Molecular dynamics simulations of the ice temperature dependence of water ice photodesorption

• Published in: The Journal of chemical physics 2010-05, Vol.132 (18)
• Main Authors: Arasa, C., Andersson, S., Cuppen, H. M., van Dishoeck, E. F., Kroes, G.-J.
• Format: Article
• Language: English
• ISSN: 0021-9606; 1089-7690
• DOI: 10.1063/1.3422213

The ultraviolet (UV) photodissociation of amorphous water ice at different ice temperatures is investigated using molecular dynamics (MD) simulations and analytical potentials. Previous MD calculations of UV photodissociation of amorphous and crystalline water ice at 10 K [S. Andersson et al., J. Chem. Phys. 124, 064715 (2006)] revealed—for both types of ice—that H atom, OH, and H2O desorption are the most important processes after photoexcitation in the uppermost layers of the ice. Water desorption takes place either by direct desorption of recombined water, or when, after dissociation, an H atom transfers part of its kinetic energy to one of the surrounding water molecules which is thereby kicked out from the ice. We present results of MD simulations of UV photodissociation of amorphous ice at 10, 20, 30, and 90 K in order to analyze the effect of ice temperature on UV photodissociation processes. Desorption and trapping probabilities are calculated for photoexcitation of H2O in the top four monolayers and the main conclusions are in agreement with the 10 K results: desorption dominates in the top layers, while trapping occurs deeper in the ice. The hydrogen atom photodesorption probability does not depend on ice temperature, but OH and H2O photodesorption probabilities tend to increase slightly (∼30%) with ice temperature. We have compared the total photodesorption probability (OH+H2O) with the experimental total photodesorption yield, and in both cases the probabilities rise smoothly with ice temperature. The experimental yield is on average 3.8 times larger than our theoretical results, which can be explained by the different time scales studied and the approximations in our model.