First-Principles Calculations of Phase Equilibria
and Transformation Dynamics of Fe-Based Alloys
Tetsuo Mohri, Munekazu Ohno, and Ying Chen
(Submitted December 29, 2004; in revised form February 28, 2005)
Theoretical procedures of first-principles calculations of phase stability and phase equilibria are
summarized. The present scheme is shown to be able to reproduce the transition temperatures
with surprisingly high accuracy for Fe-Pd and Fe-Pt systems. The main emphasis of the present
report is placed on the extension of the first-principles calculation to transition dynamics cal-
culations. This is performed by combining the cluster variation method with the phase-field
method via a coarse graining operation. The time evolution process of antiphase boundaries
associated with L1
0
ordering for Fe-Pd system is demonstrated.
Keywords cluster expansion method, cluster variation method,
coarse graining, Fe-Pd system, first-principles cal-
culation, L1
0
ordered phase, ordering dynamics,
phase field method
1. Introduction
Recent progress in first-principles calculations of phase
diagrams has been remarkable, and for certain systems, the
transition temperatures are obtained with surprisingly high
accuracy.
[1-5]
Although several approaches to first-princi-
ples phase diagram calculations are available, main schemes
based on a combination of electronic structure total energy
calculations with statistical mechanics methods such as the
cluster variation method
[6]
or Monte Carlo simulations are
well established.
Remaining problems in the future direction of the first-
principles calculations of phase stability and phase equilib-
ria can be summarized as (a) extension to a multicomponent
systems, (b) application to a low symmetry structures, and
(c) separate introduction of chemical interaction and other
interactions, such as magnetic interactions. Extension to
multicomponent systems is a requirement for practical pur-
poses, but applying a brute force first-principles calculation
to a multicomponent system is a questionable approach. In
particular, in view of excellent tools of the CALPHAD-type
software packages supported by powerful databases, one
needs to focus on the well-tuned problems for which only
first-principles calculation can provide answers.
The second subject of the low-symmetry structure, in a
sense, shares the common background with the first subject.
Most of the first-principles calculations performed so far
have centered around a high-symmetry structure, such as a
cubic structure. In order for the first-principles calculation
to be a versatile tool for a wide class of alloy systems, it is
desirable that a calculation procedure be established at least
for tetragonal and hexagonal structures. Note that symmetry
does not necessarily refer to the Bravais lattice. The lowest
energy state of a given alloy system is realized by local
atomic displacement at each lattice point, which breaks the
global symmetry of the entire lattice. Hence, even if the
global lattice symmetry is characterized in terms of the Bra-
vais lattice, local structure is generally topologically disor-
dered and atomistic calculations need to incorporate such a
low symmetry in the calculations. Recent development
in the Continuous Displacement Cluster Variation Method
(CDCVM)
[7-12]
had made it a potential theoretical tool, and
further development is urged.
The settlement of the third subject is a sheer necessity in
connection with the current study on Fe-based alloy systems.
As has been reported in the authors’ recent articles,
[1-5]
magnetism plays an essential role in the phase stability of
Fe-Ni, Fe-Pd, and Fe-Pt alloy systems. Although spin-
polarized electronic structure calculations reveal the signifi-
cance of magnetism, the separation of the atomic interaction
energies into chemical and magnetic energy contributions
have not been attempted. It is plausible that a number of
unsettled subjects in the Fe-based phase diagram calcula-
tions originate from the neglect of the separate treatment of
the two contributions. Among such problems is the signifi-
cant deviation of the congruent composition of L1
0
ordered
phase found in the Fe-Pd phase diagram.
[13]
Despite various
attempts, the authors have not been able to reproduce the
shifting of the congruent composition from 1:1 stoichiom-
etry. The double-Ising type formulation is believed to be
necessary for introducing interaction energies of different
origins.
Apart from the settlement of these subjects in the phase
diagram calculations, the other direction in which the first-
This paper was presented at the International Symposium on User
Aspects of Phase Diagrams, Materials Solutions Conference and Ex-
position, Columbus, Ohio, 18-20 October, 2004.
Tetsuo Mohri, Division of Materials Science and Engineering, Gradu-
ate School of Engineering, Hokkaido University, Sapporo 060-8628,
Japan; Munekazu Ohno, Division of Materials Science and Engineer-
ing, Graduate School of Engineering, Hokkaido University, Sapporo
060-8628, Japan and Technical University-Clausthal, Robert-Koch
Str. 42, D-38678, Clausthal-Zellerfeld, Germany; and Ying Chen,
Department of Quantum Engineering and System Science, School of
Engineering, The University of Tokyo, Tokyo 113-8655, Japan. Con-
tact e-mail: tmohri@eng.hokudai.ac.jp.
JPEDAV (2006) 27:47-53
DOI: 10.1361/105497106X92790
1547-7037/$19.00 ©ASM International
Basic and Applied Research: Section I
Journal of Phase Equilibria and Diffusion Vol. 27 No. 1 2006 47