JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. D13, PAGES 15,999-16,004,JULY 20, 1997

Measurement techniquefor the determination of photolyzable chlorine and bromine in the atmosphere G. A. Impey,P. B. Shepson,• D. R. Hastie,L. A. Bartie• Departmentof Chemistryand Centre for AtmosphericChemistry, York University,Toronto, Ontario, Canada

Abstract. A techniquehas been developed to enablemeasurement of photolyzablechlorine and bromineat tracelevels in thetroposphere. In thismethod, ambient air is drawnt•ough a cylindricalflow cell, whichis irradiatedwith a Xe arc lamp. In the reactionvessel of the photoactivehalogen detector (PHD), photolyrically active molecules Clp (including C12, HOC1, C1NO,C1NO2, and C1ONO2) and Brp (including Br2, HOBr, BrNO, BrNO2, and BrONO2) are photolyzed,and the halogenatoms produced react with properieto form stablehalogenated products.These products are thensampled and subsequently separated and detected by gas chromatography.The systemis calibratedusing low concentrationmixtures of C12and Br2 in air from commerciallyavailable permeation sources. We obtaineddetection limits of 4 pptv and 9 pptv as Br2 andC12, respectively, for 36 L samples.

1. Introduction (or C12)in the Arctic, largely as a result of the lack of suitable analyticalmethodologies. This paperreports the developmentof The episodicdestruction of groundlevel ozonein the Arctic at a measurementtechnique for the determinationof rapidly sunriseis a phenomenonthat hasbeen observed for many years. photolyzingchlorine (referred to hereas Clp) and bromine (Brp) With the onsetof polar sunrise,ozone levels are often observed speciesat pansper trillion by volume(pptv) mixingratios in the to drop from a backgroundconcentration of •40 ppbv to almost atmosphere.Impey et al. [this issue]discuss the resultsobserved zero on a timescaleof a day or less [Barrie et al., 1988] for from a field studyconducted in the Canadianhigh Arctic at Alert, periodsof 1-10 days. It has beensuggested that the extremely NorthwestTerritories, from mid-Februarythrough mid-April as rapid destructionof near-surfaceozone is associatedwith part of the Polar SunriseExperiment (PSE) 1995. bromine and chlorine atom chemistry [Barrie et al., 1988; McConnell et al., 1992; Jobsonet al., 1994]. While the primary sourceof the halogenatoms is still under investigation,current 2. Methodology information has led to the hypothesisthat the heterogeneous releaseof photochemicallyactive forms of chlorineand bromine The principleof the photoactivehalogen detector is to rapidly from sea salt aerosol, and/or recycling of adsorbedHBr and photolyzethe ambienthalogen-containing compound to releasea HOBr [Fan and ,lacob, 1992] may be responsible.Some of the halogen atom(s). The halogen atom(s) then reacts with a processesthought to be responsible,particularly as sourcesfor scavengerhydrocarbon molecule to producea stablehalogenated bromine atoms, are as follows: photolysisof BrNO or BrNO2 productthat can be collectedand measured.A schematicof the formed by reactionof NO2 or N205, respectively,with sea salt systemis shownin Figure 1. The sampleair is drawnthrough a aerosol [Finlayson-Pitts et al., 1989, !990]; photochemical lengthof-•7 m, 0.635 cm O.D. heavywalled TFE-Teflon tubing release of Br2 from Br' in the snow pack [McConnell et al., at 1 L/min to an in-line filter (Zitex 47 mm Teflon extra-fine, 1992]; or the recyclingprocess of HBr, HOBr, and BrONO2 on containedin a PFA-Teflon filter holder)that removesfine aerosol acidicaerosol to releaseBr2 [Fan and,lacob, 1992;Abbatt, 1994] from the air stream. Testsindicate that >94% of bothCI: and Br: along with BrC1 [Abbatt, 1994; Vogt et al., 1996]. These passthrough the inlet line assembly.Small flows (20 mL/min) of mechanismscan be testedwith appropriatemeasurements of the propene(C3H6) and nitric oxide(NO) from 100 ppm mixturesin precursorsand/or the halogenatedproducts. nitrogen are added to the inlet air stream resulting in To improve our understandingof which processesmay be concentrationsof ~0.5 ppm for both within the photoactive important with respect to halogen atom chemistry in the halogendetector (PHD) reactionvessel. The sampleair passes troposphere,reliable measurementmethods for photolyrically througha 2 L volumecylindrical pyrex vessel, 52 cm in length active sources of chlorine atoms (e.g., C12 and HOC1) and (sampleoutlet 17 cm from the end of the cell) with an outer bromine atoms (e.g., Br2 and HOBO are clearly needed. diameterof 7.0cm and a 0.159 cm thick borosilicateglass Althougha methodhas beenreported that enablesmeasurement window secured on either end. This allows transmission of the of C12[Keene et al., 1993; Pszennyet al., 1993], this is not the full rangeof visible light, decreasingto a transmissionof ! 0% at casefor Br2. There have, as yet, been no measurementsof Br2 280nm. Light from a 150 W xenon arc lamp (A1010 lamp housing,LPS-220 power supply, PTI Inc.), situatedat one end of •Nowat Departmentsof Chemistry and Earth and Atmospheric the cylindrical cell, is collimated by a fused silica lens Sciences,Purdue University, West Lafayette, Indiana. (01LQT0! ! PCC, D=50 mm, FL=-200 mm) and passedthrough 2Alsoat Atmospheric Environment Service, Toronto, Ontario, Canada. the cell. Photolyzablechlorine and bromine compoundswill yield free C1 and Br atomswhich reactrapidly with the added Copyright1997 by theAmerican Geophysical Union. propene,in the presenceof nitric oxide,to producechloroacetone Papernumber 97JD00850. and bromoacetoneas stableproducts. C1 andBr atomscan react 0148-0227/97/97JD-00850509.00 via threepossible pathways, as shownbelow:

15,999 16,000 IMPEY ET AL.' PHOTOLYZABLE HALOGEN MEASUREMENT TECHNIQUE

Ambient Pump Air

NO

"--I • - To Pump

MF• Zero Air

Dump Permeation o•n: Cdilxmion SOUrCeS Closed:Ambienl Samplin•

Mirror -•-'•'•'

Zero Zero Air Air

Figure 1. Schematicdiagram of thephotoactive halogen detector (PHD). (MFC, massflow controller).

(R1a) X + CH2 = CHCH3 ---> XCH2CHCH3 chloroacetone and bromoacetone. The two tubes in series ensure that this is the caseand enabletesting of "breakthrough"of the (Rlb) X + CH2 = CHCH3 -4 CH2CH(X)CH3 analytes. After samplecollection, the tubesare extractedwith 100 [tL of glass-distilledbenzene (Caledon Lab). Analysisis (Rlc) X + CH2 = CHCH3 -4 HX + CH2 = CHCH2 typicallydone immediately after extraction,but testshave shown both chloroacetone and bromoacetone to be stable in solution for The haloacetoneproduct is thenproduced as follows. at least several days. The products chloroacetoneand (R2) XCH2CHCH3 + 02 -4 XCH2CH(O2)CH3 bromoacetone were found to be stable on the PoraPak tubes for at least 19 hours. Samplingrates are controlledby mass flow (R3) XCH2CH(O2)CH3 + NO -4 NO2 + XCH2CH(O)CH3 controllers,and Gast diaphragmpumps are used to pump air throughthe photoactivehalogen detector and sampletubes. The (R4) XCH2CH(O)CH 3 + 02 -4 HO2 + XCH2C(O)CH3 extractsare analyzed for chloroacetoneand bromoacetone using a HewlettPackard 5890 seriesII gaschromatograph equipped with For chlorine atoms, approximately10-15% abstracta hydrogen a SupelcoVOCOL 60 m x 0.25 mm x 1.5 [tm film thickness from carbon3 to form HC1 [Lee et al., 1977], while the rest add columnand an electroncapture detector. A 1 [tL extractsample is to either carbon 1 or carbon2 to form, amongother products, injectedusing a 2 [tL syringeand a splitlessmode of operation chloroacetoneand 2-chloropropanal,respectively. It has been because of the extremely low solution concentrations.The shown that the yield of chloroacetoneis of the order of 40% column is temperatureprogrammed from 50øC to 100øC at [Kleindienstet al., 1989]. The mechanismis similarfor bromine 1øC/min, then at 30øC/min to 240øC. A mixture of 5% methane atoms,although abstraction does not occur. The bromoacetone and 95% argon(P5) is usedas the carriergas. Calibration of the product yield is unknown to date but is currently being gaschromatograph is doneon a regularbasis with diluteliquid determinedin our laboratory. However,detailed knowledge of standardsof chloroacetone(Aldrich) and bromoacetone (prepared the reactionyields is not necessaryif the photoactivehalogen accordingto Blatt [1950]) in benzene. The bromoacetoneis detectoris calibratedwith known concentrationsof C12and Br2. preparedby adding equimolar concentrationsof acetoneand In this system,calibration gases can be input at the reactioncell bromineto a round bottom flask containingwater and glacial inlet followedby determinationof the mixingratio of the acetic acid. The oil which separatesis collected, dried, and h/fioacetoneproduct. It is onlynecessary that the product yields fractionated. A pure product is obtained with subsequent are invariantto input sampleconditions (e.g., humidi.ty). Air distillation. from the reaction cell is sampledby pumping the air, at C12and Br2 usedfor the PHD calibrationare generatedfrom 600 mL/min,through two tubesin seriespacked with 30 mg of VICI Metronics 140-662-002 (25+50% ng/min) and 110-010- PoraPak Q (copolymer of styrene and divinylbenzene; 0003 (37+10% ng/min) permeationcells, respectively.The cells Chrompak)to trapthe chloroacetone and bromoacetone products. are containedin specially designedTeflon tubes, which are Extensiveexperiments have indicated that for 30 mg tubesheld at containedtogether in a stainlesssteel holder. Heat tape -0øC, 36 L of air can be sampledwith completetrapping of surroundsthe unit, and the outputsof the permeationsource IMPEY ET AL.' PHOTOLYZABLEHALOGEN MEASUREMENT TECHNIQUE 16,001

