XXXXporntime

δ13CStableIsotopeAnalysisofAtmosphericOxygenatedVolatileOrganicCompoundsbyGasChromatography-IsotopeRatioMassSpectrometryBrianM.
Giebel,*PeterK.
Swart,andDanielD.
RiemerUniversityofMiami,RosenstielSchoolofMarineandAtmosphericScience,4600RickenbackerCauseway,Miami,Florida33149Wepresentanewmethodforanalyzingtheδ13Cisotopiccompositionofseveraloxygenatedvolatileorganiccompounds(OVOCs)fromdirectsourcesandambientatmosphericsamples.
Guidedbytherequirementsforanalysisoftracecomponentsinair,agaschromato-graphisotoperatiomassspectrometer(GC-IRMS)systemwasdevelopedwiththegoalofincreasingsensitivity,reducingdead-volumeandpeakbandbroad-ening,optimizingcombustionandwaterremoval,anddecreasingthesplitratiototheisotoperatiomassspectrometer(IRMS).
Thetechniquereliesonatwo-stagepreconcentrationsystem,alow-volumecapillaryreactorandwatertrap,andabalancedreferencegasdeliverysystem.
Theinstrument'smeasurementpreci-sionis0.
6to2.
9‰(1σ),andresultsindicatethatnegligiblesamplefractionationoccursduringgassam-pling.
Measuredδ13Cvalueshaveaminordependenceonsamplesize;linearityforacetonewas0.
06‰ngC-1andwasbestover1-10ngC.
Sensitivityis10timesgreaterthansimilarinstrumentationdesigns,incor-poratestheuseofadilutedworkingreferencegas(0.
1%CO2),andrequirescollectionof>0.
7ngCtoproduceaccurateandpreciseresults.
Withthisdetec-tionlimit,a1.
0Lsampleofambientairprovidessufcientcarbonforisotopicanalysis.
Emissionsfromvegetationandvehicleexhaustarecomparedandshowcleardifferencesinisotopicsignatures.
AmbientsamplescollectedinmetropolitanMiamiandtheEvergladesNationalParkcanbedifferentiatedandreectmultiplesourcesandsinksaffectingasinglesamplinglocation.
Vehicleexhaustemissionsofetha-nol,andthosecollectedinmetropolitanMiami,haveanomalouslyenrichedδ13Cvaluesrangingfrom-5.
0to-17.
2‰;weattributethisresulttoethanol'soriginfromcornanduseasanadditiveinautomotivefuels.
Oxygenatedvolatileorganiccompounds(OVOCs)suchasmethanol,ethanol,acetaldehyde,andacetonearegasesfoundthroughoutthetropospherethatinuenceatmosphericchemistryinmanyways.
Thesecompoundsactasasourceofradicalsandasinkforthehydroxylradical(OH),participateintroposphericozoneformation,andareprecursorstoformaldehydeandCO.
1MixingratiosforOVOCsaretypicallyatthelowpartsperbillionbyvolume(ppbv)levelanddependonsamplinglocationandseason.
2-5MostatmosphericOVOCmeasurementshavereportedinformationonambientlevels,sourceemissionstrengths,anduxrates,2,3,5-9withtwoindividualOVOCs,methanolandacetone,receivingthemajorityoffocusthusfar.
Methanolisthesecondmostabundantorganicgasintheatmosphereaftermethaneanditsglobalbudgethasbeenstudiedextensively.
3,10-12Emissionsfromvegetationarethesinglelargestsourcetotheatmosphereandareestimatedbetween75-312Tgyear-1.
Othersourcesofmethanolexist,includingfossilfuelcombustion,biomassburning,plantdecay,andinsituatmo-sphericproductionviaoxidationofmethane.
Combined,thesesourcesareestimatedat2‰wereobserved.
Thistestwasperformeddailytoevaluatetheintegrityoftheinstrumentation.
Iftheoffsetswere0.
3-0.
5‰,theconcur-rentdataforthedaywerecorrectedbytheappropriateamount.
Iftheoffsetexceeded0.
5‰,thecapillaryreactorwasreplaced.
DynamicRangeandLinearity.
A6LelectropolishedstainlesssteelbulbwithadiptubeassemblyservedasanexponentialdilutionasktotestthedynamicrangeandlinearityofthemethodandIRMSintheabsenceofcombustion.
Thebulbcontaineda1%CO2mixtureinheliummadefromthesamesubsampledCO2usedintheproductionofworkingreferencegas.
Adiluantowofheliumenteredthesteelbulbthroughthediptubeatarateof125cm3min-1andtheoutowwasplumbedtoRV1.
Thetotalanalysisoccurredovera5.
5hperiodbrokeninto9segmentswiththeintroductionofworkingreferencegas.
Theamountofcarbonreachingtheionsourcewas0.
1-80ng.
Theδ13Cvaluesoverthisrangeareexpressedasadifferenceofthemeasured(andcorrected)exponentiallydilutedCO2fromtheworkingreferencegas'acceptedvalueandaredisplayedinFigure3a.
Ofparticularinterestistheappearanceofapositiveoffset,0.
46‰,fromzero.
Effortsweremadetominimizefractionationsduringthegastransfers,andthegasesweremadefromthesamestock.
Despitethiseffort,theoffsetstillpersisted.
(43)Apel,E.
C.
;Emmons,L.
K.
;Karl,T.
;Flocke,F.
;Hills,A.
J.
;Madronich,S.
;Lee-Taylor,J.
;Fried,A.
;Weibring,P.
;Walega,J.
;Richter,D.
;Tie,X.
;Mauldin,L.
;Campos,T.
;Weinheimer,A.
;Knapp,D.
;Sive,B.
;Kleinman,L.
;Springston,S.
;Zaveri,R.
;Ortega,J.
;Voss,P.
;Blake,D.
;Baker,A.
;Warneke,C.
;Welsh-Bon,D.
;deGouw,J.
;Zheng,J.
;Zhang,R.
;Rudolph,J.
;Junkermann,W.
;Riemer,D.
D.
Atmos.
Chem.
Phys.
2010,10,2353–2375.
Figure2.
RangeofatmosphericOVOCmixingratiosexpectedinlargeportionsofthetroposphereandtheirrelationtocarbontransmittedtotheIRMSionsource.
Theshadedareacorrespondstothemethoddetectionlimit.
EAnalyticalChemistry,Vol.
xxx,No.
xx,MonthXX,XXXXThepositiveoffsetforCO2waslikelyaresultofsmallamountsofambientCO2becomingentrainedinthesystemandtherebyenrichingthemeasuredδ13Cvalue.
Forlargesamplesizesthisappearedtohaveaminimaleffect.
However,theeffectbecamemagniedforsamplesizesbelow1ngC,wheredeviationsupto2‰areobserved.
Thelinearityoverthisrange,determinedbyordinarylinearregression,was0.
01‰ngC-1.
Thissuggeststhatvariationoftheδ13Csignaturewithsamplesizeisnegligibleandrequires10ngCtoinduceachangeof0.
1‰.
Accuracyandprecisionwerebestovertherangeof0.
2to20ngC.
Samplescontainingmorethan20ngCandlessthat0.
2ngCwereenrichedin13C.
LiquidCompoundsandSingle-ComponentGases.
Low-pressuresingle-componentgassamplesenteredRV1atarateof3cm3min-1andwereloopinjectedontotheGC-IRMSwithnocryo-focusingbeforethechromatographiccolumn.
Teninjec-tionsofeachgaswerecomparedtosixinjectionsofworkingreferencegas.
Single-componentgasestestedtheGC-IRMSinstrumentationbycomparisontotheisotopicvaluesobtainedfortherawliquidsontheelementalanalyzer(ANCA).
Thevaluesfortherawliquidsservedasthebasisofallourcomparisons.
Thecalculatedpercentdifferencebetweenthetwomeasurementsrangedbetween-0.
1and4.
8%,andtheresultsarelistedinTable1.
AcetonewassuitabletotestthedynamicrangeandlinearresponseoftheIRMSwiththeaddedstepofcombustion.
Acetonewaschosenastheanalytebecauseitshowedconsistentandexcellentreproducibilityacrossallaspectsofthisstudy.
