GEOPHYSICAL RESEARCH LETTERS, VOL. 18, NO. 1, PAGES 13-16, JANUARY 1991 ISOTOPICEXCHANGE BETWEEN CARBONDIOXIDE AND OZONEVIA O(!D) IN THE STRATOSPHERE Yuk L. Yung, W.B. DoMore,and Joseph P. Pinto Abstract. We proposea novelmechanism for isotopic Table 1 exchangebetween CO 2 and0 3 vi•aO(ID) + CO2 -->CO_•* followedby CO3' --> CO2 + O(aP). A one-•ensior•al List of essentialreactions used in the photochemicalmodel. model calculation shows that this mechanism can account for The units for photodissociationcoefficients and two-body theenrichment in 180 in thestratospheric CO2 observed by rateconstants are s -1 andcm 3 s-l, respectively.The coeffi- Gatnoeta!. [1989],using the heavy 0 3 profile-observed by cientsfor photolysisrefer to nfid-latitudedimally averaged Mauersberger[1981]. The implicationsof this mechanism values at 30 kan. All molecular kinetic data are taken frmn for otherstratospheric species and as a sourceof isotopically DoMore etal. [1990], except otherwise stated. Rate heavyCO 2 in thetroposphere are briefly discussed. constantsforisotopic species are estimated byus. Q = 180. h•troduction Rla O3+ bY-->O2 + O(1D) Jla=7.4x10 -5 RecentlyGamo et at. [1989] reportedmeasurements of Rib O2Q+ hv-->O2 + Q(!D) Jlb=1/2 J la heavyisotopes of CO2 in the stratosphereover Japan. Let O R2a O(ID)+ M--->O+ M k2a= 2.0x 10-11 e ll0/T = 160 andQ = 180.The fiQ of stratosphericCO2, as defined R2b Q(1D) + M --->Q + M k2b= k2a in Gatno et at., was found to be 2O/oo(two parts in a R3 O(ID)+ CO2---->CO2 + O k3 = 7.4x 10-11 e 120/T thousand)greater than that of troposphericCO 2 at 19 km, R4a Q(1D)+ CO2-->CO Q + O k4a= 2/3k 3 and appearsto continueto increasewith altitudeup to the R4b O(ID)+ COQ-->CO2 + Q k4b= 1/3k 3 highestaltitude (25 km) where sampleswere taken. The data are reproducedin our Figure 2. The authorsproposed R5 O(1D)+ CO-+CO+ O k5 = 4.6x 10-11 (a) two possibleexplanations, one dynamical and one chemical. R6a Q(1D)+ CO-->CQ+ O k6a= 1/2k 5 The preferred mechanism was chemical and involved R6b O(ID)+ CQ-->CO + Q k6b= 1/2k 5 transferringa heavyoxygen atom from 0 3, which was first R7 QH+ CO2---> COQ + OH k7 < 1x 10-17 Co) measuredto be isotopically enriched by Mauersberger R8 Q+ CO2-->COQ+ O k8 < 1x !0-18 (c) [I981], to carbon dioxide. However, the authors did not proposea detailedchemical scheme. The direct thermal exchangereaction CO 2 + O2Q --> COQ + 03 is inefficient (a) Cvetanovic [ 1975]. (J.-S.Wen, privatecommunication, 1990). (b) Greenblatt and Howard [1989]. We propose,instead• that the transferof Q from 0 3 to (c) Wen and Thiemens [1990]. CO2 is initiatedby O(ID) reactions. We will use a one- dimensionalphotochemical model 'mcorporatingreactions for exchangeof Q to simulatethe vertical profile of COQ Caltech-JPLmodel [Froidevauxet at., 1985], a typical observedby Gamo etal. The factors controlling the diumally-averagedone-dimensional photochemical model, COQ/CO2 ratio in the tropospherehave been studiedby alongwith our 0 3 proErie. Keeling [1961], Bottinga and Craig [1969], Francey and Considerthe reaction between O(1D) and CO 2 Tans [1987], and Fried!i etal. [1987]. We will attemptto O(!D)+ CO2 --> CO3' assessthe importanceof exchangewith the stratosphere relative to these other factors. CO3' ---->CO 2 + O Theexistence of the CO3' complex was first proposed by Chemical Kinetics Katakis and Taube [1962] in a study of the exchange Photolysisof 03 belowabout 305 nm produces O(1D), an betweenO(1D) and CO 2. Jacoxand Milligan [1971] favored excited state of the oxygen atom, by R1 (see listing and a three-memberring structurewith an O-C-O angleof 65'. numberingof reactionsin Table !). The primary fate of The complexrapidly predissociates to CO 2 + 0 [DeMote O(ID) in the atmosphereis quenchingby the ambient atmosphericgases (R2a) Becauseof the efficiencyof the quenching reactions, the abundance of O(tD) in the 8o atmosphereis very small and it has not been directly detectedin the atmosphere.Nevertheless, this excited atom plays fundamentalroles in the chemistryof HOx radicals [Levy,1971; Wofsy etal., 1972]and NO x species[McElroy and McConnell, 1971; Nicolet and Peetermans, 1972]. Figure1 presentsthe O(1D) concentration predicted by the .• 40 1Divisionof Geologicaland Planetary Sciences; Califomia Instituteof Technology,Pasadena, California 91125. 20 2Jet PropulsionLaboratory, 4800 Oak Grove Drive, Pasadena,California 91109. 3AREAL,United States Environmental Protection Agency, ResearchTriangle Park, North Carolina 27711. 0 0 5 10 15 20 25 •Q - •Qo(%ø) Copyright 1991 by the American Geophysical Union. Paper number 90GL02478 ri•;i•one-dimensionalConcentrations ofphotochemicalO(ID) and03 modelfrom the[Froidevaux Caltech- 0094-8534/90/90GL-02478503.00 et al., 1985]. 