ContentslistsavailableatScienceDirect
Journal
of
Analytical
and
Applied
Pyrolysis
j o u r n al ho me p a g e :w ww . e l s e v i e r . c o m / l o c a t e / j a a p
Catalytic
pyrolysis
of
recalcitrant,
insoluble
humin
byproducts
from
C6
sugar
biorefineries
Shilpa
Agarwal
a,
Daan
van
Es
b,
Hero
Jan
Heeres
a,∗aChemicalEngineeringDepartment,ENTEG,UniversityofGroningen,Nijenborgh4,9747AGGroningen,TheNetherlands bFood&Bio-basedResearch,WageningenUniversityandResearchCentre,6700AAWageningen,TheNetherlands
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received13August2016 Receivedinrevisedform 13December2016 Accepted17December2016 Availableonline21December2016 Keywords: Pyrolysis Humins Catalyst HZSM-5 Aromatics Biobasedchemicals
a
b
s
t
r
a
c
t
Huminsaresolidby-productsformedduringtheacid-catalysedconversionsofC-6sugarstoplatform chemicalslikehydroxymethylfurfuralandlevulinicacid.Weherereportanexperimentalstudyonthe liquefaction/depolymerisationofhuminsusingcatalyticpyrolysis.Synthetichumins(SH)andcrude industrialhumins(CIH,includingpurifiedindustrial(PIH)samples)fromtheacid-catalysedconversionof C-6sugarstoHMF/LAweretested.Thermaldegradationpatternsofbothhumintypesvarysignificantly. Majorthermaldecompositionoftheindustrialhuminswasobservedbetween50and650◦C(weight lossapprox.66wt%),whereas,majorweightlosswasobservedbetween200and800◦Cforthesynthetic humins(47wt%).Aseriesofcatalyticpyrolysistestswithsynthetichuminsanddifferentzeoliteswere performedusingaPTV-GC/MS(humintocatalystwtratioof0.2,550◦C).Bestresultswereobtained usingHZSM-5(SiO2/Al2O3=50).Forquantitativeanalysis,agramscalepyrolysisunitwasused,giving
aproductoil(9–11wt%onhuminintake)withapproximately1.5and10wt%aromaticsfromsynthetic andcrudeindustrialhumins,respectively.GPCdataontheproductoilsclearlyshowsthebreakdown ofthehuminstructureintolowmolecularweightspecies.TheHHVvalueoftheliquidproducts(upto 41MJkg−1)isconsiderablyhigherthanthatofthecrudeindustrialhuminfeed(21–24MJkg−1).
©2016TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).
1. Introduction
Due to the availability and vast abundance, lignocellulosic biomassisconsideredasapotentialfeedstockforrenewablefuels andvaluablebulkchemicalproducts[1,2].Inthelastdecade sub-stantialeffortshave beenmadetodevelop economically viable valorisationroutestoconvertbiomasstovaluableproducts[1,3–6]. Celluloseandhemicelluloseforma majorpartof lignocellu-losicbiomass(70–80wt%).Oneofthepathwaysinvestigatedto dateistheconversionofthecelluloseandhemicellulosefraction intobio-basedplatformchemicalslike5-hydroxymethylfurfural (HMF)andlevulinicacid(LA)[7,8].HMFisconsideredaversatile chemicalbuildingblockforC-6compounds,suchas alkoxymethyl-furanics,2,5-furandicarboxylicacid,2,5-dimethylfuran,adipicacid, 1,6-hexanediolandcaprolactam[8–10].ThedemandforHMFas a platformchemical has already lead totheopeningof a pilot plantbytheDutchcompanyAvantiumwiththeaimtoproduce 40tonnes/yearofHMF,whichmaysubsequentlybeusedtomake
∗ Correspondingauthor.
E-mailaddress:h.j.heeres@rug.nl(H.J.Heeres).
PEF(polyethylenefuranoate),abio-based,recyclablepolymerthat canbeusedasasubstituteforPET[11,12].Furthermore,LAhas alsobeenidentifiedasanimportantplatformchemicalforthe pro-ductionoffueladditives,monomersforplasticsandtextiles,and chemicals[13,14].
BothHMFandLAcanbeobtainedbythedehydrationofC-6 sug-arsusingacidcatalysts,butthisinevitablyleadstotheformation ofhuminbyproducts[15,16].Dependingontheprocess param-eters,suchas typeof substrate,acidcatalyst, temperature,and time,huminyieldscanvaryfrom5to50wt%[7,17].Forinstance intheBiofineprocess,asemi-commercialprocesstoproduceLA fromlignocellulosicbiomass,huminformationmaybeupto20wt% onfeed,anditsformationposesconsiderablechallengesinterms ofscaleup(e.g.fouling,complexwork-up)[18].For HMForLA productiontobecomeeconomicallyviable,thereisneedtoeither optimisetheprocessesbysuppressingtheformationofhumins, or,alternatively,toupgradethesehuminbyproductsintovaluable chemicals.
In recent studies, it has been shown that humins are car-bonaceous,heterogeneous,poly-dispersematerialsconsistingofa furan-richpolymernetworkcontainingdifferentfunctionalgroups [7,9,17,19].Basedonthisknowledge,afewvalorisationapproaches http://dx.doi.org/10.1016/j.jaap.2016.12.014
havealreadybeensuggestedinliterature,includingthermaland catalyticgasificationaswellashuminintroductioninPFA (poly-furfurylalcohol)tomakeanelaboratecompositewithenhanced properties[12,17].
