348 Appendix D * Evidence for Organosulfates in Secondary Organic Aerosol * This chapter is reproduced by permission from “Evidence for Organosulfates in Secondary Organic Aerosol” by J. D. Surratt, J. H. Kroll, T. E. Kleindienst, E. O. Edney, M. Claeys, A. Sorooshian, N. L. Ng, J. H. Offenberg, M. Lewandowski, M. Jaoui, R. C. Flagan, J. H. Seinfeld, Environmental Science and Technology, 41: 517-527 2006. Copyright 2006. American Chemical Society. 349 Environ.Sci.Technol.2007,41,517-527 Evidence for Organosulfates in (1).Particle-phaseoligomerizationleadstotheformationof high-molecular-weight(MW)species(2-4);suggestedoli- Secondary Organic Aerosol gomerizationreactionsincludethereactiveuptakeofvolatile aldehydesorketonesviaperoxyhemiacetalformation(5,6), JASON D. SURRATT,† JESSE H. KROLL,‡,X hydration,hemiacetal/acetalformation,andaldolconden- TADEUSZ E. KLEINDIENST,§ sation (7, 8). Esterification in isoprene photooxidation (9, EDWARD O. EDNEY,§ MAGDA CLAEYS,^ 10) has also been reported in SOA formation. The role of ARMIN SOROOSHIAN,| NGA L. NG,| these reactions remains in some doubt as some of the JOHN H. OFFENBERG,§ proposedreactions(e.g.,hemiacetal/acetalformationand MICHAEL LEWANDOWSKI,§ aldolcondensation)arenotthermodynamicallyfavorableat MOHAMMED JAOUI,# ambientconditions(11,12). RICHARD C. FLAGAN,‡ AND Laboratorychamberstudieshavedemonstratedthatthe JOHN H. SEINFELD*,‡ presence of acidic seed aerosol enhances the SOA yields DepartmentofChemistry,CaliforniaInstituteofTechnology, observed from the oxidation of various volatile organic Pasadena,California91125,DepartmentsofEnvironmental compounds(VOCs),suchasR-pinene(2,3,13,14),isoprene ScienceandEngineeringandChemicalEngineering, (8, 14, 15), and several model cycloalkenes (2) over those CaliforniaInstituteofTechnology,Pasadena,California withalessacidicseedaerosol.Despiterecentadvancesin 91125,NationalExposureLaboratory,OfficeofResearchand understanding particle-phase SOA chemistry, the role of Development,EnvironmentalProtectionAgency,Research particle-phaseacidityinenhancingSOAformationremains TrianglePark,NorthCarolina27711,Departmentof essentiallyunexplained.Recentstudieshavepresentedmass PharmaceuticalSciences,UniversityofAntwerp(CampusDrie spectrometricevidencethatthereactiveuptakeofglyoxal Eiken),Universiteitsplein1,BE-2610Antwerp,Belgium, andpinonaldehyde(principalgas-phaseoxidationproducts DepartmentofChemicalEngineering,CaliforniaInstituteof fromaromaticsandR-pinene,respectively)onacidicaerosol Technology,Pasadena,California91125,andAlionScience involvesorganosulfateformation(sulfateestersorderivatives; andTechnology,P.O.Box12313,ResearchTrianglePark, forsimplicity,wewillusehereafterthetermsulfateesters NorthCarolina27709 to also denote sulfate derivatives; i.e. sulfate derivatives formed from a carbonyl compound) (16-18). In addition, severalfieldstudieshavereportedsulfateestersignaturesin Recent work has shown that particle-phase reactions aerosolcollectedonfiltersusingFouriertransforminfrared spectroscopy(19,20)andelectrosprayionization(ESI)-mass contribute to the formation of secondary organic aerosol spectrometry(MS)(21,22).Nevertheless,theimportanceof (SOA),withenhancementsofSOAyieldsinthepresenceof organosulfateformationtoSOAremainsunclear.Traditional acidicseedaerosol.Inthisstudy,thechemicalcomposition analytical methods, such as gas chromatography/mass of SOA from the photooxidations of R-pinene and spectrometry(GC/MS)withpriorderivatization,maynotbe isoprene, in the presence or absence of sulfate seed wellsuitedforidentifyingorganosulfates.Itislikelythatsingle aerosol, is investigated through a series of controlled derivatization protocols, such as trimethylsilylation, GC chamberexperimentsintwoseparatelaboratories.Byusing injectionandcolumntemperaturescouldcausethedegra- electrospray ionization-mass spectrometry, sulfate dationormisinterpretationofsuchspecies(23).Ontheother esters in SOA produced in laboratory photooxidation hand,ESI-MShasbeenshownasaneffectivemethodfor experiments are identified for the first time. Sulfate esters the detection and quantification of organosulfate species (24,25). are found to account for a larger fraction of the SOA In