Table of contents

Atom-probe FIMs are micro-analytical devices of ultimate sensitivity by combining the single atom resolution of the field-ion microscope with the single particle detection of a mass spectrometer. The original ToF version may achieve a resolution ΔM/M of 1/200 and is thus adequate for routine metallurgical applications. The resolution is limited by an inadequate subnanosecond pulse front and can be improved tenfold in an energy deficit compensating ToF atom-probe. A magnetic sector also offers high resolution capability. New surface effects discovered by the atom-probe and affecting image interpretation in field-ion microscopy are the occurrence of up to fourfold charged metal ions as products of field evaporation, the field-induced adsorption of noble gases, and the formation of metal-helium molecular ions.

Two electronic spectroscopies of solid surfaces, ion-neutralization spectroscopy and ultraviolet photoemission spectroscopy, are discussed with particular reference to their joint use in determining orbital energy spectra of electrons in surface adsorption complexes. The goal is the exploitation of the differences in matrix element and surface sensitivity when the two spectroscopies are applied to the same surface. Localization of surface orbitals and factors determining the orbital energies are discussed.

The recent progress in the studies of the electronic properties of semiconductor interface is reviewed briefly. The attention is paid especially on the Landau levels of quantized sub-bands in the inversion layer of MOS. The physics involved in the de Haas-Schubnikov effect, the enhancement of spin splitting by the exchange interaction of electrons, and the interband mixing effect in the narrow-gap-semiconductor are discussed by using simplified intuitive pictures.

The shape of a crystal in equilibrium or during growth is determined by the surface free energy and its anisotropy, by transport phenomena and reactions. Inversely such type of experiments should in principle enable the determination of elementary surface quantities if they are made under well known conditions. Crystals which form the end of a tip seems to be in particular suitable for it, because (1) calculations on shape changes of tips do exists and are confirmed in important details, (2) the cleanliness of the surface or the presence of a special adsorbed layer can be controled by field emission microscopy, (3) the variation of geometrical data (tip length, curvature radius, diameter of faces) can be measured in situ in ultra high vacuum, (4) as such crystals are small (diameter roughly between 0.1 and 10 µ) one shape change experiment requires only a short time. In this paper a review is presented, mainly results on metal crystals obtained in recent years by Piquet, Vu Thien, Roux, Uzan, Pichaud and A. Müller in cooperation with the author. Described are new methods using scanning and field emission microscopes. The theory of Nichols and Mullins on shape changes of conical tips by surface self-diffusion is confirmed experimentally, in particular the existence of the critical half cone angle α=3°, the formation of single crystal drops (α<3°), the evolution of "steady state" forms and the radius-time relation. Surface self-diffusion diffusivities and energies are measured. If an adsorbed layer is added to the clean surface, an essential increase or decrease of surface self-diffusion can occur. Shape changes can be considerably influenced by free evaporation and surface reactions (corrosion) which is shown theoretically as well as experimentally. The equilibrium shape and the surface free energy anisotropy is determined for several clean metals in agreement with calculated data.

The application of field ion microscopy to the study of interfaces has been extended to Ga, In and Sn on W, Mo, Ir and Re. Regularly ordered pseudomorphs and various superstructures were observed only for Ga–W and In–W. The observed high evaporation fields of Ga, Ir and Sn indicated the large binding energies at the interfaces. The deposited atoms were imaged with various brightnesses and might be invisible in the two-dimensional alloys. The disordered surface caused by a contact or by heating exhibited a higher evaporation field than that of the substrate. The effects of the deposits on the thermal end form and surface rearrangements were observed. The observed phenomena are explained as the result of transitions between electron states, the formation of two dimensional alloys and promoted migration induced by the increased surface energy.

The field ionization probability of imaging gas atom in the presence of the field adsorbed atom in the field ion microscope is calculated. By orthogonalizing the basis set of the system, the ionization probability between the defined initial and the final state is calculated by the Fermi-Golden rule. Two kinds of the transition matrix elements which represent the field adsorption effects are derived; one due to the overlap between the basis set and the other due to the Coulomb interaction in the perturbation potential. The calculation show that the adsorbed He atoms supress the ionization probability of imaging gas considerably and the adsorbed Ne and Ar supress and enhance it a little respectively. The ratio of the resultant ionization probability by the adsorption of Ne to that of He is about 10.

The experimental results and their analysis on the energy distribution of field ionized gaseous species on tungsten surfaces are presented. With the noble gases (He and Ne) the observed energy distribution on W(110) and W(100) shows the main peaks plus a series of field induced resonance peaks which was observed first by Jason. The observed shapes and positions of these resonance peaks are discussed. With hydrogen, it was shown that the field ionization of the molecular hydrogen is non-vertical Frank-Condon transition and the orientation of H₂ at the instant of the ionization is nearly random up to the field of 4 V/Å. Also our experimental observations and subsequent analysis of H₃⁺ formation strongly support that the reaction kinetics is H₂⁺ (field ionized) + H₂ (adsorbed due to dipole-dipole interaction on the protruding tungsten atom) →H₃⁺ + H.

Field ion specimens of iridium were exposed to methane at low pressure for controlled periods of a few seconds, with the field removed, and then imaged in neon or in helium. With neon microscopy, adsorption was clearly visible after exposure to 2×10^-9 Torr for 5–20 seconds. The rate of adsorption was measured on several different crystallographic planes. The <110> zones appeared to be very resistant to adsorption. Field desorption was done by raising the specimen voltage in controlled increments, and the field required to desorb the adsorbate was measured for several different planes. Some field-induced corrosion was seen on the {311} planes. With helium microscopy, adsorption was not visible except after rather larger exposures to 1×10^-8 Torr for 25 seconds. However, field-induced corrosion was seen after smaller exposures, suggesting that adsorption had occurred but that the adsorbate had been removed by the field. The {311} planes were particularly susceptible to corrosion.

The effects of negative high-electric fields of the order of 10⁷–10⁸ V/cm on oxidation of tungsten and molybdenum single crystals were investigated at 1200–1500 K for tungsten and at 900–1100 K for molybdenum, respectively, with the aid of a field emission and a transmission electron microscope. Oxidation of field emitters in the presence of a field resulted in the growth of crystallized oxides on the emitter surface, in different manners with tungsten and molybdenum. For tungsten, oxide crystals were constructed selectively on the {110} planes, whereas, for molybdenum, an oxide overlayer with a facetted structure was formed on the entire emitter surface.

The remolding proxess of <110>, <310>, <111> and <100> oriented tungsten single crystal tips have been observed using FEM and 3 MV electron microscope. In the pressure of 10^-10 Torr, the following planes {110}, {211} and {100} are enlarged at initial stage ofremolding process in all of the single crystal wires, then the ridges will form on the {111} and {310} planes and a spot is at last formed at and near the (310) plane. Field emission current also varied in three steps according to the migration process of which have activation energy of 1.9, 1.26 and 0.64 eV respectively. The activation energy for the blunting process of the spot thus formed is 2.3 eV, which means to be a defective crystal. By the formation of faceted surfaces upon heating the tip in the pressure of less than 10^-9 Torr and the change in relative surface energy of planes around the {100} planes due to chemisorption of gases, a spot is formed on the {100} planese. Observations of remold tip with 3 MV electron microscope reveal dislocation along the wife axis, the ridge direction, the slip plane and the direction normal to surface. A spot radius measured by FEM is larger than the value obtained by 3 MV electron microscope.

Si ribbon was made and used as a specimen tip in the field ion microscope. It had double best image voltages. This phenomenon was interpreted as being due to the band structure of Si. The position of image spots at the two voltages did not coincide, though the cause of this difference remains to be investigated. Some results for Si are compared with those for Ge.

Field ion microscopic (FIM) observation was carried out on tungsten tips, which had been cleaned by field evaporation and then corroded with dilute nitric acid. Very initial stages of corrosion were observed. When the tungsten tip was corroded with the acid immediately after being cleaned by field evaporation, its radius was much reduced. The reduction of tip radius occurred also in distilled water or basic solution. This sort of corrosion may be a transition reaction to ward the steady corrosion reaction. When the tip was corroded after exposure to air more than a day, it was scarecely corroded and only occasionally showed a pit hole. This difference could be ascribed to the difference of the exposing periods to air and therefore of the gas corroded layers.

Boron has been adsorbed onto molybdenum field emitter surfaces by the thermal decomposition of boron triiodide at corrected partial pressures of 1.4×10^-9 Torr to 4.5×10^-9 Torr. Based on field emission work function data and an analytical work function versus adsorbate-coverage relationship, a quantitative measure of the monolayer coverage of boron on molybdenum has been determined; this value is 6.3×10¹⁴ atoms/cm²/monolayer. A parameter, β, is defined which is a measure of the effective yield ofadsorbed boron on molybdenum from BI₃. β is formulated in terms of the free energy of dissociation of the BI₃ molecule on and the free energy of desorption of the BI₃ molecule from the molybdenum surface. These energies were determined from the data to be -34.4±2 kcal/mole and 4.5±4 kcal/mole, respectively.

The diffusion length of K, Rb and Cs on W-(100)-vicinals has been measured using a thermal ion-microscope. The distribution of the ion-current density, which desorbs from the surface, is registered in the vicinity of a vapor deposition limit. At low temperatures the investigation concerning the plateau of this distribution shows that there exist W-(100)-islets at the grain-boundaries. Their adsorption energy surmounts that of the (100)-areas by ΔQ∼0.4 eV. Although the islets contribute only about 10^-3to the total area, most of the current desorbs from them, provided that >50 µm. The analysis of the ion-current from the (100)-faces enables the determination of the diffusion length ○ of the atoms moving to the islets. Both quantities and ○ lead to the same diffusion coefficient D=7.9 cm²s^-1exp [±0.7-(0.86±0.08)eV/kT] corresponding to the ledges in the (100)-vicinals.

If a conical metal tip is heated in ultra-high-vacuum, the curvature radius at the apex increases continuously. A measurement of this phenomenon permits to determine the surface self-diffusion coefficient using Herring's equation and Nichols and Mullins numerical data. The increase of the curvature radius at the apex is determined in situ by measuring the field electron current as a function of voltage and the final radius and profile are determined by scanning electron microscopy. The logarithm of the surface-self-diffusion coefficient of tungsten without adsorbed layer versus the reciprocal temperatures, varies linearly in the experimental region (D=3.2.10^-8 cm²/sec at 2100 K and D=3.6.10^-5 cm²/sec at 2900 K). As activation energy a value of 3.08 eV/atom is found.

While O₂ adsorption on platinum leads, at 300 K, to three apparent binding states, as detected by thermal desorption and mass spectrometry, N₂O reaction fills only the most energetic one, β. Saturation coverages and variation of sticking coefficients with coverage have been measured (300 K and 500 K). Analysis of β state desorption at different heating rates (3.3 K sec^-1 and 370 K sec^-1) has been fully performed and show that the desorption rate is second order with coverage. Optimization of the results shows that only a variable desorption energy and a variable desorption frequency factor, both depending on coverage are in agreement with the experimental results. The chemical difference between the various binding states of oxygen is, however, not yet clear: isotopic mixing is almost identical in all states, during desorption, despite apparent first order desorption kinetics for the low energy states.

Oxidation of molybdenum (110) surface has been observed by Auger electron spectroscopy (AES) and LEED. The surface was oxidized at 500°C in 10^-4∼10^-6 Torr of oxygen. The ratios of oxygen to molybdenum Auger peak intensities increased markedly at exposure of about 10^-3 Torr sec of oxygen. The ratios remained constant for further exposure, probably because of complete oxidation within the escape depth of the oxygen Auger electrons. The surface and bulk plasma loss peaks shifted with exposure, reached to the value of oxide at exposure of about 10^-2 Torr sec, thereafter remained constant. The LEED pattern became diffuse with progress of oxidation and disappeared at about 10^-2 Torr sec. These results suggest the growth of amorphous oxide layer. The initial surface was proved to be very clean by AES and LEED.

A series of successive observations of structures, formed on a clean Fe (001) surface kept at various temperatures up to 300°C during continuous O₂ exposure, was made by LEED and AES. With the oxygen-coverage equivalent to about a monolayer, the p(1×1)–O was confirmed. After exposure of 20∼60 Langmuirs, a new LEED pattern appeared and others followed it consecutively with increasing exposure, all of which have yet been unreported and could now be correlated to the structures preceded Fe₃O₄. Spinel-typed layers of Fe₃O₄ (001) were formed on the bulk Fe crystal with a slight misfit at the later stages of oxidation.

Interactions of oxygen with iron, nickel and copper at room temperature have been investigated by means of an ultrahigh vacuum type X-ray photoelectron spectrometer designed by authors. Behaviors of peak heights and chemical shifts in O ls and Fe 2p spectra caused by oxygen chemi-sorption and oxidation are reported. Similar results on copper and nickel are also reported in this paper.

Oxide formed when (110), (001) and (111) surfaces of W are oxidised grows as whiskers which are single crystals with their growth axis in <001> type oxide directions and containing closely spaced faults on {310} type planes parallel to the axis.

A short circuit diffusion model due to a mosaic structure of the oxide has been proposed on the features of the oxidation rate anisotropy of single crystals of copper and nickel in the growth stage of continuous oxide films. In order to make clear the influence of the mosaic structure on the oxidation rate anisotropy, a correlation between the mosaic structure and the kinetic behavior was studied on the (001) face of copper by means of RHEED and microbalance techniques with UHV systems. The mosaic structure which depends on the epitaxial nucleation behavior in the early stage of oxidation was controlled by changing an off angle of a specimen surface from the (001) face of copper. The results show that the nucleation and growth behaviors of oxide have a straight effect on the mosaic structure and in turn the oxidation kinetics due to the short circuit diffusion.

Kinetic studies were made of the localized crystal growth and the oxygen chemisorption on the surface of iron and copper metal powders (aver. dia. 1.3–1.7 µm) by the use of electron microscope and micro-TG analyser. In common to both metals, similar type growths of cylindrical oxide whiskers, uniform in diameter and uniformly distributed on the surface of metal particles, were observed. Surface densities and average lengths of these whiskers formed at various temperatures 300–650°C remained fairly constant and equal to 10⁹ per cm² and about 0.45 µm, respectively. Only the diameters of cylindrical whiskers varied with temperatures, controlling the integral rates of localized crystal growths. Coincidently at about 500°C, for both metal powders, a maximal temperature for the localized crystal growth and a distinct change in the temperature dependence of oxygen chemisorption rates were observed. Correlation between these phenomena is discussed on mechanistic basis.

Infrared absorption spectra of adsorbed HF on the inner surface of KBr windows have been observed at room temperature. As the equilibrium gas pressure p over the adsorbed species is changed, different spectra are obtained for the following pressure ranges: (a) p\gtrsim40, (b) 15\lesssimp\lesssim30, (c) 3\lesssimp\lesssim11 and (d) p\lesssim2 Torr. Spectral changes between adjacent pressure ranges are sharp, and these changes occur almost reversibly with variation of the pressure. The four spectra are considered to correspond to the four phases of adsorption: (a) KF·nHF (n≧4), (b) KF·3HF, (c) KF·2HF and (d) KF·HF.

