Citation

Material Information

Title:
Modeling atmospheric mineral aerosol chemistry to predict heterogeneous photooxidation of SO2
Creator:
Yu, Zechen
Publisher:
Atmospheric Chemistry and Physics
Publication Date:
Language:
English
Physical Description:
Journal Article

Notes

Abstract:
The photocatalytic ability of airborne mineral dust particles is known to heterogeneously promote SO2 oxidation, but prediction of this phenomenon is not fully taken into account by current models. In this study, the Atmospheric Mineral Aerosol Reaction (AMAR) model was developed to capture the influence of air-suspended mineral dust particles on sulfate formation in various environments. In the model, SO2 oxidation proceeds in three phases including the gas phase, the inorganic-salted aqueous phase (non-dust phase), and the dust phase. Dust chemistry is described as the absorption–desorption kinetics of SO2 and NOx (partitioning between the gas phase and the multilayer coated dust). The reaction of absorbed SO2 on dust particles occurs via two major paths: autoxidation of SO2 in open air and photocatalytic mechanisms under UV light. The kinetic mechanism of autoxidation was first leveraged using controlled indoor chamber data in the presence of Arizona Test Dust (ATD) particles without UV light, and then extended to photochemistry. With UV light, SO2 photooxidation was promoted by surface oxidants (OH radicals) that are generated via the photocatalysis of semiconducting metal oxides (electron–hole theory) of ATD particles. This photocatalytic rate constant was derived from the integration of the combinational product of the dust absorbance spectrum and wave-dependent actinic flux for the full range of wavelengths of the light source. The predicted concentrations of sulfate and nitrate using the AMAR model agreed well with outdoor chamber data that were produced under natural sunlight. For seven consecutive hours of photooxidation of SO2 in an outdoor chamber, dust chemistry at the low NOx level was attributed to 55 % of total sulfate (56 ppb SO2, 290 µg m−3 ATD, and NOx less than 5 ppb). At high NOx ( >  50 ppb of NOx with low hydrocarbons), sulfate formation was also greatly promoted by dust chemistry, but it was suppressed by the competition between NO2 and SO2, which both consume the dust-surface oxidants (OH radicals or ozone).
Acquisition:
Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Zechen Yu.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the submitter.

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Atmos.Chem.Phys.,17,1000110017,2017 https://doi.org/10.5194/acp-17-10001-2017 Authors2017.Thisworkisdistributedunder theCreativeCommonsAttribution3.0License. Modelingatmosphericmineralaerosolchemistrytopredict heterogeneousphotooxidationofSO 2 ZechenYu,MyoseonJang,andJiyeonPark DepartmentofEnvironmentalEngineeringSciences,EngineeringSchoolofSustainableInfrastructureandEnvironment, UniversityofFlorida,P.O.Box116450Gainesville,FL32611,USA Correspondenceto: MyoseonJangmjang@u.edu Received:8February2017Discussionstarted:9March2017 Revised:16July2017Accepted:17July2017Published:25August2017 Abstract. Thephotocatalyticabilityofairbornemineraldust particlesisknowntoheterogeneouslypromoteSO 2 oxidation,butpredictionofthisphenomenonisnotfullytakeninto accountbycurrentmodels.Inthisstudy,theAtmospheric MineralAerosolReactionAMARmodelwasdeveloped tocapturetheinuenceofair-suspendedmineraldustparticlesonsulfateformationinvariousenvironments.Inthe model,SO 2 oxidationproceedsinthreephasesincluding thegasphase,theinorganic-saltedaqueousphasenon-dust phase,andthedustphase.Dustchemistryisdescribedasthe absorptiondesorptionkineticsofSO 2 andNO x partitioning betweenthegasphaseandthemultilayercoateddust.The reactionofabsorbedSO 2 ondustparticlesoccursviatwo majorpaths:autoxidationofSO 2 inopenairandphotocatalyticmechanismsunderUVlight.Thekineticmechanism ofautoxidationwasrstleveragedusingcontrolledindoor chamberdatainthepresenceofArizonaTestDustATD particleswithoutUVlight,andthenextendedtophotochemistry.WithUVlight,SO 2 photooxidationwaspromotedby surfaceoxidantsOHradicalsthataregeneratedviathephotocatalysisofsemiconductingmetaloxideselectronhole theoryofATDparticles.Thisphotocatalyticrateconstant wasderivedfromtheintegrationofthecombinationalproductofthedustabsorbancespectrumandwave-dependentactinicuxforthefullrangeofwavelengthsofthelightsource. Thepredictedconcentrationsofsulfateandnitrateusingthe AMARmodelagreedwellwithoutdoorchamberdatathat wereproducedundernaturalsunlight.Forsevenconsecutive hoursofphotooxidationofSO 2 inanoutdoorchamber,dust chemistryatthelowNO x levelwasattributedto55%oftotalsulfateppbSO 2 ,290gm )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 ATD,andNO x lessthan 5ppb.AthighNO x > 50ppbofNO x withlowhydrocarbons,sulfateformationwasalsogreatlypromotedbydust chemistry,butitwassuppressedbythecompetitionbetween NO 2 andSO 2 ,whichbothconsumethedust-surfaceoxidants OHradicalsorozone. 1Introduction Thesurfaceofmineraldustparticlesisabletoactasasink forvariousatmospherictracegasessuchassulfurdioxide SO 2 / ,nitrogenoxidesNO x ,e.g.,NOandNO 2 / ,andozone O 3 / .Amongtracegases,SO 2 hasreceivedmuchattentionbecauseheterogeneousoxidationofSO 2 producesnonvolatilesulfuricacid,whichisreadilyinvolvedintheacidicationofparticlesorthereactionwithdustconstituentssuch asalkalinemetalsK C ,Na C / ormetaloxidese.g., -Al 2 O 3 andFe 2 O 3 / .Suchmodicationofthechemicalcomposition ofdustparticlescaninuencethehygroscopicpropertiesof mineraldust,whichisessentialtoactivatecloudcondensationnucleationKruegeretal.,2003;ZhangandChan,2002; Vlasenkoetal.,2006;Liuetal.,2008;Tangetal.,2016. Metaloxidese.g.,TiO 2 andAl 2 O 3 / havefrequentlybeen usedinmanylaboratoriestostudythekeyroleofmineral dustintheheterogeneousoxidationofSO 2 Goodmanetal., 2001;Usheretal.,2002;Zhangetal.,2006.However,these laboratorystudieshavebeenlimitedtoacertaintypeofmetal oxideandautoxidationofSO 2 withoutalightsource.To date,onlyafewstudieshaveattemptedtostudythephotocatalyticcharacteristicsofmineraldustintheoxidationofSO 2 andNO x .Forexample,asnotedbyParkandJang,the reactiveuptakecoefcient SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 / ofSO 2 inthepresenceof dryArizonaTestDustATDparticlesunderUVlightwas PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.

PAGE 2

10002Z.Yuetal.:AtmosphericMineralAerosolReactionmodel 1orderofmagnitudehigher.16 10 )]TJ/F100 7.5716 Tf 5.905 0 Td [(6 usinganindoor chamberwithalightmixofUV-AandUV-Blightthanthat fromautoxidation.15 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(7 / withoutalightsource.Usinganaerosolowtube,Dupartetal.observedthat theuptakerateofNO 2 byATDdustparticleswassignicantlyenhancedby4timesunderUV-Airradiationcomparedtodarkconditions.FieldobservationshavealsoreportedthepromotionofSO 2 photooxidationinthepresence ofmineraldust.Forinstance,nearBeijing,Chinagroundbasedcampaignin2009,andinLyon,Franceremotesensingcampaignin2010,Dupartetal.foundthat mineraldustwasasourceofOHradicalsunderUVradiation thatpromotedsulfateformation. Semiconductingmetaloxidese.g., -Al 2 O 3 -Fe 2 O 3 andTiO 2 / actasaphotocatalystinmineraldustparticles thatcanyieldelectron e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb / hole h C vb / pairs,andthatthey areinvolvedintheproductionofstrongoxidizers,suchassuperoxideradicalanionsO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(2 / andOHradicalsLinsebigleret al.,1995;Hoffmannetal.,1995;ThompsonandYates,2006; Cwiertnyetal.,2008;Chenetal.,2012;Dupartetal.,2014; ColmenaresandLuque,2014.Theseoxidizersenablerapid oxidationofadsorbedSO 2 andNO x onthesurfaceofmineral dustparticles.Forexample,usingtransmissionFouriertransforminfraredFTIRspectroscopyandX-rayphotoelectron spectroscopy,Nanayakkaraetal.observedtheoxidationofSO 2 bythephotocatalyticallygeneratedOHradicalsinthepresenceoftitaniumoxideparticles.Theheterogeneousformationofsulfateandnitratecanbehighlyvariable anddependentonthechemicalcharacteristicsofdustaerosol Gankandaetal.,2016.Authenticmineraldustparticlesdifferfrompuremetaloxidesinchemicalcomposition.Forexample,Wagneretal.reportedthatthecontentofmetal oxidesinSaharandustsamplescollectedfromBurkinaFaso includes14%Al 2 O 3 ,8.4%Fe 2 O 3 ,and1.2%TiO 2 Mostresearchondustphotochemistryhasbeenlimited toqualitativestudiesandlackskineticmechanismsthatare linkedtoapredictivemodel.Thetypicalwave-dependent photolysisofgas-phasetracegaseshaslongbeensubjectto atmosphericphotochemistry.Thisphotolysisrateisarstorderreactionandiscalculatedviathecouplingactinicux thequantityofphotonswiththecharacteristicscrosssectionareaandquantumyieldofalight-absorbingmolecule McNaughtandWilkinson,1997.Inordertomodeldust photochemistry,theintegrationofwavelength-dependentactinicuxwiththephotocatalyticactivityofmineraldustis needed. Inadditiontosunlightintensity,humidityalsoinuences heterogeneousdustchemistry.Humiditygovernsparticlewatercontent,whichinuencesthegasdustsorptionprocess oftracegasesNaveaetal.,2010andtheformationofdustphaseoxidants.Huangetal.foundthatthe SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 of SO 2 autoxidationinATDparticlesincreasedby142%becausetherelativehumidityRHchangedfrom15to90%. InthepresenceofUVlight,theparticlewatercontentcan actasanacceptorfor h C vb andproducesurfaceOHradicals, promotingheterogeneousphotochemistryofSO 2 onmineraldust.InthepresenceofUVlight,Shangetal. reportedthatsulfateproductiononthesurfaceofTiO 2 increasedby5timesbecauseoftheincreaseinRHfrom20 to80%.ParkandJangalsoreportedtheexponential increasein SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 astheRHincreasedfrom20to80%for bothautoxidationandphotooxidationofSO 2 inthepresence ofATDparticles.Afewstudieshaveattemptedtosimulate sulfateformationinthepresenceofmineraldustatregional scalesusinglaboratory-generatedkineticparametersTang etal.,2004;LiandHan,2010;Dongetal.,2016.However, SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 appliedtotheregionalsimulationsoriginatedfrom pureanddrymetaloxideswithoutUVlight,andthuswill differfromthoseofambientdustexposedtonaturalsunlight. Itisexpectedthatthetypicalregionalsimulationsduringdust eventsmightunderestimatetheformationofsulfate. Inthisstudy,theAtmosphericMineralAerosolReaction AMARmodelwasdevelopedtopredictatmosphericoxidationoftracegasessuchasSO 2 andNO 2 underambientconditions.Thekineticmechanismsofdust-drivenphotochemistry,includingautoxidationandphotooxidationof SO 2 ,werenewlyestablishedinthemodel.Therateconstant ofdustphotoactivation,whichformselectronholepairsand sourcesdust-drivenoxidants,wasintegratedintothemodel. Theinuenceofmeteorologicalvariables,suchashumidity,temperature,andsunlight,onSO 2 oxidationwasinvestigatedusingtheresultingAMARmodel.Themodel alsoaddressesthekineticmechanismtosimulatehowatmosphericmajorpollutantssuchasNO x andozoneare engagedintheoxidationofSO 2 inthepresenceofairbornedustparticles.Forenvironmentalscenarios,themodel wasappliedforpollutedurbanconditionse.g.,hydrocarbon ppbC = NO x ppb < 5andlow-NO x conditionse.g.,hydrocarbonppbC = NO x ppb < 5.Thereactionrateconstantsfor bothautoxidationandphotocatalyticreactionsofSO 2 were obtainedthroughthesimulationofindoorchamberdata, whichwerepreviouslygeneratedundervariousmeteorologicalandenvironmentalconditionsParkandJang,2016.The suitabilityoftheresultingAMARmodelwastestedagainst sulfateformationinalargeoutdoorsmogchamberatthe UniversityofFloridaAtmosphericPhotochemicalOutdoor ReactorUF-APHORundernaturalsunlight.TheAMAR modelofthisstudywillvastlyimprovetheaccuracyofthe predictionofsulfateandnitrateformationinregionaland globalscaleswheredustemissionisinuential. 2Experimental 2.1Chamberexperiments Theindoorchamberdataofthisstudywereobtainedfrom therecentlaboratorystudybyParkandJangtodeterminethekineticrateconstantsthatareneededtodevelop Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 3

