The mechanisms of atomic transport induced by ion irradiation generally fall into the categories of anisotropic or isotropic processes. Typical examples of these are recoil implantation and cascade mixing, respectively. Each possesses its own characteristic fluence dependence (linear and square-root, respectively), although both may occur concurrently. In this case, these processes can be distinguished by carefully examining the fluence dependence at low fluences (less that 1015cm-2). Parametric studies can produce added insight.
We have measured the interaction of these processes in the mixing of Ti/SiO2/Si, Cr/SiO2/Si and Ni/SiO2/Si multi-layers irradiated with Xe at fluences of 0.01 - 10x1015cm-2. The fluence dependence of net metal transport into the underlying layers was measured with different thicknesses of SiO2 and different sample temperatures during irradiation (-196 to 500C). There is a linear dependence at low fluences. The initial slope depends on the sample temperature. At high fluences, a square-root behavior predominates. For thin SiO2 layers (< 20 nm), the cross-over point depends on the SiO2 thickness. These results are readily interpreted in terms of competition between the flux of injected atoms and diffusion of the accumulating metal. Metal appears to be the dominant moving species in these systems. The initial linear dependence was not observed in samples without SiO2, which shows that the effect is not due to cascade overlap.
The detailed analysis allows us to speculate on the role of chemical reaction kinetics in controlling the outcome of intra-cascade processes. There is no significant correlation between the reactivity of the metal with SiO2 and the amount of mixing observed when the irradiations are performed at 25C or below. The contribution from primary recoils is quite pronounced since the gross mixing is small. A significant correlation exists between the mixing and the energy deposited through elastic collisions FD. Several models are examined in an attempt to describe the transport process in Ni/SiO2. It is likely that injection of Ni by secondary recoil implantation is primarily responsible for getting Ni into the SiO2. There were a few more objections against the other conceivable limiting cases. Secondary recoil injection is thought to scale with FD. Trends in the mixing rates indicate that the dominant mechanism for Ti and Cr could be the same as for Ni. Nevertheless, the structure of the mixed region is very different for the three metals. The product of chemical reactions were identified with TEM and XPS. The low temperature results suggest that kinetic constraints are responsible for suppressing the chemical reactions that one might anticipate.
The processes of atomic transport and phase formation clearly fail to be separable at higher temperatures. A positive correlation with chemical reactivity emerges at higher irradiation temperatures. This was not pursued beyond preliminary results reported for Ti or Cr. The temperature at which rapid mixing occurs is not much below that for spontaneous thermal reaction. Less Ni is retained in the SiO2 at high irradiation temperatures. Ni incorporated in the SiO2 by low temperature irradiation is not expelled during a consecutive high temperature irradiation. The Ni remains trapped within larger clusters during a sequential 500C irradiation.
