The Effect of Inclusions on Fatigue Behavior of Nitinol and NiTi Shape Memory Alloys for Biomedical Applications

ASTM F2063-05 requires the evaluation of microcleanliness in NiTi mill products with section size between 6.3 and 94 mm. These dimensions exclude most of finished components for biomedical applications. Thus, the inclusion assessment is carried out at an intermediate manufacturing step. To our knowledge, no rigorous method has been developed to correlate the inclusion dimensions in the final product to those in semi-finished materials. Moreover, the method of inclusion analysis must be based on a robust statistical approach to accurately define inclusion content. Finally, specifications about microcleanliness should rely on studies explaining the correlation between inclusions size and material performance, in particular fatigue. Quite recently, ASTM E2283-08 for statistical analysis of nonmetallic inclusions in steels has been introduced. Aim of the practice is to estimate the expected largest inclusion in a determined volume of material. The need of a thorough method for the analysis of inclusions followed the studies reported by Murakami, which demonstrates the negative effect of inclusions on the fatigue behavior of hard steels and the need to adopt the extreme value inclusion rating.

In our opinion, a similar approach should be considered for SMAs. For this reason, a study on the effect of inclusions in Nitinol fatigue has been undertaken. Here we present the results concerning rotational bending fatigue on superelastic wires. A comprehensive experimental campaign has been carried out on materials coming from different melting processes. The fracture surfaces of failed wires have been analyzed and data about the presence of particles and their morphology have been recorded. The focus has been set on the particle dimensions and the cycles to fracture. For each level of alternate strain, cycles to failure have been plotted against inclusion dimensions in order to evaluate possible correlations. Finally the extreme value distribution of particles for the selected materials is identified.