the present study ESI-MS is used to detect and masswhentheacidityofseedaerosolisincreased,aresult structurallyelucidatesulfateestersinSOAformedfromthe consistent with aerosol acidity increasing SOA formation. photooxidationsofisopreneandR-pineneunderdiffering Many of the isoprene and R-pinene sulfate esters combinationsofNO levelsandseedaerosolacidities.Asa x identified in these chamber experiments are also found in result, the formation of sulfate esters may be a major ambient aerosol collected at several locations in the contributortotheobservedenhancementinSOAyieldsin southeastern U.S. It is likely that this pathway is important thepresenceofacidicaerosol. for other biogenic terpenes, and may be important in the Experimental Section formationofhumic-likesubstances(HULIS)inambientaerosol. IsopreneChamberExperiments.Asummaryofexperimental Introduction conditionsforallisoprenephotooxidationexperimentscan befoundinTable1.Isoprenephotooxidationexperiments Particle-phase reactions are now understood to play an were conducted in Caltech’s dual indoor 28 m3 Teflon importantroleinsecondaryorganicaerosol(SOA)formation chambers(26,27)andinEPA’sfixedvolume14.5m3indoor chamber(15).Thetemperatures,aerosolsizedistributions, *Correspondingauthorphone: (626)395-4635;fax: (626)796- andrelativehumidities,aswellastheO,nitricoxide(NO), 2591;e-mail: [email protected]. 3 NO concentrations were continuously measured in both †DepartmentofChemistry,CaliforniaInstituteofTechnology. x ‡DepartmentsofEnvironmentalScienceandEngineeringand facilities.Theisopreneconcentrationsinbothfacilitieswere ChemicalEngineering,CaliforniaInstituteofTechnology. monitoredbyGCwithflameionizationdetection.TheCaltech §EnvironmentalProtectionAgency. experimentswereconductedinthestaticmode(i.e.,batch ^ UniversityofAntwerp. reactor) whereas the EPA experiments were conducted in |Department of Chemical Engineering, California Institute of thedynamicmode(i.e.,continuousstirredtankreactor)with Technology. theexceptionofEPA-326whichwasastaticmodeexperiment. #AlionScienceandTechnology. XCurrentaddress: AerodyneResearch,Inc.,45ManningRoad, Hydroxyl radical (OH) precursors (H2O2 or HONO) were Billerica,MA01281. employedintheCaltechexperiments(28,29).ForallCaltech 10.1021/es062081qCCC:$37.00 ª 2007AmericanChemicalSociety VOL.41,NO.2,2007/ENVIRONMENTALSCIENCE&TECHNOLOGY9517 PublishedonWeb12/07/2006 350 TABLE 1.Summary of Experimental Conditions and Sulfate Ester Formation from Isoprene Photooxidation initial initial inorganicseed OH isoprene] [NOx] average experiment aerosola precursor (ppb) (ppb) T((cid:176) C) [M-H]-detectedsulfateesterions(m/z) Caltechlow-NO noneadded H O 500 noneadded 23.7 nonedetected x 2 2 Caltechlow-NO AS H O 500 noneadded 23.9 215,333 x 2 2 Caltechlow-NO AAS H O 500 noneadded 23.8 153,155,169,215,333,451 x 2 2 Caltechhigh-NO noneadded H O /NO 500 891 24.3 nonedetected x 2 2 Caltechhigh-NO AS H O /NO 500 963 24.9 199,215,244 x 2 2 Caltechhigh-NO AAS H O /NO 500 904 24.6 139,153,155,197,199,215,244,260,301,346 x 2 2 Caltechhigh-NO noneadded HONO 500 382 20.1 nonedetected x Caltechhigh-NO AS HONO 500 366 21.4 199,215,260,333 x EPA-299stage1 ASb c 2500 200 29.0 nonedetected EPA-299stage2 AAS c 2500 200 29.0 153,155,167,169,181,197,199,215,244, 260,333 EPA-299stage3 H SO only c 2500 200 29.0 153,155,157,167,169,181,197,199,215,244, 2 4 260,333 EPA-199stage1 60ppbSO c 1598 475 24.6 197,199,215,244,260,301,317,333 2 EPA-199stage2 200ppbSO c 1598 475 24.6 155,169,197,199,215,244,260,301,317, 2 333,346 aAS)15mM(NH)SO;AAS)15mM(NH)SO +15mMHSO forCaltechexperimentsand0.31mM(NH)SO +0.612mMHSO for 42 4 42 4 2 4 42 4 2 4 EPA-299experiments;HSO only)0.92mMHSO;EPA-199hadnoseednebulizedbutinsteadusedthephotooxidationofSO togenerate 2 4 2 4 2 sulfuricacidaerosol.bDuetothelowinitialinorganicseedaerosolconcentration,thisconditionismoreconducivetonucleation.cNoOHprecursor wasused. TABLE 2.Summary of Experimental Conditions and Sulfate Ester Formation from r-Pinene Photooxidation initial initial initial initial [r-pinene] [isoprene] [toluene] [NOx] SO2 averageT [M-H]-detectedsulfate experiment (ppb) (ppb) (ppm) (ppb) (ppb) ((cid:176) C) esterions(m/z)a EPA-220 220 b b 450 b 25.5 