Corrosion Mechanisms of Magnesium for Biomedical Applications: Diffusion-Limited and Microstructure-Dependent Uniform Resorption

The resorption rate of biodegradable magnesium (Mg) implants in the body is challenging to predict using current in vitro methodologies. This difficulty arises because the corrosion product layer (CPL) formed on Mg exhibits different diffusion barrier properties under in vivo and in vitro environments. This Ph.D. thesis shows that an in vivo-relevant CPL only forms when the key physiological boundary conditions are simulated together in vitro: solution chemistry, tissue-level flow, and dynamic pH buffering.

Under physiological ionic concentrations, the initially formed CPL consists of mixed Mg- and Ca-phosphate precipitates. This CPL can equalize the corrosion rates of low-purity (LP) and ultra-high-purity (XHP) Mg. Using high-resolution cross-sectional electron microscopy and electrochemical impedance spectroscopy, the role of individual corrosion products in this equalization was elucidated. Nanoporous Mg-phosphates suppressed the cathodic reactivity of Fe-rich secondary phases in LP Mg. Calcium incorporation promoted a more compact nanostructure that limited Mg anodic oxidation but reactivated cathodic reactions on Fe-rich secondary phases. Model experiments with cathodic polarization on pure iron covered by different Mg-containing precipitated corrosion product layers explained this mechanism. Mg-phosphate hindered the hydrogen evolution reaction (HER), whereas Ca-phosphate promoted HER. Deviations from physiological inorganic ion concentrations can therefore shift precipitation kinetics and change the rate-limiting step.

A novel in vitro methodology was developed to maintain physiological precipitation kinetics. A new formulation of simulated interstitial body fluid was delivered to Mg under tissue-level flow (3.5 um/s). This approach was validated using high-purity (HP) Mg by comparing volumetric hydrogen gas (H2) evolution in vitro with in vivo data. With the developed approach, H2 formation was reduced (about 0.06 ml/(cm2·24 h)) towards the in vivo range (0.002 to 0.04 ml/(cm2·24 h)). Conventional stagnant immersion methodologies generated structurally defective CPLs and overestimated H2 evolution by a factor of 40 to 80. Applying this methodology, the CPL formed on LP-Mg effectively limited mass transport toward Fe-rich secondary phases, reducing the corrosion rate to the same range as XHP-Mg.

Finally, the body's response to foreign implant materials was studied by simulating inflammation-related pH gradients on Mg. In vitro pH produced distinct cross-sectional CPL morphologies that were matched to the explanted HP-Mg samples. Experiments at pH 6.2 showed good agreement with the CPL formed in vivo, validating that the subcutaneous implantation environment and its inflammation-related response to Mg can be mimicked in vitro. Exposure to pH 6.2 generated a defective CPL structure that persisted after returning to pH 7.4 and maintained elevated corrosion rates. In contrast, the less permeable CPL formed at pH 7.4 sustained electrochemica