Search

Energy dissipation during chemical reactions at metal surfaces may excite electron-hole pairs in the metal. Direct detection of such reaction-induced hot electrons and holes is feasible using solid state barrier devices like Schottky diodes with ultrathin metal films. While exposing the diodes to reactive gases, a chemicurrent is observed in the diodes. The concept of hot charge carrier detection by chemicurrent measurements and the dependence of the current strength on device properties are discussed in detail. Data recorded from thin film Cu/n-Si(1 1 1) and Ag/n-Si(1 1 1) diodes exposed to atomic hydrogen and atomic oxygen are presented. The current detection sensitivity is improved by a factor of 10 if the metal films are annealed to room temperature after low-temperature deposition. This annealing effect is related to a reduced scattering of hot electrons in the metal. Chemicurrents are attenuated exponentially with increasing metal film thickness. Attenuation lengths between 6 and 11 nm are observed. They are much smaller than attenuation lengths for photo- and internal photoemission currents. The results demonstrate that chemicurrents are due to hot charge carrier excitation and transport and are not attributed to surface chemiluminescence and photon reabsorption in the device.

Non-adiabatic energy dissipation during exothermic chemical reactions on metal surfaces occurs by creation of electron–hole pairs in the metal. The excited charge carriers have been directly detected using metal–silicon (Schottky) diodes with ultrathin metal films. The chemically created hot electrons travel ballistically through the metal film, traverse the Schottky barrier and are detected as a chemicurrent in the diode. Three examples are presented, i.e. the adsorption of atomic hydrogen on Ag and Fe and the chemisorption of molecular oxygen on Ag. The chemicurrent transients upon exposure are related to the kinetics of the surface reaction. The large difference in the detection efficiency between Ag/Si and Fe/Si diodes is attributed to different conditions at the metal/silicon interface.

Heteroepitaxy of InP on InGaP/GaAs(001) has been studied using atomic force microscopy (AFM) as a function of substrate temperature and misorientation. Stranski-Krastanov growth of InP on GaAs results in three types and sizes of islands. The smallest and medium-sized islands are coherently strained to the substrate, whereas dislocations are seen in the largest islands. Growth, morphology and island separation distributions have been studied for the small and medium islands. Scaling in the radial distribution function indicates a uniformly random growth of medium-sized InP dots on all the substrates. However, on substrates grown at 650°C and misoriented by 2°, clustering of islands is also observed. We suggest that is due to the preferential growth of mediumsized islands at the step edges of the substrate. Separation distributions along and perpendicular to the step edges of this surface reveal that the step edges are decorated by clusters of 6–8 islands having an average period of 2 μm between them. These clusters exhibit long-range ordering and have a potential for device applications. Radial separation distribution studies indicate that the substrate morphology and topography are crucial for achieving a desirable distribution of islands. Moreover, the small InP islands and the dislocated islands are uniformly distributed on all the substrates irrespective of the growth temperature or misorientation, suggesting a different nucleation mechanism than for the medium-sized InP islands.

Rates of desorption of CO chemisorbed on Ir(111) were investigated using time-resolved Fourier-transform IR reflection-absorption spectroscopy in combination with an isotopic jump method. Data for desorption into vacuum and desorption in the presence of an ambient CO background pressure were measured at surface temperatures from 390 to 500 K and for a wide range of CO pressures from 1 × 10 −8 to 1 × 10 −5 mbar, extending the range of control parameters substantially compared to all other previously reported investigations. The rates of desorption in both cases are essentially identical, and the existence of the frequently discussed phenomenon of “adsorption-assisted desorption” can be excluded for CO on Ir(111).

The power spectral density (PSD) of a three-dimensionally islanded InP surface resulting from coherent Stranski-Krastanov growth is examined and interpreted by applying the mathematical formalism which describes the scattering intensity in the kinematic approximation of small-angle X-ray scattering. Specifically, the morphology of the InP surface is approximated by an ensemble of two-dimensional circles superimposed on a field of zero height, which leads directly to the mathematical equivalence of the isotropic PSD and the scattering intensity. The PSD for a simple model displays oscillations around a λ3 dependence in the limit of small wavelengths, which can be explained from a purely geometrical perspective (Porod-Debye law). Deconvolution of the spatial correlation and island size distributions requires an evaluation of the interference in terms of a “hard-sphere” radial distribution function, analogous to Percus-Yevick theory for the interparticle structure factor, and yields close agreement with real-space measurements. The radius of impenetrability (0.095 μm) implied by the power spectral density agrees with the expected value corresponding to approximately twice the average island radius.

