Sorption and kinetics of CO2 and CH4 in binderless beads of 13X zeolite

内容需要下载文档才能查看

SorptionandkineticsofCO2andCH4inbinderlessbeadsof13Xzeolite

JoséA.C.Silvaa, ,KristinSchumannc,AlírioE.Rodriguesb

a

EscolaSuperiordeTecnologiaeGestão,InstitutoPolitécnicodeBragança,Apartado134,5301-857Bragança,Portugal

LaboratoryofSeparationandReactionEngineering,DepartamentodeEngenhariaQuímica,FaculdadedeEngenharia,UniversidadedoPorto,RuadoDr.RobertoFriass/n,Portugalc

ChemiewerkBadKöstritzGmbH,Heinrichshall2,07586BadKöstritz,Germany

b

articleinfoabstract

ThesorptionequilibriumofCO2andCH4inbinderlessbeadsof13Xzeolitehasbeeninvestigatedbetween313and373Kandpressureupto4atm.TheamountadsorbedofCO2andCH4isaround5.2mmol/gadsand1.2mmol/gads,respectively,http://wendang.chazidian.comparingthesevalueswiththeonesinliteraturethevalueofCO2is20%higherthaninCECA13Xbinderpellets.ItisalsofoundthatisothermsarepronouncedTypeIforCO2andalmostlinearforCH4.TheCO2isothermsweremodeledusingasimpledeviationfromLangmuirisothermthattakesintoaccountinteractionbetweenadsorbedmoleculesatadjacentsites(Fowlermodel)suggestingamoderaterepulsion.Henry’sconstantsrangefrom143to11.1mmol/gads.atmforCO2and0.45to0.27mmol/gads.atmforCH4between313and373K,respectively.Theheatsofsorptionatzerocoverageare43.1kJ/molforCO2and9.2kJ/molforCH4.

ThesorptionkineticshasbeeninvestigatedbytheZero-LengthColumntechnique(ZLC).RecipestoanalyzeZLCdesorptioncurvesinpelletsofadsorbentsarereviewedanditisderivedacriteriawhichindi-catesthatforthesorptionratebemeasuredmacroscopicallythetimeoftheexperiment(thatshouldbe

r2

aboveafewseconds)isdirectlycalculatedwiththefollowingexpression:t0:1!7:02Â10À2ccBasedonsuchcriteriaitisshownthatcrystaldiffusivityofCO2in13XcanbemeasuredmacroscopicallybyZLC,beingthesamemeasurementforCH4practicallyimpossible.ThecrystaldiffusivityofCO2measuredexperimentallyis5.8Â10À15m2/sand1.3Â10À15m2/sat373and313K,respectively.Thesevaluesarecomparabletotheonesmeasuredbyafrequencyresponseandpulsechromatographytechniquesreportedinliterature.TheZLCdesorptioncurvesforCH4weremeasuredunderanequilibriumregime.

Ó2012ElsevierInc.Allrightsreserved.

Articlehistory:

Received17February2012

Receivedinrevisedform18March2012Accepted22March2012

Availableonline30March2012Keywords:

Binderless13XzeoliteSorptionofCO2/CH4ZLCtechnique

Adsorptionequilibrium

1.Introduction

Thereductionofcarbondioxideandmethaneemissionstoatmosphereisamatterofgreatconcernnowadayssincebothgasescancontributesigni cantlytotheso-calledgreenhouseeffectthatdescribesthetrappingofheatnearearth’ssurfacebygasesintheatmosphere.Indeed,carbondioxideisnecessarybecausetherearecalculationsshowingthatifitwerenotpresentintheatmo-sphereearthwillbe30°Ccooler.ThepresenceofCO2intheatmo-sphereisruledbythecarboncyclebuttodaythatbalancehasprobablybeenupset.AtthesametimeCO2/CH4separationsareofgreateconomicalandtechnologicalimportanceintreatinggasstreamslikeland llgas,biogasandcoal-bedmethane.Accord-ingly,thereisaneedtoinvestigateonthistopicandthatcanbedonewithimprovedef cienttechnologiestoseparateorremoveCO2andCH4fromexhaustgases.

