Continuous-time random-walk simulation of H2 formation on interstellar grains

The formation of molecular hydrogen via the recombination of hydrogen atoms on interstellar grain surfaces has been investigated anew. A detailed Monte Carlo procedure known as the continuous- time random- walk method has been used. This Monte Carlo approach has two advantages over the stochastic master equation method: it treats random walk on a surface correctly, and it can easily be used for inhomogeneous surfaces. The recombination efficiency for H-2 formation as a function of surface temperature and grain size has been calculated for a variety of grain surfaces with a flux of hydrogen atoms representative of diffuse interstellar clouds. The surfaces studied include homogeneous olivine and amorphous carbon, characterized by single energies for the diffusion barrier and binding energy of H atoms, inhomogeneous versions of these two surfaces with distributions of H- atom diffusion barriers and binding energies, and a variety of mixed surfaces of olivine and carbon. For the homogeneous surfaces, we confirm that the temperature range for efficient formation of H-2 is very small. At temperatures near peak efficiency, there is little dependence on grain size. At temperatures higher than those of peak efficiency, the Monte Carlo procedure exhibits smaller efficiencies for molecular hydrogen formation than the master equation method in the limit of large grain sizes. For various types of inhomogeneous and mixed surfaces, the major effect we find is an increase in the temperature range over which the efficiency of molecular hydrogen formation is high. Efficient formation of H-2 in diffuse interstellar clouds now seems possible with inhomogeneous grains.