RelationsbetweenHomogeneousandHeterogeneousCatalysis,Berlin,2011.
[14] J. B. Binder, R. T. Raines, JournaloftheAmericanChemicalSociety2009, 131,
1979‐1985.
[15] Z.Zhang,Z.K.Zhao,BioresourceTechnology2010,101,1111‐1114.
[16] C.Z.Li,Z.H.Zhang,Z.B.K.Zhao,TetrahedronLetters2009,50,5403‐5405.
Introduction
21
[17] S.Lima,P.Neves,M.M.Antunes,M.Pillinger,N.Ignatyev,A.A.Valente,Applied
Catalysisa‐General2009,363,93‐99.
[18] A. Chareonlimkun, V. Champreda, A. Shotipruk, N. Laosiripojana, Bioresource
Technology2010,101,4179‐4186.
[19] J.Kiermayer,Chem.Zeit.1895,19,1003‐1005.
[20] G.Düll,C.J.Lintner,Chemiker‐Zeitung1895,19,216‐217.
[21] G.Düll,C.J.Lintner,Chem.Zeit.1895,19,216‐217.
[22] J.U.Nef,JustusLiebigsAnnalenDerChemie1910,376,1‐119.
[23] W.A.vanEkenstein,J.J.Blanksma,RecueilDesTravauxChimiquesDesPays‐BasEt
DeLaBelgique1906,25,153‐161.
[24] W.A.vanEkenstein,J.J.Blanksma,BerichtederdeutschenchemischenGesellschaft
1910,43,2355‐2361.
[25] W.N.Haworth,E.L.Hirst,V.S.Nicholson,JournaloftheChemicalSociety1927,
1513‐1526;W.N.Haworth,W.G.M.Jones,JournaloftheChemicalSociety1944,
667‐670.
[26] B.F.M.Kuster,Starch‐Starke1990,42,314‐321.
[27] S. H. Khajavi, Y. Kimura, R. Oomori, R. Matsuno, S. Adachi, JournalofFood
Engineering2005,68,309‐313.
[28] W.Funcke,C.vonSonntag,C.Triantaphylides,CarbohydrateResearch1979,75,
305‐309.
[29] S.Angyal,A.Stütz,Vol.215,SpringerBerlin/Heidelberg,2001,pp.1‐14.
[30] M. Jetzki, R. Signorell, JournalofChemicalPhysics2002, 117, 8063‐8073; H. E.
Vandam,A.P.G.Kieboom,H.Vanbekkum,Starch‐Starke1986,38,95‐101.
[31] S.J.Angyal,V.A.Pickles,AustralianJournalofChemistry1972,25,1695‐&.
[32] L.D.Hayward,S.J.Angyal,CarbohydrateResearch1977,53,13‐20.
[33] S.J.Angyal,G.S.Bethell,D.E.Cowley,V.A.Pickles,AustralianJournalofChemistry
1976,29,1239‐1247.
[34] C.D.Hurd,L.L.Isenhour,JournaloftheAmericanChemicalSociety1932,54,317‐
330.
[35] M. L. Wolfrom, E. G. Wallace, E. A. Metcalf, JournaloftheAmericanChemical
Society1942,64,265‐269.
Introduction
22
[36] M. L. Wolfrom, R. D. Schuetz, L. F. Cavalieri, JournaloftheAmericanChemical
Society1948,70,514‐517;M.L.Wolfrom,R.D.Schuetz,L.F.Cavalieri,Journalof
theAmericanChemicalSociety1949,71,3518‐3523.
[37] J.Lewkowski,Arkivoc2001,2;G.Machell,G.N.Richards,JournaloftheChemical
Society1960,1938‐1943.
[38] M.J.Antal,W.S.L.Mok,G.N.Richards,CarbohydrateResearch1990,199,91‐109.
[39] E. F. L.J. Anet, J.Am.Chem.Soc.1960,82,1502; E. Anet, AustralianJournalof
Chemistry1960,13,396‐403;E.F.L.Anet,AustralianJournalofChemistry1962,
15,503‐&;E.Anet,AdvancesinCarbohydrateChemistry1964,19,181‐218.
[40] E.F.L.Anet,AustralianJournalofChemistry1965,18,240‐&.
[41] G.Fodor,J.P.Sachetto,TetrahedronLetters1968,401‐&.
[42] M.S.Feather,J.F.Harris,CarbohydrateResearch1970,15,304‐&.
[43] D.W.Harris,M.S.Feather,JournalofOrganicChemistry1974,39,724‐725.
[44] C.P.Locas,V.A.Yaylayan,JournalofAgriculturalandFoodChemistry2008,56,
6717‐6723.
[45] Y.D.Chu,K.A.Berglund,Starch‐Starke1990,42,112‐117.
[46] Y.Roman‐Leshkov,J.N.Chheda,J.A.Dumesic,Science2006,312,1933‐1937;B.F.
M.Kuster,H.S.Vanderbaan,CarbohydrateResearch1977,54,165‐176;P.Vinke,
H.vanBekkum,Starch‐Starke1992,44,90‐96.
[47] N.Baccile,G.Laurent,F.Babonneau,F.Fayon,M.M.Titirici,M.Antonietti,Journal
ofPhysicalChemistryC2009,113,9644‐9654.
[48] P.Dais,A.S.Perlin,CarbohydrateResearch1987,169,159‐169.
[49] S.J.Angyal,J.C.Christofides,JournaloftheChemicalSociety‐PerkinTransactions2
1996,1485‐1491.
[50] A.S.Amarasekara,L.D.Williams,C.C.Ebede,CarbohydrateResearch2008,343,
3021‐3024.
23
II OUTLINE OF THE WORK
The literature on HMF synthesis is bulky and showed various different approaches
for the common target reaction of sugar dehydration. The majority of research efforts
went into catalyst screening of commercially available materials under different
reaction conditions. However, since there is no common agreement on the conditions of
catalyst testing, a detailed comparison of the results is hardly possible. Furthermore the
literature is leaking on detailed characterization of the catalytically active materials.
Consequently questions on catalyst stability or the structure of the actual active
component remain unsolved. Ideas on both aspects would be needed for a more
directed synthesis of the catalysts as the basis of improvements in the final HMF yields.
The highly diverse literature basis can be seen as the natural response on the
complexity of the reaction. Since the dehydration of sugars implies a broad reaction
network, the elucidation of the reaction mechanism, in particular the function of the
heterogeneous catalyst is by no means trivial. The intention of this work is to provide a
tiny but solid stepping stone on the long way of mechanistic understanding of HMF
synthesis on heterogeneous catalysts. Under this aspect one class of heterogeneous
catalysts, i. e. carbon based catalysts, has been studied in depth in the dehydration of
glucose and fructose into HMF. In a first process, suitable reaction conditions and
catalytic testing modes have been elaborated, in order to compare performance and
stability of the different carbon material. Subsequently, the effect of the carbon
Outline of the Work
24
structure and the influence of the functionalization method on stability and
performance have been investigated. The stability of the catalyst was tested in terms of
resistance of the functional groups against applied reaction conditions, site blocking of
the active component by humin accumulation and structural changes of the catalyst by
solvent or heat treatment. The latter was further investigated by in-situ XPS
experiments.
25
III HETEROGENEOUS CATALYSTS IN THE
DEHYDRATION OF FRUCTOSE TO HMF
Authors: Sylvia Reiche, Edward Kunkes, Nuruzatulifah Bt. Asari Mansor,
Xiao Chen Zhao, Koteswara Rao Vuyyuru, Alberto Villa, Jean-
Philippe Tessonnier, Dangsheng Su, Malte Behrens, Peter Strasser,
Robert Schlögl
Abstract
Several heterogeneous catalysts have been reported in the dehydration of
fructose to 5-hydroxymethyl furfural (HMF). Despite the obvious arguments for
the preferential use of heterogeneous catalyst, their performance is lower than
for the further optimized homogeneous system. In addition, the direct
comparison of literature results is hindered by the huge variety of reaction
conditions applied. We tested multiple promising heterogeneous catalyst in a
work-up-friendly one-phase system in 2-butanol and compared performance as
well as recyclability.
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
26
III.1 Introduction
The discovery of 5-hydroxymethyl furfural (HMF) and first mechanistic
ideas were reported more than 100 years ago[1]. Still, an inexpensive and
sustainable production route from biomass derived feedstocks remains a
challenge to current research. Even though, the awareness of future feedstock
changes in the chemical industry and HMF being a key molecule[2] in the pool of
biomass derived building blocks lead to an exponential increase in research
interest. Recently published reviews[3, 4] summarized the most critical points in
HMF synthesis by deficient mechanistic understanding of the complex reaction
network, the low reactivity of glucose in comparison to the less abundant
isomer fructose and the leak of efficient separation methods, due to the
unfavorably high hydrophilicity of HMF[3]. Groundbreaking progress was
achieved in the optimization of homogeneously catalyzed fructose dehydration
in the development of efficient biphasic systems[5, 6]. The addition of phase
modifiers, such as inorganic salts, typically NaCl[6], or polar aprotic solvents,
such as dimethyl sulfoxid (DMSO) or 1-methyl-2-pyrrolidinone (NMP), and/or
the hydrophilic polymer poly(1-vinyl-2-pyrrolidinone) (PVP) into the aqueous
phase and 2-butanol to the organic phase further enhanced the HMF yield[5].
However, with every additive used the purification of HMF is hampered. Hence,
practicable solutions for the complete elimination of impurities in the product
stream are still outstanding.
More efficient would be the direct use of heterogeneous catalysts. Also here,
intensive research efforts have been reported in the literature[7-9]. The type of
heterogeneous catalysts applied ranged from oxides[10, 11], over polymer resin
down[9, 12-14] to sulfonated carbon materials[15, 16]. However a direct comparison
of the data is difficult due to the variety of different reaction conditions applied.
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
27
For this reason, we retested acidic heterogeneous catalysts that gave promising
results in the literature, among them, the zeolithes H-mordenite and H-ZSM5, as
well as other oxidic materials, such as niobium oxide and sulfated zirconia.
Furthermore, niobium phosphate and vanadyl pyrophosphate were tested, as
well as the ion exchange resins Amberlyst® 15 and Nafion®. Finally the study
involves carbon catalysts based on functionalized multi-walled carbon
nanotubes (MWCNTs)[17], amorphous carbon from glucose pyrolysis (AGP)[18]
and ordered mesoporous carbon (OMC) based on polymeric precursors[19]. The
catalyst screening was performed, in order to achieve comparable data as the
basis for further studies in the direction of a better understanding of the active
surface, structural requirements of the materials and ideally the final
implementation of the findings into improvements of the catalyst synthesis. By
systematic re-run (= recycling), it was checked for often neglected irreversible
catalyst modifications.
III.2 Results
For the present study, we moved from an aqueous system into a one-phase
system in 2-butanol. This has the advantage that part of possible side reactions,
e. g. the re-addition of water to HMF under the formation of levulinic acid and
formic acid, are suppressed. Furthermore, the final extraction of the product is
not necessary and impurities due to phase modifiers can be avoided. The
advantages mentioned prevail the lower solubility of fructose in 2-butanol and
the use of lower concentrated feed solutions.
The fructose dehydration experiments were performed in a batch reactor at
130°C. For a typical experiment, 2.5 g fructose were added to 100 ml 2-butanol
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
28
and 250 mg catalyst. The results are summarized in Table 1. Comparing the
MWCNT-based catalysts (Figure 1), we could observe the highest activity for
the Baytubes ® functionalized in sulfuric acid (BTs) with 28% yield of HMF
after 3 h reaction time. The catalysts obtained by the functionalization with
boronic acid (BTb), as well as the gas phase functionalized carbon nanotubes
(BTsG200, BTsG600) did not exhibit high dehydration activities. In the recycling
run, i. e. the reuse of the material in a second experiment, the HMF yield using
BTs decreased to 4%. Hence the material deactivated strongly already after one
reaction run.
Figure 1: Catalytic performance of differently functionalized Baytube catalysts in the
dehydration of fructose into HMF (130°C, one phase system 2-butanol). At t = 0.33 h the
reaction temperature of 130°C is reached
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
29
Table 1: Overview of catalytic performance of tested materials and comparison to literature
references
Catalyst BET surface
area [m²/g]
Yield HMF [%]
(Recycling Run)
r
0
*
�
µ
mol HMF
2h∙m
²
�
Reference HMF Yield [%] / conditions
BT0 227 0 0
-
BTs 245 28 (4) 56.1
BTsG200 258 8 17.4
BTsG600 265 2 2.9
BTb 253 2 2.5
AGPs nonporous 50; 37
a
(5
a
) - -
OMCs 683 17 7.8
-
MC_TDP 654 61 (17) 51.5
MC_H2O2 596 59 (34) 55.7
Nafion® flexible
structure,
swellable in
liquid phase
36 (28) -
[9] 94 / 0.3 g F, 10 g DMSO, 0.02 g cat.,
120°C, 2 h
Amberlyst® 15
50 -
[9] 92 / 0.3 g F, 10 g DMSO, 0.02 g cat.,
120°C, 2 h
[13] 87 /
Sulfated
zirconia 424 11 12.6
[9] 92 / 0.3 g F, 10 g DMSO, 0.02 g
catalyst, 120°C, 2 h
[20]
68 / 0.045 g F, 1 ml DMSO, 0.018 g
cat, 130°C, 4 h
H-mordenite 387 4 5.7
[10, 11] 69 / 1-7 g F, 0.5-2 g cat., 35 ml
H2O, 175 ml MIBK, 10 bar N2,
165°C, 60 min
H-ZSM5 418 27b; 14c 8.6
[10] 53 / 1-7 g F, 0.5-2 g cat., 35 ml H
2
O,
175 ml MIBK, 10 bar N2, 165°C,
60 min
Nb2O5 n. d. 1 -
[21] 29 / 10 wt% F in H
2
O, 0.7-1.7 g,
100°C, 30 min
NbOPO4 141 24 (6) 89.0
[22]
23 / 6 wt% F in H2O, 0.7 g, 100 °C
(VO)2P2O7 10.8 6 239.6
[23]
42 / H2O, 50°C, 1 h
All catalytic tests were performed in a one-phase system of 2-butanol at 130°C. The yields
(calculated per mol initial fructose) refer to the amount of HMF formed after t = 3 h.
r0* specific apparent initial rate
a HMF yield for the preconditioned catalyst (catalyst was pretreated 4 times in 2-butanol
at 130°C for 15 h each)
b 500 mg catalyst
c 1 g catalyst
F fructose
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
30
The further tested mesoporous materials AGPs, MC_TDP and MC_H2O2
exhibit even higher HMF yields, up to 61% for MC_TDP, in the first reaction run.
However, also here the catalysts could not preserve their activity in the
recycling run. Comparing the specific apparent initial rate of the graphitic
Baytube-catalyst BTs with the amorphous, mesoporous material MC_H2O2, we
can conclude similar activities of both materials when normalizing to the
surface area. However, the BET isotherms show the significant differences in the
pore structures of the two materials. The direct comparison according the
surface area, neglects the influence of the pore structure on the catalytic
performance, as well possible changes in the surface area due to different
swelling behavior of material during the reaction.
Figure 2: Comparison of BET isotherms of BTs and MC_H2O2. From the graph it can be
concluded that the MC_H2O2 catalyst has the higher amount of mesopores, as well as
micropores.
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
31
The activity losses in case of the ion exchange resins are less pronounced
comparing the first run and the recycling run. In case of Nafion® the HMF yield
decreases from 36% to 28% (t = 3 h). In comparison to the carbon catalysts that
are naturally black, the ion exchange resins change their color and appearance
upon catalytic testing (Figure 4). Considering the dramatic impact of the
reaction conditions on the nafion beads, the long-term stability of the material is
questionable.
Figure 3: Comparison of different oxidic materials in the dehydration of fructose to
HMF. For the H-ZSM5 catalyst 500 mg (=2xm) and 1 g (=4xm) were used. For all other
materials 250 mg were tested. At t = 0.33 h the reaction temperature of 130°C is
reached.
For the white oxidic catalysts similar darkening of the material can be
observed (Figure 4). The color change is associated with the often reported
formation of insoluble humins[8, 24], side products of the fructose dehydration
that can be accumulated onto the catalysts surface during the reaction.