Tworecentreviewsdiscussthismatterwithgreatdetailcon-cerningtheuseofadsorbentbasedtechniquestohandleCO2cap-tureandCO2/CH4separations[1,2].AnewclassofadsorbentsCorrespondingauthor.Tel.:+351273303125.

E-mailaddress:jsilva@ipb.pt(J.A.C.Silva).

1387-1811/$-seefrontmatterÓ2012ElsevierInc.Allrightsreserved.http://wendang.chazidian.com/10.1016/j.micromeso.2012.03.042

namedMetal–OrganicFrameworks(MOFs)arefocusedbeingclearthatinfuturetheycanbeanexcellentalternativetozeoliteadsor-bentsgenerallyusednowadays.However,MOFsneedtobefurtherre nedregardingitsproductioninlargescale,chemicalandther-malstability,whicharepropertiesalreadywell-establishedinzeolites.

TherearetodayadsorptionprocessesbasedinzeoliteslikePres-sureSwingAdsorption–PSAtostoreandseparatecompoundssuchasCO2andCH4.Tobeusedasadsorbentszeolitepowderneedstotransformedintomolecularsievesandthisreducesitsworkingcapacityin20%ormorewhichistheamountofadsorptiveinertclaybindergenerallyusedtogivethenecessarymechanicalstrengthtothepelletsorbeadsinordertobeusedinpacked-col-umnsandatthesametimereducepressuredrop.Toincreasetheworkingcapacitythebindercanalsobeconvertedtozeolitematterleadingtotheso-calledbinderlesspelletsorbeads[3,4]butthistechnologyhasnotreceivedgreatattentionfromcompaniesthatproducemolecularsieves.Recently,thistechnologyhasbeenrecoveredandappliedforthesynthesisofbinderlessbeadsof13Xzeolitewherethenon-zeoliticcomponents(temporarybinder)isconvertedtozeoliteduringahydrothermalconversionafterthemanufacturingprocedure[5].Theresultingbinderlessbeadscan

220J.A.C.Silvaetal./MicroporousandMesoporousMaterials158(2012)219–228

Nomenclaturebcc0DcDpDmDKFHKLpqqm

isothermequilibriumconstant,PaÀ1

outletconcentrationoftheZLC,mol/m3

saturationconcentrationoftheZLC,mol/m3crystaldiffusivity,m2/smacroporediffusivity,m2/smoleculardiffusivity,m2/sKnudsendiffusivity,m2/spurge owratenZLC,m3/sHenry’slaw,mol/g.Pa

adsorptionequilibriumconstant(Henry’slawconstant),dimensionless

ZLCmodelparameter,dimensionlesspressure,PaÀ1

amountadsorbed,mol/kg

amountadsorbedatthesaturationoftheadsorbent,mol/kg

rcRpwRtTVsRp

crystalradius,mpelletradius,m

istheextraenergy(Fowlerisotherm),J/molidealgasconstant,J/mol.Ktime,s

temperature,K

volumeofadsorbent,m3pelletradius,m

Greeksymbolsqppelletdensity,kg/m3eppelletsporosity.DimensionlessCptortuosity,dimensionless

rootsoftranscendentalEq.(4),dimensionlessb1

hcoverageequaltoq/qm;thesameasdegreeof llingof

sites,dimensionless

increaseinthiswaytheworkingcapacitiesofexistingzeoliteadsorbenttechnologies.

Inliteraturewecan ndseveraldataandmodelingregardingthesorptionofCO2andCH4inzeolites[6–10]andMOFs[11–17].Amongthezeolitesoneofthemostinterestingiszeolite13Xduetoitslargecagesthatcanaccommodatealargeamountofmassandatthesametimethepresenceofcationsthatproduceelectric eldthatinteractswithstrongquadropolemomentmole-culessuchasCO2.Thisgivesrisetoanincreasedselectivitybe-tweenCH4(apolar)andCO2thathasbeenexploitedincyclicprocesses[18–21].

