Machine learning-guided accelerated discovery of structure-property correlations in lean magnesium alloys for biomedical applications

Abstract

Magnesium alloys are emerging as promising alternatives to traditional orthopedic implant materials thanks to their biodegradability, biocompatibility, and impressive mechanical characteristics. However, their rapid in-vivo degradation presents challenges, notably in upholding mechanical integrity over time. This study investigates the impact of high-temperature thermal processing on the mechanical and degradation attributes of a lean Mg-Zn-Ca-Mn alloy, ZX10. Utilizing rapid, cost-efficient characterization methods like X-ray diffraction and optical microscopy, we swiftly examine microstructural changes post-thermal treatment. Employing Pearson correlation coefficient analysis, we unveil the relationship between microstructural properties and critical targets (properties): hardness and corrosion resistance. Additionally, leveraging the least absolute shrinkage and selection operator (LASSO), we pinpoint the dominant microstructural factors among closely correlated variables. Our findings underscore the significant role of grain size refinement in strengthening and the predominance of the ternary Ca2Mg6Zn3 phase in corrosion behavior. This suggests that achieving an optimal blend of strength and corrosion resistance is attainable through fine grains and reduced concentration of ternary phases. This thorough investigation furnishes valuable insights into the intricate interplay of processing, structure, and properties in magnesium alloys, thereby advancing the development of superior biodegradable implant materials.

Keywords

• Magnesium alloys
• Machine learning
• Corrosion
• Mechanical properties
• Rapid characterization