Scanning tunnelling microscopy (STM) is perhaps the most remarkable new technique to have become assimilated in the surface scientist's armoury in the last 10-15 years. It provides new insights into the role of structure in a variety of surface phenomena, although the interpretation of the atomic-scale images in terms of atomic positions is far from trivial. It is extremely seductive to associate, in a routine fashion, the atomic-scale 'bumps' in images with surface atoms, but particularly in surfaces comprising two or more atomic species it is actually very difficult to identify atomic positions, even parallel to the surface, in a wholly reliable fashion. Even atomic-scale STM is therefore not primarily a technique for determining atomic positions, especially for adsorbates on metals.

The great strength of STM, like of any real-space microscopy, is in identifying inhomogeneous processes at surfaces, such as nucleation and growth of overlayers and adsorbate-induced restructuring at surfaces. The latter can be of considerable significance in determining key aspects of surface reactivity and the exact role of heterogeneous catalysts. With the emphasis on adsorbate structure determination in other parts of the Warwick programme, therefore, UHV STM studies of adsorbate-induced restructuring of metal surfaces have an important role, partly to help provide ideas for possible structures, but more importantly to link local structure to reactivity and give a clearer view of the true inhomogeneity of surfaces often assumed to be 'perfect' or laterally well-ordered.

Inhomogeneity is the raison d'etre of self-organised nanostructure growth, of course, and STM (supported by ex situ AFM and SEM imaging) is crucial to this part of the research programme. It is also a common feature of epitaxial growth, even when the desired end result is a highly ordered epitaxial layer, since during MBE growth non-equilibrium structures are formed.

---

The STM image on the right shows an InAs(001) surface terminated by an array of As dimers. The overall surface symmetry is (2x4), highlighted by red arrows in the figure (the 4x arrow is approximately 1.7 nm long). This periodicity is produced either by adjacent dimer pairs and adjacent missing dimers ('beta' structure) or by a single As dimer with three adjacent missing dimers ('alpha' structure). Diffraction techniques such as RHEED do not pick up the different local structures and the many defects on this surface, but STM can image them beautifully. However, the interpretation of the bright cigar-shaped features in the image as arsenic dimers depends on other surface structure determination techniques - the STM does not simply identify atomic species. Image taken from Surface Science 433–435 (1999) 455–459, G.R. Bell et al.