Forthemodelingofadsorptionprocessesitisoffundamentalimportancetoanalyzeconvenientlythermodynamicdata.ThebooksofBarrer[22],Ruthven[23],Guiochonetal.[24]andDo[25]highlightingreatdetailthebasicstoanalyzesuchdata.ForTypeIisothermswhicharethemostfrequentinzeolites,localizedadsorptionmodelssuchas:Langmuir,dual-site-Langmuir,Fowler,Nitta,etc.areusedextensivelyduetotheirsimplicityandatthesametimebeingthermodynamicconsistentgivinginsightintosorptioneventsinacomprehensiveway.

Themeasurementofsorptionkineticisalsofundamentalformodelingadsorptioncyclicprocessessincethetransportofmassintoandoutoftheadsorbentcanaffectsigni cantlytheperfor-manceofindustrialprocesses.Forthemeasurementofsorptionkineticsthereareseveraltechniques,oneismicroscopic(PFG,NMR)andtheotherismacroscopic(uptakerate,chromatographic)[26–28].SinceitsintroductionbyEicandRuthven[29]forthemeasurementofintracrystallinediffusivitiesinstronglyadsorbedspeciestheZero-LengthColumn(ZLC)techniquehasbeenusedextensivelyforthemeasurementofsorptionratesinporousmediaduetoitsapparentsimplicity[30–33].However,specialattentionintheuseofmodelparametersfromwhichkineticdataareob-tainedisrequiredsincetherearemodelswiththesamemathemat-icalformincompletelydifferentregimesthatwhenusedwithoutpreviouscalculationscanproduceerroneousresults[34].Exten-sionsforusingthetechniqueforliquidsystems[35,36],pelletsofadsorbents[37],analysesofin uenceofheateffects[38,39],effectofnon-linearequilibrium[40],effectofsurfacebarriers[41,42]andalsoforthemeasuringofadsorptionequilibriahavebeendevel-oped[43],beingnowpossibletousethetechniqueinabroadrangeofsystemssorbate-sorbent.

Thegoalofthisworkistoaccessdataofequilibriumandkinet-icsofsorptionofCO2andCH4onanewtypeofbinderlessbeadsof13Xzeolite.TheequilibriumdataaremeasuredinabreakthroughapparatusandthekineticdatabytheZLCtechnique.

Attentionismaderegardingthecomparisonoftheseresultswithpublisheddataonpelletsofthesamezeolitetypewithbin-der.Atthesametimethermodynamicandkineticparametersareobtainedthatareusefulforthedevelopmentofadsorptionsepara-tionprocessessuchastheonescalculatedfrommodelingofequi-libriumandkineticofsorption:heatsofsorption,Henry’sconstants,equilibriumconstants,workingcapacities,interandintracrystallinediffusivities.ThroughthisworksomeideasabouttheuseZLCtechniqueforthemeasurementthediffusivityinpor-ousadsorbentsarerevisedbyestablishingaproceduretoanalyseproperlysuchresultsintroducingasimplecriteriatoevaluatewhichkindofsystemscanbemeasuredmacroscopicallybyZLC.2.Experimentalsection2.1.Binderless13Xzeolite

Thepowderof13XfromwhichthebinderlessbeadswereformedisfromChemiewerkBasKostritzGmbH(Germany)withaSi/Alratioof1.18.Metakaolinisusedtomanufacturethebeads.Thesynthesisandcharacterizationprocedureisdescribedindetailelsewhere[5].Brie y,thebeadsformedconsistinsphericalparti-cleswithadiameterrangingfrom1.2to2.0mm.Thesizeofthezeolitecrystalsarearound2lm.Table1summarizesthecharac-teristicsofthebeads.

2.2.AdsorptionequilibriumandZLCapparatus

TheequilibriumandkineticsstudieswereperformedintheapparatusillustratedinFig.1.Brie y,itconsistsintwosections:i)agaspreparationsystem;andii)aGasChromatographwitha

Table1

Physicalpropertiesofzeolite13Xbeadsandadsorptioncolumncharacteristics.

