Microscopic simulation of methanol and formaldehyde ice formation in cold dense cores

Context. Methanol and its precursor formaldehyde are among the most studied organic molecules in the interstellar medium and are abundant in the gaseous and solid phases. We recently developed a model to simulate CO hydrogenation via H atoms on interstellar ice surfaces, the most important interstellar route to H2CO and CH3OH, under laboratory conditions. Aims. We extend this model to simulate the formation of both organic species under interstellar conditions, including freeze-out from the gas and hydrogenation on surfaces. Our aim is to compare calculated abundance ratios with observed values and with the results of prior models. Methods. Our model utilises the continuous-time, random-walk Monte Carlo method, which - unlike other approaches - is able to simulate microscopic grain-surface chemistry over the long timescales in interstellar space, including the layering of ices during freeze-out. Results. Simulations under different conditions, including density and temperature, have been performed. We find that H2CO and CH3OH form efficiently in cold dense cores or the cold outer envelopes of young stellar objects. The grain mantle is found to have a layered structure with CH3OH on top. The species CO and H2CO are found to exist predominantly in the lower layers of ice mantles where they are not available for hydrogenation at late times. This finding is in contrast with previous gas-grain models, which do not take into account the layering of the ice. Some of our results can be reproduced by a simple quasi-steady-state analytical model that focuses on the outer layer. Conclusions. Observational solid H2CO/CH3OH and CO/CH3OH abundance ratios in the outer envelopes of an assortment of young stellar objects agree reasonably well with our model results, which also suggest that the large range in CH3OH/H2O observed abundance ratios is due to variations in the evolutionary stages. Finally, we conclude that the limited chemical network used here for surface reactions apparently does not alter the overall conclusions.