TheexperimentaldesignwasidenticaltothatoftheCO2experimentdescribedpreviouslybutwithoneexception,theacetonemixturehadastartingcarbonequivalentof0.
1%beforethediluentowofheliumwasadded.
A1%mixtureofacetonewasFigure3.
(A,B)Resultsforexponentiallydiluted(A)CO2and(B)acetonesamples.
δ13CvaluesareexpressedasadifferenceofthedilutedCO2andacetonefromtheacceptedvalueofthe0.
1%workingreferencegasandacetonevalueobtainedontheelementalanalyzer.
Anoffset,oppositeinsignbutalmostequalmagnitude,existsforCO2(0.
5‰)andacetone(-0.
6‰).
ForCO2,thisisthoughttobetheresultofanambientleakwherebyatmosphericCO2entersthesystem.
Foracetone,incompletecombustionwithinthecapillaryreactormaycontributetotheobservednegativeoffset.
Table1.
Tabulatedδ13CValuesforOVOCsUsedinThisWorkaelementalanalyzerliquidcompoundsGC-IRMSsingle-componentgasGC-IRMSlow-pressure7-componentGC-IRMShigh-pressure7-componentpooledGC-IRMSmean±1σ(‰)mean±1σ(‰)95%condence(±‰)%errormean±1σ(‰)95%condence(±‰)%errormean±1σ(‰)95%condence(±‰)%errormean±1σ(‰)95%condence(±‰)%errormethanol–35.
3±0.
1–34.
1±0.
10.
13.
5–33.
0±0.
10.
16.
5–34.
4±2.
82.
12.
5–34.
0±1.
60.
73.
7ethanol–29.
2±0.
2–27.
8±0.
80.
64.
8–26.
2±0.
81.
310.
2–26.
5±0.
50.
49.
2–26.
6±1.
40.
68.
7propanal–32.
8±0.
1–31.
9±0.
20.
12.
7–35.
0±0.
61.
0–6.
9–27.
4±0.
80.
616.
5–30.
5±2.
91.
26.
9acetone–27.
5±0.
2–28.
5±0.
70.
5–3.
7–27.
9±0.
20.
3–1.
4–27.
6±0.
20.
1–0.
4–28.
0±0.
60.
2–1.
7MEK–23.
2±0.
2–23.
5±0.
90.
6–1.
3–25.
5±0.
71.
1–10.
1–22.
5±0.
30.
2–3.
0–23.
5±1.
20.
5–1.
22-pentanone–25.
0±0.
3–26.
1±0.
20.
1–4.
3–33.
7±1.
11.
8–34.
8–28.
4±0.
60.
5–13.
6–28.
6±2.
81.
1–14.
33-pentanone–30.
7±0.
2–30.
7±0.
60.
4–0.
1–34.
3±0.
30.
5–11.
6–32.
7±0.
70.
5–6.
5–32.
3±1.
50.
6–5.
1samplenumbern)3n)10n)4n)9n)23InstrumentVariablesdiluentHeHeN2sampleloopRV1RV1RV2cryo-focusnoyesyescarbonsorbentnonoyeszeroairdilutionnonoyesaAccuracyandprecisionistracedfromtheelementalanalyzerthroughthenaldesignoftheGC-IRMSsystem.
Differentvariablestestedduringeachphasearelisted.
Thesystem'stotalprecisionwascalculatedbetween0.
6and2.
9‰whencomparedtothevaluesobtainedontheelementalanalyzer.
FAnalyticalChemistry,Vol.
xxx,No.
xx,MonthXX,XXXXavoidedfortworeasons.
First,ambientsamplesarenotexpectedtobegreaterthan0.
1%,andsecond,thereismoreconcernforwhathappenstomeasuredisotopicsignaturesassmallersampleconcentrationsareapproached.
TheamountofcarbonreachingtheionsourcewasdeterminedsimilarlytotheCO2testandrangedbetween0.
8and12ng.
Theδ13Cvaluesoverthisrangeareexpressedasadifferenceofthemeasuredandcorrectedexponentiallydilutedacetonefromthevalueobtainedontheelementalanalyzer(Figure3b).