Inourpreviousworkwehavereportedthethermalpyrolysis,a matureandcommerciallyavailabletechnology[20–25],asa poten-tialroutetoconverthuminsintoapyrolysisoilenrichedwithlow molecularweightcompounds[7].Theaimofthisapproachisto convertasolid,lowenergydensityfeedintoaneasilytransportable fuelwithahighenergydensity.Thepyrolysisoilcanpotentiallybe usedasaco-feedinpowerstations,asaboilerfuel,ortoisolate highaddedvaluebulkchemicals,suchasaceticacidandformic acid.Catalyticversionsofpyrolysistechnologyhavebeen devel-opedrecentlyforarange oflignocellulosicbiomasssourcesand areconsideredaninterestingroutetoobtainlowmolecularweight aromaticslikebenzene,toluene,andxylene(BTX)[26–29].Wehere reportastudyonthecatalyticpyrolysisofhumins.Ouraimisto obtainliquefiedproductswithahighcarbonefficiencyandwhich are potentiallyenriched in valuable bulk chemicals likeBTXNE (benzene,toluene,xylenes,naphthalene,ethylbenzene).Catalytic pyrolysishastothebestofourknowledgenotbeenusedforhumin valorisation.Thisrouteismotivatedbythefactthatithasbeen shownthatfuranics,whichareconsideredthemainbuildingblocks ofhumins,canbesuccessfullyconvertedtoaromaticsinthe pres-enceofa zeolite catalyst [30].In this work,we haveused two typesofhumins:synthetichumins(SH)obtainedbytreatmentof d-glucoseinaqueousmediawithastrongmineralacids,andcrude industrialhumins(CIH)obtainedfromAvantiumChemicals,the byproductoftheconversionofC-6sugarstoHMFinmethanolas thesolvent.InadditionapurifiedformoftheAvantiumhumins (PIH)wasalsotested.
2. Materialsandmethods
2.1. Chemicals
d-Glucose(99%)andsulphuricacid(95–97%)usedfor synthe-sisofsynthetichuminswerepurchasedfromSigmaAldrichand MerckKGaA,respectively. Tetrahydrofuran(THF)wasusedasa solventandwasobtainedfromBoomB.V.Allthezeolitecatalysts (Ferrierite20,HY5.1,NaY5.1,HY80,Mordenite20,Na-Mordenite 13,Beta-25,HZSM-5(SiO2/Al2O3=23,50,80))usedinthisstudy
werepurchasedfromZeolyst.Relevantpropertiesofthecatalysts areprovidedinTablesS1andS2(Supplementaryinformation).The hydrotalcitecatalystwasobtainedfromSigmaAldrich.
2.2. Industrialhumins
Crudeindustrial(fromAvantium)humins(CIH)wereobtained asthewastestreamfromthecatalyticdehydrationofa carbohy-dratefeedstockinmethanoltoproduce,methoxymethylfurfural (MMF)andmethyllevulinate.ThisprocessisoperatedbyAvantium ChemicalsattheirpilotplantfacilitesinGeleen,TheNetherlands. Theindustrialhuminswerefurtherpurifiedusingsolvent extrac-tions to remove the residual monomers (MMF, levulinic acid). Detailsoftheextractionprocedureandcharacterisationofthe puri-fiedindustrialhumins(PIH)aregiveninreference[31].
2.3. Preparationofsynthetichuminsamples
Thesynthetichumins(SH)usedinthisstudyweresynthesised bythe acidcatalysed hydrothermal reactionof d-glucose(1M) using0.01MH2SO4[7,19].Theclearsolutionofthesugarandacid
wastransferredtoastainlesssteelautoclave(1L)equippedwithan overheadstirrerandflushedwithN2beforeheating.Thesolution
wassubjectedtoahydrothermaltreatmentat180◦C(heatingrate:
1.3◦C/min;stirringrate:120rpm)for8h.Themaximumpressure intheautoclaveat180◦Cwas15–18bar.Afterthehydrothermal treatment,theprecipitateswereseparatedusingvacuumfiltration, theresiduewaswashedwith3Ldeionisedwaterandthenvacuum driedat60◦Covernight.Driedsynthetichuminsweregroundand purifiedusingaSoxhletextractionwithwaterfor24h.Finally,the purifiedhuminswerevacuumdriedfor24hat60◦Candground intoapowder.
2.4. Analyticalmethods
Elementalanalyses(C,H,N,O,andScontent)ofthehuminsand residueaftercatalyticpyrolysiswerecarriedoutusingan auto-matedEuroVectorEA3000CHNSanalyserwithacetanilideasthe calibrationreference.Theoxygencontentwasdeterminedbythe differenceofCHNS.Allsampleswereanalysedatleastinduplicate andtheaveragevaluesarereported.
Thermogravimetricanalysis(TGA)ofallthehuminsampleswas performedusingaMettler–Toledoanalyser(TGA/SDTA851e).The huminsampleswereheatedfrom30to900◦Cinnitrogen atmo-sphere,withaheatingrateof10◦C/min.
BETsurfaceareasofthezeolitesweredeterminedbyperforming N2physisorptionexperimentsusingaMicromeriticsTristar
instru-ment.Thesamplesweredegassedinvacuumat300◦Cfor24hprior totheanalysis.
Temperatureprogrammeddesorptionofammonia(NH3-TPD):
NH3-TPD measurements were performedusing a Micrometrics
AutoChem II 2920 system (Micromeritics, Norcross, GA, USA). BeforeaTPDexperiment,thecatalyst(∼60mg)wasdegassedat 600◦Cataheatingrateof10◦C/mininaHeflowfor60min.Then thesamplewascooledto100◦CunderaHeflow.Inthenextstep, thesamplewassaturatedwithagascontaining1%ofNH3inHeflow
at100◦Cfor60min.Then,thesamplewaspurgedwithaHeflowat 100◦Cuntilaconstantbaselinelevelwasattained.TheTPD mea-surementwasperformedinthetemperaturerange100–600◦Cata rateof10◦C/minusingHeasthecarriergas.TheevolvedNH3was
detectedbyanon-linethermal-conductivitydetector,calibratedby thepeakareaofknownpulsesofNH3.
2.5. Catalyticpyrolysisexperiments 2.5.1. Small-scale(catalytic)pyrolysis
Small-scale(mg-scale)pyrolysisofhuminswasperformedusing aPTVinjector(programmabletemperaturevaporiser,Model:Optic 2fromAtlas)placedonaHP5890GCSeriesIIsystemin combina-tionwitha HP5972MSdetector(PTV-GC/MS)controlledbyan Optic2device(Fig.1).Forcatalyticexperiments,thehuminwas firstloadedinthesamplevialfollowedbythecatalyst.Typically,for thermalpyrolysis,0.5mgofhuminsamplewasused,whereasfor catalyticpyrolysis,ahumintocatalystweightratioof0.2wasused inmostexperiments(catalystscreeningusingsynthetichumins). However,forcatalyticpyrolysisexperimentscomparingdifferent humintypes,ahumintocatalystweightratioof0.05wasuseddue togeometricalconstraintsofthesamplevial.Itmustbenotedthat nosignificantdifferenceinproductyieldswereobservedwhenthe humintocatalystratiowaschangedfrom0.2to0.05(seeFig.S1in Supplementaryinformation).