Comparing the catalytic activity of the different oxide based catalysts
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
32
(Figure 3), the niobium phosphate catalyst provided the highest HMF yields per
mass catalyst. By increasing the mass of H-ZSM5 to 1 g, the HMF yield could also
be pushed to 27% (t = 3 h). However, the elevated HMF formation is combined
with more pronounced humin accumulation, as observable in the brownish
color of the material (Figure 4).
Figure 4: Characteristic color changes of the catalysts after the fructose dehydration
reaction. a) Nafion before reaction (photography), b) Nafion beads after reaction
(photography), c) one Nafion bead at magnification of 200x (optical microscope), d) H-
ZSM5 before reaction, e) H-ZSM5 2xm after reaction, f) H-ZSM5 4xm after reaction (d-f:
optical microscope, magnification 50x)
III.3 Discussion
In summary, the tested heterogeneous catalysts showed either low
performance from the very beginning or did not maintain their catalytic activity
in the recycling run. Comparing our findings to the results in the literature
b
e)
d
a)
c)
f)
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
33
(Table 1), we can conclude that in principal higher HMF yields are possible
using the same heterogeneous catalysts. However, in most cases reported in the
literature, the use of heterogeneous catalysts is combined with solvents of
intrinsic catalytic activity such as DMSO[9] or ionic liquid additives[25]. Another
possibility to enhance the HMF yields is the parallel extraction using a 2-phase
system of water with methyl isobutyl ketone (MIBK)[10, 11]. The latter can be
seen as an elegant option for the longterm use of heterogeneous catalysts.
However, the method would not be compatible with catalysts of a highly
hydrophobic surface, such as carbon based catalysts, because of their
preferential abundance in the organic phase. Hence, the use of an organic
extraction solvent would not only hinder the contact between the catalyst and
the initial product, the extended impact of the catalyst on the HMF product,
could even provoke side reactions. Furthermore, the use of an additional
extracting solvent, in general, is accompanied by the same technical difficulties
as described for the homogeneously catalyzed reactions.
The other critical point is the recycling of the material. Often only the first-
run-performance is reported leaving questions on the catalyst stability open to
the reader. Recently, the first review which includes the comparison of
recyclability of the materials was published by Rosatella et al.[3]. Only for 18 out
of 100 entries on the use of heterogeneous catalysts in the production of HMF,
data on the recyclability could be found. Most commonly the recycling capability
is reported for ion exchange resins, where our results also point to relatively
good recyclability. However, special attention is required when comparing the
reaction times and the fructose to resin ratio. If the material was reused 5 times
after a reaction time of 5-20 min per run and a fructose to resin ratio of one or
smaller, it is hard to predict the longterm stability of the material[14].
Another option of possible recycling of oxidic catalysts was reported by Yan
et al.[20] for sulfated zirconia catalysts. They demonstrated the successive
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
34
deactivation of the material over four dehydration runs in DMSO. However, by
calcination of the catalyst at 550°C for 2 h after each dehydration run, the
catalytic activity could be partially recovered. The most significant loss in
activity was observed in the first recycling (about one quarter lower HMF yield).
In subsequent reaction runs the activity was preserved by calcination.
Based on the catalytic data presented and the critical comparison of the
same to the literature, it can be concluded that deeper studies of the catalyst
stability are essential, in order to excel established homogeneous solutions and
to guarantee a long-term application. So far, little is known on possible
structural changes of the catalyst in the course of the fructose dehydration
reaction. Hence, only post-treatments can be applied for the reactivation of the
catalyst, such as calcination[20], extractive washings by different solvents[21, 26] or
recovery of surface functional group by re-treatments in diluted mineral
acids[27]. Consequently, it is the future aim of our research to further elucidate
possible deactivation mechanisms, in order to comprise structural requirements
on the heterogeneous catalyst.
The promising results on carbon based catalysts in first dehydration
experiments, strongly suggest the further exploration of the same. The
hydrothermal stability and structural diversity of carbon catalysts provide a
prospective basis, widely applied in other biomass relevant reactions, such as
esterification[18] or cellulose hydrolysis[16, 28].
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
35
III.4 Experimental Section
I.4.1 Materials
Baytubes®(BT) were purchased from Bayer® and functionalized in
concentrated sulfuric acid (97% Merck, 500 ml for 10 g BT) or nitric acid (65%
Merck, 500 ml for 10 g BT) at 100°C for 20 h. Subsequently, the material was
filtered, intensively washed by water until the pH of the washing solution was
neutral. The material was redispersed in water and stirred over night, in order
to completely desorb remaining acid. Finally the resulting sulfuric acid treated
Baytubes® (BTs) or nitric acid treated Baytubes® (BTn) were dried at 100°C
for 10 h. For the preparation of boronic acid functionalized Baytubes (BTb), 10 g
of BTn were impregnated with 0.75 M H3BO3 by incipient wetness
impregnation. Subsequently, the material was heat treated in Ar at 1500°C for
4h.
For the preparation of the gas phase functionalized sulfonated Baytubes, 3 g
BTn were treated in oleum saturated argon atmosphere at 200°C (BTsG200) or
600°C (BTsG600) for 20 h. In order to guarantee an optimal mixing during the
functionalization process, a horizontal rocking furnace was used, providing an
agitated sample bed[29]. The oleum saturation was achieved by leading the Ar
carrier gas through 50 ml oleum solution (20% fuming oleum in concentrated
sulfuric acid).
Sulfonated amorphous carbon material was prepared as reported by Hara et
al.[18]. The functionalization of the amorphous carbon obtained by glucose
pyrolysis (AGP), was performed as described for BTs. The resulting material is
sulfonated amorphous carbon from glucose pyrolysis (AGPs).
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
36
For the catalysts based on ordered mesoporous carbon (OMC), 5.5 g of
resorcinol and 5.0 g of formaldehyde solution (37 wt%) undergo an acid
catalyzed polymerization in 50 ml of a water-ethanol-mixture (1:1 by weight).
The obtained polymer is carbonized in N2 atmosphere at 350°C and 600°C for
2h, respectively. The detailed procedure of OMC synthesis is described
elsewhere[19]. Subsequently, 2 g OMC are stirred in 100 ml H2SO4 (97%) at
100°C for 20 h, resulting in sulfuric acid treated OMC, denoted as OMCs. To
prepare TDP-donated S-containig mesoporous carbon, 20 mol% of resorcinol is
replaced by 4,4’-thiodiphenol (TDP). After the carbonization at 350 and 600°C,
the TDP containing mesoporous carbon is oxidized by H2O2 in a mixture of
methanol and 2 M HCl (1:1, V:V) to obtain MC-TDP[19].
The ion exchange resins Amberlyst® 15 and Nafion® NR50 were purchased
from Sigma Aldrich. Additional commercial catalysts were sulfated zirconia
(MEL Chemicals), H-mordenite (Südchemie), H-ZSM5 (Degussa), niobium oxide
(Roth) and niobium phosphate (CBMM). The vanadyl pyrophosphate material
was produced by Ecole Polytechnique Montréal.
I.4.2 Catalyst Performance Tests
The catalytic tests were performed in a one-phase system in 2-butanol. For a
typical experiment 2.5 g fructose were loaded together with 100 ml 2-butanol. If
not stated differently 250 mg catalyst were added and the system was heated
under stirring (450 rpm) to 130°C. All reactions were performed in a 200 ml
Parr® autoclave using a teflon liner and a teflon-coated stirrer.
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
37
III.5 References
[1] J. Kiermayer, Chem. Zeit. 1895, 19, 1003-1005; J. U. Nef, Justus Liebigs Annalen
Der Chemie 1910, 376, 1-119.
[2] H. Schiweck, M. Munir, K. M. Rapp, B. Schneider, M. Vogel, Zuckerindustrie 1990,
115, 555-565.
[3] A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso, Green Chemistry
2011, 13, 754-793.
[4] M. J. Climent, A. Corma, S. Iborra, Green Chemistry 2011, 13, 520-540; X. L. Tong,
Y. Ma, Y. D. Li, Applied Catalysis a-General 2010, 385, 1-13; K. D. Vigier, F.
Jerome, in Carbohydrates in Sustainable Development Ii: a Mine for Functional
Molecules and Materials, Vol. 295, 2010, pp. 63-92.
[5] Y. Roman-Leshkov, J. N. Chheda, J. A. Dumesic, Science 2006, 312, 1933-1937; J.
N. Chheda, Y. Roman-Leshkov, J. A. Dumesic, Green Chemistry 2007, 9, 342-350.
[6] Y. Roman-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature 2007, 447, 982-
U985.
[7] L. Cottier, G. Descotes, Trends Heterocycl. Chem. 1991, 2, 233-248.
[8] B. F. M. Kuster, Starch-Starke 1990, 42, 314-321.
[9] K. Shimizu, R. Uozumi, A. Satsuma, Catalysis Communications 2009, 10, 1849-
1853.
[10] P. Rivalier, J. Duhamet, C. Moreau, R. Durand, Catalysis Today 1995, 24, 165-171.
[11] C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier, P. Ros, G.
Avignon, Applied Catalysis a-General 1996, 145, 211-224.
[12] Y. Nakamura, S. Morikawa, Bulletin of the Chemical Society of Japan 1980, 53,
3705-3706; J. N. Chheda, J. A. Dumesic, Catalysis Today 2007, 123, 59-70.
[13] C. Lansalot-Matras, C. Moreau, Catalysis Communications 2003, 4, 517-520.
[14] X. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Jr., Green Chemistry 2008, 10, 799-
805; X. H. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Industrial & Engineering
Chemistry Research 2008, 47, 9234-9239.
[15] S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato, S. Hayashi, M. Hara,
Solid State Sciences 2010, 12, 1029-1034.
Heterogeneous Catalysts in the Dehydration of Fructose to HMF
38
[16] S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato, S. Hayashi, M. Hara,
Journal of the American Chemical Society 2008, 130, 12787-12793.
[17] J. P. Tessonnier, D. Rosenthal, T. W. Hansen, C. Hess, M. E. Schuster, R. Blume, F.
Girgsdies, N. Pfander, O. Timpe, D. S. Su, R. Schlogl, Carbon 2009, 47, 1779-1798.
[18] A. Takagaki, M. Toda, M. Okamura, J. N. Kondo, S. Hayashi, K. Domen, M. Hara,
Catalysis Today 2006, 116, 157-161.
[19] X. C. Zhao, A. Q. Wang, J. W. Yan, G. Q. Sun, L. X. Sun, T. Zhang, Chemistry of
Materials 2010, 22, 5463-5473.
[20] H. P. Yan, Y. Yang, D. M. Tong, X. Xiang, C. W. Hu, Catalysis Communications 2009,
10, 1558-1563.
[21] C. Carlini, M. Giuttari, A. M. R. Galletti, G. Sbrana, T. Armaroli, G. Busca, Applied
Catalysis a-General 1999, 183, 295-302.
[22] T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A. M. R. Galletti, G. Sbrana, Journal of
Molecular Catalysis a-Chemical 2000, 151, 233-243.
[23] C. Carlini, P. Patrono, A. M. R. Galletti, G. Sbrana, Applied Catalysis a-General
2004, 275, 111-118.
[24] J. Horvat, B. Klaic, B. Metelko, V. Sunjic, Croatica Chemica Acta 1986, 59, 429-
438; J. Horvat, B. Klaic, B. Metelko, V. Sunjic, Tetrahedron Letters 1985, 26,
2111-2114; S. K. R. Patil, C. R. F. Lund, Energy Fuels 2011, 25, 4745-4755.
[25] X. H. Qi, M. Watanabe, T. M. Aida, R. L. Smith, Green Chemistry 2009, 11, 1327-
1331.
[26] A. Takagaki, M. Ohara, S. Nishimura, K. Ebitani, Chemical Communications 2009,
6276-6278.
[27] F. S. Asghari, H. Yoshida, Carbohydrate Research 2006, 341, 2379-2387.
[28] A. Onda, T. Ochi, K. Yanagisawa, Topics in Catalysis 2009, 52, 801-807.
[29] N. Giliard, Technische Universität Berlin (Berlin), 2011.
39
IV DEACTIVATION PATHWAYS OF CARBON
CATALYSTS FOR FRUCTOSE DEHYDRATION
Authors: Sylvia Reiche, Xiao Chen Zhao, Matthew Aronson, Klaus Friedel,
Edward Kunkes, Jean-Philippe Tessonnier, Dangsheng Su, Malte
Behrens, Sharifah Bee Abdul Hamid, Robert Schlögl
Abstract
A variety of acid functionalized carbon based catalysts has been synthesized and
tested for the dehydration of fructose to 5-hydroxymethyl furfural (HMF). The
stability of the materials has been studied under three different aspects: Firstly,
the stability of the acid functional groups was evaluated in leaching tests.
Secondly, the influence of surface poisoning was determined by comparing the
catalytic performance in the dehydration of fructose to the esterification of
acetic acid in ethanol. The latter was used as reference reaction, due to the
milder reaction conditions and the obviation of humin formation. Thirdly, we
tested for the influence of the reaction solvent by preconditioning the catalyst
prior to the actual reaction.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
40
IV.1 Introduction
As extensively reported in previous chapters, several crucial steps in the
upgrading of biomass-derived raw materials involve acid catalyzed reactions.
The use of homogeneous acids involves the contamination of the product with
the corrosive catalyst, as well as the introduction of other impurities, such as
chlorine or sulfur containing compounds from counter ions of mineral acids.
Hence, thorough product separation is required to avoid on the one hand
consecutive site reactions, and to eliminate potential catalyst poisons for further
processing steps. The required product separation is challenging and cost
intensive. In the case of sugar dehydration to HMF, more advanced biphasic
processes yielded promising results[1, 2]. However, also here product separation
is limited by partition coefficients between the reaction solvent water and the
extraction solvent, e. g. 2-butanol or methyl isobutylketon (MIBK). Therefore,
such a process would involve a solvent recovery and purification system, if
implemented on a large scale. Thus heterogeneous catalysts are desired in large
scale industrial processes. The application of heterogeneous catalysts is limited
by several challenges concerning the conversion of highly functionalized
feedstocks. All the more, a deeper understanding of heterogeneous catalysts,
their relevant properties for this highly complex reaction systems, catalyst
stability and causes of deactivation are essential for the development of
sustainable processes.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
41
Figure 1: Overview of carbon material basis
The focus of this study is on carbon based heterogeneous catalysts. Carbon
catalysts themselves can be produced from biomass feedstocks on a sustainable
basis[3, 4]. Furthermore, carbonized structures are hydrothermally stable, which
is critical for water based processes and a major advantage in comparison to the
more commonly used metal oxide based catalysts. In addition, carbon provides
a broad structural diversity. Bulky solvated biomolecules need wide pores for
preferential mass transport[5]. One suitable concept of carbon structures was
suggested by Lefferts and Schouten[6, 7] who described carbon nanofibers (CNF)
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
42
as an inverse of the conventional porous catalyst. Large surfaces provided by
the outside surface of tubular CNFs, which is more easily accessible than long
narrow pores. Another concept for suitable carbon structures can be found in
mesoporous carbon. By altering the choice of the template the pore size can be
adjusted according to the requirements from the reaction network. Within this
work, both, carbon nanotubes (CNTs) as well as ordered mesoporous carbon
(OMC) catalysts were applied. Finally, an unstructured amorphous carbon from
glucose pyrolysis (AGP) completes the carbon basis of the catalysts tested in this
chapter. A summary of major properties of the pristine carbon materials is given
in Figure 1.
Carbon catalysts offer a huge variety of functionalization possibilities[8, 9].
They can be acidified by treatments in concentrated acid, as applied in this
work. They can also be used as a catalyst support; for instance for metal
nanoparticles allowing for the synthesis of multifunctional catalysts. By
chemical grafting it is possible to immobilize functional groups more selectively
on the carbon structures. In the case of graphitic structures, such as CNTs, a
selective backbone functionalization is only possible if relatively harsh
conditions are applied. Examples are flourination treatments followed by
substitution reactions or diazotizations. The latter technique was also applied
here.
