Christopher R. Arumainayagam

Molecular beam studies of gas-surface collision dynamics

Abstract Recent progress in the application of supersonic molecular beam techniques to the study of gas-surface interfacial phenomena is reviewed. The experimental and theoretical studies discussed examine fundamental issues regarding both non-reactive and reactive scattering. Topics discussed include elastic scattering (thermal energy atom scattering and diffractive selective adsorption), inelastic scattering, trapping-desorption, molecular chemisorption and desorption, dissociative chemisorption (direct collisional activation, precursor-mediated dissociation, and collision-induced dissociation of adsorbed species) and chemical reactions at surfaces. The review is organized around dynamical variables rather than particular gas-surface systems to emphasize the underlying physical principles governing gas-surface interactions. An index is provided at the beginning for the reader interested in specific gas-surface systems.

Dissociative chemisorption of methane on Pt(111)

Abstract Dissociative chemisorption of methane on clean Pt(111) was studied with a supersonic molecular beam. Initial dissociative sticking probabilities from 0.01 to 0.19 were measured directly with incident total beam energies from 68 to 95 kJ/mol, surface temperatures from 500 to 1250 K, and angles of incidence from 0° to 45° measured from the surface normal. The nozzle temperature and stagnation pressure were both fixed so that the effect of translational energy at a fixed incident vibrational energy could be probed. The initial dissociative sticking probability of methane on clean Pt(111) equalled 0.06 ± 0.02 and was independent of surface temperature between 500 and 1250 K for a fixed normal incident kinetic energy of 68 kJ/mol, implying that dissociation proceeded via direct collisional activation rather than via trapping or precursor-mediated processes in this energy range. The initial dissociative sticking probability of methane on clean Pt(111) increased exponentially with increasing normal kinetic energy. The barrier height for C-H bond rupture by kinetic energy is 121 kJ/mol. The exponential dependence is consistent with a model for dissociative methane adsorption that involves quantum mechanical tunneling of a hydrogen atom through a one-dimensional, parabolic barrier of this height with a thickness at half height of 0.13 ± 0.01 A. Differences in the initial dissociative sticking probabilities observed on Pt(111) versus Ni(111) and W(110) in studies from different laboratories can be reconciled on the basis of the different vibrational energies employed, but this explanation does not account for the high reactivity on Ir(110)-(1 × 2). The activation barriers predicted by the bond order conservation theory of Shustorovich agree closely with the barrier heights estimated from the tunneling model for Pt(111), W(110), and Ir(110) if no correction for energy dissipation to the lattice is made. The activation barriers predicted by the molecular orbital analysis of Anderson and Maloney are much lower than the barrier heights estimated from the tunneling model. The barrier thicknesses show qualitative agreement with the C-H bond elongations in the transition state predicted by their molecular orbital theory, but do not correlate with the crystalline radii of the metal atoms or the stiffness of the metal lattices. Furthermore, the values of the tunneling parameters may be dependent on the vibrational and translational energies employed in the studies.

Trapping dynamics of xenon on Pt(111)

The dynamics of Xe trapping on Pt(111) was studied using supersonic atomic beam techniques. Initial trapping probabilities (S0) were measured directly as a function of incident translational energy (EinT) and angle of incidence (θi) at a surface temperature (Tins) 95 K. The initial trapping probability decreases smoothly with increasing ET cosθi;, rather than ET cos2θi, suggesting participation of parallel momentum in the trapping process. Accordingly, the measured initial trapping probability falls off more slowly with increasing incident translational energy than predicted by one-dimensional theories. This finding is in near agreement with previous mean translational energy measurements for Xe desorbing near the Pt(111) surface normal, assuming detailed balance applies. Three-dimensional stochastic classical trajectory calculations presented herein also exhibit the importance of tangential momentum in trapping and satisfactorily reproduce the experimental initial trapping probabilities.

Dynamics of molecular CH4 adsorption on Pt(111)

Abstract The dynamics of molecular methane adsorption on Pt(111) were probed with supersonic molecular beam techniques. Initial trapping probabilities were directly measured between 0.94 and 0.16 for incident total translational energies between 3.4 and 20.2 kJ/mol and angles of incidence (with respect to the surface normal) between 0° and 45° at a surface temperature ( T s ) of 100 K. The incident methane molecules were rotationally and vibrationally cold. The initial trapping probability decreases with increasing incident translational energy ( E T ) and decreasing angle of incidence (θ i ) and varies smoothly with incident normal energy ( E n = E T cos 2 θ i ), indicating a low corrugation of the molecule-surf ace interaction potential. The dependence of the initial trapping probability on incident normal translational energy agrees quantitatively with both a modified hard cube model and Leuthaussers theory at incident normal translational energies below 8 kJ/mol. At higher incident normal translational energies the observed initial trapping probabilities are higher than the values predicted by both models. Energy loss mechanisms other than surface phonon excitations may partially account for this discrepancy. A rapid decrease in the apparent adsorption probability as the surface temperature approaches 140 K is caused by the competitive influence of desorption. The temperature at which the apparent adsorption probability goes to zero agrees well with the desorption temperature measured independently by temperature programmed desorption. In accordance with the aforementioned models, the measured in-plane angular distributions suggest that the trapping probability is relatively independent of surface temperature in the range of 160 to 500 K. The relatively low intensity of methane found near the surface normal in the angular distributions may be partially explained by a wider than cosine angular distribution for the trapped-desorbed channel, which is consistent with the observation that the trapping probability increases with angle of incidence. Comparison of our initial trapping probability versus normal translational energy data to previous mean translational energy measurements of methane molecules desorbing from Pt(111) at the surface normal suggests that detailed balance applies for the non-equilibrium situation involving a collimated monoenergetic molecular beam of methane incident on a Pt(111) surface.