Physicalpropertiesofbinderless13XbeadsaSiO2/Al2O3ratio

Crystaldimensions(lm)

Beadsdimension(spherical)(mm)Averageporediameter(lm)AdsorptioncolumncharacteristicsLength(cm)

Internaldiameter(mm)

a

2.35%2

1.2–2.00.684.6

FromRef.[5].

J.A.C.Silvaetal./MicroporousandMesoporousMaterials158(2012)219–228221

TCDdetectorwheretheadsorptioncolumnisplaced.Inthegaspreparationsectionitisusedheliumastheinertgaswhichispre-viouslydehydratedinamolecularsievetype5A.Heliumentersinthesystembytwodifferentstreams:onelinetobemixedwithsor-batespecies(CO2andCH4)andtheotherwhereitispure(line1).TheHefromline(1)andthemixture(sorbate+inert)fromline(2)runtoa6-waycrossovervalve(SV)thatallowstheselectionofwhichline1or2passesbytheadsorptioncolumn.Thelinethatdoesnotenterintheadsorptioncolumnisby-passedthroughline3.Thepressureofallthesystemiscontrolledbyaback-pressure-regulator(BPR).Theef uentoftheadsorptioncolumnpassesdi-rectlybythereferencesideofaTCDthatcandetectconcentrationofallgasestoaround100ppm(ifneededapartofthisef uentcouldby-passtheTCD).

Theadsorptioncolumnconsistsofa4.6mmi.d.stainlesssteelcolumnwith80mmlength.Fortheadsorptionequilibriumstudiesthecolumnisentirely lledwithzeolitebeads.InZLCexperimentsjustthebottomofthecolumnis lledwithfewadsorbentparticles,beingtheremainingspaceoccupiedwithsmallglassspheres.Valveposition,oventemperature,mass ows,back-pressureregu-latorandTCDsignalarecompletelyautomated.

Thesorbateandinertgaseswerefurnishedbyairliquidwiththefollowingpurities:methaneN35(99.95%),carbondioxideN48(99.998%),andheliumALPHAGAZ2(99.9998%)2.3.Procedureforsorptionequilibriumexperiments

Beforetimezeroofanexperimenttheswitchvalve(SV)directspureheliumthat owsthroughline1totheadsorptioncolumn,andventthroughline3themixturethatcomesfromline2pre-paredbyacombinationofmass owsofsorbateplushelium.Attimezeroswitchvalve(SV)isswitchedallowingline2toenterinthecolumnventingatthesametimeline1.Theexperimentcon-sistsinmeasuringcontinuouslytheconcentrationasafunctionoftimeattheoutletofthebedbytheTCD.Theequilibriumloadingoftheexperimentisobtainedbyintegratingtheconcentrationpro- lesofthebreakthroughcurvewithaproceduredescribedelse-where[44].Beforethe rstruntheadsorptioncolumnisactivatedforatleast24hat493Kunderpurehelium ow.Whentheadsorptionexperiment nishesSVisswitchedagain.OncetheTCDsignalreachesitslowerlevelanotherrunisperformed.Onerunmeansoneequilibriumpointoftheisotherm.

2.4.ProcedureforZLCexperiments

FortheZLCexperimentswefollowstrictlytheguidelinespro-videdbyEicandRuthvenintheoriginalZLCpaper[29].Especially:1)Set-upalowconcentrationofsorbatetovalidateHenry’slawofisotherm;2)Blankrunstoaccountforextraneouscapacities(no-problemswerefoundsinceinoursystemthecolumnisdirectlyat-tachedtothedetector);3)Performexperimentswithdifferentpurge owratestoseetheproportionalitybetweenparameterLandthe owrate.4)set-upadifferentialbed(50mgofpelletswereusedintheexperiments).