The acidified carbon catalysts were tested for glucose and fructose
dehydration reactions in water as well as in 2-butanol. The esterification of
acetic acid in ethanol was used as a reference reaction. Both reactions, the
fructose dehydration in 2-butanol and the esterification of acetic acid, are acid
catalyzed and form water as byproduct. Additionally both reactions use an
alcoholic solvent; in case of the esterification the ethanol is solvent and reactant
at the same time. Based on these commonalities (Figure 2) the esterification is
considered to be a suitable reference reaction. The main difference between the
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
43
esterification and the fructose dehydration is given by the reactivity of the initial
product. Since carbohydrates carry multiple functional groups, they can react in
acidic aqueous media under the formation of various side products in a complex
reaction network[10, 11]. As side product formation can be excluded for the
esterification reaction, the stability of the acid functional groups and the catalyst
against heat (80°C), the alcoholic solvent and the water formed in the course of
the reaction, can be determined.
a)
b)
commonalities
acid catalyzed reactions
alcoholic solvent
byproduct water is formed
differences
reaction temperature 130°C
reaction temperature 80°C
multiple functional groups in initial product
monofunctional initial product
complex reaction network
single reaction step, known mechanism
Figure 2: Comparison of the dehydration of fructose (a) and the reference reaction, the
esterification of acetic acid in ethanol (b)
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
44
IV.2 Experimental
IV.2.1 Catalyst Preparation
Baytubes®(BT) were purchased from Bayer® and functionalized in
concentrated sulfuric acid (97% Merck, 500 ml for 10 g BT) or nitric acid (65%
Merck, 500 ml for 10 g BT) at 100°C for 20 h. Subsequently, the material was
filtered, intensively washed by water until the pH of the washing solution was
neutral. The material was redispersed in water and stirred over night, in order
to completely desorb remaining acid. Finally the resulting sulfuric acid treated
Baytubes® (BTs) or nitric acid treated Baytubes® (BTn) were dried at 100°C
for 10 h. For the preparation of boronic acid functionalized Baytubes (BTb), 10 g
of BTn were impregnated with 0.75 M B(OH)3 by incipient wetness
impregnation. Subsequently, the material was heat treated in Ar at 1500°C for
4h.
Highly graphitized carbon nanofibers (CNF) were purchased from Applied
Sciences, Inc. (ASI, material PR24XT-HHT). For the grafting of benzene sulfonic
acid onto the CNF-surface a diazotization was applied, similar to the one
reported by Dyke et al.[12]. Therein, 2.50 g PR24XT-HHT (0.21 mol) were stirred
together with 1.80 g sulfanilic acid (10.4 mmol) in a 1 l three necked round
bottom flask. Subsequently, 300 mL of 1,2-dichlorobenzene were added and the
mixture was sonicated for 20 min in order to disperse the solids in the organic
solution. The suspension was heated to 60 °C while vigorously stirred at 400
rpm. Subsequently, 2.80 mL of isoamyl nitrite (20.8 mmol) were added slowly
via a dropping funnel into the flask and the mixture was further stirred at 60 °C
for 16 h. After cooling to 45°C, 300 ml dimethylformamide were added and the
suspension was sonicated for 10 min. The solid was filtered off and washed
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
45
thorough by DMF and dichloromethane. In order to remove unreacted, low
soluble sulfanilic acid, the sample was stirred 2-time in 300 mL NaOH (1 M) for
30 minutes while being heated to 60 °C at 400 rpm. For neutralization, the
sample was stirred in 500 mL 1 M HCl solution over night. Finally the material
was filtered, washed with deionized water until the filtrate showed neutral pH
and dried at 110 °C to obtain benzene sulfonic acid grafted CNF (BS-CNF).
Sulfonated amorphous carbon material was prepared as reported by Hara et
al.[3]. The functionalization of the amorphous carbon obtained by glucose
pyrolysis (AGP), was performed as described for BTs. The resulting material is
sulfonated amorphous carbon from glucose pyrolysis (AGPs).
For the catalysts based on ordered mesoporous carbon (OMC), 5.5 g of
resorcinol and 5.0 g of formaldehyde solution (37 wt%) undergo an acid
catalyzed polymerization in 50 ml of a water-ethanol-mixture (1:1 by weight).
The obtained polymer is carbonized in N2 atmosphere at 350°C and 600°C for
2h, respectively. The detailed procedure of OMC synthesis is described
elsewhere[13]. Subsequently, 2 g OMC are stirred in 100 ml H2SO4 (97%) at
100°C for 20 h, resulting in sulfuric acid treated OMC, denoted as OMCs.
To prepare TDP-donated S-containig mesoporous carbon, 20 mol% of
resorcinol is replaced by 4,4’-thiodiphenol (TDP). After the carbonization at 350
and 600°C, the TDP containing mesoporous carbon is oxidized by H2O2 in a
mixture of methanol and 2 M HCl (1:1, V:V) to obtain MC-TDP[13].
The ion exchange resins Amberlyst® 15 and Nafion® NR50 were purchased
from Sigma Aldrich.
An overview of the catalysts tested is given in Table 1.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
46
Table 1: Overview of catalysts
Sample
Material description
BET [m²/g]
Acid sites
[mmol/g]
BT0
pristine Baytubes® from Bayer
293a, 227b
-
BTs
Baytubes® functionalized by H2SO4
304a, 245b
0.24a,c, 0.09b,c
BTn
Baytubes® functionalized by HNO3
326a
0.19a,c
BTb
Baytubes® functionalized by B(OH)3
253a
BS-CNF
PR24XT-HHT from Applied Sciences,
Inc. (ASI) grafted by benzene sulfonic
acid
36.1
AGP
amorphous carbon from glucose
pyrolysis
-
AGPs
amorphous carbon from glucose
pyrolysis functionalized by H2SO4
-
1.7c,d, 0.42c,e
OMC
ordered mesoporous carbon
789
-
OMCs
ordered mesoporous carbon
functionalized by H2SO4
683
0.45c
MC_TDP
mesoporous carbon, 20mol% of
building block recorcinol replaced by
thiodiphenyl
654
0.25c
Naf
Nafion® NR50 from Aldrich
Amb
Amberlyst® 15 from Aldrich
3.7c
a first batch of Baytubes®, b second batch of Baytubes®, c density of acid sites determined by titration of
the material (100 mg in 50 ml KCl 0.001 M) with 0.01 M NaOH, d titration of material as synthesized, e
titration result of pretreated material (4x 15 h at 130°C in 2-butanol)
IV.2.2 Catalyst Characterization
The BET surface areas were determined by measuring the N2 adsorption-
desorption-isotherms with a Quantachrome Autosorb automatic BET-
sorptometer at -196°C with nitrogen as analysis gas.
SEM micrographs were acquired using a Hitachi S-4800 microscope. For the
acquisition of TEM micrographs a Philips CM200 LaB6 microscope and a FEI
Quanta 200 FEG was used.
The number of acid functional groups was determined by titration using an
automatic titrator (Mettler Toledo). For each measurement, 100 mg catalyst
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
47
were dispersed in 10-3 mol KCl-solution and stirred over night. Following the
dispersion was titrated under Ar atmosphere, using a 0.1 M NaOH solution.
IV.2.3 Catalyst Performance Tests
Catalytic test reactions in water were performed in a fully automated reactor
system from Cambridge Reactor Design (CRD) with 5 independent 100-ml batch
reactors made of hastelloy (Figure 3a). In a typical experiment 50 ml of an
aqueous solution of 10 wt.% glucose (Sigma Aldrich) were stirred at 450 rpm
together with 250 mg catalyst at 40 bar N2 pressure.
For the catalyst performance tests in 2-butanol a 200-ml Parr autoclave
(Series 4560) made of stainless steel using a Teflon liner and a Teflon-coated
stirrer was used (Figure 3b). Fructose dehydration reactions consisted of 2.5 g
fructose (Sigma Aldrich) in 100 ml 2-butanol. The reactions were conducted at
130°C and 7 bar N2 pressure. Sampling was performed through an integrated
sampling tube. All samples were filtered with a 0.2-µm syringe filter and diluted
by water (1:50) prior to analysis.
For the analysis of sugars and HMF by liquid chromatography an Agilent
1200 HPLC instrument was used. Fructose and glucose were detected with a
refractive index detector (RID), and HMF was detected with a UV–Vis detector
(variable wavelength detector, VWD) at 284 nm (absorption maxima HMF[14]).
The column used was a Rezex™ RHM-Monosaccharide from Phenomenex®. The
method includes a mobile phase of 0.005 M H2SO4 with a flow rate of 0.6
mL/min at 80°C column temperature.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
48
Figure 3: Catalytic
testing facilities: a) 12 x
100 ml autoclaves in
automated reactor
system from CRD, b)
200 ml Parr reactor, c)
300 ml Premex reactor
The esterification reactions were carried out in a 300-ml batch reactor by
premex reactor ag®(Figure 3c). Therein, 146 ml ethanol (2.5 mol) and 1.4 ml
acetic acid (0.025 mol) were stirred at 800 rpm at 80°C. An integrated sampling
tube allowed direct sampling. All samples were analyzed by GC-MS using an
Agilent GC (6890 N) coupled to an Agilent 5975B MS detector. For the
quantification of ethanol, acetic acid and ethyl acetate a flame ionization
detector (FID) was used.
a)
b)
c)
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
49
IV.3 Results and Discussion
IV.3.1 Catalyst Testing in the Dehydration of Glucose to HMF
The dehydration of glucose in water was studied in a fully automated reactor
system developed for high-throughput screening. Since the single batch
reactors did not provide a possibility of sampling during the reaction, we
needed first to choose proper reaction conditions for a qualitative comparison
of the catalysts. For the determination of an optimal reaction temperature, 5
experiments at 130°C, 150°C, 170°C, 180°C and 190°C were performed in 40 bar
N2 for 2 h. For each experiment, the autoclaves were loaded by 50 ml 10 wt%
glucose solution together with 250 mg BTs catalyst. The optimal temperature
was determined to be 180°C for the maximum HMF yield after 2 h reaction time
(Figure 4). For better comparison we combined the quantities of glucose and
fructose, as determined by HPLC analysis, in order to plot a total sugar
conversion (G & F conversion, G… glucose, F… fructose). At 130 °C no HMF
formation could be observed, merely a glucose isomerization into fructose was
detected after 2 h reaction time. With further increase in temperature, the
activity towards glucose dehydration generally increased. However, for
temperatures higher than 180°C the selectivity to HMF drops, most likely due to
the prevailing of the formation of secondary products. This is in agreement with
studies from the literature that state that the formation of these insoluble
humins is favored at high reaction temperatures.
For the choice of a proper reaction time in the batch reactor system, three
separate experiments with of 20 min-, 60 min- and 120 min-duration were
performed. As shown in Figure 5 the HMF selectivity is decreasing over time
with increasing G&F conversion. As reaction time, we have chosen 2 h for the
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
50
following comparison of the differently functionalized carbon nanotube
catalysts, since the HMF yield is still in an increasing regime and the conversion
of < 70% allows a reasonable comparison of the different catalysts.
Figure 4: Determination of the optimal reaction temperature for the glucose
dehydration in aqueous phase (t = 120 min, 40 bar N2, 10 wt% glucose solution, 250 mg
BTs catalyst). The sugar conversion was summarized in G&F conversion (-♦-), which is
calculated based on the sum of glucose and fructose concentrations measured, and
plotted together with the HMF selectivity (···■···) and yield (--▲--).
Figure 5: G&F conversion (-♦-), HMF selectivity (···■···), and HMF yield (--▲--) for
different reaction times at 40 bar N2 for 10 wt% glucose solutions in water and 250 mg
BTs catalyst
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
51
Subsequently, we compared differently functionalized CNT catalysts in the
dehydration of a 10 wt% glucose solution in water (Figure 6). Bronsted acid
functional groups had been incorporated into the carbon surface by nitric acid-
(BTn), sulfuric acid- (BTs) and boronic acid- (BTb) treatments. Furthermore
iron nanoparticles were deposited by incipient wetness. Iron which is surface
oxidized into iron oxide under ambient conditions can react as a Lewis acid. Fig
x. depicts the sum of glucose and fructose conversion, as well as the HMF
selectivities after 2 h reaction time at 180°C. All materials tested show a higher
activity of G&F conversion compared to the BLANK experiment, i. e. a reaction
performed under the same conditions without the addition of a heterogeneous
catalyst. The greatest difference in sugar conversion and HMF selectivity
compared to the blank experiment (45 % conversion, 26 % selectivity) was
observed for the CNT catalyst treated in nitrid acid (53 % conversion, 24 %
selectivity). The activity generally was measured to be increasing in the order
blank < BTn < BTs ≈ BTb. Comparing the equally active materials BTs and BTb
in terms of selectivity towards HMF, we see a higher HMF selectivity for the
sulfuric acid functionalized BTs. One reason for this tendency could be the
occurrence of Lewis acid sites for the BTb sample. Lewis acid sites have been
reported before to be active in the dehydration of sugars. However, they can
catalyze more side reactions, e. g. the formation of humins from the reaction of
furfural products with the initial sugars[15]. Therefore the total furfural yield
decreases with the amount of Lewis acid functional groups of heterogeneous
catalysts. A similar behavior of higher activity but lower HMF selectivity could
be observed for the materials after the deposition of iron. Another observation
which agrees with the findings by Weingarten et al. [15].
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
52
Figure 6: Comparison of differently functionalized Baytube® catalysts in the
dehydration of glucose in aqueous media after 120 min reaction time at 40 bar N2 and
250 mg catalyst
Although the activity of glucose dehydration could be enhanced by the
heterogeneous catalysts we tested in aqueous phase, the selectivities of our
targeted Product HMF were poor. In order to elucidate the reasons for the low
selectivity of the carbon catalysts, all materials were characterized by BET and
SEM, and compared to the unused catalysts. The analytical techniques applied
clearly showed the coverage of the high surface area catalysts by insoluble
byproducts. This leads to coverage of surface functional groups, i. e. the active
sites in the reaction, as well to a tremendous decrease of catalyst surface area. In
the case of the BTs-catalyst the surface area decreased by 98% from 351 to
7 m²/g (Table 2).
Table 2: BET surface area before and after reaction
Sample
before ABET [m²/g]
after ABET [m²/g]
BTs
351
7.25
BTn
326
26.6
BTb
253
32.0
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
53
One reason for enhanced humin accumulation onto the catalyst surface can
be related to the surface properties of the catalysts. Since carbon nanotubes
consist of a graphitic carbon structure with highly hydrophobic properties, this
could favor high product reactivity on the catalyst surface. In order to study this
phenomenon, adsorption experiments of glucose and HMF in aqueous medium
as well as in 2-butanol were performed (Figure 7). 2-butanol is a more
hydrophobic solvent that is widely used as extraction solvent in biphasic
reaction systems in sugar dehydration processes [1, 2, 16]. For this reason, we
have chosen 2-butanol for comparison to aqueous system. In Fig. x. the HMF and
the glucose coverages are plotted as a function of their concentrations. HMF in
aqueous solution adsorbs with a capacity that is one order of magnitude greater
than that of HMF in 2-butanol solution. In contrast, the glucose adsorption onto
the hydrophobic carbon surface is small.
Figure 7: Adsorption isotherms of glucose and HMF on BTs in different solvents after
20 h at 30°C.
Based on these adsorption results, we performed further catalytic tests in
sugar dehydration both in aqueous phase as well as in a one-phase system in
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
54
2-butanol. Due to low reactivity of glucose in the direct dehydration to HMF, we
extended the catalytic tests towards the more reactive initial product fructose.
The reason for the tremendous difference in dehydration activity of both sugars
is originated from structural properties of the two isomers. The furanose form
of fructose directly provides the structural element to follow the cyclic
mechanism over the furanosyl cation. In contrast, glucose requires an
epimerization over the acyclic conformer which is abundant in only diminutive
quantities (< 0.005 %).
The comparative dehydration experiments between glucose and fructose
were performed at 150°C in a 200 ml Parr® autoclave. A sampling valve allowed
sampling during the catalytic reaction. Figure 8 shows the HMF yield versus
reaction time for the BLANK reaction compared to the use of 250 mg BTs
catalyst in the dehydration of glucose in 2-butanol, fructose in 2-butanol and
fructose in water, respectively. Under these reaction conditions glucose is
essentially inactive. For fructose dehydration, the use of 2-butanol instead of
water as solvent increases the HMF yield significantly.