Thelinearityoverthisrange,determinedbyordinarylinearregression,was0.
06‰ngC-1.
Foracetone,thisindicatesthatsamplesizecaninuencemeasuredδ13Candthatachangebetween1and10ngCcaninduceanoticeableshiftof0.
6‰inmeasuredδ13C.
Accuracyandprecisionwerebestovertherangeof0.
2-10ngC.
Alsoworthnotingistheapparentnegativeoffsetforacetone,-0.
56‰,comparedtothepositiveoffsetforCO2,0.
46‰.
Rawdataforbothexperimentswerecorrectedby0.
5‰and0.
4‰foracetoneandCO2,respectively.
However,theoffsetsstillexist.
EntrainmentofambientCO2didnotappeartoaffectacetonebecauseofitsseparationonthechromatographiccolumn.
Thenegativeoffsetforacetonewaslikelyrelatedtoincompletecombustionwithinthecapillaryreactor.
Thermo-dynamicprinciplessupport12Cbeingcombustedbefore13C;thus,ifcombustionwasincomplete,wewouldobservealighterδ13Cvalue.
CalibrantGasAnalyses.
Low-PressureSeven-ComponentGasMixture.
Alow-pressureseven-componentgasmixtureinheliumwasusedpreliminarilytotestchromatographicconditionsintheabsenceofthecarbonsorbentbyusingtheRV1loop(Table1).
Thiswasalogicalstepbetweentheuseofsingle-componentgasesandagravimetricallyprepared,high-pressure,seven-componentcalibrationgasinnitrogen.
Thelow-pressureseven-componentgasmixtureowedthroughtheRV1loopfor5minpriortostartingtheanalysis.
Theowrate(3cm3min-1)wasmaintainedbyMFC(no.
1)upstreamofRV1.
Aftertheinitial5minpurgeperiod,RV1wasmanuallyswitchedandthegaswithintheinjectionloopwasdivertedthroughRV2andcryogenicallyfocusedinliquidnitrogenforanadditional5minbeforeinjectionintothechromatographiccolumn.
Ofparticularnotearethevaluesobtainedfor2-and3-pentanone,whicharedepletedin13Ccomparedtoboththeliquidcompoundsandthesinglecomponentgasmixtures.
Thismayindicateanunknowneffectresultingfromtheanalyticalcolumn.
Thepercenterrorbetweenthismeasurementtechniqueandthatperformedontheelementalanalyzerforthepureliquidcompoundsrangesbetween1.
4and35%.
GravimetricSeven-ComponentGasMixture.
OneofthemaingoalsofthisworkwastodevelopaGC-IRMSsystemcapableofmeasuringOVOCsoverthedynamicrangefoundintheatmo-sphere.
Tomimicambientlevelsofthesecompoundsintheatmosphere,thehigh-pressurecalibrantgaswasdilutedintomoistzero-airusingadynamicdilutionsystem.
Dilutionproducedmixingratiosbetween18.
6ppbv(methanol)and7.
3ppbv(2-pentanone)forallcomponents.
Thedilutedcalibrantwascon-necteddirectlytothegasmanifold(Figure1).
Usingtherangeofmixingratiosproducedafterthehigh-pressurecalibrantgaswasdilutedinzero-air(7.
3-18.
6ppbv),thevolumeofairconcentrated(1.
0L),andtheopensplitdilution(30%),wecalculated2.
5-5ngCweredeliveredtotheionsourceforallcomponents.
ResultsforninereplicateanalysesarepresentedinTable1,andanexampleofthechromatographicresponseappearsinFigure4.
Reasonableagreementexistsforallsevencomponentscomparedtotheliquidreagentsanalyzedontheelementalanalyzer;themarginoferrorbetweenthesetwomeasurementsrangedbetween0.
4and16.
5%.
Thecomponentswiththetwolargesterrorswerepropanal(16.
5%)and2-pentanone(13.
6%).
Bothofthesepeaksaretheleadingpeakinapair(propanal/acetoneand2-penatnone/3-penatanone),andperhapsthelaterelutingcompoundsinuencethemeasuredδ13Cvaluesoftheearliercompounds.
