Intrinsic uncertainty on ab initio phase diagram and compound formation energy calculations: BCC Mo–Fe as a test case

Abstract

Ab initio electronic structure calculations, within the Kohn–Sham scheme of the density functional theory, are often considered reliable and a powerful tool to provide ground state information on intermetallic compounds. More recently, it has been used in multi‐scale intermetallic phase diagram calculations with particular importance when experimental data is lacking. In spite of this recent success, they rely on basic choices of the computational solution of the complicated quantum mechanical problem. Therefore, the calculated phase diagrams may depend on these choices. Here, we concentrate on the influence of some methodological aspects of the ab initio calculations on the resulting phase diagram, for a given statistical mechanics approach. We use the Full Potential‐Linear Augmented Plane Wave and the Projector Augmented Wave Methods to perform electronic structure calculations combined with the cluster variation method in the irregular tetrahedron approximation to calculate the BCC Mo–Fe phase diagram. It is shown that all calculated phase diagrams present similar qualitative features (an asymmetric miscibility gap), but quantitative variations are found depending on some of the basic assumptions adopted for the electronic structure calculations. Based on these results, a “natural” accuracy of the order of ± 1 kJ mol−1 can be estimated for ab initio compound formation energies.

Evaluation of the tantalum-titanium phase diagram from ab-initio calculations

“…In this study we used ab-initio calculations to investigate the Ta-Ti phase diagram and provide better estimates of the solvus location and structure. Similar methodologies of phase diagram prediction using ab-initio calculations have not been applied yet to the Ta-Ti system, but were used previously for various binary alloys and produced good agreement between the calculations and experimental observations [56,57,58,59]. In this research we utilize all-electron and pseudopotential DFT calculations to compute the ground state enthalpy for various ordered structures and for the SQS that model, employing the generic scaling factor corrections known for the hcp and bcc phases [46,47].…”

Ordering phase relationships in ternary iron aluminides

“…All calculations were performed with the non-relativistic code, for compatibility with previously published results. A recent work by some of the present authors [15] suggests that the particular settings of the ab initio calculation (like the introduction of relativistic corrections) do have an impact on the obtained formation energies, but that phase diagram topology, in particular, is not affected.…”

“…The standard procedure used here to achieve self consistency was already described in Ref. [15]. Exchange and correlation effects were treated with the generalized gradient approximation (GGA), using the implementations proposed by Perdew et al [16].…”

Thermodynamic modelling of the Nb–Ni–Si phase diagram based on the 1073 K isothermal section using ab initio calculations

“…The self-consistent processes have been performed until the differences of total energies per unit cell were smaller than 10 Ry 5 − (1 Ry ¼13.606 eV; 1 eV ¼96.485 kJ/mol) in two subsequent calculations. The standard procedure used here to achieve self-consistency was already described in a previous work by some of the authors [45]. Exchange and correlation effects were treated using the Generalised Gradient Approximation (GGA), with the implementation proposed by Perdew et al (PBE) [46].…”