Figure 8: HMF yield over reaction time for the dehydration of glucose and fructose in
aqueous phase in comparison to the one-phase system in 2-butanol at 150°C,
respectively
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
55
Based on these results, we changed our testing conditions into a one-phase
system of 2-butanol and focused on the dehydration of fructose to HMF, in order
to achieve a better comparison of the catalysts.
IV.3.2 Dehydration of Fructose in 2-butanol
In the following comparison of different acidified carbon catalysts, the 2-
butanol one phase system was used. The use of 2-butanol as the only reaction
solvent has the disadvantage that the fructose solubility is low in comparison to
that in water. For this reason we used 2.5 wt% solutions of fructose. However,
the change in reaction conditions has several advantages. First, the product
adsorption onto the catalyst surface is lower, as described above. Furthermore,
side reactions that require water as reactant are suppressed. Thus the
formation of levulinic acid and formic acid can be avoided which themselves
lead to a lowering in the pH and continuously increase the amount of catalyst;
making the reaction in this state difficult to control. The disadvantages of water
as reaction solvent have been widely discussed in the literature [17, 18] and a wide
range of different solvents have been tested for the dehydration of
monosaccharides[19]. Most effective apart from ionic liquids were DMSO, DMF or
biphasic systems using methyl isobutylketon (MIBK) or 2-butanol as extraction
solvents. However, all those solvents have certain disadvantages. Most
challenging is the product separation of the reaction mixture due to the high
boiling points, e.g. for DMSO of 189°C[10, 20-23]. Furthermore the processes lead to
impurities in the HMF-product that can be problematic in subsequent
processing steps, such as S- and N-containing solvent or salt residues from the
biphasic process. Since hydrogenolysis as well as oxidation reactions require
metal containing catalysts, that are highly sensitive against poisoning by the
impurities named, costly purification processes would be required.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
56
In the one-phase system in 2-butanol, 2.5 g fructose were loaded together
with 100 ml 2-butanol. If not stated differently 250 mg catalyst were added and
the system was heated under stirring (450 rpm) to 130°C. All reactions were
performed in a 200 ml Parr® autoclave. During the reaction the solid fructose
dissolves in 2-butanol prior to dehydration into HMF. The schematic reaction
assembly is depicted in Figure 9.
Figure 9
: Schematic reaction assembly
of one phase system in 2-butanol
In order to test for catalyst stability, four different testing modes were
applied (Figure 10). First the catalyst was tested in a standard experiment as
describes above (RUN1). Following the catalyst was filtered off and reused in a
second experiment under identical reaction conditions, which is further
referred to as recycling run (RECYCL). The stability of immobilized functional
groups was tested in a separate leaching test (LEACH). Here the catalyst was
stirred in the reaction solvent 2-butanol at 130°C for 15 h. Subsequently, the
catalyst was filtered off and 2.5 g fructose were added to the 2-butanol filtrate. If
the catalyst is stable the activity of the leaching test should be equal to the blank
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
57
experiment (BLANK), i.e. a blind test under identical reaction conditions as
RUN1 but without the addition of any catalyst. The catalyst remained from the
first step of the leaching experiment was tested separately as preconditioned
material (PC). The pretreatment in 2-butanol at 130°C for 15 h can be repeated
several times which will be further assigned to as PC1 to PC4. The letter stands
for the test of a catalyst that has been preconditioned in 4 successive
preconditioning steps.
Figure 10: Catalytic testing modes
As described before the major side reaction in sugar dehydration is the
formation of humins. In order to discriminate between side blocking by humin
formation and other catalyst deactivation modes, the catalyst was studied in a
probe reaction. The esterification of acetic acid in ethanol is performed at
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
58
milder reaction conditions (80°C) and in the absence of humin-forming side
streams. For this reason it was considered as suitable reference reaction.
The catalytic performance of the acidified carbon nanotube catalyst BTs is
summarized in Figure 11. Here the contribution of leaching experiment is
significant. A major part of the activity found in RUN1 can be assigned to
unstable acid functional groups that come off the catalyst surface during the
reaction. After one preconditioning step in 2-butanol, the BTs catalyst loses the
most of its activity. Similar behavior was found in the esterification reaction
(Figure 12). Although the esterification is performed under milder conditions
of only 80°C and the formation of humins can be excluded, the catalyst loses
almost all activity after only one reaction run. Comparing the results of the
esterification and the dehydration reactions, it can be concluded that in the case
of BTs the deactivation mainly occurs due to loss of acid functional groups or
leaching of the catalyst.
Figure 11: BTs catalyst in the dehydration of fructose in 2-butanol. Reaction
performance in first reaction run (RUN1) vs. leaching test (LEACH) and test of
preconditioned catalyst (PC1) at 130°C, 7 bar N2 using 2.5 g fructose in 100 ml
2-butanol and 250 mg BTs-catalyst. After t = 0.33 h the system reached 130°C reaction
temperature.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
59
Since the functionalization by sulfuric acid is a defect functionalization, it is
possible that strongly functionalized amorphous carbon fractions can be
dissolved from the CNT surface during the reaction. For this reason, further
attempts have been made to improve the stability of the functional groups in
using a highly graphitized carbon support, i. e. after temperature treatment at
3000°C, which shows a minimum content in amorphous carbon. Under this
aspect, a commercial carbon nanofiber (CNF) named PR24XT-HHT from Applied
Sciences® was chosen[24]. The benzene sulfonic acid was grafted by
diazotization onto the CNF surface, resulting into covalently bond acid
functional groups. The functionalized CNF catalyst, further named as BS-CNF,
was tested in the dehydration of fructose in 2-butanol (Figure 13). Similar to
BTs, the BS-CNF catalyst shows significant activities in the leaching experiment.
In that respect, the grafting of benzene sulfonic acid onto highly graphitized
CNFs did not result in more stable functional groups.
Figure 12: BTs catalyst in the esterification of acetic acid. Comparison of catalytic
performance in the first reaction run (RUN1) vs. catalyst recycling (RECYL) at 80°C, 20
bar N2 using 25 mmol acetic acid in 2.5 mol ethanol and 250 mg BTs-catalyst. At
t = 0.33 h the reaction temperature of 80°C is reached
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
60
As reference the ion exchange resins Amberlyst® 15 and Nafion® were also
tested in the fructose dehydration (Figure 14). Both materials showed leaching
activities, although they carry only covalently bond sulfonic acid functional
groups. In the case of Nafion the activity difference between the first run, the
leaching experiment and the recycling run is very similar. The simple recycling
of Nafion consequently would not point to deactivation of the material due to
unstable functional groups. Since both materials carry a high number of acid
functional groups, they can be “recycled” although they continuously lose active
sites. However, for longterm application these materials are not suitable.
Figure 13: Catalytic performance of BS-CNF in the dehydration of fructose in 2-butanol.
Comparison of leaching test (LEACH) with BS-CNF after preconditioning (PC1). At
t = 0.33 h the reaction temperature of 130°C is reached
Since it was not possible to stabilize a high number of acid functional groups
onto the tubular structured, highly graphitic BTs or BS-CNF catalysts, an
acidified amorphous carbon (AGP) was tested as alternative. The carbon
support was synthesized by pyrolysis of glucose in argon atmosphere at 400°C.
After the treatment in sulfuric acid the amorphous carbon support was
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
61
functionalized by 1.7 mmol/g acid functional group, as determined by titration,
which is in agreement with the literature[3].
The stability of the functional groups was tested over four successive
preconditioning steps at 130°C in 2-butanol for 15 h each. The remaining acid
functional groups have been determined by titration (Figure 15). The main loss
of acidity was detected after the first preconditioning. After two pretreatment
runs the amount of acid functional groups decreased to 0.4 mmol/g and this
value was retained over another two preconditioning steps. The catalyst
preconditioned in this way (AGPs_PC4) was tested in the dehydration of
fructose (Figure 16). The HMF productivity decreased from 55 % HMF yield
(t = 3 h) in the first run of AGPs to 39 % HMF yield for the preconditioned
catalyst. However, the material did not show activity due to leaching. Here
leaching was tested by filtering off the catalyst after 2 h reaction time and
continuing the reaction under the same conditions. No further HMF production
could be determined.
Figure 14: Left: Amberlyst® 15 (Amb) in the fructose dehydration. Comparison of first
run (RUN1), catalyst testing after two successive preconditioning steps (PC2), leaching
test of 2-butanol after the first pretreatment (LEACH1) and 2-butanol of the second,
subsequent pretreatment (LEACH2). Right: Nafion® (Naf) in the fructose dehydration.
Comparison of first run (RUN1), leaching test (LEACH) and reuse of the catalyst after
RUN1 in a recycling run (RECYCL)
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
62
Figure 15: Left: Comparison of first run activities of AGPs and BTs. At t = 0.33 h the
reaction temperature of 130°C is reached. Right: Number of acid functional groups
determined by titration as a function of preconditioning steps
Although the AGP material was tested to be stable against leaching after
suitable preconditioning steps, the catalyst could not be recycled. The
deactivation of the material during the fructose dehydration must have a
different origin. The reference experiment showed that AGPs_PC4 can be
recycled in the esterification of acetic acid in ethanol at 80°C. The recyclability in
the esterification (Figure 15) can be seen as further proof that the material is
stable against leaching after sufficient preconditioning. According to this
evidence, the reason for the deactivation in the fructose dehydration is most
likely humin formation and site blocking by solid byproducts.
Another carbon material tested was ordered mesoporous carbon (OMC).
OMC was chosen, since it combines an amorphous structure, predestinated for a
high degree of functionalization by sulfuric acid, with high surface area and a
mesoporous structure suitable for bulky solvated fructose molecules. After the
functionalization in sulfuric acid the OMCs catalyst carried 0.45 mmol/g acid
functional groups. However, catalytic tests of preconditioned OMCs showed
deactivation of the material in the dehydration of fructose, since the material
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
63
cannot be recycled for this reaction (Figure 17). Furthermore, OMCs showed
leaching activity already under the milder conditions of the esterification
reaction.
Figure 16: Upper: Catalytic performance of AGPs after four preconditioning step
(AGPs_PC4) in dehydration of fructose in 2-butanol. Lower: Recyclability of AGPs_PC4
in the esterification reaction
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
64
Figure 17: Left: OMCs in the dehydration of fructose in 2-butanol after preconditioning
(PC), and recycling of the preconditioned OMCs (RECYCL). Right: Leaching activity of
OMCs in the esterification of acetic acid (LEACH)
In the synthesis of OMC functional groups can be incorporated by the choice
of suitable monomers for the polymerization process. For the incorporation of
sulfur functional groups, 4,4’-thiodiphenol (TDP) was used to replace 20% of
the resorcinol in the polymer. After carbonization at 600°C, sulfonic acid
functional groups were formed by oxidation in H2O2. The obtained mesoporous
catalyst, further referred to as TDP0.2, carried 0.25 mmol/g acid functional
groups and showed high activities in the first run of fructose dehydration
(Figure 18). Already after 1 h reaction time, HMF yields of 60% were achieved,
whereas the leaching activity of TDP0.2 is comparably low. However, the
material strongly deactivated after preconditioning in 2-butanol at 130°C. The
activity of the preconditioned material is close to the one of the recycling run.
Hence the major deactivation of TDP0.2 seems to happen by the contact with
the solvent.
The catalytic performance of TDP0.2 in the esterification of acetic acid in
ethanol is depicted in Figure 19. Similar behavior as in the fructose dehydration
reaction could be detected. The leaching activity is comparably low. However,
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
65
the pretreatment in the reaction solvent ethanol leads to significant
deactivation. Deeper investigations of surface changes after exposure to
alcoholic solvents have been done using insitu-XPS and will be intensively
discussed in the following chapter.
Figure 18: Mesoporous carbon catalyst TDP0.2 in the dehydration of fructose in 2-
butanol
Figure 19: Mesoporous TDP0.2-catalyst in the esterification of acetic acid.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
66
IV.4 Summary and Conclusion
As the catalytic tests of acidified carbon materials have shown, several
deactivation procedures can be distinguished during the dehydration of
fructose. A first idea of the stability of the catalyst can be extracted from the
RUN1-activity plots. In the case of homogeneous acids (LEACHing solutions) or
the extensively pretreated catalyst AGPs_PC4, the HMF yield development over
reaction time corresponds to a logarithmic curve shape. Any deviation from the
typical curve shape points to the superposition of several processes and thus
can be seen as first hint for catalyst deactivation.
Table 3: Overview of results on catalyst deactivation tests
Catalyst
Yield
HMF
[%]
BET [m²/g]
r0*
�
µ
mol HMF
2h∙m
²
�
Leach
Active
after
PC
Stable
in
esterification
Recyclable
in
dehydration
BTs 28 245 56.1 yes no no no
BS-CNF 4.1 a 36.1 48.5 yes no n.d. no
AGPs 50; 37 b - - nob yesb yesb no
OMCs 17 683 7.8 yes no no no
MC_TDP 61 654 51.5 little no no no
Naf 36
flexible
structure,
swellable in
liquid phase
- yes yes n.d. no
Amb 50 - yes yes n.d. no
All catalytic tests were performed in a one-phase system of 2-butanol at 130°C. The yields (calculated
per mol initial fructose) refer to the amount of HMF formed after t = 3 h.
r0* specific apparent initial rate
a HMF yield for the preconditioned catalyst (catalyst was pretreated one time in 2-butanol at
130°C for 15 h each)
b HMF yield for the preconditioned catalyst (catalyst was pretreated 4 times in 2-butanol
at 130°C for 15 h each)
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
67
Furthermore, the stability of acid functional groups was found to be critical
for all materials and determined as the major deactivation process for the BTs
catalyst (compare Table 3). Hence appropriate pretreatments are required, and
leaching tests are essential for thorough proof of catalyst stability.
Even if the deactivation by leaching can be diminished by suitable
pretreatments of the catalyst, and the stability of the groups is proven by
leaching tests and recyclability in the reference reaction esterification, as in the
case of AGPs, the catalyst deactivates in the fructose dehydration. Humin
formation and accumulation of the active surface of the catalyst by those
byproducts, is most likely the reason for the catalyst deactivation. The third
aspect comprises catalyst deactivation by exposure to alcoholic solvents, as
found for the mesoporous carbon material TDP0.2.
Finally, it can be concluded that independent from the choice of the carbon
backbone, the covalently bond acid functional groups are not stable in water
and alcohol under the described reaction conditions. This result contradicts the
general hypothesis of considering carbon as hydrolytically stable support
material. The correlation of the amount of functional groups with the number of
defects in the carbon structure leads to a decrease in stability of the carbon
support upon functionalization. Further investigation and better understanding
of carbon degradation processes, i. e. carbon hydrolysis, could lead to
knowledge-based optimization of carbon functional materials in catalysis, as
well as in electrocatalytic applications.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
68
V References
[1] Y. Roman-Leshkov, J. N. Chheda, J. A. Dumesic Science. 2006, 312, 1933-1937.
[2] J. N. Chheda, Y. Roman-Leshkov, J. A. Dumesic Green Chemistry. 2007, 9, 342-350.
[3] A. Takagaki, M. Toda, M. Okamura, J. N. Kondo, S. Hayashi, K. Domen, M. Hara
Catalysis Today. 2006, 116, 157-161.
[4] B. Hu, K. Wang, L. H. Wu, S. H. Yu, M. Antonietti, M. M. Titirici Advanced Materials.
2010, 22, 813-828.
[5] R. Rinaldi, F. Schuth Energy & Environmental Science. 2009, 2, 610-626.
[6] J. K. Chinthaginjala, K. Seshan, L. Lefferts Industrial & Engineering Chemistry
Research. 2007, 46, 3968-3978.
[7] P. Wenmakers, J. van der Schaaf, B. F. M. Kuster, J. C. Schouten Journal of Materials
Chemistry. 2008, 18, 2426-2436.
[8] A. Hirsch Angewandte Chemie-International Edition. 2002, 41, 1853-1859.
[9] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato Chemical Reviews. 2006, 106, 1105-
1136.
[10] B. F. M. Kuster Starch-Starke. 1990, 42, 314-321.
[11] A. I. Torres, P. Daoutidis, M. Tsapatsis Energy & Environmental Science. 2010, 3,
1560-1572.
