Efficient prediction of equations of state and strength properties using new electronic structure methods

We have demonstrated the use of density functional electronic structure (DFT) calculations to predict dislocation-based strength in addition to the equation of state (EOS) for high energy density studies of condensed matter. Practical challenges include numerical noise in derivatives of the EOS needed for ion-thermal effects and elasticity, the validity of pseudopotentials, and uncertainty from the use of exchange-correlation (XC) functionals. The new finite element DFT program SPARC-X provides a significant gain in speed and numerical convergence, predicting elastic moduli than can often be used directly for the strength and Debye temperature. When corrected to match the STP state, the EOS at terapascal pressures is much less sensitive to the XC functional, though still relies on the accuracy of pseudopotentials. We have compared with equivalent EOS and strength models derived from all-electron muffin-tin calculations. This approach can be used as readily for compounds as elements, depending on the complexity of the unit cell. We compare EOS predictions with recent high-pressure measurements, including the first ever Hugoniot experiments on Ru and the refractory intermetallic RuAl, made at the Z facility.