Automatic Identification and Consequences of Low-Melting-Point Impurity Particles in LPBF Al–Mg–Zr Powder

Low-melting-point impurities in powder feedstock can trigger local melting phenomena in laser powder bed fusion (LPBF) parts and may initiate defects in printed components. Here, we combine bulk chemistry with automated, high-throughput particle-by-particle SEM/EDS to identify and quantify Sn-containing impurity particles in two gas-atomized Al–Mg–Zr powder batches with different bulk Sn levels. The aim was not to establish a direct batch-to-batch performance comparison, but to clarify whether Sn was uniformly distributed among the powder particles or concentrated in rare impurity particles. Although ICP analysis indicated only 0.07 ± 0.02 wt.% Sn in the Sn-higher batch and <0.01 wt.% Sn in the Sn-lower batch, automated SEM/EDS screening of 20,001 particles per batch revealed that Sn was present as a very small number of highly enriched particles with Sn > 45 wt.% (eight particles in the Sn-higher batch and three particles in the Sn-lower batch). In the Sn-higher batch, Sn-rich particles were predominantly spherical and fell within the LPBF feedstock size window (Dmax ≈ 25–40 μm), implying that standard sieving would not remove them. BSE imaging and EDS mapping of polished sections and fracture surfaces of LPBF specimens built from the Sn-higher batch revealed spatially localized Sn-rich features associated with pores and Sn-rich phases on the fracture surface, supporting a direct powder-to-part transfer. These results demonstrate that low bulk impurity levels can mask highly localized, particle-scale contamination and highlight the need for particle-level compositional screening to support robust powder qualification and reuse decisions in LPBF.