ISSN:
1089-7550
Source:
AIP Digital Archive
Topics:
Physics
Notes:
The large, growth-induced magnetic anisotropy in amorphous rare earth-transition metal alloys such as Tb-Fe are shown to depend strongly on the deposition temperature and only weakly on deposition rate or deposition technique (e.g., sputtering versus electron beam co-evaporation). These dependencies can be well fit with a thermally activated form involving minimization of surface energy during the growth by a re-orienting of adatom configurations over potential energy barriers. In this model, the growing film lowers its surface energy by a partial alignment of local clusters, presumably such as to maximize the number of in-plane bonds, although chemical effects undoubtedly also play an important role. These effects are somewhat analogous to a surface reconstruction which becomes trapped into the growing film by low bulk diffusion rates. In particular, a two-level model with a flat distribution of energy barriers is here shown to provide an excellent fit to the observations. Such a model leads to a ln(t) dependence on deposition rate and an exponential dependence on deposition temperature. We have also studied the subsequent irreversible relaxation of the anisotropy upon annealing. This relaxation is strongly influenced by the original growth temperature. In particular, the higher the original growth temperature, the more resistant the film is to subsequent relaxation. This result has important technological implications. As is commonly observed, the relaxation is well fit by a two-level model, again with a flat distribution of energy barriers over a range of energies, producing a ln(t) dependence on annealing time and a thermally activated dependence on annealing temperature. In annealing, of course, the lower energy state is isotropic, unlike the surface-induced anisotropic state produced during growth. The influence of the growth temperature on this relaxation implies that the actual process of creating the anisotropic state during the growth has the consequence of eliminating free volume in the sample, thereby raising the energy barriers to subsequent relaxation.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1063/1.356812