devicesare held constantat 30øC using a Love (model 1600) 200 temperaturecontroller. The permeationrates are determined using the hydrolysis/ionchromatography method described by 180 - Keene et al. [1993]. Typically, 1 L/min of clean air flows over 160 - the permeationsources into two impingersin seriesthat contain ---10mL of an aqueous solution of30 mM HCO3' and 5 mM 140 - HSO3-. The mixtureof HCO3-and HSO3' quantitatively converts C12and Br2to their correspondingions (e.g., 2Cl' for eachC12). 120 - Aftera specifiedsampling time, the measured volume ofsolution lOO is analyzedfor CI- and Br- using a Dionex(DX-100) ion chromatograph.We obtainedpermeation rates for C12and Br2 of 80 36(+4) ng/min and33(+2) ng/min, respectively. 60

3. Results and Discussion 40

3.1. Calibration 20

The photoactivehalogen detector is calibrated with low concentrationmixtures of C12and Br2 in air. UsingFigure 1 as a 0 20 40 60 80 100 120 140 reference,the calibrationprocess proceeds as follows:a mixture of clean air, C3H6(0.5 ppm), and NO (0.5 ppm) continuously flows at a total rate of 1 L/min throughthe reactionvessel, [Br2](pptv) controlledby a pumpat thecell output,and a by-passline just Figure 2. Photoactivehalogen detector calibration for Br2input priorto the cell inlet.The outputof theC12 and Br2 permeation concentrationsin the rangefrom 0 to 140 pptv. The equationof sourcesis dilutedwith clean air to low mixing ratios (0 to theline is given by: Y = [1.49(ñ0.04)]X+4.76(ñ5.59); R2 = 0.98. ! 00 pptv)using a separatemultiple dilution system composed of five massflow controllers,which enables accurate sampling of a fractionof the outputof thepermeation sources. The C12and Br2 mixtureis then furtherdiluted before entering the reactionvessel 0.01 pmol injected,respectively. For a 36 L samplefrom the of the PHD. The chloroacetoneand bromoacetoneproduced PHD, this correspondsto 1.2 pptv and 0.6 pptv as C12and Br2, within the cell are collectedon the adsorbent-filledsampling respectively. A typical ambientPHD samplectiromatogram tubes,extracted, and analyzed. It shouldbe notedthat although takenfrom a recentfield studyconducted in the CanadianArctic our systemis designedto enable samplecollection in remote at Alert, NorthwestTerritories (82.5øN, 63.2øW), as part of the locations,the analysiscould be doneby directGC analysisof the Polar SunriseExperiment 1995, is shownin Figure4. As shown cell contents. The gas phasechloroacetone and bromoacetone in Figure 4, the target productsare well resolvedfro TM other concentrationsin the cell are directly proportional to the (presumablyhalogenated, e.g., CHBr3) species. It is thuslikely concentrationsof Cl2 and Br2 enteringat the reactioncell inlet. that the analysistime could be significantlyshortened. The Samplecalibration curves for Br2and C12 are shownin Figures2 resolvedchloroacetone and bromoacetonepeaks correspond to and 3, respectively.We note that the levelsof chloroacetoneand 100 pptv and 20 pptv of chlorine and bromine,respectively. bromoacetoneproduced are, in the limit of only a small fraction Ambient mixing ratiosof "total photolyzablechlorine" •aried of the halogenphotolyzed, directly proportionalto the Xe arc from<9 to 100pptv and those of "totalphotolyzable bromihe" lamp light intensity. On the other hand, if all Br2 were variedfrom <4 to 38 pptvover the courseof the study[lmpey et quantitativelyphotolyzed, the slope of the calibration curve al., this issue]. would be 2(¾),with a maximumof 2 for ¾ = 1, where¾ is the bromoacetoneyield