nonedetected EPA-326 186 1108 b 248 287 20.5 249,265,294,310,412,426 EPA-211stage1 105 b b 378 b 24.3 nonedetected EPA-211stage2 105 b 1.59 378 b 24.3 nonedetected EPA-211stage3 103 820 1.58 378 b 24.3 nonedetected EPA-211stage4 117 854 1.57 378 275 24.3 265,279,294,310,326,412,426 EPA-211stage5 115 b 1.59 378 275 24.3 249,265,279,294,310,326 EPA-211stage6 b 794 1.56 378 b 24.3 nonedetected EPA-205 106 592 1.45 599 278 24.0 265,294,310,412,426 aIsoprenesulfateesterproductslikethoseinTable2werealsodetectedonlywhenSO andisoprenewerecopresent.Nodiscernibletoluene 2 sulfateesterproductsweredetected.bThiscompoundwasnotpresentduringtheexperiment. low-NO experiments,onlyHO wasadded,resultinginNO followingaerosolconditionswereused: AS,AAS,andsulfuric x 2 2 x concentrations<1ppb.Caltechhigh-NO experimentseither acid,witheachoftheaqueoussolutionsatomizedintothe x usedHO andaninitialamountofNO((cid:24)800ppb),orwith chamberbyatomization.Theinitialaerosolconcentrations 2 2 HONOandNO asasideproduct.IntheCaltechexperiments, were0.1,30.0,and30.0(cid:237)g/m3,forEPA-299stage1,EPA-299 x three initial inorganic seed aerosol conditions were used: stage2,andEPA-299stage3,respectively.Forthesecondset (1)theabsenceofaerosolwhereSOAformationwasinitiated ofEPAexperiments,EPA-199stage1andEPA-199stage2, bynucleation;(2)ammoniumsulfate(AS)aerosol;and(3) acidicaerosolwasgeneratedbyadding60and200ppbof acidifiedammoniumsulfate(AAS)aerosol.Concentrations SO,respectivelytothereactantmixture.AllEPAexperiments 2 of the aqueous solutions that were introduced into the wereconductedatarelativehumidityof(cid:24)30%.Resultsfrom chambersbyatomizationareshowninTable1.Theinitial bothEPAexperiments(i.e.,SOAyields,gas-phaseproducts, seedaerosolconcentrationsthatresultedrangedfrom(cid:24)19- trends,etc.)willbediscussedinmoredetailinforthcoming 24(cid:237)g/m3.Teflonfilters(PALLLifeSciences,47-mmdiameter, publications;evidencefororganosulfatesisthefocushere. 1.0-(cid:237)m pore size, Teflon filters) were collected for offline r-Pinene Chamber Experiments. All R-pinene experi- chemicalanalysisfromtheCaltechexperimentsatthepoint mentswereconductedintheEPAdynamicchamber(15). atwhichtheaerosolvolumereacheditsmaximumvalue,as Conditions for each experiment are listed in Table 2. The determined by the differential mobility analyzer (DMA). EPAexperimentsconsistedofoneR-pinene/NO irradiation x All experiments were carried out at relative humidities experimentalongwithaseriesofexperimentswheremixtures (RHs)<9%. ofhydrocarbonscontainingR-pinenewereirradiatedinthe InthedynamicexperimentsintheEPAchamber,reactants presenceofNO.Forsomeoftheseexperiments,SO was x 2 suchasNO,SO,andisoprenewerecontinuouslyaddedfrom addedtothechambertogenerateacidityintheaerosol.The 2 high-pressurecylinderstothereactionchamberthrougha samecollectionprotocolwasusedhereasthatemployedin mixingmanifold.Thesteady-statenatureofchamberopera- theEPAisopreneexperiments. tion allows for filter sampling for extended periods for AmbientAerosolCollection.Ambientaerosolwascol- determining the composition of the resultant SOA. Once lected from the Southeastern Aerosol Research and Char- steady-stateconditionswereattained((cid:24)24h),samplesfor acterizationStudy(SEARCH)networkandanalyzedforsulfate determiningthecompositionoftheSOAwerecollectedon esters.Thisnetworkcomprisesoffoururban-rural(orurban- glassfiberfiltersprecededbyacarbonstripdenuder.Two suburban)sitepairsatlocationsacrossthesoutheastU.S. setsofEPAexperimentswereconducted.Inthefirstset,the and was initiated in mid-1998 to carry out systematic 5189ENVIRONMENTALSCIENCE&TECHNOLOGY/VOL.41,NO.2,2007 351 FIGURE1. (-)LC/ESI-MSupfrontCIDmassspectraforselectedisoprenesulfateesterSOAproductsshowninTable2.(A)Production massspectrumforsodiumpropylsulfatestandard(anionicmassofintactpropylsulfateester)139Da).(B)Productionmassspectrum for a 2-methyltetrol sulfate ester detected in a Caltech high-NO HO AS seed photooxidation experiment. (C) Product ion mass x 2 2 spectrumforahemiacetaldimersulfateesterdetectedinaCaltechlow-NO AASseedphotooxidationexperiment.(D)Productionmass x spectrumforaC trihydroxynitratesulfateesterdetectedinEPA-299stage2. 