Carbon and Other Low-Z materials Under Extreme Conditions

This work is focused on understanding material\u27s behavior and response to extreme conditions. Under extreme conditions, which is categorized as regions of high pressures and temperatures in (P-T) space, materials can undergo multiple types of phase transitions as well as exhibit substantial damage as well as other exotic behaviors. By studying matter at these extreme conditions, we can elucidate a broad range of fundamental physics including a material\u27s energetic, mechanical, and electronic responses. This thesis describes work that makes contributions to the growing body of knowledge within these subsets of condensed matter physics. In the first thrust, crystal structure prediction methods are utilized to predict new stable structures of lithium pentazolates, as well as pathways to their synthesis. In the second thrust, a modified screened environment-dependent reactive empirical bond-order (SED-REBO) potential for carbon is developed and used to determine the impact of vacancy defects on the intrinsic strength of graphene. In the final thrust, a Spectral Neighbor Analysis Potential (SNAP) machine learning potential for carbon is developed, which demonstrates a quantum accurate description of carbon at extreme conditions. The predictive power of this quantum accurate SNAP, combined with excellent computational performance on leadership-class GPU-enabled high performance computing (HPC) systems is demonstrated by running groundbreaking simulations of synthesis of the elusive BC8 phase from amorphous carbon and shock wave propagation in diamond, which uncovered fundamental atomic-scale mechanisms of its anomalous strength and inelastic deformations in shocked diamond