resulting from Br atom reaction with 3.3. Blanks and Detection Limits propene. From Figure2 we can seethat the slopeis ~1.5. This Blanksamples are generated using the same flows th?oughout impliesthat underour operatingconditions, most (if not all) of the detector,with the lamp on, but no C12or Br2 enteringthe the Br2 is photolyzed. Under thesecircumstances, the system system. Estimationof the detectionlimit for a measurementwas will be insensitiveto changesin lightintensity. If indeedall the calculatedusing the standarddeviation (SD) from the total Br2 is photolyzed,then the bromoacetoneproduct yield is 75%. numberof blankdeterminations. The averageSDbiank•for Cl2 is In the caseof chlorine,the maximumslope attainable is 2(0.4) = 3.1 ppt and for Br2 is 1.3 ppt. The detectionlimit at the 99.7% 0.8 [Kleindienstet al., 1989]. FromFigure 3 the observedslope confidencelevel can then be expressedas 3 x SDblank or 9 pptv is -•0.53;thus a relativelysmaller fraction of the C12is photolyzed C12and 4 pptvBr2. Thiscould be improved by increasingthe gas in our system.This is consistentwith the valuesfor JBr2and Jcl2 phasesample volume. To date, we have been limited to a 36 L and the 2 min residence time within the reaction cell. We estimatevalues of Jm2and Jc• in oursystem of 1.1x10-I s'l and sampleas a resultof chloroacetonebreakthrough •om the PoraPaksample tubes. For field measurements,two tyPi:sof 1.5x10 '2 S -l, respectively. blanksamples were obtained, specifically involving either. having the lampon with zeroair addedat theinlet, or havingthe lamp 3.2. Gas Chromatography off with ambientair. Theseblank samplesexhibit bromO-and This methodwas designedto take advantageof the high chloro-acetoneconcentrations that are typically much lower than sensitivityof the electroncapture detector to chloroacetoneand for ambientor calibrationsamples; that is, we seelittle evidence bromoacetone. GC calibrations show that the detection limit for for problemswith carry over of photolyzablespecies or of the chloroacetoneand bromoacetone is of the orderof 0.02 pmoland analytechloro- and bromo-acetoneon the walls. 16,002 IMPEY ET AL.: PHOTOLYZABLEHALOGEN MEASUREMENTTECHNIQUE

60 brominatedspecies that have beensuggested as precursorsto Br atomproduction for the Arctic. Table 1 showsthe equivalent relative concentrationfor each of these speciesthat would producethe same amount of chlorine and bromine atoms as producedby Br2 and C12;that is, Table 1 indicatesthe relative concentrationof each speciesthat would produce the same measuredresponse as do Cl2 and Br2. For example,100 ppt of CHBr3is requiredto producethe samesignal as obtainedfor 1 ppt of Br2. If we considerthe 4 pptv "as Br2"detection limit, this would correspondto a 400 pptv detectionlimit for CHBr3. Thus it is very unlikely that CHBr3 would contributeto the overall responsein the Arctic, whereCHBr3 mixingratios are typically of the orderof 1-5 ppt [ Yokouchiet al., 1994]. If the exact nature of the photolyticallyactive specie(s)is 10 known, we can estimatethe overall uncertaintyfor this method. Variabilities in a number of factors which include blank levels, measurementof air volumes,analyte collectionefficiencies, and 0 inlet losses and calibrations influence the overall combined uncertaintyof the method. For the case of CI2 and Br2, the 0 20 40 60 80 100 120 uncertaintiesare as follows. Equation(1) showsthe propagation of errorsin calculatingthe C12concentration entering the reaction [C12](pptv) vesselfor a calibrationsample:

Figure 3. Photoactivehalogen detector calibration for CI2 input concentrationsin the range from 0 to 120pptv. The eckuation of (Perm.Rate :l: 10%) x (Dilution Factor:l: 2%) x K• = the line is givenby Y= [0.53(+0.01)] X-0.71(+2.35); R'-= 0.99. [Cl2], ppt (+10%) (1) 3.4. Method Uncertainty Althoughthis methodyields accurate and reproducibleresults whereK• is a constantwhich includesthe molecularweight (for for a known individualphotolyrically active speciessuch as Br2, C12or Br2),Avogadro's number, and the numberdensity at STP. there can be very large effective systematicerrors in the Dilution flow rates have uncertaintiesof the order of +1-2%; determination of an absolute concentration if there are unknown therefore,as shown,the biggestcontributor to the standard photolyricallyactive speciespresent. The PHD will respondto a uncertaintyis the uncertaintyin the outputsof the C12and Br2 variety of atmospherichalogen-containing photolyrically active permeationsources, which is knownto be +10%. The second species,e.g., in the caseof Brp: Br2, HOBr, BrNO, BrNO2, part of the calibrationinvolves measurement of the chloroacetone BrONO2, and CHBr3. This latter group representsthe produced:

5000

bromoacetone

4000 -- ---2opptv

3000-

-100 pptv as C!2

2000- chloroacetone I lOOO -

! ! , ! ! 25 30 35 40 45 RetentionTime (min)

Figure 4. Ambient sample chromatogramtaken on March 15, 1995 at 1041 LT during the Polar Sunrise Experiment at Alert, Northwest Territories, Canada. The resolved chloroacetoneand bromoacetonepeaks correspondto 100pptv and 20 pptvof chlorineand bromine, respectively. IMPEYET AL.: PHOTOLYZABLEHALOGEN MEASUREMENT TECHNIQUE 16,003