Basedonthepreliminarypyrolysistests,weobservedmaximum volatileyieldsat550◦C(resultsnotshown)andhence550◦Cwas usedasthepyrolysistemperatureinthisstudy.Thepyrolysis tem-peratureprogrammewassettorampfrom40◦Cto550◦Catarate of16◦C/sfollowedbyanisothermalperiodat550◦Cfor60s.After (catalytic)pyrolysisintheinjector,theproductsweretransferredto acapillarycolumn(AgilentTechnologiesVF–5ms,30×0.25×1.0, split50:1).Heliumwasusedasthecarriergasataflow rateof 1mL/min.ForGCmeasurementsthefollowingtemperatureprofile
Fig.1.PyrolysisGC–MS(smallscale)setup.
wasused:initialtemperatureof40◦Cfor60sfollowedbyheating witharateof10◦C/mintothefinaltemperatureof250◦C.TheMS detectorwasoperatedinthescanrangeofm/z35–400usingthe electronionisationmode(70eV)withaninterfacetemperatureof 280◦C.PyrolysisproductswereidentifiedusingtheNIST05amass spectralibrary.
Inordertogetasemi-quantitativeestimateofthe concentra-tionofGC/MSdetectedcompounds,thetotalionchromatogram (TIC)peakareasoftheMSdetectorwereusedasan approxima-tion.ItwasassumedthattheTICpeakareapercentageofacertain compoundislinearlycorrelatedtotheconcentrationofthe corre-spondingcompound.Therefore,tocalulatethesemi-quantitative yieldsofeach compound/groupofcompounds,the correspond-ingTICpeakareapercentagewasdividedbytheTICpeakareaof alltheGCdetectables.Thismethodolgyhasoftenbeenreported foranalysingthecompositionofbiomass-derivedpyrolysis prod-uct stream. For simplicity, the compounds, except BTXE, were classifiedbasedonchemicalfunctionalities,e.g.,aromatic deriva-tives(substituted benzeneandpolycyclicaromatics),furansand derivatives,napthalenes(napthaleneandsubsitutednapthalenes), phenolic derivatives (substituted), and others (mainly alkanes, alkenes,aceticacid,indanesetc).
TodeterminetheyieldsofBTXNEusingPTV-GC/MS,anexternal standardsolutionwithknownquantitiesofBTXNE,wasusedto cal-culatetheresponsefactor.Itwasassumedthattheconcentration ofthecompoundsis lineralyproportionaltothetotalion chro-matogrampeakarea.Ameasurementusinganexternalstandard solutionwasperformedpriortoeachsetofexperiments.
2.5.2. Gram-scale(catalytic)pyrolysis
Gram-scalepyrolysiswasperformedusinganin housebuilt reactorsetupconsistingoftwostainlesssteelreactors(one reac-torforthehuminsandoneforthecatalyst,Fig.2).Thepyrolysis productsfromthepyrolysisreactorpassthroughthecatalystbed
forcatalytic upgradingofthevapourstream.Immediatelyafter thecatalytic reactor,thecondensablevapourswerecollectedin twoconsecutivecondensorsimmersedinisopropanol/liquid nitro-genmixturesmaintainedat−50◦C.Non-condensablevapourphase
productswerecollectedinagasbag(SKCTedlar0.5Lsamplebag). Nitrogen(Linde3.0,11mL/min)wasusedasacarriergas.For heat-ing,thewholereactorsystemwasimmersedinafluidisedsand bathmaintainedat570◦C.Thedurationofthepyrolysisreaction was10min.Massbalancecalculationswereperformedby deter-minationoftheweightdifferenceofthereactors,coolersandtheir connectionsbeforeandafterpyrolysis.Theamountofnon condens-ables(gasphase)wascalculatedbyweighingtheamountofwater displacedwhenthegasbagwascarefullysubmergedinabeaker filledwithwater.Theliquidproductswerecollectedfromthe cool-ersandreactortubesbyrinsingwithTHF.Asthewaterandoilare bothmiscibleinTHF,asinglephaseproductliquidwasobtained.
Thewatercontentinthepyrolytichuminoilsobtainedafter gram-scale pyrolysis wasdetermined by Karl Fischer titrations using a Metrohm Titrino 758 titration device. A small amount ofsample(approximately0.02–0.06g)wasaddedtoanisolated glasschambercontainingHydranal®(KarlFischerSolvent,Riedelde Haen).ThetitrationswerecarriedoutusingtheKarlFischertitrant Composit5K(RiedeldeHaen).ThewatercontentoftheTHFwas alsomeasuredandsubtractedfromthefinalvalueofthepyrolytic huminoil.Allmeasurementswereperformedinduplicateandthe averagevalueisreported.Theamountoftheorganicphasewas cal-culatedbysubtractingthewatercontentoftheproductfromthe totalproductliquidobtainedafterreaction.
For quantification of liquid phase components, the humin oilsobtainedaftergram-scaleexperimentswereanalysedusing GC/MS-FID. GC/MS spectra were recorded using a Quadrupole Hewlett-Packard 6890 MSD connected to a Hewlett-Packard 5890gas chromatograph(GC) witha Restek RTX-1701column (60m×0.25mmi.d.and0.25mfilmthickness)andflame ioni-sationdetector(FID).Injectionanddetectionwereperformedat 280◦C,usingoventemperatureheatingprofilesfrom40to250◦C atarateof3◦C/min.PeakidentificationwasdoneusingtheNIST05a massspectralibrary.Quantificationofproductsinpyrolytichumin oilwasdeterminedbyusingaveragerelativeresponsefactor(RRF) ofeachcomponentcalculatedwithrespectton-nonane(internal standard).