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10003 theAMARmodel.Theindoorchamberoperationhasbeen reportedpreviouslyParkandJang,2016alsoseeSect.S1 intheSupplement.Theindoorchamberdataarelistedin Table1.Theoutdoorchamberexperimentswereperformed intheUF-APHORdual-chambersm 3 foreachchamber totestthesuitabilityofAMARmodeltoambientcondition. Thelightirradiationoftheindoor-UVlightandthesunlight isshowninFig.S1.Adetaileddescriptionoftheoperation oftheoutdoorchamberisalsodescribedinSect.S1.The outdoorexperimentalconditionsforSO 2 heterogeneousreactioninthepresenceofmineraldustparticlesarelistedin Table2. 2.2LightabsorptionofATDparticles TheabsorbancespectrumofATDparticleswasmeasured todevelopthereactionrateconstantsinthekineticmodel. Thedetailedprocedureforlightabsorptionmeasurementof particlesamplescanbefoundinthepreviousstudyZhong andJang,2011.TheparticlesizedistributionofATDis showninFig.S2.Thesuspendeddustparticlesweresampled onaTeon-coatedglassberlterfor20min.Themasses differenceofdustsamplewasmeasuredusingamicrobalanceMX5,MettlerToledo,Columbus,OH.ThelightabsorbanceofthedustltersampleAbs ATD / wasmeasured usingaPerkin-ElmerLambda35UVvisiblespectrophotometerequippedwithaLabsphereRSA-PE-20diffusereectanceaccessory.Theabsorbancespectrumwasnormalizedbyparticlemassandcalculatedtoobtainthemass absorbancecrosssectionSeeSect.S1.TheresultingabsorbancecrosssectionandquantumyieldofATDdustare showninFig.S3. 3AMARmodeldescription TheoverallschematicoftheAMARmodelisshown inFig.1.Inthemodel,thetotalsulfatemassconcentration[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / ispredictedfromthereactions inthreephases:thesulfateformedinthegasphase [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] gas ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / ,thesulfatefromtheaqueousphase [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] aq ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / ,andthesulfatefromdust-drivenchemistry[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] dust ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / .Thekeycomponentsofthemodel consistofthepartitioningprocessandthekineticmechanismsinthreephases. 1.Thegaseousinorganicspeciese.g.,SO 2 ,NO x and ozonearepartitionedontobothinorganic-saltsulfuricacidanditssaltsseededaqueousparticlesandmineraldustparticles.ATDparticlesareknowntobecoated withthemultilayerofwaterduetotheirhighafnityto waterGustafssonetal.,2005Sect.3.2.1.Therefore, weassumethatgasdustpartitioningoftracersonmultilayerwaterisprocessedinabsorptionmode. 2.SO 2 oxidationinthegasphaseissimulatedusingmechanismspreviouslyreportedintheliteratureByunand Schere,2006;Sarwaretal.,2013,2014;Binkowskiand Roselle,2003TableS1intheSupplement. 3.ThepartitionedSO 2 isheterogeneouslyoxidizedin theinorganic-saltseededaqueousphasebasedonthe previouslyreportedmechanismsLiangandJacobson, 1999. 4.Theformationofsulfate[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] dust / inthedustphase isapproachedusingtwokineticsub-modules:theproductionofsulfate[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / byautoxidation inopenairandsulfateformation[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / byphotocatalyticreactions.Overall,dustchemistry withinthemultilayerofwateristreatedinasimilar mannertoaqueouschemistry.However,aqueouschemistryisoperatedthroughthewholeaerosolvolumeand dustchemistryisprocessedinthewaterlayersonthe surfaceofdustparticles. ThesimulationofchamberdatausingthemodelwasperformedusingakineticsolverMorphoJeffries,1998.In thesemechanisms,thesymbolsg,aq,andddenote thechemicalspeciesinthegasphase,inorganic-saltseeded aqueousphase,anddustphase,respectively.Theunitof theconcentrationofchemicalspeciesismoleculepercubiccentimeterofair.TherateconstantsassociatedwithvariousreactionmechanismsintheAMARmodelweredeterminedbysimulatingpre-existingindoorchamberdataobtainedfromcontrolledexperimentalconditionsParkand Jang,2016.Forexample,therateconstantforSO 2 autoxidation k auto ,s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / issemiempiricallydeterminedbyttingthe predictedconcentrationofsulfatetotheexperimentaldata D1inTable1.Thegasdustpartitioningconstant K d ; SO 2 Sect.3.2.1ofSO 2 isdependentontemperature,aerosol watercontent,andacidity. K d ; SO 2 valuesweresemiempiricallydeterminedusingdataD1D3threedifferentRHs andtheliteratureparametersrelatedtotheeffectoftemperatureandacidityon K d ; SO 2 .Therateconstant k photo cm 3 molecule )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 / forthesulfateformationbyphotocatalyticreactionsissemiempiricallydeterminedusingdata L1L3threedifferentRHsinTable1.Inthepresenceof ozone, k auto and k photo aredeterminedusingdatasetsD4and L4,respectively.Inthefollowingsections,thecomponents oftheAMARmodelaredescribedindetail. 3.1SO 2 oxidationingasphaseandaerosolaqueous phase 3.1.1Gas-phaseoxidation TheoxidationofSO 2 inthegasphasehasbeenextensively studiedbynumerousresearchersBaulchetal.,1984;Kerr, 1984;AtkinsonandLioyd,1984;CalvertandStockwell, 1984;Graedel,1977;Atkinsonetal.,1997.Inthisstudy, www.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 4

10004Z.Yuetal.:AtmosphericMineralAerosolReactionmodel Table1. ExperimentconditionsandsimulationresultsforSO 2 heterogeneousphotooxidationonthesurfaceofATDparticlesatavarietyof humidityconditionsRH,lightsources,andinitialconcentrationsoftracesusingindoorchamberdata. Exp.no. a UV RH b %Temp. b K Initialconcentration Duration e min Exp. Note g ATDdust c SO d 2 NO = NO d 2 O d 3 T SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 U f T gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / ppbppbppbgm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / D1Off21.0295.9295267n/an/a1500.61 0.02 K d ; SO 2 ;k auto D2Off55.3295.04061520.1/0.61.861481.02 0.01 D3Off80.1294.52781470.9/1.60.291471.59 0.02 L1On20.4297.012387.80.3/1.70.301201.66 0.04 k OH ; O 2 k auto L2On55.2299.312082.30.2/1.91.791202.54 0.21 L3On80.7298.713178.00.2/0.40.281205.22 0.19 L4On21.0296.913078.10.1/1.3564.81204.48 0.14 k OH ; O 3 D4Off20.4296.6293101.00.7/1.965.4600.158 0.01 k auto ; O 3 a Ddenotesexperimentsunderdarkconditions.LdenotesexperimentswithUVlight.ThedatasetD1D3andL1L4wereobtainedfromtherecentlaboratorydatareportedbyPark andJang.DatasetD4wasnewlyaddedheretoestimatethekineticparameterofheterogeneousautoxidationofSO 2 inthepresenceofozone. b TheaccuracyofRHis 5%.The accuracyoftemperatureis 0.5K. c ThemassconcentrationofATDparticleswerecalculatedcombiningSMPSdata,OPCdata,thedensityofdustparticles.65gcm )]TJ/F100 5.9776 Tf 4.663 0 Td [(3 / ,andthe particlesizedistribution < 3m.Theerrorsassociatedwiththedustparticlemassconcentrationwere 6%. d TheerrorsassociatedwiththeobservationofSO 2 ,NO,NO 2 ,andO 3 were 0.9, 12.5, 6.9,and 0.2%,respectively. e Thedurationisthesimulationtimefromthebeginningoftheexperimenttotheendoftheexperiment. f Sulfateconcentrationswere measuredattheendofexperimentsusingPILS-IC.Themeasurementswerenotcorrectedfortheparticlelossratetothewallbutcorrectedfortheindigenoussulfatefromdustparticles. g Theexperimentsarenotedwiththeassociatedkineticparametersthatwereempiricallydetermined.n/a D notapplicable Table2. OutdoorchamberexperimentconditionforSO 2 heterogeneouslyphotooxidationontheATDparticlesatvarietyinitialconcentration ofSO 2 ,dustparticle,andNO x Exp.datePurposeRH a %Temp. a KSimulationtimeEST Initialconcentration b ATDdust c SO 2 NO = NO 2 O 3 gm )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 / ppbppbppb 28Mar2015SO 2 18277.1.911:10:30n/a60.10.1/0.96.3 28Mar2015SO 2 &dust15277.8.510:50:30290.156.40.1/0.70.7 16Jun2015Lowdust15286.7.008:40:3090.1100.00.1/0.70.7 16Jun201Highdust16287.0.509:30:30403.7120.11.1/1.05 12Nov2015LowSO 2 24287.8.908:40:30239.2119.00.5/2.03.0 12Nov2015HighSO 2 14287.3.0609:00:30229.0271.60.2/2.12.6 14Apr2017NO x effect33287.8.306:30:30496.288.188.9/13.53.0 25Apr2017NO x effect18283.8.606:00:00414.015.0112.0/13.22.2 25Apr2017NO x effect26284.1.706:00:00478.717.535.9/3.61.9 a TheaccuracyofRHis 5%.Theaccuracyoftemperatureis 0.5K. b TheerrorsassociatedwiththeobservationofSO 2 ,NO,NO 2 ,O 3 ,NH C 4 ,andtheconcentrationof dustparticlemasswere 0.9, 12.5, 6.9, 0.2, 5.0,and 6%,respectively.Thedetailedobservationsofthechemicalspeciesduringtheexperimentswereshownin Figs.S4andS5intheSupplement. c ThemassconcentrationsofATDparticleswerecalculatedcombiningSMPSdata,OPCdata,thedensityofdustparticles.65gcm )]TJ/F100 5.9776 Tf 4.662 0 Td [(3 / andtheparticlesizedistribution < 3m.n/a D notapplicable theoxidationofSO 2 isdescribedusingcomprehensivereactionmechanismsshowninTableS1.Themechanismscan alsobesimpliedasfollows: SO 2 g / C OH HOSO 2 ; R1 HOSO 2 C O 2 SO 3 C HO 2 ; R2 SO 3 g / C H 2 O g / C M H 2 SO 4 aq / C M; R3 HOSO 2 C OH g / C M H 2 SO 4 aq / C M: R4 3.1.2Gasaerosolpartitioning SO 2 isdissolvedintohygroscopicsulfuricacidH 2 SO 4 / whichisformedinthegasphase,viaapartitioningprocessandreactswiththeaqueous-phaseoxidantse.g.,H 2 O 2 andO 3 / toheterogeneouslyformH 2 SO 4 .Thechemical speciesthatweretreatedbythepartitioningprocessincludeSO 2 ,NO x ,O 3 ,OH,HO 2 ,H 2 O 2 ,HCOOH,CH 3 OOH, HNO 3 ,CH 3 O 2 ,HONO,CH 3 COOH,andHCHO.Inthe model,thepartitioningprocessisapproachedusingthegas particlepartitioningcoefcient K aq ; SO 2 m 3 g )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / basedon aerosolmassconcentration. K aq ; SO 2 isderivedfromHenry's lawconstantofSO 2 K H ; SO 2 D 1.2molL )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 atm )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 at298K Chameides,1984, K aq ; SO 2 D K H ; SO 2 RT aq ; where R istheidealgasconstantJK )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 mol )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / and aq gcm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / isthedensityoftheparticle,whichiscalculatedusAtmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 5

