Heterogeneous reactions of Cl2 with sea salts at ... - Wiley Online Library

12 downloads 0 Views 422KB Size Report
Nov 1, 1998 - tions of CI 2 with solid pure NaBr and KBr: 1/2C12(g ) + MBr(s). --> MCI(s) + 1/2Br2(g) (M = Na, K), have been investigated. These results show ...
GEOPHYSICAL RESEARCH LETTERS, VOL.25,NO.21,PAGES3927-3930, NOVEMBER1, 1998

Heterogeneousreactions of with sea salts at ambient temperature: Implications for halogen exchangein the atmosphere M. Mochida,J. Hirokawa,Y. Kajii, and H. Akimoto Research Centerfor Advanced ScienceandTechnology, The Universityof Tokyo,Japan

Abstract. Laboratorymeasurements of heterogeneous reac- 1998].In addition,recentstudieshaverevealedthe possibiltionsof C12(g ) on solid sea saltsat ambienttemperatureand ity of C12formationby the heterogeneous reactionof ozone low relativehumidityhavebeencarriedout usinga Knudsen with deliquesced saltaerosol[Oumet al., 1998],andBr2forcell reactor. Syntheticsea salt and commercialnatural sea salt mationwith deliquesced NaBr [Hirokawaet al., 1998]. are usedto representseasalt particles.The uptakeprobabiliIn orderto understand the halogencyclesin the tropo-

tiesfor synthetic seasalt(7 = (2.2+ 0.3)x 10'2) andfor natural sphere, it is important to take into accountthe simultaneous salt(7 = (3.1+ 1.1)x 10'2)havebeenobtained andBr2 isob- cyclesof differenthalogens[Sanderand Crutzen, 1996]. servedas the predominantproduct.The heterogeneous reac- There are a few previousresultsregardingheterogeneous reactionsof CI2 with solidpureNaBr andKBr: 1/2C12(g ) + MBr(s) tions of halogen exchange,i.e. conversionof one halogen --> MCI(s) + 1/2Br2(g)(M = Na, K), have beeninvestigated. molecule(e.g. C12)to another(e.g. Br2) [Berkoet al., 1991; Hu Theseresultsshow that syntheticand naturalsaltshave a sim-

et al., 1995]. These conversionsare importantsince CI and Br

ilar extentof reactivitycomparedwith solidpure bromidesin containing specieshave different photoactive features in the spiteof theirsmallBr fraction.Thesereactions are expectedto' troposphere.The C1 atom is expectedto contributemuch more be an importantC12sinkandBr2 sourcein the marinebound- than the Br atom to the oxidation of hydrocarbons,which ary layer.