Gasphaseswerecollectedduringcatalyticpyrolysisinagasbag (SKCTedlar0.5Lsamplebag(6×6))withapolypropylene sep-tumfitting.Quantificationofthemaingasphasecomponentswas doneuisngGC-TCDanalyses.ThesewereperformedonaHewlett Packard5890SeriesIIGCequippedwithaPoraplotQAl2O3/Na2SO4
columnandamolecularsieve(5Å)column.Theinjector tempera-turewassetat150◦Candthedetectortemperaturewasmaintained at90◦C.Theoventemperaturewaskeptat40◦Cfor2minthen heatedupto90◦Cattherateof20◦C/minandkeptatthis temper-aturefor2min.Areferencegaswasusedtoidentifythepeaksby retentiontimeandtoquantifytheproductsinthegasphase(55.19% H2,19.70%CH4,3.00%CO,18.10%CO2,0.51%ethylene,1.49%ethane,
0.51%propyleneand1.50%propane).
GPC analyses were performed on the humin oils using an HP1100equippedwiththree300×7.5mmPLgel3mMIXED-E columns(temperaturemaintainedat42◦Cduringmeasurements) inseriesusingaGBCLC1240RIdetector.Averagemolecularweight calculationswereperformedwiththePSSWinGPCUnitysoftware fromPolymerStandardsService.Polystyrenesampleswereused asthecalibrationstandard.ForGPCmeasurements,THFwasused aseluentataflowrateof1mL/minat140bar,withaninjection volumeof20L.AllthesamplesbeforeGPCanalysiswerefiltered using0.2mPTFEfiltertotrapparticles,ifany,thatmayaffectthe analysis.
Fig.2. Pyrolysisgramscalesetup.
3. Resultsanddiscussion
3.1. Characterisationofthehuminsamples 3.1.1. Elementalcompositionofthehumins
Thesynthetichumincontainsapproximately66wt%[C],30wt% [O]and4wt%[H],seeTable1fordetails.Thesevaluesagreewell with the reported values in the literature [7,19]. The elemen-talcompositionofthesynthetichuminsissignificantlydifferent fromthanthatofthestartingmaterial(d-glucose)dueto dehy-dration/condensation reactions involving sugar molecules and intermediates[7,8].
For theindustrial humins, thecarbon contentis lower than forthesynthetichumin,whereasthehydrogenandoxygen con-tents are higher. The H/C and O/C ratios of industrial humins wereobservedtobesimilartothoseforHMF,seeTable1.These observationsindicatethattheindustrialhuminsamplesareless dehydratedthanthesynthetichumins.
3.1.2. Thermogravimetricbehaviourofthehuminsamples
Togainmoreinsightintothethermalbehaviourofthehumin samples,thermogravemetricanalyseswereperformedinaninert (N2)atmosphere.Thechangeintheweightofthehuminsamples
asafunctionoftemperatureisshowninFig.3.Itisclearlyevident thatthethermalbehaviourofthesyntheticandindustrial (includ-ingpurified)huminsamplesareverydifferent.Initialweightlosses startsaround50◦Cfortheindustrialhuminsamplesandlevelsoff ataround650◦C(weightlosswasapproximately66wt%).Forthe purifiedindustrialhumins(PIH)thethermaldecomposition pat-ternisquitesimilartothecrudesample,howeverthetotalweight lossisabout57wt%.Thisdecreaseinweightlosscomparedtothe crudesampleindicatesthatlowmolecularweightcompoundsare presentinthecrudeindustrialhuminsample.Forsynthetichumins, themajorweightlosswasobservedbetween200and800◦C(about 47wt%).Theseobservationsindicatethatthegasandliquidyields forindustrialhumins(includingPIH)duringpyrolysisareexpected tobehigherthansynthetichumins.
Ascanbeseenin Fig.3 (rightplot),thehighestdegradation rateforthesynthetichuminswasobservedat430◦Cwhereasfor industrialhuminssharppeakswereseenat180and200◦C.The presenceofthissharpdegradationpeakincaseofindustrialhumins pointstowardsthepresenceofconsiderableamountsofvolatile,
Table1
Elementalcompositionofthehuminsamplesandrelevantmodelcompounds.
Humins C(wt%) H(wt%) O(wt%) H/C O/C Synthetic 65.7 4.4 29.9 0.80 0.34 Crudeindustrial 56.7 5.4 37.9 1.12 0.50 HMF 57.1 4.8 38.1 1.0 0.50 Levulinicacid 51.7 6.9 41.3 1.6 0.6 Glucose 40 6.7 53.3 2.0 1.0
lowmolecularweightcomponents.Afterpurificationofthe
indus-trialhumins,thesharppeakat180◦Cdisappearsandthispoints
towardsthepartialremoval of(trapped) low molecularweight
compounds.
3.2. Thermalpyrolysisofhumins
Thermal, non-catalyticpyrolysis wasperformed on
represe-nativesamplesat550◦C usingPTV-GC/MS(mg-scale)toobtain
insightinthecompositionofthevapourphase.Thechromatograms
forallthehuminsamplesareshowninFig.4.Mainlyfuranics,such
as2-methylfuran,2,5-dimethylfuran,1,2-furanylethanone were observed inthe case ofsynthetic humins,confirming that syn-thetichuminscontainaconsiderableamountoffuranicmoieties [7,19].Interestingly,fortheindustrialhuminsandPIH,inaddition tothefuranics,monomericspecieslikeHMF,levulinicacidandtheir methylesterswerealsoobserved.Thisobservationindicatesthat industrialhuminsnotonlyconsistofcondensedfuranicsringsbut alsocontaintrappedmonomericintermediate/products.Thisisin agreementwiththeelementalanalysis(Table1)andTGA(Fig.3) data.Furthermore,afterpurificationoftheindustrialhumins,these monomerswerestillpresentinthevapourphase,whichindicates thatsomeofthesemonomersareactuallyincorporatedinthecore structureofthehumins.