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10005 Figure1. TheoverallschematicoftheAMARmodeltosimulateheterogeneousSO 2 oxidation.Forthedescriptionofchemicalspecies, gasphase,aqueousphase,anddustphasearesymbolizedasgas,aq,anddust,respectively.SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 _T,H 2 SO 4 _gas,SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 _aq,and H 2 SO 4 _dustarethetotalsulfateformationandtheformationofsulfatefromgasphase,aqueousphase,anddustphase,respectively. SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 _d_saltandNO )]TJ/F100 7.5716 Tf 0 -7.412 Td [(3 _d_saltaretheneutralizedsulfateandnitrateinthedustphase. inganinorganicthermodynamicmodelE-AIMIIClegget al.,1998;WexlerandClegg,2002;CleggandWexler,2011 basedonhumidityandinorganiccomposition.ThepartitioningprocessofSO 2 oninorganicaerosolIn aq ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / isexpressedas SO 2 g / C In aq SO 2 aq / C In aq k abs ; SO 2 ; aq m 3 g )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /; R5 SO 2 aq / SO 2 g / k des_SO 2 ; aq s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /; R6 where k abs ; SO 2 ; aq m 3 g )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 and k des ; SO 2 ; aq s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 / arethe uptakerateconstantandthedesorptionrateconstant,respectively,andarecalculatedasfollows: k abs ; SO 2 ; aq D f abs ; aq SO 2 f aq ; S_M 4 ; k des ; SO 2 ; aq D k abs ; SO 2 ; aq K aq ; where f aq ; S_ M 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(4 ,m 2 g )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / isthecoefcienttoconverttheaerosolmassconcentrationgm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / tothesurface areaconcentrationm 2 m )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 / forparticlesizenear100nm. f abs ; aq isthecoefcientforuptakeprocessand SO 2 isthe meanmolecularvelocityms )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / ofSO 2 andcanbecalculatedasfollows: SO 2 D r 8 RT MW ; whereMWismolecularweightkgmol )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / .Inourmodel, f abs ; aq wassetat2 10 4 inEq.tohavefastpartitionwww.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 6

10006Z.Yuetal.:AtmosphericMineralAerosolReactionmodel ingprocess.TableS2summarizesthecharacteristictimethat isestimatedfordiffusion,partitioning,andthereactionsof majorspecieswithOHradicalsingas,aqueous,anddust phases.Ingeneral,thecharacteristictimesofapartitioningprocessorderof10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(7 sismuchfasterthangas-phase oxidationorderof10 6 s,aqueous-phaseoxidationorderof 10 3 4 s,anddust-phaseoxidationorderof10 2 3 sat presenceof200gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 ofdustparticles.Themassconcentrationgm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / ofinorganicseededaqueousphaseabove theeforescentrelativehumidityERHisalsodynamicallycalculatedfortheSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 NH C 4 H 2 Osystem.Colberg etal.semiempiricallypredictedERHbyttingtothe experimentaldatabasedontheammonia-to-sulfateratioin theSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 NH C 4 H 2 Osystem.AMARmodelutilizesthese parameterizationstopredictERHdynamically.Ammoniais inevitableinourchamberstudyandmainlyactsasacarryoverfrompreviouschamberexperiments.Thus,H 2 SO 4 is fullyorpartiallyneutralizedbyammonia. 3.1.3Aerosolaqueous-phasereaction TheAMARmodelimplementsaqueous-phasechemistry thatoccursininorganicsaltedaqueousaerosolSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 NH C 4 H 2 OsystemwithoutdusttoformSO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 aq / and NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 aq / .Weemployedthepreexistingaqueous-phasekineticreactionsinvolvingSO 2 LiangandJacobson,1999 andNO x chemistryLiangandJacobson,1999;Hoyleetal., 2016.Thus,oursimulationinheritsallthepossibleuncertaintiesembeddedintheoriginalkineticdata. TheSO 2 dissolvedintheaqueousphaseishydrolyzed intoH 2 SO 3 anddissociatestoformionicspeciesHSO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 and SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(3 / .SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 aqisformedbyreactionsofthesulfurspecies inoxidationstateIVS IV / aq / / withOHaq,H 2 O 2 aq, orO 3 aqTableS1.ThedissolvedHONOcanalsodissociatetoformNO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(2 aq / andresulttoNO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 aq / .EachchemicalspeciesinS IV / aq / hasadifferentreactivityforoxidationreactions.Thedistributionofchemicalspeciesisaffectedbyaerosolacidity,whichiscontrolledbyhumidity andinorganiccomposition.Hence,theformationofsulfate isverysensitivetoaerosolacidity.Forexample,mostofthe SIVisconsumedbyH 2 O 2 atpH < 4,whereasmostofit isconsumedbyO 3 atpH > 4.Somestronginorganicacids, suchassulfuricacid,inuenceaerosolacidity.InAMAR, aerosolacidity[H C ]isestimatedateachtimestepbyEAIMIICleggetal.,1998;WexlerandClegg,2002;Clegg andWexler,2011correctedfortheammonia-richconditionLietal.,2015;BeardsleyandJang,2016;LiandJang, 2012asafunctionofinorganiccompositionmeasuredby aparticle-into-liquidsamplercoupledwithionchromatographyPILS-IC.Whentheammonia-to-sulfateratioisgreater than0.8,thepredictionof[H C ]iscorrectedbasedonthe methoddescribedbyLiandJang.AthighNO x levels,NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(2 aq / competeswithS IV / aq / forthereactionwith OHaq,O 3 ,orH 2 O 2 TableS1Maetal.,2008.However, theHONOconcentrationbecomeshighathighNO x levelsandenhancesSO 2 oxidationintheinorganic-saltseeded aqueousphaseduetotheformationofOHradicalsviaphotolysisofHONO. 3.2Heterogeneousoxidationinthepresenceofmineral dustparticles TheheterogeneouschemistryinthepresenceofdustparticleshasbeennewlyestablishedintheAMARmodel. Thedust-phasemoduleconsistsofapartitioningprocess Sect.3.2.1andheterogeneouschemistryforSO 2 andother tracegasesozone,HONO,andNO 2 / Table3Fig.1.The heterogeneouschemistryofSO 2 ishandledbyautoxidation Sect.3.2.2andphotooxidationunderUVlightSect.3.2.4. Indust-phasephotochemistry,thecentralmechanismfor SO 2 oxidationisoperatedbythesurfaceoxidantse.g., OHd,whichisgeneratedviathephotoactivationprocess ofsemiconductivemetaloxidesindustparticlesSect.3.2.3. 3.2.1Gasdustparticlepartitioning Inanadsorptivemode,watermoleculessuppresspartitioning ofSO 2 becausetheycompeteforadsorptivesiteswithtracersCwiertnyetal.,2008.However,theformationofthe sulfateassociatedwithATDincreasedwithincreasingRH asshowninTable1,suggestingthatgasdustpartitioningis morelikelyoperatedbyabsorptiononthemultilayercoated dustwithwatermolecules.ATDcontainshygroscopicinorganicsaltsthatformthethinwaterlmonthesurfaceofATD particleswhenthesaltsaredeliquescentoraboveERH. Somesaltssuchasmagnesiumsulfateandcalciumsulfate canbehydratedevenatlowhumidityBeardsleyetal.,2013; Jangetal.,2010.Gustafssonetal.reportedthatATD particlesshowedasubstantiallyhighafnitytowatercomparedtopureCaCO 3 particles.Intheirstudy,thewatercontentofATDparticles,whichwasmeasuredusingthethermogravimetricmethod,rangedfromtwomonolayerstofour monolayersbasedontheBETsurfaceareabetween20and 80%relativehumidity.Thiswaterlayerinuencesgasdust partitioningofatmospherictracerssuchasSO 2 andNO 2 Thegasdustpartitioningconstant K d ; SO 2 ,m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 / ofSO 2 isdenedas K d ; SO 2 D T SO 2 U d T SO 2 U g A Dust m 3 m )]TJ/F100 7.5716 Tf 5.905 0 Td [(2 /; where A dust m 2 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / isthegeometricsurfaceconcentration ofATDdustparticlesandiscalculatedbymultiplyingthe dustmassconcentrationgm 3 / byageometricsurface-mass ratio f dust ; S_ M / ofATDparticles.066 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(6 ,m 2 g )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / TheSO 2 absorptionanddesorptionprocessesforthedust phaseareexpressedas Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 7

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10007 Table3. Dust-phaseheterogeneousreactionsandtheirrateconstantsinthepresenceofATDparticles. Reaction a Rateconstant b Coefcientsofrateconstants b K c a Reference d Note e k 1 k 2 Partitioning 1SO 2 C Dust SO 2 d / C Dust k abs 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(8 AR05,HZ15ReactionR7 2SO 2 d / SO 2 k des 1 10 9 31000.013AR05,HZ15ReactionR8 3O 3 C Dust O 3 d / C Dust k abs 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(8 MU03,US01 4O 3 d / O 3 k des 3 10 10 27000MU03,US01 5NO 2 C Dust NO 2 d / C Dust k abs 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(8 CW84 6NO 2 d / NO 2 k des 1 10 10 25000CW84 7HNO 3 C Dust HNO 3 d / C Dust k abs 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(8 SW81,Sc84 8HNO 3 d / HNO 3 k des 1 10 15 870015.4SW81,Sc84 9HONO C Dust HONO d / C Dust k abs 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(8 BK96 10HONO d / HONO k des 1 10 10 49000BK96 11N 2 O 5 C Dust HNO 3 d / C Dust k abs 7 : 3 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 WS09 Dustphase 1Dust C h Dust C e hk j e h j T ATD U Sect.3.2.3ReactionR10 2 e h energy k recom 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 Sect.3.2.3ReactionR11 3 e h C O 2 OH d /k OH ; O 2 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(22 2.3RHSect.3.2.3ReactionR12 4SO 2 d / SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 d /k auto 5 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(6 Sect.3.2.2ReactionR9 5SO 2 d / C OH d / SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 d /k photo 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 Sect.3.2.4ReactionR13 6SO 2 d / C O 3 d / SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 d / C O 2 k auto ; O 3 2 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(11 Sect.3.3.1ReactionR14 7 e h C O 3 d / OH d / C O 2 k OH ; O 3 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 Sect.3.3.1ReactionR15 8NO 2 d / NO )]TJ/F100 7.5716 Tf 0 -7.412 Td [(3 d /k auto ; NO 2 6 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(5 Sect.3.3.2ReactionR18 9 e h C NO 2 d / HONO d /k e h; NO 2 6 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 Sect.3.3.2ReactionR16 10HONO d / C h OH d / C NO k j HONO j T HONO_to_OH U BK91,AB97ReactionR17 11NO 2 d / C OH d / NO )]TJ/F100 7.5716 Tf 0 -7.411 Td [(3 d /k photo ; NO 2 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(10 Sect.3.3.2ReactionR19 a Theunitofthechemicalspeciesexceptdustismoleculecm )]TJ/F100 5.9776 Tf 4.662 0 Td [(3 forbothpartitioningprocessanddust-phasechemistry.Theunitofthedustformodelinputismassconcentration gm )]TJ/F100 5.9776 Tf 4.663 0 Td [(3 / andismultipliedbyafactorof2.45 10 10 forsimulation. b Theunitofreactionrateconstantsiss )]TJ/F100 5.9776 Tf 4.663 0 Td [(1 fortherst-orderreactionsandcm 3 molecule )]TJ/F100 5.9776 Tf 4.662 0 Td [(1 s )]TJ/F100 5.9776 Tf 4.663 0 Td [(1 forthe second-orderreactions. k abs isuptakerateconstant. k abs D k 1 !f dust ; S_M = 4 ,where D p 8 RT=. MW / ms )]TJ/F100 5.9776 Tf 4.662 0 Td [(1 / and f dust ; S_M D 3 .066 10 )]TJ/F100 5.9776 Tf 4.662 0 Td [(6 m 2 g )]TJ/F100 5.9776 Tf 4.662 0 Td [(1 / R istheidealgas constantandMWgmol )]TJ/F100 5.9776 Tf 4.662 0 Td [(1 / isthemoleculeweightofchemicalspecies. k des isdesorptionrateconstant. k des D k 1 exp )]TJ/F53 5.9776 Tf 6.48 4.014 Td [(k 2 T =.F water 1 C K a = H C // ,where F water iscalculated usingEq..[H C ]isdynamicallycalculatedbasedonthermodynamicmodelE-AIMIICleggetal.,1998;WexlerandClegg,2002;CleggandWexler,2011.Therateconstants k/ fordust-phasereactionsis k D k 1 exp .k 2 / k j e h and k j HONO arephotocatalyticreactionrates.ThecrosssectionsandquantumyieldsofdustareestimatedseeSect.2.2. c Coefcient K a isaciddissociationconstantsee k des / d Therateconstantparameters,whicharenotedasthisstudy,aredeterminedusingthesimulationofindoorchamberdata ParkandJang,2016seeSect.3.AB97,Atkinsonetal.;AR05,Adamsetal.;BK91,Bongartzetal.;BK96,Beckeretal.;CW84,Chameides; HZ15,Huangetal.;MU03,Micheletal.;Sc84,Schwartz;SW81,SchwartzandWhite;US01,Underwoodetal.;WS09,Wagneretal.. e Thereactionsarenotedwiththenumbersassociatedwiththereactioninthemaintext. SO 2 g / C A Dust SO 2 d / C A Dust k abs_SO 2 ; dust m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /; R7 SO 2 d / SO 2 g / k des_SO 2 ; dust s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /; R8 where k abs_SO 2 ; dust m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / and k des_SO 2 ; dust s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / arethe absorptionrateconstantandthedesorptionrateconstant,respectively.Atequilibrium,theabsorptionrateR7equals thedesorptionrateR8.Thus, K d ; SO 2 canbeexpressedas K d ; SO 2 D k abs_SO 2 ; dust k des_SO 2 ; dust m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 /: The K d ; SO 2 valueat20%RHissetat1.63m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 / based ontheliteraturedatadustparticlesat20%RHAdams etal.,2005;Huangetal.,2015.Thecharacteristictimeto reachtoequilibriumisveryshortSect.3.1.1.Inkinetic mechanisms, k ads_SO 2 ; dust wassetat1 : 7 10 3 m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 fordryparticles%RHusingthesameapproachas Eq..Theresultingcharacteristictimefor k ads_SO 2 ; dust is 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(6 s.ThecharacteristictimeofthereactionofSO 2 withan OHradical 6 moleculescm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / isabout10 6 7 singas phaseand10 5 6 sinbothaqueousphaseanddustphase. Toconsidertheeffectoftemperatureon K d ; SO 2 ,thetemperaturedependencyof k des_SO 2 ; dust Eq.6isderivedfrom theHenry'sconstantChameides,1984. K d ; SO 2 Eq.5is alsoinuencedbyaerosolwatercontentZuendetal.,2011 aswellasthedissociationofH 2 SO 3 ,whichisoperated byaerosolacidity[H C ]andanaciddissociationconstant Ka SO 2 / MartellandSmith,1976.Thus, k des_SO 2 ; dust isexwww.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 8