Thisissupportedbytheobservationthatanalysisofthesingle-componentgasesforthesamecompoundsontheGC-IRMShadalowererror(25%ofU.
S.
cornproductionandthatethanolconstitutes99%ofallbiofuelsintheUnitedStates.
47,48UtilizingC4photosynthesis,whichdiscriminateslessagainst13C,cornandotherC4plantsaregenerallyenrichedintheisotopecomparedtoC3plants.
Bulkcarbohydrateanalysesbetweenthetwoplanttypesshowanenrichmentof15‰incarbohydratesextractedfromC4plantmaterial.
49Investigationsofindustriallyproducedethanoloriginatingfromcornhavebeenshowntohaveδ13Cvaluesof-10.
71±0.
31‰.
49ThevaluesweobservedintheScoutsamplesare5‰heavierand,consideringthewidespreaduseofethanol(7.
5billiongallonsareexpectedtobeusedinfuelby201248),mayserveasatracerfortransportationrelatedsourcestotheatmosphere.
Somebiogenicsamplesinthisstudy,suchassandliveoakandorangecitrus,hadsubstantiallydepletedvaluesformethanolandagreewithincubatedemissionsfromvariousdeciduoustreesandgrassesmadebyKeppleretal.
33(Table2).
However,thisobservationisnotconsistentacrossallsamplesandsuggestthatvariationsinδ13Cvaluesmayresultfrominterspeciesdiffer-ences,microbeinteractionontheleaf'ssurface,prey/injuryresponse,thepotentialpresenceofamethanolutilizationpathwaywhichoxidizesmethanoltoformaldehydeandformicacid/formate,50,51andotherlesserknownmetabolic,formation,andlosspathwayswithinplants.
52Finally,awoundresponsemaybeobservedbetweentheclippedandintactphilodendronandseagrapesamples.
Inonedistinctcase,acetaldehydeemittedfromclippedseagrapespecimenswereenrichedby4‰comparedtothefossilfuelemissions.
AmbientMeasurementResults.
Considerabledifferencesinδ13Careobservedbetweenambientsamplinglocations(Table3).
ResultsfromMiamiInternationalAirportarereectiveofanaveragedvalueforfreshvehicularsources.
Themeasuredδ13Crangeforairportsamplesisbetween-12.
3±3.
7‰(ethanol)to-35.
3±1.
7‰(3-pentanone).
Withtheexceptionofethanol,whichhasaδ13CvalueconsistentwithitsC4plantsource,and2-and3-pentanone,themeasuredrangeattheairportagreeswiththatestablishedforNMHCsfromtrans-portation-relatedsourcesbyRudolph,namely,-21.
9to-31.
3‰.
25Ouracetaldehydevalueisconsistentwiththerange(44)Iannone,R.
;Koppmann,R.
;Rudolph,J.
J.
Atmos.
Chem.
2007,58,181–202.
(45)Rudolph,J.
;Anderson,R.
S.
;Czapiewski,K.
V.
;Czuba,E.
;Ernst,D.
;Gillespie,T.
;Huang,L.
;Rigby,C.
;Thompson,A.
E.
J.
Atmos.
Chem.
2003,44,39–55.
(46)Rudolph,J.
;Czuba,E.
;Norman,A.
L.
;Huang,L.
;Ernst,D.
Atmos.
Environ.
2002,36,1173–1181.
(47)Barnett,M.
O.
Environ.
Sci.
Technol.
2010,44,5330-5331.
(48)Farrell,A.
E.
;Plevin,R.
J.
;Turner,B.
T.
;Jones,A.
D.
;O'Hare,M.
;Kammen,D.
M.
Science2006,311,506–508.
(49)Ishida-Fujii,K.
;Goto,S.
;Uemura,R.
;Yamada,K.
;Sato,M.
;Yoshida,N.
Biosci.
,Biotechnol.
,Biochem.
2005,69,2193–2199.
(50)Cossins,E.
A.
Can.
J.
Biochem.
1964,42,1793–1802.
(51)Gout,E.
;Aubert,S.
;Bligny,R.
;Rebeille,F.
;Nonomura,A.
R.
;Benson,A.
