Application of image analysis to quantify maceral composition of source rocks: Examples from the Devonian New Albany Shale and Marcellus Shale

Abstract

The composition of sedimentary organic matter (OM) is an important parameter that determines the hydrocarbon potential and reveals the depositional conditions of organic-rich shales. Several automated analysis methods have been developed to determine the maceral composition of coals, but few studies have applied these techniques to assess the composition of OM in source rocks. This research developed an image evaluation method that combines maceral identification with machine-learning algorithms to quantify OM compositions. Three Devonian shales, two samples of New Albany Shale and one of Marcellus Shale, ranging from marginally mature to overmature were selected to evaluate the thermal evolution of maceral components, including vitrinite, inertinite, liptinite, and secondary products (i.e., solid bitumen and pyrobitumen). The method provides an efficient approach for identifying pyrobitumen and alginite in samples and is superior to automated coal analysis methods. Comparison of traditional point-counting methods with the new approach validates the effectiveness of image analysis in quantifying vitrinite and inertinite contents. However, the challenge of extracting amorphous OM mixed with mineral matter from the background requires further refinement. This methodological advancement provides a new tool for assessing the composition, sources, and thermal evolution of OM, offering valuable data to complement organic geochemical interpretations of depositional environments.