10008Z.Yuetal.:AtmosphericMineralAerosolReactionmodel pressedas k des_SO 2 ; dust D 1 10 9 exp )]TJ/F100 8.4682 Tf 7.801 5.936 Td [(3100 T = F water 1 C Ka SO 2 H C !! s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /: Ka SO 2 is0.013molL )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / at298KMartellandSmith, 1976.Theinuenceofthedissociationofinorganicacidon K d ; SO 2 isaccountedforbytheterm C Ka SO 2 [ H C ] / inEq.. Theestimationof[H C ]istreatedinthesamewaysasaqueouschemistrySect.3.1.3. Inordertoestimate K d ; SO 2 atdifferentRH, F water coefcientofthemassfractionofwatertodustparticleswas introducedintothemodel.Thehygroscopicpropertyofmineraldustdynamicallychangesbecausedustcanbesubstantiallymodiedbydirectreactionofsomeofitscomponentse.g.,CaCO 3 / withinorganicacidssuchasH 2 SO 4 and HNO 3 .WhendustformsCaNO 3 / 2 ,dustbecomesmorehygroscopic.NitratesaltsdeliquesceatverylowRH% Kruegeretal.,2003,2004.CaSO 4 is,however,relatively hydrophobic.Nitratesaltsexistonlywhensulfateconcentrationsisverylow.Inthemodel, F water isassociatedwith thehygroscopicpropertyofindigenousdustrsttermin Eq.8,theinorganicnitratesformedfromthereactionof absorbedHNO 3 withdustsecondterm,andtheinorganic sulfateSO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 NH C 4 H 2 Osystem,thirdterm. F water D exp 4 : 4RH / C 3 : 7 f dust ; S_ M exp 4 : 4RH / NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 d_salt / A Dust C f dust ; S_ M M in ; water A Dust ; where M in ; water isthewaterconcentrationgm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / associatedwithinorganicsulfateandcalculatedusingE-AIM II.Both NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 d_salt / and M in ; water arenormalizedby themassconcentrationofATDparticles[Dust],gcm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / F water isrstdeterminedusingchambersimulationofSO 2 heterogeneousoxidationrstandthirdtermsinEq.8D1 D3inTable1undervariedRHlevelsandextendedtoSO 2 oxidationinthepresenceofNO x Exp.14April2017inTable2.Amongtemperature,RH,andaerosolacidity,themost inuentialvariableisRHduetothevariationin F water see sensitivityanalysisinSect.5. 3.2.2AutoxidationofSO 2 ondustsurface Typically,autoxidationofSO 2 isanoxidationprocessvia thereactionofabsorbedSO 2 ReactionsR7andR8withan oxygenmolecule.Inthemodel,[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto isdenedasthe sulfateresultedfromanyoxidationreactionsautoxidationin openairandoxidationwithozoneofSO 2 withoutUVlight Fig.1.Inautoxidation,thereactionofSO 2 dwiththeoxygenmoleculesistreatedastherst-orderreactionassuming theconcentrationofoxygenisconstantas2 10 5 ppm. SO 2 d / O 2 g / )167(! SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 d / k auto D 5 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(6 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / R9 Figure2. Uptakecoefcient / ofSO 2 inthepresenceoftheATD particlesunderdarkconditionsandUVlightconditions.Thevaluesof wereobtainedbykineticmodelusingindoorexperimental data.The SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; light iscorrelatedtoconcentrationofOHradicals andRH%.The SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; dark isafunctionofRH.Theerrorbarof wasderivedfromthemodeluncertainty. Underdarkconditions,theformationofsulfateismainly sourcedfromautoxidationofSO 2 .Forcomparisonwith otherstudies,weestimatethereactiveuptakecoefcient SO 2 )]TJ/F100 5.9776 Tf -2.988 -5.942 Td [(4 ; auto / ofSO 2 ontoATDdustintheabsenceofozone andNO x Fig.2. SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; auto D 4 K d ; SO 2 k auto SO 2 ; SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; auto isproportionalto K d ; SO 2 ,andinuencedbyhumidityEq.7. 3.2.3Photoactivationofdustparticlesand heterogeneousformationofOHradicals ThereactiveuptakeofSO 2 onparticlesistraditionally treatedasarst-orderprocessUllerstametal.,2003;Liet al.,2007.Suchanapproachisappropriateforsimpleautoxidationmechanisms,butnotforthecomplexheterogeneous photooxidationofSO 2 .IntheAMARmodel,theheterogeneousphotooxidationofSO 2 isapproachedinthreesteps: theformationofan e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pairviaphotoactivationof dustparticles,theformationofOHdviathereactionof an e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pairwithawateroroxygenmolecule,andthe reactionofabsorbedSO 2 withtheresultingOHdsecondorderreactionsTableS1. Thephotoactivationofdustparticlesandtherecombinationreactionofanelectronholepair e h areaddedinto themodel. Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 9

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10009 Dust h )167(! Dust C e hk j e h D j [ ATD ] ; R10 e h )167(! energy k recom D 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /; R11 where k j e h isthephotoactivationrateconstanttoform e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pairsand k recom isthereactionrateconstantofrecombinationheatradiationofanelectronandahole.Thevalueof k recom issetatalargenumbertopreventtheaccumulationof electronholepairs.TheformationofOHdisexpressedas e h C O 2 g / )167(! OH d /k OH ; O 2 D 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(22 exp 2 : 3RH / cm 3 molecules )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 /; R12 where k OH ; O 2 isthereactionrateconstanttoformOHdand isrstestimatedusingindoorchamberdataL1L3inTable1atRH20,55,and80%andthenregressedagainstRH. ThestudybyThiebaudetal.reportedtherecombinationofOH d / neartoTiO 2 surfaces.Inourmodel,themechanisticroleofthecatalyticformationoftheelectronhole pairsReactionR10andtheirrecombinationReactionR11 compensatestheformationandtheself-reactionofOHradicals. InReactionR10, k j e h istheoperationalrateconstantfor thephotoactivationofdustparticlesandisdependentonthe photolysisrateconstant, j [ ATD ] s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 / .Likethetypicalphotolysisofagaseousmolecule,thephotocatalyticproduction of e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pairsislineartoboththeactinicux I./ ,photonscm )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 nm )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / originatingfromthelightsourceand thephotocatalyticpropertyofdustparticles.Thevalueof j [ ATD ] isdeterminedby I./ ,theabsorptioncrosssection ./ ,cm 2 g )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / ,andthequantumyield .// ofdustconductingmatterateachwavelengthrange ,nm, j T ATD U D 2 Z 1 I / / / d: Inthemodel, ./ isthelightabsorptionneededtoactivate dust-phasesemiconductingmetaloxidesexcitationfroma groundenergyleveltoaconductingband,and ./ isthe probabilityofyieldingthe e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pairinthedustphase. Both ./ and ./ cannotbedirectlymeasuredbecauseof complexityinthequantityofphotoactiveconductingmatter industparticlesandtheirradiationprocessesofthe e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb pair.Inordertodealwith ./ ./ ,wecalculatedthe massabsorptioncrosssectionofdustparticlesMAC ATD m 2 g )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / ,whichwasdeterminedusingtheabsorptioncoefcientofATDparticles b ATD ,m )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 / withtheparticleconcentration m ATD ,gm )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 / : MAC ATD D b ATD m ATD : InEq., b ATD canbecalculatedfromtheabsorbanceof dustltersampleAbs ATD ,dimensionlessmeasuredusinga reectiveUVvisiblespectrometerFig.S3: b ATD D Abs ATD A fV ln ; where A D 7.85 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(5 m 2 / isthesampledareaonthelterand V m 3 / isthetotalairvolumepassingthroughthe lterduringsampling.Inordertoeliminatetheabsorbance causedbyltermaterialscattering,acorrectionfactor f D 1.4845isobtainedfromapreviousstudyZhongandJang, 2011andcoupledintoEq..Thepreliminarystudy showedthattheeffectofaerosolscatteringonthe b abs valuesoftheaerosolcollectedonthelterwasnegligible. Further,Bondreportedthatparticlelightscattering doesnotsignicantlyinuencespectralabsorptionselectivity.TheMAC ATD ofdustparticlesoriginatesfromphotocatalyticconductingmattere.g.,TiO 2 / aswellaslightabsorbingmattere.g.,gypsumandmetalsulfate.Thus,the MAC ATD spectrumisadjustedusingtheknownTiO 2 absorptionspectrumReyes-Coronadoetal.,2008andappliedto ./ ./ Fig.S3.Theresulting ./ ./ spectrum isappliedtoEq.tocalculate j [ ATD ] ReactionR10. 3.2.4HeterogeneousphotooxidationofSO 2 SO 2 isoxidizedbyOHdonthesurfaceofATDparticlesas follows: SO 2 d / C OH d / SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 d / k photo D 1 : 0 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 cm 3 molecule )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 /; R13 where k photo isthereactionrateconstantofSO 2 withOHd andisestimatedfromgas-phaseReactionR1.Combining Eq.,,ReactionsR11andR15,thereactiveuptake coefcient SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo / ofSO 2 onATDparticlesunderUV lightcanbewrittenas SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo D 4 K d ; SO2 )]TJ/F53 9.9626 Tf 4.015 -8.01 Td [(k photo [ OH d / ] C k auto SO 2 ; where SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo istheconstantatagivenconcentration ofOHdforagivenlightsource,dustconcentration,and humidityReactionsR10andR12.Figure2illustrates SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo valuesatthreedifferentRHs,whichwereobtainedusingindoorchamberdata. SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo issignicantly inuencedbybothUVlightandhumidity.Forexample, SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo is1orderofmagnitudehigherthan SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; auto atlowNO x levels < 5ppb,and SO 2 )]TJ/F100 5.9776 Tf -2.989 -5.942 Td [(4 ; photo increasedfrom 2.0 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(5 to1.24 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(4 whentheRHchangedfrom20to 80%. 3.3ImpactofozoneandNO x onheterogeneous chemistryofSO 2 Todate,moststudiesoftheeffectofNO x onsulfateformationhavebeenlimitedtothereactionunderdarkconditions. www.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 10