13 14 Yung et al.' CarbonDioxide and Ozone andDede, 1970]. There has been great interest in CO3'in enrichmentproduced by Q(1D). However,we believethat comxectionto the CO2 stabilityproblem on Mars [McElroy the results of Jaffe and Klein are in error. First, it is and Hunten, 1970; Noxon, 1970], which was subsequently extremelyimprobable that k 8 > 100 k7, sinceOH is gener- solved by a different scheme [Parkinson'and Hunten, 1972; ally morereactive than O. Second,a qualitativeupper lhnit McElroyand Donahue, 1972].. for k8 may be obtainedby modelingan experimentper- The rapid quenchingof O(tD) by CO9 [DeMote eta!., formedby Wen and Thiemens[1990]. Wen (1990, private 1990]with rate coefficient, k3 = 1.1x I0-'10 c•m 3 s -1 atT = conununication)deduced k8 < 1.5x 10-17cm 3 s-1 at T = 298 K, undoubtedlyproceeds via the CO3* intermediate 363K. Weconclude that k 8 < 10-18 cm 3 s-! atstratospheric [DeMote and Dede, 1970]. For comparisonwe notethat the temperatures,and hence R8 is nothrtportant. correspondingrate constants for quenchingof O(ID) by Ar andKr are7 x 10-13 and8 x 10-12 cmTs-1, respectively PhotochemicalModeling [Cvetanovic,1975]. The existenceof CO3' wouldbe The Caltech-$PLone-dimensional photochemical model expectedto result in an exchm•gebetween the incidentand is describedelsewhere [Allen et ai., !981; Froidevauxet al., emergentO atoms. Indeed,in an isotopicallylabelled study usingCO 2 photolysisat 147nm asthe source of O(ID), 1985;Michelangeli etal., 1989] i In the present investigation Baulch and Breckenridge[1966] showedthat the efficiency we fix the concentrationsof O(D) and0 3 (seeFigure 1) to the values in Froidevauxet al. [1985]. This is justified yield for isotopicexchange is 0.69, closeto the expected valueof 2/3ff theejection of anO atomfrom CO3' were a becauseCO 2 has negligibleinfluence on oxygen.photo- purely statisticalprocess. Other kinetic studies[Yamazaki chemistry.The relevantphotochemistry of O, O(XD),03, and Cvetanovic, 1964; Weissberger et al., 1967] have CO2, CO, andtheir isotopes is summarizedin Table !. The generally confim•edthis result. rate coefficientsfor the isotopicspecies are for the mostpart The starflingdiscovery of an 180 enrictunentin strato- unknown, and we have to make reasonableguesses to arrive spheric O by Mauersberger[1981] has been confmxted at thepreferred values given in Table !.. Verticaltransport is [Carlieta•., 1982; deZafra etal., 1984; Mauersberger, 1983, parameterized by eddydiffusion. 1987; Rinsland et al., 1985; Goldman et aI., 1989], and Let thestandard Q/O ratiobe f. Sincestratospheric 0 3 is studiedin the laboratory [Heidenreichand Thiemens, 1986; enrichedin Q, we have [O.2Q]/[O3] = 2 f E where E = Thiemensand Jackson,1987; Yang and Epstein, 1987a,b; enrichmentfactor. SinceO(XD) is in photochemicalequilib- Morton et al., 1989; Andersoneta!., 1989]. The theory of rium with 03, we have, assumingthe reactionsof Table 1, isotopicenrichment for 0 3 was initiatedeven before the [Q(ID)]/[O(ID)]= f E. Thechoice of E in ourmodel repre- atmosphericobservations [Cicerone and McCmmb, 1980], sentssome difficulty. The observedenric!maent factors axe but this and all subsequenttheories [Kaye and Strobel,1983; not reproducible,either due to uncertaintiesin the experi- Blake et aI., 1984; Kaye, 1986, 1987] completelyfail to ments or a possible dependenceon the solar cycle accountfor either the atmosphericor laboratorymeasure- (Mauersberger,1990, private communication). It must be ments. The suggestionsof Valentini et al. [1987] and Bates emphasizedthat there is no theoreticaljustification for a [ 1986] arenot widely accepted.It is not thepurpose of this solar cycle dependence. In case A, we set the enrichment factorE A to be the measurementsof Mauersberger[1981]. paperthatstratospheric toenter this great0 3 is debate. enhanced Wein acce•,t ! Oas andempirical pursue factits To set a lower bound for E we also use the later measure- mentsto deriveanother profile E B = 1 + (EA - !)/3. All implicationsfor the isotopicenrichment of CO2 and other otheruncertahaties in the model are probablysmaller than species. that of E. Considerthe followingsequence of reactionsinitiated by The photochemicalmodel was mn to steadystate in the the photolysisof a heavyozone molecule, diurna/lyaveraged mode. We adoptf = 1/500 [Kaye, 1987] The lowerboundary conditions at the groundare [CO2] = O2Q+hv--> 0 2 +Q(!D) 340 ppmmad [COQ] = 1.36ppm. At theupper boundary (80 km), the fluxes of CO2 and COQ are zero. In the standard Q(1D)+ CO2 --> CO2Q* runE = 1.0. In runsA andB we useas input the profries E^ CO2Q*--> COQ+ O madE B. Theresults for •JQ - ;SQofor CO 2 are.shown in Fig- ure 2 (bQo= troposphericvalue = 40.7O/oorelative to The net result is the transfer of Q from ozone to carbon SMOW at 24øC),along with Gamoet al.'s measurements.It dioxide.The reverseprocess