Table of contents

Quantitative surface analysis using ESCA is considered from three perspectives: 1) some limitations on its use for quantitative surface measurements, 2) use with depth profiling to study SnF₂ treated dental enamel and 3) utility of ESCA for studying surface species on catalysts. Enhancement of the surface sensitivity of ESCA by making measurements at low angles is restricted to very thin films. Optimum is a film thickness of ca. one-half the photoelectron escape depth. The utility of a quantitative measurement is influenced by both the distribution of the measured species in the photoelectron escape depth. ESCA is valuable for trace-metal analysis, although calibration curves are non-linear. Dental enamel treated with SnF₂ shows three distinct zones of composition. The outermost layer consists almost entirely of Sn(OH)₂, produced by hydrolysis. The next layer is composed of a mixture of species produced by several reactions. The bulk of the tooth consists primarily of a mixture of apatite and fluoroapatite. The nature of the nickel species on a catalyst surface varies with treatment and with support. On silica, the dominant species is NiO. On γ-alumina two species are present. One is a pseudoaluminate, Ni(+2) ions occupying tetrahedral sites in the alumina lattice. The second is an oxide which is reducible in hydrogen at 450°C.

Core level energies of free metal atoms have been determined experimentally by means of photo-electron and Auger electron spectroscopy and are compared with theoretical and estimated "free atom" values. Theoretical values agree with our experimental energies within 1 eV whereas the estimated "free atom" values deviate from them by as much as 8 eV. The occurrence of strong correlation effects in the Auger spectra of the alkali-earth atoms is demonstrated at the 3d-Auger spectrum of Sr and 4d-Auger spectrum of Ba. The super Coster-Kronig and Coster-Kronig transitions of free Zn atoms have been measured. From a quantitative analysis of these spectra, the level widths Γ(3p_3/2) = Γ(3p_1/2) have been determined and are compared with experimental solid state values and with theoretical ones.

A new photoelectron spectroscopic method is described directed towards the measurement of the photoelectron spectroscopy of trapped ions. An electrostatic retarding field electron velocity analyser incorporating a channel plate multiplier is included with an ion trap in a strong magnetic field. A method is described for combining photoelectron spectroscopic with ion cyclotron resonance experiments to give positive identification of those features in the photoemission associated with double ionization processes of particular mass selected ions. The theoretical form of the retarding field photoelectron curve and its dependence upon β is discussed and preliminary experiments reported which show an abnormality in the angular distribution for argon and xenon.

Photoelectron spectra of the monomeric species of several classes of ionic molecule have been obtained by crossing a molecular beam of the material with a photon beam containing photons of energies up to 50 eV (He I and He II lines). In this way electrons from the s², p⁶ and d¹⁰ shells of the relevant atoms (or ions) were ejected and studied by photoelectron spectroscopy. The spectra are found to be explicable in terms of simple ionic models in which polarization of both negative and positive ions but mainly the former are taken into account. Work is reported on the alkali halides and the halides of Tl, In, Sn and Pb.

The vacuum ultraviolet photoelectron spectrum of the t-butyl radical produced by pyrolysis of 2,2' azoisobutane has been investigated. The band associated with the first ionization potential is the only one for which vibrational structure could be resolved and this has been analysed in terms of ionization of a vibrationally excited t-butyl molecule. From the observed Franck-Condon envelope of this band, it is concluded that the t-butyl radical has a pyramidal C_3v geometry in its ground electronic state and the barrier to inversion is estimated to be (900±100) cm^-1 with an out-of-plane distance of the central carbon atom of (0.30±0.05) Å. The first vertical and adiabatic ionization potentials of the t-butyl radical are estimated to be (6.90±0.01) eV and (6.58±0.01) eV respectively. The higher ionization potential region is interpreted with the aid of semiempirical molecular orbital calculations.

Recent work concerning the photoelectron spectra of unsubstituted and substituted acetylenes and polyacetylenes is reviewed. It is shown that these systems are ideal cases for the investigation of some of the principles which govern photoelectron spectra of organic compounds.

