Formation of diamond and/or lonsdaleite phases under the effect of a nanoindenter on multilayer graphene: machine learning simulation

The formation of diamond in multilayer graphene under a local mechanical stress induced by the indentation of a film is theoretically studied. To describe this effect, the machine learning potential based on the representation of the atomic environment in the form of moment of inertia tensors is parameterized. The results demonstrate that the atomic structure of emerging diamond is determined by both graphene stacking and indenter size, providing an interplay of cubic and hexagonal diamond phases observed in the arising structure. The study also reveals a much lower phase transition pressure for graphene with the AA' stacking as compared to the ABC stacking in the films with a number of layers below hundred.