[12] C. A. Dyke, J. M. Tour Journal of the American Chemical Society. 2003, 125, 1156-
1157.
[13] X. C. Zhao, A. Q. Wang, J. W. Yan, G. Q. Sun, L. X. Sun, T. Zhang Chemistry of Materials.
2010, 22, 5463-5473.
[14] J. P. Yuan, F. Chen Food Chemistry. 1999, 64, 423-427.
[15] R. Weingarten, G. A. Tompsett, W. C. Conner, Jr., G. W. Huber Journal of Catalysis.
2011, 279, 174-182.
[16] Y. Roman-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic Nature. 2007, 447, 982-
U985.
[17] H. E. Van Dam, A. P. G. Kieboom, H. Van Bekkum Starch. 1986, 38, 95-101.
[18] X. L. Tong, Y. Ma, Y. D. Li Applied Catalysis a-General. 2010, 385, 1-13.
Deactivation Pathways of Carbon Catalysts for Fructose Dehydration
69
[19] A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso Green Chemistry.
2011, 13, 754-793.
[20] P. Vinke, H. van Bekkum Starch-Starke. 1992, 44, 90-96.
[21] D. W. Brown, A. J. Floyd, R. G. Kinsman, Y. Roshanali Journal of Chemical Technology
and Biotechnology. 1982, 32, 920-924.
[22] G. A. Halliday, R. J. Young, V. V. Grushin Organic Letters. 2003, 5, 2003-2005.
[23] X. H. Qi, M. Watanabe, T. M. Aida, R. L. Smith Industrial & Engineering Chemistry
Research. 2008, 47, 9234-9239.
[24] J. P. Tessonnier, D. Rosenthal, T. W. Hansen, C. Hess, M. E. Schuster, R. Blume, F.
Girgsdies, N. Pfander, O. Timpe, D. S. Su, R. Schlogl Carbon. 2009, 47, 1779-1798.
70
V REACTIVITYOFMESOPOROUSCARBON
AGAINSTWATER–ANIN‐SITUXPSSTUDY
Authors: Sylvia Reiche, Raoul Blume, Xiao Chen Zhao, DangshengSu,
EdwardKunkes,MalteBehrens,RobertSchlögl*
Abstract
Thiodiphenol(TDP)modifiedmesoporouscarboncatalystscanbeusedin
theacidcatalyzeddehydrationoffructoseto5‐hydroxymethylfurfural(HMF).
However,strongdeactivationcanbeobservedafterpreconditioning of the
materialinthesolvent2‐butanol.Surfacechangescausedbythepretreatment
have been studied by a XPS. The comparison of the pristine sampleandthe
pretreated carbon sample showed similar distribution of oxygen functional
groupsbyex‐situXPS,aswellassimilarbehaviorduringheating in vacuum.
However,theadditionof0.1mbarvaporpressureandsubsequentheatingto
130°CexhibitedprominentdifferencesintheevolutionoftheO1s,aswellasfor
theC1sspectraofthetwosamples.Changesinthesurfacetermination and
hydrophobicity of the materials are discussed under the aspect of possible
reactionsofsurfacefunctionalgroupswiththealcoholicsolventandwater.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
71
V.1 Introduction
Mesoporouscarbonmaterialsfromresorcinol‐formaldehydeaerogelshave
attractedenormousattentionduetotheiradvantageouspropertiesandmultiple
potentialapplications,rangingfromelectrodematerialsinsupercapacitorsand
fuel cells, over filters, down to catalytic supports or catalysts[1, 2]. Several
studies on material texture and pore structure are reported[3]. However, the
detailedchemicalsurfacestructureisstillunderinvestigation.Surfaceanalytical
methods,suchasX‐rayphotoelectronspectroscopy(XPS),havebeen used to
quantifythetotalsurfaceoxygencontentofthemesoporouscarbon[4].However,
theidentificationofthenatureandtheabundanceofspecificoxygenfunctional
groups are still not understood.Basedonthestructuralelements of the
monomersandthepolymerizationmechanismsuggestedbyPekala
[5]
(Figure1),thereisageneralagreementonthefunctionalgroupsoftheaerogel.
Isotope exchange experiments combined with 13C‐NMR confirmed the
abundanceofmethylenebridgesandetherbridgesinthepolymer
[6]. The
elucidationofthesurfacefunctionalgroupsofthecarbonizedmaterialwasthe
intentionofthepresentstudy.Thefocuswasonanimprovedunderstandingof
observations from prior catalysis tests of acidified mesoporouscarbon
materials[7].
Acidifiedcarboncatalystscanbeappliedintheacidcatalyzeddehydrationof
fructoseto 5‐hydroxymethyl furfural(HMF,Figure2).Thegeneralfindingof
the catalytic tests, as reported elsewhere[7, 8], we observed a significant
deactivationofthecarbonbasedmaterialsalreadyinthefirstrecyclingrun.The
deactivationwasfoundtohavetwocauses;thelossofacidfunctionalgroups,
which was determined by leaching tests, and surface poisoning by insoluble
humins,asconcludedfromcomparisontoareferenceesterificationreaction.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
72
Figure1:Possiblereactionmechanismfortheformationoftheresorcinol‐
formaldehydepolymerastheprecursorforOMCsynthesis[5,9]
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
73
Figure2:Fructosedehydrationinto5‐hydroxymethylfurfuralunderliberation of 3
moleculesofwater
For one group of materials, the thiodiphenol (TDP) modifiedmesoporous
carbon,wecouldobserveverypromisingHMFyieldsof60%after1hreaction
time(Figure3,left).Theleachingactivitywascomparablylow.However,the
catalyst deactivated after the preconditioning in 2‐butanol. Inasubsequent
catalytictestofthepreconditionedmaterialweobservedadelayedinitialization
ofHMFformation,aswellasalowertotalactivity.Thebehavior of the
preconditioned material is very similar to the catalyst in the recycling run.
Hence, major contribution to the deactivation of the catalyst occurred from
exposuretothesolvent.Thebetter understanding of this deactivation
procedureisnecessaryforthedevelopmentofstableheterogeneouscatalysts.
We further found that the initial activity of TDP0.2 is not, asoriginally
expected,afunctionofthethiodiphenylcontent(Figure3,right).TheTDP‐free
catalyst MC_1 shows identical activity as TDP0.2 and TDP0.6. The oxidation
treatmentinhydrogenperoxidemostlikelyintroducessurfacespeciesthatare
responsible for the dehydration activity of the materials. Hence all further
considerationsarebasedonTDP‐freesamplesofmesoporouscarbon.
Inordertoinvestigatechangesinsurfacestructureandoxygenfunctional
groups during the pretreatment in 2‐butanol, we have chosen two
representative samples for insitu XPS experiments. The first mesoporous
carbon sample was functionalized in hydrogen peroxide at pH 1 and
corresponds to an active catalyst in the dehydration of fructose into HMF
(MC_1),asshowninFigure3. Secondly, a sample MC_2 was obtained by a
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
74
pretreatmentofMC_1inreactionsolventatreactiontemperature, i. e. in 2‐
butanolat130°C.Inaddition,areferencesampleMC_0wasexaminedex‐situby
XPS.MC_0isanorderedmesoporouscarbonsamplewhichhasnotbeen
oxidizedbyH2O2.
Weinvestigateddynamicchangesinsurfacestructuresandinthe
distribution of oxygen functional groups under vapor exposure and heat
treatments to the reaction temperature of 130°C by in‐situ XPS.Thevapor
pressureof0.1mbarwasappliedinordertostudythereactivity of the
functionalized carbon surface towards water. Furthermore, the addition of
vaporsimulatesthewaterevolutionof3H2OmoleculespersynthesizedHMF
moleculeduringthedehydrationreactionoffructose.
Figure3:Catalyticperformanceofmesoporouscarbonmaterialsinthedehydrationof
fructosetoHMF.Left:Deactivationofthematerialafterthepretreatmentin2‐butanol
(PCTDP0.2).Right:ComparisonofsamplesofdifferentTDPcontent.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
75
V.2 Experimental
V.2.1 Materialsynthesisandpreliminarycharacterization
For the synthesis of MC_0, 5.5 g of resorcinol and 5.0 g of formaldehyde
solution (37 wt%) undergo an acid catalyzed polymerization in 50 ml of a
water‐ethanol‐mixture(1:1byweight).Theobtainedpolymeriscarbonizedin
N2atmosphereat350°Cand600°Cfor2h,respectively.Thedetailedprocedure
ofOMCsynthesisisdescribedelsewhere[2].
After the carbonization, the mesoporous carbon is oxidized by H2O2ina
mixtureofmethanoland2MHCl(1:1,V:V)toobtainMC_1[2].
ThepreconditionedmaterialMC_2wassynthesizedofMC_1bystirringin2‐
butanolat130°Cfor15h.
All materials are mesoporous carbons. The BET surface areas were
determined by measuring the adsorption‐desorption‐isotherms with a
QuantachromeAutosorbautomaticBET‐sorptometerat‐196°Cwithnitrogenas
analysisgas.Fordataevaluation the Quantachrome software Autosorb1
(version1.54)wasused.AfulllistofBETisothermsandplotsfortheBJHpore
sizedistributioncanbefoundinthesupplementaryinformation.
Thenumberofacidfunctionalgroupswasdeterminedbytitrationusingan
automatic titrator (Mettler Toledo). For each measurement, 100 mg catalyst
weredispersedin10‐3molKCl‐solutionandstirredovernight.Followingthe
dispersionwastitratedunderAratmosphere,usinga0.1MNaOHsolution.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
76
Table1:SummaryofmaterialbasisincludingBETareasandtitrationresults
SampleMaterialdescriptionBET[m²/g]Acidsites
[mmol/g]
MC_0orderedmesoporouscarbon789‐
MC_1mesoporouscarbon,functionalizedby
H2O2atpH=15960.24a
TDP0.2mesoporouscarbon,20mol%ofbuilding
blockrecorcinolreplacedbythiodiphenyl6540.25a
MC_2
mesoporouscarbon,MC_1after
preconditioningin2‐butanolat130°Cfor
15h
604‐
adensityofacidsitesdeterminedbytitrationofthematerial(100mgin50mlKCl0.001M)with0.01MNaOH
V.2.2 Instrumental
In‐Situ XPS experiments were performed at the synchrotron radiation
sourceBESSYIIoftheHelmholtzZentrumBerlin(HZB).Thein‐situ chamber
was designed by FHI[10, 11]. A schematic sketch of the set up is depicted in
Figure4.Thehigh‐pressurereactioncellisseparatedfromtheX‐raysourceby
anX‐raytransparentwindow.Theemittedelectronsattainthehemispherical
electronanalyzerthroughadifferentiallypumpedaperture.
During in‐situ XPS analysis of MC_1 and MC_2, the samples were heated
stepwiseto130°Cinvacuumandsubsequentlyexposedtowaterat0.1mbar.In
a second experiment, the samples were heated from RT to 130°C in water
atmosphere(0.1mbar).SpectraoftheC1s,O1sandCl2pregionsaswellastheir
respective Fermi edges were recorded with an electron kinetic energy of
~150eV.
All spectra are normalized to the background on the high binding energy
side for better comparison of the peak shape of the main component. The
spectrawerefittedwithasetofpeaksderivedfromadifferentialspectrasurvey
ofalargenumberoffunctionalizedandunfunctionalizedcarbonsamples[12].
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
77
Figure4:SchematicdrawingofthehighpressureXPSsystematBESSYII[11].
V.3 ResultsandDiscussion
V.3.1 Differencesinoriginalcarbonsamples
InafirststepthesamplesMC_1andMC_2wereinvestigatedex‐situ and
comparedtothereferencesampleMC_0thatrepresentsthemesoporouscarbon
beforeanyoxidationtreatment.AccordingtoBlumeetal.[12],the O1s spectra
werefittedby6differentcomponentsofthepeakpositions530.5eV,531.2eV,
531.9eV,532.7eV,533.5eVand534.2eV(Table2).Anadditionalfeaturefor
themeasurementsunderwatervaporwasdetectedat535eVandcorresponds
tothewatergasphasepeak[13].
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
78
Table2:O1speakassignmentaccordingtotheliterature
No.Peakposition[eV]OxygenfunctionalgroupSupportingReferences
I 530.5C=Oinquinones
[13‐16]
[17,18]
II 531.2C=Oinketones,aldehydes[17,18]
III 531.9C‐O‐Cinaromates(furan),
orketo‐enoltautomers
[13‐15,18]
IV 532.7
OH inphenoloraliphatic
alcohols
[16]
chemisorbedH2O
V 533.5C‐O‐Cinethers,esters,
anhydride
[16,19]
VI 534.2
C‐OHincarboxylicacid[16,18]
chemisorbedH2O
[13‐15]
[19]
VII 535gasphaseH2O[13]
The assignments of specific features in the O1s peak are intensively and
partiallycontroversiallydebatedintheliterature.Itisgenerallyagreedtothe
discrimination of minimum two different oxygen species, the double bonded
oxygenatlowerbindingenergies(~531eV)andthesinglebondedoxygenat
higher binding energies (~533 eV)[20]. Additionally, a third species at higher
bindingenergiesiscommonlyconsideredinthefit,assignedtoadsorbedwater
oroxygen[13‐15].Clarketal.performedmoredetailedstudiesinthesystematic
comparisonofwelldefinedpolymersandsuggestedthediscriminationoffour
differentoxygenspecies:double‐bondedoxygeninesters,carbonatesandacids
(~532.8–532.9eV),oxygeninketons,ethersandalcohols(~533.6–533.7eV),
single‐bonded oxygen in acids and esters (~534.3 eV) and single‐bonded
oxygenincarbonates(~535.0–535.2eV)[21].ThehighresolutionofmodernXPS
instruments,theuseofsynchrotronradiationsources,andtheconsiderationof
thethermostabilityofdifferentoxygenfunctionalgroupsleadtoverificationof
theClarkmodel[18,22]andfurtherdifferentiationoftheO1speak[16,17,19,23].
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
79
Forthepresentworkwealsotriedtoinvolveconsiderationsonstructural
elements that are predetermined by the synthesis procedure. Based on the
polymeric precursor of the mesoporous carbon (compare Figure1) a broad
variety of oxygen functional groups is possible for the carbon samples. Most
likely the material still contains phenolic and aliphatic OH‐groups, as well as
ethergroupsafterthecarbonizationprocedure.Duetopossiblecondensation
reactions at the applied temperatures (600°C for 2 h during carbonization),
lactones,furansorquinonesarefurtherpossiblestructuralelements.Duringthe
oxidationinhydrogenperoxideanincreaseincarbonylandcarboxylgroupsis
expected,aswellasfurtherhydroxylfunctionalgroupsthroughtheoxidationof
doublebonds.
Figure5:Schemeofpossiblecarbonstructureforthesamplesinvestigated in
considerationofthestructuralelementsofthepolymericprecursorandideasreported
intheliterature[23,24].Inblue:functionalgroupsthatcaninteractwithwatermolecules
undertheformationofhydrogenbondsorinhydrolysisreactions
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
80
Comprisingtheknowledgeofthebuildingblocks,thesynthesisprocedure
andexperiencesreportedintheliterature[23, 24],aschematiccarbonstructure
canbeproposedforthematerials(Figure5).Inthefollowing,thisstructural
ideawillbecomparedtotheresultoftheXPSexperiments.
Figure6:O1sfitsofMC_0,MC_1andMC_2ofex‐situXPSmeasurements
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
81
Ex‐situmeasurementsofMC1andMC_2showO1sspectraofsimilar line
shapeforbothsamples(Figure6).Thereisonlyasmalldifferenceinthetotal
oxygencontentofabout1%(Table3).Thesmallintensitydifferenceisequally
distributedbetweenthespeciesII‐V.However,thecomponentVIat534.2eV,
whichcanbeassignedtocarboxylicacidfunctionalgroups[16,18]orchemisorbed
water[13‐15]accordingtotheliterature,isabundantinslightlyhigherproportion
insampleMC_1incomparisontothesampleMC_2afterthebutanoltreatment.
EitherassignmentcouldsupporttheideaofpossiblesurfacechangesofMC_1
duringthetreatmentinalcoholicsolvents.Inthecaseoftheassignment to
carboxylic acid functional groups, the lower content for MC_2 could be
explainedbypossibleesterificationreactionsinthealcoholicsolvent.Ifassigned
to adsorbed water, the higher content of the species at 534.2 eV points to a
higherhydrophilicityforMC_1whichcanbeseenasindirectevidence for
surfacemodificationsby2‐butanolinthecaseofMC_2.