A comprehensive review is given of the K-LL Auger spectrum for members of the third row of the periodic table. This group contains the lightest elements wherein an Auger process can occur without the direct participation of the valence shell. Recent Auger spectra induced by x-ray photoionization on Mg, Al, and Si metals and their oxides are given as well as preliminary results on salts of Na and K, and for comparison, the L-MM spectra of RbCl. Our results have been combined with previous experimental data and theory in order to obtain an overview. Four topics concerned with Auger processes are discussed: Energies, relative intensities, chemical shifts, and satellite structure. Interlaced throughout the entire paper is a discussion of the role played by the chemical environment and electron correlation.

Using a newly constructed charged particle oscillator He II discharge lamp, we have obtained high resolution (⩽0.03 eV instrumental resolution) spectra of the outermost metal core d levels in alkyl compounds of Cd, In, and Pb. The Cd and In 4d spectra of Me₂Cd (Me = CH₃), Et₂Cd (Et = C₂H₅) and Me₃In are split into five peaks due to the combined effects of spin orbit splitting (5/2 λ) and an asymmetric crystal field (the C₂⁰ term which transforms like the electric field gradient). As expected, the spin orbit split Pb 5d levels in the tetrahedral molecule Et₄Pb (C₂⁰ = 0) show no noticeable crystal field splitting. The new photoelectron effect enables electric field gradients to be obtained in the gas phase. Our derived C₂⁰ values are consistent with previously measured nuclear electric field gradients in Cd and In compounds; and the measured C₂⁰ value for Me₂Cd is in qualitative agreement with that calculated by H. Basch from an ab initio self-consistent field (SCF) calculation. Because the measured linewidths are mainly due to the inherent natural linewidths of the d levels, we can readily obtain natural d widths from these spectra. The high resolution spectra of the Cd 4d levels of Et₂Cd enable us to resolve Cd 4d peaks due to Cd atoms and EtCd· radicals. The Cd 4d spectra of Me₂Cd and Et₂Cd taken at 21.2 eV, 40.8 eV, and 51.0 eV photon energies show marked changes in the relative intensities of the five 4d peaks.

Angular distributions of photoelectrons are conveniently studied with the application of a fourfold mirror polarizer on a conventional UV-source. This technique provides plane polarized radiation which promotes maximum sensitivity in the determination of the asymmetry parameter β. As an example polarized HeI radiation has been used to induce the electron spectrum of benzene.

Valence electron spectra of the methyl halides and of methyl alcohol have been induced by means of the HeI resonance radiation at 21.22 eV. A large number of new vibrational bands are reported. The observation of these was facilitated by a high instrumental resolving power and fast data acquisition. In most cases the new bands are observed in ¹A₁→²E transitions but also the electron band of the ¹A₁→²A₁ transition of CH₃I exhibits extensive progressions of closely spaced vibrational bands not previously reported. The vibrational structure of the ¹A₁→²E transitions is affected by vibronic coupling. The structure of the 2e electron band of CH₃Cl probably reflects the operation of the Ham effect. Calculations of the band structure including spin-orbit coupling are performed according to a model which is briefly described.

He I induced valence electron spectra are presented from CF₃Cl, CF₂Cl₂ and CFCl₃. A high resolution electron spectrometer with a computerized multichannel detector system is used for detailed data recordings. Several new electronic and vibrational bands are reported.

The photoelectron spectrum of nitric monoxide NO has been studied with the 73.58 nm line of Ne I. Strong perturbation of the X¹Σ⁺ and a³Σ⁺ bands is ascribed to autoionisation from a Rydberg state. The molecular characteristics of the R-state are discussed on the basis of selection rules, comparison between experimental vibrational progressions and Franck-Condon factors calculations, and a ²Σ⁺ symmetry is deduced. Electronic branching ratios are also examined.

Expressions relevant to optical and electron spectroscopy are presented for calculation of the vibronic energy levels and Franck-Condon factors due to excitation of a Jahn-Teller active e vibrational mode in an electronic E state of C_3v type molecules. Both A → E and E → A electronic transitions are investigated with allowance of different vibrational energies in the initial and final states. Energy levels and Franck-Condon factors including consideration of hot band transitions are explicitly given for some selected cases. Application is made to the fourth electron band at 18 eV of the electron spectrum of SO₃. The vibronic band structure is well reproduced when the Jahn-Teller coupling parameters D₃ and D₄ of the ν₃ and ν₄ active modes are 0.06 and 0.35, respectively. A theoretical estimation is also made of the Jahn-Teller stabilization energy of SO₃⁺. Finally, the X → B Rydberg transition of CF₃I is discussed and a value of 0.01 is suggested for the Jahn-Teller coupling parameter D₆.