Adsorbate‐assisted adsorption: Trapping dynamics of Xe on Pt(111) at nonzero coverages

The trappingdynamics of Xe on Pt(111) has been probed as a function of Xe coverage with supersonic molecular‐beam techniques. Adsorption probabilities were directly measured at a surface temperature of 95 K at coverages ranging from zero to monolayer saturation at incident translational energies between 6 and 63 kJ/mol and incident angles between 0° and 60°. In apparent agreement with the predictions of the original Kisliuk model, the adsorption probability at the lowest incident translational energy (6 kJ/mol) remains almost constant with coverage up to near monolayer saturation. However, in contradiction to the original Kisliuk model, at higher incident translational energies, the trapping probability increases nearly linearly with xenon coverage up to near monolayer coverage. For example, the trapping probability increases from 0.06 to 0.42 for an incident translational energy of 63 kJ/mol at normal incidence as the coverage is increased from zero to saturation monolayer coverage. This behavior can be explained adequately by a model that incorporates enhanced trapping onto the monolayer compared to the clean surface, a property of the model that is confirmed directly by experiments presented herein. The angular dependence of the adsorption probability shows progressive deviation from normal energy scaling with increasing Xe surface coverage, proving that the degree to which parallel momentum participates in the adsorption process increases with adsorbate coverage. The initial trapping probability of Xe onto the monolayer is independent of incident angle indicating total‐energy scaling. The above findings are qualitatively identical to our previous results for the molecular adsorption of ethane on the same surface, suggesting that these phenomena occur, in general, for weak molecular adsorption regardless of molecular shape and internal degrees of freedom, at least for small molecules.

Molecular propane adsorption dynamics on Pt(111)

Abstract Supersonic molecular beam techniques and temperature programmed desorption (TPD) were used to study the adsorption dynamics of propane onto clean and propane-covered Pt(111). The propane sticking probability was measured directly as a function of incident translational energy, E i , incident angle, θ i , and propane coverage, θ cov , at a surface temperature of 95 K. Under these experimental conditions, propane adsorbs molecularly onto Pt(111) terrace sites. Non-normal energy scaling is observed at all propane coverages indicating the importance of parallel momentum in the adsorption process. At all incident translational energies and angles studied, the sticking probability on a propane covered surface, S (θ cov ), increases with increasing propane coverage. Upon saturation of the Pt(111) terrace sites, spontaneous desorption is observed in the direct adsorption probability experiments.

IR spectroscopy of adsorbed dinitrogen : a sensitive probe of defect sites on Pt(111)

Fourier transform infrared reflection absorption spectroscopy (FT-IRAS) has been used to probe the non-dissociative adsorption of N2 on an atomically clean Pt(111) single crystal. In contradiction to a previous IRAS study of nitrogen adsorption on a Pt(111) foil at 120 K, no nitrogen infrared (IR) band was observed on a fully annealed Pt(111) surface at 90 K. Following Ar+ ion bombardment, adsorption of nitrogen at 90 K produces an intense IR band at ∼2222 cm−1 attributed to the NN stretching mode of molecular nitrogen adsorbed on defect sites produced by ion bombardment. Annealing the Ar+ ion sputtered surface to a temperature above ∼750 K completely suppresses the adsorption of nitrogen at 90 K. Based on these and other results, we postulate that nitrogen adsorbs at 90 K mainly on monovacancies on platinum. We suggest that this specific adsorption occurs by sigma donation from nitrogen to the base of monovacancy sites which possess a low d-electron density compared to surface Pt atoms.

The dynamics of precursor adsorption : ethane on Pt(111)

Abstract We have studied ethane trapping into a second-layer state on Pt(111) using supersonic molecular beam techniques to investigate the dynamics of extrinsic precursor adsorption. Initial trapping probabilities of ethane on an ethane covered Pt(111) surface were measured directly as a function of incident translational energy and incident angle at a surface temperature of 95 K. At all incident translational energies and angles the initial trapping probability into the second-layer state is higher than on a clean surface. In addition the initial trapping probability into the second layer decreases less with incident translational energy than the initial trapping probability onto the clean surface. In contrast to previous findings for non-dissociative weak adsorption on clean surfaces showing the initial trapping probability to increase with incident angle, the initial trapping probability into the second layer is independent of incident angle indicating “total” energy scaling. A dynamical corrugation of the adsorbed layer is postulated to rationalize this strong deviation from the “normal” energy scaling implicit in one-dimensional theories of trapping.

Molecular beam studies of adsorption dynamics

We have investigated the trapping dynamics of C1‐C3 alkanes and Xe on Pt(111) using supersonic molecular beams and a direct technique to measure trapping probabilities. We have extended a one‐dimensional model based on classical mechanics to include trapping and have found semiquantitative agreement with experimental results for the dependence of the initial trapping probability on incident translational energy at normal incidence. Our measurements of the initial trapping probability as a function of incident translational energy at normal incidence are in agreement with previous mean translational energy measurements for Xe and CH4 desorbing near the surface normal, in accordance with detailed balance. However, the angular dependence of the initial trapping probability shows deviations from normal energy scaling, demonstrating the importance of parallel momentum in the trapping process and the inadequacy of one‐dimensional models. The dependence of the initial trapping probability of Xe on incident trans...

An improved method to spot-weld difficult junctions

Recent advances in spot-welding technology such as high frequency direct current inverter welders provide an improved and reproducible method to spot-weld difficult junctions. The importance of removing the oxide layers on metal surfaces, accurately delivering the weld pulse profile, and controlling the force applied to the materials during the welding process are discussed in the context of resistance spot-welding a molybdenum crystal to a tantalum loop and attaching a tungsten–rhenium thermocouple to the crystal.