Table3:QuantificationofoxygenspeciesofMC_0,MC_1andMC_2byex‐situXPS[%]
Sample530.5eV531.2eV531.9eV 532.7eV 533.5eV 534.2eV Totaloxygen
content
MC_00.3 0.5 0.8 1.61.9 0.25.3
MC_10.4 1.5 1.43.04.00.811.1
MC_20.4 1.8 1.63.44.40.512.1
Incontrasttotheoxidizedsamples,theO1sspectrumofthereference
sampleMC_0differsinlineshapeduetothelowerinfluenceoffeaturesatlower
binding energies. After the oxidation, the carbonyl species at 531.2 eV
contributedstrongertotheoveralllineshapeforthesamplesMC_1andMC_2.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
82
Table4:C1speakassignmentaccordingtoliterature
No.Peakposition[eV]CarbonfunctionalgroupSupportingReferences
I 284.4C=C,sp²carbon[15‐17,25]
II 284.7C‐C,sp³carbon[15‐17,25]
III 285.2aliphatischeC‐H,C‐Oin
alcohols,phenols,ethers
[15‐17,25]
IV 285.9C=C‐Oinketo‐enolicequilibria
orfurans[15,17]
[16,25]
V 286.6C=O[16,25]
VI 287.9COOH,COORincarboxylic
acidsoresters
[15‐17,25]
VII 288.5
carbonate[15,16]
[26]
VIII 289.1[26]
SimilartotheO1sspectra,theC1speaks(Figure7)werefittedbyasetof
fittingparametersderivedfromadifferentialspectrasurveyofalargenumber
offunctionalizedandunfunctionalizedcarbonsamples[12].WithintheC1speak,
eightfeaturesatthepositions284.4eV,284.7eV,285.2eV,285.9eV,286.6eV,
287.9eV,288.5eVand289.1eVwerediscriminated(Table4).However,dueto
thedominantinfluenceoftheasymmetryofthesp²carbonpeak(Imax=
284.4eV)uptohighbindingenergies,thequantificationoftheoxygenspecies
intheC1sisratherdifficult.Smallerrorsinthesubtractionofthegraphiticpeak
canleadtolargeerrorsintheevaluationofthefunctionalgroups,inparticular
duetotheirminorcontribution[27].
ComparingthefitsoftheC1sspectraforthethreesamples,acleardifference
intheproportionofgraphiticcarbontoamorphouscarbonand/orunsaturated
bondscanbefound.ThehigherpercentageofgraphiticsurfacecarbonforMC_2
couldbeexplainedbypartialremovalofamorphouscarbonspeciesduringthe
pretreatment in 2‐butanol. Another explanation could be partialstructural
reorientationofthesurfaceasaconsequenceofpretreatment.Theesterification
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
83
of carboxylic acid functional groups, for instance, could lead to a decrease of
structure determining hydrogen bonds (compare Fig.). Asa consequence, the
influence of π‐π‐stacking interactions increases which could result in more
denselypackedregimesofgraphiticcarbon.ThepeaksIII‐VIIIcorrespondtothe
oxygenfunctionalgroups.Basedonthequantificationofthesame,loweroxygen
contentcouldbedeterminedforthereverencesampleMC_0(Table5)whichis
inagreementtothefindingsoftheO1sspectraevaluation.
Figure7:C1sfitsofMC_0,MC_1andMC_2ofex‐situXPSmeasurements
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
84
Table5:QuantificationofcarbonspeciesintheC1spectraofMC_0,MC_1andMC_2
obtainedbyex‐situXPS[%]
Sample284.4eV284.7eV285.2eV 285.9eV 286.6eV 287.9eV 288.5eV 289.1eV
Total
carbon
content
MC_040.0 27.7 4.72.42.91.85.91.294.8
MC_137.026.7 4.93.04.02.40.61.888.6
MC_243.0 15.7 5.12.93.42.30.51.187.7
V.3.2 Differencesduringheatinginvacuum
TheO1sspectrashowsimilarbehaviorforMC_1andMC_2whileheatingthe
materialsinvacuum.Forbothsamples,adecreaseinthetotaloxygencontentof
aboutonequarterforMC_1andaboutonethirdforMC_2canbedetermined
(Table3).Eventhewaterexposureat130°Cdoesnotchangethetrendofthe
gradualdecreaseofmostoxygencomponents.Themajorchangeintheoxygen
speciesdistributionisobservedforthefeatureat532.7eVwhich can be
assignedtotemperaturesensitiveOH‐groups,accordingtotheliterature.Dueto
the mild temperatures applied, the loss of adsorbed water species from the
highlyporousmaterials(compareBETsurfaceareasinTable1)seemstobe
anotherplausibleconclusion.TG‐MSexperimentswereperformedinorderto
check for carbon decomposition reactions under CO2 evolution at the
temperatures applied in the in‐situ experiments. I could be confirmed that
water is the only released species within the investigated temperature range
(Figure8). The most pronounced intensity after the heating experiments
correspondstoanoxygenspeciesat533.5eV.Asdescribedbeforethisfeature
canbeassignedtoether‐likeoxygenfunctionalgroupswhichwereexpectedto
bestableintheappliedtemperaturetreatment.Duetothepronouncedintensity
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
85
lossinthisregimeandwaterbeingtheonlyspeciesevolvedbelow130°C,the
533.5eVpeakhastobeassociatedwithwater.
Table6:QuantificationofoxygenspeciesofMC_1andMC_2duringheatinginvacuum
andsubsequentadditionofwater(0.1mbar)at130°C
Figure8: TG‐MS experiment of MC_1. Left: mass loss over temperature treatment,
Right:MS‐signalforCO,CO2,OHandH2O
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
86
DespitethesimilarityintheevolutionoftheO1sspectra,differencescanbe
observedcomparingtheC1sspectraofthetwosamples(Figure9).Incaseof
MC_1thecarbonstructureofthematerialremainsalmostunchanged during
heatinginvacuum.Incontrary,anincreaseinamorphoussurfacecarboncanbe
determinedforMC_2.Generallytheformationofamorphoussurfacecarboncan
beexplainedbytheremovalofoxygenfunctionalgroupswhileless ordered
amorphous structures are retained. However, since the O1s spectra of both
samples show similar losses in oxygenfunctionalgroups,thisidea does not
provideasatisfyingexplanationoftheobservationsintheC1sspectra.Forthis
reason,wewouldliketorefertotheinitialdifferencesofthe C1s spectra as
discussedfortheex‐situXPSmeasurements.Theeffectofpossibleesterification
reactionsduringthepretreatmentin2‐butanolcouldbepartiallyreversedby
the water evolved during heating. Thus, hydrolysis and re‐formation of
hydrogenbondscanleadtostructuralchanges.Theincreaseinwettabilityand
intercalationofwatercanresultintheexfoliationofthestructure.Itispossible
thatlessstableamorphouscomponentscanbepartiallytransported towards
thesurfacebythewaterevolved,resultinginthedetectionofhigheramountsof
amorphoussurfacecarbon.
Sincefornoneoftheoxygenrelatedfeaturesathigherbindingenergiesof
the C1s spectra a pronounced loss was observed, the strong changes in O1s
spectraforbothsamplesareunlikelytoberelatedtothecleavageofC‐O‐bonds.
Moreplausiblefortheexplanationofthechangesinthefeatureat532.7eVis
theremovalofstructuralwater,e.g.fromwaterincorporationoftheporesor
thesolvationofoxygenfunctionalgroups.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
87
Figure9: Evolution of C1s components during heating in vacuum and subsequent
additionofwater(0.1mbar)at130°C
V.3.3 Behaviorduringheatinginwateratmosphere
Comparedtotheexperimentsinvacuum,therearesignificantdifferencesin
the behavior of the two samples MC_1 and MC_2 while heating under water
vaporof0.1mbar.Thepresenceofwaterinducestheadditionalreactivityof
hydrolysis to the sequence of chemicaleventsuponthermaltreatment. The
lattercommonlyleadstocondensationreactions.Whereastheoveralloxygen
content for MC_2 does not change significantly under the heat treatment in
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
88
vapor(evenshowsatendencytoriseupto80°C),thetotalO‐contentforMC_1
decreasesfrom11.1%to9.5%(Table7).
Table7: Quantification of oxygen species for MC_1 and MC_2 during heating in
0.1mbarvapor
Thesignalat533.5eV,whichwasstableforbothsamplesduringtheheat
treatmentinvacuum,decreasesforMC_1whileheatinginwateratmosphere.
ThisindicatesthatpartsoftheetherorestergroupsofMC_1are sensitive
againsthydrolysis.IncaseofMC_2,thestructuralchangesofthesamplesduring
the pretreatment in 2‐butanol, i. e. the reordering towards a more graphitic
surface termination (compare result of ex‐situ XPS,
Figure7),seemstostabilizethematerialagainsttheintrusionofwater and
hydrolyticeffectsoftheheattreatmentsinvapor.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
89
Figure10: Reactivity of different oxygen functional groups towards water[28]and
consequentialchangesinthesolvationchemistry.
Hydrolysisreactionshaveanoxygenintroducingeffect.Thusthe small
changesintheoveralloxygencontentforMC_2couldbetheresult of two
simultaneousprocesses:Firstly,thelossofoxygenspecies(waterremoval)due
totheincreasingtemperatureandsecondlythepartialintroductionofoxygen
functional groups by possible hydrolysis reactions. Figure10summarizes
potentialhydrolysisreactionsforthepriorsuggestedsurfacefunctionalgroups
ofthematerials.Asaresultofthehydrolysis,thesurfacesolvationisenhanced,
due to the increase in functional groups that are capable of hydrogen bond
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
90
formation. Solvation or the incorporation of water molecules could be one
explanationofrestructuringforcesasdiscussedbefore.
Figure11: Evolution of C1s components during heating in vacuum and subsequent
additionofwater(0.1mbar)at130°C
AsacommontrendoftheO1sspectra,bothsamples,MC_1andMC_2,exhibit
adecreaseofthepeaklocatedat532.7eV,alsoobservedundervacuum.Atthe
sametime,thefeatureatthebindingenergyof534.2eVincreasesinintensity.
The opposite trend of development of both peaks supports the idea of two
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
91
simultaneousprocesses.Thecorrelationofthepeakat534.2eVtoOH‐species
incarboxylfunctionalgroupscorrespondstoanoftenreportedassignmentin
the literature[15‐17,25]. Inaddition, carboxyl groups are preferentially formed in
most hydrolysis reactions (Figure10). Based on the reaction conditions
applied,i.e.mildtemperaturesandvaporatmosphere,thedecreaseinthepeak
at532.7eVismostlikelyrelatedtoadsorbedwaterfromporeincorporationor
materialsolvation.
The C1s spectra for the experiments in water atmosphere reveal
astrongercontributionofamorphouscarbonforbothsamples(Figure11),in
comparisontothewater‐freeexperiments.IncaseofMC_1adramaticincrease
oftheamorphouscarboncanbefollowedduringheatinginvapor.Similartothe
observationsfromtheO1sspectra,MC_1seemstobelessstableagainstwater.
V.3.4 Climpuritiesandtheirevolutionduringheatandvaportreatment
TheacidicpHduringtheoxidationtreatmentinhydrogenperoxide is
achieved by the addition of hydrochloric acid to the reaction mixture. As
observedinthesurveyspectra,partofthehydrochloricacidreactedwiththe
carbon material and formed chloro‐functionalized surface species on the
material.InordertogiveacompletecomparisonofthetwosamplesMC_1and
MC_2intermsofcatalyticactivity,itisimportanttoalsoconsiderdifferencesin
the amount of the Cl‐species and their changes under reaction conditions.
ThereforetheCl2ppeakshavebeenexaminedindetail.
ThemainfeatureoftheCl2pislocatedat~200.4eV.Thiscanberelatedto
chlorinated polymers like polyethylene or chlorinated benzene‐like
molecules[29].SampleMC_2exhibitsasmalleramountofClthansampleMC_1at
roomtemperature(Table8)andonlyMC_1showsadecreaseoftheClsignal
with increasing T under vacuum conditions. This effect proceedsalsounder
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
92
water exposure at 130°C, where the intensity of MC_1 reaches values
comparabletotheaverageClsignalofMC_2.
Table8:QuantificationofClspeciesofMC_1andMC_2duringin‐situXPS
Experiment:Heatinginvacuumand
subsequentadditionof0.1mbarvapor
Processstep
TotalClcontent
[%]
MC_1
RT0.4
80°C0.2
130°C0.2
130°Cvapor0.1
MC_2
RT0.2
80°C0.1
130°C0.1
130°Cvapor0.1
Experiment:Heatingin0.1mbarvapor
pressure
Processstep
TotalClcontent
[%]
MC_1
RTvapor0.3
80°Cvapor0.3
130°Cvapor0.2
MC_2
RTvapor0.2
80°Cvapor0.2
130°Cvapor0.3
Theheatingofthecarbonmaterial under water atmosphere in seems to
stabilizetheCl‐functionalgroupsinboth samples.Minordifferencescouldbe
detected both for the different samples MC_1 and MC_2, as well as for the
differenttemperaturesteps.
Althoughthetotalamountofchlorinefunctionalgroupsiswith<0.5%small
incomparisonto12%oxygencontent,itcannotbeexcludedthatthedifferences
intheCl‐contentbetweenMC_1andMC_2influencedthecatalyticperformance
ofthetwomaterials.Thenucleophilicattackofthealcoholicsolvent on the
chlorinated carbon could lead to the release of HCl under the formation of
butylethergroupsatthecarbonsurface.Theamountof0.4%Clfor MC_1
correspondstoanacidityofapproximately0.3mmol/g,incaseofathorough
substitutionoftheCl‐functionalgroupsbyetherification.Thisvalueisinclose
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
93
agreementwiththeresultsofthe titration experiments. However,sincethe
leachingactivity,asshownforTDP0.2(Figure3),waslowincomparisontothe
activity of the oxidized carbon material, the question of the function of the
heterogeneouscatalystremains.Itispossiblethatthein‐situformationofHClis
more effective, since side reaction, e. g. with the solvent, can be avoided.
Anotherpossibleexplanationisgivenbytheoxygenfunctionalgroups.Brønsted
acidiccarboxylicacidfunctionalgroupscouldalsobetheactivespeciesinthe
dehydration reaction. However, their esterification with the alcoholic solvent
couldleadtofastdeactivationoftheacidifiedcarbon.Experimentsunderthe
additionofhomogeneousorganicacids,suchasaceticacidandoxalicacid,atthe
samereactionconditionsleadtodiminutiveHMFyieldsincomparisontothe
use of homogeneous mineral acids. The reason also here is the competitive
reactionofsolventesterification.
V.4 Conclusions
Theresultsofthepresentedin‐situXPSexperimentcanbeevaluatedunder
variousaspects.Firstwewantedtogainfurtherinsightintothe deactivation
mechanismofacidifiedmesoporouscatalystinthedehydrationoffructose.Here
we found evidence for two different deactivation procedures, whereas the
weightingofthefinalinfluenceofthesamewouldneedfurtherinvestigationsby
complementarymethods.
OntheonehandthedifferenceinCl‐contentcouldbeanexplanationforthe
unequalcatalyticperformanceofMC_1andMC_2.However,thelowactivityin
the leaching tests points to other influencing factors. On the other hand the
catalyticallymoreactivesampleMC_1exhibitsahighercontentoftheoxygen
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
94
speciesat534.2eVwhichcanbeassignedtocarboxylicacidaccordingtothe
literature[15‐17,25]. The deactivation of the catalyst during the pretreatment in
alcoholic solvents can be due to possible esterification reactions. As indirect
evidenceofsurfacechangesbyesterification,thestructuraldifferencesobtained
fromtheex‐situXPSandthedifferentbehaviorofthetwosamples under
heatinginwateratmospherewerediscussed.
Figure12:InsituXPSspectraofsampleMC1(left)andMC2(right).Theline in the
figurecorrespondstothemainintense533.4eVpeak.Thearrow points to a new
featuredevelopingwhilewaterexposure.