A quantitative treatment invoking first-order Jahn-Teller interaction together with spin-orbit coupling is employed to determine the energy levels of ²E electronic states of C_3v type molecules. Both the static and dynamic nature of the interactions are discussed. The vibronic energies as well as vibrational transition probabilities of ¹A₁→²E transitions are presented from two sets of calculations corresponding to different magnitudes of the spin-orbit coupling. An application is made to explain the vibronic structure of the 2e electron band of the high resolution electron spectra of CH₃Cl, CH₃Br and CH₃I. The magnitude of the vibronic coupling is determined in terms of a Jahn-Teller coupling parameter. The result for CH₃Cl⁺ suggests that the spin-orbit splitting is quenched and the geometry of the ground ionic state is distorted to C_s or C₁ symmetry.

The complete KLL and KLM Auger electron spectra of argon have been recorded at high resolution. The satellite lines in the KLL spectrum have been investigated and relativistic multiconfigurational Dirac-Fock calculations have been made to suggest assignments for these structures. Investigations of the neon KLL spectrum are employed to interpret part of the shake-off satellites.

High resolution KLL Auger electron spectra of H₂S, SO₂ and SF₆ are presented and analyzed. The experimental equipment, using a recently designed electron retardation system, is described. For the analysis of the spectra, ab initio computations have been performed including configuration interaction. The discussion is mainly focussed on the satellite structure. The major satellites are due to shake-off in the valence shell accompanying the K-ionization and are found to be chemically dependent. Extra lines are also observed, which are assigned to autoionization processes.

Recent developments of molecular X-ray emission spectroscopy and its interrelation with electron spectroscopy are discussed. Deep-lying level energies obtained by X-ray emission and ESCA are compared. Valence electron structure of small molecules is studied as well as multiply excited inner vacancy states of free atoms.

The present status of experimental and theoretical binding and transition energy determinations is reviewed. Experimental data and the most recent theoretical predictions are compared for the energies of Kα₁ X-rays, M series X-rays, K-LL Auger electrons, K, L₃, M and N levels, and the 4f spin-orbit splitting. In addition, the Kα₁ and L_a data are fitted by Moseley-type diagrams, and data on the shallow levels and the valence bands of actinide oxides are discussed. Comparison shows that the single-particle Dirac-Fock theory and the inclusion of quantum-electrodynamic contributions predicts energies of the innermost levels generally within the accuracy of data, that is in the order of magnitude of 1 eV. However, in the N, O... shells large deviations do occur presumably due to strong many-electron interactions. The inclusion of many-electron effects in the relativistic theory remains a challenge, as do experimental investigations affording an accuracy of better than 1 eV for the various electronic levels.

Calculations of the photoionization cross-sections of atomic and molecular levels for the A1Kα and MgKα lines depending on the wave function of the final and the initial states are analysed. The ns, np and nd photoionization cross-sections of atomic shells for Z ⩽ 36 and the HF, NH₃, H₂O, CH₄, CH₃F, CH₂F₂, CHF₃, CF₄, N₂, CO and C₂H₄ molecular levels for the YMζ line and the photoionization cross-sections of atoms for Z ⩽ 36 and the CH₄, N₂ and H₂O molecular levels for the ZrMζ line are calculated. The theoretical cross-sections of atoms for Z ⩽ 10 for the FKα line are presented too. Photoionization cross-section calculations using Hartree-Fock-Slater potential are in better agreement with the experiment than those using orthogonalized plane waves for photo-electron wave functions.

It is demonstrated that a 5s core hole in Xe is strongly influenced by (mainly) 5 ⇌5²md super Coster-Kronig dipole fluctuations. This leads to a 5 level shift of ∼ 3.1 eV to lower binding energy (monopole relaxation gives ∼ 0.8 eV) and to a prominent satellite spectrum.

