The surface structure of gold

Abstract

The anomalous nature of the first two or three atomic layers of a gold surface has an important effect in a number of technical fields, including catalysis and the bonding to semiconductors, and could be responsible for its resistance to oxidation. This paper reviews our present knowledge of the single crystal surfaces of gold and attempts to reconcile some conflicting ideas that have been put forward on the origin of the anomalous structure observed on the cube face.Gold, because of its relative inertness and resistance to chemical attack, has been traditionally a popular choice by research workers wishing to examine the behaviour of a metallic surface unaffected by oxides or corrosion products. However, the modern surface research worker with extreme criteria for cleanliness on an atomic scale must question all assumptions regarding the composition of any surface. In studies of mechanical, electrical or chemical properties, the main questions to be answered about any clean surface are: "How clean is it?"; "What is the atomic structure and how does the surface deviate from the bulk material ?". The gold { 100} surface in particular has produced a puzzle that, at present, is not completely resolved.Surface physics and chemistry received a boost in the 1960's due to the ready availability of commercial ultra-high vacuum apparatus. In particular, the availability of equipment for low energy electron diffraction (LEED) (1, 2, 3) and for reflection high energy electron diffraction (RHEED) (4) permitted determination of the atomic structure and spacing of the outermost layer of surfaces. LEED utilises electrons of energy in the range 30 to 1000 ev, at which energies the atomic scattering factors are large. Consequently the penetration of low energy electrons in a solid is very low (typically a few atomic spacings) and the diffracted beams give structural information about the outermost atomic layers. The most commonly used LEED apparatus is the direct display type in which the backscattered electron diffraction pattern is observed on a fluorescent screen. The angles between the diffracted beams can be determined from photographs of the diffraction pattern and this gives directly the spacing of atoms in the plane of the surface. In principle, the spacing of atoms normal to the surface can be obtained from measurements of the intensities of diffracted beams as a function of primary beam energy. However, a complete dynamical theory that takes into account multiple scattering of the diffracted beams is needed before such information can be extracted. Such a theory should be available within the next few years.RHEED utilises electron beams of energy 20 keV to 100 keV incident at grazing incidence on surfaces. Even though penetration of high energy electrons is relatively high, the use of grazing incidence (typically of the order of one degree) means that only the outermost two or three atomic layers contribute to the diffraction pattern. Results from RHEED are generally complementary to those from LEED. 15