Recent Trends in the Application of Low-energy Electron Diffraction

BY R. W. JOYNER AND G. A. SOMORJAI

1 Introduction

In recent years surface science has become one of the fastest growing fields of the physical sciences. The application of low-energy electron diffraction (LEED) to determine surface structure, and to correlate it to other chemical, electronic, and transport properties of surfaces, is partly responsible for this remarkable development. The importance of surface phenomena in areas ranging from biology to solid state physics does not have to be elaborated. The interdisciplinary character of surface science, however, may be masked when viewed in the framework of a departmentalized academic or industrial structure. Surface science is in the early stages of development when compared with our understanding of the chemistry (structure and dynamics) of gas-phase molecules or with our knowledge of many fundamental properties of the solid state. The atomic structure and the chemical composition that were commonly available in these other fields and have served as the foundation for studies of all other physical chemical properties have been lacking in surface science. In the absence of this information many important surface phenomena have been awaiting interpretation.

Techniques of scattering of electrons from surfaces have provided the means to determine both the surface structure and the surface composition on an atomic scale. Low-energy electron diffraction permits us to develop surface crystallography that, similarly to X-ray diffraction, will be utilized to determine structure of surfaces and of adsorbed molecules of ever-increasing complexity. The analysis of the inelastically backscattered electrons yields, via the Auger spectrum, the chemical composition of the surface.

Electron scattering techniques for surface studies have overcome the difficulty of detecting a very small number of surface atoms (1013 — 1015 atoms cm-2) in the background of a very large concentration of bulk atoms (1022 cm-3). Since the chemical composition of the surface must be known to provide unique interpretation of surface structure, Auger electron spectroscopy which can be carried out in combination with low-energy electron diffraction has aided greatly the development of surface crystallography.

Low-energy electron diffraction plays a pre-eminent role in the determination of surface structure on an atomic scale. In addition, low-energy electron diffraction studies have been utilized to establish correlations between surface structure and many other surface phenomena, i.e. phase transformations, adsorption, condensation (epitaxy), and catalysis. It is the purpose of this Report to review the development of low-energy electron diffraction in the past few years and to indicate future trends of this important field of surface science.

2 The Theory of Surface Structure Analysis by Low-energy Electron Diffraction

The nature of low-energy electron scattering from surfaces is the question that must be answered in developing the theory to be used for the analysis of the diffracted electron beam intensities. Electron scattering by a single atom or by a periodic potential is interesting in its own right, and the interpretation of various properties of the scattered diffraction beams (their absolute intensities, shape of the scattered beam, surface resonance effects) are under intensive investigation. Nevertheless, the main thrust of theoretical work is to develop a simple but viable theory which permits computation of the relative intensities of the diffracted electron beams using a scattering model in which the only adjustable parameters are the positions of surface atoms. To some extent, the development of such a scattering theory that would enable one to perform surface crystallography is less demanding than the computation of various other beam properties. In the region of 30 — 120 eV, where most of the experimental data on the surface structures are detected, there are several diffraction beams and a great deal of experimental information is available. Just as in crystal structure analysis using X-ray beams, one needs to determine only the intensity ratios of the various diffracted electron beams, instead of the absolute intensities. Since the recent publication of a review on low-energy electron diffraction, there has been a great deal of progress in computing the diffracted electron beam intensities. Most of the special characteristics of low-energy electron scattering have been recognized. It has been shown that the magnitude of the backscattered intensity is more than 40 — 50% of the magnitude of the forward scattered beam intensity in the energy range of interest. Owing to the strong interaction between the incoming electrons and the ion cores in the solid, cross-sections for both elastic and inelastic scattering of electrons are about 106 times larger than for X-rays. The strong inelastic scattering effectively removes electrons from the incident or from the diffracted electron beam so that the elastically scattered fraction (which contains the diffraction information) that leaves the surface is 1 — 5 % of the total scattered intensity. The total reflectivity is low, of the order of 1%. The peak widths of the diffraction beams are broad, 4 — 20 eV, and there is a significant amount of multiple scattering. In spite of the multiple scattering events that complicate the intensity analysis, the large inelastic scattering restricts backscattering to about four atomic layers at the surface and greatly reduces the contribution of multiple scattering to the total scattered intensity. A scattering theory that takes into account all of the features of low-energy electrons diffracted from surfaces has been developed and applied successfully to computation of beam intensities from various aluminium crystal surfaces. The computer program to calculate surface structures from the diffracted beam intensities that appears at present to be the simplest uses the T-matrix formalism developed by Beeby and extended by Duke and Tucker to include inelastic damping of the electron beam. The outgoing beams each correspond to a two-dimensional reciprocal lattice vector, g, in the plane of the surface, and we consider only the elastic case where E represents incident and emerging electron...