Table 1. EstimatedEquivalent Concentrations of Chlorineand different reactor residence times, two equationswith two BromineSpecies Under CurrentCell Conditions unknowns can be obtained that can enable determination of both species(e.g., Br2 and HOBr). We currentlyoperate the reactor Equivalent Equivalent with a 2.0 min residence time. We estimate that the reactor CI Species Concentration Br Species Concentration photodissociationrate coefficientsfor C12and HOC1are Jc12= 1.5x10-2s -I andJHOCI -' 3.6x10 -3S-I. From equation (4), where z [C12] [C12] [Br2] [Br2] is the residencetime in the cell, [HOCI] 5.0[C12] [HOBr] 3.4[Br2] [C1NO] 2.2[C12] [BrNO] 2.0[Br2] [CINO2] 4.2[C12] [BrNO2] 2.0[Br2] [C12]' =e-Jc'2• (4) [CIONO2] 20.0[C12] [BrONO2] 3.1[Br2] [Cl:]0 [CHBr3] 100.0[Br2] we find that =80% of the C12entering the cell is photolyzed. Similarly,we estimatethat for thisresidence time, only =35% of the HOCI is photolyzed. If we then increasethe residence (AreaCounts +2%) time by a factor of 2, to 4.0 min, the fraction that each will (GCCalibration Slope+3%)x (K 2+ 3%)= photolyzechanges to 97% and 58%, respectively. Thus individual [C|2] and [HOC1] mixing ratios could then be [Haloacetone+5%], ppt (2) determinedin the field by measurementof chloroacetoneat these two residencetimes and solving the following two equations' where, again, K2 is a constantwhich includes the solvent extractionvolume (+2%), Avogadro'snumber, the air number [chloroacetone]•,=((1-e-aC'2•')x2[C12] + density,and the volume of air sampled(+2%). In equation(2) the resultingpercentage uncertainty is 5%. The individualblank (l-e_d•ioclZ 1)x [HOCi]) • (5) sampledeterminations also carry a +5% uncertainty;however, theactual blank uncertainty depends mostly on thevariability of the actualblank levels of [haloacetone]in the reactor.Therefore [chloroacetone]=2: ((1- e-JC'2= 2)x 2[C12] + (6) the combined uncertainty in the determination of the [haloacetone]that is produced must take into account the 0-e-J"øc'•2 )x[HOCI]) •t standarddeviation of a seriesof blankdeterminations. Typically, whereJ representsthe photodissociationrate coefficientfor the the blankcorrection is importantonly near the methoddetection respectivespecies in the reactioncell and •t is the yield of limit. For ambientsamples collected in the field, the reactor chloroacetone.Thus, to differentiatebetween C12 and HOC1 in haloacetoneconcentration (blank subtracted)is determinedfirst ambient samples, it is only necessary to measure the using equation(2), and the resultinghalogen concentration is [chloroacetone]for two different reactor residencetimes, and obtainedby substitutioninto the appropriatePHD calibration solve for [C12]and [HOCI], assumingthat dcl2and JHOCIare equation: known. We plan to conductmeasurements of di•ocland JHOBr, chloroacetone+5%- intercept +X =[C12 ] +Z ppt (3) after synthesisof HOCI and HOBr, as describedin Taylor and slope+ Y Bostock [ 1912], Kirk-Othmer [ 1993], Orlando and Burkholder [1995] and Mabenet al. [1995], so that thesefour halogenatom Here the uncertaintyin the slopecontains a contributionfrom the precursorscan be quantitativelymeasured in the atmosphere. uncertainty in the permeation source output, i.e. from equation(1). For samplesnear the detectionlimit, the largest 4. Conclusions contributionsto the overalluncertainty are derivedmostly from the calibrationslope and interceptuncertainties, rather than the The need for measurementsof rapidlyphotolyzable halogen 5% uncertaintyfrom the GC determination.From the recentfield atom precursors has become increasingly important to study conductedin Alert, NorthwestTerritories, Canada, the developmentof a betterunderstanding of the complexchemistry combineduncertainties in the measured concentrations for Clp of the Arctic region. In responseto this need,we developeda andBrp range from 10 to 35% and15 to 100%,respectively, the measurementtechnique that may provide answersto the surface higheruncertainty corresponding to levelsclose to the detection ozonedepletion phenomenon found with the onsetof 24 hoursof limit. However,these values only apply if themeasured analytes daylight, as well as the more generalquestion regarding the are in fact C12and Br2. global importanceof halogen atom chemistryin the marine Althoughthis methodresponds to all rapidlyphotolyzable boundarylayer. The methoddeveloped enables measurement of speciesand is thusa potentiallyuseful technique for detecting photolyticallyactive sources of Br and CI atomsat the low parts their presence,it is in the sameway limitedfor quantitative per trillion level,and enables detection of the relativeimportance measurementswhen numerous such species are present of Br and C1 atomsin the atmosphere.This instrumentwas used simultaneously,as indicatedfrom the informationin Table 1. in a recentfield studyconducted in the Canadianhigh Arctic at However,it is in principlepossible to distinguishbetween two Alert (82.5øN, 63.2øW) as part of the Polar SunriseExperiment knownspecies with significantlydifferent photodissociation rate 1995. The data are discussedby Impey et al. [this issue]. coefficients(J values),e.g., Br2and HOBr (or C12and HOCI), Ambient mixing ratios of "total photolyzablechlorine" varied whichare suspectedof beingthe dominantphotolytically active from <9 to 100 pptv and thoseof "total photolyzablebromine" speciesin the Arctic [Fan and Jacob, 1992]. A theoretical from <4 to 38 pptv. calculationof the speciesphotolytic lifetimes within the reaction Acknowledgments.We thankthe Natural Sciencesand Engineering vessel,based on the lamp spectraloutput and the published ResearchCouncil of Canadaand the AtmosphericEnvironment Service absorptionspectra, shows that by obtainingambient samples with for their financialsupport. 16,004 IMPEY ET AL.: PHOTOLYZABLE HALOGEN MEASUREMENT TECHNIQUE