5 measurementsoftemporalandspatialvariabilityofPM,in lium2-(2-methoxyethoxy)ethylsulfate(Sigma-Aldrich,95% particularPM ,gasesrelevanttosecondaryO formation, purity)wereanalyzedontheLC/MSinstrumenttodetermine 2.5 3 and surface meteorology (22). Twenty-four h composite common product ions associated with sulfate esters. Ad- quartz filters were taken on 4 days at four different sites ditionalsampleswerecollectedwiththeparticle-into-liquid during June 2004: Birmingham, AL (BHM, urban site), sampler (PILS) with subsequent offline analysis by ion Centerville,AL(CTR,ruralsiteoutsideofBHM),Jefferson chromatography(IC)(31).ThePILS/ICtechniqueallowsfor Street(JST,neardowntownAtlanta,GA),andPensacola,FL thequantitativemeasurementofwater-solubleinorganicions (PS, marine influenced urban site). Details of these sites inaerosol;noICpeakscouldbeattributedtoorganosulfates, (terrain,vegetation,transportation,andindustrialsources), sofortheseexperimentsonlyinorganicionsaremeasured. sample collection and handling procedures, and specific aerosol and gas-phase measurements obtained are given Results elsewhere(22,30). SulfateEsterStandards.(-)ESI-MSstudieshaveshownthat Filter Extraction and Chemical Analyses. Detailed ex- sulfate esters produce abundant [M-H]- ions, and upon tractionproceduresforTeflonandquartzfiltersaredescribed collisionalactivationoftheseions,yieldm/z97(HSO-)and elsewhere (9, 22). Glass-fiber filters were extracted in the 4 80(SO-(cid:226))productions(21,24,25).Inconjunctionwiththe samemannerasTeflonfilters(9),exceptresultantextracts 3 knownisotopicdistributionofsulfur,whichcontains34Swith werefilteredthroughaPALLLifeSciencesAcrodiscCR25 mm syringe filter (PTFE membrane, 0.2 (cid:237)m pore size) to anaturalabundanceof4.2%,theseproductionscanbeused toidentifysulfateesters.Asconfirmation,authenticstandards removefilterfibers.Allsampleextractswereanalyzedbya ofsodiumpropylsulfate(anionicmass)139Da),sodium Hewlett-Packard1100seriesHPLCinstrument,coupledwith lauryl sulfate (anionic mass ) 265 Da), and 1-butyl-3- aquadrupolemassspectrometer,andbydirectinfusiononto methylimidazolium 2-(2-methoxyethoxy)ethyl sulfate (an- aThermoElectronLCQiontrapmassspectrometer(ITMS), ionicmass)199Da)wereanalyzedbythe(-)LC/ESI-MS bothequippedwithanESIsourceoperatedinthenegative (-)ionizationmode.Detailsoftheoperatingconditionsfor techniqueinthefullscanmodeofanalysisfollowedbythe upfrontCIDmodeofanalysistogenerateMSandMS/MS these instruments are described elsewhere (9). Briefly, all data,respectively.Asshownforthesodiumpropylsulfate sampleswereanalyzedontheLC/MSinstrumentinthefull esterstandardinFigure1A,theseauthenticstandardsyielded scanmodeandupfrontcollision-induceddissociation(CID) m/z97and80productions. modeofanalysis.Comparisonoftheresultingmassspectra producedfromthesetwomodesofanalysesontheLC/MS Sulfate Esters from Isoprene Oxidation. Previously instrument allows for some structural information to be characterized(9,10)isopreneSOAproductswereobserved obtainedonthedetectedSOAcomponents.Sampleswere intheseexperimentsandareshowninTable3;however,the alsoanalyzedontheThermoElectronLCQITMSinstrument focus here will be on the identification of sulfate esters. toconfirmtheseresults,andinsomecases,providefurther Comparisonof(-)ESI-MSdatacollectedfromexperiments structural elucidation. Sulfate standards of sodium propyl employingnosulfateaerosoltothosewithsulfateaerosol sulfate (City Chemical, 98% purity), sodium lauryl sulfate showedthatnumerouscompoundsweredetectedonlywhen (CityChemical,98%purity),and1-butyl-3-methylimidazo- sulfate aerosol was present. To understand the nature of VOL.41,NO.2,2007/ENVIRONMENTALSCIENCE&TECHNOLOGY9519 352 TABLE 3.Proposed Isoprene Sulfate Ester SOA Products aPositionalisomerscontainingnitrateorsulfategroupsatotherhydroxylatedpositionsarelikely.bIsopreneSOAproductspreviouslyidentified inpriorstudiesbySurrattetal.(9)and/orSzmigielskietal.(10)and/orEdneyetal.(15).cDetectedinambientaerosolcollectedfromSEARCH network(summer2004)forfirsttime.dConsideredmajorproductduetolargeMSabundanceinchamberstudies.eCompoundslistedinparentheses areneutrallossesobservedupon(-)ESI-MS/MS.fInferredprecursorduetotheMS/MSfragmentationofitsrespectiveorganosulfateproduct; thisparentisopreneproductgoesundetectedby(-)ESI-MSandGC/MSmethods.gSomeevidenceforitsexistenceinfirst-ordermassspectra. hDetectedinambientaerosolbyMatsunagaetal.