Decoupling of atomic interactions for accurate thermodynamic prediction in FCC alloys

Abstract

Accurate prediction of thermodynamic properties and phase equilibria in multicomponent Ni-based superalloys and high/medium-entropy alloys (HEAs/MEAs) poses persistent challenges due to complex atomic interactions and data scarcity. Here we present a simple yet powerful solution: a CALPHAD framework that bypasses computational and experimental bottlenecks by strategically decoupling nearest-neighbor (NN) and long-range (LR) interactions in face-centered cubic (FCC) alloys. The core innovation lies in a four-sublattice compound energy formalism (4SL-CEF) that embeds strong NN interactions into a physics-based “reference surface” derived from computationally efficient quasi-harmonic approximation (QHA) calculations, while confining excess terms to weak LR interactions—constrained to narrow, physically reasonable ranges, serving solely to refine phase equilibria. This divide-and-conquer strategy achieves both rigor and efficiency: for a 13-component Ni-based superalloy, only 1820 cost-effective QHA calculations (replacing thousands of empirical fittings) resolve NN interactions, while approximations (fitting/truncation/extrapolation) are applied exclusively to weak LR terms. Validated against the Ni–Co–Al ternary system, the model achieves good agreement with experimental thermodynamic data and phase equilibria, outperforming traditional CALPHAD methods. Furthermore, the framework enables seamless integration with atomic-scale simulations, revealing hidden mechanisms in complex alloys. For Cr–Co–Ni MEAs, we uncover a metastable L1₀ superstructure formed via an order–disorder phase transformation, resolving ambiguities in “diffuse scattering” and thermodynamic anomalies. This discovery challenges the prevailing chemical short-range order (CSRO) interpretation and directly links abrupt heat capacity changes to phase transformation. Our work demonstrates how decoupling interactions and minimizing computational efforts can unravel the complexities of multicomponent alloy modeling, offering a scalable, physics-based tool for accelerating the design of superalloys and HEAs/MEAs.