Mechanisms behind the formation of microstructural defects in powder bed fusion processes: a review

Abstract

Powder bed fusion (PBF) is a leading metal additive manufacturing technique capable of producing complex, high-precision components with superior material efficiency. However, the technology’s widespread adoption is limited by the presence of microstructural defects such as porosity, cracks, and textured columnar grains. These defects, originating from both feedstock characteristics and process instabilities, significantly undermine the structural integrity, fatigue resistance, and reliability of manufactured parts. To systematically address these challenges, this review classifies microstructural defects into feedstock-induced and process-induced categories, examining their formation mechanisms and impact on mechanical properties. Feedstock-induced defects, including satellite formation, internal porosity, and surface contamination, stem from irregularities in feedstock morphology and handling conditions. Process-induced defects, such as lack of fusion pores, keyhole porosity, and cracking, are strongly correlated to thermal gradients, melt pool dynamics, and energy density variations. Furthermore, this work examines defect mitigation strategies, including feedstock optimization (powder atomization, particle morphology control), process parameter refinement (energy density regulation, scan strategy optimization), and post-processing techniques (hot isostatic pressing, surface treatments). A summarized classification of defects, their mechanisms, effects on mechanical properties, and mitigation strategies is provided as a comprehensive reference for researchers and practitioners. Through the integration of experimental findings, multi-physics modeling, and advanced in-situ diagnostics, this review establishes a framework for understanding defect formation and its influence on process-structure-property relationships. The insights provided address current limitations in defect modeling and control, guiding future research toward achieving higher repeatability and scalability in industrial applications.