A.
;Douce,R.
PlantPhysiol.
2000,123,287–296.
(52)Fall,R.
InReactiveHydrocarbonsintheAtmosphere;Hewitt,C.
N.
,Ed.
;AcademicPress:SanDiego,CA,1999;pp43-97.
Table2.
δ13CValuesforCompoundsEmittedfromVariousTropicalPlantsandaFossilFuelCombustionSourceaplanttypesandliveoakQuercusgeminataorangeCitrussinensislemonCitruslimonphilodendronPhilodendronselloumseagrapeCoccolobauviferaKeppleretal.
33fossilfuelcombustionacetaldehyde-29.
9(2.
3)-25.
7(0.
1)-22.
4(1.
4)*-17.
5(0.
5)-30.
7(1.
1)*-24.
9(2.
2)-20.
9(0.
4)methanol-41.
9(3.
1)-59.
7(2.
9)-37.
8(2.
6)*-27.
5(0.
5)-30.
7(1.
0)*-68.
2(11.
2)-16.
9(1.
3)ethanol-41.
5(0.
8)-37.
5(0.
3)-30.
6(0.
2)-36.
5(0.
2)*-29.
4(2.
6)-5.
0(0.
4)propanal-25.
6(2.
7)isopreneoffscaleb-26.
9(3.
7)-35.
2(3.
5)-33.
8(2.
6)-23.
0(2.
6)*-16.
7(1.
2)-32.
6(0.
9)*acetone-35.
7(4.
1)-37.
4(2.
4)-32.
8(1.
2)-38.
8(1.
1)-29.
3(1.
5)*-33.
8(0.
8)-31.
3(0.
8)*-28.
1(2.
5)-25.
6(0.
5)2-pentanone-35.
2(1.
4)benzene-26.
9(0.
3)toluene-27.
5(0.
6)aAlsoincludedarevaluesforpreppedandincubatedbiogenicsamplesfromKeppleretal.
.
33Allvaluesarereportedastheaverage(standarddeviation).
Allsamplesn)5,exceptthefossilfuelsourcewheren)3.
Allbiogenicsamplesarewounded/clippedbranches,exceptwherenoted(*),whichrepresentsanintactbranchonthesamplespecimen.
Thefossilfuelsourcewascollectedfroma1972ScoutInternationalwithnocatalyticconverterataconstantcruise.
bIsoprenewaspresent;however,itsaturatedthedetectorsandthesignalresponsewasoffscaleandtheδ13Cvaluecouldnotbecalculated.
HAnalyticalChemistry,Vol.
xxx,No.
xx,MonthXX,XXXXpresentedbyWenetal.
whomeasuredvaluesviaaderivati-zationprocedureof-21.
0‰and-29.
2‰forsamplescollectedatabusstationandpetrochemicalrenery,respec-tively.
36SomeobservationsatMiami'sFinancialDistrictarebetween2.
2and4.
4‰enrichedin13CcomparedtothesamecompoundsatMiamiInternationalAirport,andagainweobserveananomalouslyenrichedvalueforethanol(-17.
2±4.
1‰).
Samplesfromtheairportaregeneralδ13CvalueswecanexpectforOVOCsfromtransportationrelatedsourceswithoutadditionfromothersourcesandlossescausedbysolarradiationandreactionwithOH.
Miami'sFinancialDistrictislocatedwithin0.
1mileofBiscayneBayand1mileofthePortofMiamiandwasdominatedbyanonshorebreezeduringthesamplecollection.
Therefore,wecanexpectvaluesfromthenancialdistricttobeenrichedsincetheδ13Csignatureforeachcompoundwillreectacombinationofvehicular,biogenic,andpossiblymarinesourcesand,additionally,lossesattributabletoreactivitywithOHandphotolysis.
Isotopicvaluesforsamplesfromthenancialdistrictareboundwithinthereportedrangeof-15.
8to-37.
4‰forNMHCssampledatamoderatelypollutedwaterfrontinWellington,NewZealand.
25Incomparisonwithautomobileexhaust(Table2),themeanvaluesobservedatMiamiInternationalAirportandMiami'snancialdistrictaregenerallydepletedin13C.