10010Z.Yuetal.:AtmosphericMineralAerosolReactionmodel Forexample,previouslaboratorystudiesusingpuremetal oxidesreportedtheaccelerationoftheheterogeneousoxidationofSO 2 byNO x underdarkconditionsMaetal.,2008; Liuetal.,2012.Fortheeffectofozone,therecentchamber studybyParkandJangshowedsignicantenhancementofheterogeneousphotooxidationofSO 2 .IntheAMAR model,theformationofsulfateisalsomodulatedbytheinvolvementofozoneandNO x inbothautoxidationandphotochemistryonthesurfaceofdustparticlesFig.1. 3.3.1Dust-phaseozonechemistry Thegasdustpartitioningcoefcientofozoneisscaled using K d ; SO 2 andtheratiooftheHenry'slawconstantofSO 2 K H ; SO 2 ,Eq.1tothatofozone K H ; O 3 D 1.2 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 molL )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 atm )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 at298KChameides,1984, K d ; O 3 D K d ; SO 2 K H ; O 3 K H ; SO 2 D 7 : 7 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(7 F water exp 2700 T / m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 /: Thepartitioningprocessisalsotreatedbytheabsorption desorptionkineticmechanismasshowninReactionsR7 andR8Table3:partitioning.Ozonecandecaycatalyticallyinthedustphase,forminganoxygenmoleculeand surface-boundatomicoxygenUsheretal.,2003;Changet al.,2005.TheformedatomicoxygenreactswithSO 2 dto formsulfateUllerstametal.,2002;Usheretal.,2002: SO 2 d / C O 3 d / )167(! SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 d / C O 2 k auto ; O 3 D 2 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(11 cm 3 molecules )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 /: R14 Inthepresenceof300gm 3 ofATDparticlesand60ppb ofozone,theconcentrationofO 3 disestimatedas 2.4 10 7 moleculecm )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 .Underthiscondition,thecharacteristictimeoftheautoxidationbyozoneReactionR14is 2 10 3 sandismuchfasterthantheautoxidationbyoxygenReactionR9,2 10 5 s.Atnighttime,inthepresenceof ozone,theautoxidationofSO 2 dyieldsasignicantamount ofsulfate. UnderUVlight,ozoneisalsoinvolvedintheproduction ofthesurfaceoxidantsO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 ,HO 3 radicals,andOHradicals thatfurtherpromoteheterogeneousoxidationofSO 2 .O 3 d actsasanacceptorfor e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb h C vb andformsOHd. e h C O 3 d / )167(! q OH d / C O 2 k OH ; O 3 D 1 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 cm 3 molecules )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 / R15 3.3.2Dust-phaseNO x chemistry ThegasdustpartitioningcoefcientofNO 2 K d ; NO 2 / is treatedasthesameapproachwithozone,using K d ; SO 2 andtheratioof K H ; SO 2 Eq.1totheHenry'slaw constantofNO 2 K H ; NO 2 D 1.2 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 molL )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 atm )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 at 298KChameides,1984. K d ; NO 2 D K d ; SO 2 K H ; NO 2 K H ; SO 2 D 1 : 5 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(6 F water exp 2500 T / m 3 m )]TJ/F100 7.5716 Tf 5.906 0 Td [(2 / TheabsorbedNO 2 rstreactswith e )]TJ/F100 7.5716 Tf -0.199 -7.595 Td [(cb d / or q O )]TJ/F100 7.5716 Tf 0 -7.595 Td [(2 d / onthe dustsurfaceReactionR10andformsHONOdMaetal., 2008;Liuetal.,2012;SalibaandChamseddine,2012;Salibaetal.,2014.InAMAR,theformationofHONOdis simpliedinto e h C NO 2 d / )167(! HONO d / k e h; NO 2 D 6 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(12 cm 3 molecules )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /: R16 HONOdisfurtherdecomposedthroughphotolysisand yieldsOHd: HONO d / hv )167(! q OH d / C NO k j HONO D j [ HONO ] s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 /: R17 ThephotolysisrateconstantofHONOdistreatedwiththe oneforgaseousHONO j T HONO U / .Similartoautoxidationof SO 2 Sect.3.2.2,NO 2 dautoxidizestoformnitrate: NO 2 d / O 2 g / )167(! NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 d / k auto ; NO 2 D 6 10 )]TJ/F100 7.5716 Tf 5.906 0 Td [(5 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 /: R18 NO 2 reactswithOHd: NO 2 d / C OH d / NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 d / k photo ; NO 2 D 1 10 )]TJ/F100 7.5716 Tf 5.905 0 Td [(10 cm 3 molecules )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 s )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 /: R19 k auto ; NO 2 and k photo ; NO 2 wasdeterminedusingthesimulationofoutdoorchamberdataExp.14April2017inTable2.TheestimationofthegasdustpartitioningcoefcientsofHONO K d ; HONO / Beckeretal.,1996andHNO 3 K d ; HNO 3 / SchwartzandWhite,1981wasapproachedusingthesimilarmethodforSO 2 Table3.N 2 O 5 formsnitrate viaareactiveuptakeprocessasshowninTable3partitioningReaction11. 4SimulationoftheAMARmodel AtthebeginningofthedevelopmentoftheAMARmodel, thekineticparameterstopredicttheformationofsulfateand nitrateinthepresenceofATDparticleswereleveragedusinganindoorchamber.Inordertotestthefeasibilityofthe resultingAMARmodel,theUF-APHORdatausingnaturalsunlightweresimulatedTable2.ThechamberdilutionmeasuredbyCCl 4 / andthewallprocessofgaseous compoundse.g.,ozone,SO 2 ,HONO,NO 2 / andparticles wereintegratedwiththekineticmechanismstosimulateUFAPHORdataSect.S1.AsshowninFig.1,themodelinputs aretheconcentrationofchemicalspecies,theamountofdust, andthemeteorologicalvariablesthatarecommonlyfoundat regionalscales.Thedualchambersallowfortwocontrolled experimentstobeperformedsimultaneouslyunderthesame meteorologicalconditions. Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 11

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10011 4.1Simulationsfordifferentdustloadings Figure3showsthatthepredicted[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T isingoodagreementwithexperimentalobservations,whichwereperformed underlow-NO x conditionsNO x < 5ppbfortwodifferent dustloadingsaswellastwodifferentSO 2 levels.Thegreater increasein[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T appearedwiththehighersunlightintensitybetween11:00and14:00.InFig.3a,thepredicted [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T increasedby63%at3PMwith290gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 of ATDparticlescomparedtothe[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T withoutdustparticles.Figure3bconrmsthatthelargerdustparticleloading yieldsmore[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T .InFig.3c,[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T waspredictedwith highandlowinitialconcentrationsofSO 2 foragivendust loading.Thetimeprolesofthesimulationofconcentrations ofNO x ,ozone,SO 2 ,anddustareshowninFig.S4. Becauseofthelargesizeofdustparticles,thewallprocessese.g.,settlingandwalldepositionofdustparticles isgreaterthanthatofthesulfateparticlesoriginatedfrom [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] aq nodust.Hence,thefractionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] dust to [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T declinesoverthecourseofthechamberexperiment.Toestimatehowthepredicted[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T isattributedto [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] aq C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] gas non-dustsulfateand[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] dust withoutwallprocesses,Fig.3d,e,andfarereconstructedfrom Fig.3a,b,andc,respectively.Asshownintheinnerpie chartofFig.3d,asignicantfractionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T isattributed todust-phasechemistry[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo :0.58.In Fig.3e,thefractionofnal[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo to[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T increases from0.28to0.72withtheincreaseindustloadingfrom90 to403gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 .Theincreaseddustloadingpromotesboththe absorptionofSO 2 ontodustparticlesandtheproductionof dust-phaseoxidantsandthusyieldsmoresulfateproduction. WiththeincreaseintheinitialconcentrationofSO 2 from 119to272ppbinFig.3f,thefractionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo and [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] gas C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] aq arenotmuchchanged,while[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T increasesfrom16.6to30.1gm )]TJ/F100 7.5716 Tf 5.905 0 Td [(3 .TheelevationoftheconcentrationofSO 2 producesmoresulfateinallthreephases gas,aqueous,anddustphases.Thesulfuricacidformedin theaqueousphaseishydrophilicandcreatesapositivefeedbackloopwhichaggravatesthegrowthofaqueousaerosol. Overall,thevariationindustconcentrationismoreinuentialon[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo thanthatofSO 2 4.2SimulationofNO x effect Figure4showsthatthemodelperformswellinpredicting [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T invariouslevelsofNO x .Figure4disreconstructed fromFig.4a,b,andctoillustratehow[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T isattributed totheaqueous-phasereaction[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] gas C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] aq / ,dustphaseautoxidation[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto / ,anddustphotochemistry [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo / .ComparingFig.4bwithc,[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo issuppressedathighNO x levelsbecauseNO 2 competesforthe consumptionofdust-phaseOHradicalswithSO 2 .Thereductionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T intheafternoonisduetotheparticleloss atthelowconcentrationsofSO 2 .ThesimulatedconcentrationsofNO x ,ozone,SO 2 ; anddustareshowninFig.S5. Thetimeprolesofthepredicted[NO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 ] T arealsoshown inFig.4a,b,andc.Inthemorning,NO 2 quicklyoxidizes toaccumulatenitricacidinthedustphase.Thedust-phase nitricacidmightrapidlyreactwithalkalinecarbonatese.g., K,Na,CaandMgionsinthedustphaseandformnitrate saltsNO )]TJ/F100 7.5716 Tf 0 -7.595 Td [(3 d_salt / inReactionS23ofdust-phasereactions inTableS1.AsdescribedinSect.3.2.1,thesenitratesalts areveryhygroscopicandfurtherenhancegasdustpartitioningofgaseousspeciesincludingHNO 3 ,SO 2 ,andHONO athighhumidityinthemorning.Withincreasingsunlight intensity,thetemperatureincreasesbuthumiditydecreases %,Fig.S6andthusincreasethedesorptionofHNO 3 .In additiontometeorologicalconditions,theformationoflowvolatilitysulfuricacidcandepletenitrateviaevaporationof volatilenitricacidSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 d_salt / inReactionsS24andS25 ofdust-phasereactionsinTableS1fromthedustsurface. ThecapacityofATDparticlestoformnitratesaltsorsulfatesaltsislimitedbytheamountofcarbonatesandmetal oxidesonthesurfaceofdustparticles.Thiscapacityisestimatedtobe0.6ppbthenumberconcentrationofreactive sitesinair,whichwasdeterminedbycomparingtheactual aerosolacidity,asmeasuredbycolorimetryintegratedwitha reectanceUVvisiblespectrometerC-RUV,totheaerosol aciditypredictedbytheinorganicthermodynamicmodelEAIMIIusingtheinorganiccompositionfromPILS-ICLi etal.,2015;BeardsleyandJang,2016.AsshowninFig.4, theeffectofHNO 3 ontheheterogeneousreactionisnegligibleduringthedaytimebecausesulfuricacid,astrong acid,depletespartitioningofHNO 3 Eq.15.Attheendof thephotooxidation,nitrateisslightlyunderestimatedbecause someobservednitratemaybetrappedunderthelayerofhydrophobicalkalinesulfateformedviaagingofATDparticles eforesced.ThesurfaceHONOd,whichformedviathe photocatalyticprocessofNO 2 ReactionR16,caninuence theproductionofOHd.However,themodelanalysisoriginatedfromtheintegratedreactionrateIRR,anaccumulated uxofchemicalformation,suggeststhatthecontributionof HONOdtoOHdproductionisrelativelysmallcompared tothedirectphotocatalyticprocesscausedbydustparticles showninSect.3.2.3. 5Sensitivityanduncertainties Thesensitivityofsulfatepredictiontomajorvariablese.g., temperature,humidity,sunlightprole,theconcentrationof SO 2 andNO x ,anddustloadingisillustratedinFig.5.To avoidthesuppressionofdustchemistryathighNO x levels,themostsensitivitytestswereperformedatlowlevels ofNO x .Thestackedchartnormalizedwith[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ]inFig.5 showshow[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T isattributedto[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto ,[SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] photo and[SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] aq C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] gas non-dustchemistry. www.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 12

