PSI - Issue 82

Hang Su et al. / Procedia Structural Integrity 82 (2026) 131–137 H. Su et al. / Structural Integrity Procedia 00 (2026) 000–000

135

5

Fig. 2 presents representative SEM morphologies of the AM titanium alloy under R = –1, highlighting both the overall fracture surfaces and the crack initiation regions at different stress amplitudes. At the higher stress amplitude of σ ₐ = 500 MPa (Figs. 2a and 2c), the specimen failed after a relatively short lifetime of 1.09 × 10⁵ cycles. Crack initiation in this case occurred from a near-surface LoF defect, where stress concentration led to rapid crack nucleation and propagation through the surrounding acicular α′ martensitic structure. The area around the initiation site shows a mixed ductile–brittle transition, characterized by cleavage-like facets and shallow dimples, indicating limited plastic deformation before failure. The size and morphology of the defect, together with the inhomogeneity of microstructural distribution around it, contributes to the formation of such fracture surface morphology, as an obvious ridge going through the defect can be clearly identified in Fig. 2a. This highly probability-dependent feature indicates that LoF defect and gas pores themselves play an important role in fatigue damage development. In contrast, under a lower stress amplitude of σ ₐ = 250 MPa (Figs. 2b and 2d), the specimen survived up to 8.81 × 10⁷ cycles and exhibited an internal-origin fracture mode typical of VHCF behavior. The crack initiated from an internal pore or LoF-type defect in the center of the specimen surface. A typical FGA developed around the defect, followed by a distinct FiE region extending outward toward the final fracture zone, consistent with the internal crack evolution features observed in gradient-structured titanium alloys (Pan et al., 2020b). The FGA surface morphology suggests fatigue crack growth under near-threshold conditions, where cyclic plastic strain is highly localized and dominated by microstructural slip incompatibility (Pan et al., 2020a; Du et al., 2021; Qian et al., 2020). An interesting point is that the morphology of the defect (Fig. 2d) is more irregular compared to the one in Fig. 2c, which means higher stress concentration around the defect. However, the fracture morphology around the defect indicates a less defect-morphology-dependent trend compared with the previous one (Fig. 2c). Therefore, the size and morphology effect of defects on fatigue damage decreases along with increasing cycles from HCF regime to VHCF regime. Low stress levels lead to a more uniform development of microstructural damage regardless of the size and morphology of defects. In the CP equiaxed alloy, the micro- and nano-structural evolution near the initiation site follows a similar trend but originates from microstructural incompatibility rather than void defects. EBSD and TEM observations identify localized grain fragmentation and misorientation accumulation along α/β interfaces, producing a nanograin layer with distinct orientation gradients (Pan and Hong, 2019; Su et al., 2017; Pan et al., 2024b). The absence of large pores results in a smaller FGA size but a comparable level of nanoscale refinement. This indicates that the formation of nanograins is a universal response to long-term cyclic micro-plasticity, regardless of the initial microstructure or manufacturing route (Heinz and Eifler, 2016; Nikitin et al., 2016). Overall, the micro- and nano-structural evidence demonstrates a common fatigue-damage evolution mechanism for both AM and CP titanium alloys: crack initiation begins either at intrinsic defects (AM) or at microstructural heterogeneities (CP), followed by the development of an ultrafine-grained zone through cyclic plastic refinement. This process bridges the macroscopic S–N behavior and microscopic damage accumulation, providing insight into the transition from defect-dominated to microstructure-controlled fatigue mechanisms in the very-high-cycle regime. 4. Conclusion In summary, the comparative investigation of AM and CP titanium alloys under fully reversed loading reveals that the fatigue behavior transition from microstructure-controlled to defect-dominated mechanisms fundamentally defines their HCF and VHCF performance. The AM alloy, characterized by lamellar martensitic microstructure and inherent LoF or gas-pore defects, exhibits a duplex S–N response and internal-origin fractures with smooth FGA and FiE morphologies. In contrast, the CP equiaxed alloy presents a monotonic S–N curve with surface or α/β interface crack initiation and irregular FGAs, indicating microstructure-sensitive fatigue governed by slip incompatibility rather than porosity. Despite their different microstructural origins, both materials display nanoscale grain refinement within the crack-initiation zone, evidencing a universal mechanism of cyclic plastic accommodation in the VHCF regime. These findings highlight that controlling the population and morphology of void-type defects, together with understanding micro- and nano-scale fatigue damage evolution, is essential for improving the fatigue reliability of AM titanium alloys and for establishing a unified framework linking defect statistics, microstructure, and long-life performance.