The impact of gaseous H atoms at D covered Pt(111) surfaces at 85 K leads to the formation of gaseous HD and D2 products. The kinetics of formation of these products was measured simultaneously with H exposure for different initial D coverages. The HD and D2 rates as a function of H fluence from the reaction start exhibit common characteristics; a rate step, a rate maximum, and a subsequent exponential rate decay. The HD and D2 rate steps were observed not to scale linear with the D coverage and to increase if on D covered surfaces H was coadsorbed prior to reaction. Of the observed phenomena, only the exponential decay of the HD rate is in line with expectations if an Eley–Rideal mechanism acts in the present reaction. D2 formation, the HD rate step as a function of D coverage, and the presence of a H coadsorbate contradict the Eley–Rideal picture. The results suggest that the reactions towards HD and D2 proceed via hot atom type mechanisms.

Recent experiments on the abstraction of D adsorbed on metal surfaces with gaseous hydrogen atoms revealed a kinetics of HD formation which is not compatible with the operation of Eley–Rideal (ER) mechanisms. Furthermore, homonuclear products were observed during abstraction, which are not expected through an ER reaction scheme. It was therefore suggested that hot-atom (HA) mechanisms are more appropriate to explain the measured kinetics and products. Random walk calculations of the abstraction kinetics are presented based on a model which exclusively relies on elementary reaction steps which are HA mediated processes. Within this model, the ratio of two variables, the probabilities for hot-atom sticking at empty sites $p_s$ and hot-atom reaction with adsorbed species $p_r,$ was found to control the kinetics of HD and ${\mathrm{D}}_2$ formation. The essential features of measured kinetic data at Ni(100), Pt(111), and Cu(111) surfaces were reproduced by simple and reasonable assumptions on $p_s{\mathrm{/p}}_r.$