In situ dissolution–reprecipitation of TiC in SLM-fabricated functionally graded 316L/TiC composites: microstructural evidence and strengthening mechanisms

• In-situ dissolution–reprecipitation of TiC revealed in SLM 316L/TiC composites. • Particle-size-driven transition from columnar to equiaxed grain morphology. • Dual TiC role: heterogeneous nucleation and load transfer strengthening. • Marangoni convection controls secondary TiC particle segregation patterns. • Highest strength in fine-TiC layer via grain refinement and nanoparticle pinning. This study provides the first direct multimodal evidence of dissolution and reprecipitation of TiC during selective laser melting (SLM) of 316L/TiC composites. Functionally graded samples were fabricated with a three-layer architecture: pure 316L SS (L1), 316L + 10 wt% fine TiC (L2), and 316L + 10 wt% coarse TiC (L3). Defect-free samples thereby enabled an isolated study of the particle-size effects on solidification, phase evolution, and strengthening mechanisms. EBSD revealed a transition from coarse columnar grains in L1 to fully equiaxed grains in L2, driven by TiC-induced heterogeneous nucleation and Zener pinning. High-resolution SEM, XRD, and EDS confirmed two distinct populations of secondary TiC: fragmentation-derived intragranular particles (∼100–300 nm) and nanoscale intergranular precipitates formed via dissolution–reprecipitation. Fine TiC reinforcement yielded the most refined microstructure, with the highest high-angle grain boundary fraction (96.6 %). Fine-TiC composites achieved the highest yield strength (847 ± 18 MPa) and ultimate tensile strength (1042 ± 10 MPa), representing ∼ 90 % and ∼ 62 % improvements over pure 316L, respectively, with reduced ductility. Strengthening arose from grain refinement (Hall–Petch), Orowan looping, and load transfer. These results clarify the particle-size-dependent mechanisms governing microstructure–property relationships in SLM-fabricated metal-matrix-composites (MMCs) and offer guidelines for reinforcement engineering.