An innovative technique for energy-efficient CO2

Solid acid catalysts for low-temperature regeneration of non-aqueous sorbents: An innovative technique

for energy-efficient CO2 capture processes

Item Type Article

Authors Barzagli, Francesco;Bhatti, Umair Hassan;Kazmi, Wajahat W.;Peruzzini, Maurizio

Citation Barzagli, F., Bhatti, U. H., Kazmi, W. W., & Peruzzini, M. (2023).

Solid acid catalysts for low-temperature regeneration of non- aqueous sorbents: An innovative technique for energy-efficient CO2 capture processes. Carbon Capture Science & Technology, 8, 100124. https://doi.org/10.1016/j.ccst.2023.100124

Eprint version Publisher's Version/PDF

DOI 10.1016/j.ccst.2023.100124

Publisher Elsevier BV

Journal Carbon Capture Science and Technology

Rights Archived with thanks to Carbon Capture Science and Technology under a Creative Commons license, details at: http://

creativecommons.org/licenses/by/4.0/

Download date 2023-12-04 19:24:28

Item License http://creativecommons.org/licenses/by/4.0/

Link to Item http://hdl.handle.net/10754/692904

ContentslistsavailableatScienceDirect

Carbon Capture Science & Technology

journalhomepage:www.elsevier.com/locate/ccst

Full Length Article

Solid acid catalysts for low-temperature regeneration of non-aqueous sorbents: An innovative technique for energy-efficient CO ₂ capture processes

Francesco Barzagli

^a,∗

, Umair H. Bhatti

^b,∗

, Wajahat W. Kazmi

^c

, Maurizio Peruzzini

^a

aNational Research Council, ICCOM Institute, via Madonna del Piano 10, 50019 Sesto F.no, Florence, Italy

bKAUST Catalysis Center (KCC), King Abdullah University of Science and Technology (KAUST), Thuwal 23955, Saudi Arabia

cKorea Institute of Energy Research, University of Science and Technology, 217 Gajeong-ro Yuseong-gu, Daejeon 34129, Korea (the Republic of)

a r t i c le i n f o

Keywords:

CO ₂capture Non-aqueous sorbent Catalytic regeneration

13C NMR speciation 2-(2-aminoethoxy)ethanol

a b s t r a ct

ChemicalabsorptionofCO₂fromfluegasesbyusingaqueousaminesorbentsisregardedasthemostmatureand effectivetechnologyforreducingCO₂emissions,butthehighenergycostofsorbentregenerationhassofargreatly limiteditsindustrialapplication.Oneofthebeststrategiesdevelopedbyresearchersinrecentyearstoreduce energyrequirementsistoimproveCO₂ desorptionkineticsbyaddingcatalystsduringsorbentregeneration.

Moreover,non-aqueoussorbentshaverecentlybeengainingincreasingattentionasalternativestoconventional aqueousaminesastheyhavethepotentialtolowertheenergydemandforsorbentregeneration.Inthispaper, weproposeaninnovativetechniquethatcombinesandfurtherdevelopsthesetwoemergingtechnologies,non- aqueousabsorbentandcatalyst-assistedregeneration,toenablelessenergy-intensiveCO₂ captureprocesses.

Inparticular,inthisscreeningstudy,weevaluatetheregenerationbehaviourofaCO₂-saturatedsolutionof 2-(2-aminoethoxy)ethanol(DGA)indiethyleneglycolmonomethylether(DEGMME)whenheatedto85°Cin theabsenceandpresenceofdifferenttypesofsolidacidcatalysts.Inordertoassessthepotentialbenefitsof thisinnovativetechniqueoverconventionalsystems,theresultsobtainedwerecomparedwiththedesorption performanceofanaqueoussolutionofDGAunderthesameoperatingconditionsandwiththesamecatalysts, whichhighlightedthepossibilityofobtainingrapiddesorptionatrelativelylowtemperatureswhenconducted withsuitableacidcatalystsusingaminesinorganicdiluents.

Introduction

TheincreasingconcentrationofCO₂ intheatmosphere isconsid- eredamajorcauseoftheevidentglobalwarmingandrelatedclimate change.Thedevelopmentanddeploymentoftechniquestolimitanthro- pogenicCO₂ emissionsiscrucialtokeeptheglobaltemperaturerise within1.5°Cofthepre-industriallevel,asenvisagedinthe2015Paris AgreementandsubsequentConferencesoftheParties(COP)(UNFCCC- UnitedNationsFrameworkConventiononClimateChange,2021,2015).

The post-combustion CO₂ absorption by chemical capturebased on aqueous amines is todayregarded asthemost mature andeffective methodforindustrialuse.Inparticular,MEA30%(aqueoussolutions ofethanolamine,30%byweight)havebeenthoroughlytestedonpilot andoperatingplants(Buietal.,2018;ElHadrietal.,2017;Feronetal., 2020;Zhangetal.,2017).However,thewidespreadapplicationofthis

∗Correspondingauthors.

E-mailaddresses:francesco.barzagli@iccom.cnr.it(F.Barzagli),umair.bhatti@kaust.edu.sa(U.H.Bhatti).

technologyisstillchallenging,mainlybecauseoftheenormousenergy requirement forthesorbent regenerationprocess(CO₂ desorptionat temperatureshigherthan120°C)(Heetal.,2023;Liangetal.,2015).

Inrecentyears,researchershavefocusedtheireffortsondeveloping severalinnovativestrategiestoreducetheenergydemandfordesorp- tionwhilemaintainingaCO₂ absorptionperformancecomparableto conventionalaqueousMEA(Atzorietal.,2023;GuangjieChenetal., 2022;GuangyingChenetal.,2022;Shietal.,2021;Wangetal.,2019).

amongstthesestrategies,theuseofsolidacidcatalyststoimproveCO₂ desorptionandtheuseofnon-aqueoussorbentsareparticularlypromis- ing.

