Mechanisms of self and impurity diffusion in Fe-Al intermetallic compounds

Vacancies, impurities and precipitates have a significant influence on material properties. They can severely enhance or decrease hardness, ductility or yield stress to name just a few examples. Thus it is crucial both from a scientific point of view as well as for industrial applications to understand their formation and diffusion behavior. Kinetic Monte Carlo simulations are very well suited for this task. They however rely on accurate data for the diffusion barriers which can be provided by ab initio theories [1].

As an example we focus our diffusion studies on the iron aluminides which are a promising material class for industrial high temperature applications due to their excellent oxidation and sulphidation resistance. They are also lightweight and offer a good strength-to-weight ratio. Currently the widespread use of FeAl is mainly inhibited by its limited ductility at ambient temperatures which is thought to be related to defects in the material since FeAl shows a very high vacancy concentration (several percent at T>1000 K).

The diffusion mechanisms in ordered binary alloys are more complicated than in materials with only one atom species where diffusion is usually mediated by random nearest neighbor (NN) vacancy jumps. This simple mechanism cannot be the dominant one for FeAl, which has B2 structure, because it would lead to a breakdown of long-range order. Several more sophisticated mechanisms have been described in the literature (see e.g. [3], [4]). The simplest generalization of the NN jump is the next-nearest neighbor (NNN) jump of a vacancy, i.e. it remains on the same sublattice and therefore long-range order is maintained (see figure).

Another very promising candidate for the diffusion in FeAl is the so-called triple defect mechanism. It has been suggested by several authors to be dominant at intermediate temperatures. The corresponding migration path is also schematically shown in the figure. The migration barrier for these and other diffusion mechanisms will be calculated using the climbing image nudged elastic band method [5]. In this way, the energetically most favorable diffusion mechanism will be identified. The determined parameters will also be used by our Chinese partner to develop reliable potentials for large scale simulation of diffusion processes [2].

[3] R. Drautz and M. Fähnle, Acta Mater. 47, 2437 (1999)

[4] N. A. Stolwijk, M. van Gend and H. Bakker, Philos. Mag. A 42, 783 (1980)

[5] G. Henkelman, B. P. Uberuaga and H. Jónsson, J. Chem. Phys. 113, 9901 (2000)