Evaluation of advanced SEM-EDX tools for classification of complex particles in respirable dust

Occupational lung disease remains a serious concern among miner workers, underscoring the need for improved characterization of respirable dust. Scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX) enables high-resolution analysis of filter samples, but accurate identification of complex, multi-constituent particles like agglomerates during direct-on-filter (DOF) analysis remains challenging. This is because standard tools for automated SEM-EDX treat each dust entity as an independent particle. In this study, we evaluated two advanced SEM-EDX tools for their potential to classify complex particles in respirable coal mine dust (RCMD) samples. One tool applies mineral liberation analysis (MLA), which collects EDX data pixel-by-pixel to enable classification of chemically distinct 'grains'. The other uses phase separation analysis (PSA) to enable classification of chemically and/morphologically distinct 'segments'. We tested both tools on five RCMD samples previously analyzed by standard automated and manual SEM-EDX methods. Both tools were generally capable of identifying and classifying complex particles, but certain strengths and limitations were observed. The MLA tool was more efficient and user-friendly, but it generally could not identify like-grained clusters (e.g., an agglomerate with two silica particles). The PSA could identify some such clusters, but required more extensive setup and data interpretation.