3.3. Catalyticpyrolysisofhumins
3.3.1. Catalyticpyrolysisusingsynthetichumins
Aseriesofpyrolysistestswithsynthetichuminsanddifferent zeoliteswereperformedinthePTV-GC/MS(humintocatalystwt ratioof0.2)at550◦Ctoidentifythebestcatalysttoobtainliquefied productswithahighcarbonefficiencyandpotentiallyenrichedin valuablebulkchemicalslikeBTXNE.
Fig.3.TGA(left)andDTG(right)plotsofcrudeindustrial(CIH,reddashline),purifiedindustrial(PIH,bluedash-dot)andsynthetic(SH,blacksolidline)humins.(For interpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)
Fig.4. PyrolysisGC/MSchromatogramofSH,PIHandCIHat550◦C.
Asmentioned earlier,furanicswerepredominantly observed afterpyrolysisofsynthetichuminsat550◦Cintheabsenceof cat-alyst,andnotraceofaromaticswasobserved(Fig.4).However, onpyrolysisat550◦Cinthepresenceofacatalyst,theproduct distributionsignificantlychanged,ascanbeseeninFig.5and con-siderableamountsofBTXNEareformed,theexactamountbeinga functionofthezeoliteused.
Webelievethatthehuminsfirstdepolymerizethermallyaswell ascatalyticallyontheoutsidesurfaceofthezeolites.These depoly-merizedfragmentsreactfurtheratthesurfaceandinsidethepores togivearomaticproducts.Asimilarmechanismhasbeenreported forthecatalyticcrackingofpolymerslikelowdensitypolyethylene onzeolites[37].Thearomatics aremostlikelyformedby Diels-Alderadditionreactionsbetweenfuranicsandolefines(Scheme1). Inamodelcompoundstudy[30,38],ithasalreadybeen demon-stratedthataromaticsareformedbyco-feedingolefinsandfuranics overforinstanceaHZSM-5catalyst.
Remarkable differences in aromatics (BTXNE) yields were observedforthevariouscatalysts.LowestBTXNEyieldswerefound
forNaY,Na-mordeniteandthebasichydrotalcitecatalyst(Fig.5). Thesefindingsrevealthatacidic,protonatedzeolitesaredesirable for thearomatizationof thepyrolytic vapours[32–35].For the acidiczeolites,theBTXNEyieldsincreaseintheorder:Ferrierite 20<HY–80<HY5.1<H-Mordenite20<H-ZSM-5(Fig.5).These dif-ferencesareexpectedtoberelatedtozeolitepropertieslikepore dimensionsandaverageporesize(TableS1),BETsurfaceareaand acidity(TableS2).
Inthisrespect,thepoorperformanceofFerrierite20maybedue tothesmallporedimensionsandtheporesizeofthiszeolite(Table S1).Thisisagreementwithliteraturedata,wherethelowaromatic yieldforthiszeoliteisassociatedwithdiffusionlimitationsinthe catalystporesystemduetothesmallporesizes[32].
AclearrelationbetweentheBETsurfaceareaofthecatalysts (TableS2)andtheBTXNEyieldisabsent,thoughthehighestyields wereobtainedwithzeoliteswithhighBETsurfaceareas.Upon sug-gestionofoneofthereviewers,acatalyticpyrolysisexperimenthas beenperformedwithBeta-25zeoliteinthemgscaleset-up.The
Scheme1.Proposedreactionpathwayforthecatalyticpyrolysisofhumins.
Fig.5. RepresentativeGCdetectablemonomericvolatilesreleasedduringthecatalyticpyrolysisofsynthetichumins(SH)withdifferentzeolitecatalysts(Humintocatalyst wt.ratioof0.2)at550◦C.
BTXNEyieldwasabouthalfofthatobservedforH-ZSM-5(Fig.S2, Supplementaryinformation).
ThegoodperformanceoftheH-ZSM-5catalystsislikelyrelated toitsacidityincombinationwiththeporestructure/sizes.Studies intheliteraturehave alreadyshownthatHZSM-5hasthe opti-malzeoliteporestructuretoobtainthehighestyieldsofaromatics fromthecatalyticpyrolysisofbiomass[36].Assuch,itappearsthat HZSM-5hastheoptimumgeometryforfuranics,theprimary ther-malproductsfromhuminpyrolysis,toentertheporesandtobe convertedintoaromatics.
3.3.2. Catalyticpyrolysisofsyntheticandindustrialhuminsusing HZSM-5
AsHZSM-5wasshowntobethemostpromisingcatalystfor thesynthetichumins,thenextstepinthisstudywastoinvestigate theeffectoftheacidityofHZSM-5duringthecatalyticpyrolysisof bothindustrialandsynthetichumins.Weperformedthecatalytic pyrolysistestsusingHZSM-5withdifferentSiO2/Al2O3ratios(23,
50and80)andassuchwithdifferentacidity(highestfor23).It isapparentfromFig.6thatHZSM-5withaSiO2/Al2O3ratioof50
givesthehighestyieldofaromaticsforbothtypesofhumins.Witha furtherincreaseintheacidity(from50to23)adecreaseinthe aro-maticyieldwasobserved.ItisevidentthathigherBrönstedacidity isdesirableforaromaticyields,howeveratoohighvalueleadsto loweryields.
Toestimatetheyieldsofaromaticsonhuminintake,an exter-nal standard solution with known BTXNE concentrations was used.Basedonthesmall-scalepyrolysis(Py-GC/MS)experiments (Fig.7),theBTXNEyieldissignificantlyhigherforindustrialhumins (10wt%)incomparisontothesyntheticone(1.5wt%).Asexpected forPIHsamples,theBTXNEyields(5wt%)areintermediatebetween thecrudeindustrialandsynthetichumins.Wefurtherdetermined theBTXNEyieldforkraftligninundersimilarpyrolysisconditions withtheincentivetocomparetheyieldsforhuminsandlignin,the latterbeingconsideredapromisingsourceforBTXNE.Withkraft lignin,aBTXNEyieldof6wt%(onligninintake)wasobtainedby catalyticpyrolysis.ThehigherBTXNEvaluesforindustrialhumins clearlyindicatethatindustrialhuminshavepotentialtobeusedas feedsfortheproductionofbio-basedaromatics.