Theimplementationofsolidacidcatalystsinthedesorbertoenhance sorbentregenerationisastrategydevelopedveryrecentlyforaqueous aminesolutions.Intheseprocesses,suitableacidcatalystspromotethe CO₂releasefromloadedsorbentsatlowertemperatures,thusdecreasing thetotalenergyconsumption(Alivandetal.,2020).Inparticular,the

https://doi.org/10.1016/j.ccst.2023.100124

Received4May2023;Receivedinrevisedform12June2023;Accepted12June2023

2772-6568/© 2023TheAuthor(s).PublishedbyElsevierLtdonbehalfofInstitutionofChemicalEngineers(IChemE).ThisisanopenaccessarticleundertheCC BYlicense(http://creativecommons.org/licenses/by/4.0/)

F. Barzagli, U.H. Bhatti, W.W. Kazmi et al. Carbon Capture Science & Technology 8 (2023) 100124

additionofcatalystenablessorbentregenerationattemperaturesof85–

100°C,wellbelowconventionaltemperatureusedaround120–140°C, resultinginsignificativeenergysavings.Generally,metaloxides,zeo- lites,modifiedclays,andsilicamaterialswerethecatalyststhatshowed the most beneficial effect in the solventregeneration (Bhatti et al., 2021b,2020;Lietal.,2023;Zhangetal.,2023b).

Theotherinnovativestrategyconsidered hereinvolvestheuseof water-freesorbents. Traditionally,themost studiedsorbents forCO₂ capturewereaqueousamine,mainlybecausewateristhemostcommon andavailablediluent.However,highsensibleandlatentheatsofwater maketheregenerationprocessofaqueous sorbentsextremelyenergy intensive.Owingtothat,recently,anincreasingnumberofresearchers areconsideringusingofamine-basedsorbentswithdiluentsotherthan water(Wanderleyetal.,2021).Indeed,replacingwaterwithorganicsol- ventsredirectsthereactionbetweenamineandCO₂towardslessstable carbonatedspeciesthatcansubsequentlybedecomposedmoreeasily (atT<100°C).Moreover,thelowerheatcapacityandheatofvaporiza- tionoforganicdiluentswithrespecttowatercouldsignificantlyreduce theenergydemandoftheentireCO₂captureprocess(Alkhatibetal., 2022;Barzaglietal.,2022;Heldebrantetal.,2017;Karlssonetal.,2020; Svenssonetal.,2014).

WiththeaimofdevelopingatechniqueforCO₂capturewithhigh efficiencyandlowenergyrequirement,wedecidedtocombinethetwo strategiesabove,resultinginanewtechniquebasedonabsorptionby non-aqueousamine-based solutionsanddesorptioncatalysed byacid solids.Toverifythesuitabilityofthisinnovativeapproach,wecarried outadetailedinvestigationofthedesorptionperformanceatlowtem- peratureofthesameamineinbothaqueousandnon-aqueoussolutions, withandwithout catalysts.The amineused forthis studywas2-(2- aminoethoxy)ethanol(DGA),a highlyabsorptive primaryaminethat hasalongstoryforacidgasremoval,withfewercorrosionproblems thanMEA,resultinginpotentialcapitalandoperatingcostsavings,as wellasimprovedoperationatrelativelylowpressure(Al-Juaiedand Rochelle,2006;SalkuyehandMofarahi,2012).Moreover,inourrecent studiesDGAhasdemonstratedexcellentCO₂absorptionefficienciesin bothwaterandorganicdiluent (Barzaglietal.,2018).TwoDGAso- lutionswerepreparedwithaconcentrationof3mol‧L⁻¹:onein wa- ter,theotherinDEGMME(diethyleneglycolmonomethylether)asor- ganicdiluent.DEGMMEexhibitsinterestingchemical-physicalcharac- teristics,as itisable tosolubiliseboth amineandits reaction prod- uctswithCO₂,hasahighboilingpoint(194°C),viscositycompatible withuse inconventionalequipment(3.5cPat25°C)and,aboveall, amuchlowerheatcapacitythanwater(2260J‧kg⁻¹‧K⁻¹,withrespect to4184J‧kg⁻¹‧K⁻¹ofH₂O),whichensureslowerenergydemanddur- ingheating(Bhattietal.,2021a).Thetwoaqueousandnon-aqueous solutionsof DGAwerefirstsaturatedwithpure CO₂,andthentheir desorptionperformance withandwithout catalysts wasstudiedat a temperatureof 85°C(significantlylowerthanconventional tempera- tures),evaluatingtheCO₂released,therateofCO₂desorptionandthe percentageofsorbentregeneratedasafunctionofthedesorptiontime.

Fivedifferentsolidacidcatalystswereusedfortheseexperimentalstud- ies,andinparticulartwometaloxides(V₂O₅andWO₃)andthreesil- icatederivatives,namelyHClactivatedmontmorillonite(HCl-MONT), 20%Mo/SBA-15,and20%Mo/SAPO-11.Theseacidsolidsandtheirpre- cursorshavedemonstratedinpreviousstudiestheircatalyticactivityin theregenerativeprocessesofaqueoussolutionsofamines,andinpar- ticularwithMEA(BarzagliandMani,2021;Bhattietal.,2018,2017; Lietal.,2023).Moreover,V₂O₅andWO₃hadalreadybeensuccessfully testedtoregeneratenon-aqueoussolutionsofDGAat90°C(Bhattietal., 2021a).Finally,adetailedspeciationstudy(qualitativeandquantitative assessmentofspeciesinsolution)wascarriedoutduringthedesorption processusing¹³CNMRanalysisfortheDGA/H₂OandDGA/DEGMME solutionsbothinthepresenceandwithoutthe20%Mo/SAPO-11cat- alyst,in ordertobetter understandthedifferent mechanismsofsor- bentregenerationunderspecificoperatingconditions(Huetal.,2020; Karlssonetal.,2021;Perinuetal.,2018).

