Irradiation-Driven Recrystallization in Fusion-Grade Tungsten: A Mesoscale, Microstructure-Aware Model

Tungsten (W) is the leading candidate material for plasma-facing components in fusion reactors, yet its upper operational temperature is limited by premature grain growth and recrystallization processes. Irradiation adds further complications by generating defect clusters and transmutation products that alter both the driving forces and kinetics of grain boundary motion. In this work, we develop a physics-based, multiscale framework that couples crystal plasticity, stochastic cluster dynamics, and discrete grain boundary dynamics to model the co-evolution of plastic deformation, irradiation damage, and grain growth in fusion-grade tungsten polycrystals. The approach enables simulations on realistic microstructures with arbitrary grain size and misorientation distributions, without recourse to mean-field simplifications. The model captures (i) the spatial heterogeneity of dislocation density distribution during hot working; (ii) irradiation-induced defect accumulation under fusion conditions, and (iii) the buildup of chemical and elastic driving forces for grain boundary migration and microstructural evolution. Parametric studies demonstrate the dominant influence that temperature has on thermally activated grain-boundary mobility, a weaker dependence on prior strain, and a pronounced retardation of recrystallization by rhenium segregation arising from neutron transmutation. Under fusion energy irradiation conditions, our framework predicts a substantial reduction of the effective recrystallization temperature relative to unirradiated microstructures, while Re production restores and even elevates this limit. By providing quantitative projections of recrystallization kinetics and in-service recrystallization temperatures, this work establishes a predictive tool for assessing the lifetime and operational envelope of W-based plasma-facing materials under fusion conditions.

💡 Research Summary

**
This paper presents a physics‑based, multiscale computational framework that simultaneously captures crystal plasticity, stochastic radiation‑induced defect evolution, and grain‑boundary migration in fusion‑grade tungsten polycrystals. The authors integrate three distinct modules: (1) a finite‑element crystal plasticity (CPFE) model that resolves slip on multiple crystallographic systems, incorporates temperature‑dependent lattice friction, forest hardening, grain‑size‑limited hardening, and a dispersed‑barrier hardening term representing interactions with radiation‑generated defect clusters; (2) a stochastic cluster dynamics (SCD) module that solves master equations for point defects, self‑interstitial clusters, vacancy clusters, and transmutation products (notably rhenium) under prescribed neutron fluxes, temperatures, and local dislocation densities; and (3) a two‑dimensional vertex dynamics (VD) model that discretizes the grain‑boundary network into vertices and triple junctions, and drives boundary motion through a mobility tensor multiplied by a vector of driving forces, including chemical potential differences (grain‑boundary energy, solute segregation) and shear stresses acting in the boundary plane.

The coupling is fully bidirectional. The CPFE simulation of hot working generates spatially heterogeneous dislocation density fields, which feed into the SCD module to modify defect production and recombination rates (through a strain‑enhanced recombination term). The SCD output—defect concentrations and cluster size distributions—updates the irradiation‑hardening term τ_irr in the CPFE constitutive law, thereby altering slip resistance locally. Simultaneously, the stored energy from dislocations and radiation defects, together with solute segregation gradients (especially Re produced by neutron transmutation), constitute the chemical and elastic driving forces supplied to the VD module. The VD module then evolves the grain‑boundary network, creating new recrystallized grains and moving existing boundaries. Updated grain orientations and boundary geometries are fed back into the CPFE model, closing the loop.

Key findings from parametric studies are: (i) temperature dominates grain‑boundary mobility; a modest increase from 1100 K to 1300 K raises the mobility by orders of magnitude, leading to rapid recrystallization with Avrami exponent ≈1, indicative of simultaneous nucleation and growth under irradiation assistance. (ii) Prior plastic strain (i.e., the initial dislocation density) modestly accelerates nucleation but its effect is secondary to temperature. (iii) Neutron irradiation dramatically lowers the effective recrystallization temperature by up to 200 °C because point‑defect accumulation raises the stored energy that drives boundary migration. (iv) Rhenium segregation, however, strongly retards boundary motion: the mobility tensor is reduced proportionally to the local Re concentration, and the chemical driving force is diminished by Zener pinning. Consequently, Re production can restore or even raise the recrystallization temperature by 100–150 °C, counteracting the irradiation‑induced softening.

The model reproduces experimental observations of rapid, irradiation‑assisted recrystallization with Avrami exponents near unity and captures the experimentally reported increase in recrystallization temperature for W‑Re alloys. Unlike earlier mean‑field approaches that assume fixed grain‑size distributions, fixed misorientation, and constant nucleation rates, this framework resolves spatial heterogeneity in dislocation density, defect clusters, and solute segregation, allowing realistic prediction of microstructural evolution under complex service conditions.

Limitations include the use of a 2‑D vertex representation, which neglects out‑of‑plane grain‑boundary curvature and triple‑line dynamics present in three dimensions. Extending the VD module to full 3‑D topology would improve quantitative accuracy. Additionally, several material parameters (e.g., the dispersed‑barrier hardening coefficients, recovery parameters in the Kocks‑Mecking law, and the mobility prefactor) require calibration against dedicated high‑temperature, high‑dose experiments.

Practically, the framework offers a predictive tool for designers of plasma‑facing components: by inputting operating temperature, neutron flux, pre‑strain history, and alloy composition (e.g., Re content or other Zener‑pinning particles), the model can forecast the time‑dependent grain‑size distribution, recrystallization onset, and effective ductile‑to‑brittle transition temperature. This enables optimization of processing routes (hot working schedules, alloying strategies) to maximize the lifetime of tungsten‑based armor under the extreme conditions of future fusion reactors.

In summary, the authors have successfully combined crystal plasticity, stochastic radiation damage kinetics, and grain‑boundary vertex dynamics into a coherent multiscale model that captures the coupled effects of temperature, deformation, irradiation, and alloying on recrystallization kinetics in tungsten. The dominant role of temperature, the secondary influence of prior strain, and the pronounced retardation of recrystallization by Re segregation constitute the central insights, providing a solid foundation for future alloy design and lifetime assessment of fusion‑grade tungsten components.