The adsorption, surface reaction and mutual replacement of CH₄, CO and NO on the films of titanium and tungsten evaporated under ultra-high vacuum condition have been studied with a combined Knudsen flow capillary method and mass spectrometer system. When CH₄-covered surfaces are exposed to gaseous CO or NO at room temperature and 77 K, the appearance of both CH₄ and complex produced by the reaction of CH₄ with CO or NO in the gas phase is noted. However, when the surface with preadsorbed CO or NO are exposed to gaseous CH₄, CH₄ does not remove CO, NO and the complex from the surfaces.

Surface layers of chemisorbed chlorine were obtained on the (111) face of a silver single crystal by decomposition of ethylene dichloride vapour. The structure of the surface was investigated by LEED and the concentration was estimated and controlled by Auger-electron spectroscopy. The formation of two surface structures, corresponding to different amounts of chemisorbed chlorine was observed. At lower coverages a (√3×√3)–R30° superstructure, which is stable up to high temperatures, was found. At higher surface concentrations, a (3×3) superstructure was formed, which was stable in ultra-high vacuum up to about 200°C. Models of the two structures are proposed in qualitative agreement with the experimental results.

An initial oxidation process on iron single crystals is studied by means of photoelectric work function measurements combined with Auger electron spectroscopy in the oxygen pressure of 10^-9 to 10^-5 Torr at room temperature. The crystal surfaces were initially observed the work function and Auger spectrum, and exposed to each oxygen atmosphere in step wise. The work function and Auger signals (Fe-46 eV, S-148 eV and O-508 eV) were observed in each step. Where the 46 eV-iron Auger peak showed a slow decrease and followed a rapid decrease in an intensity with increasing oxygen exposure, and finally the 46 eV peak split into two peaks (41 eV and 49 eV) as a result of formation of oxides. Each experimental run in the oxidation process showed different results for different initial condition of the sample surface. That is, chemically clean surface was oxidized in early stage of the oxidation. On the contrary, a cathode protection effect was observed on the surface which was increased in work function due to sulfur coverage.

An important contribution towards the well-known ionization of alkali atoms chemisorbed at low coverages on transition metal surfaces comes from the upward shift in the alkali valence level due to the image force. The Hartree-Fock approximation for this upward shift, in which the image of the adatom electron cloud is assumed static, may differ by up to a factor of two from that using the result that the self-image energy of an electron at a distance d from a metal surface is given by v=e²/4d. The resolution of this problem depends on the lifetime of an electron in the adatom orbital and the dielectric response of the metal surface. A simple model Hamiltonian is put forward which includes these effects, and which is solved in limiting cases. In particular it is found that the HF approximation should overestimate the upward shift in the case of a substrate such as tungsten. Possible implications for the screening of the Coulomb interaction between electrons in an adatom orbital are discussed.

The relation between certain parameters in the expression for catalytic reaction rates and the response or fluctuation characteristics of the substrate is discussed for the case of equilibrium reactions, as well as for the practically important case in which the equilibrium rate is not attained. A Kubo type formula for the equilibrium rate is established, and we indicate how this should be related to the usual transport theory of reactions.

A theoretical consideration of the surface deformations observed by FEM, FIM, etc. when electronegative gases such as carbon monoxide are chemisorbed on metals is presented in this paper. The existence conditions for localized states are exaimed by investigating a one-dimensional model which consists of semi-infinite linear chain and one adsorption centre, via MO-TBA method. The results qualitatively show that such chemisorption-induced surface deformations might occur under some conditions as the electronegativity of the adsorbate and the strength in chemisorption bond formed become large. The effects of surface deformation on the bond order which is associated with IR spectral band position are also considered.

The ordered structure of adatoms is discussed in the light of a theory which determines rigorously the ground state of the extended range lattice gas model. Several examples are discussed to show that an appearance or absence of an ordered structure at a particular density yields informations about the adatom interaction. The intermediate phase which may appear at finite temperatures is discussed briefly.

The increase of electrical resistance of Pt wires (dia., 9.8 µ; thermally treated by hydrogen and oxygen) was observed at 78 K and 90 K owing to adsorption of oxygen. Its relative increase (order of 10^-3) was analyzed by improving Fuchs' parameter p (the probability of specular reflection) in the Boltzmann-Fuchs theory under the assumptions: (1) geometrical surface roughness scatters the electrons and (2) the adsorbed oxygen atom additionally scatters the electrons. Effect (1) is described by a new Fuchs-form parameter, p(cos θ)=exp (-Ccos ²θ), (Ziman-Soffer theories), where θ is the incident angle and C (a function of the roughness) is experimentally determined as 3.0. (2) is evaluated in accordance with Greene-O'Donnell theories. The scattering cross section of adsorbed oxygen atoms was thus estimated to be 120 Ų for 0.03 coverage, which decreased with the increase of coverage.

The use of thin (\leqslant1 µ) nearly single-crystal semiconductor films as model adsorbents is discussed. In such cases, a quartz-crystal microbalance combined with field effect measurements would allow simultaneously recording of both the adsorbed quantity and the energy spectrum of surface states. The possibilities of a set up of this kind are analysed. It will be suitable for elucidation of problems related to electronic mechanisms in chemisorption: specification of the bonding between adsorbed particle and semiconductor surface (local chemisorption level, adsorption on surface defects, mere change in the parameters of nonadsorptive surface states); evaluation of the amount of chemically and physically adsorbed particles; effect of dispersity on the adsorption properties of semiconductors.

In order to understand the considerable irreversible ageing, annealing and adsorption effects on the various physical properties of amorphous (a–) Ge films, the effect of adsorption of N₂, O₂, Ar and air on the transport properties of obliquely deposited a–Ge films has been studied. The resistivity increases parabolically with time for about 30 minutes and thereafter increases slowly. Similarly, the resistivity rises rapidly with increasing pressure, upto ∼10^-2 Torr. The largest effect is observed with oxygen. The resistivity changes are drastically enhanced with increasing angle of oblique deposition. The temperature dependence of the resistivity is qualitatively similar for films with and without adsorption effects. This paper will discuss the various observations and their interpretation in terms of adsorption induced microstructural rearrangement of anisotropic films.

If vapour atoms have a finite adatom mobility on a substrate during the vacuum condensation process, the oblique incidence of vapours results in an anisotropic growth of vacuum evaporated films. The anisotropic structure enhances the surface area and adsorption effects and consequently changes in the electrical and optical properties of metal films occur. Dramatic and anomalous changes in the electrical conductivity, thermoelectric power, optical absorption and photovoltaic effect are exhibited by obliquely deposited amorphous semiconducting films. This paper will review our work on the electrical transport and optical properties of amorphous Ge films.

Comparative features of the chemisorption of reactive gases on the three principal low index clean single crystal faces of nickel and interpretations based on several typical atom-surface interaction models are presented and discussed. The conclusions based on measurements of microscopic parameters are reviewed in terms of data correlated from LEED patterns, computational analysis of LEED spectra, fine structure of Auger peak shapes and variations in work function changes as a function of coverage, exposure and heat treatment. The observed adsorption features are used to distinguish chemisorption phenomena in terms of differences in chemical binding, surface structure and charge transfer. Chemical overlayer systems involving well ordered single surface phases formed on unperturbed substrates most easily lend themselves to quantitative crystallographic analysis. Some comparisons of the results of this study to other experimental and theoretical studies on clean and chemisorbed nickel surfaces are considered.

Oxygen or cesium adsorption and coadsorption on a (100) tungsten crystal at room temperature have been studied by work-function measurements φ (Kelvin Method), Auger electron spectroscopy (AES) and LEED. Cesium adsorption results are in accordance with already known studies, except for new features in the Auger spectrum. For oxygen adsorption a close agreement is found with Madey φ-measurements, and a complex evolution in LEED patterns suggests the onset of oxidation. The increase in the dipole moment of cesium adsorbed on oxygen covered W(100) accounts for the lowering of the work function minimum and for the increase in the cesium monolayer density shown by AES, consistently with work-function values in qualitative agreement with Lang theoretical approach. However the cesium coverage at φ minimum seems not to be reduced by oxygen. A new peak in the Auger spectrum, associated with coexisting oxygen and cesium, suggests the existence of cesium oxide.

The angular distribution of Auger electrons from clean Fe(100) surface and from Fe(100) surfaces having various adsorbed structures has been investigated by means of a simple new method of Auger electron spectroscopy using a combined LEED and scanning photoelectron multiplier system. The experimental results indicate that the typical three types of angular distribution of Auger electrons are existed; 1) the lobular distribution due to the diffraction effect in the case of the 45 eV Fe Auger electron emission from well-ordered surfaces, 2) the cosine squared distribution in the case of the 598 eV Fe Auger electron emission, 3) the cosine distribution in the case of Auger electron emission from adatoms located at the outermost surface, for example, S Auger electrons from the Fe(100) c(2×2)–S and the Fe(100)–disordered–H₂S structures.

The interaction of sulfur with the (100) and (111) faces of platinum was studied in the temperature range from 25° to 1000°C. The electrochemical cell Pt/Ag₂S/AgJ/Ag served as a source of elemental sulfur (S₂). The Auger signal was calibrated by comparison to results obtained by a radiotracer method on polycrystalline foils. The LEED patterns obtained at different sulfur coverages and substrate temperatures could then be correlated to a number of adsorbate structures. These structures correspond in part to those obtained by others after exposing the surface to H₂S. On either surface two states of adsorption can be distinguished. They are characterized by a different degree of order and by different desorption temperatures.

We have used a focused electron beam to scan the surface of a substrate sample. The resulting electron induced desorbed ion is mass analyzed and the signal thus obtained is used to intensity modulate an output display. This technique enables us to obtain spatial distribution information of the adsorbed species. We have used this technique to monitor the spatial distribution of the F⁺ signal desorbed by electron interaction with a tungsten ribbon substrate. Results obtained indicate considerable surface migration of the fluorine or more probably the fluorine containing molecule. Rapid (flash) heating causes primarily thermal desorption while slow heating results in rapid surface migration to the sample support rods.

The dissociative chemisorption of iodine and bromine on tungsten single crystal surfaces has been studied by LEED. Iodine adsorption on W(110) produced long chain structures which could be divided into three families of phases (β₁, β₂ and β₃) distinguishable by the arrangement of atoms within the chain. Continuous changes in coverage occurred by sheets of these wellordered chain structures shearing relative to each other to produce packing faults at the resulting shear line. Although bromine on W(110) produced two comparable chain structures, shearing did not occur with changing coverage. Iodine on W(100) failed to produce either chain or shear structures. The lateral adatom interactions responsible for the stability of these structures is discussed.

Oxygen adsorption on tungsten has been studied at high temperature by Auger electron spectroscopy using a special object stage with small cylindrical object and electron bombardment heating. No temperature effects has been observed on the character of Auger electron spectra up to 3000 K. The study of the decarbonisation process of tungsten shows that small contamination of carbon on the surface influences strongly the equilibrium oxygen coverage. Adsorption isotherms of carbon free surfaces has been measured up to 10^-4 Torr in the temperature range from 1600 K to 2700 K on polycrystal faces. The measured isotherms can be interpreted assuming rather a second order desorption. A specially formed object has been used to investigate whether the electron bombardment heating is influencing the adsorption equilibrium but no influence on adsorption isotherms has been found.

Ionic and neutral evaporation from thick layers of KF on a Pt ribbon was studied as a function of the surface temperature (T) and the mean molecular layer thickness (θ). At a given T ranging from 300 to 500°C the neutral evaporation decreased monotonically with decreasing θ. The positive ion emission increased gradually with decreasing θ, and reached a maximum at θ≃1. The ionization efficiency increased from 10^-7 to 10^-2 according as θ decreased from 10² to 1. The desorption energies of the positive ion and the neutral particle at θ>1 were 2.7±0.3 and 2.4±0.2 eV, respectively. Negative ion emission was too weak to be detected at T<500°C. Thermochemical consideration of these results suggests that Saha-Langmuir's equation is not applicable to the above thick layer system.

The change of work function on adsorption of ethylene, acetylene and benzene on Ni (111) single crystal surface was followed by the retarding potential method. The work function changes were studied as a function of increasing exposure. On clean surface we obtained at 3×10^-9 Torr: -0.3 eV for ethylene, -0.6 eV for acetylene and -1.1 eV for benzene adsorbed. The same results are observed for a pressure of 10^-8 Torr for ethylene and benzene. For acetylene the change of work function attains to -1 eV suggesting the polymerization of the hydrocarbon. A part of benzene is reversibly adsorbed. After pumping off this fraction we observed a change of φ=-0.8 eV due to the adsorption of irreversibly adsorbed benzene. These changes were correlated with LEED patterns and thermodesorption data.

A new model for predicting work functions of binary compounds is proposed which is based on the idea of gas adsorption on metal surfaces. The work function of a compound AB is determined by the higher work function element, B, which is perturbed by "adsorption" of lower work function element, A, as φ_AB=φ_B-0.88r_iA(0.74 φ_B-0.44 φ_A-0.64), eV where r_iA is the ionic radius of the element A in Angstrom. The equation is tested for the work functions of rare earth borides, sulphides, and carbides. Agreement is found to be good. The model is also compared with our previously developed "perturbation" model.

The adsorption of CO on tungsten is discussed in the light of thermal and electron impact desorption, as well as photoemission and field emission spectroscopy. The evidence for the existence of several low and intermediate temperature binding states is discussed, and the nature of the states interpreted in terms of undissociated CO. The high temperature binding modes are also discussed and it is suggested that the present evidence does not permit a distinction between dissociated and undissociated adsorption although both C and O seem to be in contact with W atoms.

Ultraviolet photoelectron spectra (hv=40.8 eV) were recorded from clean and CO covered (110) and polycrystalline Pd surfaces, respectively. Adsorption of carbon monoxide leads to the appearance of two maxima at 7.9 and 10.8 eV below the Fermi level E_F which are associated with chemisorption levels derived from the 5σ- and 1π-orbitals of free CO. These states are broadened to 1.5 and 1 eV, respectively, and the vibrational structure present in the UPS data from gaseous CO may no longer be resolved. An observed decrease of the emitted intensity just below E_F is interpreted as being caused by an additional level in this energy range which is due to back-donating of metallic d-electrons into the empty 2π^*-orbital of CO as predicted theoretically.

The adsorption of CO on Ru(001) has been studied using a combination of techniques: LEED/Auger, Kelvin probe contact potential changes, and flash desorption mass spectrometry. Adsorption of CO is reversible at temperatures and pressures as high as 700 K and 10^-4 Torr, respectively. Two binding states of CO are identified, and isosteric heats determined both from work function changes and from LEED intensity measurements agree well with flash desorption energies. A disordering of the CO layer due to repulsive interactions between neighbors at high coverages is postulated. Electron beam-induced LEED pattern changes are characterized and found to have a high cross section (∼7×10^-17 cm²) at ∼110 eV.

The characterization of the W(100) surface adsorbing nitric oxide at room temperature has been conducted with the conventional LEED and AES technique. Nitric oxide on the W(100) has shown ordered (2×2) and (4×1) patterns at the exposure of ≃0.8 L and ≃1.6 L respectively without any heat treatment. The sequential heating of the fully NO-covered surface has revealed (2×2), (4×1) and (2×1) structures. In the AES, as the exposure was increased the nitrogen and oxygen peaks did not grow up at the same rate, the latter kept nearly constant over the exposure of about 3 L. The surface potential of nitric oxide on W(100) was -1.52±0.03 V by the retarding field method.