Experimentalsection Materials

The primary amine2-(2-aminoethoxy)ethanol (DGA)and theor- ganicsolventdiethyleneglycolmonomethylether(DEGMME)weresup- pliedbySigma-Aldrichandusedwithoutfurtherpurification.Thenon- aqueoussorbentusedwaspreparedbydissolvingDGAintoDEGMMEin ordertoobtainanaminesolutionwithmolarconcentrationof3mol‧L⁻¹. Moreover, for comparisonpurposes, a3 mol‧L⁻¹ DGAaqueous solu- tionwasalsopreparedbyusingdeionizedwater.Previousstudieshave shownthatthis amineconcentrationrepresentsanexcellentcompro- misebetweenhighcaptureefficiencyandacceptableviscosity(related toeaseofhandling)oftheCO₂-loadedsolution(Bhattietal.,2021a).

PureCO₂(RivoiraSpa)wasusedtosaturateaminesolutions,whileN₂ (RivoiraSpa)wasusedasgas-carrierduringthedesorptionprocess.Two flowmeters(ColeParmer)wereusedtocontroltheflowratesofthetwo gases.AVarianCP-4900gaschromatograph(GC)wasusedtocontinu- ouslymeasuretheamountofCO₂releasedduringdesorption.TheGC wasequippedwithaPoraPLOTUcolumn(Agilent)andaTCDdetector, andwascalibratedwith15%and40%v/vCO₂/N₂referencemixture and100%CO₂referencegas(RivoiraSpa).

Regardingcatalysts,tungstentrioxide(WO₃)anddivanadiumpen- toxide(V₂O₅)weresuppliedbyFisherScientificandwereusedasre- ceived.ThevirgincatalystsupportusedinthisstudywasSAPO-11,a molecularsieveobtainedfromACSMaterialsLLC.Tetraethylorthosili- cate(TEOS,99.0%),pluronicP123(M_avg5800),HCl(37%),montmo- rillonite,andammonium-molybdatetetra-hydrateweresourcedfrom Sigma-Aldrich.

Catalystspreparation

TheproductionofmesoporoussilicaSBA-15wasperformedusinga methodoutlinedinpreviousliterature(Moogietal.,2020).Foratyp- icalbatch,calculatedamountsofpluronicP-123(EO₂₀PO₇₀EO₂₀)dis- solvedin waterfollowedbytheaddition of2MHClsolution.TEOS wasthenaddeddropwisewhileundercontinuousstirring.After24hof stirringat40°C,themixturewastransferredintoatetrafluropolyethy- lene(TFPE)bottleandkeptinanovenforhydrothermaltreatmentat 100°C for48h.TheresultingSBA-15micelleswerecooleddownto ambienttemperature,filtered,washed,anddriedfor12hat105°C.

SBA-15micelleswerethencalcinedinairflow(0.1L‧min⁻¹)at550°C for6h.Theincipientwetnessimpregnationmethodwasusedtoprepare molybdenum-supported catalysts,wherecalculatedamountsofmetal precursors (20%Mo) weredissolvedindistilledwater andadded to SAPO-11 andSBA-15supportswhile beingstirred continuously.The catalystsweredriedat105°Candthencalcinedat550°Cinanitrogen flowfor6h.HCl-MONTwasalsopreparedusingamethoddescribed inourpreviouswork.Briefly,5gofMONTclaywasslowlyaddedwith 100mLof aqueousHClsolution(4mol‧L⁻¹)inaround-bottomglass flask,thenstirredvigorously(600rpm)at100°Cfor4htoallowthe acidactivation.Thesamplewasthencooledtoabout20°Candwashed severaltimeswithdeionizedwater,untilthefiltratepHwas∼7.The collectedHCl-MONTwasfinallydriedat120°Cfor10h.

CO₂saturationanddesorptionexperiments

DGA/H₂OandDGA/DEGMMEsolutionsweresaturatedwithpure CO₂(flowrate=0.230molCO₂‧h⁻¹)usingaDrechselwashingbottle withasinteredglassdiffuser (16–40𝜇mpores),maintainedat40°C withathermostattedbath(ThermoScientificAC200-A25).Toobtain evidenceofsaturation,theweightoftheabsorberwasconstantlymon- itoredduringabsorptionuntilaconstantweightwasreached.TheCO₂ loadingvalue(molCO₂captured/molamine)wascalculatedfromthe maximumincreaseinweightoftheCO₂-saturatedsolution.

2

Fig.1.SimplifiedsketchoftheinstrumentationusedforCO₂thermalrelease experiments.

CO₂desorptionexperimentswereperformedaccordingtoaprevi- ouslyvalidatedprocedure(Bhattietal.,2021a).Theinstrumentation usedconsistedofa50mLconicalflaskconnectedtoacondenserand anoilbathequippedwithamagnetic stirrer(IKA heatingbathHBR 4).Asimplifiedrepresentationof theexperimental setupis depicted inFig.1.Briefly,25mLoftheCO₂-saturatedsolution(maintainedat 20°Cbeforebeingused)wasaddedtotheflaskandheatedto85°C.

ThereleaseofCO₂wasfacilitatedbymagneticstirring(800rpm)and bythecontinuousblowingofastreamofN₂(flowrate=410mL‧min⁻¹), whichactedasagascarrier.TheCO₂+N₂gasmixtureleavingthesolu- tionpassedsequentiallythroughacoldcondensertolimitevaporation lossesof amineanddiluent,aconcentratedsolutionofH₂SO₄ tore- moveanytracesofamine,andadesiccantcolumnfilledwithP₂O₅to retainanytracesofdiluent.Finally,thegasmixturewascontinuously analysedbytheGC,whichprovidedtheamountofCO₂desorbed,and consequentlytheCO₂desorptionrate,throughoutthedurationofthe experiment(60min).Exceptforthedesorptionprocessesconductedon thetwosolutionswithoutcatalysts(blank),thecatalystunderinvesti- gationwasaddeddirectlyintotheCO₂-loadedsolutionbeforeheating (1.25g,about5%byweight).Eachexperimentwasrepeatedtwiceand thedeviationfromtheaveragevalueofthereleasedCO₂wasalwaysin therangeof0.5–1.5%.

