Molecular level structure development of Indian coal using experimental, ML and DFT techniques

•Molecular level presentation of coal.•Molecular structure development using machine learning (ML) and density functional theory (DFT) methods.•ReaxFF MD simulations to understand thermal decomposition of coal.•DFT IR spectra.•Characterization of coal using FTIR and NMR techniques. The nature of interaction between coal surface and reagent affects fine coal flotation yield. Hence, it is crucial to understand the nature of coal surface to improve flotation yield. Therefore, determination of coal surface at the molecular level is important to optimize the flotation performance. But coal is a complex material, and its molecular presentations depend highly on its geographical origin. In this work, various experimental techniques such as ultimate analysis, proximate analysis, TGA, FTIR and solid state 13C NMR were used to characterize the structure of low-rank Indian coal. The data extracted from these experiments were then used to develop molecular level presentation of coal using machine learning (ML) and DFT. Five stable molecular level structures were proposed for this Indian coal. DFT calculated FTIR spectra matches reasonably with the experimental FTIR data. Finally, a 3D model of the coal sample was developed, and reactive force field (ReaxFF) molecular dynamics simulations were performed for thermal decomposition analysis. [Display omitted]