10012Z.Yuetal.:AtmosphericMineralAerosolReactionmodel Figure3. TimeprolesoftotalsulfateconcentrationSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / intheUF-APHOR.Expdenotestheexperimentallyobservedsulfate [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.411 Td [(4 ] T / andModeldenotesthemodel-predictedsulfate.HandLrepresentthehighandthelowinitialconcentrationsofchemical species.Theerrorsassociatedwiththeconcentrationofsulfateis 10%originatedfromthePILS-ICmeasurement. a Sulfateformation withandwithoutATDparticlesSO 2 60ppbvs.SO 2 56ppbanddust290gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / b Thehighandlowloadingsofdustparticlesdust 90gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 andSO 2 100ppbvs.dust404gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 andSO 2 120ppb. c ThehighandthelowconcentrationsofSO 2 SO 2 119ppband dust239gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 vs.SO 2 272ppbanddust230gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / .For a b ,and c ,thesimulationsincludedthechamberdilutionandthewall processofgaseouscompoundsandparticlesSect.S1.For d e ,and f ,thewallprocessfortheparticlelosswasexcludedtoestimate theinuenceofATDparticlesonsulfateformationwithoutthechamberartifacts.In d e ,and f ,totalsulfatewasdecoupledintothe sulfateoriginatedfromdustchemistry[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] dust D [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] photo C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] auto / .Thepiechartsinsertedinto d e ,and f illustratehow totalsulfateisattributedtomajorpathwaysattheendoftheexperiments. Figure5aillustratesthatthereductionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T ata highertemperaturevs.298Kisascribedtothedecrease inthepartitioningprocess.Figure5bshowsthat[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T increasesbyafactorof2.8withRHincreasingfrom25to 80%.Humidityplaysanimportantroleinthemodulation ofbothaerosolacidityandliquidwatercontent,andultimatelyinuencesthepartitioningprocesse.g.,SO 2 partitioningondustsurfaceanddust-phasechemistrye.g.,productionofOHd.InthestackedcolumnchartofFig.5b, thecontributionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] dust to[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T increasesfrom 0.73to0.86withincreasingRH,suggestingthatdustchemistryismoresensitivetohumiditythanaqueous-phasechemistry.Figure5cpresents[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T attwodifferentsunlight intensitieswinteron12November2015vs.summeron 25April2017inGainesville,Floridalatitude/longitude: 29.64185 = )]TJ/F100 9.9626 Tf 7.771 0 Td [(82.347883 .AsshowninFig.5d,withSO 2 concentrationsincreasingfrom20to100ppb,[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T increasesbyafactorof4.4inthegivensimulationcondition.TheeffectoftheconcentrationofSO 2 on[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T hasbeendiscussedinSect.4.1above.Figure5eshowsthe sensitivityof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T totheATDloading,200,and 400gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / .Withtheincreasingofdustloading,thecontributionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo to[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T alsoincreases.Figure5f illustrateshowsulfateformationissuppressedbydifferent NO x levelsalsoseeSect.3.3.2. TheinorganicthermodynamicmodelE-AIMIIwas employedtoestimate[H C ]andtheliquidwatercontent M in ; water / fortheSO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 NH C 4 H 2 Osystemexcluding SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 d_salt / inReaction13ofTable3:dustphaseEq.8 inbothinorganic-saltseededaqueous-phaseanddust-phase chemistry.Theuncertaintyin M in ; water and[H C ]inuences thepartitioningtracersandconsequentlycausestheuncertaintyin[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T .Theuncertaintiesinthepredictionof[H C ] usinginorganicthermodynamicmodelsarelargebecauseof thelimiteddataCleggetal.,1998;WexlerandClegg,2002. Inthisstudy,[H C ]isestimatedbyE-AIMIICleggetal., 1998;WexlerandClegg,2002;CleggandWexler,2011and correctedfortheammonia-richconditionLietal.,2015;Li andJang,2012.Thereporteduncertaintyin[H C ]associated withtheC-RUVmethodis 18%.FigureS7illustratesthe Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 13

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10013 Figure4. Timeprolesoftotalsulfateconcentration[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] T ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / andnitrateconcentration[NO )]TJ/F100 7.5716 Tf 0 -7.411 Td [(3 ] T ,gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 / inthedual-chamber experimentsusingUF-APHORatdifferentNO x levels.TheconcentrationsofsulfateandnitrateweremeasuredusingPILS-ICduringthe experiments.Theerrorbarsoftheconcentrationofsulfateandnitrateis 10%originatedfromthePILS-ICmeasurement.Thedetailed experimentalconditionsof a b ,and c areshowninTable2.Panel d showshowtotalsulfateisattributedtoaqueous-phasereaction sulfateformationingasphase C sulfateformationininorganicsaltedinorganicaqueousphase[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] aq C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 ] gas / ,dust-phaseautoxidation[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.411 Td [(4 ] auto / ,anddustphotochemistry[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.411 Td [(4 ] photo / attheendoftheexperiments.Expdenotestheexperimentalobservationand ModeldenotesthesimulationusingtheAMARmodule.Thechamberdilutionandthewallprocessofgaseouscompoundsandparticles wereincludedinthesimulationSect.S1. uncertaintiesofthemajormodelparameters[H C ], F water K d ; SO 2 k auto and k OH ; O 2 / andthepredictionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T Theuncertaintyin F water rangesfrom )]TJ/F100 9.9626 Tf 7.771 0 Td [(20to30%dueto theuncertaintyinthemeasurementofsulfate 10%and ammoniaions 10%usingPILS-IC.ThepropagationerrorinEqs.andisusedtoestimatetheuncertainty in K d ; SO 2 )]TJ/F100 9.9626 Tf 7.771 0 Td [(20to30%.Theuncertaintyin k auto )]TJ/F100 9.9626 Tf 7.771 0 Td [(48to 42%and k OH ; O 2 )]TJ/F100 9.9626 Tf 7.771 0 Td [(20to22%areestimatedbysimulating [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T withintheuncertaintyin K d ; SO 2 andthemeasurementofsulfate 10%.Theuncertaintyin[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T was estimatedtobe 12%attheendofthesimulation.Inthis model,theHenry'slawconstantthatisappliedtogasdust partitioningoftracersSect.3.2.1mayleadtosomebiases inthepredictionofsulfateornitrate. 6Conclusionandatmosphericimplication TheAMARmodelofthisstudywasdevelopedtopredict theoxidationofSO 2 andNO x usingcomprehensivekineticmechanismsinthegasphase,inorganicseededaqueousphase,anddustphase.Thethermodynamicparameters engagedinthepartitioningprocessbetweengas,inorganic saltedaqueousaerosol,anddustphaseswereobtainedfrom knowndataintheliteratureTable3,andthekineticparametersfordustchemistrywereestimatedusingpreviously reportedindoorchamberdataParkandJang,2016.Overall, theAMARsimulationswereconsistentwithexperimentally observedoutdoorchamberdataFigs.3and4underambientsunlight.AsdiscussedinthesensitivityanalysisSect.5, boththe[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T andtherelativedistributionofmechanismbasedsulfateformationaresensitivetoallmajorvariables www.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 14

10014Z.Yuetal.:AtmosphericMineralAerosolReactionmodel Figure5. SensitivitytestofAMARmodelto a temperatureat273and298K; b RHat25,50,and80%; c sunlightprolesofsummertimeApril2017andwintertimeNovember2015atGainesville,Floridalatitude/longitude:29.64185 = )]TJ/F100 8.9664 Tf 6.994 0 Td [(82.347883 ; d the concentrationofSO 2 ; e theconcentrationofdustparticles;and f theNO x concentrationinitialNO:NO 2 D 1 V 1.Thestackedcolumn chartineachgureillustrateshowtotalsulfateisattributedtomajorpathwaysattheendofeachexperiment.Forthesensitivitytest,the chambersimulationisconductedwith100ppbofinitialSO 2 ,2ppbofinitialNO 2 ,2ppbofinitialO 3 ,and200gm )]TJ/F100 7.5716 Tf 5.906 0 Td [(3 ofATDparticles at T D 298KandRH D 40%underambientsunlighton25April2017.NO x rateofux D 2.7 10 6 ,s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / andisoprenerateofux D 2.7 10 6 ,s )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 / wereconstantlyaddedtosimulatechamberdilution.Thesimulationwasperformedwithoutconsideringtheparticlelossto thechamberwall. modelinputs,includingtemperature,humidity,sunlightintensity,thequantityofdustloading,andconcentrationsof NO x andSO 2 Inordertoassesstheimportanceofdustchemistry inambientconditions,thepredictionofsulfateformationinthepresenceofATDdustneedstobeextended to24hsimulationsundervariousenvironmentalconditions.FigureS8showstheoutputsimulatedfor24hwith 200gm 3 ofATDparticleloadingunderurbanppb NO x ;VOC = NO x < 5;20ppbSO 2 / andruralatmospheres ppbNO x ;VOC = NO x > 20;2ppbSO 2 / .Atnighttime, whenthetemperaturedropsandhumidityincreases 90%,Fig.S6,thecontributionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto to[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] T becomeslargerthanthetypicalchambersimulationduringthedaytime.Inaruralenvironment,[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo isstill themostinuentialonsulfateformation.76fractionof [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T inFig.S8a.Forthesimulationinapollutedarea Fig.S8b,thefractionof[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo to[SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T signicantlydecreases.61becauseofthesuppressioninduced byNO x Sect.3.3.2,butthefractionof[SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] auto to [SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] T increases.28.Withdecreasingsunlightintensity after5PM,Fig.S8showstherapidincreasesin[SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.595 Td [(4 ] auto duetothereactionofdust-phaseSO 2 withozone,whichis theresultofdaytimephotooxidationSect.3.3.1.FigureS8 suggeststhatthefailuretopredictsulfateformationwithout accuratedustchemistry[SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] auto C [SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.595 Td [(4 ] photo / canlead tosubstantialunderestimationofthequantityoftotalsulfateatregionalorglobalscales.SO 2 autoxidationalonemay partiallyimprovethepredictionofsulfateinthepresenceof mineraldust,butsulfateproductioncanstillbelargelyunderestimatedandincorrectlypredictedintimeserieswhenheterogeneousphotocatalyticreactionsinkineticmechanisms arenotconsidered. TheATDparticlesinthisstudyhavechemicalandphysicalpropertiesdifferentfromambientmineraldustparticles. Ingeneral,theuptakecoefcientofSO 2 inauthenticmineral dustparticlese.g.,GobiDesertdustandSaharandustis knowntobehigherthanthatofATDparticlesCrowleyetal., 2010.Thus,theeffectofambientdustparticlesonheterogeneousphotocatalyticoxidationwouldbemuchmoreimportantthanthatoftheATDparticlesofthisstudy.Toextendthe AMARmodelderivedwithATDparticletoambientdustparticles,themodelparametersrelatedtotherateconstantsand Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 15