Apartfromtheresultsrelevantforthecatalyticapplicationofthematerials
inbiomassconversionreactions,theexperimentsprovidedcrucialinformation
ongeneralpropertiesofoxygenfunctionalized carbon in contact of water.
Figure12revealsthedifferentbehaviorofthetwocarbonmaterialsMC_1and
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
95
MC_2bydirectcomparisonoftheO1sspectra.Despitethesimilarities in the
oxygencontentandtheevolutionoftheO1speakduringheatinginvacuum,the
additionofwaterandadjacentheattreatmentsrevealedsignificantdifferences
inthebehaviorofthetwosamples.
Followingitcanbeconcludedthatcarbonmaterialsarenotnecessarilyinert
towardswateriftheycarryoxygenfunctionalgroups.Carbon,ifnotgraphitic,
exhibitsadynamicbehaviorundertheinfluenceofwater.Sincethein‐situXPS
experiment have been performed in only 0.1 mbar vapor pressure and mild
temperaturesofmaximal130°C,moredrasticchangesinthestructureofcarbon
materials can be expected under “real” hydrothermal or electrochemical
conditions.Severalconclusionsofthebehaviorofcarbonmaterialsunderthe
influenceofaqueousmedia,e.g.forelectrodematerials,canbereconsidered
basedonthefindingofthepresentedXPSexperiments.
V.5 References
[1] H.Probstle,C.Schmitt,J.Fricke,JournalofPowerSources2002,105,189‐194;L.
Wencui,G.Reichenauer,J.Fricke,Carbon2002,40,2955‐29592959;C.Moreno‐
Castilla, F.J.Maldonado‐Hodar,J.Rivera‐Utrilla,E.Rodriguez‐Castellon,Applied
Catalysisa‐General1999,183,345‐356.
[2] X. C. Zhao, A. Q. Wang, J. W. Yan, G. Q. Sun, L. X. Sun, T.Zhang,Chemistryof
Materials2010,22,5463‐5473.
[3] Y.Hanzawa,H.Hatori,N.Yoshizawa,Y.Yamada,Carbon2002,40,575‐581;W.C.
Li,A.H.Lu,S.C.Guo,Carbon2001,39,1989‐1994.
[4] P. V. Samant, F. Goncalves, M. M. A. Freitas, M. F. R. Pereira,J.L.Figueiredo,
Carbon2004,42,1321‐1325.
[5] R.W.Pekala,JournalofMaterialsScience1989,24,3221‐3227.
[6] A.W.Christiansen,JournalofAppliedPolymerScience2000,75,1760‐1768.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
96
[7] S. Reiche, X. C. Zhao, M. Aronson, K. Friedel, E. Kunkes, J.‐P. Tessonnier, M.
Behrens,D.S.Su,R.Davis,S.B.AbdulHamid,R.Schlögl,2012.
[8] S.Reiche,E.Kunkes,N.B.AsariMansor,X.C.Zhao,K.R.Yuyyuru,A.Villa,D.S.Su,
M.Behrens,P.Strasser,R.Schlögl,2012.
[9] G. C. Ruben, R. W. Pekala, T. M. Tillotson, L. W. Hrubesh, JournalofMaterials
Science1992,27,4341‐4349.
[10] H.Bluhm,M.Haevecker,A.Knop‐Gericke,M.Kiskinova,R.Schloegl,M.Salmeron,
MrsBulletin2007,32,1022‐1030;A.Knop‐Gericke,E.Kleimenov,M.Haevecker,
R.Blume,D.Teschner,S.Zafeiratos,R.Schloegl,V.I.Bukhtiyarov,V.V.Kaichev,I.
P.Prosvirin,A.I.Nizovskii,H.Bluhm,A.Barinov,P.Dudin,M.Kiskinova,Advances
inCatalysis,Vol522009,52,213‐272;M.Salmeron,R. Schlogl,SurfaceScience
Reports2008,63,169‐199.
[11] E.M.Vass,M.Haevecker,S.Zafeiratos,D.Teschner,A.Knop‐Gericke,R.Schloegl,
JournalofPhysics‐CondensedMatter2008,20.
[12] R.Blume,inunpublishedwork,2011.
[13] G. Ketteler, P. Ashby, B. S. Mun, I. Ratera, H. Bluhm, B. Kasemo, M. Salmeron,
JournalofPhysics‐CondensedMatter2008,20.
[14] R. Schlogl, G. Loose, M. Wesemann, SolidStateIonics1990, 43, 183‐192; M. T.
Martinez,M.A.Callejas,A.M.Benito,M.Cochet,T.Seeger,A.Anson,J.Schreiber,
C.Gordon,C.Marhic,O.Chauvet,J.L.G.Fierro,W.K.Maser,Carbon2003,41,
2247‐2256;E.Desimoni,G.I.Casella,A.Morone,A.M.Salvi,SurfaceandInterface
Analysis1990,15,627‐634.
[15] S. Biniak, G. Szymanski, J. Siedlewski, A. Swiatkowski, Carbon1997, 35, 1799‐
1810;H.Darmstadt,C.Roy,S.Kaliaguine,Carbon1994,32,1399‐1406.
[16] J.L.Figueiredo,M.F.R.Pereira,CatalysisToday2010,150,2‐7.
[17] M.E.Schuster,M.Havecker,R.Arrigo,R.Blume,M.Knauer,N.P.Ivleva,D.S.Su,R.
Niessner,R.Schlogl,JournalofPhysicalChemistryA2011,115,2568‐2580.
[18] J.L.Figueiredo,M.F.R.Pereira,M.M.A.Freitas,J.J.M.Orfao,Carbon1999,37,
1379‐1389.
[19] B.Frank,A.Rinaldi,R.Blume,R.Schlogl,D.S.Su,ChemistryofMaterials2010,22,
4462‐4470.
[20] S.Kundu,Y.M.Wang,W.Xia,M.Muhler,JournalofPhysicalChemistryC2008,
112,16869‐16878.
ReactivityofMesoporousCarbonAgainstWater–anIn‐SituXPSStudy
97
[21] D. T. Clark, B. J. Cromarty, A. Dilks, JournalofPolymerScienceParta‐Polymer
Chemistry1978,16,3173‐3184;D.T.Clark,A.Dilks,JournalofPolymerScience
Parta‐PolymerChemistry1979,17,957‐976.
[22] D.Rosenthal,M.Ruta,R.Schloegl,L.Kiwi‐Minsker,Carbon2010,48,1835‐1843;
G.P.Lopez,D.G.Castner,B.D.Ratner,SurfaceandInterfaceAnalysis1991,17,
267‐272.
[23] A.Stein,Z.Wang,M.A.Fierke,AdvancedMaterials2009,21,265‐293.
[24] M.S.Akhter,A.R.Chughtai,D.M.Smith,AppliedSpectroscopy1985,39,143‐153;
D.M.Smith,A.R.Chughtai,ColloidsandSurfacesa‐Physicochemicaland
EngineeringAspects1995,105,47‐77;P.E.Fanning,M.A.Vannice,Carbon1993,
31,721‐730.
[25] U.Zielke,K.J.Huttinger,W.P.Hoffman,Carbon1996,34,983‐998.
[26] M.E.Schuster,TechnischeUniversitätBerlin(Berlin),2011.
[27] E.Desimoni,G.I.Casella,T.R.I.Cataldi,A.M.Salvi,T.Rotunno,E.Dicroce,Surface
andInterfaceAnalysis1992,18,623‐630.
[28] E.Breitmaier,G.Jung,OrganischeChemie‐Grundlagen,Stoffklassen,Reaktionen,
Konzepte,Molekülstruktur,5ed.,GeorgThiemeVerlag,Stuttgart,NewYork,2005.
[29] J. Moulder, W. Stickle, P. Sobol, HandbookofXRayPhotoelectronSpectroscopy,
1992.
98
VI FINALDISCUSSIONANDOUTLOOK
Severalheterogeneouscatalystshavebeentestedintheacidcatalyzeddehydration
offructoseto5‐hydroxymethylfurfural(HMF)withtheresultofaseveredeactivation
ofthematerialsalreadyafterthefirstreactionrun.Thecritical comparison of
measuredcatalyticactivitiestoliteraturedatahasshownalackofstabilitystudieson
heterogeneous catalyst applied in this reaction. Furthermore, the variety of catalytic
testconditionshampersthedirectcomparisonofdata.Oftenappliedmulti‐component
systems lead to additional complexity hindering the basic understanding of the
influence of single reaction parameters. We developed a one‐phase solution for the
dehydrationoffructose.Theuseof2‐butanolastheonlyreactionsolventinhibitsthe
productaccumulationonhydrophobiccatalysts(e.g.functionalizedcarbonnanotubes),
asconfirmedbyadsorptionstudies.Inaddition,therehydrationtothesideproducts
levulinicacidandformicacidaresuppressed.Theprocessavoids impurities in the
HMF‐productthatcanbeproblematicinsubsequentprocessingsteps,suchasS‐andN‐
containingsolventsorsaltresiduesappliedinthebiphasicprocess[1].Hencetheone‐
phase system in 2‐butanol would require lower purification costs than other
establishedprocesses.
Comparative studies on carbon based heterogeneous catalysts revealed three
different deactivating processes: (1) the leaching of instable acid functional groups,
(2)thesurfacecoverageorsideblockingbyinsolublepolymericbyproducts(humins)
FinalDiscussionandOutlook
99
and(3)thesurfacepassivationbythesolvent.Thestabilityofacidfunctionalgroups
was investigated by activity tests of the solvent after preconditioning of the catalyst
(leaching tests). It was found that leaching is a common problem for all post‐
functionalized,e.g.sulfonated,catalystsandappropriatepretreatmentsarerequired,in
ordertostartwithastablefractionofacidfunctionalgroups. Leaching tests are
essentialforthoroughproofofcatalyststabilityandhencearesuggestedtobeadapted
ascommontestforheterogeneouscatalystsinfructosedehydration.
Furthermore we established the comparison to the reference reaction, the
esterificationofaceticacidinethanol,assuccessfultoolfortheestimationofcatalyst
stability. The deactivation by leaching or solvent reactions can be traced by the
referencereaction.However,thesuccessfulrecyclingintheesterificationreactiondoes
notguaranteeastablecatalystinthefructosedehydration.Itcouldbeshownthatthe
preconditionedcatalystAGPs(sulfonatedamporphouscarbonfromglucosepyrolysis)
can be recycled successfully in catalyzing the conversion of the monofunctionalized
reactantinthereferencereaction.However,thecascadereactionfructosedehydration
(Figure1)cannotbecatalyzedinacontrolledway,avoidingtheformationofsurface
coveringhumins.
Underthefocusofgainingfurtherinsideintothesolventdeactivatingprocess(3),
in‐situXPSstudieshavebeperformed.Itwasshownthatthecatalystpretreatedinthe
solvent2‐butanolexhibitslessstructuralchangesintheO1saswellastheC1sspectra
while heating in vapor atmosphere. Consequently, we concluded ahigher
hydrophobicity for the pretreated sample. As possible passivating reaction the
esterificationofsurfacecarboxylicacidgroupsbythealcoholicsolventwasdiscussed.
Theformedestergroupsleadtoamorehydrophobicsurfaceterminationwhichcould
preventtheintrusionofwater.Asconsequenceforthedehydrationreaction,thelower
abundanceofacidfunctionalgroupscanexplainthedecreaseincatalyticactivity.
100
Figure1:Mechanismoffructosedehydrationcatalyzedbyhomogeneousacids
Futurestudiescouldbedirected into more mechanistic understanding of the
fructosedehydrationonheterogeneouscatalysts.Figure1showsthemechanismofthe
dehydration reaction of fructose (possible side reactions neglected) for the
homogeneously catalyzed case. In comparison to the homogeneous case, where the
proton concentration is constant over the total reaction medium, the protons at the
surfaceoftheheterogeneouscatalystarelocalizedatthe“activesite”.Theconstrained
localenvironmentoftheactiveprotonplaysamajorroleinthereactivityofthecatalyst
and the selectivity of the total reaction. The reactant can be stabilized in certain
conformations by complexation by surface functional groups. Additionally, side
reactionscanbeprovokediffurtherreactivegroupsaresituatedincloseproximityto
theactivesite.Sofar,neithertheactualnortheideallocalenvironmentofanactivesite
isknownforfructosedehydration.Somegroupsdiscussedthegeneral positive or
FinalDiscussionandOutlook
101
negative effect of additional Lewis acid sites on the reaction with controversial
conclusions[2].
Due to the incomplete knowledge on the requirements of the heterogeneous
catalyst, model systems carrying only one type of acid functional groups would be
needed. Preferentially, the support should be variable in pore sizeandinerttothe
reaction conditions, as well as to the reactant fructose. In the case of oxides, the
stabilityagainsthydrolysisislimitedandtheinfluenceofLewisacidandbasicsitesadd
complexity.Furthermore,theredoxactivitycannotbeexcludedforsomesystems,such
asthepromisingniobium‐basedcatalysts. For this reason we have chosen carbon
catalysts, although the recovery of by calcination is not possible.Undertheaimof
further mechanistic understanding of the heterogeneously catalyzed fructose
dehydration,carbonstaysaninterestingsupportmaterial.However,higherprecision
inthefunctionalizationprocessisrequiredwhichincludesthethoroughpurificationof
an–atbest–graphiticcarbonsupportandtheselectiveintroductionononlyonetype
offunctionalgroup.Thereareongoingresearcheffortstowardsthisdirectioninour
department.
Based on a controlled functionalization process, the optimal density of acid sites
couldbedetermined.Accordingtothepresentstateoftheliterature,itisnotknown
howmanyacidsitesparticipateintheformationofoneHMFmolecule.Itispossible
thatmorethanoneacidsitecanbeinvolvedinthecourseofthereaction.Hence,the
local proton density would affect the devolution of the cascade reaction of three
subsequent dehydration steps (Figure1)andconsequentlycanbeseenasan
importantselectivitycriterion.
Underthefocusoftheshorttermdevelopmentofaneconomicprocessforfructose
dehydrationtheuseofcalcinable,oxidiccatalystsprovideapossiblestartingpoint.The
useoftwoparallelreactorswouldprovidethepossibilityofalternatingprocessingand
reactivation/calcinationsstepsunderconstantHMFproduction.Alsohere,thebasisof
successfulprocessdevelopmentisthecarefulinvestigationofthedeactivationbehavior
ofthecatalyst.Typicaldeactivationproceduresandessentialstabilitytestshavebeen
describedinthisthesis.
102
[1]Y.Roman‐Leshkov,C.J.Barrett,Z.Y.Liu,J.A.DumesicNature.2007,447,982‐U985.
[2]R.Weingarten,G.A.Tompsett,W.C.Conner,Jr.,G.W.Huber JournalofCatalysis.