The adsorption of gas-phase atomic hydrogen on potassium-precovered Ir(111) surfaces was investigated. Even very low coverages of potassium adatoms strongly inhibit the dissociative adsorption of molecular hydrogen. However, using gas-phase atomic hydrogen allows us to overcome the activation barrier for dissociative hydrogen adsorption. In addition, abstraction of hydrogen adatoms by impinging atomic hydrogen occurs. The probabilities and cross-sections for both reactions and the maximum number of hydrogen adsorption sites are derived and compared to data obtained on other surfaces. Furthermore, a kinetic isotope effect in the desorption of hydrogen and deuterium was observed. Implications of these results with respect to the potassium-hydrogen interaction are discussed.

The adsorption of CO on Ir(111) has been investigated with Fourier transform infrared reflection-absorption spectroscopy, temperature programmed desorption, and low-energy electron diffraction. At sample temperatures between 90 and 350 K, only a single absorption band, above 2000 cm−1, has been observed at all CO coverages. For fractional coverages above approximately 0.2, the bandwidth becomes as narrow as 5.5 cm−1. The linewidth is attributed mainly to inhomogeneous broadening at low CO coverages and to the creation of electron-hole pairs at higher CO coverages. The coverage-dependent frequency shift of the IR band can be described quantitatively using an improved dipolar coupling model. The contribution of the dipole shift and the chemical shift to the total frequency shift were separated using isotopic mixtures of CO. The chemical shift is positive with a constant value of approximately 12 cm−1 for all coverages, whereas the dipole shift increases with coverage up to a value of 36 cm−1 at a coverage of 0.5 ML.

Atomic force microscopy has been used to investigate the metalorganic chemical vapor deposition of InP on nominally flat InGaP GaAs (001) substrates as a function of deposition temperature. In particular, the deposition time required for island formation as a function of deposition temperature is determined at a constant reactant flux, and the effect of InP deposition temperature on the island shape and size distributions is explored. The qualitative behavior of the growth changes very little with temperature, whereas the size and spatial distributions are significantly affected. In addition, the effect of InGaP surface morphology on the size and density of InP islands is compared at constant deposition times. From a detailed analysis of the islands on the various surfaces, the island density and base area are observed to be strong functions of surface morphology, while the height of the islands remains unchanged. These observations indicate that the size, uniformity, and density of the coherent islands are not dominated by strain energy, and, thus, they can be manipulated by changes in substrate morphology and growth conditions.

Atomic force microscopy and transmission electron microscopy have been used to investigate the metalorganic chemical vapor deposition of InP on nominally flat InGaPGaAs(001) substrates. From a detailed structural analysis of the deposited InP, we observe coherent Stranski-Krastanov growth, i.e., after the initial formation of a two-dimensional layer, two types of coherently strained islands appear above a defined thickness. At higher coverages a third type of larger, incoherent and defected island is observed. The two types of coherently strained islands can be characterized by their distinct sizes. For low coverages the coherently strained islands possess irregular dimensions in the plane of the substrate and have a maximum height of 6 nm. At higher coverages the second type of coherently strained islands appears. These are cap-shaped and have a rather narrow distribution of base diameter (120 ± 10 nm) and height (22.5 ± 2.5 nm). An appreciable density of incoherent and defected relaxed islands also becomes evident at increased coverages. These have a hexagonal shape exposing clear facets and are elongated along the [110] crystallographic direction of the substrate.

We have carried out measurements of the steady-state rate of decomposition of formaldehyde on the Pt(110)-(1 × 2) surface, and we have interpreted these measurements in such a way as to permit an incisive probe of the molecular precursor to dissociation. The rate coefficient of the decomposition reaction of molecularly adsorbed formaldehyde (H2CO) was found to be 2 × 1012 exp[( −13.4 kcal mol )/kBT] s−1.

