Exploration of the Optimal Model for Raman Spectral Characterization of Bituminous Coal Structure: Based on Multipeak...

Area of Science:

• Geochemistry
• Analytical Chemistry
• Materials Science

Background:

• Raman spectroscopy is vital for coal structure characterization.
• Lack of standardized peak-fitting models leads to inconsistent results.
• Overfitting undermines the physical significance of coal analysis.

Purpose of the Study:

• To compare four peak-fitting models (GaussAmp, GLC, Lorentz, Voigt) for Raman spectroscopy of bituminous coal.
• To determine the optimal number of subpeaks for accurate analysis.
• To establish a framework combining statistical and physical interpretability.

Main Methods:

• Two bituminous coal samples of varying metamorphic grades were analyzed.
• Peak fitting was performed using 2-5 subpeaks for each model.
• Statistical indicators (F-test, AIC, BIC) and physical reasonableness of parameters were used for evaluation.

Main Results:

• Increasing subpeaks improved statistical fit but led to overfitting in the five-peak model.
• The Gaussian-LorenCross (GLC) four-peak model yielded R² > 0.999 and low RSS.
• Fitting parameters were physically reasonable, indicating mixed line-shape characteristics.

Conclusions:

• The GLC four-peak model is optimal for Raman spectroscopy of the studied bituminous coals.
• A combined statistical and physical interpretability framework is proposed.
• Further research will include a wider range of coal samples.