Abstract
The electron beam in an electron microscope produces lattice displacements when the accelerating voltage exceeds a critical value which depends on the specimen material. The available data on threshold energies and the rates of point defect production in the common elements are reviewed. It has been observed that under certain conditions defect clusters nucleate when point defects are formed. A simple theory of homogeneous nucleation of interstitial clusters is developed which fits the experimental data on copper. The theory is used to predict the time dependence of cluster growth, the width of the denuded zone close to interfaces, the apparent variation in threshold energy for cluster formation with foil thickness and temperature, the interstitial and vacancy concentrations and the enhanced diffusion which they produce. It is concluded that it is possible to avoid loop formation by a suitable choice of conditions, in particular by increasing the temperature and restricting the foil thickness. It is shown that the enhanced diffusion rate is unlikely to produce observable effects when loops are produced unless a very sensitive technique is used.