The heterogeneously catalyzed reaction between NO and D2 to produce N2, ND3 and D2O over Ir(110) was investigated under ultra-high vacuum conditions for partial pressures of the reactants between 5 × 10−8 and 1 × 10−6Torr, total pressures between 10−7 and 10−6 Torr, and surface temperatures between 300 and 1000 K. Mass spectrometry, LEED, UPS, XPS and AES measurements were used to study this reacting system. In addition, the competitive coadsorption of NO and deuterium was investigated via thermal desorption mass spectrometry and contact potential difference measurements to gain further insight into the observed steady state rates of reaction. Depending on the ratio of partial pressures ( R  P D 2 P NO ), the rate of reduction of NO to N2 shows a pronounced enhancement when the surface is heated above a critical temperature. As the surface is cooled, the rate maintains a high value independent of temperature until a lower critical temperature is reached, where the rate drops precipitously. This hysteresis is due to a change in the structure and composition of the surface. For sufficiently large values of R and for an “activated” surface, N2 and ND3 are produced competitively between 470 and 630 K. Empirical models of the different regimes of the steady state reaction are presented with interpretations of these models.

The adsorption of H atoms on metal (jellium) surfaces has been investigated using linear response theory within the density functional formalism. The adsorbate is represented initially by a 1S orbital on the H atom, which perturbs the jellium surface and indirectly the adsorbate itself. The interaction energy curves, atomic binding energies, induced dipole moments, chemical shifts associated with the adsorbate, and vibrational excitation energies at the equilibrium internuclear separation have been calculated for a single H atom chemisorbed on metal surfaces. The sum of the atomic binding energy and the ionization potential of the H 1S level may be regarded as the initial state energy in the case of photoemission from the chemisorbed H. The rather satisfactory overall agreement between the theory and the experimental results for binding energies, vibrational excitation energies, and dipole moments suggests that this simple formalism could also have useful applications in more complicated chemisorbed systems.

The molecular chemisorption of N2 on the reconstructed Ir(110)-(1 × 2) surface has been studied with thermal desorption mass spectrometry, XPS, UPS, AES, LEED and the co-adsorption of N2 with hydrogen. Photoelectron spectroscopy shows molecular levels of N2 at 8.0 (5σ + 1π) and 11.8 (4σ) eV in the valence band and at 399.2 eV with a satellite at 404.2 eV in the N(1s) region, where the binding energies are referenced to the Ir Fermi level. The kinetics of adsorption and desorption show that both precursor kinetics and interadsorbate interactions are important for this chemisorption system. Adsorption occurs with a constant probability of adsorption of unity up to saturation coverage (4.8 × 1014 cm−2), and the thermal desorption spectra give rise to two peaks. The activation energy for desorption varies between 8.5 and 6.0 kcal mole−1 at low and high coverages, respectively. Results of the co-adsorption of N2 and hydrogen indicate that adsorbed N2 resides in the missing-row troughs on the reconstructed surface. Nitrogen is displaced by hydrogen, and the most tightly bound state of hydrogen blocks virtually all N2 adsorption. A p1g1(2 × 2) LEED pattern is associated with a saturated overlayer of adsorbed N2 on Ir(110)-(1 × 2).

The chemisorption of oxygen on Ir(110) has been investigated under ultra-high vacuum conditions with thermal desorption mass spectrometry, contact potential difference measurements, Auger electron spectroscopy, and LEED. Oxygen may adsorb in three distinct chemical states: a molecularly chemisorbed species that is stable below 100 K, a dissociatively chemisorbed species, and a surface oxide that forms rapidly above 700 K. The oxide layer saturates at a coverage of 0.25 ML (1 ML = 9.6 × 10 14 atoms / cm 2 ) and orders to form a (1 × 4) LEED pattern. Different LEED patterns of dissociatively chemisorbed oxygen are observed on clean and oxidized Ir(110). A p(2 × 2) pattern forms on the clean surface near 0.25 ML coverage whereas a c(2 × 2) pattern forms on the oxidized surface near 0.5 ML coverage. Oxygen desorbs molecularly from Ir(110) with an activation energy of 45–70 kcal/mole, decreasing continuously with increasing coverage. The adsorption kinetics are described by a second-order precursor model for surface temperatures between 300–700 K. Oxygen chemisorption is not activated since the initial sticking probabilities on the clean and the oxidized Ir(110) surfaces are equal to 0.28 and 0.4, irrespective of the surface temperature. The dipole moment and polarizability of dissociatively chemisorbed oxygen change at 0.25 ML coverage on the clean surface and at 0.5 ML coverage on the oxidized surface. Although the dipole moment for any coverage is independent of temperature, the polarizability is inversely proportional to temperature. The activation energy for the dissociation of molecularly chemisorbed oxygen is 8 kcal/mole.