The spectral function for a 3 core hole (M_2,3 vacancy) in atomic Kr is calculated within the framework of many-body theory. The resulting position and width of the core level is in good agreement with experiment. In particular, we find that super Coster-Kronig dipole fluctuations give an additional relaxation energy shift of ∼ 3 eV, which is ∼ 30% of the shift due to monopole relaxation. We also study the variation of the M_2,3 width in free atoms as a function of atomic number Z (for the elements Ge, Br, Kr, Rb and Zr) and we discuss the influence of solid state effects.

The energy levels of the TSeF molecule have been calculated by the scattered wave method and compared with the experimental photoemission spectrum for TSeF vapor. There is good overall agreement between theory and experiment. Scattered wave calculations for model TSeF dimers lead to a value of 0.59 eV for the infinite stack bandwidth for the TSeF stack in crystalline TSeF-TCNQ (compare with our earlier values of 0.47 eV for TTF and 0.58 eV for TCNQ bandwidths in TTF-TCNQ). The dependence of the bandwidth on stacking distance is also reported for TTF-TCNQ and TSeF-TCNQ.

Recent theoretical and experimental developments in the ESCA research programme at Durham are reviewed. Detailed investigations reveal the structural dependence of sample charging phenomena which therefore forms an interesting extra information level for the technique. Direct measurements have been made of electron mean free paths as a function of kinetic energy in both linear and cross linked polymeric systems and these reveal a close similarity to typical metals and semi conductors. The large range of information levels available from the ESCA experiment allow the development of quantitative aspects of the surface science of polymeric systems. Theoretical work has been directed towards the investigation of various aspects of relaxation phenomena accompanying core ionization and data is presented on the enhancement of weak interactions in going to the core hole state manifold and in the study of vibrational fine structure for both direct photoionization and shake up states.

The electronic structure (core and valence levels) of various simple polymers has been investigated by ESCA (Electron Spectroscopy for Chemical Analysis), with the aid of original EHCO (Extended Hückel Crystalline Orbitals) calculations giving corresponding density of states and band structure schemes. Simple monosubstituted homopolymers derived from polyethylene ([CH₂-CHX]_n) have been chosen in order to study the influence of X substituents (X = -F, -Cl, -OH, -CH₃, -CH₂CH₃, -C₆H₅, -C₆H₁₁ corresponding respectively to poly(vinyl fluoride), poly(vinyl chloride), poly(vinyl alcohol), poly(propylene), poly(1-butene), poly(styrene), and poly(vinyl cyclohexane)) through the chemical shifts measured for the core levels, and the alteration of the valence band structures. Assignment of the valence molecular orbitals is made with the aid of the theoretical calculations and with comparison of spectra recorded for molecules similar to the monomeric units. Emphasis is put on the very significant information provided by the valence band spectra for polymers containing only carbon and hydrogen atoms. Valence band data are used to calculate physical parameters (first ionization potential, energy gap, work function) of importance to deduce electrical properties for the polymers.

Uv-photoelectron spectroscopy has been used for the qualitative and quantitative analysis of polymer degradation products. The degradation can be carried out under flowing or static conditions and with different atmospheres. Typically, a few μg of decomposition products gives rise to an acceptable spectrum, and decomposition rates of 5 μg min^-1 (0.05% per hour) are easily monitored. The method has been applied to PVC, polycarbonates and PTFE. It is well suited to the detection and determination of small molecules such as HCl, CO₂ and tetrafluoroethylene, which are major decomposition products and N₂ and O₂ whose role in degradation is of interest.

By comparison of the E.S.C.A. valence band spectra recorded in the solid phase for polyacenes (benzene, naphthalene, anthracene, and tetracene) with gas phase results, important energy parameters (energy gap, work function) of these compounds have been obtained. The trends observed in the evolution of the main C1s shake-up positions and intensities serve to assign the π satellites to transitions of the type nb₁ → b₁*.