Infrared spectra of CO adsorbed on SiO₂ supported Pd–Ag and Ni–Cu alloys were measured in order to estimate the relative importance of the geometric and the electronic effect of alloy catalysts in adsorbing CO. The IR bands due to the adsorption of CO on Pd or Ni atoms were observed in two frequency regions; above and below 2000 cm^-1. The intensity of the bands below 2000 cm^-1 decreased strongly upon increasing the content of inactive Ag or Cu in these alloys,while the band positions remained unchanged. From these results it is concluded that the assignment of CO bands by Eischens, i.e. the linear and bridge bonding complexes of CO on transition metals, is correct and that the change in CO adsorption due to alloying is predominantly caused by the geometry of the adsorbing surface, while the electronic factor is less important for these alloys. The latter conclusion is confirmed by results of CO adsorption on Pd hydride and by the constancy of the free energy of activation for CO desorption for alloys of different composition.

ESD and TD techniques were used to study the adsorption and thermal decomposition of ammonia on a polycrystalline molybdenum sample. The TD results suggest a two stage mechanism for the thermal decomposition of ammonia into hydrogen and nitrogen.

Backscattered electron energy distribution have been measured using a high resolution cylindrical electron spectrometer. Characteristic energy losses were observed for a clean Mo(100) surface, and for the same surface covered by oxygen. The surface plasmon loss peak was identified by its position and intensity dependence on oxygen adsorption, and using optical data. Interband transitions were also observed, which may be classified in three categories: a) Bulk transitions of Mo in agreement with optical data, b) Transitions very sensitive to the surface state, they disappeared with adsorption of small quantities of hydrogen and oxygen. Their energy and sensitivity suggest their identification as surface state transitions, c) New transitions due to the adsorbed molecules.

A pure SnO₂ sample outgassed in vacuum gives an ESR signal (g=1.896). The signal is attributed to the donor electrons. Adsorption of oxygen results in the formation of three paramagnetic centers at room temperature and of two centers at high temperatures. These centers are characterized by different desorption energies. Hydrogen adsorption results in an increase in the g=1.896 signal intensity at room temperature or above, but no change occurs at 77 K. When a small amount of NO₂ adsorbs on SnO₂, the NO₂ molecule dissociates into O₂ and NO. A large amount of NO₂ gives rise to a rather complicated spectrum with hfs from nuclear spins of Nand Sn.

The adsorption state of hydrogen on platinum was studied by measurement of NMR line width and the Knight shift of proton. The line width was about 8 G at 4.2 K and decreased from 2.5 G at 150 K to 0.9 G at 290 K gradually. This decrease might be attributed to surface migration of loosely bonded hydrogen adatoms, whose activation energy was found to be 0.5 kcal/mol. At temperature above 290 K the line width decreased rapidly and the apparent activation energy in this temperature region was 6.8 kcal/mol. The Knight shift was found to be negative as ΔH/H=-0.004% and was independent of temperature and magnetic field strength within the experimental accuracy. The observed resonance line was resolved into two components. It was concluded that it correlates with different states of adsorbed proton.

H₂ adsorption studies have been carried out on a series of supported Ni–Cu alloys ranging from pure Ni to 60 Cu at.%. Reduction and homogeneity of the metallic phase are controlled by Curie point and spontaneous magnetization determinations. When H₂ is admitted, the 4 K saturation magnetization decreases linearly as the quantity of adsorbed gas increases. Desorption and adsorption curves suggest only one state of chemisorption. The decrease of magnetization per adsorbed molecule (α) is twice as large as the magnetic moment of Ni in the alloys. This result rules out the collective model to describe H₂ chemisorption on Ni, and suggests that Ni sites are magnetically decoupled from subjacent moments upon chemisorption. A complete demetallization of Ni atoms would lead to another expression for α. Thus, it may be concluded that Ni atoms remain in their metallic state. This result is compared to magnetic effect observed on other ferromagnetic catalysts (Fe, Co): in all cases, H₂ chemisorption results in a change of the superficial magnetic structure.

Reflection-absorption spectroscopy has been shown to yield, in some cases, the infrared spectrum of as little as a few percent of a monomolecular layer of adsorbate on an extended metal surface. In order to maximize the sensitivity of the technique, there is an optimum number of reflections which depends on the angle of incidence and the optical constants of the metal. In a number of cases, using only one reflection, we can obtain a sensitivity comparable with that produced by using the optimum number. With only one reflection the experimental procedure is simplified and the reflection-absorption experiment can be applied to a sample studied simultaneously by other surface probes.

The quasiequilibrium treatment of heterogeneous reactions suggested by Batty and Stickney and first order desorption kinetics are combined to formulate a model of the equilibrium coverage of oxygen on tungsten at high temperatures (1200<T<2500K) and low oxygen partial pressures (10^-9\leqslantP_o2\leqslant10^-5 Torr). The results are compared with existing data and the agreement is fairly good in view of the extreme simplicity of the theoretical model.

The physisorption of light hydrocarbons on the basal plane of graphite has been theoretically studied; potential energies are calculated in the pairwise approximation, using Lennard-Jones potentials, by summing up all interactions between the carbon atoms in the graphite crystal and the hydrogen and carbon atoms in the adsorbed molecule. Potential minima are determined by a process of simultaneous minimization in the six parameters which define the configuration of the rigid molecule with respect to the graphite crystal. A normal coordinate analysis is then carried out in the harmonic approximation, which proves satisfactory for fundamental and singly excited states. In the cases here considered, the residual energy represents an appreciable fraction of the adsorption energy, a large part being due to hindered rotations at the adsorbing surface.

The Auger electron spectrum from xenon adsorbed at 77–85 K on evaporated nickel films has been investigated by means of a retarding grid analyser with post-monochromator. The surface concentration of xenon is quantitatively monitored by the 37 eV xenon peak and by the attenuation of the 60 eV nickel peak. The effect of the incident electron beam on the adsorbate has been measured by quantitative comparison with xenon adsorption isotherms determined volumetrically on the same type of surface. The xenon coverage is found to depend on the current density and electron energy of the incident beam. For the conditions typically used for Auger analysis it is shown that at high equilibrium xenon pressures (>10^-8 Torr) the temperature rise due to the power input to the surface is the important effect. At lower pressures electron beam desorption is found to be significant. These effects are discussed in connection with the quantitative analysis of physisorbed layers.

Motional states of argon, oxygen, nitrogen and carbon dioxide sorbed in mordenite, were studied from statistical mechanical points of view and by EPR techniques. Above the dry ice temperature, these molecules were freely translating, as a one-dimensional gas, along the pore channel in the mordenite. The rotational freedom was seriously affected by the magnitude of the electric quadrupole moment of the molecule, which interacts with the crystal field in the pore. Oxygen molecule, having a small electric quadrupole moment, was freely rotating in the pore. Nitrogen molecule has a considerably large quadrupole moment, and rotated freely only above 18°C. In the case of carbon dioxide, the rotational hinderance was more serious, on account of its much larger quadrupole moment. Below 50°C, its molecular axis was orientated parallel to the pore axis. Above 160°C, it behaved as a plane rotator in the cross-sectional plane of the pore.

Experimental adsorption isobars of xenon on nickel (110) in thermodynamical equilibrium conditions are obtained between 110 and 150 K. The behaviour of the adsorption heat as a function of coverage (6000 cal/mole for θ=0, plateau of 5000 cal/mole for θ>10^-2) and the applicability of the Dubinin-Radushkevitch isotherm to only the lower part of the coverage (θ<10^-2) can be explained by the existence of a nonuniform part of the surface of about 1%. The calculated virial coefficients apply essentially to this region. The magnitude of the isosteric heat on the plateau and its decrease at higher coverages are explained in terms of an intrinsic surface electric field of the order of 5×10⁹ V/m, inducing electrostatic attraction and mutual adatom dipole repulsion.

In the study of the process of the deposition of potassium atoms on a mica surface by means of molecular beams, the phenomenon that the sticking coefficient changes with the exposure time and has a minimum at room temperature was observed. To interpret this phenomenon, simultaneous rate equations of adsorption were proposed and it was found that the sticking coefficient is expressed by S=1-\frac{N}{I}∑_i=1^∞\frac{\theta_{i}}{\tau_{i}}. The sticking coefficient curve having a minimum as observed in the experiment can be produced from the solutions of the equations related to the BET model. If the surface migration of adatoms is taken into account, the minimum of the theoretical sticking coefficient curves becomes less deep. If b≡N/Iτ₂>1 the sticking coefficient curve monotonically tends to zero as observed in the experiment at higher temperature.

Recent calorimetric studies of helium monolayers adsorbed on basal plane graphite substrates have disclosed two-dimensional gas, liquid and solid phases as well as an ordered epitaxial array believed to be √3 R 30°. Previous failures to observe these phases are ascribed to the heterogeneity of typical adsorbents. The new results suggest several ways in which He films can be used as very sensitive probes of certain properties of solid surfaces.

An Einstein model is used to calculate the vibrational entropy of adsorption of a krypton or xenon monolayer adsorbed on the (0001) graphite surface. The force constants occuring in the theory are deduced from Lennard Jones interaction laws. The computed value for xenon ΔS≃-1.2 cal mole^-1 K^-1 is compared to the one evaluated measured by Auger electron spectroscopy i.e. ΔS=-2±1 cal mole^-1 K^-1. We conclude that the vibrations in the adsorbed phase are partly hindered in comparison to the vibrations of atoms of a bulk xenon crystal. For krypton, no experimental value of ΔS is available, but the computation shows that the vibrations are hindered or amplified according to the direction.

X-ray photoelectron spectroscopy (ESCA) is a powerful technique for the study of the bonding of adsorbed molecules to surfaces. We have employed an ultrahigh vacuum ESCA spectrometer to study the adsorption of Xe on a W(111) single crystal disk at ∼100 K. The photoelectron spectrum of Xe is compared with flash desorption measurements of Xe surface coverage and desorption energy. The ESCA measurement has a limiting sensitivity of ∼0.05 monolayer of Xe, and involves digital summation of multiple scans. As Xe coverage increases in the range 0.05\leqslantθ_Xe\leqslant1.0 the Xe (3d_5/2) level is shifted monotonically upward in binding energy by 0.5 ± 0.1 eV, relative to the Fermi level of W(111). Part of this shift may be due to preferential Xe adsorption on extraneous sites other than W(111), the remainder being related to weak interactional effects between Xe molecules on W(111). Above θ_Xe\cong0.2 the desorption energy of Xe is constant on this W(111) crystal, in agreement with previous work where a constant desorption energy of 9.3 kcal.mole^-1 was measured. The work function decreases by 1.0 ± 0.1 eV upon adsorption of Xe to monolayer coverage at ∼100 K.

A thermodynamic introduction is presented to describe observations of a large surface effect produced by small amounts of an adsorbate. The essential point of the paper is the use of a very large (∼10^-11 cm²) second virial coefficient, A model, using surface atomic vibrations is discussed in connection herewith.

The interactions at room temperature of O₂, H₂O, NO and NH₃ with thermally cleaned Si(111) surfaces have been investigated using the combined Auger electron spectroscopy and electron impact desorption techniques. The Auger yields and the electron impact desorption are related to the structural models of the adsorption complexes. The ion species observed by electron bombardment are O⁺ ions from Si(111)–O₂, H⁺ and O⁺ ions from Si(111)–H₂O, and O⁺ ions from Si(111)–NO systems. The energy distributions of these ions, and the adsorption kinetics measurements monitored by the ion currents are presented and discussed.

Adsorption of NO on Germanium films was studied by means of infrared spectroscopy. The spectra were recorded at 77 K. Absorption bands were obtained at 2214 cm^-1, 1270 cm^-1, 580 cm^-1 and a band at 760 cm^-1 which shifted to 780 cm^-1 -840 cm^-1 on adsorption of additional amounts of NO. The first three absorptions are assigned to N₂O molecule formed on the surface of the adsorbent as a result of the reaction: 2NO→N₂O+O, whereas the last band is due to GeO. On heating the system to 200 K desorption of N₂O occurred and no change in the 780–840 cm^-1 band was noticed. On adsorption of additional doses of NO, after the whole surface became covered with GeO, two additional bands at 1853 cm^-1 and 1750 cm^-1 appeared; these are ascribed to NO dimers formed on the surface.

We have carried out ultra-high vacuum studies of the effects of oxygen on transport properties of epitaxial films of PbS which show that the differential Hall mobility remains precisely constant during oxidation even though large changes occur in the number of carriers in the film. However, the kinetics of the electrical changes obey an Elovich rate law (log t) and the pressure dependence follows a Temkin isotherm (log P), suggesting that the oxygen remains on the surface. These apparently contradictory observations can be explained in terms of an entirely new type of surface mechanism in which there are no localized charged states and the electrical effects are due to changes in the surface stoichiometry. This surface stoichiometry effect (SSE) is closely related to a bulk mechanism which was recently proposed by Parada and Pratt to explain the lack of carrier freeze-out in the lead chalcogenides.

We have measured the photoeffect on the oxygen adsorbed on SnO₂ and TiO₂ under different surface treatments and pressures in the range 10^-4 to 10^-1 Torr, with particular attention to the photoeffect inversion. In TiO₂ at room temperature the photoeffect inversion pressure is found to be about 10^-3 Torr, the photoadsorption and photodesorption occuring above and below this pressure, respectively. This inversion is interpreted as to be related to the transition from a gas-adsorbate equilibrium situation to a non-equilibrium situation, due to the pinch-off mechanism. In agreement with our interpretation, we found an hysteresis effect in the photoeffect vs pressure relation.

It is demonstrated that medium resolution NMR can be used, as in the homogeneous phase, for the study of the heterogeneous chemical equilibrium of NH₃ adsorbed on silica surfaces and for the determination of the exchange rate of protons.

The EPR Technique has been used to investigate the isotopic exchange between oxygen from the adsorbed species and the oxygen from either the lattice or the gas phase. The reactions are as follows: (¹⁶O₂^-)_s+¹⁷O²\rightleftarrows(¹⁷O₂^-)_s+¹⁶O_2,g (¹⁷O₂^-)_s+¹⁶O_l^2-→(¹⁶O¹⁷O^-)_s+¹⁷O_l^2- where the subscripts s, g and l refer to adsorbed, gaseous or lattice oxygen respectively. The hyperfine lines due to the nuclear spin (I=5/2) of ¹⁷O (in oxygen enriched to 58 atom% of ¹⁷O) have been used to monitor the isotopic exchange. The predicted behaviour of the different possible mechanisms is compared to the experimental results which show, on MgO, that reaction (1) is reversible and occurs via an (O₄^-)_s intermediate. The possible geometries of such a species are compared and discussed. On the other hand, reaction (2) does not appear to be reversible and is favoured by UV or gamma irradiation.

The treatment of exchange energy in inversion layers is extended to take the layer thickness into account when only the lowest subband is occupied by electrons. The exchange energy of electrons in Si is close to the zero-thickness limit obtained previously. Wave functions are only slightly contracted because of exchange. The exchange energy in excited subbands is smaller in magnitude, and has a smaller variation with wavevector, than in the lowest subband. Estimates based on a three-dimensional local density approximation give the approximate magnitude of the exchange energy but do not give its dependence on the spatial extent of the wave function correctly and cannot account for image effects.

