As the performance demands on metal alloys of interest to the aerospace community continue to increase, the complexity of these alloys has tended to grow as well, with corresponding increases in the cost and time needed to develop future generations of materials. While the traditional empirical approach to designing these materials has proven effective, as the complexity of the materials grows, such methods will become less efficient, limiting prospects for further improvement. The incorporation of theoretically- and computationally-based input into the design process has the potential to both improve the quality of, and reduce the time to design, new materials.

Among the most fundamental issues in new materials design, the prediction of the crystallographic structure of solids has long been a goal of researchers in the field of computer simulation. To become an effective tool for assisting in the process of alloy design, a computational scheme must make use of an energy method fast enough to allow consideration of a large number of candidate structures, and an efficient energy minimization scheme.

In this work, we discuss our application of Monte Carlo methods, using the BFS semi-empirical energy method developed in our laboratory, to the problem of prediction of the crystallographic structure of multi-component metal alloys. Using a simple Metropolis Monte Carlo method in conjunction with an accurate and transferable method for computing energetics, we have investigated the crystallographic structure of a variety of fcc- and bcc-based intermetallic compounds, and have considered the influence of alloying additions in compounds having up to five constituents. We discuss the successes and limitations of the methodology, and the prospects for obtaining accurate prediction of physical properties from our results. We also emphasize the role of this type of methodology as part of a synergistic experimental-theoretical effort for alloy design.