(32). 5209ENVIRONMENTALSCIENCE&TECHNOLOGY/VOL.41,NO.2,2007 353 FIGURE2. TimeevolutionoftheSO2-andNH+aerosolmassconcentrationsfromthePILS/ICanalysis.(A)Caltechhigh-NO HO isoprene 4 4 x 2 2 experimentwithASseedaerosol.(B)Caltechlow-NO isopreneexperimentwithAASseedaerosol.Acontrolexperimentwasconducted x inwhichseedaerosolisatomizedfromasolutionof0.015MASintotheCaltechexperimentalchamber,andnootherreactantssuch asVOCsorNO werepresent.ThiscontrolexperimentproducedasimilarresulttothatofFigure2A(althoughnotevidentfromthetime x scalepresented,SO2-andNH+decayby(cid:24)20and14%,respectively,over9h),indicatingthattheonlylossmechanismforsulfatein 4 4 this case was wall loss. Of the Caltech isoprene experiments, only the low-NO AAS seed aerosol experiment showed a significant x decreaseintheSO2-aerosolmassconcentration,indicatingthatitwaslikelylosttoreaction. 4 thesecompounds,tandemMStechniqueswereemployed. experiments.Forexample,inthelow-NO experiments,the x Figure 1B-D shows the LC/ESI-MS upfront CID mass LC/MSpeakareaforthem/z215sulfateesterwasfoundto spectracollectedforlargechromatographicpeakscommon doublewhenAASaerosolwasusedratherthannonacidicAS tomanyofthesulfateaerosolexperimentslistedinTable1. aerosol.Althoughquantitativedatacouldnotbeobtained The [M-H]- ions associated with these chromatographic for either the sulfate ester concentrations or the effective peaksincludem/z215,333,and260,respectively.Aswasthe acidityofthereactionsystem,theseresultssuggestthatsulfate caseforthesulfateesterstandards,thecollisionalactivation ester formation is enhanced by the presence of an acidic ofthese[M-H]-ionsyieldedm/z97and80productions. sulfateaerosol,andthatthisenhancedsulfateesterformation In addition, these ions also had an isotopic distribution may be contributing to the increased SOA mass detected commontosulfur,andasaresult,thesecompoundswere previouslyunderacidicconditions.Furtherworkisneeded identifiedassulfateesters.Theotherproductions,observed in order to accurately quantify these sulfate esters. It was inthespectrashowninFigure1,providedfurtherinformation foundthat(-)LC/ESI-MScalibrationcurvesgeneratedby onthechemicalstructuresoftheidentifiedsulfateesters. surrogate standards lacking sulfate groups (such as meso- Proposedsulfateesterstructuresfortheseionsandallother erythritol) were not suitable for quantifying the identified ionslistedinTable1aregiveninTable3.Manyofthesulfate sulfateesters,resultingfromthesestandardshavinglower esters shown in Table 3 were formed from previously (-)ESI-MSsensitivities.Thesulfateesterstandardslistedin identified isoprene SOA products, including aldehydes, the experimental section were also not suitable for quan- dicarbonyls,hydroxycarbonyls,alcohols,andacidscontain- tification because these compounds had retention times inganalcoholmoiety(9,10,32).Sulfateestersformedfrom much greater than (1 min of the retention times for the smallvolatileoxidationproducts,suchasglyoxal,hydroxy- isoprene sulfate esters. Also, these standards lack many acetone,andglycolaldehyde,wereonlydetectedinexperi- structuralfeaturescommontotheidentifiedsulfateesters; mentsinvolvingthehighestaerosolacidity.Allsulfateesters therefore, synthesis of more representative standards is listedinTable3elutedfromthereverse-phaseLCcolumn neededinordertoquantifysulfateestersby(-)LC/ESI-MS. within 3 min, indicating their high water solubility. For Sulfate ester aerosol was atomized from a standard example,them/z215sulfateesterhadaretentiontimeof solutionofsodiumpropylsulfateandanalyzeddirectlyby (cid:24)1.4min,closetothatoftheinorganicsulfate(firstpeakto the PILS/IC technique; no significant levels of inorganic elute).Ontheotherhand,them/z260sulfateesterwasfound sulfate were detected, suggesting that organosulfates are tobeslightlylesspolarwithisomerselutingat2.4,2.7,and thermallystableattheoperatingconditionsofthisinstru- 2.9min.Thepresenceofanitrategroupwasconfirmedby ment.Inaddition,nochromatographicpeakintheICdata itseven-mass[M-H]-ionandtheobservationofa63Da couldbeattributedtothesulfateesterstandard.Theseresults (HNO)neutrallossshowninFigure1D.Sulfateesterproducts suggestthatthePILS/ICtechniquewillobservedecreasesin 3 given in Table 3 containing nitrate groups were detected inorganicsulfateifsulfateesterformationoccurs. onlyinexperimentscontainingNO. ThetimeevolutionoftheSO2-andNH+aerosolmass x 4 4 Inseveralpreviousstudies,thepresenceofacidicsulfate concentrationsobtainedusingthePILS/ICtechniquefora aerosolwasfoundtohaveapronouncedeffectonthequantity Caltechlow-NO isopreneAASseedaerosolexperimentis x ofSOAformedbythephotooxidationofisoprene(9,14,15). compared to that of a Caltech high-NO AS seed aerosol x Inthepresentstudy,itappearsthatsulfateesterformation experimentinFigure2.AsshowninFigure2A(althoughnot maybesimilarlyenhancedbytheintroductionofanacidic evidentfromthetimescalepresented,SO2-andNH+decay 4 4 sulfateaerosol.Intheabsenceofsignificantlevelsofinorganic by(cid:24)20and14%,respectively,over9h),atypicalprofilefor sulfate, no sulfate esters were detected by LC/ESI-MS in mostCaltechisopreneexperimentsinTable1,ammonium anyoftheisoprenesystemsconsideredhere.Whenexperi- andsulfatetypicallydecreasedslowlywithtimeduetowall- mentswerecarriedoutinthepresenceofASaerosols,afew lossprocesses.However,intheexperimentshowninFigure sulfateestersweredetected,including[M-H]-ionsatm/z 2B,inwhichsulfateesterconcentrationswereexceedingly 199, 215, 244, 260, or 333. Experiments carried out under high,theSO2-aerosolmassconcentrationdecreasedmuch 4 acidicconditionswithAASaerosolproducedaconsiderably faster(i.e.,SO2-decayedby(cid:24)60%over6h)thanwallloss, 4 wider array of detectable sulfate ester compounds. In suggestinganextralossprocess,mostlikelyduetochemical addition,thepeakareasofseveralionsobservedinboththe reaction.ItshouldbenotedthattheinitialNH:SO molar 4 4 ASandAASexperimentswerefoundtobelargerintheAAS ratioinFigure2Awasnotexactlytwoduetoaknownsource VOL.41,NO.2,2007/ENVIRONMENTALSCIENCE&TECHNOLOGY9521 354 FIGURE3. (-)ESI-ITMSproductionmassspectraforsulfateestersofr-pineneoxidationproducts.(A)Productionmassspectrumfor m/z294detectedinEPA-211stage5.(B)Productionmassspectrumform/z265detectedinEPA-211stage4.(C)Productionmassspectrum form/z310detectedinEPA-211stage5.(D)Productionmassspectrumform/z326detectedinEPA-211stage5.Thesesulfateesterswere alwayspresentwhenr-pinenewasphotooxidizedinthepresenceofSO. 2 ofammoniumvolatilizationpreviouslycharacterized(31). thussupportingthatthe[M-H]-ionsatm/z294,310,and The significant decrease in the SO2- mass concentration 326containasulfategroup.Theseresultswereconfirmed 4 observedfortheCaltechlow-NO AASseedaerosolexperi- ontheLC/MSinstrumentoperatedintheupfrontCIDmode x mentisconsistentwithpreviouslyobservedincreasesinSOA ofanalysis. yields (9), strongly suggesting that particle-phase sulfate SulfateestersfromR-pinenewerefoundtoelutefromthe esterificationisatleastpartlyresponsibleforthis“acid-effect.” reverse-phaseLCcolumnatmuchlaterRTs((cid:24)10-26min) FiltersamplingformostCaltechisopreneexperimentsbegan thanthoseformedinisopreneoxidation,indicatingdiffer- (cid:24)5-7haftertheexperimentwasinitiated;sulfateestersare encesinwatersolubility.IdentifiedR-pinenesulfateesters formed by this point in the experiments as shown in listed in Table 4 were formed from the reactive uptake of Figure2. previouslyidentifiedgas-phaseoxidationproducts(33,34), Sulfate Esters from r-Pinene Oxidation. As in the consistentwithpreviouswork(17,18).Exceptforpinonal- isopreneexperiments,R-pinenesulfateesterswerefoundto dehyde,nosulfateestershavebeenidentifiedtoformfrom produceabundant[M-H]-ions,corresponding34Sisotopic previously identified R-pinene SOA products; however, ions,m/z97and80productions,andwerenotobservedin furtherinvestigationiswarranted.Forqualitycontrolpur- experimentswithoutSO (Table2).Isoprenesulfateesters poses,solid-phaseextraction(SPE)wasusedonduplicate 2 wereformedalsoinexperimentsinvolvingisopreneandSO filters collected from selected experiments (EPA-211) to 2 photooxidation;forsimplicity,theseestersarenotlistedin remove excess inorganic sulfate; it was found that the Table2.ShowninFigure3A-Darethe(-)ESI-ITMSproduct R-pinenesulfateesterswerestilldetected,andinsomecases