Abstract
In this paper, a hybrid three-stage methodology based on in vitro experiments, simulations, and metaheuristic optimization is presented to enhance the corrosion resistance of hydroxyapatite (HA)-coated magnesium implants in biomedical applications. In the first stage, we add cerium (Ce) to HA and present a new coating (named HA+Ce) to improve the resistance of the coating to corrosion. Then, various HA+Ce compounds with different factors (e.g., concentration, pH, immersion time, and temperature) are generated and their propensity for corrosion is examined in a physiological environment using EIS and DC polarization tests in a simulated body fluid solution. Eventually, a comprehensive dataset comprising 1024 HA+Ce coating samples is collected. In the second stage, machine learning using random forest (RF) is used to learn the relation between the input factors of the coating and its corrosion resistance. In the third stage, a metaheuristic algorithm based on the whale optimization algorithm (WOA) is utilized to find the best HA+Ce compound with the maximum corrosion resistance, while the objective function of WOA for a new unseen coating solution is estimated using the trained RF model. Finally, the morphology and composition of the best coating solution are inspected using FE-SEM. According to the obtained results, the HA+Ce coating with an immersion time of 60 min, concentrations of 0.9 for Ce and 1.2 for HA, pH of 4.1 for solution, and temperature of 70 °C demonstrated the highest level of corrosion resistance among all experiments and simulations. The final optimized HA+Ce coating solution has obtained a corrosion resistance of 14,050 Ω·cm2, which resulted in a gain of 14.9% compared to the HA-coated Mg implants.