Valence band X-ray photoelectron spectra of the rutile oxides TiO₂, RuO₂, and IrO₂ show a distinct evolution as a function of the increasing number of metal d-electrons. The observations are found to be in excellent agreement with recent theoretical models. Intensity analysis and photo-ionization cross sections are used to confirm the nature of the band responsible for conducting properties. The result is in accordance with a crystal field model proposed earlier. The deduced number of "free electrons" in the valence band is used to correlate the asymmetry recorded for the core level lines.

ESCA has been used to study the complexation of the alkali cations Na⁺, K⁺ and Rb⁺ by the crown-ether, dibenzo-18-crown-6. The binding energies of cations as well as the counter ions (Cl^-, Br^- and I^-), oxygens and carbons have been measured. The results reveal strong backbonding effects. A comparison of the electron binding energies for the complexed and uncomplexed alkali halides was also performed. The differences are discussed within the Born model for ionic solids and the hypothesis that the crownether complexes constitute a standard reference system for the comparison of electron binding energies in different alkali halides was tested.

The binding energies of Co_2p3/2, N_1s and C_1s and Cl_2p3/2 electrons have been measured for the compound Coen₃ (ClO₄)₃ where en = ethylenediamine. The perchlorate ion has been used as an internal standard. The main theme of the paper is whether it is possible or not to use sub-standards that are related to a certain main standard. In the present case the latter is the tetraphenylphosphonium ion. As sub-standards K⁺, ClO₄^- and Cl^- are tested. In this way the charge distribution is determined for the Coen₃³⁺ and Co(CN)₆^3- ions. The effective charges of cobalt are found to be q_Co = +0.7±0.3 and q_Co = +0.6±0.4, respectively.

The core electron binding energies have been measured for a series of 16 gold compounds. They allow the identification of the rare Au II formal oxidation state in a number of these compounds. The binding energies of the Au 4f electrons are correlated with the partial atomic charges calculated by means of the electronegativity equalization procedure.

The XPS spectra of 19 triphenylphosphorus derivatives are measured in the solid state. Net atomic charges, derived from electronegativity equalization considerations in simple structures provide a better correlation with experimental binding energies when the potential model is used neglecting the calculated molecular potential. Overall correlation is not improved upon introduction of more sophisticated π-bonding characteristics except in a few cases where very pronounced back donation to phosphorus occurs.

It is expected that keratin, the protective protein of the skin, has capability to bind materials for which it is exposed e.g. inorganic ions. To test this hypothesis for sodium ions we studied the Na Auger ¹D line in NaCl-keratin combinations. The keratin methylene C-1s line served as an internal reference and the shift, S, to the Na ¹D line was measured. It was found that S was dependent on P, the molar fraction of NaCl (on a NaCl-peptide bond basis). The experimental values suggest that S depends on P according to the eq.:

S = 13.0-3.2√P

with S expressed in eV. This result indicates the following: (1) There exists some kind of bonding, or association, between NaCl and keratin. (2) NaCl seems to be associated to the peptide regions of the protein. (3) The bonding is second order in the sense that (probably) two moles of bonds occur per mole NaCl. (4) Possible models can be presented for the mechanisms involved and it is tentatively suggested that electrostatic bonding occurs between Na⁺ and the carbonyl oxygen and, under certain circumstances, between Cl^- and the nitrogen of the peptide bond. Besides keratin, a number of other biochemicals and cervical mucus were studied. Cervical mucus associates, according to our experimental findings, sodium chloride, probably in a similar fashion. This has diagnostic interest in human fertility problems. It is not yet known which of the constituents in the cervical mucus that is mainly responsible for the bonding, and this and several other related questions are the subject of continued studies.

The X-ray and UV photoelectron spectra of several nitronaphthylamine isomers are reported. In some cases, the nitro N_1s signal is accompanied by an intense satellite band 1.5-2.0 eV from the main photoline having 60-150% of its intensity; also in the corresponding O_1s energy region a satellite band is present having a larger energy separation from the main peak (2.0-2.3₅ eV) and smaller relative intensity (I_sh = 10-30%). The difference between the ionization energy values of the nitro and amino N_1s signal (ΔN_1s) is found to follow the same trend found for the intensity of the satellite and log , the optical absorption coefficient of the HOMO-LUMO charge-transfer (CT) transition.