2011,279,174‐182.
i
VII SUPPLEMENTARY INFORMATION
Supplementary Information
ii
VII.1 Database Sample Numbers
Abbreviation
Material description
Sample Number
BT0
pristine Baytubes® from Bayer
3832, 8819
BTs Baytubes® functionalized by H2SO4 8234, 8032, 8879
BTsG200 Baytubes® gas phase functionalized in oleum
at 200°C 9121
BTsG600 Baytubes® gas phase functionalized in oleum
at 600°C 9187
BTn Baytubes® functionalized by HNO3 8240
BTb Baytubes® functionalized by B(OH)3 8241
BS-CNF PR24XT-HHT from Applied Sciences, grafted
by benzene sulfonic acid 12001
AGP amorphous carbon from glucose pyrolysis 9922, 10121, 10122
AGPs amorphous carbon from glucose pyrolysis
functionalized by H2SO4 9933, 10279
OMC (MC_0) ordered mesoporous carbon 10921
OMCs ordered mesoporous carbon functionalized by
H2SO4 10956
MC_TDP
(TDP0.2)
mesoporous carbon, 20mol% of building block
recorcinol replaced by thiodiphenyl 11108, 11454
MC_H2O2
(MC_1)
mesoporous carbon, functionalized by H2O2
at pH = 1 11466
MC_H2O2_2B
(MC_2)
mesoporous carbon, after functionalization by
H2O2 at pH = 1 and susequent treatment in 2-
butanol (130°C, 15 h)
11690
Nafion ®
Nafion® NR50 from Aldrich
10281
Amberlyst®
Amberlyst® 15 from Aldrich
12375
Sulfated zirconia Sulfated zirconia from MEL Chemicals 3006
H-mordenite
H-mordenite (H-MOR 14) from Südchemie
6188
H-ZSM5
H-ZSM5, Degussa, Si/Al=19, #KM-426
9378
NbPO
4
niobium phosphate from CBMM
12318
(VO)2P2O7 vanadyl pyrophosphate from Ecole
Polytechnique Montréal 12351
Supplementary Information
iii
VII.2 BET Isotherms
Table 1: BET isotherms of mesoporous carbon samples investigated by XPS
Sample
SN
BET isotherm
BJH poresize distribution
MC_0 10921
TDP0.2 11108
MC_1 11466
MC_2 11690
Supplementary Information
iv
VII.3 XPS
Figure 1: O1s fits (left panel) and C1s fits (right panel) of MC_0, MC_1 and MC_2
obtained by ex-situ XPS
Supplementary Information
v
Figure 2: Fits for O1s spectra of MC_1 and MC_2 during heating in vacuum
Figure 3: Fits for C1s spectra of MC_1 and MC_2 during heating in vacuum
Supplementary Information
vi
Figure 4: Fits for O1s spectra of MC_1 and MC_2 during heating in 0.1 mbar vapor
Figure 5: Fits for C1s spectra of MC_1 and MC_2 during heating in 0.1 mbar vapor
Supplementary Information
vii
Figure 6: Cl 2p spectra for MC_1 and MC_2 during heating in vacuum
Figure 7: Cl 2p spectra for MC_1 and MC_2 during heating in vapor
Supplementary Information
viii
Table 2: Quantification of carbon species in the O1s peak during heating in vacuum and
subsequent addition of water at 130°C (carbon fraction in %)
Sample Process
step 530.5 eV 531.2 eV 531.9 eV 532.7 eV 533.5 eV 534.2 eV Total oxygen
content
MC_1
RT 0.4 (4) 1.5 (14) 1.4 (13) 3.0 (27) 4.0 (36) 0.8 (7) 11.1
80°C 0.4 (4) 1.3 (13) 1.2 (12) 2.2 (22) 3.8 (39) 0.9 (9) 10.3
130°C 0.4 (4) 1.4 (15) 1.2 (13) 1.9 (20) 3.7 (39) 0.9 (10) 9.4
130°C aq 0.2 (2) 1.3 (15) 1.1 (13) 1.3 (15) 4.1 (48) 0.6 (7) 8.5
MC_2
RT 0.4 (3) 1.8 (15) 1.6 (13) 3.4 (28) 4.4 (36) 0.5 (4) 12.1
80°C 0.4 (4) 1.7 (16) 1.4 (13) 2.6 (24) 4.3 (40) 0.5 (5) 10.9
130°C 0.5 (5) 1.5 (15) 1.1 (11) 2.0 (20) 4.1 (42) 0.6 (6) 9.8
130°C aq 0.2 (2) 1.4 (17) 1.1 (13) 1.3 (16) 4.1 (49) 0.2 (2) 8.3
Table 3: Quantification of carbon species in the C1s peak during heating in vacuum and
subsequent addition of water at 130°C (carbon fraction in %)
Sample Process
step 284.4 eV 284.7 eV 285.2 eV 285.9 eV 286.6 eV 287.9 eV Total carbon
content
MC_1
RT 37.0 (42) 26.7 (30) 4.9 (6) 3.0 (3) 4.0 (5) 2.4 (3) 88.6
80°C 37.2 (42) 27.0 (30) 5.4 (6) 3.6 (4) 3.6 (4) 2.4 (3) 89.4
130°C 38.8 (43) 25.5 (28) 5.4 (6) 3.6 (4) 3.6 (4) 1.8 (2) 90.4
130°C aq 38.6 (43) 28.9 (32) 5.8 (6) 3.2 (4) 3.6 (4) 1.9 (2) 91.3
MC_2
RT 43.0 (49) 15.7 (18) 5.1 (6) 2.9 (3) 3.4 (4) 2.3 (3) 87.7
80°C 41.0 (46)
22.2 (25) 6.4 (7) 2.3 (3) 2.9 (3) 1.8 (2) 88.9
130°C 41.5 (46) 24.0 (27) 5.4 (6) 3.0 (3) 2.4 (3) 1.8 (2) 90.1
130°C aq 40.6 (44) 27.5 (30) 5.9 (6) 2.6 (3) 2.6 (3) 1.3 (2) 91.6
Supplementary Information
ix
Table 4: Quantification of carbon species in the O1s peak during heating in 0.1 mbar
vapor pressure (carbon fraction in %)
Sample Process
step 530.5 eV 531.2 eV 531.9 eV 532.7 eV 533.5 eV 534.2 eV Total oxygen
content
MC_1
RTaq 0.1 (1) 1.6 (13) 1.4 (12) 3.3 (28) 4.6 (39) 0.9 (8) 11.8
80°Caq 0.2 (2) 1.0 (12) 1.3 (15) 1.9 (22) 3.2 (37) 1.1 (13) 8.7
130°Caq 0.5 (5) 1.6 (16) 1.3 (13) 1.7 (17) 3.4 (34) 1.5 (15) 9.9
MC_2
RTaq 0.2 (2) 1.1 (12) 1.4 (15) 2.4 (26) 3.7 (26) 0.3 (3) 9.2
80°Caq 0.1 (1) 1.1 (11) 1.4 (14) 2.3 (24) 4.1 (42) 0.8 (8) 9.8
130°Caq 0.1 (1) 1.1 (12) 1.4 (1.6) 1.6 (17) 4.0 (42) 1.4 (15) 9.5
Table 5: Quantification of carbon species in the C1s peak during heating in 0.1 mbar
vapor pressure (carbon fraction in %)
Sample Process
step 284.4 eV 284.7 eV 285.2 eV 285.9 eV 286.6 eV 287.9 eV Total carbon
content
MC_1
RTaq 27.3 (31) 33.3 (38) 9.1 (10) 3.6 (4) 4.8 (6) 2.4 (3) 87.9
80°Caq 25.3 (28) 36.0 (40) 10.7 (12) 3.8 (4) 5.1 (6) 2.5 (3) 91
130°Caq 23.7 (26) 36.2 (40) 11.9 (13) 3.7 (4) 4.9 (6) 2.5 (3) 89.9
MC_2
RTaq 32.3 (36) 30.4 (34) 7.4 (8) 3.7 (4) 4.3 (5) 1.9 (2) 90.6
80°Caq 36.7 (41)
26.9 (30) 7.1 (8) 3.1 (3) 3.1 (3) 1.9 (2) 90
130°Caq 35.1 (39) 29.5 (33) 8.2 (9.1) 3.8 (4) 3.1 (3) 1.9 (2) 90.2
Acknowledgment
A good friend kept telling me: “If you really want something the whole universe will
come together and help you to make it happen.” This phrase gives a very suitable
expression for the dimensions of help and support during my PhD, I would like to
gratefully acknowledge for in fewer words than probably adequate.
First I would like to thank Prof. Robert Schlögl for the opportunity to work on this
challenging and multi-disciplinary project in such a great environment. The scientific
input as well as continuous encouragement and motivation helped me to keep track and
to overcome the one or other hurdle. Furthermore, I thank my two group leaders during
my time at FHI, Dr. Dangsheng Su and Dr. Malte Behrens for scientific as well as moral
support. I would like to thank my first supervisor Dr. Jean-Philippe Tessonnier for
sharing his broad knowledge in carbon research and catalysis. The open office
discussions facilitated my start at FHI and generated a fruitful spirit of interactive
exchange. In this regards, I am also grateful for my permanent office mate Weiqing
Zheng who patiently handled my curiosity and openly answered all kind of questions on
Chinese culture, which widened my view towards other perspectives. Special thank goes
to Dr. Alberto Villa for his supervision and his experience in the preparation of metal
colloids and liquid phase catalysis.
I would like to thank Nuruzatulifah Bt. Asari Mansor, Xiao Chen Zhao, Klaus Friedel
and Matthew Aronson for their collaboration in the material synthesis and the PIRE
program, in particular Keenan Deutsch, Prof. Brent Shanks and Prof. Robert Davis for
fruitful exchange of experience in biomass conversion chemistry. Furthermore, I highly
appreciate the collaboration with Prof. Sharifah Bee Abdul Hamid who gave me the
opportunity to use the high-throughput facilities at Universiti Malaya in Kuala Lumpur
for my experiments.
I am particularly grateful for the support by Dr. Edward Kunkes. His long-ranging
experience in the catalytic conversion of biomass, as well as the talent to oversee
complex problems with an immediate focus on critical points, gave me a big push
forward. I highly appreciate the scientific input and all the efforts in teaching me the
engineer’s point of view, as well as the time taken for the correction of my English
writing.
For their help in the analysis of the materials I would like to thank Gisela Lorenz for
the BET measurements, as well as Edith Kitzelmann and Dr. Andrey Tarasov for TG-MS.
The microscopy group, in particular Achim Klein-Hoffman, Gisela Weinberg, Norbert
Pfänder and Dr. Marc Willinger I would like to thank for their support and the
introduction into the microscopes. My special thank goes to Dr. Raoul Blume for
technical support at BESSY, the fits of the XPS spectra and intensive scientific
discussions.
I thank all members of the AC department of FHI for creating this characteristic
scientific spirit that I enjoyed very much. Countless discussions set new impulses in the
project and led to a continuous learning process. For my “daily lecture” during coffee
break or lunch I thank Christian Heine, Andreas Östereich, Klaus Friedel, Pierre Kube,
Stefan Zander, Gregor Wowsnick, Julia Neuendorf, Antje Ota und Steffi Kühl. In
particular in the last months some daily routine gave me stability and mental strength
in stressful times.
At the Technische Universität Berlin I thank Prof. Peter Strasser and Koteswara Rao
Vuyyuru for the collaboration and fruitful input. I am grateful that I could be a part of
UniCat and the BIG-NSE, and want to thank for the financial support, as well as the
scientific interactions. I send my special thanks to my BIG-NSE-batch, including Kirstin
Hobiger, Sara Bruun, Subhamoy Bhattacharya, Sardor Mavlyankariev, Stanislav Jaso,
Changzhu Wu, Manuel Harth and Carlos Carrero, as well as to Dr. Jean-Philippe Lonjaret
who spared no efforts to make this BIG-NSE time enjoyable and memorable.
From the beginning till the end, I could always rely on Steffi Kühl and Antje Ota who
offered help intuitively even before I could ask for. I wanna thank for all efforts and
support, and for the little reminders that prevented me from trouble multiple times.
I thank my family – my island of continuity, stability and love that I can always land
on. I wanna thank my parents and my brother Martin because they always stand beside
me, no matter what. And I thank my fiancé Christoph for his love and all the weight he
carried on his broad shoulders, in particular during the last months.
Danke
CURRICULUM VITAE
- Sylvia Reiche -
- born on November 24th 1983 in Oschatz (Germany) -
EDUCATION
From
7/2008
PhD thesis
Insights into Heterogeneous Catalysts for HMF synthesis
from Biomass
Prof. Dr. Robert Schlögl (co-
advisor, Fritz Haber
Institute), Prof. Dr. Peter Strasser (co-advisor, Technische
Universität Berlin)
Fritz Haber Institute
Berlin, Germany
Technische Universität
Berlin, Germany
From
10/2008
Berlin International Graduate School of Natural Sciences
and Engineering (BIG-NSE)
Technische Universität
Berlin, Germany
04/2008
M. Sc.
Preparation and Characterization of Silver Nanoparticles
for Printable Electronics
Prof. Dr. Dan Goia (co-advisor, Clarkson University), Prof.
Dr. Berthold Kersting (co-advisor, Universität Leipzig)
Universität Leipzig
Leipzig, Germany
Clarkson University
Postdam, NY, United
States
10/2005
B. Sc.
Determination of the Degree of Phosphorylation of the
TAU-Protein by Antibodies
Prof. Dr. Ralf Hoffmann
Universität Leipzig
Leipzig, Germany
PROFESSIONAL EXPERIENCE
10/2009
Research visit
Combinatorial Technologies and Catalysis Research
Centre (COMBICAT), Universiti Malaya
Prof. Dr. Sharifah Bee Abd. Hamid
Universiti Malaya
Kuala Lumpur, Malaysia
06/2006–
03/2007
Employee
R&D of Contact Materials (Dept: TM-ETM-FE),
preparation of silver dispersions, scale up
Dr. Bernd Kempf, Dr. Sebastian Fritzsche
Umicore AG & Co. KG
Hanau, Germany
02/2007
and
10/2007–
03/2008
Research visit
Center for Advanced Material Processing (CAMP),
cooperated partner of Umicore AG & Co. KG
Prof. Dr. Dan Goia
Clarkson University
Postdam, NY, United
States
HONORS AND AWARDS
10/2008 – 10/2011
Scholarship from Berlin International Graduate School of Natural Sciences
and Engineering (BIG-NSE), graduate school of the Cluster of Excellence
"Unifying Concepts in Catalysis" (UniCat)
08/2010
Invitation to BASF International Summer Courses 2010
REFERENCES
Oral presentations:
S. Reiche, M. Aronson, X. C. Zhao, K. Friedel, R. Blume, E. Kunkes, J.-P. Tessonnier, M. Behrens, D. S. Su,
R. Davis, R. Schlögl, Acidified Carbon Catalysts for Liquid Phase Reactions in Biomass Conversion Chemistry,
EuropaCat X, August 2011, Glasgow, Scotland
S. Reiche, M. Aronson, X. C. Zhao, K. Friedel, R. Blume, E. Kunkes, J.-P. Tessonnier, M. Behrens, D. S. Su,
R. Davis, R. Schlögl, Heterogeneous Catalysts in the Dehydration of Fructose to HMF, AIChE Annual Meeting,
October 2011, Minneapolis (MN), United States
Poster presentations:
A. Demund, D. Wett, S. Reiche, R. Szargan, R. Denecke, Study of formation and thermal stability of Fe
layers on ZnO surfaces
(Poster), DPG Annual Meeting, February 2008, Berlin, Germany
S. Reiche, M. Aronson, X. Ch. Zhao, E. Kunkes, J.-P. Tessonnier, M. Behrens, D. S. Su,
R. Davis, R. Schlögl,
Acidified Carbon Catalysts for Liquid Phase Reactions in Biomass Conversion
Chemistry
, 44. Jahrestreffen Deutscher Katalytiker, March 2011, Weimar, Germany
Publications:
S. Sahin, P. Maki-Arvela, J.-P. Tessonnier, A. Villa, S. Reiche, S. Wrabetz, D. S. Su, R. Schlögl, T. Salmi, D. Y.
Murzin,
Palladium catalysts supported on N-functionalized hollow vapor-
grown carbon nanofibers: The
effect of the basic support and catalyst reduction temperature, Applied Catalysis A – General
2011
, 408,
1
-
2, 137-147
S. Reiche, E. Kunkes, N. B. Asari Mansor, X. C. Zhao, K. R. Vuyyuru, A. Villa, J.-P. Tessonnier, D. S. Su,
M. Behrens, P. Strasser, R. Schlögl, Heterogeneous Catalysts in the Dehydration of Fructose to HMF (to be
submitted)
S. Reiche, X. C. Zhao, M. Aronson, K. Friedel, E. Kunkes, J.-P. Tessonnier, M. Behrens, D. S. Su,
R. Davis, S. B. Abdul Hamid, R. Schlögl, Deactivation Pathways of Carbon Catalysts in the Dehydration of
Fructose (to be submitted)
S. Reiche, R. Blume, X. C. Zhao, E. Kunkes, D. S. Su, M. Behrens, R. Schlögl, Reactivity of Mesoporous Carbon
Against Water: An In-Situ XPS Study (to be submitted)
„Groß ist die Aufgabe, die vor mir steht und
bescheiden sind die Kenntnisse und Kräfte, die für
ihre Bewältigung ausreichen sollen. Aber Aufgaben
sind da, um gelöst zu werden, und welcher
Schlachtruf wäre wohl besser geeignet, den
Ermatteten mit neuem Mut zu erfüllen, als das Wort:
Energie?“
Wilhelm Ostwald
