Kerogen H:C ratio calculation based on chemometric analysis of Raman spectroscopic results

There are several publications focused on calculating the maturity of carbonaceous matter (CM) from Raman spectroscopy data. Different criteria or parameters have been cited on bibliography as a tool to characterize the CM found in geologic samples. For instance, G band center, D1 center, G FWHM or Id/Ig parameters have been used with the mentioned aim.

One of the latest works¹ developed a Raman microscopy method to calculate the H:C ratio of the CM on fossils. In this way, the maturity of the CM could be studied. However, the method supposes the deconvolution or decomposition of Raman bands which can vary largely depending on the method, the software employed or the user. Besides, data should be treated spectrum by spectrum and a similar spectroscopic pre-treatment should be applied to the whole dataset. All these factors affect the final results and a wide uncertainty is introduced in the result.

In this work a chemometric method is developed to avoid these uncertainty sources. For the method, the whole spectrum is considered, not only certain regions. Moreover, spectra pre-treatment is not required. The method do not provide an absolute value but it allows comparing different CM to know which is more mature or which has less H:C ratio.

For the method development Raman microscopy was employed. Several coal working standards were used: SH95-S1-A, UWMA-1, UWLA-1, UWHA-1, PSOC1468, UWJA-1². All these standards were analyzed in a SIMS mount by micro-Raman spectroscopy 10 times. The employed Raman system was an InVia Renishaw spectrometer and a 514 nm laser was used. The conditions were the following ones: 50x objective, 1% laser power, 10s acquisitions, 1800l/mm grating and a spectral range from 100 to 3200 cm^-1. Principal Component Analysis (PCA) was performed with the whole range of all spectra without any spectral treatment.

The resulting plot (Figure 1) revealed that the samples were grouped by coal standard and they were ordered depending on their maturity. In addition, the method allowed the detection of outliers. Once the outliers were removed, PC1 + PC2 explained the 99%of the total variance.

Real samples could be introduced in the method to estimate their maturity and H:C ratio according to the standards values.

References

1. N. Ferralis et al. , Carbon 2018, 108: 440