C. S. Fadley

Angle-resolved x-ray photoelectron spectroscopy

Abstract In this review, various aspects of angle-resolved x-ray photoelectron spectroscopy (ARXPS) as applied to solid state- and surface chemical- studies are discussed. Special requirements for instrumentation are first considered. The use of grazing-emission angles to enhance surface sensitivity and study surface concentration profiles of various types is then discussed. Various effects that may limit the accuracy of such measurements such as surface roughness, electron refraction, and elastic scattering are considered. Several examples of surface-specific electronic structure changes as studied by grazing-emission ARXPS (e.g., valence-band narrowing and core-level shifts) are also reviewed. The use of grazing-incidence geometries for surface enhancement is also briefly considered. Single-crystal studies providing additional types of information via ARXPS are next discussed. For core-level emission from single-crystal substrates or adsorbed overlayers, x-ray photoelectron diffraction (XPD) is found to produce considerable fine structure in polar- or azimuthal- scans of intensity. Such XPD effects can be very directly related to the atomic geometry near a surface, for example, through simple intramolecular or intermolecular scattering processes. A straightforward single scattering or kinematical theory also appears to describe such effects rather well, thus far permitting several structures to be solved by analyses of azimuthal intensity scans. Likely future developments and possible limitations of such XPD structure studies are also discussed. Finally, valence-band ARXPS is considered, and it is shown that pronounced direct-transition effects can be observed provided that the specimen Debye-Waller factor is not too small. A simple free-electron final-state model is found to predict these direct-transition effects very well, and future studies at low temperatures and with higher angular resolution seem promising.

Strong interlayer coupling in van der Waals heterostructures built from single-layer chalcogenides

Significance A new class of heterostructures consisting of layered transition metal dichalcogenide components can be designed and built by van der Waals (vdW) stacking of individual monolayers into functional multilayer structures. Nonetheless, the optoelectronic properties of this new type of vdW heterostructure are unknown. Here, we investigate artificial semiconductor heterostructures built from single-layer WSe2 and MoS2. We observe spatially direct absorption but spatially indirect emission in this heterostructure, with strong interlayer coupling of charge carriers. The coupling at the hetero-interface can be readily tuned by inserting hexagonal BN dielectric layers into the vdW gap. The generic nature of this interlayer coupling is expected to yield a new family of semiconductor heterostructures having tunable optoelectronic properties through customized composite layers. Semiconductor heterostructures are the fundamental platform for many important device applications such as lasers, light-emitting diodes, solar cells, and high-electron-mobility transistors. Analogous to traditional heterostructures, layered transition metal dichalcogenide heterostructures can be designed and built by assembling individual single layers into functional multilayer structures, but in principle with atomically sharp interfaces, no interdiffusion of atoms, digitally controlled layered components, and no lattice parameter constraints. Nonetheless, the optoelectronic behavior of this new type of van der Waals (vdW) semiconductor heterostructure is unknown at the single-layer limit. Specifically, it is experimentally unknown whether the optical transitions will be spatially direct or indirect in such hetero-bilayers. Here, we investigate artificial semiconductor heterostructures built from single-layer WSe2 and MoS2. We observe a large Stokes-like shift of ∼100 meV between the photoluminescence peak and the lowest absorption peak that is consistent with a type II band alignment having spatially direct absorption but spatially indirect emission. Notably, the photoluminescence intensity of this spatially indirect transition is strong, suggesting strong interlayer coupling of charge carriers. This coupling at the hetero-interface can be readily tuned by inserting dielectric layers into the vdW gap, consisting of hexagonal BN. Consequently, the generic nature of this interlayer coupling provides a new degree of freedom in band engineering and is expected to yield a new family of semiconductor heterostructures having tunable optoelectronic properties with customized composite layers.

Surface analysis and angular distributions in x-ray photoelectron spectroscopy

Abstract A theoretical and experimental study of the application of x-ray-photoelectron angular distribution measurements to quantitative surface characterizations is presented. The basic theoretical model that has been used previously to analyze such angular distributions from flat surfaces in the absence of electron-diffraction (channeling) effects is discussed, including certain new generalizations and special cases pertinent to surface analysis. Previous experimental work is reviewed. The predictions of this model are also found to be consistent with new experimental data obtained from gold specimens with carbon-containing surface layers and from aluminum specimens with successive oxide- and carbon-containing layers. An order of magnitude increase in surface-layer relative intensities is observed at low electron escape angles relative to the surface. Also, effects due to x-ray refraction and reflection are found for very low angles of incidence, and these lead to approximately a four-fold increase in surface-layer relative intensities. Extensions of the theory to include the effects of non-uniform x-ray flux, a more realistic spectrometer acceptance function, non-uniformity of surface layers, and surface roughness are also considered, and numerical calculations for the specific case of a sinusoidally rough surface are presented. It is shown that rough-surface intensities will equal flat-surface intensities provided that both surfaces are clean and that no x-ray shading occurs. If surface layers are present, however, rough-surface angular distributions are predicted to deviate markedly from flat-surface distributions. By means of angular distribution measurements, it thus appears possible to selectively enhance near-surface contributions to photoelectron spectra, as well as to obtain information concerning electron mean free paths, surface layer thicknesses and uniformity, and perhaps surface roughnesses.

