Ab initio calculations for high explosive equations of state

Modeling detonation in a hydrodynamic simulation requires a reactant and product equation of state (EOS). The latter EOS is particularly challenging to obtain from ab initio density functional theory (DFT) simulations of the condensed phase due to the complications of chemical bond breaking and formation. Leveraging a novel nested Monte Carlo simulation framework with a machine learning derived reference potential, we extend our original analysis of the high explosive PETN to two other high explosives, namely, HMX and DAAF. We find DFT is again not quantitatively accurate but we discuss a possible first principles correction to improve the quantitative accuracy.