Thetwomostobviousdifferencesamongthesesamplesthatmayinuencetheobservationsarethefuelsourceandthepresenceofacatalyticconverter.
Emissionscollectedattheairportareamixofrenedpetroleumanddiesel,whereastheScoutInterna-tionalwasfueledbyunleadedgasoline.
Furthermore,vehicleemissionsattheairportareassumedtobeproducedbyengineshavingacatalyticconverter.
However,theScoutlackedaconverter,andthespeedsoftheenginesproducingtheemissionswereverydifferent.
Trafcthroughtheairport'slowerroadwaymovedatanidlepaceandrarelyexceeded15mph.
TheScoutsampleswereobtainedwiththeengineundersignicantloadandataconstantrevolutionperminute(2000rpm)andcruisespeed(80kph).
Toourknowledge,nostudiesexistshowinghowthepresenceofacatalyticconverterorenginespeedmayinuencetheδ13Cofemittedhydrocarbons.
SamplesfromEvergladesNationalParkspannedalargerangefrom-19.
0to-36.
3‰.
MeasuredmethanolfromwithintheNationalParkwas-36.
3±3.
7‰,considerablydepletedandconsistentwithothervaluesobtainedinthetropicalplantenclosurestudies(i.
e.
,sandliveoakδ13CMethanol)-41.
9±3.
1‰)andwiththeresultspresentedearlierfromKeppleretal.
33Similarly,δ13CvaluesforisoprenereleasedfromC3plantsrangefrom-26to-29‰.
45IsoprenevaluesattheNationalParkarelighter(-30.
3±2.
1‰)thantherangepresentedbyRudolphetal.
However,whentheprecisionofthemeasure-mentisconsidered,theisoprenevaluesmeasuredfromtheEverglades'samplesoverlaptherangeobservedwiththatpreviouswork.
AcetoneandacetaldehydevaluesfromwithintheNationalParkaremoreenrichedthananticipated.
Themeanδ13Cvaluesforthesecompoundsare-23.
7‰and-19.
0‰,respectively.
Eachareenrichedapproximately7.
5‰comparedtosamplescollectedatMiamiInternationalAirportandarefairlyconsistentwithsamplesfromMiami'snancialdistrictandfossilfuelcombustion.
Whenestimatedatmosphericlifetimes(τ)areconsideredforthesecompoundsinthetroposphereforlossescausedbyreactivitywithOH(τacetoneOH)66days;τacetaldehydeOH)11h)andphotolysis(τacetonehν)38days;τacetaldehydehν)5days),theseobservationscanbeexplained,especiallyfortheenrichmentofacetaldehydeoveracetone(5‰).
Fewstudiesofambientδ13Cforacetaldehydeexist,29,36andonlyoneexistsforacetone.
37ForsamplescollectedwithinabiospherereserveinChina,Guoetal.
measuredacetaldehydevaluesbetween-31.
6and-34.
9‰.
Thesevaluesaredepletedin13Ccomparedtoourmeasurements.
However,theyreportweakphotolyticlossofformaldehydeinthesamestudy,andconsideringformaldehyde'slifetimeagainstphotolysisisshorter(4h)comparedtoacetaldehyde(5days),weassumethistobetrueforacetaldehydeatthesamelocation.
Guoetal.
usedaderivatizationmethodtocalculateδ13Cvaluesforacetonecollectedataforestedsite(-31‰)andatthetopofa10mbuildinginuencedbyvehicleemissions(-26‰).
TheacetonevaluesfromEvergladesNationalParkareenrichedby2-7‰comparedtothevaluespresentedbyGuoetal.
Isotopicvaluesobtainedfromtheforestmayreectthesignatureoffreshacetoneemissionsfrombiomass,whilevaluesforEvergladesNationalParksamplesmaybemorestronglyinuencedbyphotochemistry.
ThemeasuredvaluesforacetoneandacetaldehydefromwithintheEvergladesmayalsoindicatecontributionsfrominsituatmosphericproductionviaoxidationandphotolysisofhigherorderhydrocarbons.