TheZLCexperimentissimilartotheoneperformedfortheadsorptionequilibriummeasurements.BeforetimezeroSVdirectsthe owofline2(whichcontainsaverylowconcentrationofthesorbatespeciesaround0.01atm)tothecolumn.AttimezeroSVisswitchedandline1(purehelium)isallowedtopassthroughtheadsorptioncolumninitiatingthedesorptioncurve.Thesignalpro-ducedbytheTCDiscontinuouslymonitoredbyacomputerforfu-turedatatreatmentinordertoset-upofthedesorptioncurveintermsofconcentrationversustime.

3.Theoretical

3.1.Purecomponentsisotherms

AsimpleandsuitablemodeltorepresenttypeIisothermsistheLangmuirequation:

1h

¼b

pð1ÀhÞ

ð1Þ

whereh=q/qmisthedegreeof llingofsites,bisanequilibriumconstant,pthepressure,qtheamountadsorbedandqmistheamountadsorbedatthesaturationoftheadsorbent.TheLangmuirequationisperfecttorepresentsorptioninahomogeneoussitessurfacewhereasorbatemoleculeoccupiesoneactivesitewhenitadsorbswithnointeractionbetweenadsorbedmolecules.

ToaccountforlateralinteractionsbetweenadsorbedmoleculesFowler[45]proposedthefollowingequation

1h

¼bexpðÀ2wh=RTÞ

pð1ÀhÞ

ð2Þ

222J.A.C.Silvaetal./MicroporousandMesoporousMaterials158(2012)219–228

wherewistheextraenergywhensorbatemoleculesoccupyadja-centsites(positiveforrepulsion,negativeforattraction),RtheidealgasconstantandTthetemperature.

Thevalidityofbothmodelscanbeeasilyveri edbyplotting

log1h

againsth.Ifwe ndahorizontallinetheLangmuirmodelisvalid.InthecaseofFowlerisothermequationwewill ndastraightlinewithslopeÀ2w/RT.Tobettertestbothmodelsitisconvenienttoknowaprioritheparameterqmaxbutitcanbere-laxedduringveri cationifnodataareknown.3.2.ZLCmodels

TheZero-LengthColumn(ZLC)technique[29]formeasuringdiffusivitesinadsorbentsconsistsinadifferentialbedofporousparticlesthatis rstsaturatedwiththesorbatespeciespreferablyataverylowconcentrationunderthevalidityoftheHenry’slawoftheisotherm.Attimezerothecarriergas(freeofsorbate) owsthroughtheZLCandthedesorptioncurveismeasuredasafunc-tionoftime.

AsexplainedbyEicandRuthvenmodelparameterscanbeeas-ilyobtainedusingtheinformationofdesorptioncurvesatlong

times.ThemodeloftheZLCbasedontheFick

´slawofdiffusionre-ducesatlongtimestothefollowinglinearequationinasemi-logplot,

ln c

¼ln2L!0bÀb2Dc

12t1þLðLÀ1Þ

ð3Þ

cwhere,

b1cotðb1ÞþLÀ1¼0ð4Þ¼

1Fr2Lc

3KVð5Þ

sDc

InthepreviousequationscistheoutletconcentrationoftheZLC,c0isthesaturationconcentration,tisthetime,b1aretherootsoftranscendentalEq.(4),Dcisthecrystaldiffusivity,rcthecrystalradius,Lamodelparameter,Fthepurge owrate,VsthevolumeofadsorbentandKadimensionlessadsorptionequilibriumconstant(Henry’slawconstant).Thepreviousmodelisvalidforthemea-surementofintracrystallinediffusivity.Ifexperimentsaremadeinpelletsandifthecontrollingmechanismofdiffusionistheoneinthemacroporesofthepelletsthepreviousequationsareslightlymodi edbeing,

ln c

2L!c¼lnðLÀ1ÞÀb2Dp

0b21R2

tð6Þ

1þLpð1þKÞL¼

1FR2p

3Vð7Þ

sDp

whereDpisthemacroporediffusivityandRpisthepelletradius.ConsideringthatdiffusioninmacroporesisKnudsendiffusivityinserieswithmoleculardiffusion,theporediffusivitycanbeesti-matedby,

Dp¼

C

1p11ð8Þ

m

þ

K

whereCpisthetortuosity,DmisthemoleculardiffusivityandDKistheKnudsendiffusivity.