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10015 physicalcharacteristicse.g.,surfaceareaandhygroscopic propertiesofdustparticlesneedtobemodiedusinglaboratorystudies.Photocatalyticcapacitiesofauthenticmineral dustwouldbedifferentfromATDduetothecompositions andquantityofconductivemetaloxides.Hence,thephotoactivationrateconstant k j e h inSect.3.2.3toformelectron holepairsshouldalsoberevisitedtoapplythepresentmodel todifferentmineraldustsystems.Inadditiontoreactionsof inorganicspeciesi.e.,SO 2 andNO x / ,thatoforganicspecies e.g.,HCOOH,HCHO,andCH 3 CHOondustsurfaceneeds tobeinvestigatedinthefuture. Dataavailability. Thedatasetsgeneratedduringthecurrentstudy areavailablefromthecorrespondingauthoronreasonablerequest. TheSupplementrelatedtothisarticleisavailable onlineathttps://doi.org/10.5194/acp-17-10001-2017supplement. Competinginterests. Theauthorsdeclarethattheyhavenoconict ofinterest. Acknowledgements. Thisworkwassupportedbygrantsfromthe NationalInstituteofMetrologicalScienceNIMS-2016-3100, theMinistryofScience,ICT,andFuturePlanningatSouthKorea M3C8A5032316andtheFulbrightScholarProgramfrom USAtoMongolia. Editedby:BarbaraErvens Reviewedby:twoanonymousreferees References Adams,J.W.,Rodriguez,D.,andCox,R.A.:TheuptakeofSO 2 on Saharandust:aowtubestudy,Atmos.Chem.Phys.,5,2679 2689,https://doi.org/10.5194/acp-5-2679-2005,2005. Atkinson,R.andLloyd,A.C.:Evaluationofkineticandmechanisticdataformodelingofphotochemicalsmog,J.Phys.Chem. Ref.Data,13,315,https://doi.org/10.1063/1.555710,1984. Atkinson,R.,Baulch,D.,Cox,R.,HampsonJr.,R.,Kerr,J.,Rossi, M.,andTroe,J.:Evaluatedkineticandphotochemicaldatafor atmosphericchemistry:supplementVI,IUPACsubcommitteeon gaskineticdataevaluationforatmosphericchemistry,J.Phys. Chem.Ref.Data,26,1329,1997. Baulch,D.,Cox,R.,HampsonJr.,R.,Kerr,J.,Troe,J.,andWatson, R.:Evaluatedkineticandphotochemicaldataforatmospheric chemistry:supplementII.CODATAtaskgroupongas20phase chemicalkinetics,J.Phys.Chem.Ref.Data,13,1259, https://doi.org/10.1063/1.555721,1984. Beardsley,R.,Jang,M.,Ori,B.,Im,Y.,Delcomyn,C.A.,andWitherspoon,N.:Roleofseasaltaerosolsintheformationofaromatic secondaryorganicaerosol:yieldsandhygroscopicproperties, Environ.Chem.,10,167,https://doi.org/10.1071/En13016, 2013. Beardsley,R.L.andJang,M.:SimulatingtheSOAformationof isoprenefrompartitioningandaerosolphasereactionsinthe presenceofinorganics,Atmos.Chem.Phys.,16,5993, https://doi.org/10.5194/acp-16-5993-2016,2016. Becker,K.H.,Kleffmann,J.,Kurtenbach,R.,andWiesen,P.:SolubilityofnitrousacidHONOinsulfuricacidsolutions,J.Phys. Chem.,100,14984,1996. Binkowski,F.S.andRoselle,S.J.:Models-3CommunityMultiscaleAirQualityCMAQmodelaerosolcomponent1.Modeldescription,J.Geophys.Res.-Atmos.,108, https://doi.org/10.1029/2001JD001409,2003. Bond,T.C.:Spectraldependenceofvisiblelightabsorptionbycarbonaceousparticlesemittedfromcoal combustion,Geophys.Res.Lett.,28,4075, https://doi.org/10.1029/2001gl013652,2001. Bongartz,A.,Kames,J.,Welter,F.,andSchurath,U.:Near-UVabsorptioncrosssectionsandtrans/cisequilibriumofnitrousacid, J.Phys.Chem.,95,1076,1991. Byun,D.andSchere,K.L.:Reviewofthegoverningequations, computationalalgorithms,andothercomponentsofthemodels3CommunityMultiscaleAirQualityCMAQmodelingsystem, Appl.Mech.Rev.,59,51,https://doi.org/10.1115/1.2128636, 2006. Calvert,J.G.andStockwell,W.:Mechanismandratesofthegas phaseoxidationsofsulfurdioxideandthenitrogenoxidesinthe atmosphere,OhioStateUniv.,ColumbusUSA,Dept.ofChemistry,1983. Chameides,W.L.:Thephotochemistryofaremotemarinestratiformcloud,J.Geophys.Res.-Atmos.,89,4739,1984. Chang,R.Y.W.,Sullivan,R.C.,andAbbatt,J.P.D.:InitialuptakeofozoneonSaharandustatatmospheric relativehumidities,Geophys.Res.Lett.,32,L14815, https://doi.org/10.1029/2005gl023317,2005. Chen,H.H.,Nanayakkara,C.E.,andGrassian,V.H.:Titanium DioxidePhotocatalysisinAtmosphericChemistry,Chem.Rev., 112,5919,https://doi.org/10.1021/cr3002092,2012. Clegg,S.andWexler,A.S.:DensitiesandApparentMolarVolumes ofAtmosphericallyImportantElectrolyteSolutions.2.TheSystemsH C -HSO )]TJ/F100 7.5716 Tf 0 -7.411 Td [(4 -SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.411 Td [(4 -H 2 Ofrom0to3molkg )]TJ/F100 7.5716 Tf 5.905 0 Td [(1 asaFunctionofTemperatureandH C -NH C 4 -HSO )]TJ/F100 7.5716 Tf 0 -7.412 Td [(4 -SO 2 )]TJ/F100 7.5716 Tf -3.785 -7.412 Td [(4 -H 2 Ofrom0 to6mol,kg )]TJ/F100 7.5716 Tf 5.906 0 Td [(1 at25 CUsingaPitzerIonInteractionModel,and NH 4 HSO 4 -H 2 OandNH 4 / 3 HSO 4 / 2 -H2OovertheEntire ConcentrationRange,J.Phys.Chem.A,115,3461,2011. Clegg,S.L.,Brimblecombe,P.,andWexler,A.S.:Thermodynamic modelofthesystemH C -NH C 4 -SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 -NO )]TJ/F100 7.5716 Tf 0 -7.412 Td [(3 -H 2 Oattropospheric temperatures,J.Phys.Chem.A,102,2137,1998. Colberg,C.A.,Luo,B.P.,Wernli,H.,Koop,T.,andPeter,Th.: Anovelmodeltopredictthephysicalstateofatmospheric H 2 SO 4 /NH 3 /H 2 Oaerosolparticles,Atmos.Chem.Phys.,3, 909,https://doi.org/10.5194/acp-3-909-2003,2003. Colmenares,J.C.andLuque,R.:Heterogeneousphotocatalytic nanomaterials:prospectsandchallengesinselectivetransformationsofbiomass-derivedcompounds,Chem.Soc.Rev.,43,765 778,https://doi.org/10.1039/c3cs60262a,2014. Crowley,J.N.,Ammann,M.,Cox,R.A.,Hynes,R.G.,Jenkin, M.E.,Mellouki,A.,Rossi,M.J.,Troe,J.,andWallington,T.J.:Evaluatedkineticandphotochemicaldataforatwww.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017

PAGE 16

10016Z.Yuetal.:AtmosphericMineralAerosolReactionmodel mosphericchemistry:VolumeVheterogeneousreactions onsolidsubstrates,Atmos.Chem.Phys.,10,9059, https://doi.org/10.5194/acp-10-9059-2010,2010. Cwiertny,D.M.,Young,M.A.,andGrassian,V. H.:Chemistryandphotochemistryofmineraldust aerosol,Annu.Rev.Phys.Chem.,59,27, https://doi.org/10.1146/annurev.physchem.59.032607.093630, 2008. Dong,X.,Fu,J.S.,Huang,K.,Tong,D.,andZhuang,G.:Modeldevelopmentofdustemissionandheterogeneouschemistrywithin theCommunityMultiscaleAirQualitymodelingsystemandits applicationoverEastAsia,Atmos.Chem.Phys.,16,8157, https://doi.org/10.5194/acp-16-8157-2016,2016. Dupart,Y.,King,S.M.,Nekat,B.,Nowak,A.,Wiedensohler,A.,Herrmann,H.,David,G.,Thomas,B.,Miffre, A.,Rairoux,P.,D'Anna,B.,andGeorge,C.:Mineral dustphotochemistryinducesnucleationeventsinthepresenceofSO 2 ,P.Natl.Acad.Sci.USA,109,20842, https://doi.org/10.1073/pnas.1212297109,2012. Dupart,Y.,Fine,L.,D'Anna,B.,andGeorge,C.:HeterogeneousuptakeofNO 2 onArizonaTestDustunderUV-Airradiation:Anaerosolowtubestudy,AeolianRes.,15,45, https://doi.org/10.1016/j.aeolia.2013.10.001,2014. Gankanda,A.,Coddens,E.M.,Zhang,Y.,Cwiertny,D.M.,and Grassian,V.H.:Sulfateformationcatalyzedbycoalyash,mineraldustandironiiioxide:variableinuenceoftemperature andlight,Environ.Sci.,18,1484,2016. Goodman,A.L.,Li,P.,Usher,C.R.,andGrassian,V.H.:Heterogeneousuptakeofsulfurdioxideonaluminumandmagnesiumoxideparticles,J.Phys.Chem.A,105,6109, https://doi.org/10.1021/jp004423z,2001. Graedel,T.:Thehomogeneouschemistryofatmosphericsulfur, Rev.Geophys.,15,421,1977. Gustafsson,R.J.,Orlov,A.,Badger,C.L.,Grifths,P.T., Cox,R.A.,andLambert,R.M.:Acomprehensiveevaluation ofwateruptakeonatmosphericallyrelevantmineralsurfaces: DRIFTspectroscopy,thermogravimetricanalysisandaerosol growthmeasurements,Atmos.Chem.Phys.,5,3415, https://doi.org/10.5194/acp-5-3415-2005,2005. Hoffmann,M.R.,Martin,S.T.,Choi,W.Y.,andBahnemann,D.W.:EnvironmentalApplicationsofSemiconductorPhotocatalysis,Chem.Rev.,95,69, https://doi.org/10.1021/cr00033a004,1995. Hoyle,C.R.,Fuchs,C.,Jrvinen,E.,Saathoff,H.,Dias,A.,El Haddad,I.,Gysel,M.,Coburn,S.C.,Trstl,J.,Bernhammer, A.-K.,Bianchi,F.,Breitenlechner,M.,Corbin,J.C.,Craven,J., Donahue,N.M.,Duplissy,J.,Ehrhart,S.,Frege,C.,Gordon,H., Hppel,N.,Heinritzi,M.,Kristensen,T.B.,Molteni,U.,Nichman,L.,Pinterich,T.,Prvt,A.S.H.,Simon,M.,Slowik,J.G., Steiner,G.,Tom,A.,Vogel,A.L.,Volkamer,R.,Wagner,A.C., Wagner,R.,Wexler,A.S.,Williamson,C.,Winkler,P.M.,Yan, C.,Amorim,A.,Dommen,J.,Curtius,J.,Gallagher,M.W.,Flagan,R.C.,Hansel,A.,Kirkby,J.,Kulmala,M.,Mhler,O.,Stratmann,F.,Worsnop,D.R.,andBaltensperger,U.:Aqueousphase oxidationofsulphurdioxidebyozoneinclouddroplets,Atmos.Chem.Phys.,16,1693,https://doi.org/10.5194/acp16-1693-2016,2016. Huang,L.B.,Zhao,Y.,Li,H.,andChen,Z.M.:KineticsofHeterogeneousReactionofSulfurDioxideonAuthenticMineralDust:EffectsofRelativeHumidityandHydrogenPeroxide,Environ.Sci.Technol.,49,10797, https://doi.org/10.1021/acs.est.5b03930,2015. Jang,J.,Jang,M.,Mui,W.,Delcomyn,C.A.,Henley,M.V.,and Hearn,J.D.:Formationofactivechlorineoxidantsinsalineoxoneaerosol,AerosolSci.Tech.,44,1018,2010. Jeffries,H.E.,Gary,M.W.,Kessler,M.,andSexton,K.G.:Morpheculereactionmechanism,MORPHO,ALLOMORPHICsimulationsoftware,1998. Kerr,J.andCalvert,J.:ChemicalTransformationModulesforEulerianAcidDecompositionModels,VolumeI:Thegasphase chemistry,Boulder,Colorado,1984. Krueger,B.J.,Grassian,V.H.,Laskin,A.,andCowin,J.P.: Thetransformationofsolidatmosphericparticlesintoliquiddropletsthroughheterogeneouschemistry:Laboratoryinsightsintotheprocessingofcalciumcontainingmineraldust aerosolinthetroposphere,Geophys.Res.Lett.,30,1148, https://doi.org/10.1029/2002gl016563,2003. Li,J.andJang,M.:Aerosolaciditymeasurementusingcolorimetry coupledwithareectanceUV-visiblespectrometer,Aerosol.Sci. Tech.,46,833,2012. Li,J.,Jang,M.,andBeardsley,R.L.:Dialkylsulfateformationin sulfuricacid-seededsecondaryorganicaerosolproducedusing anoutdoorchamberundernaturalsunlight,Environ.Chem.,13, 590,https://doi.org/10.1071/EN15129,2015. Li,J.W.andHan,Z.W.:Amodelingstudyoftheimpactof heterogeneousreactionsonmineralaerosolsurfacesontroposphericchemistryoverEastAsia,Particuology,8,433, https://doi.org/10.1016/j.partic.2010.03.018,2010. Liang,J.Y.andJacobson,M.Z.:Astudyofsulfurdioxideoxidationpathwaysoverarangeofliquidwatercontents,pHvalues, andtemperatures,JGeophysRes-Atmos,104,13749-13769,Doi 10.1029/1999jd900097,1999. Linsebigler,A.L.,Lu,G.Q.,andYates,J.T.:Photocatalysison Tio2SurfacesPrinciples,Mechanisms,andSelectedResults, Chem.Rev.,95,735,https://doi.org/10.1021/cr00035a013, 1995. Liu,C.,Ma,Q.X.,Liu,Y.C.,Ma,J.Z.,andHe,H.:Synergistic reactionbetweenSO 2 andNO 2 onmineraloxides:apotential formationpathwayofsulfateaerosol,Phys.Chem.Chem.Phys., 14,1668,https://doi.org/10.1039/c1cp22217a,2012. Liu,Y.J.,Zhu,T.,Zhao,D.F.,andZhang,Z.F.:Investigationof thehygroscopicpropertiesofCaNO 3 / 2 andinternallymixed CaNO 3 / 2 /CaCO 3 particlesbymicro-Ramanspectrometry,Atmos.Chem.Phys.,8,7205,https://doi.org/10.5194/acp-87205-2008,2008. Ma,Q.X.,Liu,Y.C.,andHe,H.:SynergisticeffectbetweenNO 2 andSO 2 intheiradsorptionandreaction ongamma-alumina,J.Phys.Chem.A.,112,6630, https://doi.org/10.1021/jp802025z,2008. Martell,A.E.andSmith,R.M.:Inorganiccomplexes,Plenum Press,1976. McNaught,A.D.andWilkinson,A.:IUPAC,Compendiumof ChemicalTerminology,theGoldBook,WileyBlackwell,2nd revisedEdn.,1997. Michel,A.E.,Usher,C.R.,andGrassian,V.H.:Reactiveuptake ofozoneonmineraloxidesandmineraldusts,Atmos.Environ., 37,3201,https://doi.org/10.1016/S1352-23104,2003. Atmos.Chem.Phys.,17,1000110017,2017www.atmos-chem-phys.net/17/10001/2017/

