An Advanced Modeling Framework with WellDesigned Experiments to Unveil Molten Pool Dynamics and Particle Spattering Behavior

Abstract

Laser powder bed fusion (LPBF) has emerged as a pivotal metal additive manufacturing technology, enabling near-net-shape fabrication with superior design flexibility and material efficiency. However, the process remains challenged by complex defect formation mechanisms—most notably spatter generation and melt pool instabilities—which hinder component quality, dimensional accuracy, and process repeatability. This work presents a comprehensive and multi-physics investigation that bridges powder-bed behavior, melt pool thermofluidics, and vapor-particle interactions to elucidate the underlying physics of defect evolution in LPBF. At the powder scale, a validated Discrete Element Method (DEM) model successfully captures powder spreading uniformity and surface roughness, serving as the foundation for simulating layer quality and initial bed conditions. At the melt pool scale, a high-fidelity CFD model incorporating heat conduction, Marangoni convection, vapor recoil pressure, surface tension effects, and phase change dynamics is developed. The results reveal that the rate of interfacial heat extraction at the solid–liquid boundary and localized thermal inhomogeneities—not total energy input—are the dominant parameter governing melt pool morphology and defect initiation. At high scan speeds, increased local thermal extraction steepens temperature gradients, driving hump and balling via thermally-induced surface retraction. Conversely, at low scanning speeds, prolonged laser exposure tends to induce frequent thermal accumulation events. Although the solid–liquid interfacial area increases—leading to greater total heat dissipation—the net heat accumulation within the melt pool can sometimes become negative. Under such conditions, vapor bubbles generated within the keyhole are more likely to be trapped by the solidification front, significantly increasing the risk of porosity formation. Furthermore, a novel heat transfer perspective redefines the origin of these defects: spatial thermal inhomogeneity along the melt pool generates localized capillary instabilities, leading to melt pool break-up. The role of preheating is also critically reassessed; while it improves thermal uniformity and suppresses surface instabilities at high scan speeds, it aggravates porosity at low speeds due to reduced Marangoni convection and enhanced vapor entrapment. In-depth analysis of powder spattering dynamics reveals a two-stage detachment mechanism, where particles first undergo micro-scale rolling or sliding driven by shear and vapor forces, then overcome interfacial friction and adhesion to become airborne. High-speed imaging, SEM observations, and CFD modeling collectively show that vapor-induced drag and turbulence dominate particle ejection, with rearward angular shift and clustering behavior becoming pronounced at elevated scan speeds. The final fate of particles-reincorporation or escape-is strongly influenced by melt pool wettability and plume oscillations. A modified spatter prediction model incorporating melt pool capillarity and vapor-plume coupling substantially improves agreement with experimental trends. This study offers a unified framework for understanding and predicting defect evolution in LPBF by tightly integrating powder-scale mechanics, melt pool thermodynamics, and gas-particle coupling. The insights provide a scientific basis for defect suppression strategies, real-time control, and digital twin development in next-generation metal AM platforms.