Resumen:

Biodegradable magnesium (Mg) alloys represent a promising class of materials for next-generation biomedical implants, particularly endoluminal devices. Their major limitation, however, lies in the difficulty of controlling corrosion in the aggressive physiological environment, where chloride-rich solutions promote rapid dissolution and localized damage. While extensive in vitro and in vivo experiments have been carried out to characterize Mg degradation, predictive computational models are still required to unravel the underlying mechanisms and provide design guidelines for reliable device performance.
This work presents a comprehensive phase-field framework to simulate corrosion processes in Mg alloys under physiologically relevant conditions. In this approach, the sharp interface between the metallic core and the electrolyte is replaced by a smooth transition zone described by an auxiliary variable, the phase-field parameter φ. This formulation avoids explicit front-tracking and naturally captures complex morphological evolutions, including interface merging, pit nucleation, and irregular dissolution fronts. Furthermore, the framework can be seamlessly extended to incorporate mechanical fields, enabling the study of chemo–mechanical interactions during degradation.
From the numerical perspective, the governing equations are implemented in the finite element library FEniCSx, exploiting its Unified Form Language (UFL) for variational formulations. One-dimensional and two-dimensional problems are solved monolithically, ensuring robustness at moderate computational cost. For three-dimensional and fully coupled chemo–mechanical cases, a staggered algorithm is introduced, alternating between the mechanical and chemical sub-problems until convergence. Adaptive time-stepping strategies and MPI-based domain decomposition allow the simulation of long degradation times in large-scale geometries.
To capture localized corrosion phenomena such as pitting, spatial heterogeneity is introduced in the interfacial mobility coefficient via stochastic fields constructed from trigonometric mode superpositions. This enables a seamless transition from uniform to localized degradation within a unified formulation.
Validation is performed against in vitro immersion experiments, including hydrogen release measurements, mass loss quantification, and pH monitoring under both unbuffered and CO₂-buffered NaCl solutions. The model successfully reproduces the strong dependence of corrosion kinetics on environmental conditions, highlighting the critical role of interfacial mobility calibration. Importantly, the full 3D simulations demonstrate the capability of the framework to simultaneously predict localized corrosion morphologies and the mechanical response of the degrading material, thereby providing direct insight into the loss of structural integrity under service-relevant conditions.
Overall, the proposed phase-field model provides a versatile computational tool that bridges experimental observations with device design requirements. By combining physics-based modeling with advanced numerical strategies, it enables virtual testing of alloy compositions, device geometries, and operating environments, thereby supporting the development of safe and reliable biodegradable implants.