IfexperimentsinZLCareperformedunderanequilibriumre-gimeL<0.1thepreviousequationscanbesubstitutedwiththefol-lowingequation[43],

ln c c¼ÀF

kVt:ð9Þ

0s

RecipesforusingZLCtechniquecarefullyaregivencomprehen-sivelyandwithgreatdetailintheoriginalpaperbyEicandRuth-ven[29].

4.Resultsanddiscussion4.1.Sorptionisotherms

Theexperimentalsinglecomponentsorptionisothermsmea-suredforCO2andCH4in13XareshowninFig.2aandb,respec-tively.Datawerecollectedatthreetemperatures:313,343and373Kandforpartialpressures(heliumasinert)upto4atm.WecanobserveinFig.2thatCH4isothermsarepracticallylinearandmarkedtypeIforCO2intherangeoftemperatureandpressurestudied.Asexpected[6,9,10],Fig.1showsthatCO2isthemorestrongadsorbedcomponentwithanamountadsorbedthatreachesmorethan5mmol/gadsat313Kandpartialpressurearound4atm.ForthesamepressureandtemperaturetheamountadsorbedofCH4ismuchsmallerbeingalmost1.2mmol/gads.ThismeansalsoaselectivityCO2/CH4around4.2.Sinceweusebinderlessbeadsitisexpectedthatthesorptioncapacityincreases,beingofinteresttocompareourdatawith13Xzeolitewithbinder.LookingatdatameasuredbyMulgundmathetal.[9]inCECA13Xzeoliteofmeshsize20–60ourdatacomparesto4.0mmol/gadsforCO2and1.27mmol/gadsforCH4.RegardingthedataonCECA13Xextru-datesof1.6mm[10]thedataofthepresentworkcomparesto

J.A.C.Silvaetal./MicroporousandMesoporousMaterials158(2012)219–228223

4.05mmol/gforCO2at308Kand1.32mmol/gforCH4at308K.Thismeansthatthebinderless13Xstudiedinthisworkincreasesthesorptioncapacityin20%forCO2beingthevaluesforCH4prac-ticallythesame.

TocalculatedirectlyHenry’sconstantsfromexperimentaldataweplotp/qversuspaccordingtoaVirialPlot.ExtrapolationofdatatozerocoveragegivesusthereciprocalofHenry’sconstants.Fig.3showssuchplotsinsemi-logcoordinatesforbothCO2andCH4andTable2thecalculatedvalues.ForCO2theHenry’sconstantsrange

Table2

IsothermmodelparametersforsorptionofCO2andCH4inbinderlessbeadsof13Xzeolite.

IsothermmodelsCO2

CH4

Fowler

Langmuirqm(mmol/gads)7.4a

7.4a

ÀDH(kJ/mol)43.19.2w(kJ/mol)6.1–

313KHb(mmol/gads.atm)1430.45b

(atmÀ1)

21.30.0643343KHb(mmol/gads.atm)38.50.34b

(atmÀ1)

4.760.0462373KHb(mmol/gads.atm)11.10.27b

(atmÀ1)

1.49

0.0374

aFromRef.[2,10].

b

Henry’sconstantcalculatedfromVirialplots.

from143to11.1mmol/gads.atmbetween313and373K,respec-tively.TheseveryhighvaluesareclearlyduetothestrongbondsbetweenCO2andthecationsofzeoliteinspiteofthelargequadro-polemomentofCO2.ForCH4thevaluesaremuchsmallerandrangebetween0.45to0.27mmol/gads.atminthesametempera-tureintervallevel.Asaresult,wefoundthattheselectivityCO2/CH4atlowpartialpressuremeasuredbytheratiooftheHenry’sconstantsat313Kis143/0.45equalto318.Thisisaveryhighvalue.

TomodelsorptiondatawedecidedtouselocalizedadsorptionmodelswheretheLangmuirmodelisthesimplerone.Fig.4shows