Molecular evolution in collapsing prestellar cores. II. The effect of grain-surface reactions

The molecular evolution that occurs in collapsing prestellar cores is investigated. To model the dynamics, we adopt the Larson-Penston solution and analogs with slower rates of collapse. For the chemistry, we utilize the new standard model with the addition of deuterium fractionation and grain-surface reactions treated via the modified rate approach. The use of surface reactions distinguishes the present work from our previous model. We find that these reactions efficiently produce H2O, H2CO, CH3OH, N-2, and NH3 ices. In addition, the surface chemistry influences the gas-phase abundances in a variety of ways. For example, formation of molecular nitrogen on grain surfaces followed by desorption into the gas enhances the abundance of this gas-phase species and its daughter products N2H+ and NH3. The current reaction network along with the Larson-Penston solution allows us to reproduce satisfactorily most of the molecular column densities and their radial distributions observed in L1544. The agreement tends to worsen with models that include strongly delayed collapse rates. Inferred radial distributions in terms of fractional abundances are somewhat harder to reproduce. In addition to our standard chemical model, we have also run a model with the UMIST gas-phase chemical network. The abundances of gas-phase sulphur-bearing molecules such as CS and CCS are significantly affected by uncertainties in the gas-phase chemical network. In all our models, the column density of N2H+ monotonically increases as the central density of the core increases during collapse from 3 x 10(4) to 3 x 10(7) cm(-3). Thus, the abundance of this ion can be a probe of evolutionary stage. Molecular D/H ratios in assorted cores are best reproduced in the Larson-Penston picture with the conventional rate coefficients for fractionation reactions. If we adopt the newly measured and calculated rate coefficients, the D/H ratios, especially N2D+/N2H+, become significantly lower than the observed values.