Nitinol, a binary alloy of Ni and Ti, is the most widely used shape memory alloy (SMA). It is used in medical technology as well as engineering applications such as thermally triggered micro-actuators. The shape memory effect relies on a reversible, stress- or thermally induced transformation from austenite to martensite. SMAs are typically produced as sheets or wires with a pronounced texture resulting from the manufacturing process. The texture strongly influences the deformation behavior, the mechanical properties, and the extent of the shape memory effect.

Electron beam powder bed fusion (PBF-EB) is a promising alternative to the established thermo-mechanical processing methods as it allows near-net-shape production with minimal impurity pick-up. However, the build process induces directional solidification, evaporation, and internal stresses, all of which may affect the functional properties. In this study, we investigate the effects of the beam power on the microstructure and the mechanical properties, as well as the influence of the build orientation on superelasticity and the shape memory effect of Ti-rich Nitinol. The material is processed on a freely programmable Freemelt One PBF-EB machine.

First, a processing window is rapidly established using only in situ electron optical imaging (ELO) of the build surfaces. The effect of the heat input is examined by systematically varying the area energy. Increasing energy input causes the formation of Ti2Ni precipitates due to the evaporation of Ni. These precipitates increase the stiffness and reduce the ductility.

Samples built standing, flat, and at 45° orientation were tested under tension and cyclic compression using a Gleeble 3500. The load direction relative to the build direction markedly influenced the mechanical strength. Furthermore, the percentage of recoverable strain under cyclic load was similarly affected, ranging from about 50 % for 45° relative orientation to 95 % for loading parallel to the build direction. Therefore, the strength of the shape memory effect shows a pronounced dependence on the texture. In the future, controlling the texture may enable the design of components with locally tailored shape memory properties.