The oscillatory enhancement of the g factor due to the exchange interaction among electrons in the inversion layer made on Si (100) surface is calculated under strong magnetic fields. The degree of the splitting of each Landau level observed in Schubnikov-de Haas measurements is shown to be explained by the theory of the level broadening of Landau levels if such enhancement is taken into account. The g factor is also calculated under tilted magnetic fields, and the theoretical one is in excellent agreement with the experiment by Fang and Stiles, and with the similar one performed recently by Kobayashi and Komatsubara. It is proposed that the valley splitting is enhanced due to just the same mechanism as that of the case of the spin splitting and that due to such enhancement the valley splitting is observed experimentally in low-lying Landau levels.

A description of experimental results is given of the electronic properties of carriers at the interface in a silicon-silicon dioxide field effect transistor as it exists at low temperatures. The case of a p-type inversion layer and the case for an n-type inversion type layer are both discussed. Properties in the low magnetic field range where the cyclotron frequency times τ, i.e., the ω_cτ is much less than 1 are described as well as the intermediate field case (ω_cτ∼1) and the high field case (ω_cτ≫1) for the n-type inversion layer are discussed. Single particle properties, as well as many-body effects are included to illustrate the unique potential that such studies have in providing information, as one can vary the contrast between the single particle and many-body effects in the single system by varying the density.

The transverse magnetoconductance of a n-type inverted (100) silicon surface was measured at 4.2 K and up to 7.5 Tesla in order to examine the dependence of the amplitude of its oscillatory part on the surface electron density and on the magnetic field. A phenomenological description of the results is given. At electron densities sufficiently small to satisfy the electric quantum limit condition, the observed width of the Landau levels is in agreement with the momentum relaxation time of the electrons in the inversion layer. At higher surface electron densities, however, an additional level broadening is found which increases with the number of electrons in the first excited electric subband. The dependence of the amplitude of the oscillatory magnetoconductivity on the Fermi energy has been used to determine the Landé g factor.

The detailed measurements of transverse conductivity on silicon (100) surface in n-inversion layer were done under the titled magnetic field. The g value of electrons was obtained as a function of carrier concentrations. Two kinds of g value were obtained; one is the case where the level separation between the adjacent levels are equal and the other is the one where the two levels (N, ↑) and (N+1, ↓) coincide with each other. The g value for the former is much larger than two and has a strong dependence on carrier concentrations, the larger its value at lower carrier concentrations. This increase of the g value is explained by the theory of Ando and Uemura and the good agreement with this theory was obtained. Whereas the g value for the latter has a weak dependence on carrier concentrations and further analysis for this reason is necessary.

Self-consistent calculations for energy levels are performed for n-type inversion layers of silicon with magnetic field perpendicular to the (100) surface. Physical quantities are shown to have an oscillation as the magnetic field is varied. An "apparent" g factor, g^*, obtained from a period of the oscillation of states density at the Fermi level is plotted as a function of the Γ/2βH, where Γ is the width of Landau levels. The results show that g^*∼5 for Γ <2βH and g^*∼g for Γ≫2/βH.

Surface Shubnikov-de Haas (SDH) oscillations have been measured for the first time in p-type channels of(1 10) silicon field effect transistors. Quantum oscillations were recorded as a function of magnetic fields up to 14.6 Tesla (146 kG) at different constant gate voltages or at fixed magnetic fields as a function of surface electric fields up to 1.2×10⁶ V/cm. The analysis of the data revealed two electric subbands. The effective masses m of the holes in both bands were determined, in the first subband m_c increased with increasing gate voltage. In addition, the angular dependence of the SDH-effect was studied. It was found that the spin splitting of the Landau-levels was dependent on the angle between surface and magnetic field.

Quantized surface states in an inversion layer of a narrow-gap semiconductor (NGS) with the zinc-blend (ZB) structure are investigated theoretically by the effective mass approximation (EMA) for nearly degenerate bands. When the band gap is small, and the spin-orbit interaction is large, the surface potential causes large k-linear term in the two dimensional dispersion relation, which removes the spin degeneracy. And the g-factor becomes so large that the inversion of the order of Landau levels is expected. The theory is applied to the analysis of the Schubnikov-de Haas (SdH) measurements in an inversion layer of Hg_0.79 Cd_0.21 Te (ε_G=68 meV, m₀^*=0.006 m). According to this analysis, the splitting of the spin degeneracy is as much as several meV, and the inversion of the order of Landau levels seems to be realized. The surface potential of the sample in this experiment is described in the parametrized form, and values of the parameters are estimated by using the experimental data.

Surface Shubnikov-de Haas oscillations were studied in (0001) and (100) accumulation layers of pure tellurium single crystals in external electric fields up to 10⁶ V/cm. Whereas in accumulation layers of purely chemical origin at most two electric subbands could be identified, up to four subbands have been detected with the application of additional electric surface fields. The energies of the subbands were derived from the data and compared with a theoretical model assuming an exponential surface potential. The agreement is reasonably good, having in mind the complicated valence band structure of tellurium. The effective mass m_c of the surface holes was determined from the temperature dependence of the amplitude of the quantum oscillations. It was found that m_c had increased with respect to the bulk values. The Dingle temperature varied between 3 and 6 K for different samples.

Low temperature mobilities (µ) of holes in the Si surface layer of an MOS structure are presented to illustrate the important influence of the SiO₂–Si interface. Results at 4.2 K show that µ, measured at identical inversion layer carrier concentration (N_inv), is higher for samples with a thicker oxide layer. This apparently contradicts the theory that the average distance (Z_ave) of carriers from the surface is a function of N_inv only, and independent of the oxide thickness; a number of possibilities to explain the discrepancy is given, but is not resolved. Specially processed samples with different degrees of surface graininess give indirect evidence for scattering by surface roughness. Results at very weak inversion show that the mobility increase is steeper for samples with large surface charge densities, supporting the theory of scattering by charge inhomogeneities.

For the interpretation of observed low temperature mobility in n-channel inversion layers of Si MOS, two scattering mechanisms are discussed theoretically. The one is the scattering due to interface roughness and a simplified model is investigated. The other is the scattering by the charged centers. The former can explain the carrier concentration dependence of the mobility in high carrier concentration region. The latter is introduced for the interpretation of the mobility at low carrier concentration with some considerations of the charge distribution.

Cyclotron resonance of electrons bound in quantized states in space charge layers on a (100) surface of Si has been observed. We have seen the electron resonance in both inversion and accumulation layers. For the case of inversion the cyclotron mass m_c^* is found to differ substantially from the bulk value 0.1905 m₀ and to vary with the surface density of electrons (Q_s). We find m_c^*=0.230 m₀ for Q_s≈0.2 ×10¹² electrons/cm², which decreases to m_c^*=0.230 m₀ for Q_s ≈2.0×10¹² electrons/cm². The relaxation time from cyclotron resonance exhibits a maximum near Q_s≈0.8×10¹² electrons/cm² for the inversion layer. In recent work^* we have discovered in addition to the fundamental resonance an approximately subharmonic sequence of peaks, which we attribute to cyclotron resonance perturbed by charged trap states at the Si–SiO₂ interface. We also find cyclotron resonance of electrons in an accumulation layer on n-type (100) Si with mass m_c^* close to the bulk value.

Cyclotron resonances in n-type silicon inversion layers has been observed at 4.2 K at 45.407 and 32.12 cm^-1. Both real and imaginary parts of the channel conductance are measured with a homodyne detector as the magnetic field is swept from 0 to 100 kG. In this way the resonant electron density, mass and scattering rate have been measured for electron densities between 0.38×10¹¹ and 9.0×10¹²/cm². The mass is essentially constant, (0.20±01) m₀, and does not increase at low electron densities. The far-infrared mobility drops monotonically with increasing density in marked contrast to the dc mobility which exhibits a strong maximum in the range 1.5×10¹²/cm². The resonant electron density is proportional to gate voltage over the entire range of concentrations. The results indicate that the decrease in dc mobility at low densities is not due to carrier freeze out or increase in the microscopic scattering rate but may be related to a break up of the channel.

Electron paramagnetic resonance detects primarily the presence of unpaired electrons, and provides information, sometimes in great detail, about their environment. In the case of semiconductor surfaces prepared by crushing in ultra high vacuum, unpaired electrons have been detected on Si, Ge, Ge–Si alloys and on graphite. All other high vacuum crushed semiconductors studied, show negligible surface resonance. The most extensive work has been done on Si, including cleaved surfaces. A detailed model has been proposed but some features of the behaviour of the resonance upon oxygen adsorption remain puzzling. In many cases adsorbed paramagnetic species display resonances that reveal properties of the substrate. In particular, adsorbed O₂^- ions have revealed properties about various surfaces of catalytic interest and have allowed the wave functions of(110) surface electrons on AlSb and GaAs to be deduced.

Unreconstructed intrinsic semiconductor (111) surfaces are, to a good approximation, two-dimensional metals. The Fermi surface has flat portions, leading to a very large density of states at the Fermi energy, and to a quasi-1 dimensional electron dynamics. The condition for the stability of the surface metallic state is examined and shown to be violated; a surface phase transition to an insulating state follows. The insulating state can be either a magnetically ordered state (spin density wave) or a Peierls superlattice (charge density wave), depending on the relative strength of the surface electron-hole and electron-phonon interaction. The periodicities of many of the higher n×n reconstructions observed on Si and Ge (111) agree with those expected for the charge density wave case. It is pointed out that SDW magnetic ordering on surfaces might be detectable by polarized LEED, or otherwise.

Clean surfaces produced by cleavage or ion bombardment and annealing mostly show atomic steps. With a new derivation of the LEED pattern of a stepped surface it is now possible to predict and evaluate also the pattern of a surface on a non-primitive lattice. With the use of high precision LEED data depression of the edge atoms of about 0.25 Å is supported by several independent observations. On cleaved silicon and germanium surfaces the step density is determined for each spot of the surface. At clean cleaved silicon surfaces the photo surface voltage has been measured by scanning the surfaces with an electron beam. A correlation between the photo surface voltage and the angle of inclination of the respective spot of the surface towards the (111) face reveals the existence of one-dimensional surface states ("edge states"). It is therefore important to know the existence and properties of atomic steps on clean surfaces in both structural and electronic respects.

Clean, cleaved Si (111) surfaces exhibit a 2×1 superstructure, which converts to a 7×7 structure upon annealing. We have simultaneously studied the changes of LEED pattern, of surface conductivity, and of work function at cleaved Si surfaces after annealing at temperatures as high as 500°C. The results may be summarized as follows: 1. Correlated with the density of cleavage steps the conversion temperature T_c, for the above irreversible transition of the cleaved silicon surfaces varied between 330 and 390°C. The experimental points are fitted by T_c^*-T_c∼w, where T_c^*=425°C is the limiting conversion temperature for small terrace width w of the steps. 2. The disappearance of the Si (111)–2×1 structure is correlated with a decrease of the surface conductivity as well as of the work function. With developing 7×7 structure both quantities increased again. The results show that due to the structural conversion not only do band bending and surface state distribution change but also the electron affinity is lowered by 0.25 eV compared to the freshly cleaved surface.

The edge component of dislocations in semiconductors of the diamond structure corresponds to a one dimensional surface. We have measured spectral conductivity and band bending at dislocations in <123> deformed Si single crystals with a dislocation density of 5×10⁷ cm^-2. The result is, compared with those for clean Si surfaces. There is in part a striking quantitative agreement between the two systems, demonstrating the same dominating effect of dangling bonds in both systems.

The total current of photoemitted electrons has been accurately measured as a function of photon energy in the threshod range on a set of ultrahigh vacuum cleaved silicon samples of various doping levels. Highly doped p-type samples allow the determination of the work function. Samples at other doping levels show a structure in the lower energy part of the threshold which is due to surface states and gives their energy distribution. The effect of 10^-2 Torr. s O₂ is briefly discussed.

Electronic surface states of Si, Ge, W and Mo have been investigated by the measurements of ultraviolet photoemission spectra. Emission from the surface states is observed from the cleaved Si (111) and Ge (111) surfaces. Changes are observed of the spectra by annealing these surfaces, and the results are presented in connection with the surface reconstruction. Surface sensitive peaks have been observed from W and Mo, whose peak energies are in agreement with the band calculation.

LEED, AES, photoemission, and surface photovoltage studies have been carried out an a range of materials which crystallise with layered structures. In all cases cleaved basal faces are extremely inert to common contaminants. Very slight bombardment with low energy argon ions leads to a dramatic increase in the reactivity of these faces suggesting anisotropy ratios, in some cases, of around 10¹², for sticking coefficients on non-basal to basal faces. For clean surfaces no photovoltage or bending of the energy bands is generally observed but for p-type crystals slight contamination by a fraction of a monolayer of carbon leads to band bending. These results are discussed with respect to surface states in these materials.

Two photoinduced ESR absorption lines are observed at high resistivity silicon surfaces which are chemically etched and followed to rinse with water. It is found that these centers are associated with dissociative water adsorption on surface structural hydroxyl groups and become parama-gnetic by trapping photoexcited excess free carriers.

The consequences of tunneling from surface states on the total energy distribution of field emitted electrons from a clean germanium surface have been investigated in some detail using a model for the surface states originally suggested by Handler. At low emitted current densities (<10² A·cm^-2) the emitted electrons come mostly from the surface states and from the valence band. The shape of the calculated distribution is in reasonable agreement with available experimental data. At current densities of the order of 10³ A·cm^-2 emission from the conduction band becomes appreciable but it is still a small proportion of the emitted current. We expect that at higher current densities (<10⁴ A·cm^-2) corresponding to applied fields \gtrsim0.5 Volts/Å the conduction band will eventually become degenerate at the surface. For these current densities however the transport of electrons and holes in the interior of the semiconductor play an important role and the theory in its present form becomes inapplicable.

Radiation associated with surface centers has been observed in GaAs single crystals. The characteristics of the surface luminescence in relation to surface treatment have been investigated. Surface luminescence peaks have been observed in cleaved surfaces, in Au-doped surfaces, in Si₃N₄-covered surfaces and in γ-irradiated surfaces. The possibility of the investigation of surface phonons and excitons are demonstrated.

Germanium surfaces contacted by an electrolyte are covered by OH-groups after anodic, with H-particles after cathodic polarization and can be covered even with a mixed layer of different compositons. With alkaline electrolytes (above pH 11) a surface state can be produced, the energy level of which shifts in dependence of the mole fraction of the OH-groups within the mixed layer. This has to be concluded from measurements of the differential interface capacity.

The conductivity and the electron mobility in inversion layers on etched Si surfaces covered with Cs have been measured at temperatures between 80 and 400 K. The temperature dependence of the electron concentration can be explained by following energy level scheme at the surface. (1) Acceptor-like trap levels of 1.9×10¹² cm^-2 exist at 0.3 eV below the conduction band. (2) Cs-associated donor levels whose concentration decreases from 1.9×10¹² cm^-2 to zero with the increase of the adsorbed Cs lies at 0.098∼0.070 eV. (3) Cs = associated donor levels whose concentration increases from zero to 2×10¹³ cm^-2 which is almost equal to the maximum number of adsorbed Cs lies at 0.01∼0.005 eV.

The electron tunneling effects of the single crystal of HgTe, which is a typical zero-gap semiconductor, were investigated at the liquid helium temperatures. The junctions have Pb-insulator-HgTe structures. Phonon-assisted tunneling effects were observed in the d²J/dV²-V characteristics. The optical phonon energies of HgTe are estimated to be 14 meV (transverse) and 17 meV (longitudinal). The dJ/dV-V characteristics showed asymmetric behaviors with a minimum in the negative bias range. The bias position of the minimum is about -50 mV in most samples. These asymmetric behaviors are explained by taking into account the upward band bending and the peculiar band structure of HgTe.

