Machine Learning Force Field Predictions of Structural and Dynamical Properties in HOPG Defects and the HOPG-Water...

Understanding the electronic, structural, and dynamical properties of highly oriented pyrolytic graphite (HOPG) and its interface with water is crucial, as its layered structure, vacancy defects, and elemental doping offer untapped potential across various fields. This study employed ab initio (DFT) methods and state-of-the-art "on-the-fly" machine learning force fields (MLFFs) to investigate pristine HOPG, vacancy-engineered and doped systems (N, O, S), graphene nanoribbon (GNR) interfaces, and the HOPG-water interface. Projected, total, and local density of states (PDOS, TDOS, LDOS) analyses revealed that defects and dopants significantly modified the electronic band structure by introducing midgap states, altering orbital overlaps, shifting the Fermi level, and inducing local magnetic moments. GNRs interfaced with HOPG exhibited hybridized electronic states, forming distinct bands via π-orbital interactions, while dopants such as C-N improved conductivity, C-O introduced midgap states, and C-S generated localized defect states affecting conductivity and reactivity. Defect-induced crystallographic distortions were captured in 3D band structures, and mean squared displacement (MSD) analysis confirmed atomic stability in solid HOPG and structured water layering at the HOPG-water interface due to disrupted hydrogen bonding. The integration of MLFFs enabled efficient and accurate simulations of large atomic systems, significantly reducing the computational cost of ab initio molecular dynamics. Overall, this work offers critical insights into how defects and interfacial environments influence HOPG's electronic behavior, providing a robust computational framework for the rational design of defect-engineered carbon-based materials. These findings have direct implications for advancing HOPG in catalysis, energy storage systems such as batteries and supercapacitors, nanoelectronics including field-effect transistors, as well as in sensors, functional coatings, and other next-generation electronic applications.