Functional fatigue and one-way shape memory effect in nanocrystalline TiNi shape memory alloys by gradient grains

Functional fatigue during cyclic loading limits the practical application of TiNi shape memory alloys (SMAs), as simultaneously achieving high strength, low residual strain, and high damping capacity is difficult. In this work, molecular dynamics simulation was performed to systematically compare the cyclic one-way shape memory effect (OWSME) and functional fatigue behavior of homogeneous nanograined (HNG) and gradient nanograined (GNG) TiNi SMAs under thermomechanical cycling. The HNG models show clear grain-size dependence. Grain refinement increases peak stress and reduces residual strain, but simultaneously suppresses martensitic transformation and hysteresis, which reveals an intrinsic trade-off between cyclic stability and energy dissipation. In contrast, the GNG structure maintains a large and stable hysteresis loop while retaining a high stress level and a low residual-strain level comparable to those of the HNG counterpart with the same average grain size. This superior performance originates from gradient-layer synergy: coarse-grained layers facilitate martensitic transformation and energy dissipation, whereas the fine-grained layers suppress dislocation accumulation and irreversible plastic deformation. As a result, the GNG architecture effectively slows functional degradation, limits residual-strain accumulation, and retains high damping capacity. These results demonstrate that grain-size gradients provide an effective microstructural route for simultaneously improving OWSME stability and functional fatigue performance in TiNi SMAs.