Ion-beam mixing of thin marker layers in amorphous silicon and germanium was studied using irradiations with Xe ions at temperatures of 34K and 77K. The marker species, ion energies and doses were: in silicon, markers of Ge and Pt irradiated with 200-keV Xe up to 2.7xiol 6 ions cm"^; and in germanium, markers of Al and Si bombarded with 295-keV Xe up to 1.63xlO 16 ions cm" 2 . In silicon, Pt markers were found to broaden at about the same rate at 34K and 77K; and the rate of broadening was similar to that found by other workers when expressed as an efficiency of mixing, i.e., when dependence on ion dose and deposited energy was factored out. However, a Ge marker irradiated at 34K did not broaden from its original thickness. In germanium, markers of both Al and Si were mixed by irradiation at 34K, but at 77K only the Al marker broadened; the Si marker did not. His calculations are based on the theory of random flights, using power law cross sections to compute the recoil ranges, while the energy spectrum of moving atoms follows the usual low energy 1/E 2 law. However Sigmund and Gras Marti's calculation (9) of an ostensibly similar quantity (the contribution of low-energy recoils: the so-called "cascade mixing") yields a rate of increase of the variance of a thin marker which is about an order-of-magnitude smaller than experimentally measured values.When the contribution of rare high energy recoils is included bySigmund and Gras-Marti, the rate of increase of variance becomes much larger, but arises mainly from an extended tail formed on the marker distribution. This tail is not included in most experimental measurements of the variance, which are typically obtained by fitting a Gaussian to the marker profile.The identification of the mechanism of broadening of marker layers as ballistic mixing mainly rests on the absence of any temperature dependence of the diffusion coefficient when irradiations are performed at sufficiently low temperature (2,10), and on the linear dependence on damage rate already mentioned. An additional contribution of defect-mediated diffusion has been noted at higher temperatures, generally near or above room temperature in silicon (2,3). where it occurs, defect-mediated motion can yield diffusion coefficients larger than than those due to ballistic mixing, and allows the formation of crystalline phases which would be absent if ballistic diffusion were the only operative transport mechanism.Unfortunately, observation of a temperature-independent diffusion coefficient is not sufficient to guarantee that ballistic mixing is the dominant mechanism, because under certain conditions diffusion by interstitialcy mechanisms can , in principle, also be temperature independent t 11 ). It is therefore quite difficult to rule out a point defect contribution, except by irradiating at such low temperatures that inters itials are immobile.In the present study ion beam mixing of thin marker layers in Si and Ge was investigated using BBS. By comparing mixing rates of different solutes at 34K and 77K it wa...