These satellites are ascribed to a shake-up process involving a CT transition from the naphthylamine part of the molecule towards a π* MO localized on the nitro group on the basis of a simple model we have previously proposed for nitroanilines and other aromatic derivatives.

The trend of ΔN_1s is thought to derive mainly from a greater "interatomic" valence band relaxation accompanying nitro N_1s ionization than amino N_1s ionization, rather than from variation of the ground state charges at the two nitrogen atoms.

The UPS spectra indicate that there is no sizeable variation of the ground state mesomeric interaction between the two substituents on changing the position of the nitro group, and suggest mutual interaction of the two substituents when they are adjacent.

Satellite structure which occurs in ESCA spectra of transition metal, lanthanide and actinide compounds is rationalized in terms of a simple model. The relationship between high intensity and "covalency" is pointed out and it is shown that the word "covalency" has to be used with caution in this context.

Equations are proposed which relate spin density to intensity ratio and width of interatomic shake-up satellites in the 2p region in ESCA spectra of Cu(II) compounds. In a number of cases these equations are followed. In other cases two distinct satellites appear in the 2p_1/2 as well as in the 2p_3/2 region.

The applicability of the YMζ (hv = 132 eV) and the FKα (hv = 676 eV) X-ray lines for generating photoelectron spectra is demonstrated. The N₂ valence electron spectra excited by YMζ and A1Kα are studied in regard to relative intensities and shake-up structure. Relative intensities of electron levels in Au excited by FKα and A1Kα are measured. Cross-section ratios for FKα and A1Kα excited spin-orbit components in Ar and Au are determined. Experimental and theoretical cross-section ratios of the Hg 5d spin-orbit components are reviewed for excitation energies between 21 eV and 2 700 eV.

Photoemission studies for hv < 400 eV are used to illustrate some of the advantages of synchrotron radiation for studying the surface electronic structure and chemistry of solids. Studies of GaAs and Pt are used to illustrate the following advantages (all of which depend on a source of radiation with a continuously tunable photon energy): (1) both valence and core states can be examined, (2) photon energies can be chosen so that the electron escape length can be minimized, restricting the studies to the first several atomic or molecular layers of the solids, (3) since the escape depth minimum is rather shallow the matrix element for excitation from a given set of levels may be maximized consistent with probing the last few layers, (4) when two sets of levels are degenerate or near degenerate in energy, e.g. the solid valence levels and the orbits of an adsorbed gas, the photon energy dependence of the matrix element can be used to separate the photoemission from the two states, and (5) good energy resolution (often 0.2 eV or better) can be obtained. For GaAs, studies of the surface valence electronic structure (hv ≈ 21 eV) in conjunction with LEED results indicated that rearrangement of the atoms within the last unit cell of the surface plays an important role in determining the electronic structure associated with this last layer of the crystal. In fact, it appears that measurement of this electronic structure may prove to be one of the most sensitive ways of determining the details of the atomic rearrangement. Studies of the Ga and As 3d core shifts on chemisorption of oxygen indicate that electrons are transferred from the As to the oxygen. The next step in oxidation, the formation of true As and Ga oxides, has also been followed using the core shifts. CO adsorbed on metals has been studied extensively previously using photoemission methods; however, neither the CO 3σ levels nor the "shake up" structure has been clearly seen in the past due to background produced by the strong Pt 5d valence photoemission. Making use of the Cooper minimum in the 5d excitation cross section at hv = 150 eV, the CO 3σ and "shake up" structures have been clearly seen for the first time.

With the advent of synchrotron radiation in the 32-280 eV range at the Stanford Synchrotron Radiation Project, it has become possible to elucidate the transition from ultraviolet to X-ray-induced photoemission. This has been accomplished by studies of noble metals. Polycrystalline copper shows a valence-band (VB) profile that approaches the X-ray induced shape at hv ∼ 100 eV. In polycrystalline silver, the 4d cross section follows the atomic curve, with a reversal of VB peak intensities near hv = 110 eV. Strongly anisotropic behavior is observed in copper single crystals, using angle-resolved photoemission (ARP). Normal ARP spectra from Cu(100), (110), and (111) crystals follow the band dispersion through the Brillouin Zone, including a dramatic resonance between the Fermi level (E_F) and 2 eV binding energy for hv = 43-52 eV. High temperature and high photon energy studies demonstrate the importance of the Debye-Waller factor in photoemission leading to a breakdown of the direct transition model. In adsorption studies of CO on Ni and Pt, CO is shown to stand up with oxygen out. For Pt, electrons are found to flow from t_2g orbitals near E_F to CO, and the CO 1π and 5σ binding energies are reversed relative to the gas phase. At higher photon energies, hv = 150 eV, the CO orbitals are very prominent on a Pt substrate. An inversion of the angular distribution of these orbitals and energy-dependent resonances in their intensities provide evidence for final state scattering effects at photon energies above 40 eV.

