Electronic Structure of Alloys and Surfaces
R. E. Watson, M. Weinert, and M. Alatalo, Brookhaven National Laboratory
Research Objectives
To understand and predict the relative stability of multicomponent metallic alloys.
Computational Approach
The fundamental quantity determining the relative stability of different phases at zero temperature is the heat of formation. The heats for observed and unobserved competing alloy phases are obtained from first-principles local density calculations employing the full-potential linearized augmented Slater-type orbital method. To get a consistent set of heats of formation, all the structural parameters (volume, c/a ratios, internal atomic positions, etc.) must be optimized. The self-consistent solution of the set of effective one-particle equations are solved by a basis of single-particle wave functions expanded in a set of Slater-type orbitals in the interstitial, augmented by numerical solutions of the effective potential in spheres surrounding each atom. The large numbers of calculations needed to adequately investigate the competing alloy compositions, different crystal structures, and structural parameters were done on the cluster of Cray J90s at NERSC.
Accomplishments
The heats of formation for a large number of binary and ternary transition-metal and aluminum alloys (at various compositions and in a diverse set of crystal structures characteristic of the observed phases) were calculated. The calculated competition among ordered Ni-Ti-Al alloys is shown in the accompanying ternary phase diagram. The triangles connecting different stable phases represent three-phase regions; relatively small changes in the heats of the binaries and/or ternary heats can result in different connections among the phases or even cause phases to disappear. Of the three predicted stable ternary phases inside the triangle, only the NiTiAl phase has not been reported, possibly losing out to as-yet-unconsidered ordered or disordered phases, several of which are currently under investigation.
Experimental heats are available for only Ni2TiAl and some of the ordered binaries around the perimeter of the triangle. The agreement between experiment and theory is typically a few hundredths of an eV/atom, the same order as the differences among reported experimental values. Similar agreement is obtained for the aluminides of V and of the 4d elements Y to Pd. For nonmagnetic iron alloys, the discrepancy with experiment is an order of magnitude greater, suggesting that the local spin-density approximation underestimates the on-site magnetic energy contribution of bulk Fe to the heat of formation by a factor of 2, and that gradient corrections reduce, but do not remove, this error.
Calculated phase boundaries for ternary Ni-Ti-Al alloys.
Significance
The relative phase stabilities of multicomponent alloys at different compositions and in a diverse set of crystal structures were predicted from first-principles determined heats of formation. Transition-metal aluminum alloys are an important class of technologically important alloys that include super alloys, light-weight structural alloys, and phases displaying shape-memory effects. As a set, these calculated heats can be used to model the phases and precipitates that might form during material preparation, and thus provide guidance for the future development of new materials.
Publications
R. E. Watson and M. Weinert, "Transition-metal aluminide formation: Ti, V, Fe, and Ni aluminides," Phys. Rev. B 58, 5981 (1998).
R. E. Watson, M. Weinert, and M. Alatalo, "Ternary transition metal aluminide alloy formation: The BiF3 structure," Phys. Rev. B 57, 12134 (1998).
M. Weinert and R. E. Watson, "Hybridization-induced band gaps in transition-metal aluminides," Phys. Rev. B 58 (1998).