Observing Surface Diffusion at the Atomic Level

Since ancient times many have no doubt been fascinated by dust dancing in a sunbeam, but the long history of studies of individual particle motion in diffusion perhaps begins with Brown’s studies in 1827 of the motion of pollen grains seen under the microscope. In these early studies the particles undergoing random motion, or Brownian motion as we now term it, were macroscopic and their motions could be observed and recorded using optical microscopes. In thinking about surface diffusion, however, we are interested in the motions of individual atoms and to study them we require microscopes with atomic resolution. For these we can use the field ion microscope (FIM) if we are interested in metals, while for atoms diffusing on non-metal surfaces scanning transmission electron microscopes (STEM) can be used. Despite the much greater resolution compared to optical microscopy, the phenomena to be observed have many similarities to the motion of macroscopic particles and approaches used in their study remain valid. It is fascinating, for example, to compare the mapping of individual colloid particle trajectories by Jean Perrin1 with the atomic resolution trajectory mapping undertaken more recently by Tsong2, which can be used to define the atomic site lattice for the surface even when this cannot be seen3. Such comparisons emphasise the close links between present day studies of surface diffusion and discoveries in the early part of this century, a feature that seems characteristic of recent developments in surface science generally. In the following sections, however, this point will not be pursued and attention is restricted to recent studies using atomic resolution microscopy.