To explore the use of Raman spectroscopy in analyzing crystalline silica content within coal dust, the Raman spectroscopy techniques were investigated for quantitative analysis of silica in coal dust samples from China based on unary linear regression (ULR) and partial least squares regression (PLSR) method. In the experimental setup, the formulated coal dust samples with known crystalline silica content were meticulously prepared and subjected to an ashing procedure using a muffle furnace. The cooled mixtures were placed in water for magnetic stirring before being dispensed onto an aluminum substrate via a pipette gun, ultimately forming dry sample deposits. According to the Raman spectrum data from these dry sample deposits, a calibration curve was generated and served as the foundation for predicting crystalline silica content in real coal dust samples. In the quantitative analysis of the crystalline silica load in the formulated coal dust samples deposited within a diameter of about 1 cm, the calibration curve R² value of the ULR model is 0.90, with a limit of detection (LOD) is 0.96% (i.e., 19.17 μg). Meanwhile, the R² value of the PLSR model can reach 0.988, and concurrently, the LOD can be reduced to 0.18% (3.65 μg). Moreover, the improved PLSR model, which established with coal dust samples collected from a broader range of geological conditions, demonstrates high accuracy and good agreement between the analysis results of the prediction set and the measurements obtained from the standard infrared (IR) spectrometric method. This study indicates the potential of Raman spectroscopy for measuring crystalline silica in coal dust.

coal dust,crystalline silica,quantitative analysis,Raman spectroscopy