Vacancy Diffusion Across FeCrAl Alloy Composition Space for Accident-Tolerant Fuel Cladding

Technical Summary: Vacancy Diffusion Across FeCrAl Alloy Composition Space for Accident-Tolerant Fuel Cladding

Problem Statement
Iron-chromium-aluminium (FeCrAl) alloys are leading candidates for accident-tolerant fuel (ATF) cladding in light-water reactors (LWRs) due to their superior high-temperature oxidation resistance compared to Zircaloy. However, the in-reactor lifetime of these claddings is governed by radiation-induced microstructural evolution, where point defect transport—specifically vacancy diffusion—is the dominant mechanism. While the formation of protective oxide layers is well-characterized, the fundamental defect physics of FeCrAl under irradiation remain poorly understood. Existing theoretical treatments often rely on idealized or compositionally simplified environments, failing to account for the complex local chemical fluctuations inherent in concentrated solid-solution alloys. These local environments significantly alter vacancy binding, migration pathways, and collective behaviors such as clustering and trapping. There is a critical need to disentangle the individual contributions of constituent elements (Fe, Cr, Al) to overall vacancy transport to predict long-term material evolution, phase stability, and creep under combined thermal and irradiation loading.

Methodology
The authors employ a species-resolved kinetic Monte Carlo (KMC) framework to model vacancy diffusion across a systematic range of Fe/Cr compositions (from Fe-rich to Cr-rich) with a fixed 5 at.% Al concentration. The methodology proceeds through the following steps:

1. Potential and Data Generation: Interatomic interactions are evaluated using an Embedded Atom Method (EAM) potential fitted to a Density Functional Theory (DFT) database. Vacancy migration data is generated using the "Hop-Decorate" workflow, where a BCC Fe reference structure is randomly decorated with Fe, Cr, and Al atoms according to target global compositions. Minimum energy paths are computed using the Climbing-Image Nudged Elastic Band (CI-NEB) method for global compositions of Fe$_{80}$Cr$_{15}$Al$_5$, Fe$_{47.5}$Cr$_{47.5}$Al$_5$, and Fe$_{15}$Cr$_{80}$Al$_5$.
2. Surrogate Modeling: To avoid the computational cost of on-the-fly NEB calculations during KMC, linear surrogate models are trained on the generated barrier database. The local chemical environment for each hop is encoded using descriptors counting Fe, Cr, and Al atoms within fixed radial shells around the saddle point and the migrating atom. Separate linear regression models are fitted for each migrating species (Fe, Cr, Al) to capture species-specific barrier dependencies.
3. KMC Simulations: The trained surrogate models drive residence-time KMC simulations at temperatures ranging from 500 K to 1500 K. Simulations track single vacancy trajectories over $10^5$ steps to calculate mean squared displacement (MSD) and extract diffusion coefficients ($D$), activation energies ($E_a$), and pre-exponential factors ($D_0$).

Key Results

• Migration Barriers and Local Environment: The study confirms that migration barriers are highly sensitive to the local chemical environment. While pure Fe exhibits a barrier of ~0.63 eV, Cr is ~0.95 eV, and Al is ~0.69 eV. In alloys, the distributions of barriers broaden significantly, particularly for Cr and Al, indicating that local chemical asymmetry modulates migration pathways.
• Role of Aluminum: Despite its low concentration (5 at.%), Al exerts a disproportionate influence on migration barriers. In all three surrogate models, the presence of an Al atom at the first nearest-neighbor (1NN) saddle shell significantly raises the energy barrier (+0.048 eV for Fe, +0.073 eV for Cr, and +0.151 eV for Al).
• Compositional Dependence of Diffusivity:
  • Fe-Rich to Equiatomic: In compositions ranging from Fe$_{80}$Cr$_{15}$Al$_5$ to Fe$_{32.5}$Cr$_{62.5}$Al$_5$, vacancy diffusivity is relatively insensitive to global composition. Activation energies converge to a range of 0.70–0.84 eV, suggesting that Fe migration dominates transport in these regimes.
  • Cr-Rich Regime: The Cr-rich composition (Fe$_{15}$Cr$_{80}$Al$_5$) exhibits a markedly elevated activation energy of 1.135 eV. Consequently, its macroscopic diffusivity is orders of magnitude lower than Fe-rich alloys at lower temperatures (e.g., ~$10^{-17}$ m$^2$ s$^{-1}$ at ~570 K compared to ~$10^{-13}$ m$^2$ s$^{-1}$ for Fe-rich alloys).
• Percolation and "Sluggish" Diffusion: The results suggest a percolation threshold where the connected network of low-barrier paths (dominated by Fe hops) breaks down as Cr content increases. In the Cr-rich regime, vacancy migration is dominated by high-barrier Cr hops, leading to "sluggish diffusion." The authors observe that vacancy trajectories in chemically rough landscapes involve escape from trapping pockets, resulting in non-linear mean squared displacement behavior before reaching a diffusive regime.

Significance and Claims
The paper claims that the macroscopic diffusivity of FeCrAl alloys cannot be predicted by a simple composition-weighted average of pure end-member barriers; rather, it is set by the heterogeneity of the local energy landscape and the specific connectivity of low-barrier migration paths.

• Design Implications: The study identifies a distinct compositional regime (Cr-rich) where vacancy mobility is suppressed. This suppression implies that in Cr-rich $\alpha'$ phases (known to form under irradiation), vacancies are less mobile, potentially reducing the rate of vacancy-interstitial recombination and altering defect accumulation kinetics.
• Accident Tolerance: Under Loss-of-Coolant Accident (LOCA) conditions where temperatures rise above 1200 K, the compositional dependence of diffusivity becomes pronounced. The authors suggest that the Cr-rich composition, with its high activation energy and suppressed diffusivity, may offer favorable vacancy kinetics for accident-tolerant cladding, potentially complementing its known oxidation resistance.
• Methodological Contribution: The work demonstrates that reduced-order kinetic Monte Carlo frameworks, parameterized by species-resolved linear surrogate models trained on NEB databases, can effectively map defect transport across complex composition spaces. This approach provides a mechanistic link between atomistic processes and mesoscale models used in fuel performance codes, moving beyond empirical trial-and-error in alloy design.

Limitations Noted by Authors
The authors explicitly note several limitations: the model treats the lattice as static and frozen, ignoring local relaxation and radiation-induced compositional evolution (segregation) over time. The linear surrogate model cannot capture directional or symmetry-breaking effects, nor magnetic contributions to the energy landscape. Furthermore, the study focuses solely on vacancy transport and does not address interstitial cluster transport or the radiation-induced transformation of dislocation loops, which may also influence defect sink densities.