The chemisorption of NO on Ir(110) has been studied with thermal desorption mass spectrometry (including isotopic exchange experiments), X-ray and UV-photoelectron spectroscopies, Auger electron spectroscopy,LEED and CPD measurements. Chemisorption of NO proceeds by precursor kinetics with the initial probability of adsorption equal to unity independent of surface temperature. Saturation coverage of molecular NO corresponds to 9.6 × 10 14 cm −2 below 300 K. Approximately 35% of the saturated layer desorbs as NO in two well separated features of equal integrated intensity in the thermal desorption spectra. The balance of the NO desorbs as N 2 and O 2 with desorption of N 2 beginning after the low-temperature peak of NO has desorbed almost completely. Molecular NO desorbs with activation energies of 23.4–28.9 and 32.5–40.1 kcal mole −1 , assuming the preexponential factor for both processes is between 10 13 –10 16 s −1 . At low coverages of NO, N 2 desorbs with an activation energy of 36–45 kcal mole −1 , assuming the preexponential factor is between 10 −2 and 10 cm 2 s −1 . Levels at 13.5, 10.4 and 8.5 eV below the Fermi level are observed with HeI UPS, associated with the 4σ, 5σ and 1π orbitals of NO, respectively. Core levels of NO appear at 531.5 eV [O(1s)] and 400.2 eV [N(1s)], and do not shift in the presence of oxygen. Oxygen overlayers tend to stabilize chemisorbed NO as reflected in thermal desorption spectra and a downshift in the 1π level to 9.5 eV.

Inelastic electron tunneling spectroscopy has been applied to characterize both supported reduced metallic particles and supported homogeneous cluster catalysts. In particular, the interaction of C2H5OH with clusters of Ag supported on Al2O3, and of Zr(BH4)4, [RhCl(CO)2]2 and Ru3(CO)12 on Al2O3 have been examined, including the interaction of C2H4, C3H6 and C2H2 with the supported polymerization catalyst, Zr(BH4)4.

The structure of the reconstructed Ir(110)-(1 × 2) surface has been analyzed by low-energy electron diffraction. The reliability (R)-factor analysis, proposed by Zanazzi and Jona, has been applied to determine quantitatively the level of agreement between the experimental and calculated beam intensities for different models proposed for the (1 × 2) structure. The models tested were the following: (1) the missing row model, (2) the missing row model with a slight movement of the second layer, (3) the paired-rows model, and (4) the buckled surface model. Based on the results of the R-factor analysis, the missing row model with a topmost layer spacing of 1.16 ± 0.07 A, which corresponds to approximately 15% contraction of the bulk interlayer spacing of 1.36 A, is the preferred structure.

Isotopic exchange measurements have shown that at surface temperatures below 800 K, the upper limit on the probability of dissociation/recombination of CO on Rh(111) is approximately 10 −3 . Related Auger spectroscopic measurements, monitoring the accumulation of atomic carbon on the surface, have shown that the maximum probability of dissociation of CO per collision with the Rh surface in a temperature range between 300 and 870 K is 10 −4 . Hence, it has been demonstrated that the probability of dissociation of CO on Rh(111) is negligible compared to the probability of molecular desorption, in contrast to recent reports in the literature concerning both polycrystalline Rh and stepped surfaces of Rh.