13CNMRspeciationstudy

Thespeciationofbothaqueousandnon-aqueoussorbentswasdeter- minedby¹³CNMRanalysisofsamplescollectedfromtheCO₂-saturated solutionandduringdesorption,after5,10,15,25,40,and60minof heatingat85 °C.The¹³CNMRspectrawererecordedwithaBruker AvanceIII400spectrometer(100.613 MHz) accordingtoa protocol validatedinourlaboratories(Barzaglietal.,2013;Bhattietal.,2021a).

Si(CH₃)₄wasusedasexternalstandard,CH₃CNasinternalreference, andD₂O(inasealedglasscapillary)wasintroducedintotheNMRtube toprovide the locksignal. The acquisitionparameters used are:ac- quisitiontime=1.3632s,delaytime=2–30s,pulseangle=90.0°,

scans=280–320,datapoints=65K.Byintegratingthesignalsatabout 40 ppm(relativetotheC–CH₂–Ncarbon intheDGAstructure), the relativeamountsofDGAcarbamateandDGA/DGAH⁺(freeandproto- natedamineinrapidequilibrium)canbedeterminedwithanestimated error<2%.Similarly,meticulousintegrationofpeaksintherange158–

165ppmprovidesagoodestimate(error<5%)oftherelativequantity ofthepossiblecarbonylspecies,namelyDGAcarbamate,DEGMMEcar- bonateandthefast-equilibratingHCO₃^–/CO₃^2–ions.AlthoughDGAand DGAH⁺provideauniquepeak(becauseoftherapidprotonscrambling), theirrelativequantitiescanbeassessedfromthechemicalshift(𝛿,ppm) oftheirownpeakafterconstructingasuitablecalibrationline.Thelimit values foundforfreeDGAandfullyprotonatedDGAH⁺ inDEGMME were41.48ppmand39.51ppm,respectively,whileinaqueoussolu- tiontheywere40.24ppmand39.27ppm.Similarly,thepositionofthe uniquesignalofHCO₃^–/CO₃^2–canbeexploitedtoevaluatetheirrela- tiveamountinaqueoussolution(Zhangetal.,2023a):thelimitvalues foundforcarbonateandbicarbonateions(obviouslyonlyinwater)were 168.10ppmand160.25ppm,respectively.

Resultsanddiscussion

ChemicalequilibriainCO₂captureprocesses

Beforestudyingtheirdesorptionperformancewithandwithoutcat- alysts,theDGA/H₂OandDGA/DEGMMEsolutions(amineconcentra- tion=3mol‧L⁻¹) weresaturatedwithpureCO₂atatemperatureof 40 °C,asdescribed inSection2.3,andtheirCO₂ loadings (molCO₂ captured/molamine)weredeterminedfromtheincreaseinweightof the CO₂-saturated solution at theend of the experiment. The load- ingvalueobtainedinaqueoussolutionwas0.647,whilethatinDEG- MMEwas0.576.Moreover,solutionsampleswereanalysedby¹³CNMR spectroscopytoqualitativelyandquantitativelydeterminethespecies present. The spectra obtained are shown in Fig. 2, andthe species presentarethoseexpectedfromreactions1–6,whichdescribetheCO₂ capturebyDGAindifferentdiluents(ElHadrietal.,2017;Zhangetal., 2016).

2R-NH₂+CO₂⇄R-NHCO₂⁻+R-NH₃⁺ (1)

R-NHCO₂⁻+CO₂+2H₂O⇄2HCO₃⁻+R-NH₃⁺ (2)

R-NH₂+CO₂+H₂O⇄HCO₃⁻+R-NH₃⁺ (3)

HCO₃⁻+R-NH₂⇄CO₃²⁻+R-NH₃⁺ (4)

CO₂+CO₃²⁻+H₂O⇄2HCO₃⁻ (5)

R-NH₂+CO₂+R₁-OH⇄R-NH₃⁺+R₁-OCO₂⁻ (6)

Accordingtoreaction1,theprimaryalkanolamine DGA,denoted hereasR-NH₂(R=–CH₂CH₂–O–CH₂CH₂–OH),reactswithCO₂topro- duceDGAcarbamate(R-NHCO₂⁻)andtheprotonatedspeciesDGAH⁺ (R-NH₃⁺),irrespectiveofthetypeofdiluentused.Incontrast,carbonate andbicarbonateionscanonlybeproducedinwater(reactions2–5).Or- ganicdiluentswithahydroxylgroupsuchasDEGMME,denotedhereas R₁–OH(R₁=–CH₂CH₂–O–CH₂CH₂–O–CH₃),canalsoreactwithboth CO₂,whenalkalinespeciessuchasDGAarepresent,toformthecor- respondingalkylcarbonateR₁-OCO₂⁻(reaction6).Thisreactioniski- neticallyandthermodynamicallylessfavourablethantheothersand,in fact,generallyoccursduringprolongedabsorption,whentheamountof freeamineislow;moreover,thealkylcarbonateformedhaslowstability

F. Barzagli, U.H. Bhatti, W.W. Kazmi et al. Carbon Capture Science & Technology 8 (2023) 100124

Fig.2.¹³CNMRspectraofDGA/DEGMMEandDGA/H₂OsolutionsafterCO₂saturation.ThecarbonatomsofDGAandDGAH⁺(uniquesignal)areindicatedby numbers,andasterisksdenotethecarbonatomsofDGAcarbamate.DindicatesDEGMMEsignals.ThecarbonylcarbonsofDEGMMEcarbonate,HCO₃⁻/CO₃²⁻and DGAcarbamateareindicatedbya,b/candC,respectively(peakintensitynotscaled).

(Barzaglietal.,2019;Karlssonetal.,2021).ThespectrainFig.2clearly showsthatCO₂isabsorbedalmostexclusivelyasDGAcarbamate(re- action1,DGA/CO₂stoichiometry=2/1)whenthereactiontakesplace inDEGMME,whereasinaqueoussolutionCO₂iscapturedbothascar- bamateandbicarbonate(reaction2,DGA/CO₂stoichiometry=1/1),in comparableamounts:thisexplainsthehigherloadingvalueobtainedin H₂Othaninorganicsolvent.

