The present study investigated the influence of pore size of strut-based Diamond and surface-based Gyroid structures for their suitability as medical implants. Samples were made additively from laser powder bed fusion process with a relative density of 0.3 and pore sizes ranging from 300 to 1300 μm. They were subsequently examined for their manufacturability and mechanical properties. In addition, non-Newtonian computational fluid dynamics and discrete phase models were conducted to assess pressure drop and cell seeding efficiency. The results showed that both Diamond and Gyroid had higher as-built densities with smaller pore sizes. However, Gyroid demonstrated better manufacturability as its relative density was closer to the as-designed one. In addition, based on mechanical testing, the elastic modulus was largely unaffected by pore size, but post-yielding behaviors differed, especially in Diamond. High mechanical sensitivity in Diamond could be explained partly by Finite Element simulations, which revealed stress localization in Diamond and more uniform stress distribution in Gyroid. Furthermore, we defined the product of the normalized specific surface, normalized pressure drop, and cell seeding efficiency as the indicator of an optimal pore size, in which this factor identified an optimal pore size of approximately 500 μm for both Diamond and Gyroid. Besides, based on such criterion, Gyroid exhibited greater applicability as bone scaffolds. In summary, this study provides comprehensive assessment of the effect of pore size and demonstrates the efficient estimation of an in-silico framework for evaluating lattice structures as medical implants, which could be applied to other lattice architectures.

In the field of medical engineering, Triply Periodic Minimal Surfaces (TPMS) structures have been studied widely owing to their physical attributes similar to those of human bones. Computational Fluid Dynamics (CFD) is often used to reveal the interaction between structural architectures and flow fields. Nevertheless, a comprehensive study on the effect of manufacturing defects and non-Newtonian behavior on the fluid responses in TPMS scaffolds is still lacking. Therefore, the present study fabricated Gyroid TPMS with four relative densities from 0.1 to 0.4. Non-destructive techniques were used to examine surface roughness and geometric deviation. We found that the manufacturing defects had a minor effect on fluid responses. The pressure drop comparison between defect-containing and defect-free models could be differed up to 7%. The same comparison for the average shear stress showed a difference up to 23%, in which greater deviation between both models was observed at higher relative density. On the contrary, the viscosity model played a significant role in flow prediction. By comparing the Newtonian model with Carreau-Yasuda non-Newtonian model, the resulting pressure drop and average wall shear stress from non-Newtonian viscosity could be higher than those of the Newtonian model by more than a factor of two. In addition, we matched the fluid-induced shear stress from both viscosity models with desirable ranges of shear stresses for tissue growth obtained from the literature. Up to 70% from the Newtonian model fell within the desirable range while the matching stress reduced to lower than 8% for the non-Newtonian results. Furthermore, by correlating geometric features with physical outputs, the geometric deviation was seen associated with surface curvature while the local shear stress revealed a strong correlation with inclination angle. Overall, the present work emphasized the importance of the viscosity model for CFD analysis of the scaffolds, especially when resulting fluid-induced wall shear stress is of interest. In addition, the geometric correlation has introduced the alternative consideration of structural architectures from local perspectives, which could assist the further comparison and optimization among different porous scaffolds in the future.