During catalytic pyrolysis,the colorof theHZSM-5-50 cata-lystchangedfromoffwhite/greytoblack,anindicationforcoke formation,whichiswellestablishedforcatalyticpyrolysisof ligno-cellulosicbiomass.Tocheckwhethercokeremovalispossibleand doesnotleadtoirreversibledeactivation,thespentcatalystwas calcinedat550◦Cinairfor4h.Nodifferenceinthecatalyticactivity wasobservedwhenusingtheregeneratedcatalyst,aswasevident fromtheamountsoflowmolecularweightproductsformed. Fur-therworkissuggestedtoinvestigatethelongtermstabilityofthe catalystafterrepeatedcatalytic-oxidativework-upcycles.
Fig.6.EffectoftheSiO2/Al2O3ratiooftheHZSM-5zeoliteontheproductdistributionforthecatalyticpyrolysisofcrudeindustrial(CIH)andsynthetic(SH)humins(Humin
tocatalystweightratioof0.05)at550◦C.
Fig.7.BTXNEyieldsonhuminintakeobtainedduringthepyrolysisofsynthetic(SH),crude/purifiedindustrial(CIH/PIH)huminsandkraftlignin(KL)withHZSM-5-50 (Humin:Catalystweightratio=1:20)at550◦C.
3.4. Gram-scalepyrolysis
Inadditiontothesmallscalecatalyticpyrolysisexperiments, catalytic pyrolysisexperiments wereperformedwith industrial and synthetic humins at gram-scale to obtain humin oils for detailedproductanalysisandtoestimatethemassbalance clo-sures.ItmustbenotedthatTHFwasusedtocollecttheproducts fromthereactortubesandcoolersandthishamperstheaccuracy oftheoilyielddeterminations.
Based on mass-balance calculations (Fig. 8), an oil yield of approx.11–14wt%wasobtainedforthecatalyticpyrolysisofboth industrialandsynthetichumins.Thesolidresidueforthesynthetic
humins(58wt%)afterreactionismuchhigherthanforthe indus-trialones(38wt%onhuminintake),inagreementwiththeTGA data(Fig.3).
TheoilphasewasfurthercharacterisedusingGC/MS-FIDtogain detailedinsights intoproductyieldsandcomposition (Table2). Thetotalamountofdetectablearomaticspeciesforindustrialand synthetic humins were at max. 9 and 2wt% on huminintake, respectively, which confirms the results from the PTV-GC/MS experiments.Asalargefractionofthehuminoils(especiallyinthe caseofsynthetichumins)wasnotobservedbyGCandislikely non-volatileandofhighermolecularweight,wefurtheranalysedthe productoilsusingGPC(Fig.9).TheGPCspectradonotclearlyshow
Fig.8. Productdistributionforthecatalyticpyrolysisofcrudeindustrial(CIH)andsynthetichumins(SH)withHZSM-5-50at550◦C(Humintocatalystweightratioof0.2)
intheg-scalereactor.
Fig.9.Gelpermeationchromatograms(GPC)ofhuminoilsobtainedafternon-catalyticandcatalytic(withHZSM-5-50)pyrolysisofcrudeindustrial(left)andsynthetic (right)huminsat550◦C.Molecularweightvaluesarerelativenumbersbasedonpolystyrenestandards.
Table2
Distributionofpyrolysisproducts(wt%onhuminintake)forthecatalyticpyrolysis ofvarioushuminssamplesusinggramscalepyrolysis.
Compoundclass CrudeInd.humins(CIH) Synthetichumins(SH)
Aromatics 5.64 1.35 Napthalenes 2.00 0.70 Indane/Indenes 0.80 0.01 Alkanes 0.09 0.01 Polyaromatics 0.26 0.11 Total 8.79 2.18
thepresenceofhighermolecularweightspeciesintheproductoils,
thoughsometailing,associatedwithhighermolecularweight
com-pounds,cannotbeexcluded.Assuch,thedifferencesinoilyield
basedonmassbalancecalculationsandtheGCyieldsforBTXNE
arelikelyduetothecomplexwork-upoftheoilphase(usingTHF)
afterpyrolysiswhichispronetosomeexperimentalerror.
AcomparisonbetweentheGPCdataforthecatalyticand
non-catalyticpyrolysisclearlyrevealsthatthepresenceofacatalysts
leadstoareductionofthemolecularweightoftheproductoils.As
such,thecatalystsnotonlycatalysedeoxygenationreactionsbut
alsoareabletobreakdownandcrackthepolymericstructureof
thehumins.
Finally,theelementalcompositionsofhuminfeedsandthesolid
residuesobtainedaftercatalytic pyrolysisweredeterminedand
resultsareplottedintheformofavanKrevelendiagram(Fig.10).
BothhuminfeedshavesignificantlydifferentH/CandO/Cratios, howeverafterpyrolysisthecompositionoftheresiduefromboth huminsourcesisquitesimilar.Inbothcases,theoxygenand hydro-gencontentsdecreasessignificantlyuponpyrolysis.
Inaddition,thehigherheatingvalue(HHV)ofthehuminsand pyrolytichuminoilswasestimatedusingtheMilneequation(Eq. (1),whereC,H,OandSarecarbon,hydrogen,oxygenandsulphur in weight percentages, respectively. The Milne equation semi-quantitativelypredictstheHHVoffuelsorfuelresourcesobtained fromcoal,biomass,pyrolysisoilandbiodieselusingelemental com-positions.ForHHVcalculationoftheproductoil,wecalculatedthe HHVvalueusingtheelementalcompositionoftheproducts
identi-Fig.10.vanKrevelendiagramofthehuminfeedsandsolidresiduesobtainedaftercatalyticpyrolysis.
fiedbyGC–MS/FID(seeTableS3insupplementaryinformationfor details).