Catalyst-assistedsorbentregeneration

To evaluatethe potential benefits of combining the use of non- aqueoussorbentswiththeuseofcatalyststofacilitateregenerationcom- paredtoconventionalsorbents,theCO₂desorptionperformanceofsatu- ratedsolutionsofDGA/DEGMMEandDGA/H₂Owithoutandwithfive distinctcatalystswasstudied.Thecatalystsusedweretwometalox- ides,V₂O₅andWO₃,andthreesilicatederivatives,namelyHCl-MONT, 20%Mo/SBA-15,and20%Mo/SAPO-11,whichhavealreadyprovento beactivewithseveralaqueoussolutionsofaminesand,inthecaseof metaloxides,alsowithnon-aqueousDGAsorbents(Bhattietal.,2021a, 2020).DesorptionexperimentswerecarriedoutatT=85°Conaliquots (0.025L)ofthepreviouslysaturatedsolutionsaccordingtotheproce- duredescribedinSection2.3.Foreachsolution,desorptionwasper- formedwithoutcatalysts(blank)andwith1.25g(5%wt)ofeach of thefiveselectedcatalysts.Duringthe60-minuteexperiment,theCO₂ releasedwas continuouslymeasuredby theGC,asdescribed inSec- tion2.3;fromthesedata,therateofCO₂desorption(mmolCO₂‧min⁻¹), thecumulativequantityofCO₂desorbed(mmol)andthepercentage ofregeneratedsorbentwerecalculated.Allresultsobtainedwithand withoutcatalystsduringthedesorptionexperiments foraqueousand non-aqueousDGAsolutionsaregivenintheSupportingInformationin TablesS1andS2,respectively.Moreover,themostsignificantdatafor comparingthedifferentregenerationperformancesareshownherein Table1.

Oncetheexperimentswerecompleted,thecatalystswererecovered fromtheregeneratedsolutionsbyfiltration,withtheexceptionofV₂O₅, whichissolubleinbothwaterandDEGMMEatthedesorptiontemper- atureandcannotbecompletelyseparatedevenaftercooling.

For both DGA solutions, the change in CO₂ desorption rate (mmol‧min⁻¹)intheabsenceandpresenceofthedifferentcatalystsover thedesorptiontimeisillustratedinFig.3.

Eitherinaqueousandnon-aqueoussolutions,allthecatalyststested wereabletosignificantlyincreasetherateofCO₂desorptionwithre- specttothesolutionwithoutcatalysts(blank)duringthefirstfewmin- utesofdesorption.Astheregenerationproceeded(andtheresidualCO₂ inthesolutionsdecreased),thedesorptionratesforallsolutionstended to decrease,andsimilar values wererecorded. Inthe DGA/H₂Oso- lution,themaximum desorptionrateobtainedwiththecatalysts de- creases in theorder 20%Mo/SAPO-11> 20%Mo/SBA-15 > V₂O₅ >

WO₃≈HCl-MONT.Inparticular,20%Mo/SAPO-11and20%Mo/SBA- 15recordedthehighestabsolutedesorptionratevalues,of2.70and2.24 mmol‧min⁻¹,respectively(Table1),andamaximumdesorptionratein- creaseofupto520%and415%,respectively,comparedtotheaqueous blank (TableS1).Differently,in DEGMME,thebest performingcata- lystsaremetaloxides,andthebeneficialeffectonsorbentregeneration decreasesintheorderV₂O₅>WO₃>20%Mo/SAPO-11>20%Mo/SBA- 15 > HCl-MONT. Theaddition of V₂O₅ totheDGA/DEGMME solu- tionresultedinamaximumdesorptionratevalueof3.19mmol‧min⁻¹ (Table1)andamaximumdesorptionrateincreaseof86%comparedto thenon-aqueousblank(TableS2).

Mostnotably,Fig.3clearlyshowsthattherateofCO₂desorptionat 85°CofDGAintheorganicdiluentissignificantlyhigherthaninwater underthesameoperatingconditions.Withoutadditionofanycatalyst, theCO₂desorptionrateofblankDGA/DEGMMEishigherthanthatof blankDGA/H₂O(themaximumvaluesare2.16and1.75mmol‧min⁻¹, respectively).Theadditionofcatalystssignificantlyincreasesthedes- orptionrateforboth solutions,however thevaluesobtained innon- aqueous solution arestillhigher. Moreover, these higher values are maintainedforlongerperiodsoftime,thus allowingmoreCO₂ tobe desorbedforthesameamountoftime.Thissituationcanbebetterun- derstoodbylookingatFig.4,whichshowsthepercentageofsorbent regeneration(ratio oftheamountof CO₂ desorbedtotheamountof CO₂absorbedinthesaturatedsolution)asafunctionoftimeforboth aqueousandnon-aqueoussolutions.

Due to the highest (and longer lasting) desorption rate values, DGA/DEGMMEisalready50%regeneratedwithin14minwithoutcat- alystsandevenwithin11minwiththeadditionofV₂O₅.Thesameper- centageofregenerationinDGA/H₂Oisonlyachievedbetween27min (with20%Mo/SAPO-11)and33min(blank).Inorganicdiluent,regen- erationat85 °C isaround90%attheend ofthe60-minuteheating

4

Table1

Parametersforevaluatingtheregenerationperformanceofaqueousandnon-aqueousDGAsolutionswithandwithout catalysts:maximumdesorptionrate,totalamountofCO₂released,CO₂ loadingafterdesorption,percentageoffinal regeneration,timerequiredtoregenerate50%ofthesorbent.

DGA/H 2O solution –CO 2loading after saturation: 0.647

catalyst

Max des.

rate (mmol ‧ min ⁻¹)

Time at max des. rate

(min) Total CO 2

des. (mmol) CO 2

Loading

after des. Sorbent regen.