ionspectraforrepresentativeR-pinenesulfateesters;these at higher [M-H]- ion abundances, indicating that these includeesterscontaininga[M-H]-ionatm/z294,265,310, sulfateestersarenotaresultofinorganicsulfateclustersin and 326, respectively. Analogous to some of the isoprene themassspectrometer. sulfate esters, the m/z 294, 310, and 326 R-pinene sulfate SulfateEstersinAmbientAerosol.Figures4A-Ccompare estersalsocontainnitrategroupsassuggestedbytheireven- theLC/MSextractedionchromatograms(EICs)ofm/z215 mass[M-H]-ionsandobservedneutrallossesof63(HNO) obtainedfromaCaltechlow-NO isopreneAASexperiment 3 x and/or47Da(HONO).TheMS/MSspectraofthe[M-H]- to that of two SEARCH field samples (JST and BHM, ions for the four R-pinene sulfate esters yielded m/z 97 respectively). Both the RTs and mass spectra of the chro- productions.However,them/z80productionwasobserved matographic peaks shown in the EICs of m/z 215 are the only for the m/z 265 ion due to mass range limits on the sameinallsamples,stronglysuggestingthatthisisoprene massspectrometer.ItshouldbenotedthattheMS3spectra sulfate ester is present in ambient aerosol. In addition, ofhigh-massproductionsshowninFigure3(e.g.,m/z250 ambient aerosol recently collected at K-puszta, Hungary inFigure1A)didyieldthem/z97andm/z80productions, indicatesthatthem/z199and260isoprenesulfateesters 5229ENVIRONMENTALSCIENCE&TECHNOLOGY/VOL.41,NO.2,2007 355 TABLE 4.Proposed r-Pinene Sulfate Ester SOA Products aPositionalisomerscontainingnitrateorsulfategroupsatotherhydroxylatedpositionsarepossible.bCompoundslistedinparenthesesare neutrallossesobserveduponESI-MS/MS.cPreviouslydetectedR-pineneoxidationproductbyAschmannetal.(33,34).dProposedbyLiggioet al.(17)toformfrompinonaldehydereactiveuptakeontoacidicseedparticles;however,structurewasnotconfirmed.Inthecurrentstudy,we confirmitsstructurewith(-)ESI-MS.eDetectedinambientaerosolcollectedfromSEARCHnetworkJune2004forfirsttime.fESI-MScannot differentiatebetweenwhichproductisbeingdetected;forcompletenessbothstructuresareshownhere.gObservedinanambientstudybyGao etal.(22).hNostructuralinformationwasprovidedinthepreviousstudybyGaoetal.(22). arepresent(M.Claeys,unpublishedresults);however,these to the lack of detailed connectivity of specific functional compoundsareonlyweaklydetectedonsomedaysanalyzed groups(e.g.,sulfateestersandhydroxyls)providedbyESI- fromtheSEARCHnetwork. MS/MSmethods. Figure 5A-C compares the LC/MS EICs of m/z 294 Theresultsabovesuggestthatthechemistryoccurringin obtainedfromtwoR-pineneexperiments(EPA-205andEPA- ourlaboratoryexperimentsarerelevanttotheconditionsin 326,respectively)andwithaSEARCHfieldsamplecollected thesoutheasternU.S.,eventhoughthelaboratoryaerosol attheBHMfieldsiteinJune2004.Thisfigureindicatesthat wasgeneratedfrommuchhigherVOCmixingratios,lower theR-pinenem/z294sulfateesterisaconstituentofambient RHs, and likely higher aerosol acidities observed in the aerosol,consistentwithpreviouswork(22).OtherR-pinene southeasternU.S. sulfateestersidentifiedinthisstudyhavebeenobservedin ESI-MSQualityControlTests.Toensurethatthesulfate ambientaerosolinthesoutheasternU.S(22).Itispossible esterselucidatedinthisstudywereformedonlyduringSOA thatthem/z294sulfateesterintheambientaerosolcould formationandnotonthefilterorduringtheESIprocess, alsoresultfromtheoxidationofothermonoterpenesowing several quality control tests were conducted. First, a filter VOL.41,NO.2,2007/ENVIRONMENTALSCIENCE&TECHNOLOGY9523 356 FIGURE4. (-)LC/ESI-MSextractedionchromatogramsform/z215.Theretentiontimesofthem/z215EICsarethesameaswellasthe massspectraassociatedwitheachchromatographicpeak;therefore,thecomparisonoftheseEICssuggeststhatthephotooxidationof isopreneinthepresenceofacidseedproducesthesesulfateestersobservedintheambientaerosol.Inallchamberexperimentsinvolving isopreneinthepresenceofASseedaerosol,AASseedaerosol,orSO,them/z215ionwasdetected. 