Predicting Melt Curves of Energetic Materials Using Molecular Models.

0557891 - ÚCHP 2023 RIV DE eng J - Journal Article
Tow, G.M. - Larentzos, J.P. - Sellers, M.S. - Lísal, Martin - Brennan, J.K.
Predicting Melt Curves of Energetic Materials Using Molecular Models.
Propellants Explosives Pyrotechnics. Roč. 47, č. 8 (2022), č. článku e202100363. ISSN 0721-3115. E-ISSN 1521-4087
Grant - others:ARL(US) W911NF-21- 2-0177; ARF(US) W911NF-20-2-0203; HPCMP(US) HIP-20-021
Institutional support: RVO:67985858
Keywords : HMX * melting point * molecular dynamics
OECD category: Physical chemistry
Impact factor: 1.8, year: 2022
Method of publishing: Limited access

In this work, the solid-liquid coexistence curves of classical fully flexible atomistic models of alpha-RDX and beta-HMX were calculated using thermodynamically rigorous methodologies that identify where the free energy difference between the phases is zero. The free energy difference between each phase at a given state point was computed using the pseudosupercritical path (PSCP) method, and Gibbs-Helmholtz integration was used to evaluate the solid-liquid free energy difference as a function of temperature. This procedure was repeated for several pressures to determine points along the coexistence curve, which were then fit to the Simon-Glatzel functional form. While effective, this method is computationally expensive. An alternative approach is to compute the melting point at a single pressure via the PSCP method, and then use the Gibbs-Duhem integration technique to trace out the coexistence curve in a more computationally economical manner. Both approaches were used to determine the coexistence curve of alpha-RDX. The Gibbs-Duhem integration method was shown to generate a melt curve that is in good agreement with the PSCP-derived melt curve, while only costing similar to 10 % of the computational resources used for the PSCP method. For alpha-RDX, the predicted melting temperature increases significantly more for a given increase in pressure when compared to available experimental data.
Permanent Link: http://hdl.handle.net/11104/0331771