Experimental (LEED) intensity-voltage (I–V) beam profiles for clean Ni(110), Al(110) and Ag(110) have been analyzed by the convolution-transform method. Results indicate that the first layer spacing of Ni(110), Al(110) and Ag(110) is contracted by 5%, 4% and 7% of their bulk spacing, respectively. These results are in good agreement with the results obtained earlier using dynamical calculations.

The oxidation of carbon monoxide over a Ag(111) catalyst has been studied by XPS and UPS. The kinetics have been determined over the temperature range of 180 to 400 K and found to be of the Langmuir-Hinshelwood type, although the Eley-Rideal mechanism is mimicked. A negative activation energy, −1.7 kcal/mole, and a preexponential, 6 × 10 −18 cm 2 , are found. The former corresponds to the difference in the activation energies for carbon monoxide desorption and for carbon monoxide oxidation (leading to CO 2 desorption). At 90 K, upon carbon monoxide exposure to the active oxygen precovered surface, the O ls and C ls spectral regions show the formation of CO 2 -like and carbonate species; the latter is stable to at least room temperature. That is, at 90 K, the residence time and mobility of CO 2 formed at the surface permits a new surface reaction — the formation of stable surface carbonate. The identifications are based on C and O coverages and on line positions from the literature for Cu/CO 2 and several bulk carbonates. With UPS, the 1π g , the unresolved doublet 1π u and 3σ g , and the 4σ g molecular orbitals of adsorbed CO 2 -like species are identified, as well as the unresolved triplet 1α′ 2 , 1e″ and 4e′ and the unresolved triplet 3e′, 1α″ 2 and 4a′ molecular orbitals of the carbonate species. Surface CO 2 -like species formed by surface oxidation of CO seem to be more strongly bound than reversibly adsorbed CO 2 .

We have investigated the insulating barrier of A1-oxidized A1Pb tunnel junctions. This study was effected by measuring the inelastic electron tunneling (IET) spectra of various junctions in which the insulator was formed by oxidizing the aluminum either in an O2 H2O or D2O plasma discharge at various temperatures between room temperature and 250°C. We found that those insulators formed at all temperatures in 02 and those formed at high temperatures (≳150°C) in H2O and D2O are aluminum oxide with no evidence for any appreciable bulk hydroxide. The surfaces of these insulators do have adsorbed hydroxyl groups present, however. The insulators made at lower temperatures (

Election beam induced perturbations of CO chemisorbed on Ir(111) have been measured using LEED and AES. The total interaction cross-section for electron-stimulated desorption and dissociation is found to be 0.8 to 1.7 × 10−17 cm2 near 1 3 monolayer coverage at a beam energy of 86 eV. This total cross-section is estimated to be 1 × 10−17 cm2 when defined with respect to the primary flux of a 2.5 keV beam. Electron-stimulated dissociation is found to occur at 1–2% of the rate of stimulated desorption.

Elastic low-energy electron diffraction (LEED) intensity versus voltage ( I - V ) measurements for the clean Ir(111) surface have been obtained. Seven specular I - V spectra were measured from 15 to 975 eV at incident angles from 7° to 62.5° relative to the surface normal. The outermost atomic layer spacing of the unreconstructed Ir(111) surface was determined both by the convolution-transform method we have presented previously (including certain convenient modifications) and by dynamical calculations. Results from the analysis of the I - V spectra by the convolution-transform method indicate that the outermost Ir(111) layer spacing is either unrelaxed or contracts by 4% of its bulk value depending upon whether the θ=7° data or the θ = 25° data are used. In agreement with this, the dynamical calculations show that the outermost Ir(111) layer spacing contracts by 2.5± 5% and, in addition, that the registry of the first layer of the crystal surface is not shifted, maintaining the fcc structure.

The adsorption of a proton on metal (jellium) surfaces has been investigated using a similar approach to that described in a previous paper for hydrogen chemisorption. The screening parameter of the induced electronic charge around the proton was determined by minimizing the total interaction energy variationally. This parameter was then used to obtain the induced electronic charge density and to evaluate the electrostatic potentials involved in chemisorption. The calculated binding energy was then regarded as the excited final state energy in the case of photoemission from the chemisorbed hydrogen. Incorporating the initial state energies and chemical shifts which were obtained in a previous paper, the resonance levels and relaxation energies associated with the hydrogen chemisorbed on different jellium substrates have been obtained.