A review is presented of a series of electron spectroscopic measurements investigating the surface properties of aluminum by its interaction with molecular oxygen. The studies performed at the Stanford Synchrotron Radiation Laboratory have used the unique properties of synchrotron radiation. The tunability of the excitation source in the spectral range 30-700 eV has made it possible to obtain maximum surface sensitivity and measure core-level photoabsorption.

The interaction of NO with copper has been investigated by electron spectroscopy. Nitric oxide is shown to be molecularly adsorbed in a linearly bonded configuration at 85 K; this adlayer is unstable above 120 K, resulting in an exclusively oxygen adlayer at 295 K, the nitrogen being desorbed. At 85 K N(1s) peaks observed at 401 and 406 eV are associated with two different surface species. We suggest that the former dissociates while the latter desorbs at a binding energy of 406 eV on thermal activation of the adlayer.

Interaction of nitric oxide with Cu at 295 K results in dissociative chemisorption with both oxygen and nitrogen retained at the surface. The surface concentration suggests that each nitrogen and oxygen adatom is bridge-bonded to two surface copper atoms. Estimates of absolute surface coverages are made by comparing the intensities of the N(1s) and O(1s) peaks with the Cu(2p) substrate intensity, while valence level spectroscopy (He I and II radiation) supplements core-level spectroscopy for discriminating between molecular and dissociated surface species.

Clean evaporated potassium films have been studied in a combined UV-X-ray photoelectron instrument. The valence band and the K 3p level were studied using 21.2 and 40.8 eV radiation. In the 40.8 eV excited spectrum a structure attributed to surface and volume plasma losses was resolved. Using AlKα excitation the binding energies of the core levels in potassium were determined. Potassium films at a temperature of 77 K were exposed to oxygen at a pressure of about 10^-5 Pa for increasing time periods. The intensity and shape of the core levels as well as of the valence band were studied after each exposure. The shapes of the O 1s line and the valence band excited by 40.8 eV radiation were followed as a function of increasing temperature after an initial exposure of 45 L O₂. For comparison spectra from a sample exposed to about 45 L H₂O were recorded.

Clean and O₂ exposed surfaces of PbSe have been studied by UV and X-ray photoelectron spectroscopy. The O₂ reaction was found to be quite slow. Considerations regarding relative peak intensities of the core levels indicated a diffusion of oxygen into the lattice. The sticking coefficient of O₂ was estimated to be of the order of 10^-10-10^-11.

The surface compositions of stainless steels–Fe, Cr, Ni, Mo-alloys–after scraping in vacuum, mechanical polishing in air, electropolishing and ion etching have been studied by ESCA. The most representative surface composition is obtained after scraping the sample with a diamond tool in UHV. All other methods produce enrichment of some species on the surface. Mechanical polishing in air leads to a preferential oxidation of Fe and Cr and a concurrent accumulation of Mo and Ni atoms underneath the oxide. During electropolishing in perchloric acid Fe is selectively dissolved into the solution, while Cr and Mo are enriched in their metallic states in the surface region. (It is suggested that it is the enrichment of these alloy components which gives electropolished stainless steels their excellent corrosion properties.) Ion etching lowers the Cr and Ni content in the outermost layer of the metal by selective sputtering mechanism.

