D2 dissociative adsorption on and associative desorption from Si(100): Dynamic consequences of an ab initio potential energy surface

Abstract

Dynamical calculations are reported for D 2 dissociative chemisorption on and associative desorption from a Si͑100͒ surface. These calculations use the dynamically relevant effective potential which is based on an ab initio potential energy surface for the ''pre-paired'' species. Three coordinates are included dynamically; the distance to the surface, the D-D bond length and a Si phonon coordinate. Other coordinates ͑multidimensionality͒ have been included via a static approximation. Both an asymmetric and symmetric reaction paths are considered. While energetics favors the asymmetric path, phase space favors the symmetric one. Under the conditions of many experiments, either could dominate. The calculations show quite weak dynamic coupling to the Si lattice for both paths, i.e., weak surface temperature dependences to dissociation and small energy loss to the lattice upon desorption. These calculations do not support previous suggestions that either a strong coupling to the lattice or ''entropic'' effects can reconcile the apparent violation of detailed balance obtained by comparing experimental dissociation to desorption barriers. In fact, the results reported here do not agree with several experimental findings. We discuss several possibilities for this disagreement, including experimental artifact, limitations in the dynamical model and even the possibility that electronically adiabatic dynamics involving the ''pre-paired'' species is not relevant to experiments on real systems.

Ab initio derived kinetic Monte Carlo model ofH2sdesorption from Si(100)-2×1

“…Both the temperature dependence of H 2 adsorption probabilities and the translational energy of desorbing H 2 were inconsistent with this mechanism. 16 In our previous work, we offer a detailed explanation of why we believe desorption occurs from isolated dihydrides, a mechanism which is also supported by the theoretical calculations of Nachtigall, Jordan, and Sosa 14 and Jing and Whitten. 15 It is our purpose in the present work to investigate whether desorption via this mechanism can follow first-order kinetics as we 7,8 and others 14,15 have proposed.…”

Interaction ofH2withSi(001)−(2×1): Solution of the Ba

“…The 2H mechanism dominates only at very low coverages, and the likely involvement of substrate vibrations could explain the pronounced activation of the low-coverage sticking probability by the sample temperature [9]. However, the requirement that this phonon-mediated mechanism also give rise to translationally accommodated desorbing molecules (which is in conflict with theoretical results [5,10]) can now be relaxed. Further study is required to determine the detailed dynamics of both pathways, as well as possible contributions of other pathways.…”

“…Because the buckling angle of surface Si dimers changes considerably during adsorption/ desorption of H 2 , it was hypothesized that during desorption the excess potential energy of the adsorption barrier may be transferred into lattice vibrations rather than the degrees of freedom of the H 2 molecule, explaining relatively cold desorbing molecules [4,7]. Ab initio molecular dynamics simulations [5,10] have made it clear, however, that coupling to the buckling coordinate does not correctly account for the observed desorption dynamics. The simulations show that although a fraction ͑ϳ0.1 eV͒ of the barrier energy can be transferred to the lattice, much more energy ͑ϳ0.55 eV͒ ends up in the translational energy of the desorbing molecule, contrary to experimental observation [5].…”

“…The discussion of this problem has recently focused on a promising proposed mechanism called "phonon-assisted sticking" [5,[8][9][10]. Because the buckling angle of surface Si dimers changes considerably during adsorption/ desorption of H 2 , it was hypothesized that during desorption the excess potential energy of the adsorption barrier may be transferred into lattice vibrations rather than the degrees of freedom of the H 2 molecule, explaining relatively cold desorbing molecules [4,7].…”

“…Surprisingly, however, H 2 molecules desorbing from Si(100) monohydride show no signs in their final state distributions of having traversed such a barrier [4,7], in apparent contradiction with microscopic reversibility. This paradoxical situation has puzzled the gas-surface dynamics community for many years (see [2] for a review), and has raised questions about the applicability of the principle of detailed balance [4].The discussion of this problem has recently focused on a promising proposed mechanism called "phonon-assisted sticking" [5,[8][9][10]. Because the buckling angle of surface Si dimers changes considerably during adsorption/ desorption of H 2 , it was hypothesized that during desorption the excess potential energy of the adsorption barrier may be transferred into lattice vibrations rather than the degrees of freedom of the H 2 molecule, explaining relatively cold desorbing molecules [4,7].…”