The structure of appalachian coal: Experiments and Ab initio modeling

• Published in: Carbon (New York) Vol. 225; no. C; p. 119086
• Main Authors: Ugwumadu, Chinonso, Drabold, David A., Smith, Natasha L., Trembly, Jason, Olson, Rudolph, Shereda, Eric, Al-Majali, Yahya T.
• Format: Journal Article
• Language: English
• Published: United Kingdom Elsevier Ltd 01.05.2024; Elsevier
• Subjects: Coal; DFT; FTIR; NMR; ReaxFF; XPS; ReaxFF; NMR; XPS; DFT; FTIR; Coal
• ISSN: 0008-6223; 1873-3891
• DOI: 10.1016/j.carbon.2024.119086

This study focuses on the construction of small-scale atomistic representations of three Appalachian coals of different ranks– high volatile bituminous, low volatile bituminous, and semi-anthracite. Ultimate, proximate, 13C NMR, FTIR, and XPS data were used to infer 2D and construct multiple 3D molecular models for each coal. A new simulation strategy for structure optimization, termed “Sectioned Optimization”, is introduced to ensure energetically stable configurations of the 3D coal models. Density functional methods are applied to the models, improving their quality, and giving insight into the limitations of empirical simulations in this complex system. The structural models were validated ex post facto using a density functional code, VASP. We describe the distribution of local atomic environments in the models to emphasize the significant structural distinctions between the three coal ranks. We also report the first ab initio vibrational density of states calculations for coal and compare it to FTIR and Inelastic Neutron Scattering experiments and identify vibrational modes using ab initio calculations. We also provide animations of the vibrational normal modes in coal that are consistent with FTIR band assignments. We show that the specific heat capacity compares favorably with experimental data. [Display omitted] •2D structures of three Appalachian coals were inferred from NMR, FTIR, and XPS.•The 2D structures were then used to construct realistic 3D coal models.•3D models were first optimized using ReaxFF via sectioned optimization protocol.•DFT post-treatment significantly improved the structural features of the models.•The unicity, vibrations, and specific heat capacities of DFT models were analyzed.