Toward first-principles description of amorphous monolayer carbon growth

The synthesis and properties of amorphous monolayer carbon have attracted extensive attention, yet the computational simulation of its growth mechanism poses great challenges. In a recent study, Monte Carlo simulation based on a hybrid potential model has enabled the atomic-scale growth of amorphous monolayer carbon, revealing a novel chaining-branching-networking-filling (CBNF) mechanism. In this study, we further push forward the accuracy of the growth simulation by combining a MACE-based machine-learning force field with three-dimensional continuous-space Monte Carlo sampling. The simulations shed light on the atomistic origins of the CBNF mechanism on the Cu(100) substrate: restricted carbon mobility, out-of-plane buckling, and transient subsurface carbon excursions induce kinetic trapping near deposition sites, favoring local chain extension before compact island formation; meanwhile, enhanced chain-end extension suppresses side branching, leading to elongated carbon chains and delayed network formation. These near-first-principles descriptions provide a reliable foundation for our further understanding of the structure and growth process of amorphous carbon.