Abstract
Laser powder bed fusion (LPBF) enables simultaneous control of epitaxial grain growth and solidification microstructure because it combines directional solidification with a programmable, scan-strategy-guided thermal field. In principle, epitaxial inheritance from a single-crystal (SX) substrate or previously deposited layers can be used to achieve stable crystallographic orientation and controlled cellular/dendritic growth within melt pools. However, in practice, epitaxy in LPBF often persists for only a limited number of layers before epitaxial stability is lost and new grains with different orientations emerge, disrupting both crystallographic continuity and the associated solidification microstructure. This instability limits reliable texture control and can introduce spatially varying deformation behavior because crystallographic orientation governs direction-dependent mechanical, thermal, and electrical response, while cellular solidification microstructures strongly influence dislocation motion and strengthening. The consequences are especially significant for applications where grain boundaries are detrimental, such as SX superalloy turbine blades for high-temperature creep resistance. Because these components are costly and time-intensive to manufacture, laser-based additive repair is an attractive pathway for the manufacturing and repair of SX-like microstructures while significantly reducing material waste, replacement cost, and downtime. Therefore, to achieve predictable epitaxial stability and better control of properties, it is necessary to understand when and why crystallographic epitaxial growth becomes unstable during processing, and how this instability is correlated with the solidification microstructure.
This dissertation advances the fundamental understanding of crystallographic instability in LPBF by identifying how process-induced misorientation forms, accumulates, and ultimately contributes to epitaxy loss and new grain formation. Using designed experiments on SX Stainless Steel 316L substrates, combined with quantitative orientation mapping, defect analysis, and computational modeling, this dissertation systematically reveals three central findings. First, epitaxial growth dominates microstructure development in LPBF, and preferential growth-variant selection is governed primarily by the local temperature-gradient magnitude at the solid--liquid interface. Second, crystallographic misorientation within and around a melt pool arises from two sources: (i) solidification-morphology-dependent misorientation that localizes at cell walls and colony boundaries, and (ii) morphology-independent misorientation generated by internal-stress-induced plastic deformation. This plasticity-driven contribution further extends the misorientation approximately 100 μm beyond the melt pool boundary into the substrate. Building on this melt-pool-level foundation, this dissertation shows that repeated remelting progressively accumulates twist-like crystal rotation, increases geometrically necessary dislocation (GND) density, and promotes subgrain-boundary development, ultimately reaching a point at which a subgrain boundary becomes energetically more favorable for growth than the deformed parent grain, resulting in epitaxy loss and the formation of new grains with distinct crystallographic orientations and solidification growth directions. Collectively, these findings provide a fundamental basis for designing and stabilizing crystallographic texture and solidification microstructure in LPBF, thereby enabling precise texture programming and single-crystal repair strategies.