(%) Time for 50%

regen. (min)

Blank (no cat.) 1.75 7 ′ 31.17 0.232 64.2 33‘

WO 3 1.78 5 ′ 32.93 0.208 67.8 29 ′

V 2O 5 1.90 5 ′ 32.87 0.209 67.7 28 ′

HCl-MONT 1.76 5 ′ 31.44 0.228 64.8 31 ′

20%Mo/SBA-15 2.24 3 ′ 31.54 0.227 65.0 32 ′

20%Mo/SAPO-11 2.70 3 ′ 33.01 0.207 68.0 27 ′

DGA/DEGMME solution –CO ₂loading after saturation: 0.576

catalyst Max des.

rate (mmol ‧ min ⁻¹)

Time at max des. rate (min)

Total CO 2

des. (mmol) CO 2

Loading after des.

Sorbent regen.

(%)

Time for 50%

regen. (min)

Blank (no cat.) 2.16 9 ′ 37.35 0.078 86.4 14 ′

WO 3 2.69 5 ′ 39.65 0.048 91.7 12 ′

V 2O 5 3.19 5 ′ 38.93 0.057 90.1 11 ′

HCl-MONT 2.21 7 ′ 38.43 0.064 88.9 13 ′

20%Mo/SBA-15 2.35 7 ′ 39.33 0.052 91.0 13 ′

20%Mo/SAPO-11 2.49 7 ′ 39.56 0.049 91.5 12 ′

Fig.3.VariationoftherateofCO₂desorption(mmolCO₂‧min⁻¹)intheabsenceandpresenceofthedifferentcatalystsoverthedesorptiontime(at85°C).

Fig.4. Variationofsorbentregenerationpercentage(%)intheabsenceandpresenceofthedifferentcatalystsoverthedesorptiontime(at85°C).

time,whereasinwaterregenerationvaluesarearound65%(Table1).

SincetheCO₂loadingofDGAinorganicdiluentislowerthaninwater, thepercentageofregenerationalonemaynotbesufficienttoassessthe desorptionperformance:itistherefore necessarytoverifytheactual amountofCO₂desorbed(mmol)attheendoftheprocess.However, eventhesevaluesconfirmthatdesorptionishigherinDEGMME(total

CO₂desorbedbetween37and40mmol)thanintheDGA/H₂Osolution, whichdesorbedbetween31and33mmolofCO₂(Table1).CO₂load- ingvaluesattheendofdesorptionalsoconfirmthatnon-aqueoussolu- tionsaresubstantiallyregeneratedandreadyforsubsequentabsorption (loadingsbetween0.048and0.078),whereasaqueoussolutionsexhibit higherCO₂loadings,intherange0.207–0.232(Table1)

F. Barzagli, U.H. Bhatti, W.W. Kazmi et al. Carbon Capture Science & Technology 8 (2023) 100124

Fig.5. Variationovertimeoftheconcentration(mol‧L⁻¹)ofthedifferentspeciesinDGAaqueousandnon-aqueoussystems,withoutandwith5wt%20%Mo/SAPO- 11catalyst,calculatedby¹³CNMRanalysis.

Theseresultsconfirmthepotentialenergysavingsresultingfromthe useof aminesin organicdiluents,whicharecapableof almostcom- pleteregenerationatmuchlowertemperaturesthanthoseconvention- allyusedinindustrialplantswithaqueousMEA(120–140°C).Further- more,itisclearfromFigs.3and4thattheeffectofsolidacidcatalysts istosignificantlyenhancedesorption(andthusregeneration)especially inthefirst10–15minofheating:fasterregenerationisofcrucialimpor- tanceinthedesignofcontinuousCO₂absorptionanddesorptionplants.

13CNMRstudyofthedesorptionmechanism

Inordertobetterunderstandthedifferentbehaviourduringdesorp- tion,adetailed¹³CNMRstudywascarriedouttoidentifythespecies presentinthesolutions duringtheregenerationanddeterminetheir concentration.Specifically,solutionsamplesweretaken(andanalysed) after5,10,15,25,40,and60minof heatingat85°C.Thecompar- isonwasmadeforbothDGAsolutionsintheabsenceandpresenceof 20%Mo/SAPO-11,whichprovedtobethemostefficientcatalystinH₂O, andalsoworkedinorganicdiluent.

Theresults ofthisstudyarepresentedinFig.5.Inparticular,for eachsolution-catalystcombination,thegraphsshowthechangeincon- centration(mol‧L⁻¹) over thedesorptiontimefor freeamine(DGA), protonatedamine(DGAH⁺),DGAcarbamate,bicarbonateion(onlyin aqueoussolutions)andDEGMMEcarbonate(onlyinorganicdiluent).

Theinitialvalues(attime=0min)arethoseforsaturatedsolutions (shownin Fig.2).Thecarbonateion(presentonlyin water)hasnot beenreportedasitsconcentrationisnegligibleattimezeroandisnot detectableduringregeneration.

Thespeciationprofilesaresignificantlydifferentbetweenaqueous andnon-aqueous solutions.Inagreementwiththedesorption perfor- manceresults,Fig.5highlightsthattheconcentrationofDGAincreases veryrapidlyinDEGMME,untilitreachesaconcentrationverycloseto 3mol‧L⁻¹,i.e.,theinitialconcentration.Onthecontrary,inaqueousso-

lution,theconcentrationofDGAincreasesmoreslowlyandafter60min barelyreaches1.7mol‧L⁻¹,eveninthepresenceofthecatalyst.

Another difference clearlyseen in Fig. 5concerns thedecompo- sition of thecompoundsproduced in theCO₂-saturated solution. In DGA/DEGMMEsolutions,themainabsorptionproductisDGAcarba- mate.Duringtheregenerationprocess,itdecomposesquiterapidlyin thefirst15min,andthenslowlyreachesconcentrationsclosetozero asdesorption continues.Thepresence ofthe catalystaccelerates the decompositionofDGAcarbamate,asevidentbycomparingitsconcen- trationsbetween5and25minwithorwithout20%Mo/SAPO-11.

