The uptake of atomic oxygen on ice films: Implications for noctilucent clouds

Abstract

Rocket-borne measurements show that atomic oxygen is often depleted by more than an order of magnitude in the vicinity of noctilucent clouds (NLCs). These are clouds that occur in the upper mesosphere around 82 km, during the summer at high latitudes. The volumetric surface area of NLCs is large enough for the heterogeneous removal of atomic O potentially to account for the observed depletion. In order to explore the possibility that the ice surface provides an efficient catalyst for the recombination of O with O or O₂, the uptake coefficient (γ) of atomic O on ice was measured in a fast flow tube under conditions relevant to NLCs. Ice was deposited from the vapour at 90 K, and then annealed at 160 K to form a cubic-crystalline ice film suitable for performing kinetic uptake measurements between 112 and 151 K. The effective ice surface area was determined using the BET isotherm technique with Kr. O atoms were produced by the microwave dissociation of N₂ followed by titration with NO, and monitored by resonance fluorescence at 130 nm. For the purposes of atmospheric modelling, the overall uptake rate is then given by γ(112–151 K) = 7 × 10⁻⁶ + 1.5 × 10⁻¹⁰ exp(11.4 kJ mol⁻¹/RT) [O₂], with an uncertainty of ±24%. In a separate set of experiments, pulses of atomic O were generated in the flow tube by the photolysis of O₃ at 248 nm, and a chromatographic analysis of the pulse shape indicated that O was bound relatively strongly (binding energy = 35–40 kJ mol⁻¹) to some sites on the ice surface. Quantum calculations using the ONIOM technique were employed to show that O adsorbs weakly (binding energy = 11 kJ mol⁻¹) on a perfectly crystalline ice surface by bonding to a single dangling H (and a second H bound within the surface), but can adsorb much more strongly at a disordered site where two dangling H's are in close proximity. A kinetic model of O uptake was then developed consistent with both the experimental and theoretical results. Finally, γ is almost certainly too small to explain the depletion of O around NLCs. Instead, it appears that O and O₃ are removed by gas-phase catalytic cycles driven by elevated concentrations of odd hydrogen species (H and OH), produced by the Lyman-α photolysis of the layer of enhanced H₂O vapour around an NLC, and by the direct photolysis of the ice particles.