lowers ozone destructionefficiency by forming inactive HC1. In contrast, the Br atom has no reactivity with alkanes and thus contributesto ozone destructionmore effectively. Using Introduction a train droplet technique,Hu et al. [1995] studiedreactivities of Cl2 and Br2 with NaBr and NaI solutionsand reported the Inorganic halogen-containingcompoundsare regarded as existenceof an interestinginterfacialreactionbetweenthe gas important speciesin ozone destructionand in oxidation cyand liquid phasewhich can not be explainedby conventional cles of organicsin the atmosphere.C12is one of many inorbulk liquid phase reaction. ganic halogen compoundsin the atmosphere.The active C1 In this paper we investigatedthe reactivity of Cl2 with atom,whichis formedfrom C12duringday time by photolysis, solid synthetic sea salt and natural sea salt using a Knudsen along with the CIO radical has been shown to contribute to reactionon ozone destructionin the stratosphereby catalytic cycles with cell reactor in order to demonstrateheterogeneous sea salt aerosols in the troposphere.In order to clarify the and without the existence of polar stratospheric clouds chemicalprocessesof this reaction,the uptake probabilitiesof [Solomon, 1990]. In contrast to halogen chemistry in the stratosphere,however, little is known yet about halogen cy- the heterogeneousreactions of Cl 2 with pure solid NaBr and KBr, 1/2C12(g ) + MBr(s)-->MCI(s)+ 1/2Br2(g ) (M = Nal K), cles in the troposphere[Graedel and Keene, 1995]. were also investigated. The impact of this reaction on deliReflectingthe recentinterestin tropospherichalogenchemquesced sea salt aerosols in the marine boundary layer (MBL) istry, two research groups have been measuring inorganic was estimated assuming that the reactivities on solid salts halogen species by different methods. Using a tandem mist give a lower limit to thoseon deliquescedsea salt aerosols. chambermethod, C12', includingCl2 and HOCI not trappedin the acidic mist chamber,has been measuredin the range < 26 to 254 pptv CI in the marine atmosphere[Keene et al.. 1993: Experimental Pszenny et al., 1993]. Using a photoactive halogen detector All' experimentswere conductedusing a Teflon coated with scavengerhydrocarbonmolecules, photolytically active Knudsen cell reactor coupled with a differentially pumped moleculesClp (includingC12,HOCI, CINO, C1NO 2, and quadrupolemass spectrometer(ULVAC MSQ-400). This inC1ONO2) have been detectedin the range < 9 to 100 pptv as strument is basically the same as other instrumentsdescribed C12in the CanadianArctic at Alert [Impeyet al., 1997a,b]. elsewhere [Caloz eta/., 1997; Fenter et al., 1997; Mochida et Regarding halogen sourcesin the troposphere,other studal., 1998]. The reactantgas,C12,was injectedthroughan inlet ies have suggestedthat C12 and other halogen containing capillary into the Knudsen cell reactor where C12 collides compounds are released to the atmosphere from sea salt with a solid salt surface and/or escapesthrough a small exit aerosols via heterogeneous reactions of variousNOy species orifice. Part of the effusedC12moleculesand gaseousproducts [e.g. Finlayson-Pitts et al., 1989] or HOBr [Mochida et al., to the differentialchamberwere choppedby a rotationalchopper and then detectedby the quadrupolemass spectrometer Copyright1998by theAmericanGeophysical Union. (MS). The modulatedsignalwas amplified by a lock-in amplifier (Stanford Research Systems SR830) and recorded by a PapernumberGRL-1998900100. 0094-8276/98/GRL-

1998900100505.00

computer.

3927

3928

MOCHIDA

ET AL.' HETEROGENEOUS

volume

estimatedsurfacearea (total) surfacearea (sample) gasnumberdensity

OF CL2 WITH SEA SALTS

Results

Table 1. Knudsen Cell Parameters reactor parameter

REACTIONS

value

2050 cm3 1600 cm•19.6 cm•-

Figure1 showsan exampleof reactiveuptakemeasurements of Cl2 with syntheticsea salt. The isolationplungerwhich isolates the samplesaltsubstrates fromthe Cl2 gaswaslifted during the period indicatedas "ReactionOn." As is seen in

Figure1, onliftingtheplunger, theMS signalof Cl2+ at m/e70 whichis proportional to the numberdensityof Cl2 in thereacsample collision frequency b co= 1.8Ah(T/M)i/•s-i tor decreased instantaneously becauseof the continuous upescape rateconstant (• = 8mm)c 0.48x (T/M)•/2s.• takeof Cl2 on theNaBr surface.The observed gradualdecrease aCalculated using therelation F i= Vkesc[M], where F i istheflowof of uptakewasdueto surfacesaturation.The rapiduptakewas by the appearance of Br2, whichwasthe predommolecules, Vthereactor volume, and[M]thenumber density. b•/h,the accompanied sample surface area,whichis 19.6cm 2 forthisexperiment. • Values inant productof this heterogeneous reaction.A slight indetermineddirectlyby experiment. crease of MS signalof BrCl+ at m/e 116wasalsoobserved (not shown in Figure 1). The appearanceof iodine containing specieswasnot observed. Figure2(a) showsthe uptakeprobThe first-orderescaperate constantof the reactorkes c, was ability of Cl2 on syntheticseasalt (solid symbols)and natural determinedfrom the exponentialdecayof escapeflow of Cl2 seasalt(opensymbols) asa functionof injectedCl2 flow rate. when the continuousflow into the reactorwas stopped.This As is shownin Figure 2(a), the uptakeprobabilitydid not Knudsen cell reactor was equippedwith a sample isolation showany dependence on the reactantflow. The uptakeproba(1-100)x 10løcm'3a

plungerso that the exposure/isolation of Cl2 to the salt surface bilitieswerecalculated to be 7 = (2.2+ 0.3)x 10-2and(3.1 +_ could be controlled without changingthe CI2 flow into the 1.1)x 10'2forthesynthetic andnatural seasalt,respectively. reactor. Characteristicparametersof the Knudsencell reactor In orderto get insightinto the expectedchemicalconverare presentedin Table 1.