Innon-aqueoussystems,thedecompositionreactionofDGAcarba- mateisthereverseofreaction1:

R-NHCO₂⁻+R-NH₃⁺⇄2R-NH₂+CO₂ (7)

Asreportedinoneofourpreviousstudies(Bhattietal.,2021a),the presenceofanacidcatalystpresumablyincreasesthenumberofprotons availableforthereactionwithcarbamate,therebyincreasingtherate ofdecompositionaccordingtoreactions8and9(Alivandetal.,2020; Zhangetal.,2018).

R-NHCO₂⁻+catH⁺⇄R-NH₂+cat+CO₂ (8)

R-NH₃⁺+cat⇄R-NH₂+catH⁺ (9)

Finally,DEGMMEcarbonatewasnotdetectedinanyofthesamples collectedduringregeneration:duetoitslowinitialconcentrationand poorstability,itwasfullydecomposedwithinthefirst5minofdesorp- tion,accordingto(reverse)reaction6.

Differently, intheaqueous solution,theregenerationof DGAde- pendsonthedecompositionofbothcarbamateandbicarbonate,which areinsimilarconcentrationswhenthesolutionissaturatedwithCO₂. Thespeciationstudyhighlightsaninterestingbehaviour,nottobeex- pectedconsideringonlythespeciesdecompositionreactions(inverseof 6

reactions1–5):infact,duringthefirst10 minof heating,thecarba- mateconcentrationincreasesinsteadofdecreasing.Atthesametime,a decreaseinDGAH⁺and,evenmoremarkedly,inHCO₃⁻canbeseen.

Thesetrendssuggestthata reactionbetweenbicarbonateandproto- natedspeciestoformDGAcarbamateisprominentintheinitialphase ofdesorption,accordingtoreaction10(Barzaglietal.,2009).

2HCO₃⁻+R-NH₃⁺⇄R-NHCO₂⁻+CO₂+2H₂O (10)

ThisreactionexplainsboththeCO₂ desorptionrecordedwithgas chromatograph(Fig.3)andtheverylimitedregenerationofDGAatthe beginningofthedesorption(Fig.5).After10minthebicarbonatecon- centrationisalmostzero,andtheDGAcarbamatebeginstodecompose.

Theincrementintherateofdesorptionrecordedinthepresenceofthe 20%Mo/SAPO-11(Fig.3)isattributabletothemorerapiddecreasein thebicarbonateconcentration:bycomparingthespeciationcurves,the concentrationofHCO₃⁻after5minofheatingissignificantlylowerin thepresenceofthecatalystthanintheblank.Inthiscase,theacidsolid presumablycatalysestheprotonationofbicarbonate(reaction11),and isrestoredbysubtractingaprotonfromDGAH⁺(reaction12),whichis alsoconfirmedbythehigherconcentrationofDGApresent.

HCO₃⁻+catH⁺⇄H₂O+CO₂+cat (11)

R-NH₃⁺+cat⇄R-NH₂+catH⁺ (12)

Conclusions

Toovercomethecurrentconstraintsofconventionalcarboncapture technologies,mainlyrelatedtothehighenergycostofdesorption,in thisstudyweproposedaninnovativetechniqueforCO₂capturethat combinestheuseofnon-aqueous sorbentswiththeuse ofacidcata- lystsforregeneration.Tosummarize,theregenerationperformanceof theprimaryamineDGAwasstudiedwithtwoco-solvents,traditionally usedwaterandorganicDEGMME,inthepresenceandabsenceoffive differentacidiccatalysts.Opposedtotheconventional120–140°Cre- generationtemperature,thesorbentregenerationwasstudiedatamild temperatureof85°C,andtherateofCO₂ desorption,thequantityof CO₂released,andthepercentageofsorbentregeneratedwereusedas thekeyperformanceindicators.Asafindingofthisstudy,theproposed newtechniquecombiningnon-aqueoussorbentswithacidiccatalysts fordesorptiondemonstratedsignificantlysuperiorregenerativeperfor- mancetobothnon-aqueoussorbentswithoutcatalystsand,toagreater extent,tosolutionsof thesameaminein water.Theuse ofV₂O₅ in thedesorptionoftheDGAsolutioninDEGMMEresultedinthehigh- estdesorptionratevalueof3.19mmol‧min⁻¹,whichwassignificantly betterthanboththatobtainedwiththesamesolutionwithoutcatalyst (2.16mmol‧min⁻¹)andthoseobtainedinaqueousDGAsolutions(1.75 mmol‧min⁻¹ without catalyst,2.70mmol‧min⁻¹ with 20%Mo/SAPO- 11).SimilarresultswerealsoobtainedwithWO₃,withtheadvantage overV₂O₅of beinginsolubleandthuseasily separableattheend of desorption.Intheorganicdiluent,itwasobservedthatthedesorption ratevaluesremainedhighforalongertimethaninwater:asaconse- quence,inthenon-aqueoussolution,itwaspossibletoregeneratethe sorbentby50%after11–14min,whereasinwaterthispercentagewas obtainedonlyafter27–33min.Moreover,theadditionofcatalystsin non-aqueoussorbentallowedregenerationvaluesofaround90%tobe achievedafter60minofheatingto85°C,significantlybetterthanthose obtainedinwater(65–68%).The¹³C NMRspeciationanalysisofthe desorptionprocessesclarifiedthedegradationmechanismoftheamine carbamate,thusexplainingthedifferentbehaviourinthedifferentDGA- diluent-catalystcombinations.

Inconclusion,initialevaluationsofthisproposednewapproachsug- gestgreatpotentialforCO₂captureefficienciesfromgaseousmixtures,

coupledwithregenerativeprocessesatasignificantlylowerenergycost thanconventionalaqueousMEAtechnologies.Furtherinvestigationwill berequiredtoidentifythebestcombinationofamine,organicdiluent, andcatalysttominimiseenergyexpenditurewhilemaximisingCO₂cap- tureefficiency,aswellastoassessitssustainedperformanceovertime, includingaminestabilityunderthechosenoperatingconditionsandcat- alystreusability.

