JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. D2, PAGES 3739-3743,FEBRUARY 20, 1994 Branching ratios for the O(IO) -[-N20 reaction Christopher A. Cantrell, Richard E. Shetter, andJack G. Calvert Atmospheric Kinetics and Photochemistry Section, Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, Colorado Laboratory studies ofthe branching ratio of the reaction ofelectronically excited oxygen atoms, O(ID), with nitrous oxide, N20, were performed bymeasuring the loss of N20 and the formation of products in a static optical cell interfaced to a Fourier transform spectrometer system. The results agree well with the most recentmeasurements, and a critical evaluationof the available data resultsin a factor of 2 reductionin the uncertainty. Theratio klb/k 1 is the fraction of reaction leading to the production of NO (reaction (lb)) relative to the total reaction of O(ID) with N20(reaction (1)). The present recommendation for this ratio asit pertains to atmospheric reactions is 0.61 + 0.08 (95% confidence interval). The consequences of this result in terms of stratospheric NOy production and field measurements of ozone photolysis rates are discussed. INTRODUCTION The atmospheric chemistry of nitrous oxide is of considerable interest becauseunderstanding the fate of this molecule is critical in order to assess anthropogenic influence on the budget of stratospheric odd nitrogen. In this vein, scientists are attempting to understand the reasons for the observed global averageincrease of 0.25 to 0.31% per year in theN20 tropospheric concentration [Prinn et al., 1990]. N20 has no known loss processes that operate at any significant rate in the troposphere [Stedmanet al., 1976]. It is removed from the troposphere by transport to the stratosphere, followed by reaction with O(1D), or at higher altitudes bydirect solar photolysis. N20 +O(1D) --• N 2+0 2 (la) --• 2NO (lb) N20 + hv --• N 2+O(ID) (2) These processes have been the subject of a number of laboratory studies, the earliest dating to 1928 [MacDonald, 1928], with subsequent investigationsreported in the 1950s, 1970s, and early 1980s. The possibility of N atom or NO production from N20 photolysis has beeninvestigated and found to be unimportant [Paraskevopoulos and Cvetanovic, 1969; Preston and Barr, 1971; Greenblatt and Ravishankara, 1990]. Also, upper limits for other exothermic channels for reaction (1) have been set at less than about 4% of the total reaction [Scott et al., 1971; Davidson et al., 1979]. Thus the odd nitrogen-forming and nitrogen/oxygen-forming channels constitute the major, if not exclusiveroutesfor this reaction. In order to assess the rate of odd nitrogen productionfrom reaction (lb), we must know the rate coefficient for reaction (1b), kl/,, or alternatively, we need R (=klb/kl)' We have undertakena study to perform additional experimentsto determine R, and to evaluate the reports of R in the literature. Copyright 1994by theAmerican Geophysical Union. Paper number 93JD02659. 0148-0227194193JD-02659 $05.00 EXPERIMENTAL PROCEDURES All the experiments were performed with a temperature- variable stainless steel reaction cell and BOMEM Fourier transform spectrometer system that has been discussed previously[$hetter et al., 1987; Cantrell et al., 1988]. Most of the experiments were performedbeginningwith 1.4 to 3.5 x 1015 N20 molecules cm '3in760 torr of helium. Helium bath gas was used because it is ineffective at quenching excited oxygen atoms (O(1D)) to ground state ones (O(3p)). Ozone was addedcontinuously by desorption off of silica gel at-77øC. The mixture was photolyzed with a 300 W Xe arc. The photolysis beam was filtered with 1 cm of water and a 7-54 color glass filter (Corning), which transmits wavelengths from 240 to 410 nm. Typically, the mixturewas photolyzed for 10 to 24 hourswith 25 to 75% conversion. Infrared spectra were collected hourly, and the concentrations of N20, 03, and oxides of nitrogen (HNO 3, NO, NO 2, N20 5) were determined through comparison with standard spectra. The absorbances of the standard spectra were within 20% of those of the sample being analyzed. This is necessary because of the nonlinear relationship between concentration and absorbance for molecules with structured s[•ectra atthe resolution which they were collected here (0.5cm"). Theconcentrations of NO, NO 2, and N20 5 were alwaysbelow the detection limit of the spectrometer system, about 2 to4 x 1012 molecules cm '3,and the only oxide of nitrogen with a concentration above the detection limitwas HNOy We believe therefore that we could account for greaterthan 98% of the odd nitrogen present in the system, not counting possible calibration errors. The formation of HNO 3 arises from theoxidation of the NO product formed in reaction (lb) to NO 2 by ozone. Thisis followed by further oxidation to NO 3, and finally, N205 is formed from the association of NO 2 and NO 3. In oursystem, water is formed during thelong photolysis periods and is accompanied by CO 2 formation. We suspect that small levels of hydrocarbon contamination adsorbed onto the cell walls are oxidized by hydroxyl radicals formed heterogeneously on the cell walls, therebyyielding water as a product. This explanation seems more plausiblethan simply an adsorbed monolayerof water because our calculations indicate that this would produce only about 10 TM molecules cm -3 in the gas phase, not enough to 3739