As a reactant,commerciallyavailableCI2 (Nihon Sanso> 99.8%) was used without further purification.The synthetic sea salt was a commercialproduct,"Aqua Ocean"(JapanPet Drugs) whosecomponentsare claimedto be sameas seawater. The commercialnaturalseasalt was a seasoning product"Sel de Gu6rande,"producedin France.Salt samplesof pure NaBr and KBr were preparedin two ways, as powdersubstrates and as spray-depositedsubstrates.The powder substrateswere milled salt with averagediametersbetween10 to 100 microns, whereasspray-deposited substrateswere preparedby spraying NaC1 solution in methanolusing an atomizer on glass platesheatedto a temperatureof 420K resultingin thin salt layers. Scanning electron microscope measurementshave shown that these thin salt films are coherent and that the dif-

ferencebetweenthe real surfacearea andthe macroscopic glass surfacearea is at most a factor of two [Fenter et al., 1996]. Thus we usedthe geometricglass surfacearea as the real salt surfacearea when calculatinguptake probability 7, according to the equations,kun i = (Si / Sf- 1) kescand y= kun i / (_O. Here, Si and Sf are MS signalsof reactantgas before and during the heterogeneousreaction, and to is the collision frequencyto the total surfacearea derivedfrom gaskinetictheory. When using the powder substrates, we used a porous model [Keyser and Leu, 1993] in order to take into account the diffusion of reactantgas into internal voids. The observed mass dependenceof the Cl2 initial uptakeon the numberof NaBr monodispersegrain layers can be fitted by this model usingthe true densityPt., the volumetricallymeasuredapparent bulk density Pb of salt grains and the tortuosityfactor ß fixed at a value of 2.0 [Fenter et al., 1996]. Although this conversioncould translatesome systematicerrors into calculated values, the agreementbetween 7 calculatedfor powder substratesand for spray-depositedsubstratesindicates that this error was smallerthan a randomerror due to the MS signal noise and the reproducibilityof individual salt samples.For synthetic sea salt and natural salt, only powder substrates were used in the experiments.All the uptake probabilitiesin this paper are initial uptake probabilities on salt samples which were exposedto Cl2 for the first time, andall errorlimits given in this paper are one standarddeviation.

sionof Cl2to Br2 onseasalt,theuptakeprobabilities of Cl2 on pureNaBr andKBr werealsomeasured. The heterogeneous reactioninvolvedis expressedas:

1/2C12(g ) + MBr(s)-->MCI(s)+ 1/2Br2(g ) (M = Na, K). (1)

Calibration of the MS signalsof C12 + (m/e=70)and Br•_ + (m/e--160)revealedthat the Br2 yield of (92 + 8) % wasindependentof the exposuretime. A slight increaseof BrC1+ (m/e=116) was also observed.As BrCI is thoughtto react with bromides,some portion of BrC1 thus releasedfrom the surfacemaycollideagainwiththe surfaceto forma partof the observedBr2 as follows.

C12(g ) + MBr(s) --->MCI(s) + BrCl(g) (M = Na, K)

(2)

BrCl(g)+ MBr(s)--->MCI(s) + Br2(g) (M = Na, K)

(3)

The nearunityyield of Br2 meansthatthe accumulation of reactantmoleculeson the surfacewas negligible,thus the observeddisappearance of Cl2 in Figure 1 was solelydueto the

5x10TM C12(m/e=70) •

Reaction On

-

/•Br 2(m/e=160) !

0

50

100

150

200

250

300

350

Time (s)

Figure 1. Steady-state experimentfor a typical C12uptakeon syntheticseasalt powder.Escapeflow rate of CI2 is monitored at m/e 70, and Br2 at m/e 160 in the 8mm-orificereactor.BrCI is also observedas a small fractionbut not shownhere. Injected

Cl2 flow intothe reactoris keptat 4.2 x 10TMmolecules/s whereasescapeCl2 flow decreasesduring the reactiondue to the uptakeof Cl2 on the saltsurface.