PAGE 17

Z.Yuetal.:AtmosphericMineralAerosolReactionmodel10017 Nanayakkara,C.E.,Pettibone,J.,andGrassian,V.H.:Sulfurdioxideadsorptionandphotooxidationonisotopically-labeledtitaniumdioxidenanoparticlesurfaces:rolesofsurfacehydroxyl groupsandadsorbedwaterintheformationandstabilityofadsorbedsulteandsulfate,Phys.Chem.Chem.Phys.,14,6957 6966,https://doi.org/10.1039/c2cp23684b,2012. Navea,J.G.,Chen,H.H.,Huang,M.,Carmichel,G.R.,and Grassian,V.H.:Acomparativeevaluationofwateruptake onseveralmineraldustsources,Environ.Chem.,7,162, https://doi.org/10.1071/En09122,2010. Park,J.andJang,M.:Heterogeneousphotooxidationofsulfurdioxideinthepresenceofairbornemineraldustparticles,RSCAdvances,6,58617,2016. Reyes-Coronado,D.,Rodrguez-Gattorno,G.,Espinosa-Pesqueira, M.,Cab,C.,DeCoss,R.,andOskam,G.:Phase-pureTiO 2 nanoparticles:anatase,brookiteandrutile,Nanotechnology,19, 145605,https://doi.org/10.1088/0957-4484/19/14/145605,2008. Saliba,N.A.andChamseddine,A.:Uptakeofacid pollutantsbymineraldustandtheireffecton aerosolsolubility,Atmos.Environ.,46,256, https://doi.org/10.1016/j.atmosenv.2011.09.074,2012. Saliba,N.A.,Moussa,S.G.,andElTayyar,G.:Contributionof airbornedustparticlestoHONOsources,Atmos.Chem.Phys. Discuss.,14,4827,https://doi.org/10.5194/acpd-14-48272014,2014. Sarwar,G.,Fahey,K.,Kwok,R.,Gilliam,R.C.,Roselle,S.J., Mathur,R.,Xue,J.,Yu,J.,andCarter,W.P.:Potentialimpacts oftwoSO 2 oxidationpathwaysonregionalsulfateconcentrations:aqueous-phaseoxidationbyNO 2 andgas-phaseoxidation byStabilizedCriegeeIntermediates,Atmos.Environ.,68,186 197,2013. Sarwar,G.,Simon,H.,Fahey,K.,Mathur,R.,Goliff,W.S.,and Stockwell,W.R.:ImpactofsulfurdioxideoxidationbyStabilizedCriegeeIntermediateonsulfate,Atmos.Environ.,85,204 214,2014. Schwartz,S.andWhite,W.:Solubilityequilibriaofthenitrogen oxidesandoxyacidsindiluteaqueoussolution,Adv.Environ. Sci.Eng.,UnitedStates,4,1,1981. Schwartz,S.E.:Gas-PhaseandAqueous-PhaseChemistryofHO 2 inLiquidWaterClouds,J.Geophys.Res.-Atmos.,89,1589 1598,https://doi.org/10.1029/JD089iD07p11589,1984. Shang,J.,Li,J.,andZhu,T.:Heterogeneousreactionof SO 2 onTiO 2 particles,Sci.ChinaChem.,53,2637, dio:10.1007/s11426-010-4160-3,2010. Tang,M.J.,Cziczo,D.J.,andGrassian,V.H.:InteractionsofWaterwithMineralDustAerosol:WaterAdsorption,Hygroscopicity,CloudCondensation,andIceNucleation,Chem.Rev.,116, 4205,https://doi.org/10.1021/acs.chemrev.5b00529,2016. Tang,Y.H.,Carmichael,G.R.,Kurata,G.,Uno,I.,Weber,R.J., Song,C.H.,Guttikunda,S.K.,Woo,J.H.,Streets,D.G.,Wei, C.,Clarke,A.D.,Huebert,B.,andAnderson,T.L.:Impactsof dustonregionaltroposphericchemistryduringtheACE-Asiaexperiment:Amodelstudywithobservations,J.Geophys.Res.Atmos.,109,D19s21,https://doi.org/10.1029/2003jd003806, 2004. Thiebaud,J.,Thevenet,F.,andFittschen,C.:OHRadicalsand H 2 O 2 MoleculesintheGasPhaseneartoTiO 2 Surfaces,J.Phys. Chem.C,114,3082,https://doi.org/10.1021/jp9102542, 2010. Thompson,T.L.andYates,J.T.:SurfacesciencestudiesofthephotoactivationofTiO 2 newphotochemicalprocesses,Chem.Rev., 106,4428,2006. Ullerstam,M.,Vogt,R.,Langer,S.,andLjungstrom,E.: ThekineticsandmechanismofSO 2 oxidationbyO 3 on mineraldust,Phys.Chem.Chem.Phys.,4,4694, https://doi.org/10.1039/b203529b,2002. Underwood,G.M.,Song,C.H.,Phadnis,M.,Carmichael,G. R.,andGrassian,V.H.:HeterogeneousreactionsofNO 2 and HNO 3 onoxidesandmineraldust:Acombinedlaboratoryand modelingstudy,J.Geophys.Res.-Atmos.,106,18055, https://doi.org/10.1029/2000jd900552,2001. Usher,C.R.,Al-Hosney,H.,Carlos-Cuellar,S.,andGrassian,V. H.:Alaboratorystudyoftheheterogeneousuptakeandoxidationofsulfurdioxideonmineraldustparticles,J.Geophys.Res.Atmos.,107,4713,https://doi.org/10.1029/2002jd002051,2002. Usher,C.R.,Michel,A.E.,Stec,D.,andGrassian, V.H.:Laboratorystudiesofozoneuptakeonprocessedmineraldust,Atmos.Environ.,37,5337, https://doi.org/10.1016/j.atmosenv.2003.09.014,2003. Vlasenko,A.,Sjogren,S.,Weingartner,E.,Stemmler,K.,Gggeler, H.W.,andAmmann,M.:Effectofhumidityonnitricaciduptake tomineraldustaerosolparticles,Atmos.Chem.Phys.,6,2147 2160,https://doi.org/10.5194/acp-6-2147-2006,2006. Wagner,C.,Schuster,G.,andCrowley,J.:Anaerosolowtube studyoftheinteractionofN 2 O 5 withcalcite,Arizonadustand quartz,Atmos.Environ.,43,5001,2009. Wagner,R.,Ajtai,T.,Kandler,K.,Lieke,K.,Linke,C.,Mller, T.,Schnaiter,M.,andVragel,M.:ComplexrefractiveindicesofSaharandustsamplesatvisibleandnearUVwavelengths:alaboratorystudy,Atmos.Chem.Phys.,12,2491, https://doi.org/10.5194/acp-12-2491-2012,2012. Wexler,A.S.andClegg,S.L.:AtmosphericaerosolmodelsforsystemsincludingtheionsH C ,NH C 4 ,Na C ,SO 2 )]TJ/F100 7.5716 Tf -3.786 -7.412 Td [(4 NO )]TJ/F100 7.5716 Tf 0 -7.412 Td [(3 ,Cl )]TJ/F100 8.9664 Tf 6.404 -3.255 Td [(,Br )]TJ/F100 8.9664 Tf 6.404 -3.255 Td [(,andH 2 O,J.Geophys.Res.-Atmos.,107, https://doi.org/10.1029/2001JD000451,2002. Zhang,X.Y.,Zhuang,G.S.,Chen,J.M.,Wang,Y.,Wang,X., An,Z.S.,andZhang,P.:Heterogeneousreactionsofsulfurdioxideontypicalmineralparticles,J.Phys.Chem.B,110,12588 12596,https://doi.org/10.1021/jp0617773,2006. Zhang,Y.-H.andChan,C.K.:UnderstandingthehygroscopicpropertiesofsupersaturateddropletsofmetalandammoniumsulfatesolutionsusingRamanspectroscopy,J.Phys.Chem.A,106, 285,2002. Zhong,M.andJang,M.:LightabsorptioncoefcientmeasurementofSOAusingaUV-Visiblespectrometerconnected withanintegratingsphere,Atmos.Environ.,45,4263, https://doi.org/10.1016/j.atmosenv.2011.04.082,2011. Zuend,A.,Marcolli,C.,Booth,A.M.,Lienhard,D.M.,Soonsin,V.,Krieger,U.K.,Topping,D.O.,McFiggans,G.,Peter,T.,andSeinfeld,J.H.:Newandextendedparameterization ofthethermodynamicmodelAIOMFAC:calculationofactivitycoefcientsfororganic-inorganicmixturescontainingcarboxyl,hydroxyl,carbonyl,ether,ester,alkenyl,alkyl,andaromaticfunctionalgroups,Atmos.Chem.Phys.,11,9155, https://doi.org/10.5194/acp-11-9155-2011,2011. www.atmos-chem-phys.net/17/10001/2017/Atmos.Chem.Phys.,17,1000110017,2017