Reduced graphene oxide membrane models for hydrogen storage

A database of 600 atomistic models of reduced graphene oxide (rGO) and hydrogenated rGO (H-rGO) was constructed, comprising 120 rGO and 480 H-rGO structures. The dataset spans a broad range of oxygen concentrations, –O–:–OH (epoxy/ether to hydroxyl) ratios, and hydrogenation levels. The database is designed for applications in computational modeling, machine learning, and structure–property analyses. The models were generated following a three-step simulation protocol:
– Step 1 | Generation of Pseudo-GO Models: Graphene sheets were functionalized with oxygen groups (–O– or –OH), randomly distributed according to three –O–:–OH ratios (25:75, 50:50, and 75:25). Functionalization was applied symmetrically on both sides of the sheet, avoiding adjacent carbon atoms.
– Step 2 | Thermal Reduction: Each pseudo-GO structure was subjected to annealing at four different temperatures (1000, 1500, 2000, and 2500 K). After initial relaxation at 300 K, the models were heated, equilibrated, annealed back to room temperature, and re-optimized. This step produced 120 rGO models exhibiting a range of oxygen contents, functional group distributions, and defect patterns.
– Step 3 | Hydrogenation of rGO Models: rGO models were exposed to atomic hydrogen using a combined Grand Canonical Monte Carlo and molecular dynamics (GCMC/MD) scheme at 300 K and four hydrogen pressures (1, 2, 4, and 8 bar). Hydrogenation was performed with 50 GCMC exchange steps every 1 ps of MD, followed by relaxation and removal of unstable species. This step generated 480 H-rGO models with varying hydrogen coverage and bonding configurations.
All simulations were performed using the LAMMPS software package with the ReaxFF reactive force field.