Temperature dependence of isotope fractionation in N2O photolysis

Stratospheric ultraviolet (UV) photolysis is the dominant sink reaction and main origin of isotopic enrichment for atmospheric nitrous oxide (N2O). To a large extent, the flux of isotopically heavy N2O from the stratosphere is responsible for the enrichment of tropospheric N2O relative to its sources at the Earth's surface. In order to simulate the stratospheric enrichments quantitatively in atmospheric models and to examine the global N2O cycle using isotope measurements, knowledge of the fractionation constants is required. However, to date, all experimental studies of isotopic enrichment in N2O photolysis have been performed at room temperature only. Here we report the first temperature- dependent (193 < T/K < 295) measurements of O-18 and position-dependent N-15 fractionation constants obtained by broadband photolysis at wavelengths of relevance to the stratospheric UV window. For a given extent of reaction, we find higher enrichments at lower temperatures, qualitatively in agreement with theoretical predictions. The relative changes are in the order (NNO)-N-14-N-15 > (N2O)-O-18 > (NNO)-N-15-N-14, similar to the absolute values. If temperature was the only parameter of influence, not only the fractionation constants themselves, but also the ratio of fractionation constants at the central to terminal nitrogen sites, eta = (15)epsilon(2) /(15)epsilon(1), should decrease along the vertical stratospheric temperature gradient. These temperature effects do not help to explain the lower eta values observed in the lower stratosphere, but they are nevertheless essential ingredients for models of atmospheric isotope chemistry. We also investigate a hitherto unexplained artefact in laboratory measurements of N2O photolysis: At high degrees of conversion, N2O loss by the reaction with O(D-1) becomes important, presumably due to the photochemical production and subsequent photolysis of NO2 in the reaction cell. The effect gains importance with increasing concentration and in the present study, it caused decreases in the measured fractionation constants requiring correction for initial N2O mixing ratios of 4 mmol mol(-1)