HHVMilne
MJ.kg−1
=0.341×C+1.322×H−0.12×O−0.12×N+0.0686×S−0.0153×residue (1) TheHHVvaluetheforindustrialandsynthetichuminswas cal-culatedtobe21and24MJkg−1.Aftercatalyticpyrolysis,theHHV valuesignificantlyincreasedtoabout41MJkg−1forthehuminoils (industrial/synthetic).Moreover,theHHVofthecharobtainedafter pyrolysisofhuminswasfoundtobeabout30MJkg−1,which is againhigherthanthestartingfeedvalueandcanhencebe poten-tiallyusedasaco-feedinpowerstationsorforcombustionduring pyrolysis.Theseresultsclearlyprovidetheproofofprinciplethat liquefactionofhuminsbycatalyticpyrolysistogivealiquid prod-uctwithahigherenergydensityiswellpossible,thoughtheyield ofhuminoils(11–14wt%)islow.
4. Conclusions
Wehaveprovidedtheproofofprinciplefortheconversionof humins(industrialandsyntheticones)tohuminoils(11–14wt%on huminintake)enrichedinvaluablearomaticsusingcatalytic pyrol-ysis.HZSM-5(SiO2/Al2O3=50)wasfoundtobethemostpromising
catalyst.Quantitativeanalysisofthehuminoilsshowsthepresence ofapproximately2and10wt%ofaromaticsbasedonsyntheticand industrialhuminintake,respectively.TheHHVvalueofthehumin oilswasestimatedtobeabout41MJkg−1.
Themain componentsin thehuminoils arehighvalue low molecularweightaromatics.Assuch,theproductshavepotential tobeusedasanadditiveforbiofuels.Inaddition,isolatedindividual componentsmayserveasbuildingblocksforrenewablepolymers. Basedonthestoichiometryofthereaction,themaximum theorit-icalyieldofaromaticsfromthe(catalytic)pyrolysisofhuminsis approximetely50wt%onhuminintake.Inthisworkwereportan aromaticyieldof9wt%fortheindustrialhuminsamples,whichis about18wt%ofthetheoreticalmaximum.
Acknowledgements
Thisresearchwasperformedwithintheframeworkofthe Catch-Bioprogram.Theauthorsgratefullyacknowledgethesupportof theSmartMixProgramoftheNetherlandsMinistryofEconomic Affairs,AgricultureandInnovationandtheNetherlandsMinistry
of Education, Culture and Science, project number 053.70.358. AuthorsacknowledgeAvantium Chemicals BV,TheNetherlands forprovidingindustrialhuminsandWillemVogelzangBSc. (WUR-FBR) for purifying the industrial humins. The authors are also gratefultoAnneAppeldoorn,ErwinWilbersandMarceldeVries (UniversityofGroningen)fortechnicalsupport,HansvandeVelde (UniversityofGroningen)fortheelementalanalysis,LeonRohrbach (UniversityofGroningen)fortheNH3-TPDmeasurementandK.
Altena-SchildkampfortheBETmeasurements(CPM,Universityof Twente).
AppendixA. Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.jaap.2016.12.014.
References
[1]A.Corma,S.Iborra,A.Velty,Chemicalroutesforthetransformationof biomassintochemicals,Chem.Rev.107(2007)2411–2502.
[2]A.J.Ragauskas,C.K.Williams,B.H.Davison,G.Britovsek,J.Cairney,C.A.Eckert, W.J.Frederick,J.P.Hallett,D.J.Leak,C.L.Liotta,J.R.Mielenz,R.Murphy,R. Templer,T.Tschaplinski,Thepathforwardforbiofuelsandbiomaterials, Science311(2006)484–489,6185.
[3]D.Carpenter,T.L.Westover,S.Czernik,W.Jablonski,Biomassfeedstocksfor renewablefuelproduction:areviewoftheimpactsoffeedstockand pretreatmentontheyieldandproductdistributionoffastpyrolysisbio-oils andvapors,GreenChem.16(2014)384–406.
[4]A.J.Ragauskas,G.T.Beckham,M.J.Biddy,R.Chandra,F.Chen,M.F.Davis,B.H. Davison,R.A.Dixon,P.Gilna,M.Keller,P.Langan,A.K.Naskar,J.N.Saddler,T.J. Tschaplinski,G.A.Tuskan,C.E.Wyman,Ligninvalorization:improvinglignin processinginthebiorefinery,Science344(2014).
[5]C.O.Tuck,E.Pérez,I.T.Horváth,R.A.Sheldon,M.Poliakoff,Valorizationof biomass:derivingmorevaluefromwaste,Science337(2012)695–699.
[6]G.W.Huber,S.Iborra,A.Corma,Synthesisoftransportationfuelsfrom biomass:chemistrycatalysts,andengineering,Chem.Rev.106(2006) 4044–4098.
[7]C.B.Rasrendra,M.Windt,Y.Wang,S.Adisasmito,I.G.B.N.Makertihartha, E.R.H.vanEck,D.Meier,H.J.Heeres,Experimentalstudiesonthepyrolysisof huminsfromtheacid-catalyseddehydrationofC6-sugars,J.Anal.Appl.Pyrol. 104(2013)299–307.
[8]R.-J.vanPutten,J.C.vanderWaal,E.deJong,C.B.Rasrendra,H.J.Heeres,J.G. deVries,Hydroxymethylfurfuralaversatileplatformchemicalmadefrom renewableresources,Chem.Rev.113(2013)1499–1597.
[9]J.Horvat,B.Klai ´c,B.Metelko,V. ˇSunji ´c,Mechanismoflevulinicacid formation,TetrahedronLett.26(1985)2111–2114.
[10]T.Buntara,S.Noel,P.H.Phua,I.Melián-Cabrera,J.G.deVries,H.J.Heeres, Caprolactamfromrenewableresources:catalyticconversionof 5-hydroxymethylfurfuralintocaprolactone,Angew.Chem.Int.Ed.50(2011) 7083–7087.
[11]A.J.J.E.Eerhart,W.J.J.Huijgen,R.J.H.Grisel,J.C.vanderWaal,E.deJong,A.de SousaDias,A.P.C.Faaij,M.K.Patel,Fuelsandplasticsfromlignocellulosic
biomassviathefuranpathway;atechnicalanalysis,RSCAdv.4(2014) 3536–3549.
[12]J.-M.Pin,N.Guigo,A.Mija,L.Vincent,N.Sbirrazzuoli,J.C.vanderWaal,E.de Jong,Valorizationofbiorefineryside-streamproducts:combinationof huminswithpolyfurfurylalcoholforcompositeelaboration,ACSSustainable Chem.Eng.2(2014)2182–2190.
