Previous qualitative studies of ion beam mixing of Ni-Ti and Fe-Ti multilayers at room temperature have shown the Ni-Ti samples to mix considerably faster than the Fe-Ti, in apparent contrast with theory. Furthermore, the Fe-Ti mixing was strongly inhibited by previous charging of the sample with hydrogen, whereas only a small effect was seen for Ni-Ti. We have quantified the mixing and extended the study to four more systems (Al-Ti, Co-Ti, Cu-Ti, and Pd-Ti) and lower temperatures. This allows some important conclusions to be drawn. Predictions based on a thermal spike model underestimate the larger room temperature mixing rates (Cu-Ti, Ni-Ti, and Pd-Ti), apparently because of contributions from a temperature dependent mechanism such as radiation enhanced diffusion. The lower mixing rates (Fe-Ti, Co-Ti, and Ni-Ti at -80 K) are overestimated by a factor of 2-3.5, possibly because of hydrogen contamination of the as-deposited samples. For the Al-Ti sample, the experimental mixing rate was in good agreement with predictions. Except for the Cu-Ti sample, results were seen to vary with heat of solution, rather than heat of mixing, suggesting significant contributions from the lower temperature after-spike regime. Hydrogen charging was found to reduce the Fe-Ti mixing rate by a factor of 7 at room temperature, whereas the Co-Ti and Ni-Ti rates were only reduced by a factor of 2, and the mixing of the Pd-Ti was influenced very little. Near liquid nitrogen temperature the Ni-Ti mixing rate was more strongly reduced (by a factor of 3-4). Our results suggest that the original hydrogen contamination in as-deposited samples may also cause significant reduction of mixing rates in some materials.