The sampling depth of ESCA is about two to three times the mean escape depth of the photoelectrons; thus, in general, a surface layer with a thickness of about 100 Å contributes to the signal. Within this range, information concerning the depth distribution of elements and compounds can be obtained non-destructively. ESCA can therefore be used to reveal the nature of the artefacts that are inevitably formed during the removal of surface layers by ion bombardment. Examples are given of the following effects: selective sputtering (accumulation of Cu on brass, depletion of Cr in the removal of oxide layers on steels), implantation of primary ions and surface atoms, and ion-induced reactions (destruction and formation of compounds). Particular attention is paid to the processes occurring in the bombardment of the oxide-metal interface.

Ultraviolet and x-ray excited photoelectron spectra and soft X-ray emission spectra have been measured for the compounds As₂S₃, Sb₂S₃, As₂O₃, Sb₂O₃ and Bi₂O₃. All compounds exhibit quite similar photoelectron spectra of the density of valence states which can be interpreted with the same model based on the atomic electronic properties of the compound constituents accounting for the chemical order and the chemical composition of the compounds. This model is confirmed by the X-ray emission spectra. Cluster calculations using the CNDO approximation have been performed for As₂O₃ and Sb₂O₃. The influence of the cluster symmetry, the cluster size and the CNDO parameters on the calculated spectra was investigated. Reasonable agreement between theory and experiment could be obtained using suitable chosen parameter sets.

High resolution KLL Auger electron spectrum from magnesium, L_2,3M_4,5M_4,5 spectrum from zinc, and M_4,5N_4,5N_4,5 spectrum from cadmium have been measured for both vapour and solid phases using electron impact excitation. In order to observe in detail solid state effects in these Auger spectra and to minimize the contribution of instrumental differences, the measurements for the two phases were carried out with the same spectrometer. From decomposition of the line groups into individual lines was obtained line broadening on going from vapour to solid, (0.20 ± 0.10) eV for magnesium, (0.45 ± 0.10) eV for zinc and (0.47 ± 0.10) eV for cadmium. The kinetic energy shifts of the line components between vapour and solid were found to be (16.0 ± 0.5) eV, (13.7 ± 0.5) eV, and (12.2 ± 0.5) eV for magnesium, zinc, and cadmium, respectively. The calculated vapour-metal Auger energy shifts using Shirley's extra-atomic relaxation model are in agreement with our experimental results.

X-ray photoelectron spectra of valence and core electrons are obtained for AgI, Ag₂S, Ag₂Se, Ag₂Te, Ag₂HgI₄, Cu₂HgI₄ and Cu₂S–compounds which reveal a lattice instability at temperature increase. X-ray photoelectron spectra of valence and core electrons in the chalkogenides AgAsS₂, AgSbS₂, and AsGeS in crystalline and glass-like states were studied. For AsGeS in crystalline and glass-like state the X-ray emission L_II,III-spectra of sulphur were investigated. K- and L_II,III-emission spectra of pure sulphur were also obtained. When the compounds go from the crystalline state into the α-phase, a smoothing takes place of structural singularities in the photoelectron spectra in the region of energy localization of the states of the non-metallic components. The structural singularities are clearly resolved in the X-ray photoelectron spectra of AgAsS₂ and AgSbS₂ crystals but are smoothed in the spectra of the glass-like state. The X-ray photoelectron spectrum of the glass-like state has two maxima near the valence band top while the spectrum of the crystal has only one maximum. For AsGeS in its glass-like state the singularities of the fine structure observed in the L_II,III-emission spectrum of sulphur are less pronounced than in the corresponding spectrum of the crystalline state. When the investigated compounds of Ag and Cu go from the crystalline state into the α-phase and when the chalcogenides AgAsS₂, AgSbS₂, and AsGeS are transferred from the crystalline into the glass-like state the atom charge state does not undergo perceptible changes. The alterations observed in the X-ray photoelectron spectra when compounds transfer from a crystalline state into one with a lattice instability or into a glass-like state are due to the effect produced on the energy structure of valence electrons by the short- and long-range order in the arrangement of atoms of metal and nonmetal components in the compound.

The angular variation of elastically scattered electrons excited by X-rays from various electron levels of silicon and surface or near surface impurity atoms of nickel and gold are reported. The valence band profile of the Si (111) surface is presented as a function of electron take-off angle. Evidence is presented that the Si(2p) binding energy is dependent on the crystal surface geometrical ordering. A discussion of some of the implications of these studies is presented.