The heterogeneously catalyzed reaction of CO and O2 to form CO2 over Ir(110) has been studied through measurements of the transient kinetics of the various elementary reactions that may limit the steady state rate. Rate expressions for these elementary reactions — the desorption of CO, the oxidation of CO via the Langmuir-Hinshelwood mechanism, the adsorption of CO and the adsorption of oxygen — were developed using thermal desorption mass spectrometry. Several phenomena were observed: (1) the activation energies for CO desorption and CO oxidation depend markedly upon the composition of the adlayer; (2) diffusion in the adlayer may limit the rates of CO desorption and CO oxidation; (3) the formation of a surface oxide modifies these four rate processes; and (4) chemisorbed CO blocks sites for oxygen adsorption, but chemisorbed oxygen does not block sites for CO adsorption.

In an attempt to determine the influence of the presence of adjacent islands on LEED beam profiles, an excluded area model has been proposed to describe the relative positions of ordered islands on a crystal surface. On the basis of this model, overlayers of different island density have been created on a finite lattice using a simple computer algorithm. The LEED from these overlayers has been calculated kinematically. Although the placement of the islands is not random, this does not perturb the LEED beam profile observably. Therefore, the kinematically diffracted intensity depends solely on the distribution of island sizes. Using this result, the intensity and half-width as functions of coverage have been calculated for one model of island growth.

An Ir(110)-c(2 × 2)O structure has been prepared by adsorbing a half-monolayer of oxygen at room temperature on an unreconstructed (1 × 1)Ir surface stabilized by a quarter-monolayer of randomly adsorbed oxygen. Results of the low energy electron diffraction structural analysis indicate that the ordered oxygen atoms are residing on the short-bridged sites on the (110) surface. The Ir-O interlayer spacing is 1.37 ± 0.05 A, and the bond length is 1.93 ± 0.07 A. The topmost substrate interlayer spacing is found to be 1.33 ± 0.07 A rather than 1.26 ± 0.07 A which is the topmost interlayer spacing of the unreconstructed (1× 1)Ir surface.

Surface temperature modulation has been employed to investigate the adsorption, desorption and oxidation of CO on the Pt(110)-(1 × 2) surface. Examination of the adsorption-desorption equilibrium allows a determination of the kinetic parameters describing the desorption reaction. In the limit of zero coverage, the activation energy of desorption is 36±1.5 kcal mol−1 with a first-order preexponential factor of 3 × 1014 s−1. As the coverage of CO increases, both parameters decrease in magnitude. The oxidation of CO has been studied with an emphasis on examining the kinetics of the elementary “bimolecular” surface reaction CO(a)+O(a) → CO2(g). By varying both the reactant partial pressure ratio and the steady-state temperature of the surface, the reaction rate parameters have been evaluated for a variety of combinations of adsorbate concentrations occurring at steady state. The magnitudes of both the activation energy and the preexponential factor of the reaction rate coefficient have been found to correlate well with the oxygen adatom concentration. In the low-coverage limit, the activation energy is 22±1 kcal mol−1, and the preexponential factor is 3 × 10−3 cm2 s−1. As the oxygen adatom concentration increases beyond a fractional coverage of approximately 0.15, the magnitudes of both parameters drop precipitously to values given, respectively, by 8±1 kcal mol−1 and 10−10 cm2 s−1. The decrease in the activation energy at high oxygen coverages is explained best by a decrease in the binding energies of the adsorbed reactants. The apparent decrease in the preexponential factor could well be due to the local configuration of the reactants that occurs at high oxygen coverages.

The adsorption of oxygen in Ir(111) has been investigated under ultrahigh vacuum conditions. Thermal desorption spectra of O 2 indicate a coverage-dependent activation energy for desorption of E d ≅ (65−10 Θ ) kcal/mole. The adsorption proceeds via a highly mobile precursor state giving a constant value for the sticking probability until saturation coverage is nearly reached. The work function change at saturation coverage of oxygen is Δφ = 0.56 ± 0.04 eV. The induction period necessary for the formation of a (2 × 2) LEED pattern indicates that the activation energy for diffusion of the oxygen adatoms on Ir(111) is in the range 16–19 kcal/mole Upon heating the crystal to moderate temperatures (e.g., ∼700K), the surface oxygen is partially converted into an unreactive oxide which is present in the near surface region.

