Effect of incidence kinetic energy and surface coverage on the dissociative chemisorption of oxygen on W(110)

The dissociative chemisorption of oxygen on W(110) has been studied using molecular beam techniques. Chemisorption probabilities have been measured as a function of incidence angle, θi, and kinetic energy, Ei, and of surface coverage and temperature. In addition, angular scattering distributions have been measured for a range of conditions and LEED has been used to examine surface structure. The initial (zero coverage limit) sticking probability is found to depend strongly on the incidence energy, scaling with En=Ei cos2 θi. This probability is ∼10% at En =0.1 eV, rising to essentially unity above En =0.4 eV. At half a monolayer coverage of atomic oxygen, the sticking probability is close to zero up to a threshold of ∼0.25 eV, above which it rises to over 50% by 1.3 eV. In most cases, the sticking probability is found to fall roughly linearly with increasing surface coverage. However, a less-than-linear fall-off is observed for En ≥1 eV and for En ≤0.03 eV, the sticking probability actually rises with increasing coverage reaching a maximum at ∼0.2 ML. These results indicate that while dissociation may proceed via a classical precursor at the lowest energies, such a state can play little role for En≥0.1 eV. For En≤0.3 eV, the chemisorption probability falls to less than 5% for a coverage of about 0.5 ML; however, this apparent saturation coverage rises to 0.75 ML at ∼0.25 eV and to about 1.0 ML at about 0.85 eV. These ‘‘favored’’ coverages of 0.5, 0.75, and 1.0 ML are found to be associated with p(2×1), p(2×2), and p(1×1) LEED patterns, respectively. Angular scattering distributions recorded with a differentially pumped rotatable mass spectrometer, revealed predominantly quasispecular peaks, and velocity distributions are also characteristic of direct-inelastic scattering. The variation of the sticking probability with En is analyzed to obtain barrier height distributions for the clean and half-covered surfaces and these results are used to predict the sticking probability as a function of coverage and also to predict the initial sticking probability for adsorption from ambient gas or an effusive molecular beam. Reasonable agreement is obtained with available data in each case.