J.N. Miller

Photoemission studies of the surface states and oxidation of group IV semiconductors

Photoemission utilizing synchrotron radiation has been used to study the electronic structure of surfaes and/or interfaces in the photon energy range up to 400 eV. The Si (111) surface exhibits well‐established filled surface states which disappear after about 103 langmuirs of O2. At this exposure, only a broad shoulder has appeared on the high‐binding‐energy side of the Si 2p levels. This shoulder continues to grow in magnitude until much higher exposures, about 108 langmuirs of O2, when a 2.6‐eV shifted peak begins to grow, indicating a higher state of oxidation of the Si. After exposure to activated oxygen, to stimulate the oxidation process, the chemical shift of the Si 2p level is found to be 3.7 eV, which is the same as observed in bulk SiO2. Similar results were obtained for the oxidation of the Ge(111) surface.

Oscillations in the compositional depth profile of Cu/Ni alloys: A study by UPS

Abstract The composition in the surface region of a (110) oriented single crystal of Cu/Ni alloy was studied by ultraviolet photoemission spectroscopy (UPS) in the energy range hv = 32–240 eV. The data indicate a surface enrichment of Cu and that the composition, as a function of depth below the surface, does not approach the bulk value in a monotonic manner but shows at least one oscillation. The depth sampled by the UPS experiment was varied by changing the photon energy and thus the escape length of the photoemitted electrons. Two surfaces of a single bulk crystal (bulk atomic composition 90% Ni and 10% Cu) were studied. One was the high temperature equilibrium surface (AES measurements yielded a surface composition of ~65% Cu and 35% Ni); the second surface was higher in Ni concentration (AES measurements gave ~65% Ni and 35% Cu) and was obtained by annealing at a lower temperature. Theories based on the quasi-chemical and regular solution models predict a surface enrichment of Cu for the alloy, but do not predict any oscillation in composition below the surface.

An angle-resolved photoemission study of the chemisorption of chalcogens on Cu(100) III. Cu(100) + p(2 x 2)S

Abstract Starting with the three-step direct-transition model of ARPES for bulk materials, which was examined in the preceding paper, we propose a framework for describing changes in the photoemission spectra due to chemisorption. Normal emission ARPES data for Cu(100) with a c(2 × 2)O overlayer were obtained in the photon energy range hv = 11 to 34 eV. These spectra have been compared within the proposed framework with those obtained from clean Cu(100). Changes were found in the Cu emission features which could be explained by the relaxation of momentum conservation perpendicular to the surface in the optical excitation step and by the relaxation of momentum conservation parallel to the surface in the escape step. These changes include a photon energy dependent broadening of the d-band peak and the preferential attenuation of the sharp direct-transition feature associated with the sp-band. Some evidence for a surface resonance at the top of the d-bands has been obtained. Changes in the spectrum of scattered electrons were related to modifications of evanescent final states. A 1.3 eV wide band derived from the oxygen p x,y -orbitals was deduced from spectra obtained at normal emission and along the Γ X and Γ M lines of the surface Brillouin zone. On the other hand, no emission was clearly detected from the oxygen p z -orbitals. Oxygen induced emission above the Cu d-bands was observed and attributed to antibonding states. This emission was directed towards the bulk [011] directions.

Studies of Surface Electronic Structure and Surface Chemistry Using Synchrotron Radiation

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.

Photoemission of adsorbed CO and H2O on Pt: Valence band studies

Abstract In this paper, using monochromatic He and Ne discharge lines, we report the ultraviolet photoemission results for CO and H2O adsorbed on a Pt crystal. For chemisorbed CO, we use the energy dependence of the photoionization cross section to deduce that the 5σ−1π splitting is approximately 1 eV and that the 5σ is located at about −8.1 eV while the 1π is at −9.1 eV. The distribution of metal electrons changes in two ways upon CO chemisorption: a d character peak is found at −4 to −5 eV and emission at Ef is strongly depleted. For the weakly adsorbed water molecule, we find preferential attenuation of Pt states near Ef that has heretofore been observed for strongly chemisorbed systems.

Absence of band-gap surface states on clean amorphous silicon

Abstract Photoemission studies have been made of the clean and oxidized amorphous Si surface at hv = 10·2, 21·2 and 40·8 eV. For the clean surface, no emission from surface states in the band gap is found and exposure to small amounts of oxygen results in no preferential decreases in emission near the valence band edge. We place an upper limit of 4 × 1013 surface states/cm2 in the gap and conclude that any residual states in the gap are not intrinsic to the amorphous Si vacuum interface. A small amount of oxygen induces a broad valence band peak 2 to 5 eV below the valence band maximum; this we associated with a SiOx structure. Furthermore, we find that exposure to 108 L oxygen produces valence band structure similar but not identical to SiO2 and a depletion of the SiOx emission.

Photoemission studies of clean and oxidized Nb and Nb3Sn

In conjunction with Auger electron spectroscopy measurements, we have developed a procedure for obtaining a clean Nb surface using an electron bombardment heating technique. We have then studied the valence band photoemission as a function of photon energy. We find that calculations for the bulk theoretical density of states are in basic agreement with the ultraviolet photoemission spectroscopy data. We have also studied the initial oxidation of the clean Nb surface. Structure in the energy distribution curves suggests three stages in the oxidation process. Furthermore, the valence band energy distribution curves for clean Nb3Sn prepared ’’in situ’’ were measured and showed a strong resemblance to pure Nb as has been observed previously in x‐ray photoemission. Evidence is found for the existence of Nb2O5 and SnO2 in the surface oxides of Nb3Sn.

A photoemission study of clean and oxidized Nb3Sn

Abstract The oxidation properties of Nb3Sn surfaces were studied at room temperature using photoemission spectroscopy. We found that the oxidation proceeded slower on the Nb3Sn surfaces as compared to elemental Nb and that the composites of Nb and Sn oxides are more protective than those of pure Nb.

Angle‐resolved photoemission studies of oxide formation on Cu(100)

We performed ARPES studies of the oxidation of Cu(100). The results are in agreement with a model where oxygen chemisorption above the surface is followed by incorporation into the bulk and finally copper (I) oxide formation. The spectra obtained for Cu2O show little angular anisotropy. They are in good agreement with earlier angle‐integrated measurements of Evans using a polycrystalline sample and ARPES measurements of Lloyd, Quinn, and Richardson also using Cu(100). However, the latter authors appear to have misplaced the Fermi level in their spectra.

Surface Composition of Copper-Nickel Alloys

The surface composition of alloys is of great importance in determining their chemical and catalytic properties. In this paper, a presentation is given of the application of Auger electron spectroscopy and photoemission spectroscopy to the study of the surface composition of copper-nickel alloys. An enrichment of copper in the surface region is observed for annealed surfaces. The effects of sputtering and chemisorption on the surface composition are also treated. By applying a nondestructive technique like photoemission, it has been demonstrated for the first time that there is at least one oscillation in the depth profile of the composition. This observation cannot be explained readily by the regular solution theory which, for instance, correctly predicts the copper enrichment at the surface.