The detailed energetics embodied in the interaction of CO 2 and Rh are quantified using thermodynamic and kinetic information which were measured on initially clean Rh surfaces together with the principle of microscopic reversibility. At low pressures (of order 10 −6 Torr) and at room temperature, the probability of dissociative adsorption of CO 2 is shown to be on the order of 10 −15 , i.e. approximately 14 orders of magnitude lower than that reported recently in the literature. However, it is shown also that at a pressure of one atmosphere. CO 2 does begin to dissociate at a sensibly rapid rate on Rh at temperatures near 500 K. The methodologies outlined here have utility beyond the particular example of the interaction of CO 2 with Rh.

A model describing adsorption which combines features of the bond-energy bond-order relationships of gas phase kinetics and spectroscopy, and the concept of crystal field orbitals has been formulated. This model is thus termed the crystal field surface orbital — bond energy bond order (CFSO-BEBO) model of chemisorption. The calculation is empirical in nature since single bond energies for metal-metal and metal-gas bonds are deduced from available data on bulk compounds or transition metal complexes. Alternatively, the single order bond energies may be estimated from suitable mixing rules or deduced from sublimation energies and spectroscopic gas phase data. Spectroscopic data are also used for the bond energies of the stable gas phase species and for the correlation between bond energy and bond order. The bond-energy bond-order correlation is assumed to be linear for metal-metal and metal-gas bonds. This model is used to describe the interaction of hydrogen with platinum (111) in a detailed way, i.e., the potential energy of interaction along the reaction (or adsorption) coordinate. The trapping probability of hydrogen on platinum is consistent with this CFSO-BEBO model, and the energies of the two predicted chemical adsorption states correspond to the observed experimental heats of adsorption, 17 and 31 kcal/mole, with the low energy state being molecular. LEED and flash desorption experiments also give strong support for the calculated potential energy along the reaction coordinate.

The Crystal Field Surface Orbital-Bond Energy Bond Order (CFSO-BEBO) model of chemisorption is applied to the interaction of carbon monoxide, oxygen and carbon dioxide with a (111) platinum surface; and the interaction of oxygen with a (111) nickel surface. No activation energy for molecular adsorption of carbon monoxide on platinum is predicted; however a large activation energy for dissociative chemisorption is calculated. The molecular state has a binding energy of approximately 28 kcal/mole, and vibrational stretching frequencies of 1935 and 1975 cm−1 are calculated by combining the CFSO-BEBO model with Badger's Rule. The adsorption of oxygen on (111) platinum and (111) nickel are predicted to be different in the following respects: (1) There is an activation energy of 2.1–4.5 kcal/mole on platinum, whereas adsorption on nickel is unactivated; and (2) The dissociative heat of chemisorption on platinum is 58–68 kcal/mole, whereas on nickel it is substantially larger, 112–116 kcal/mole. The adsorption of carbon dioxide on (111) platinum is predicted to be not only highly activated but also endothermic. All of the calculated results are essentially in quantitative agreement with available experimental data.

The structure and chemistry of ethylene adsorbed on a platinum (111) surface has been studied. Ethylene adsorbs irreversibly and dissociatively with unity sticking coefficient in the first monolayer, which is ordered in a (2 × 2) array. A second monolayer in registry with the first is adsorbed reversibly. Self-hydrogenation takes place during dynamic adsorption-desorption which poisons the reversible adsorptivity. The reversibly-held ethylene desorbs at the lowest temperatures (Ed = 25 kcal) followed by dissociated hydrogen (Ed = 30 kcal) and the decomposition of hydrogen in the acetylenic first monolayer (Ed = 38 kcal). Residual carbon coalesces to graphite at temperatures in excess of 500°C. CFSO-BEBO calculations qualitatively predict all of the binding states observed. Treating PtC valency as a single adjustable parameter, all four binding energies and four of the five activation energies are in quantitative agreement with experiment.