MOCHIDA

ET AL.' HETEROGENEOUS

REACTIONS

heterogeneous replacementreactionand not to adsorptionon

OF CL2 WITH

SEA SALTS

3929

0.07

the sea salt surface.

(a) _

0.06

Figure 2 (b) showsthe uptakeprobabilityof the reactionof C12with pureNaBr as a functionof flow rate of CI2. Initial uptake probability was independentof C12 flow rate, i.e. C12 number density in the reactor. Using spray-depositedsub-

0.05

0.04

[]

[]

0.03

strates, uptakeprobabilities of y= (2.4__1.3)x 10-2forNaBr andy= (2.3_+0.7)x l0-2forKBrwereobtained. Thelargerun-

0.02

certaintyof y obtained using spray-depositedsubstrateswas due to the irreproducibility inherent in the sample surface preparation.The uptake probabilitiesfor powder substrates

i n

OOl o.oo

werey= (2.1_+0.4)x l0'2andy= (3.7__0.7)x l0'2 forNaBr

• } •'•

'

i

'

i

i

i

i i i I

« 3 4 5 678

10TM

and KBr, respectively.Thus, it was confirmedthat the uptake probability of the halogenexchangereaction is independent

1015

C12Flow (molecules/s)

of counter cation and the form of substrate. We also checked

the reactivityof CI2 with NaCl and KCI, and observedthat the uptake probability was one order of magnitudesmaller than that of bromides.Sinceno productwas observedand the uptake rapidly saturated,this suggeststhat the uptake was due to adsorptionon the salt surface.The uptake probabilitiesobtainedin this studyare summarizedin Table 2.

0.07 /. 0.06• 0.05}-

(b)' t

0.04 -

0.03 ß

Discussion

0.02-

ß

ß ß

ß

ß A

ß

One hypothesisto explain the mechanismof this heteroge0.01 neousreactionis that it proceedsin surfaceadsorbedwater as , , , , , ill i I i i , , i •l 0.00 ' proposedby Beichert and Finlayson-Pitts [1996] in the case 2 3 456 2 3 456 2 of the heterogeneous reactionof HNO3 with solid NaC1. Our 10TM 1015 previous experimentsinvolving H20 vapor [Mochida et al., C12Flow (molecules/s) 1998] have shown that at least on the order of tens of monolayers of H20 was stronglyadsorbedon salt substratesin our Figure 2. The initial uptakeprobabilityof C12reactingon (a) experimentalconditions.Peters and Ewing [1996] proposed synthetic sea salt powder (solid squares) and natural salt that heterogeneousreactionson dry alkali halides occur on powder (open squares),and on (b) NaBr powder substrates (open triangles) and NaBr spray-depositedsubstrates(solid defects,but consideringthe existenceof surfaceadsorbedwatriangles) as a function of injectedCI2 flow rate in the 8ramter in our cases,we proposethat CI2 reactedwith Br' in the orifice reactor. Uptake probability on powder substratesare surface adsorbed water as an ionic reaction. If this surface adobtainedby calculatingthe correctionfactor accordingto the sorbed water was in the form of a saturated solution of both pore diffusionmodel [Keysetand Leu, 1993]. chloride and bromide,the ratio [Br-]/[CI'], which is 1/660 for sea water, would insteadexceedunity. Even if [Br'] in surface adsorbed water is not saturated,there could be some extent of

fractionationof the ratio [Br']/[C1-]. Hu et al. [1995] showeda positive dependenceof C12uptakeprobabilityon Br' concentration in the liquid phase.Thus this fractionationcan explain the similar reactivityof C12with syntheticsea salt and natural salt to those with pure bromides.However, anotherexplanation involveshigher reactivityof syntheticsea salt compared

with pure alkali halides as reported in the reaction with HNO3 andNO2 [De Haan and Finlayson-Pitts,1997]. They assumedthat the hydrateof MgCI2 contributed to the higher reactivity.Suchcomponents might haveenhancedthe reactivity in our casesalso. It is unknownwhich mechanism,surface

adsorbed wateror otherreactivecomponents governsthe high reactivityof C12on syntheticand naturalseasalt. But in either

case,thereare importantimplications for atmospheric condi-

Table 2. ExperimentalResults

tions.

substrate

Yobs'

synthetic sea-salt c

0.138+ 0.036

(2.2+ 0.3)x 10'2

natural sea-saltc

0.110 + 0.011

•3•