Master thesis : Process validation in additive manufacturing for custom bone implants: Strategy and Optimization

Abstract

Patient-specific bone implants produced via additive manufacturing offer a promising solution for reconstructing maxillofacial bone defects that exceed the body’s natural healing capacity. To ensure clinical reliability and be registered with the Food and Drug Administration, it is essential to validate the manufacturing process in accordance with regulatory standards. While additive manufacturing enables precise control over geometrical parameters, limited research at Cerhum has explored how key process parameters affect the dimensional accuracy and mechanical performance of the hydroxyapatite implants fabricated via vat photopolymerization.

This study investigated the effects of slurry condition, internal structure orientation, build position and furnace load on the properties of hydroxyapatite gyroid-structured samples, as well as the printability and dimensional accuracy of a worst-case sample. Three slurry conditions (new, reused, expired) were tested, along with different orientations (X,Y,Z), build positions (intraand intertile) and furnace loads (light, medium, maximum), in order to cover a representative range of manufacturing conditions. Dimensional and mechanical measurements followed ISO 13175-3 standards, complemented by analyses of phase purity, porosity, heavy metal content and carbon residues.

Results showed that all tested samples met phase purity and porosity requirements. Mean dimensions remained within Cerhum’s tolerance range and indentation stresses corresponded to physiological values for trabecular bone, providing preliminary validation of the manufacturing process. However, the experimental design limited the reliable conclusions regarding the individual influence of process parameters and their interactions, hindering the study’s objectives. Preliminary findings suggested an impact of intertile position on dimensional and mechanical properties. Orientation along the Z-axis yielded the best dimensional accuracy, while its effect on mechanical performance varied across slurries but tended to be detrimental. Slurry condition also proved influential, with the expired batch unexpectedly showing improved mechanical performance. Importantly, all tested implants nevertheless remained within acceptable dimensional and mechanical ranges, supporting the robustness of the process.

Further work should focus on revising the worst-case sample and refining the experimental design by reducing the number of parameters and levels to extract reliable and interpretable results.