RESUMO
In this work we perform computer simulations of the ice surface in order to elucidate the role of nitrogen in the crystal growth rates and crystal habits of snow in the atmosphere. In pure water vapor at temperatures typical of ice crystal formation in cirrus clouds, we find that basal and primary prismatic facets exhibit a layer of premelted ice, with thickness in the subnanometer range. For partial pressures of 1 bar, well above the expected values in the troposphere, we find that only small amounts of nitrogen are adsorbed. The adsorption takes place onto the premelted surface, and hardly any nitrogen dissolves within the premelting film. The premelting film thickness does not change either. We quantify the resulting change of the ice/vapor surface tension to be in the hundredth of mN m-1 and find that the structure of the pristine ice surface is not changed in a significant manner. We perform a trajectory analysis of colliding water molecules, and find that the attachment rates from direct ballistic collision are very close to unity irrespective of the nitrogen pressure. Nitrogen is however at sufficient density to deflect a fraction of trajectories with smaller distance than the mean free path. Our results show explicitly that the reported differences in growth rates measured in pure water vapor and a controlled nitrogen atmosphere are not related to a significant disruption of the ice surface due to nitrogen adsorption. On the contrary, we show clearly from our trajectory analysis that nitrogen slows down the crystal growth rates due to collisions between water molecules with bulk nitrogen gas. This clarifies the long standing controversy of the role of inert gases on crystal growth rates and demonstrates their influence is solely related to the diffusion limited flow of water vapor across the gas phase.