References McConnell, J.C., G.S. Henderson, L.A. Barrie, J. Bottenheim, H. Niki, C.H. Langford, and E.M. Templeton, Photochemicalbromine Abbatt,J.P!D., Heterogeneous reactionof HOBr with HBr and HC1 on ice productionimplicated in Arctic boundary-layerozone depletion, surfacesat 228K, Geophys.Res. Lett., 21, 665-668, 1994. Nature, 355, 150-152, 1992. Barrie, L.A., J.W. Bottenheim, R.C. Schnell, P.J. Crutzen, and R.A. Orlando, J.J., and J.B. Burkholder,Gas-phase UV/visible absorption Rasmussen,Ozone destructionand photochemicalreactions at polar spectraofHOBr andBr20, J. Phys.½hem., 99, 1143-1150,1995. sunrise in the lower Arctic atmosphere,Nature, 334, 138-14l, 1988. Pszenny,A.A.P., W.C. Keene, D.J. Jacob, S. Fan., J.R. Maben, M.P. Blatt,A. H• (Ed.), OrganicSyntheses, collective vol. 2, p. 88, JohnWiley, Zetwo, M. Springer-Young,and J.N. Galloway, Evidenceof inorganic New York, 1950. chlorine gasesother than hydrogenchloride in marine surfaceair, Fan, S.-M. and D.J. Jacob, Surface ozone depletionin Arctic spring Geophys.Res. Lett., 20, 699-702, 1993. sustainedby bromine reactionson aerosols,Nature, 359, 522-524, Taylor, R.L., and C. Bostock,Researches on bleachingpowder, II, The 1992. action of dilute acids on bleachingpowder, J. ½hem.Soc., 101, 444- Finlayson-Pitts,B.J., M.J. Ezell, and J.N. Pitts, Formationof chemically 457, 1912. activechlorine compounds by reactionsof atmosphericNaC1 particles Vogt, R., P. J. Crutzen,and R. Sander,A mechanismfor halogenrelease with gaseousN2Os and C1ONO2,Nature, 337, 241-244, 1989. from sea-saltaerosol in the remote marine boundary layer, Nature, Finlayson-Pitts,B.J., F.E. Livingston,and H.N. Berko,Ozone destruction 383, 327-330, 1996. and bromine photochemistryat ground level in the Arctic spring, Yokouchi, Y., H. Akimoto, L.A. Bartie, J.W. Bottenheim, K. Anlauf, and Nature, 343, 622-625, 1990. B.T. Jobson, Serial gas chromtographic/massspectrometric Impey, G.A., P.B. Shepson,D.R. Hastie,L.A. Barrie, and K.G. Anlauf, measurementsof some volatile organic compoundsin the Arctic Measurementsof photolyzablechlorine and brominesources during atmosphereduring the 1992 Polar SunriseExperiment, J. Geophys. the Polar SunriseExperiment 1995, J. Geophys.Res., this issue. Res., 99, 25,379-25,390, 1994. Jobson,B.T., H. Niki, Y. Yokouchi, J. Bottenheim,F. Hopper, and R. Leaitch,Measurements of C2-C6 hydrocarbonsduring the 1992 Polar SunriseExperiment: Evidence of Cl-atom and Br-atom chemistry, J. Geophys.Res .... 99, 25,355-25,368,1994. L. A. Barrie, AtmosphericEnvironment Service, 4905 Dufferin St., Keene, W.C., J.R. Maben, A.A.P. Pszenny, and J.N. Galloway, Downsview,Ontario, Canada M3H 5T4. (e-mail: len.barrie•ec.gc.ca) Measurementtechnique for inorganicchlorine gasesin the marine D. R. Hastie and G. A. Impey, Departmentof Chemistry, York boundarylayer, Environ. Sci. Technol.,27, 866-874,1993. University, Room 124 CCB, 4700 Keele St., North York, Ontario, Kirk-Othmer, Encyclopediaof Chemical Technology,Vol. 5, pp. 938- CanadaM3J 1P3. (e-mail:hastie•yorku.ca; [email protected]) 943, JohnWiley andSons, Toronto, 1993. P. B. Shepson,Department of Chemistry, Purdue University, 1393 Kleindienst, T.E., P.B. Shepson,and CM. Nero, The productionof Brown Bldg., West Lafayette, IN 47907-1393. (e-mail: chlorine atoms from the reaction of OH with chlorinated pshepson•chem.purdue.edu) hydrocarbons,Int. J. Chem.Kinetics, 21, 863-884, 1989. Lee, F.S.C., and F.S. Rowland, Thermal chlorine-38 reactions with propene,J. Phys.Chem. 81, 1222-1229,1977. Maben, J.R., W.C. Keene, A.A.P. Pszenny,and J.N. Galloway, Volatile inorganicC1 in surfaceair overeastern North America,Geophys. Res. (ReceivedSeptember 27, 1996;revised February 19, 1997; Lett.. 22, 3513-3516, 1995. acceptedMarch 7, 1997.)