The correlations between the crystallographic polarity of II–VI and III–V compounds and the barrier height of Schottky diodes are investigated. The diodes are fabricated by evaporating various metals (Au, Ag, Pt, Pd) in ultra high vacuum both on (0001) A and (000) B surfaces of a CdS crystal, and on (111) A and () B of GaP. The barrier heights are determined from the measurements of both current-voltage and capacitance-voltage characteristics. A difference of the barrier height is found between A and B surfaces. The origins of the difference are discussed in terms of the following effects; 1) the converse piezo effect due to the electric field in depletion layer, 2) the piezo effect due to the mechanical stress introduced by evaporated metal films, and 3) the interface states caused from the differences in intrinsic or adsorbed atoms or molecules.

Gallium arsenide-insulator interface properties were investigated from an analysis of conventional characteristics of MIS structures. Energy distributions of interface state density were U shape having minimum values near the mid-gap energy. The values of minimum density were dependent upon conduction types of substrates and deposition conditions of insulator films but independent of crystal orientations. They were from 0.9 to 20×10¹¹ cm^-2 eV^-1 for p-type substrate and from 1 to 4×10¹² cm^-2 eV^-1 for n-type substrate. An anomalous frequency dispersion of capacitance was observed in the accumulation region on n-type substrate and was explained by deep traps formed by oxygen atoms doped during oxide film deposition.N-channel GaAs-MISFET's were fabricated and the maximum field effect mobility of electron in the inversion layer was 1480 cm² V^-1 sec^-1.

Magnitude and sign of photon induced changes in surface conductance depend on the type of surface oxide which is grown after different surface treatments. The type of surface oxides is investigated by RHEED. The RHEED patterns exhibit independent spot and ring systems respectively. The spots are attributed to monocrystalline InSb. The rings indicate that the surface layers consist of various polycrystalline oxides of the underlying semiconductor depending on the surface treatment. The results of the structural measurements are correlated to the electrical properties of the surfaces as investigated by combined ac field effect.

The avalanche injection in silicon gate-controlled diodes is analysed in terms of the physical theory of hot electrons in silicon as developed by Bartelink, Moll and Meyer. Numerically computed universal plots are given for the calculation of the hot-carrier injection ratio η_i (gate current over junction current) at given maximum interface electric field at breakdown, \mathscrE. \mathscrE=1.4×10⁶ V/cm is experimentally determined, which is substantially higher than previously reported values. η_i≈2×10^-3 is measured, in good agreement with the calculated value.

A detailed investigation has been made of the structures which can be formed when gold is deposited onto atomically clean Si(111) surfaces which are held at elevated temperatures. Three LEED patterns are observed, two of which are associated with essentially two dimensional structures. One of these is a structure not previously reported, and is interesting in so far as the LEED pattern shows an unusual combination of spots and lines which can be interpreted in terms of a limited amount of disorder. The third structure occurs when relatively large amounts (20–30 monolayers) of gold are deposited. The LEED and Auger measurements suggest this structure to be a surface stabilised bulk phase. After forming these structures, Schottky barriers are made and the effect of the structure of the interface on the electrical properties of the diode is investigated.

It is well known that fractional monolayers of sodium can be electrolytically deposited on the Si–SiO₂ interface, but the resulting electronic structure has not been fully determined. We present thermionic emission and photoemission measurements of uniformly coated interfaces which clearly show a reduction of the contact barrier due to an interface dipole layer. The photoelectric threshold, determined by the photoyield corrected for optical absorption Y/α, is reduced by sodium to a saturation value of 2.6 eV at a full monolayer coverage. This is consistent with the thermionic emission results which show an electronic barrier of 2.7 eV at the full monolayer coverage. A simple mathematical model is given to explain this dipole layer.

Usual assumption that participants of an elementary reaction (step) behave statistically independent of one another leads to a relation between the order of the step and the number of its participants, which is used as a handy key to the analysis of reactions. This assumption was shown, however, unreliable when applied to reactions on solid surfaces. The statistical mechanical method of dealing with the rate of steps was obliged to be generalized to take the interaction of participants on one another coherently into account. The generalized method thus developed was summarized and its application were exemplified in treatments of the hydrogen electrode reaction, the flash desorption spectrum and the isotopic shift of sticking probability respectively of hydrogen on W(100).

Nitric oxide is adsorbed on the (110) plane of tungsten with a high sticking coefficient independent of coverage. Programmed thermal desorption of a W(110) surface with adsorbed NO shows no desorbed NO, but only nitrogen. Two peaks are observed in the resulting nitrogen desorption spectra. The first, appearing at low coverages of NO, shifts to lower temperatures with increasing initial coverage. The second appears at higher coverages and shifts slightly to higher temperatures with increasing coverage. The temperature region covering both states is 900–1350 K for the temperature rates employed. A simple, consistent model is developed. Work function-relative coverage measurements for NO on W(110) show an initial slow Δφ increase, a linear portion, and a maximum. The maximum Δφ is 0.85 eV and the saturated coverage value is 0.72 eV.

Even if the interaction with the phonons does not mix the initial and final state of a reaction, it may help it by supplying an energy reservoir. Phonon couplings are different in the initial and final states. We demonstrate they may give or absorb the reaction energy by ways of multiple interactions. This process is a leading one for catalysis with weak mixing, i.e. a large class of reactions. Moreover the same phonon interaction is also responsible for desorption of the atoms or molecules. When catalysis and desorption are both taken into account, we show a compensation law occurs among a lot of similar catalysts.

A new technique has been utilized in the study of surface adsorption and desorption. The technique involves the deposition of transmitting and receiving surface-wave interdigital transducers on a piezoelectric substrate such as single crystal quartz. When a surface wave is excited, certain adsorbed species are observed to desorb. The desorption is very strongly dependent on the peak amplitude of the applied surface wave rather than the average power. This technique may provide a useful tool for studying adsorption-desorption on insulating materials (which are not readily amenable to most other surface analysis techniques), and for thin metal films.

The thermal dissociation of surface OH^- and OD^- groups on high-purity MgO was studied by means of electron spin resonance and by mass spectrometry. The results indicate that residual OH^- and OD^- groups neighboring a vacancy in the oxygen lattice may decompose thermally evolving molecular hydrogen and, at the same time, forming paramagnetic V centers. Schematically: 2OH^-→2O^-+H₂ and 2OD^-→2O^-+D₂ respectively. The dissociation probability of OD^- groups appears to be higher than that of OH^- groups by a factor of about 4.

The oxidation degree of nickeloxide-electrodes in aqueous solutions depends on the electrode potential, as it is known from the surface of passive nickel-electrodes. In addition to current-potential characteristics and the impedance of the solid state/electrolyte interface this behavior was investigated using modulated reflectance measurements. By modulating the electrode potential this method results in changing the oxidation degree in connection with the optical properties of a surface layer. The investigation of Li-doped (0.5–5 mol%) nickeloxide crystals shows, that the change of the oxidation degree is catalytically influenced by the Li-doping. Measurements of the modulated reflectance result that the electrochemical activity of the doped crystals increases approximately parabolic with rising Li-amount.

The various types of surface site exposed on a metallic catalyst are classified and discussed. The distributions of these sites depend upon the type of catalyst, and may be different for polycrystalline bulk material, evaporated films or dispersed particles. Suggestions are made as to the types of reaction which might be demanding or facile.

Behavior of hydrogen and deuterium on copper surface was studied in detail. Desorption spectra, adsorption isotherms and the kinetics of exchange reaction predicted that hydrogen is adsorbed dissociatively and the H₂–D₂ equilibration proceeds via the recombination between adsorbed atoms. The optimum values of rate constants for adsorption and desorption were determined by simulating the reaction time course. Obtained constants were found to coincide with those derived by the transition state theory. Influence of molecular motion on adsorption was examined by analyzing the computed trajectories of hydrogen on a dual Cu site.

In order to study the influence of crystalline faces on the catalytic activity, supported catalysts have been prepared with well facetted Ni crystallites presenting particular faces at the surface. The reduction of synthetic antigorite of nickel gives rise to Ni crystallites oriented with either (111) or (110) planes parallel to the support. Orientations and forms of Ni crystallites were determined by electron diffraction and electron microscopy. The metallic areas were determined from electron microscopy and chemisorption measurements. It was observed that oriented Ni supported silica catalysts are more active for ethylene hydrogenation than randomly oriented catalysts.

The relation between catalytic activity to the hydrogen-deuterium exchange reaction as well as work function and alloy composition of clean surface of copper-nickel alloys was studied. Clean surface of alloys was prepared by argon ion-beam bombardment and annealing, and the surface composition of alloys was determined by an Auger spectroscopy. The catalytic activity was found to be greatest in pure nickel and then to decrease continuously as copper content in the alloys increases. The catalytic activity K_m(X) could be expressed by an equation, log K_m(X)=aX+b, where X is nickel content in the alloys and a and b are constants, while the work function Φ of copper-nickel alloys changed linearly with their alloy composition, i.e., Φ(X)=cX+d (c and d:constants). The catalytic activity factors in copper-nickel alloys were discussed on the bases of electronic structure and behaviours of hydrogen adsorption which have been presented by several workers.

The extended Hückel molecular orbital theory is applied to the dissociative adsorption of a hydrogen molecule onto linear Cu/Ni alloy chains composed of four and five metal atoms. It is concluded that Ni atoms compose an active center in the alloy and the catalytic activity of the alloy is determined by the nature of the highest occupied molecular orbital (or the Fermi level). The present calculation is consistent with the individual atom approach of Sachtler et al.

The initial oxidation of the Fe (100) clean surface at room temperature and the reaction kinetics of hydrogen reduction of the FeO(100)/Fe(100) surface at elevated temperatures have been investigated with a combined LEED-AES-MS system. The progressive oxidation was observed even at room temperature and the activation energy of the surface-to-bulk diffusion of oxygen was determined to be 21±2 kcal/mol. The FeO(100)/Fe(100) surface, formed from the Fe(100)-disordered-O structure by heating, was not reactive to molecular hydrogen, however, when preadsorbed hydrocarbon was existed on this surface, the hydrogen reduction proceeded successively with the activation energy of ∼12 kcal/mol. This reaction can be explained by the Langmuir-Hinshelwood mechanism.

A review is given on recent results of field ion mass spectrometry and other techniques concerning the interaction of oxygen (and some other gases) on silver, sulfurand selenium on platinum and sulfur on tungsten surfaces. Field electron emission and field ionization microscopy as well as kinetic, IR, and EPR measurements indicate unambiguously different adsorption states of oxygen on silver. The oxygen adsorption is influenced by the electric field. Although stable silver-oxygen ions are known, no field ions of these surface compounds can be desorbed, if molecular oxygen is interacting. On platinum the two stable adsorption states of sulfur are the strongly adsorbed α-state (desorbing at T>1200 K) and an adsorbed β-state (desorbing at T>570 K). A third transient phase of adsorbed sulfur is observed only at saturation of the stable states if S₂ is further supplied. On tungsten the two sulfur adsorption states are a less mobile (in average T>400 K) strongly chemisorbed form and a second less stable layer. Field ion mass spectra in combination with "field pulse" techniques indicate the presence of S₅, S₆, S₇ and S₈ on W. A comparison is made with field ion mass spectra of selenium where even higher ion clusters occur.

The chemisorption of NH₃ and N₂, and nitride formation on clean and pretreated molybdenum surfaces were studied by quantitative and high resolution Auger Electron (AES) and Energy Loss Spectroscopic techniques at temperature of 25°C and 450°C so as to elucidate the mechanism of ammonia decomposition on the surfaces. Intensity changes of the Auger peak due to the surface nitrogen during the adsorption of NH₃ revealed that the molybdenum nitride was formed at 450°C while the NH₃ was chemisorbed predominantly at room temperature. The concentrations of the surface species were estimated from the ratio of the nitrogen to the molybdenum peak intensities. The high resolution AES demonstrated that the KLL spectrum of the surface nitrogen at 450°C differed clearly from that at room temperature. The high and low temperature spectra could be explained in terms of a semiempirical calculation on the assumption that the nitride is formed at 450°C and the ammonia is chemisorbed on the surface at room temperature, respectively. Energy loss spectra associated with the chemisorption of NH₃ is also presented. In addition, the adsorption characteristics of NH₃ were examined on sulfur, carbon and oxygen covered surfaces, being related to the catalytic activities for the NH₃ decomposition.

Acetylene interactions with rhénium ribbons and evaporated films has been studied in an ultra high vacuum apparatus by uptake measurements, and thermal desorption monitored by mass spectrometry and Auger electron spectroscopy. Results are similar on both types of samples despite strong differences in cristallographic structures. At room temperature, the initial sticking coefficient is close to unity. C₂H₂ is dissociated on the surface at low coverages, leading to mobile hydrogen atoms. At saturation, the composition of the residue left on the surface is C₂H_1,5 at 293 K and C₂H₂ at 200 K. Thermal desorption shows two hydrogen peaks coming from the residue decomposition. Results of acetylene and ethylene adsorption are compared and shown to be similar. A model, assuming several surface sites, with different adsorption properties is proposed.

Mass spectrometric flash desorption methods were used to study the formic acid decomposition following adsorption at 37°C on clean nickel (110). The surface cleanliness was verified by AES and LEED. Flash desorption of preadsorbed DCOOH showed no H₂, HD, or H₂O peaks, indicating that the acid hydrogen desorbed immediately upon formic acid adsorption. Isothermal decomposition showed that a surface explosion occurred to form CO₂ and D₂. The kinetics of the explosion were consistent with a mechanism the rate of which was dependent on the local surface concentration of formate complexes and bare metal atoms. The first-order rate constant for the DCOOH decomposition was 1.6×10¹⁵ exp {-26.6(kcal/gmole)/RT}sec^-1.

In the range of 2100 to 3900 Å wavelength irradiation carbon monoxide is oxidized by oxygen into carbon dioxide at room temperature on the surface of titanium dioxide (anatase) non porous particles. Kinetic and isotopic exchange results show two simultaneous paths for this oxidation of CO, one requiring oxygen from the lattice of TiO₂, the second involving adsorbed and dissociated oxygen.

By a careful adjustment of the partial pressures in a mixture of carbon monoxide and oxygen, a gas composition can be realized which gives stoichiometric coadsorption of both species on platinum films. Under such conditions the reaction proceeds explosively and oscillations are observed. The nature of the adsorbed phase is deduced from the electrical properties of a Schottky diode.

Effect of activating treatment on the evaporated films of nickel has been investigated on hydrogen adsorption and hydrogen-deuterium exchange reaction in the pressure range of 10^-6–10^-3 Torr at room temperature. Adsorbed hydrogen can be distinguished into reversible and irreversible parts, the former can be removed easily by pumping at room temperature and the latter can not. Freundlich isotherm is applicable to the reversible hydrogen and the irreversible adsorption is decribed by Langmuir isotherm with dissociation. The ion-bombardment gave drastical increases of hydrogen adsorption, although surface area of nickel film measured by xenon adsorption showed no change by both bombardment and annealing. The rate of exchange reaction was found to obey the rate law, V = kP^1.3 in the both cases after bombardment and annealing. It is concluded that the rate of the exchange is controlled by the reaction step, 2H(a)_irr+D₂(a)_rev=2 HD, and that small part of the dissociative hydrogen which is adsorbed irreversibly might be effective on the exchange reaction near the room temperature.