Anexactassessmenttoseparatedirectemissionsfromphoto-chemicalproductionandlossisnotpossibleatthistimesincefractionationsassociatedwiththesepathwaysarenotknown.
CONCLUSIONSAnewmethodformeasuringδ13Cvaluesoflow-molecularweightOVOCsfromdirectsourcesandambientsampleswasdeveloped.
Themethodincorporatedacarbonsorbent,alow-volumecapillaryreactor,watertrap,andbalancedworkingreferencegasdeliverysystem.
Themethod'stotalprecisionrangedbetween0.
6and2.
9‰,andnegligiblesamplefraction-ationoccurredwhilesamplingandtrappinggases.
Furthertestingshowedthatmeasuredδ13CvalueshadlittledependenceTable3.
AmbientMeasurementResultsforSamplesCollectedfromMetropolitanMiamiandEvergladesNationalParkaδ13C±1σ(‰)MiamiInternationalAirportMiamiFinancialDistrictEvergladesNationalParkacetaldehyde-26.
7±0.
7-26.
8±1.
2-19.
0±2.
7methanol-36.
3±3.
7ethanol-12.
3±3.
7-17.
2±4.
1isoprene-30.
3±2.
1propanal-28.
4±1.
5-26.
2±2.
4acetone-31.
0±3.
5-26.
6±0.
4-23.
7±0.
4MEK-28.
3±2.
1-25.
9±1.
92-pentanone-34.
8±6.
5-29.
4±0.
13-pentanone-35.
3±1.
7-37.
8±1.
8toluene-33.
7±2.
0aMiamiInternationalAirport,n)5;Miaminancialdistrict,n)4;EvergladesNationalPark,n)3.
IAnalyticalChemistry,Vol.
xxx,No.
xx,MonthXX,XXXXonsamplesize(0.
06‰ngC1-),andlinearitywasbestovertherangeof1-10ngC.
Themethodwassensitive,requiring>0.
2ngCintotheionsourcetoproduceaccurateandpreciseresults.
Theanalysisofambientsamplesrequiredsmallsamplevolumes,with1.
0Lofgasprovidingsufcientcarbonforanalysis.
Cleardistinctionsinδ13Cwereobservedbetweenemissionsreleasedfromplantsandautomobiles.
Inparticular,ethanolemissionsfromautomotiveexhaustandmetropolitanMiamiweresignicantlyenrichedin13C.
Thisisrelatedtoethanol'sC4plantoriginanduseasafueladditive.
Ambientsamplescanbedifferentiated,butthevariationinδ13Cvalueswasnotasgreatasforthesourcesamples.
Ambientsamplessufferfromadditionalcomplexitywithmultiplesourcesandsinksaffectingsinglesamplinglocations.
Clearly,morestudiesofsourcesandambientsamplingarerequiredtodeneandcharacterizeOVOCsinthetropospherealongwithlaboratorystudiestodeterminethekineticisotopeeffectsassociatedwithOVOCs'insituproductionandlossfromreactionwithOHandphotolysis.
Asitstandsnow,thistechniquecanbeusedtodifferentiateOVOCsourcesandtoassessthecarbonisotopicvaluesforOVOCsinambientair.
Itshouldserveasausefulwaytoinvestigatetransformationsoforganicgasesintheatmosphere.
ACKNOWLEDGMENTWethankTomBrennaandHerbertTobiasforhelpfuldiscus-sionsindevelopingthismethodandRichIannoneforprovidingatemplateforrawdatacalculations.
WeacknowledgeJohnMakandZhihuiWangfortheworkingreferencegasinterlabcompari-son.
WeappreciatetheeffortsofKevinPolkandhis1972InternationalScout.
Finally,wegratefullyacknowledgethehelpfulcommentsmadebytwoanonymousreviewersandsupportprovidedbyNSFGrantNo.
0450939.
SUPPORTINGINFORMATIONAVAILABLEAdditionalinformationasnotedintext.
ThismaterialisavailablefreeofchargeviatheInternetathttp://pubs.
acs.
org.
ReceivedforreviewMarch23,2010.
AcceptedJuly6,2010.
AC1007442JAnalyticalChemistry,Vol.
xxx,No.
xx,MonthXX,XXXX