DeclarationofCompetingInterest

Theauthorsdeclarethattheyhavenoknowncompetingfinancial interestsorpersonalrelationshipsthatcouldhaveappearedtoinfluence theworkreportedinthispaper.

Acknowledgements

ThisworkwassupportedbytheCNR-ICCOMInstitutethroughthe projectSPICCO2(projectcodeDCM.AD004.109).

Supplementarymaterials

Supplementarymaterialassociatedwiththisarticlecanbefound,in theonlineversion,atdoi:10.1016/j.ccst.2023.100124.

References

Al-Juaied, M., Rochelle, G.T., 2006. Absorption of CO2 in Aqueous Diglycolamine. Ind.

Eng. Chem. Res. 45, 2473–2482. doi: 10.1021/ie0505458 .

Alivand, M.S., Mazaheri, O., Wu, Y., Stevens, G.W., Scholes, C.A., Mumford, K.A., 2020.

Catalytic Solvent Regeneration for Energy-Efficient CO2 Capture. ACS Sustain. Chem.

Eng. 8, 18755–18788. doi: 10.1021/acssuschemeng.0c07066 .

Alkhatib, I.I.I., Galindo, A., Vega, L.F., 2022. Systematic study of the effect of the co- solvent on the performance of amine-based solvents for CO2 capture. Sep. Purif. Tech- nol. 282, 120093. doi: 10.1016/j.seppur.2021.120093 .

Atzori, F., Barzagli, F., Varone, A., Cao, G., Concas, A., 2023. CO2 absorption in aqueous NH3 solutions: novel dynamic modeling of experimental outcomes. Chem. Eng. J.

451, 138999. doi: 10.1016/j.cej.2022.138999 .

Barzagli, F., Giorgi, C., Mani, F., Peruzzini, M., 2019. Comparative study of CO2 cap- ture by aqueous and nonaqueous 2-amino-2-methyl-1-propanol based absorbents car- ried out by 13C NMR and enthalpy analysis. Ind. Eng. Chem. Res. 58, 4364–4373.

doi: 10.1021/acs.iecr.9b00552 .

Barzagli, F., Giorgi, C., Mani, F., Peruzzini, M., 2018. Reversible carbon dioxide capture by aqueous and non-aqueous amine-based absorbents: a comparative analysis carried out by 13C NMR spectroscopy. Appl. Energy 220, 208–219.

doi: 10.1016/j.apenergy.2018.03.076 .

Barzagli, F., Lai, S., Mani, F., 2013. CO2 capture by liquid solvents and their regeneration by thermal decomposition of the solid carbonated derivatives. Chem. Eng. Technol.

36, 1847–1852. doi: 10.1002/ceat.201300225 .

Barzagli, F., Mani, F., 2021. Direct CO2 air capture with aqueous 2-(ethylamino)ethanol and 2-(2-aminoethoxy)ethanol: 13C NMR speciation of the absorbed solutions and study of the sorbent regeneration improved by a transition metal oxide catalyst. Inor- ganica Chim. Acta 518, 120256. doi: 10.1016/j.ica.2021.120256 .

Barzagli, F., Mani, F., Peruzzini, M., 2009. A 13C NMR study of the carbon dioxide absorp- tion and desorption equilibria by aqueous 2-aminoethanol and N-methyl-substituted 2-aminoethanol. Energy Environ. Sci. 2, 322–330. doi: 10.1039/b814670e . Barzagli, F., Peruzzini, M., Zhang, R., 2022. Direct CO2 capture from air with aqueous

and nonaqueous diamine solutions: a comparative investigation based on 13C NMR analysis. Carbon Capture Sci. Technol. 3, 100049. doi: 10.1016/j.ccst.2022.100049 . Bhatti, U.H., Ienco, A., Peruzzini, M., Barzagli, F., 2021a. Unraveling the Role of Metal

Oxide Catalysts in the CO2 Desorption Process from Nonaqueous Sorbents: an Exper- imental Study Carried out with 13C NMR. ACS Sustain. Chem. Eng. 9, 15419–15426.

doi: 10.1021/acssuschemeng.1c04026 .

Bhatti, U.H., Kazmi, W.W., Muhammad, H.A., Min, G.H., Nam, S.C., Baek, I.H., 2020.

Practical and inexpensive acid_activated montmorillonite catalysts for energy-efficient CO2capture. Green Chem 22, 6328–6333. doi: 10.1039/d0gc01887b .

Bhatti, U.H., Nam, S., Park, S., Baek, I.H., 2018. Performance and Mechanism of Metal Ox- ide Catalyst-Aided Amine Solvent Regeneration. ACS Sustain. Chem. Eng. 6, 12079–

12087. doi: 10.1021/acssuschemeng.8b02422 .

Bhatti, U.H., Shah, A.K., Kim, J.N., You, J.K., Choi, S.H., Lim, D.H., Nam, S., Park, Y.H., Baek, I.H., 2017. Effects of Transition Metal Oxide Catalysts on MEA Solvent Regen- eration for the Post-Combustion Carbon Capture Process. ACS Sustain. Chem. Eng. 5, 5862–5868. doi: 10.1021/acssuschemeng.7b00604 .

Bhatti, U.H., Sultan, H., Min, G.H., Nam, S.C., Baek, I.H., 2021b. Ion-exchanged montmo- rillonite as simple and effective catalysts for efficient CO2 capture. Chem. Eng. J. 413, 127476. doi: 10.1016/j.cej.2020.127476 .

Bui, M., Adjiman, C.S., Bardow, A., Anthony, E.J., Boston, A., Brown, S., Fennell, P.S., Fuss, S., Galindo, A., Hackett, L.A., Hallett, J.P., Herzog, H.J., Jackson, G., Kem- per, J., Krevor, S., Maitland, G.C., Matuszewski, M., Metcalfe, I.S., Petit, C., Puxty, G., Reimer, J., Reiner, D.M., Rubin, E.S., Scott, S.A., Shah, N., Smit, B., Trusler, J.P.M.,