There is a need to develop cheaper catalytic materials with improved surface interaction properties. These materials are important for the creation of energy and the utilization of energy products (for example, coal gasification and auto exhaust purification). A new technique for determining the structure of catalysts and relationships between catalytic atoms and reactive species is presented. The techniques involve the Fourier analysis of extended X-ray absorption fine structures (EXAFS) with necessary supporting measurements such as EPR, LEED, etc. Examples of data from EXAFS and EPR measurements are given and catalytic relationships discussed.

The study of interaction of molecules with surfaces is of great current interest with a view to understanding the nature of the surface bond, chemisorption and catalysis. We wish to investigate a prototype system–namely the interaction of formaldehyde with a silicon substrate. A "surface-molecule complex" or cluster model concept is used to carry out the electronic structure calculations using an SCF-procedure based on the statistical exchange approximation and muffin-tin potentials. An appropriate cluster consisting of formaldehyde molecule and several silicon atoms was chosen. By comparing the results obtained for this cluster with the electronic energy level structure of the free molecule, it was possible to recognize an "adsorption orbital" which may indicate the formation of a surface bond. The charge distribution on the remaining orbitals of formaldehyde remains essentially unaltered. Similar conclusions have been arrived at by Eastman and Demuth from their photoemission data of hydrocarbons on a nickel surface.

The competitive co-adsorption of CO and NO on the principal faces of Pt has been studied at 293K, as well as the Langmuir-Hinshelwood reaction between them which occurs when the surface is raised to higher temperatures. A variety of methods were used including LEED, Auger spectros copy, flash desorption, and surface potential measurements. Both gases adsorb nondissociatively at room temperature, and co-adsorb with a mutual perturbation of binding states. CO appears to displace NO with a rather high efficiency, and the kinetics of this process were studied. The surface reaction between ¹²CO and ¹⁴NO leads to products at 28, 32, and 44 AMU. This information is ambiguous because of certain obvious mass interferences; these interferences were completely resolved by experiments using ¹³CO and ¹⁵NO which showed that N₂ and CO₂ were the major products with N₂O and O₂ as minor products. Studying the product distribution as a function of relative reactant concentrations led to an overall mechanism which involves the thermal dissociation of NO as the first step.

The oxidation of primary, secondary and tertiary aromatic amines on oxide surfaces has been studied, using ESR spectroscopy. All the amines give adsorbed radicals whose spectra are not well defined. Addition of a polar molecule displaces the surface species. The nature of the desorbed product depends on the oxide surface involved, e.g., with diphenylamine on alumina the desorbed product is diphenylnitroxide, whereas from a silica-alumina (∼13% Al₂O₃) surface N,N'-diphenylbenzidine is obtained. These facts and other evidence indicate that the electronacceptor sites on silica-aluminas of low alumina content differ substantially from those on alumina. On hydrogen Y zeolite two types of site capable of oxidizing aromatic amines exist, depending on the degree of dehydroxylation of the zeolite surface. One of these sites corresponds to an oxidizing site present in silica-alumina; the other is more closely related to the oxidizing centre found on the alumina surface.

When thermal energy Helium or Hydrogen atoms are scattered from a perfect crystal surface quantum mechanical elastic diffracted scattering is possible. The Debye-Waller factor is the proportion of the total scattering which is elastic and recent experimental results confirm an earlier proposition of the author that the long-range, attractive atom-surface potential should be treated differently from the short-range repulsive part. It was suggested that the attractive part of the potential increases the kinetic energy of the atom but does not cause inelastic scattering. It is shown in this paper that such a result follows from a proper scattering treatment of the potential even in the absence of the line-narrowing phenomenon previously invoked. Formal results are given for the Debye-Waller factor.

It has been shown previously by the author (J. Phys. C6 (1973) 1229) that the quantum mechanical scattering of an atom from a surface represented as a set of infinitely repulsive spheres can be calculated by a method similar to the KKR band structure technique. In this paper the method is extended to allow for the spheres to have a larger radius, as is appropriate for certain experiments. When the spheres are supposed moving according to the phonon motion of the surface the new procedure is shown to allow an improvement in the accuracy of the calculations. Extensions of the technique to diatomic surfaces and more general potentials are discussed.

The vibrational structure associated with the one-phonon inelastic scattering of light particles from the (001) surface of LiF are studied and expressed in terms of surface projected phonon densities. The surface vibrations and the related densities for the semi-infinite ionic lattice are calculated in detail by means of the Green function method, in the framework of the breathing shell model. The inelastic reflection coefficient due to the creation of a phonon of given frequency and surface wave-vector is calculated within a semiclassical description of the particle motion, for Lennard-Jones interaction potentials. The nature and the amount of information on the surface dynamics obtainable from the velocity analysis of the scattered beam are discussed.

A simple theory of atom scattering from solid surfaces can be obtained by representing the atom-solid potential as an infinite step located at the surface, whose shape is taken to be periodic according to the crystal symmetry. When the surface shape is sinusoidal and the corrugation is slight, the diffracted intensities can be evaluated in terms of Bessel functions, as was done by Rayleigh for sound and light scattering. Most relevant is the presence of two (groups of) peaks, located near the classical "rainbow angles" introduced by McClure, and much more intense than the other peaks. This is the quantum surface rainbow. Between the rainbow angles the diffraction intensities oscillate, while on the outer sides the intensities fall very rapidly to zero. Extensions to metal surfaces and inelastic scattering are considered.

By using an atom-surface scattering apparatus which takes advantage of the use of liquid helium temperatures, the scattering of nearly monoenergetic He, Ne and H₂ from the cleaved (001) surface of LiF has been studied with good angular resolution and sensitivity. Owing to the low temperature surface (10 K for He and 80 K for Ne and H₂), elastic diffraction peaks were resolved up to high order. The observed elastic diffraction patterns are qualitatively explained by the quantum theory of surface rainbow developed by Levi et al. In the case of H₂ scattering, inelastic peaks related to energy exchange between translational and rotational motion of the incident particle are observed and discussed.

The inelastic scattering of He atoms from a cleaved (001) surface of NaF has been studied at different surface temperatures. The study was similar to that for the He–LiF system in that the intensity of the inelastic streaks in the vicinity of diffraction peaks was explored as a function of detector location and crystal rotation. At 140 K the phonon absorption events were less probable than phonon emission events, i.e. atoms tended to lose energy while at higher surface temperatures the probability of energy gain increased. The position of the streaks moved predictably in space by varying the incident momentum of the beam, a parameter which also influenced the intensity of the streaks.

The scattering of ⁴He atomic beams from several alkali halide crystal surfaces has been studied. The surfaces are prepared by cleavage in situ at pressures usually in the low 10^-10 Torr range and are held at room temperature throughout. The atomic beam is highly monochromatic and has a relatively long wavelength (1.04 A). LiF has previously been found to give a relatively simple pattern of principal selective adsorption transitions with a superimposed "fine structure." NaF is more complex; more transitions occur and only a tentative indexing can be given. NaCl is still more complex and no indexing can be given as yet.

Velocity selected hydrogen and deuterium atoms with kinetic energies from 20 meV to 140 meV have been scattered and diffracted from a NaF (001) cleavage plane. Highly resolved minima of selective adsorption yield binding energies and the gas-atom-surface interaction potential perpendicular to the surface. Specular intensity versus incident energy curves and specular intensity versus angle of incidence curves for different surface temperatures are compared to available theories. From these measurements the Debye-Waller-factor is extracted, which depends on the angle of atomic incidence and is characterised by a surface Debye temperature of θ_D⊥=370 K and the depth of the attractive potential D₀=17.9 meV. The strength of the (first order) periodic part of the potential is found to be D₁=0.025 D₀.

After a general introduction on various processes causing surface faceting a phenomenon is demonstrated which is caused by surface sputtering (Ne⁺ on Cu (100) surface) and which is extremely angular dependent. A variation of one degree in the range of 37° to 41° angle of incidence relative to the normal on the surface can cause dramatic changes in the optical appearance. At 39°, and room temperature, with Ne⁺ 10 to 20 keV, there is a milky shine which is due to the Tyndall effect. The facets on the surface are very regular at an interdistance of the wave tops of about 8000 Ångstroms. This appearance is thought to be due to the building up of dislocation arrays due to perpendicular replacement focusing into the bulk of the Cu crystal. This process is maximum at 39° angle of incidence. The beautiful development of (10) planes during this sputtering process is bound to temperatures lower than 350 K. For higher temperatures the etching becomes more roling, which is also the case in case of Ar⁺ ion bombardment. Ar⁺ produces a roling surface, Ne⁺ sharp facets, and He⁺ blisters. Doses are very high (10¹⁹ ions cm^-2).

It is the purpose of this paper to extend the information that exists on interpreting molecular beam data to the case of a second order surface reaction and to describe simple tests which can be made which distinguish second-order surface processes from other classes of chemically reasonable mechanisms. A numerical simulation, analogous to the operation performed in the actual experiment, has shown that second-order kinetic surface reactions are easily detected by their dependence on beam pressure and/or background pressure.

The mean residence time () of potassium atoms, ions or their intermediate states on clean and contaminated tungsten surfaces was studied by means of a pulsed molecular beam method. Taking into consideration the shape of the gate function, which can be determined from the experiment, and assuming the transient adatoms (M) and adions (M⁺) are in equilibrium, we estimated by a new method of analysis from the measured intensity function of desorbed ions. Using an Arrhenius-type expression =;⁰exp (/kT) and taking in to account the temperature dependence of the ionization coefficient β, we found the experimental desorption energies () to be 2.30 eV for a clean surface and 1.85 eV for a contaminated surface. The corresponding pre-ex-ponential factors (⁰) were found to be 3.24×10^-14 secand 2.2×10^-12 sec, respectively.

The angular distribution of hydrogen scattered from a (100) oriented tungsten crystal has been measured at room temperature for both the clean surface and the hydrogen-covered surface. Of the molecules incident on the clean surface, 81% were chemisorbed, 16% were scattered diffusely according to the cosine law, and 1% were specularly reflected. The remaining 2% appeared in a secondary peak, the origin of which is not fully understood. Scattering from the hydrogen-covered surface was dominated by an intense specular reflection. Diffraction was also observed and eight first-order diffraction peaks were identified.

A knock-on effect during ion bombardment was studied from computer simulation. The preliminary results obtained for the bombardment of argon ions with energies of 5 keV and 10 keV on silicon a target having a random lattice are described.

Intervention of adsorbed films between a bare metal substrate and a contiguous non-adsorbing gas causes readily observable and even spectacular changes in the thermal accommodation coefficient (AC) of the gas on the surface. Methods of measuring such effects and some resulting observations are described. Applications of such observations to gain knowledge of adsorption phenomena are illustrated. Particular attention is given to following the kinetics of adsorption of alkali metals on originally clean and on H₂ covered tungsten. An application of AC observations to determine the heat of sublimation of K is made. A unique case of decreasing AC with increasing coverage is observed during the second half of monolayer formation of alkali metals on tungsten. The effects on the AC such as described provide a wide-open field for theoretical investigation.

A rather comprehensive experimental examination of the effect of surface impurities (potassium and cesium) on the thermal accommodation of rare gases on a tungsten surface has been made by Roach and Thomas. The present work examines these effects from a theoretical viewpoint with particular emphasis on the effect on the thermal accommodation of helium on a tungsten surface caused by the adsorption of cesium ions and/or atoms. For this examination, the theory of Allen and Feuer has been utilized using reasonable values of the interaction parameters for helium with the adsorbate as well as the force constant for the adsorbate-substrate bond. Resulting values of the accommodation coefficients are compared with experimental values.

A dynamical theory of the relaxation of gas atoms in contact with a surface of a crystal lattice is presented. The theory is classical-mechanical and is based on the non-equilibrium statistical mechanics developed by Prigogine and his collaborators. The model gas-crystal system treated here is very similar to the model systems used by Bak et al. and Rice and Frisch in their dynamical theories of diffusion of atoms in crystals. The energy distribution function of adsorbed atoms on a surface of a crystal lattice is investigated. A differential equation that governs the temporal evolution of the distribution function is obtained. The distribution function can be used to calculate the desorption rate and other quantities.

Energy-loss spectra for the scattering of noble-gas ions by solid surfaces are determined to analyse these surfaces. By choosing the appropriate experimental conditions it is possible to study both the composition and the structure of a surface. The direction of the polar axis of single crystals with a noncentrosymmetric structure such as ZnS is easily established by comparing the ion scattering spectra for the (111) and () face. For crystals having unstable polar faces the polarity can also be determined. This is illustrated for the stable (110) face of GaP by comparing the scattering of ions incident in the <111> azimuth with the scattering of ions incident in the <> azimuth. Other applications of ion scattering such as a study of hydrogen adsorption on metal surfaces are presented. A study of oxide cathodes provides a better insight into the interaction between ions and electrons.

Ions liberated from an adsorbed layer by electron stimulated desorption are shown to have strongly focused and symmetric angular distributions. We have examined the Electron Stimulated Desorption Ion Angular Distributions (ESDIAD) for O⁺ ions desorbed from oxygen adsorbed on W(100); the ESDIAD patterns were displayed on a phosphor screen and photographed. ESDIAD patterns have been examined as a function of oxygen coverage and surface heat treatment; in most cases, patterns of 4-fold symmetry were observed. Dependent on oxygen coverage and sample temperature, the patterns were oriented with the 4-fold axes either parallel to or at 45° to the rows of surface atoms in the W(100) plane, as determined by LEED in the same apparatus. In addition, systematic variations in the appearance of the patterns (sizes and intensities of 4-fold "lobes") were seen for heat treatments at T\lesssim1000 K and may be related to the formation of surface oxides. We propose that the details of symmetry and intensity observed in ESDIAD patterns may provide new insights into the determination of adsorption sites and the nature of the bonding of atoms and molecules to surfaces.

A theory of the determination of characteristic surface vibration temperatures (so-called `surface Debye temperatures') by molecular beam scattering is outlined. The basis of the theory is the reduction in the intensities of the specular and diffracted beams which is due to the thermal motions of the surface atoms (the `Debye-Waller effect'). A combination of the theory of this effect and the diffraction theory of N. Cabrera, V. Celli, F. O. Goodman and J. R. Manson (Surface Sci. 19 (1970) 67) allows calculation of the specular and the diffracted intensities as functions of the experimental parameters (in particular, as functions of the ratio of (a) the surface temperature and (b) the characteristic surface vibration temperature). The theory is applied to some experimental data of H. Hoinkes, H. Nahr and H. Wilsch (Surface Sci. 33 (1972) 516) on the system H–LiF(001).

Helium scattering from the (112) face of tungsten has been shown by Stickney to,exhibit diffraction when the incident beam is in the (10) direction, i.e. perpendicular to the ridge-trough structure of the surface. This study reports a more complete study of helium diffranction from this surface. The positions of the (0) and (10) beams are shown to follow the Bragg relation for incident angles from 20° to 70° and the relative intensities of the diffraction beams to the specular beam are in agreement with simple diffraction theory. Debye-Waller analysis of the thermal attenuation of the specular scattering leads to unrealistically high values of the surface Debye temperature while electron scattering from the same surface gives Debye temperatures of 1/2 to 2/3 the bulk. Neon scattering in the (10) direction gives rainbow scattering while argon is scattered in a single pseudo-specular lobe. Attempts to deduce interaction energies from measurements of selective helium adsorption will also be described.