[13]D.J.Braden,C.A.Henao,J.Heltzel,C.C.Maravelias,J.A.Dumesic,Productionof liquidhydrocarbonfuelsbycatalyticconversionofbiomass-derivedlevulinic acid,GreenChem.13(2011)1755–1765.
[14]J.C.Serrano-Ruiz,D.Wang,J.A.Dumesic,Catalyticupgradingoflevulinicacid to5-nonanone,GreenChem.12(2010)574–577.
[15]A.A.Rosatella,S.P.Simeonov,R.F.M.Frade,C.A.M.Afonso,
5-Hydroxymethylfurfural(HMF)asabuildingblockplatform:biological properties,synthesisandsyntheticapplications,GreenChem.13(2011) 754–793.
[16]K.M.Rapp,in,GooglePatents(1988).
[17]T.M.C.Hoang,L.Lefferts,K.Seshan,Valorizationofhumin-basedbyproducts frombiomassprocessing—aroutetosustainablehydrogen,ChemSusChem6 (2013)1651–1658.
[18]D.J.Hayes,S.Fitzpatrick,M.H.B.Hayes,J.R.H.Ross,Biorefineries-Industrial ProcessesandProducts,Wiley-VCHVerlagGmbH,2008,pp.139–164.
[19]I.vanZandvoort,Y.Wang,C.B.Rasrendra,E.R.H.vanEck,P.C.A.Bruijnincx,H.J. Heeres,B.M.Weckhuysen,Formation,molecularstructure,andmorphology ofhuminsinbiomassconversion:influenceoffeedstockandprocessing conditions,ChemSusChem6(2013)1745–1758.
[20]A.V.Bridgwater,Renewablefuelsandchemicalsbythermalprocessingof biomass,Chem.Eng.J.91(2003)87–102.
[21]A.V.Bridgwater,Reviewoffastpyrolysisofbiomassandproductupgrading, BiomassBioenergy38(2012)68–94.
[22]S.Chu,A.V.Subrahmanyam,G.W.Huber,Thepyrolysischemistryofa beta-O-4typeoligomericligninmodelcompound,GreenChem.15(2013) 125–136.
[23]W.Mu,H.Ben,A.Ragauskas,Y.Deng,Ligninpyrolysiscomponentsand upgrading—technologyreview,BioEnergyRes.6(2013)1183–1204.
[24]H.Yang,R.Yan,H.Chen,D.H.Lee,C.Zheng,Characteristicsofhemicellulose celluloseandligninpyrolysis,Fuel86(2007)1781–1788.
[25]R.French,S.Czernik,Catalyticpyrolysisofbiomassforbiofuelsproduction, FuelProcess.Technol.91(2010)25–32.
[26]X.Li,L.Su,Y.Wang,Y.Yu,C.Wang,X.Li,Z.Wang,Catalyticfastpyrolysisof KraftligninwithHZSM-5zeoliteforproducingaromatichydrocarbons,Front. Environ.Sci.Eng.6(2012)295–303.
[27]D.Shen,J.Zhao,R.Xiao,S.Gu,Productionofaromaticmonomersfrom catalyticpyrolysisofblack-liquorlignin,J.Anal.Appl.Pyrolysis111(2015) 47–54.
[28]M.Inaba,K.Murata,I.Takahara,Y.Liu,Catalyticfastpyrolysisofeucalyptus usingzeolite,J.Chem.Eng.Jpn.47(2014)345–351.
[29]M.Inaba,K.Murata,I.Takahara,Y.Liu,Productionofaromaticandphenolic compoundsbyfastpyrolysisofeucalyptususingzeolitecatalysts,J.Jpn.Inst. Energy93(2014)944–952.
[30]Y.-T.Cheng,G.W.Huber,Productionoftargetedaromaticsbyusing Diels-AlderclassesofreactionswithfuransandolefinsoverZSM-5,Green Chem.14(2012)3114–3125.
[31]S.Constant,W.Vogelzang,R.K.P.Purushotaman,D.S.vanEs,P.C.A.Bruijnincx, B.M.Weckhuysen,CharacterisationofSolubleIndustrialHumins,tobe sumitted(2017).
[32]D.J.Mihalcik,C.A.Mullen,A.A.Boateng,Screeningacidiczeolitesforcatalytic fastpyrolysisofbiomassanditscomponents,J.Anal.Appl.Pyrolysis92(2011) 224–232.
[33]N.Nikbin,S.Feng,S.Caratzoulas,D.G.Vlachos,p-Xyleneformationby dehydrativearomatizationofaDiels-AlderproductinLewisandBrønsted acidiczeolites,J.Phys.Chem.C118(2014)24415–24424.
[34]D.Li,L.Wang,M.Koike,Y.Nakagawa,K.Tomishige,Steamreformingoftar frompyrolysisofbiomassoverNi/Mg/Alcatalystspreparedfrom hydrotalcite-likeprecursors,Appl.Catal.B102(2011)528–538.
[35]T.R.Carlson,J.Jae,Y.-C.Lin,G.A.Tompsett,G.W.Huber,Catalyticfastpyrolysis ofglucosewithHZSM-5:thecombinedhomogeneousandheterogeneous reactions,J.Catal.270(2010)110–124.
[36]J.Jae,G.A.Tompsett,A.J.Foster,K.D.Hammond,S.M.Auerbach,R.F.Lobo,G.W. Huber,Investigationintotheshapeselectivityofzeolitecatalystsforbiomass conversion,J.Catal.279(2011)257–268.
[37]A.Bonilla,D.Baudouin,J.Pérez-Ramírez,Desilicationofferrieritezeolitefor porositygenerationandimprovedeffectivenessinpolyethylenepyrolysis,J. Catal.265(2009)170–180.
[38]Y.-T.Cheng,J.Jae,J.Shi,W.Fan,G.W.Huber,Productionofrenewablearomatic compoundsbycatalyticfastpyrolysisoflignocellulosicbiomasswith bifunctionalGa/ZSM-5catalysts,Angew.Chem.Int.Ed.51(2012)1387–1390.