2 extractfromaCaltechlow-NO isoprenenucleation(i.e.,no aldehydes(e.g.,formaldehyde)intheuppertroposphereand x inorganicseedaerosolpresent)experimentwasdividedinto lower stratosphere have been suggested to occur in the twoparts.Onepartwasspikedwithahighconcentrationof presenceofsulfateaerosols(35-37),whereincreasedacidity (NH)SO andtheotherpartwasspikedwithpureHSO. wasfoundtoincreasetheiruptake.Someofthesestudies 42 4 2 4 Bothofthesesampleswereanalyzedby(-)LC/ESI-MS.The proposed that the observed uptake of the alcohols and sulfateesterslistedinTable1werenotdetectedinthesetwo aldehydeslikelyoccurredbysulfateesterformation,although testsamples,demonstratingthatthesulfateestersdetected noproductstudieswereconducted.Ithasalsobeenshown inthisstudyarelikelynotartifactsformedintheESIinterface. that reactive uptake of butanol and ethanol onto sulfate The use of reverse phase chromatography allowed for aerosolsoccursatroomtemperature(38,39).Esterification inorganic sulfate not to be confused with organosulfates, was recently shown to occur from the photooxidation of whereinorganicsulfatewastheveryfirstpeaktoelutefrom isopreneinthepresenceofNO fromcondensationreactions x the column. In the two test samples discussed above, the oforganicacidswithalcohols(9,10).Thelargeamountsof inorganicsulfatepeakwasfoundtohavesometailing,which organic acids formed during the photooxidation were wasverysimilartotheseededexperimentslistedinTable proposedtodrivethesereactions. 1; however, this tailing seems to have no effect on the Figure 6 shows the general reactions proposed for the formationofsulfateesters. formationofsulfateestersfromalcohols(2-methyltetrolused As a second test, a meso-erythritol (a surrogate for the asmodelcompound)andsulfatederivativesfromcarbonyl 2-methyltetrolsproducedfromisopreneoxidation)standard compounds(pinonaldehydeusedasmodelcompound).In wasdividedintotwoparts,whereonewasspikedwith(NH4)2- thecaseofsulfateesterformationfromalcohols,theproposed SO4andtheotherwithpureH2SO4.Asforthefirstquality reactionslikelyinvolvenucleophilicsubstitution(SN1),where controltestabove,thesetwosamplesproducednosulfate thesulfuricacidprotonatesthealcoholicgroup,makingwater estersin(-)LC/ESI-MS. the leaving group. The resulting carbocation becomes a Last, a filter extract from an EPA R-pinene experiment nucleophilicsitefortheunsharedpairofelectronsonone conductedwithoutSO2(thusnosulfateaerosolpresent)was oftheoxygenatomsofthesulfate(40).Duetothelowrelative spiked with pure H2SO4. Again, no sulfate esters were humiditiesoftheseexperiments,thislikelyshiftstheequi- detected.Theseresultsstronglysuggesttheorganosulfates librium in favor of sulfate ester formation. In the case of (sulfateesters)wereformedintheaerosolphaseandarenot sulfateesterformationfromaldehydes,theproposedreaction anartifactofsamplingormeasurement. likely involves the electron pair of the carbonyl oxygen acceptingaprotonfromsulfuricacid,producingtheoxonium Discussion ion,andmakingitmoresusceptibletonucleophilicattack SulfateEsterificationReactionMechanism.Reactiveuptake fromanunsharedpairofelectronsfromoneoftheoxygen of gas-phase alcohols (e.g., methanol and ethanol) and atomsonsulfate.Itshouldbestressedthatotherreaction 5249ENVIRONMENTALSCIENCE&TECHNOLOGY/VOL.41,NO.2,2007 357 FIGURE5. (-)LC/ESI-MSextractedionchromatogramsform/z294.Theretentiontimesofthem/z294compoundswerethesameaswell asthemassspectraassociatedwitheachchromatographicpeak;therefore,thecomparisonoftheseEICssuggeststhatthephotooxidation ofr-pineneinthepresenceofNO andacidseedproducesthesesulfateestersinambientaerosol.Nom/z294compoundsweredetected x inexperimentsinvolvingonlyisopreneandacidseed(orSO). 2 FIGURE6. Proposedreactionsfortheformationofsulfateestersfrom2-methyltetrolandpinonaldehyde,arepresentativealcoholand aldehydegeneratedbythephotooxidationofisopreneandr-pinene,respectively.Solidboxesindicate(-)ESI-MSdetectedspecies. Dashedboxesindicateotherproposedproductspossiblyformed. mechanisms are also possible, including: electrophilic place in the case of polyols; however, (-)ESI-MS is not additionofHSO toanaliphaticdoublebond,additionof sensitivetosuchneutralspecies,nosuchproductshaveyet 2 4 SO throughsomeradicalprocess,orbysomeotherunknown been identified. In prior work (9) we reported oligomeric 3 mechanismcurrentlynotunderstood.Sulfatediester(ROSOR) signatures(14,16,18Dadifferences)andcompoundswith 3 formationandsulfateesteroligomerizationcouldalsotake masses up to(cid:24)620 Da in matrix-assisted laser desorption VOL.41,NO.2,2007/ENVIRONMENTALSCIENCE&TECHNOLOGY9525