A Raman spectroscopic method for measuring the crystalline silica content in coal dust

Abstract

The applicability of Raman spectroscopy for quantitative analysis of crystalline silica content within coal dust was investigated. We prepared the formulated coal dust samples with known crystalline silica content and ashed them using a muffle furnace, followed by redeposition onto aluminum substrates to form dry sample deposits. These samples were then analyzed using Raman spectroscopy. Both univariate and multivariate calibration models were constructed for relating the Raman spectra from these dry sample deposits to the crystalline silica contents. The R² value of the unary linear regression (ULR) model is 0.900, with a detection limit of 0.96 %. Meanwhile, the R² value of the partial least squares regression (PLSR) model can reach 0.988, and the detection limit can be reduced to 0.18 %. A PLSR model for field coal dust samples collected from a broader range of geological conditions was established and used for predicting the crystalline silica content in unknown coal dust samples. The measurement results agree well with those obtained from the standard infrared (IR) spectrometric method, with a root mean square error of 2.35 %. This study demonstrates the potential of Raman spectroscopy for accurately measuring crystalline silica content in coal dust.