Methods of LEED-intensity calculation are reviewed. The discussion is mainly concerned with the theoretical principles. The basic physical ideas are clarified both in terms of the Bloch-wave theory and in terms of the multiple-scattering theory. It is shown how the theory is constructed in close connection to the theory of electron scattering by atoms and molecules, band theory, theory of surface states, and the dynamical theory of X-ray and high-energy-electron diffraction.

A method for the calculation of shape and height of potential barriers in interfacial layers is proposed. The transmission function of complex structures can be calculated for arbitrarily chosen potential models, while the actual transmission function is derived from the energy distribution of electrons having traversed the given potential barrier. Therefore, the potential is determined by adjusting model parameters by optimization techniques to match the model and the actual transmission functions. This method has been applied to low work function surfaces. The main features of the method in connection with the data obtained for Cs₂O covered surfaces of p⁺-GaAs are discussed.

A LEED theory based on the Born series for the diffraction amplitude in the momentum representation is presented. The diffraction amplitude is calculated by an iteration scheme starting with the zeroth-order solution taking into account all the forward scatterings exactly, which are dominant in the multiple scattering processes. LEED spectra for the (0, 0) and (1, 0) beams from the copper (100) surface have been calculated with good agreement with experiments. The fifth order solutions are in good agreement with the higher-order solutions within about two percent. A highly simplified method of Tait, Tong, and Rhodin is made more accurate by including all the forward scatterings exactly within the same simplified scheme.

With a triple axis low-energy electron diffractometer, the intensity curves of the specular reflection from the cleavage face (001) of magnesium oxide single crystal are measured against the glancing angle, for the electrons in the energy range from 100 eV to 180 eV and various crystal azimuths. Structures of the intensity curves are interpreted in terms of the surface-wave resonance effect.

Effects of the excitation of tightly bound valence electrons were studied on the inelastic low energy electron scattering (ILEES) in insulating crystals by means of a diagrammatic method. The effective interaction between an electron in the conduction band and a hole in the valence band was obtained by considering the diagrams used in the random phase approximation for metals. The characteristic plasmon frequency was modified in this case by the energy gap. Contrary to the case of metals the exciton state is proved to exist. Using the two-band model with simplified assumptions on the wavefunctions estimates were made of these effects on the polarization propagator which is directly related to the differential cross section of ILEES.

The contribution of backscattered electrons to the Auger and secondary electron yields was examined by in-situ measurements while evaporating Be onto a Cu substrate at constant rate, 1.2 Å/min. The ratios, β_A and β_S, of the Auger (KVV) and secondary yields of Be due to backscattered electrons, to those due to primary electrons, were experimentally obtained from the Auger and true secondary electron yields versus backscattering coefficient η curves, and were evaluated to be ∼1.6 and ∼4.4 at a primary energy of 1.0 keV, respectively. The backscattering factor r for Be (KVV Auger) was experimentally found to be ∼1.1 at a primary energy of 1.0 keV.

The Monte Carlo simulation was applied in order to calculate the escape process of the secondary electrons in a metal. The calculated energy distribution of the secondary electrons agrees satisfactorily with the experimental results. The lateral distribution of the secondary electrons is calculated to be of a breadth of ∼10 Å for normal incidence, which determines the theoretical resolution limit of SEM images formed by the secondary electrons. The energy distribution of the secondary electrons excited by Auger electrons is spread over a wide range from the energy of Auger electrons to zero.

Major advances occurring during the past two years in the development of the theory of low-energy electron diffraction (LEED) are surveyed. The quantitative nature of mathematical models of LEED from periodic systems is indicated by reference to analyses of observed LEED intensities from low-index faces of aluminum, nickel and copper. Extensions of these models to describe elastic LEED from partially disordered adsorbed overlayers are noted. Applications of model calculations to ascertain the atomic geometry of adsorbed overlayers are described. Analyses of observed LEED intensities for c (2×2) overlayers of oxygen and sulphur on the (100) faces of nickel and copper serve as examples of such applications. The analysis of inelastic LEED intensities to determine the dispersion relation of surface plasmons is illustrated for Al (111).

Iodine was adsorbed on silver (111) at room temperature forming a (√3×√3)-30° structure. Thermal desorption experiments, mass analysis and measurements of the change in work function were carried out in addition to LEED to get a realistic adsorption model: The iodine was found to be adsorbed atomically with one atom per unit mesh. The position of the iodine atom within the unit mesh must be deduced from the LEED theory using experimental intensity data. For this purpose intensity-versus-energy curves of several diffraction spots were obtained by means of a spot photometer.

By comparison of measured LEED intensity versus energy curves with intensities of dynamical model calculations, the adsorption site of an iodine atom in a √3×√3/30° superstructure on a silver (111) surface has been determined.

The origin of spin-polarisation in low energy electron diffraction is reviewed. The density matrix description of spin-dependent scattering is introduced and applied to the measurement of spin-polarisation in a Mott analyser. Theoretical intensity and polarisation versus energy curves are presented for LEED from W (001). The results indicate that spin-polarisation analysis could enhance the power of LEED for surface structural analysis.

Some observations on the angular-energy-distribution of the ion induced electron emission are explained as an effect of electron motion in the crystal lattice. Using two-beam and four-beam approximations of the dynamical electron-diffraction theory, minimum emission from Cu (111) is explained as being due to a Bragg reflection of the electrons. A reasonable agreement of calculation with experiments is found.

The (100)-cleavage surface of EuO has been investigated by LEED and AES. The cleavage was prepared in a LEED-system at a background pressure of about 3×10^-11 Torr. LEED patterns of just cleaved surface showed square structure composed of strong reflections with h+k= even and week reflections with h+k=odd. The odd reflections disappeared with time and the pattern become to that expected for the NaCl-type crystals. The Auger-electron spectrum of the surface shows a group of strong peaks at about 79, 100, 120 and 138 eV. The 100 eV-peak is about five times stronger than the 515 eV-oxygen peak for a primary electron energy of 2.5 keV. This peak group can be identified as N_4,5-spectrum of europium. At low primary energies a strong loss peak at 142 eV loss energy and weaker peaks at 258 eV and 359 eV have been observed, which can be identified as N_4,5, N₃ and N₁ loss peaks of europium.

Electron impact ionization threshold properties of surface atoms can be investigated by observing the Auger electron yield as a function of primary electron energy. Such measurements for L-shell ionization of Cl, Ti and Cu exhibit threshold structures which are related to the respective density of states for the scattered electrons. For example, Ti exhibits strong ionization threshold enhancement caused by the high density of state sjust above the Fermi level. It also is found that the density of states can influence the ionization cross-section (or Auger electron yield) to energies well beyond threshold. Comparisons are made to Auger yield data using L-shell ionization theory which is modified to approximate the effects of near threshold density of states.

A fast, direct transform method for the determination of surface structures from Low-Energy-Electron Diffraction (LEED) intensity data is presented. The basic principles of a proper, complex transform construction have been studied via model calculations of intensity profiles. It is shown that the limited range and characteristics of data obtained in a typical LEED experiment produces large truncation errors in the transform. These can be regarded as due to the convolution of an "ideal" transform, containing structural information only, with the transform of a "data window" defined by the range and characteristics of the experimental measurements. A deconvolution procedure is described for recovery of the "ideal" transform and hence precise determination of structural parameters. Results of structural analysis of clean surface and overlayer systems using the transform-deconvolution method applied to experimental LEED intensity data is presented.

The following areas will be reviewed with particular reference to theoretical research in Cambridge: (1) The experimental evidence and theoretical conditions for the existence of surface states on solids including semiconductors, transition metals and Schottky barriers. (2) The concept of a local density of states and the termination of bulk band states at the surface, with application to photo-emission and other properties. (3) The practical use of low energy electron diffraction for the determination of the structure of adsorbates on solid surfaces. (4) The application of pseudopotential theory to the lattice spacing of aluminium surfaces.

A general theory is presented which shows how electronic surface states can in principle be calculated for solids in which the potential is represented as a set of non-overlapping muffin tin potentials. It is shown that in general analytical solutions can not be obtained except in certain simple cases. These limiting cases are discussed and the relationship of the result to the one-dimensional treatment of Shockley demonstrated. An approximation technique is described which, it is believed, will enable calculations in the general case to be greatly simplified.

The local density of states on the (100) surface of an FCC tight-binding solid has been determined for an s-band and for a d-band model. Effects due to chemisorption are also treated for the s-band simple cubic substrate.

Electron transfer and bonding between an approaching CO molecule and the (100) surface of nickel has been studied, using recently developed, discrete variational molecular cluster methods.

The self-consistent determination of the ground-state properties of metal surfaces is described within the context of very simple models. Attention is focused on the electron density distribution, work function, and surface energy. The use of this framework to analyze the static response of such surfaces to external charges and the process of alkali chemisorption is discussed.

Hohenberg and Kohn have dealed with the ground state of an interacting electron gas in an external potential and show that there exists an universal functional of the electron density n(r), G{n(r)}, independent of an external potential. A formal theory for deriving the form of this functional is developed. The method of this derivation is based on the use of a nonlinear response theory. We shall derive the full expression of G{n} which is written explicitly by the density n(r). Applying this functional form to the problems of metal surfaces, we shall discuss Friedel oscillations and screening effects in the metallic surface region.

Optical second harmonic generation (SHG) from centro-symmetric crystals is restricted to the surface. A free electron model is used to derive the relevant source terms in the nonlinear polarization. Results of SHG from alkali metal covered surfaces of Ge, Cu and Ag are presented. A mono-layer of Na on Ge caused a tenfold increase in SHG. Drastic variations in SHG with Na deposition on Cu surfaces are attributed to the shifting of surface electronic states relative to the Fermi level.

A simple self-consistent calculation of the electronic energy of solid-solid interfaces between two dissimilar transition metals is described in order to study the adhesion of these metals.

The response of the superconducting surface sheath of pure vanadium to low frequency (<1 kHz) ac fields of small amplitude (h₀<10 Oe), has been studied. The effect of different surface treatments, especially of ion implantation on the flux penetration and on the energy dissipation is analysed.

In the adiabatic approximation the sum of the direct ion-ion interaction energy and the ground state energy of the electrons in the field of the ions is the effective potential for lattice vibrations. We have determined this potential formally for a semi-infinite metal slab. The electronic ground state energy was calculated in the Hartree approximation to second order in the bare electron-ion pseudopotential. The basis states used are those of an electron gas confined by infinite walls some distance above and below the surfaces of the slab. When the effective potential function is expanded in the displacements of the ions from infinite crystal equilibrium positions, explicit expressions for the first and second order atomic force constants are obtained. The former are nonvanishing only in the vicinity of the surfaces, and give rise to changes in the interplanar spacings in the surface layers.

The Mössbauer source nucleus, Co⁵⁷, was utilized as a microprobe for studying the magnetic properties of ferromagnetic metal surface. Carrier-free Co⁵⁷ atoms were electrolytically deposited on a surface of iron and the Mössbauer source spectrum was taken at 4.2 K. The spectrum consisted of broadened six lines and the distribution of the hyperfine field was estimated to be 240∼340 kOe (cf. the bulk value, 340 kOe). The reduction of the hyperfine field suggests that the magnetic moment in the surface has shrunk in some degree.

In order to elucidate the band structure and the energy band gap effects for the electron spin polarization (ESP) in photoemission of ferromagnetic metals, photocurrent is investigated in detail for a one-dimensional model of ferromagnetic metals with two energy bands. Discussions are made for the effect of the band splitting due to the exchange interaction and of the difference of related matrix elements for up and down spin electrons. It is shown that the surface effect, which are proved to be of indirect nature, is important in the positive contribution to ESP. Relation to Ni metal is briefly discussed.

The change in the spin-polarization near the surface of ferromagnetic transition metals due to adsorbed Cs-atoms is studied theoretically.

Work function changes occurring during submonolayer adsorption of rubidium on tungsten substrates were monitored using a photoelectric technique. During adsorption on polycrystalline tungsten, the work function fell rapidly from the clean value to a minimum of φ_min=1.70 eV at θ=0.6. No subsequent rise in work function was observed. For a (100) tungsten substrate, the work function fell to φ_min=1.50 eV at θ=0.6 and then rose to φ_final=2.05 eV at θ=1.0. The values of φ_min and φ_final, together with those obtained by other workers for the adsorption of lithium, sodium, potassium and caesium on tungsten, are compared with those predicted theoretically.

The electron work function variation of tungsten samples has been experimentally studied in a strontium vapor environment: single crystal sample in a thermoelectronic emission microscope, polycrystalline sample in a strontium filled cylindrical diode. The data show the emission current density versus the reciprocal sample temperature T^-1, and work function variation versus the ratio T/T_Sr, T_Sr being the strontium reservoir temperature. Comparison of these results obtained by two methods in different strontium pressure ranges and the study of oxygen influence are of interest for thermionic conversion.

Auger Electron Spectroscopy (AES) measurements of very thin ( \lesssim150 Å) gold (Au) films deposited on clean surfaces of silicon (Si) substrates are made in order to study Si/Au interface. The AES spectra taken from the surfaces of these Au films are composed of both Si and Au spectra suggesting that the interfaces are diffuse instead of sharp. The Si spectra show splitting and are different from those of pure Si. Since the same splitting can be observed in the specimen obtained by quenching from Si–Au eutectic liquid, X-ray Photoelectron Spectroscopy (XPS) is applied to the same kind of quenched specimen to investigate the origin of the splitting, through the changes in the XPS spectrum of Au d-band compared with that of pure Au. Itis proposed that Si in the diffuse interface region is metallic with 4 valence electrons (3s and 3p) to interact with 6s and 5d-electrons of Au.

A quantitative theory of surface and bulk plasmon excitation in several electron spectroscopies (transmission, XPS, and Auger from thin film sandwiches) is presented. It is shown under what circumstances, the usual exponential attenuation law is valid.

The potential felt by a charged particle near an electrolyte (or intrinsic semiconductor)–vacuum interface is calculated. The Poisson-Boltzmann equation is used to describe the mobile charge in both types of substrate together with a classical dielectric description of the aqueous medium (or filled valence band). For charge-surface separations greater than or equal to the inverse screening length in the substrate, a shifted image potential is obtained. The shift distance depends on the surface potential, screening length and dielectric constant of the material. The potential due to the compressibility of any charged molecular adsorbate layer (or surface states) present is estimated and found to be small for a wide range of parameters.

The local atomic environment of the constituent elements of alloys, compounds, and chemisorbed complexes can be probed by transitions involving their core levels. However, dynamic screening of a suddenly created or annihilated core hole complicates interpretation of the lineshapes associated with these transitions, since (1) the threshold shape is modified by coupling to electron-hole pairs, (2) strong satellite structure can be produced by coupling to characteristic excitations of conduction electrons, and (3) screening is more efficient at the surface, which decreases the core electron binding energies relative to the bulk. To illustrate the dissimilar coupling for different spectroscopies, we contrast K-shell spectra of graphite obtained by appearance potential spectroscopy, characteristic loss spectroscopy, and Auger electron spectroscopy.