A differentially pumped electrostatic lens system for photoemission studies in the millibar range

A high pressure photoemission system is described that combines differential pumping with an electrostatic lens system. This approach allows optimized differential pumping without loss of signal, thereby increasing the high-pressure performance by at least 2 orders of magnitude compared to passive differential pumping systems. A general analysis of aperture-based high-pressure photoemission is presented, followed by a description of the prototype system which has operated at pressures up to 7 mbar on a synchrotron beamline. Using this approach, photoemission experiments should be possible up to 100 mbar. Example data are presented for dielectric samples in gas atmospheres, for a copper catalyst under reaction conditions, and for liquid water in equilibrium with its vapor.

Properties of oxidized silicon as determined by angular-dependent X-ray photoelectron spectroscopy

Silicon with thermally-grown oxide overlayers in the thickness range 15–89 A is studied by angular-dependent XPS. Electron attenuation lengths at 1382 eV are found to be 37 ± 4 A in SiO2 and 27 ± 6 A in Si. Single-crystal effects and thin-layers anomalies are also discussed.

Photoionization cross-sections for atomic orbitals with random and fixed spatial orientation

Abstract Atomic photoionization subshell cross-sections and asymmetry parameters necessary for determining the differential cross-sections of randomly-oriented atoms have been calculated within the one-electron, central-potential model and the dipole approximation for all subshells of C, O, Al, Si, S, Ni, Cu, Ga, Ge, As, Se, In, Sb, Cs, Ba, Ce, Ta, W, Pt, Au, and Pb for a photon energy range from 20 to 1500 eV, and the relevant Cooper minima located to within 10 eV. These values are tabulated for general use, together with the associated radial matrix elements and phase shifts. Differential photoionization cross-sections for fixed-orientation s- , p- and d -orbitals have also been derived within the same model for a completely general experimental geometry, and closed-form expressions depending on radial matrix elements and phase shifts are given. For the special geometry of a polarized excitation source with polarization parallel to the electron emission direction, it is further shown that such oriented-atom cross-sections are exactly proportional to the probability distribution of the initial orbital, a result equivalent to that derived by using a plane-wave final-state approximation. However, detailed numerical calculations of cross-sections for oriented Cu 3 d and O 2 p orbitals in various general geometries and at various energies exhibit significant differences in comparison to plane-wave cross-sections. By contrast, certain prior angular-resolved X-ray photoemission studies of single-crystal valence bands are found to have been carried out in an experimental geometry that fortuitously gave cross-sections close to the plane-wave predictions.

CHEMICAL EFFECTS ON CORE ELECTRON BINDING ENERGIES IN IODINE AND EUROPIUM

A theoretical and experimental study was made of the shift in atomic core‐electron binding energies caused by the chemical environment. Two models are presented to account for these “chemical shifts.” The first uses an energy cycle to break the core‐electron binding energies into a free‐ion contribution and a classical Madelung energy contribution. The Madelung energy contributes a significant part of the binding‐energy shift. It can, in principle, be evaluated rigorously although there is some ambiguity as to a surface correction. The reference level for binding energies must also be considered in comparing theory with experiment (or in comparing experimental shifts with one another). Electronic relaxation could also introduce errors of ∼1 eV in shift measurements. The second, more approximate, model consists of a “charged‐shell” approximation for bonding electrons in atomic complexes. It gives semiquantitative estimates of shifts and demonstrates the relationship between bond polarity and core‐electron ...

Instrumentation for surface studies: XPS angular distributions

Abstract The application of XPS angular distribution measurements to surface studies is discussed. The basic principles behind the measurement and interpretatio

The premelting of ice studied with photoelectron spectroscopy

We address the century-old puzzle of the existence of a liquid-like layer at the ice surface near its melting point with new photoelectron spectroscopic tools using synchrotron radiation. Near-edge x-ray absorption shows that a liquid-like film exists at temperatures as low as -20 °C. Near 0 °C this film is about 20 A thick, i.e., six bilayers. With high-pressure x-ray photoelectron spectroscopy, we have further investigated the effect of surface contamination, which is ubiquitous in natural environments, on the state of the ice. Our results show that the premelting of ice can be strongly enhanced by the presence of hydrocarbon contamination.

Surface structure of MBE-grown α-Fe2O3(0001) by intermediate-energy X-ray photoelectron diffraction

Abstract We have used intermediate-energy X-ray photoelectron diffraction to determine the surface structure of epitaxial α-Fe 2 O 3 (0001) grown on α-Al 2 O 3 (0001). Comparison of experiment with quantum mechanical scattering theory reveals that the surface is Fe-terminated, and that the first four layer spacings are −41, +18, −8, and 47% of the associated bulk values, respectively. These results agree reasonably well with the predictions of molecular mechanics and spin-density functional theory previously reported in the literature for the Fe-terminated surface. However, we find no evidence for an O-terminated surface predicted to be stable by spin-density functional theory.