Predicting yield-stress anomalies in L1 2 alloys: Ni 3Ge–Fe 3Ge pseudo-binaries

Abstract

The L1 2-based pseudo-binary (Ni 1 − c Fe c ) 3Ge is an ideal system to study yield-strength anomaly and its origin as it has a solid-solution phase vs. c and Ni 3Ge exhibits an anomaly while Fe 3Ge does not. Using two ab initio electronic-structure techniques, we calculate the planar-fault energies on the γ-surface, i.e., antiphase boundaries (APB) and stacking faults, both complex and superlattice intrinsic (SISF), for (Ni 1 − c Fe c ) 3Ge as a function of c. Generally, we use the fault energies combined with elasticity theory to predict occurence/loss of the yield-strength anomaly and show that the loss of anomaly occurs due to APB(1 1 1)-to-SISF(1 1 1) instability. Assessing the stability of APB(1 1 1) on the γ-surface within linear elasticity theory, we predict the transition from anomalous to normal temperature dependence of yield strength for c ≳ 0.35 (or 26 at.% Fe), as is observed, after which type-II, rather than type-I, dissociation is energetically favorable. Hence, first-principles calculations can predict reliably the existence/loss of anomalous yield-strength. Finally, we show that (0 0 1) and (1 1 1) APB energies of the binaries and pseudo-binaries agree quantitatively with measured values when chemical antisite disorder, intrinsic to the samples characterized, is included, whereas they are too large by a factor of two in perfect L1 2. We investigate three types of disorder: thermal and off-stoichiometric antisites, as well as chemical disorder vs. Fe-content in pseudo-binaries.