Absorption cross sections of the ClO dimer

K. J. Huder and W. B. DeMore*Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109Received: January 10, 1995; In Final Form: February 3, 1995_NASA-CR-200336The absorption cross sections of the C10 dimer, C1OOCI, are important to the photochemistry of ozone depletionin the Antarctic. In this work new measurements were made of the dimer cross sections at 195 K. Theresults yield somewhat lower values in the long wavelength region, compared to those currently recommendedin the NASA data evaluation (JPL 94-26). The corresponding solar photodissociation rates in the Antarcticare reduced by about 40%.IntroductionThe C10 dimer is formed by the self-reaction of C10 radicalsat the low temperatures of the earth's polar winter. Whensunlight is present, the dimer is thought to take part in catalyticozone destruction by yielding chlorine atoms upon photolysis,followed by attack of the chlorine atoms on ozone.C1OOC1 + hr _ CI + CIOO (1)CIOO _ C1 + 02 (2)CI + 03 _ C10 + O 2 (3)To model the role of C1OOCI in polar photochemistry, it isnecessary to know its rate of formation, the photolytic crosssections, and the photolysis products. In addition to the pathyielding C1 and C1OO radicals, another path yielding C10 ÷C10 is possible.C1OOC1 + hv _ C10 + CIO (4)This path would not contribute to ozone loss, because C10is inert to ozone. Although it is normally assumed in atmo-spheric models that only the C1 ÷ CIOO branch is important,the experimental basis for this assumption is somewhat uncer-tain. Cox and Hayman I concluded that C1 and CIOO are themajor dimer photolysis products by modeling the disappearanceof C120 and the appearance of dimer in the photolysis of C120at 254 nm, at temperatures of 203-233 K. However, theyplaced an uncertainty of approximately a factor of 2 on the resultfor the C1 + CIOO yield.In the most direct study at a wavelength appropriate foratmospheric photolysis of the dimer, Molina et al) reported aquantum yield of approximately unity for the C1 ÷ C1OO pathat 308 nm. That result was based on measurements of the C1atom LIF signal in the photolysis of the dimer (235 K, 20 Totr)compared to the CI atom LIF signal when photolyzing C12 (alsoat 308 nm). However, the authors used a value of 22 for theratio of the dimer cross sections at 245 and 308 nm, asdetermined in their own work. If one takes that ratio from thecurrent NASA recommended dimer cross sections, 3 the valueis 13 and the corresponding C1 ÷ C1OO quantum yield wouldbe only 0.6. That would imply that the CIO ÷ C10 path makesa 40% contribution. Thus, there is a direct relationship betweenthe cross sections and the photolysis products.Eberstein 4 argued that dimer photolysis should proceed bypath (4) on the grounds that the O-O bond in C1OOC1 is much®Abstract published in Advance ACS Abstracts, March 1, 1995.0022-3654/95/2099-3905509.00/0weaker than the C1-O bond. However, Eberstein assumed thatthe C1-O bond energy in C1OOC1 is similar to that of the C10radical (about 64 kcal/mol). Actually, the CI-OOC1 bondenergy is only about 21 kcal/mol, just slightly greater than theC10-OC1 bond (18 kcal/mol). 3 Thus, an argument based onpreferential breaking of the weaker bond is not compelling inthis case.Previous studies of the dimer spectrum h5"6 (but not that ofMolina et al. 2) have used C120 as the source of C1OOCI. Thedimer absorption cross sections are difficult to measure, becausethe compound is stable only at low temperatures and cannot beprepared in the pure state. The spectrum has only been obtainedin the presence of other molecules such as C12, 03, or C120,and the net C1OOC1 spectrum must be extracted by deconvo-lution of the composite spectra. This process is somewhatqualitative in nature and is subject to error. The situation issimilar to the case of HOC1, for which previous cross sectionmeasurements have been erroneous due to contamination withC120 and C12.7 Some of the largest cross section errors occurin the "tail" region above 300 nm, where the spectrum is weak.The errors arise from the necessity to correct the spectrum forC12 and from base line errors in the total spectrum. Thiswavelength region is the most important for determining theatmospheric photolysis rates of C1OOC1.©Experimental SectionApparatus. The C10 dimer was prepared as before 6 byphotolysis of C120 mixtures. The C120 was produced byoxidation of C12 over HgO at low temperature: The experi-ments were performed in a low-temperature, 5 cm path lengthspectrophotometric cell of quartz construction. The cell wascooled by a surrounding jacket containing a dry ice/methanolmixture. Some early experiments were performed in a metalcell of similar construction, with halocarbon coating on thewalls. However, dimer was found to be more stable in thequartz cell. The cell was evacuated and filled with C120 at apressure of about 10 Torr and then pressurized to about 3 atmwith nitrogen to reduce diffusion to the walls. A low-pressuremercury lamp was used for irradiation at 254 nm. The UVabsorption spectra were taken from 200 to 400 nm with a CaryModel 4E spectrometer.Dimer Spectrum. Since it is difficult to obtain reliable crosssections at longer wavelengths by direct experimental measure-ment (for reasons discussed above), we have concentrated onobtaining accurate data at shorter wavelengths. Those valuescan then be used for extrapolation above 310 nm. We havelooked only at relative cross sections, since the absolute values1995 American Chemical Society