Photoelectron energy distributions from clean surfaces of Cu, Ag and Au have been determined using 40.81 eV ultraviolet photons. The residual gas adsorption time dependence for the emitting surface at a working pressure of 5×10^-7 Torr was directly determined by monitoring the intensity of photoelectrons from adsorbed gas(es) and found to be consistent with normal models of physi/chemi-sorption. Typical contamination times of 5–10 seconds were observed and photoelectron spectra for the above samples were consequently acquired in a time interval \lesssim 1 second after in situ cleaning the sample surfaces by a scraping technique. Effects due to instrumental broadening of the spectra were removed by a deconvolution process and the results compared with a Monte Carlo simulation and with similar previous measurements using high resolution X-ray photoelectron spectroscopy.

Precise binding energies for the outer valence bands of the alkali (A) and Halogen (H) ions in the Li, Na, K, Rb and Cs halides have been determined using 40.81 eV ultraviolet photoelectron spectroscopy (UPS). Following corrections for polarization effects in the region of photoexcitation according to the model of Mott and Littleton, the data is compared with the Born model for strongly ionic crystals. The direct determination of the band gap in these materials allows ionicity values to be assigned using the thermochemical model due to Pauling.

The photoelectric yield and the reflectance of solid Ar films on an Au substrate have been measured simultaneously for photon energies 10–30 eV. The thickness ranged from 30–300 Å. The synchrotron radiation of DESY was used as a light source. Even at photon energies below the photoelectric threshold of solid Ar there is some photoemission due to contributions from both the Ar film and the gold substrate. The spectra of the yield per absorbed photon exhibit peaks at the exciton energies. The height of the peaks is almost independent of the film thickness. This indicates that exciton decay at the Ar-vacuum or Ar–Au boundary contributes to photoemission.

X-ray photoelectron spectra were observed of NiO and its Li-doped systems, Li_xNi_1-xO (0≤x≤0.3). The profiles of Ni2p_3/2 and Ols spectra significantly change with Li content. It is shown that the Li-doping causes a marked effect on the state of surface oxygen.

The change of the core electron binding energies of various metals and alloys by heating at 400°C in a vacuum were observed. Pure metallic surface state of Fe, Ni, Cu, Ti, Zr and Ta metals got appeared extensively by heating over 5 hrs. Two plausible mechanisms, chemical reduction process and the diffusion process, were proposed. The X-ray photoelectron spectrum from Ti in Ti–O solid solution alloys changes into the various types of spectral patterns during heating.

With the help of second order transfer matrices the focusing properties and second order image aberrations of a spherical electrostatic analyser have been calculated including all effects of the fringing fields. Using these results a new ultrahigh vacuum electron spectrometer has been constructed. Theoretical data for the image aberrations and the window function of the spectrometer are presented and compared with experimental results. A brief description of the apparatus is given.

UV-electron spectra were taken of metal and semiconductor surfaces using a duoplasmatron discharge for the source of the He I-line. The electron energy analyzer consisted of a 4-grid-retarding system. Polycrystalline Au, Ag and Cu, as well as a single crystal of Au, were measured. Energy calibration and work function determination were carried out by means of polycrystalline Pd. The influence of the surface preparation was demonstrated by a MoS₂ single crystal.

The attenuation length for electrons in Al₂O₃ has been determined to vary from 6.3A at 157 eV to 16.7A at 1404 eV by means of an X-ray photoelectron spectroscopic technique. The attenuation of photoelectron and Auger electron lines from a substrate was monitored as the thickness of an overlayer of Al₂O₃ was increased, the results being corrected for the asymmetry of photoelectron angular distributions from various subshells and for effects due to the geometry of the spectrometer. The results have been compared with a number of theoretical models of electron scattering. The major energy loss process has been demonstrated to be due to inelastic scattering involving valence band electrons.

The influence of impurities on the electrical breakdown between electrodes in vacuum has been investigated using Auger electron spectroscopy as a probe for determining the elemental composition of the surface region. Experimental results show that during breakdown copper chloride is the main constituent of the material transported from a contaminated copper anode to a titanium cathode. A simple model based upon increased surface diffusion and build-up phenomena in the high electric field is presented to explain the experimental results.

An argon ion gun is introduced to an Auger emission microanalyzer for the purpose of analyzing element distribution in depth direction of solid specimen. The ion gun device utilized is a model of PHI 04–131. Cooperating ion etching with Auger emission microanalysis to a test specimen of iron plate, a three dimensional element distribution is demonstrated by a set of Auger emission micrographs obtained at each plane parallel to the initial surface of the specimen. Characteristics of the Auger emission microanalysis method are discussed being related to other microanalysis methods.

Secondary ion mass Spectrometry and Auger electron Spectrometry are combined to investigate surface processes on copper-beryllium during heat treatment and oxidation. Using low energy Auger peaks, data are obtained that can be related to SIMS results concerning information depth and chemical binding. It is shown that a copper-beryllium surface undergoes major changes during bake-out.

Dynamic Background Subtraction (DBS) involves multiple differentiation followed by multiple integration of some experimental variable. The technique is applied to Auger electron and soft X-ray appearance potential spectroscopies to remove large background signals allowing one to make meaningful quantitative comparisons from such data. It is found that double differentiation of electron current distributions followed by double integration is sufficient to remove backgrounds in these studies. Using DBS, the predicted quadratic dependence of Auger signal strength with modulation amplitude for second derivative current detection in a four-grid retarding potential analyzer is experimentally verified over a large range of modulation amplitude.

Polycrystalline gold films have been examined by Auger Electron Spectroscopy and LEED. The AES studies have been carried out at medium energy resolution and are part of a programme aimed at the extraction of information from line shapes. Here we report results for the N₅VV Auger transition on the correspondence of the line shape with that to be expected from simpleminded deductions from valence band density-of-states functions.

Depth distributions of B, P and F ions implanted into Si have been determined through the use of Auger electron spectroscopy (AES) combined with Xe ion sputtering. AES sensitivity has been enhanced by combination with a signal averaging calculator. This technique is not affected by the electrical properties of the implanted species. Profiles of both as-implanted and annealed samples are obtained and considerable broadening by annealing is observed.

An attempt was made to determine by AES the surface features of Si(100) samples prepared by different cleaning methods. Spectra were taken in the temperature range 20–1200°C at pressures in 10^-8–10^-10 Torr range without bakeout. The sample etched in HCl vapor showed the cleanest surface after heating in situ.

Argon ion bombarded KCl(100) surfaces are compared with air and vacuum cleaved surfaces by LEED and AES experiment. When Ag was evaporated on both argon ion bombarded and air cleaved KCl surfaces, the former showed better epitaxial growth. From these surfaces, Cl signals were always observed from AES curve. Therefore, a great probability of K and Cl diffusion into Ag film can be expected. When a total secondary electron curve was taken from KCl surface, elastic peak heights at higher voltages (up to 1 KeV) were very small. They are probably due to dissociation of the KCl surface. This is clearly related to LEED observable voltage regions. Alkali halides are materials from whose surfaces LEED patterns are observable up to several hundred voltages. However, Ta and MgO samples can be observed by LEED up to 1 KeV.

We describe adsorption studies of serveral inorganic and organic species on single crystal Ni(111) using ultraviolet photoemission spectroscopy (hv=21.2 eV). Adsorbate orbital ionization energies and line shapes have been measured and surface reactions have been studied. Ionization energies for chemisorbed unsaturated hydrocarbons (C₂H₂, C₂H₄, and C₆H₆) exhibit large surface-induced relaxation shifts (∼1–3 eV) relative to their gas phase counterparts as well as π-orbital bonding shifts (∼0.9–1.5 eV). We estimate these π-d bonding interaction strengths and chemisorption energies using Mulliken's donor-acceptor theory as described by Grimley for weak chemical bonds and show that an observed surface reaction, i.e. the dehydrogenation of chemisorbed ethylene (C₂H₄) to chemisorbed acetylene (C₂H₂) for T\gtrsim230 K, becomes exothermic only for the chemisorbed species due to the π-d electron interaction.

Energy-loss spectra of 90∼150 eV primary electrons reflected from the nickel (100) plane were measured as a function of coverage by oxygen in a retarding-field analyzer with an energy resolution of ΔE/E=0.3%. A characteristic energy loss at 5.8 eV, which was not detected in the clean Ni(100) surface, was observed in the 1/4 coverage. This loss peak split into 5.4 and 6.4 eV in the 1/2 coverage. It is concluded by use of a simple model similar to the simple Hückel molecular orbital theory that the loss peaks are caused by excitation of electrons in a chemisorption bond of Ni–O and the split of the peak corresponds to the formation of divalent Ni–O bonds. The bond nature is discussed in relation with the observed changes in the work function and in the characteristic energy-loss spectra due to core electron excitation of Ni 3p.

Using a combination of ultrahigh vacuum instruments it has been possible to apply the techniques of LEED, XPS and UPS in situ to the same sample. The oxygen/W(100) system has been chosen as model. The results from the three methods are correlated with each other, and compared to electron impact desorption data.

The surface compounds Pt–O and Pt–H were studied by means of soft X-ray spectroscopy using a variety of supported Pt-catalysts. Unexpectedly large chemical shifts were observed in the M-series, amounting for certain bands to as much as +7 and -27 eV with respect to bulk platinum. Other M-bands remain virtually unchanged. The intensity of the shifted bands can be correlated with the degree of dispersion. A model based on surface molecular orbitals will be discussed.

Chemical bonds formed between substrate and adsorbed atoms are characterized by new electronic energy levels which are derived from the atomic orbitals of each constituent. These energy levels are being observed experimentally through the techniques of field emission energy distributions, uv photoemission spectroscopy, and ion neutralization spectroscopy. These spectroscopies will be considered in the light of present day theoretical understanding of both the measurement and the chemisorption process.

Theories which calculate the energy spectrum of electrons emitted in field emission from metallic surfaces with adsorbate have normally considered the atoms to be adsorbed independently of each other. However, in such experiments the surface has appreciable coverage of the adsorbate and interactions between adsorbed atoms are of importance. Calculations have been performed in which such interactions are included. Comparison is made between the results from a "no interaction" model and those from the present one which treats adatom interactions. The interpretation of experimental data can be affected by the surface coverage through changes in the positions and widths of the resonance peaks which appear in the experimental energy spectrum.

Inelastic electron tunneling spectroscopy has been used to probe the irreversible chemisorption at 295 K of phenol, catechol, resorcinol and hydroquinone on thin amorphous alumina films in Al/Al₂O₃/Pb sandwich structure tunneling junctions. This spectroscopy yields a representation of the vibronic excitations of the chemical bonds in the adspecies. Phenol chemisorbs predominantly as C₆H₅O^- on Al⁺³ sites with a very small amount of associative chemisorption at high coverages. Catechol and resorcinol both adsorb predominently as the di-ion C₆H₄O₂^-2, whereas hydroquinone chemisorbs both as the di-ion and as the single ion 1, 4-C₆H₄(OH)O^- at high surface coverages. Hydrogen bonding is observed among the adsorbates and between the adsorbates and the OH groups on the partially hydroxylated Al₂O₃ surface. The experimental results are explicable in terms of the geometry and chemical properties of the adsorbates and the oxide surface.

Taking into account the hopping and exchange interactions between adatoms within a model analogous to the Anderson model, the local density of states at adatom site has been calculated for the H/W(100) system, and assignment is made to the FEED peaks for this system observed by Plummer and Bell. The double peaks ∼1 eV below ε_F for the β₂ state of H adatom and the dispersed peaks for the β₁ state can be assigned to the virtual bound states of nonmagnetic and ferromagnetic H adatoms, respectively. The present calculation agrees with the speculation that both β₂ and β₁ are atomic adsorbed states.

Recently interest has arisen in waves which propagate along the edge of a wedge formed by the intersection of two plane surfaces and are localized at the apex of the wedge. Such waves have been termed edge modes. They have technological applications because wedges can be used as waveguides for acoustical and optical edge modes. A survey will be presented of recent theoretical results concerning the properties of one particular type of edge mode, namely vibrational edge modes in elastic and pieozoelectric media, as well as of existing experimental data bearing on them.

Starting with a discussion of lattice structure and force-constants microscopic calculations for surface modes in diamond and zinc-blende structures are given. It is shown that the details of the modes depend strongly on the structure of the surface; a very detailed knowledge of the geometric surface structure is necessary to make definite statements on frequencies of surface modes. The difference between true and quasi-surface modes as well as the influence of a superstructure is discussed. In some cases a comparison with experimental results is possible; this allows for some statements on the force constants in the surface. Methods for calculation of force-constants in the surface are indicated.

A review is given of phenomena exhibited by surface polaritons associated with surface magnetoplasmons in semiconductors. The dispersion curves for surface polaritons are calculated using Maxwell's equations and the appropriate electromagnetic boundary conditions. The results of specific calculations for n-type InSb are presented for the cases of magnetic field perpendicular and parallel to the surface. Of particular interest is the appearance of gaps in the dispersion curve for certain values of magnetic field and carrier concentration. The situation involving a material with multivalley energy band structure is illustrated by calculations for p-type PbTe where additional surface modes occur. When the surface magnetoplasmon frequency is near the surface optical phonon frequency, effects arising from the interaction of the two modes are found.

Based on the theory of surfons, we present a formalism to calculate the attenuation rate of elastic surface waves valid in any frequency region. The attenuation rate due to cubic anharmonic terms in the elastic energy of an isotropic elastic continuum is given by means of the temperature-dependent Green's function. In the low frequency approximation, our result shows the same ωT⁴ dependence as the attenuation rate of Maradudin and Mills.

Convolution signal produced by a nonlinear interaction between acoustic surface waves on GaAs has been found to be markedly enhanced by light illumination. Spectral response of the convolution showed, besides a gradual change due to ordinary acoustoelectric coupling, a sharp "resonant" peak which can supposedly be attributed to the enhancement of elasto-optical coupling by near-band-gap illumination.

A theory of magnetoplasma surface wave is developed in the case when charge carriers are located at the boundary surface in addition to the bulk plasma. The configuration in which the wave propagates perpendicular to the external magnetic field oriented parallel to the surface is investigated. The dispersion is essentially altered by the coupling to the surface carriers. In the low frequency limit (ω≪\varOmega_s) ω=c|k|, while in the high frequency region (ω≫ω_p and \varOmega_s) ω²=\varOmega_sc|k|/ε^*, where \varOmega_s=4πN_se²/m^*c is the characteristic frequency of the surface carrier and ε^* is expressed in terms of the dielectric tensor. Frequency gaps and isolated branches appear. Particularly, for a definite sense of the direction of propagation a slow wave branch with phase velocity (ω_c\varOmega_s/ε_xω_p²)c exists. The dispersion curve is illustrated numerically for the clean surface of p-type InAs.

Experimental results about radiative decay of surface plasmons induced by 1 8 keV-electrons at sodium surfaces are discussed. The metal-films were evaporated in UHV on substrates cooled down to 77 K. The peak wavelength of an evaporated film (thickness about 0.7 µm) on sapphire is 300 nm (4.13 eV) with a half-width of 80 nm. When using a stainless steel substrate the spectrum is strongly broadened up to a half-width of 135 nm due to a greater surface roughness of the sodium film. The measurements are compared with the theoretical results about the surface roughness coupling between electron induced surface waves and the electromagnetic field. The spectrum calculated with the optically measured optical constants of sodium has a half-width of only 15 nm.