Molecular dynamics-guided material model for the simulation of shock-induced pore collapse in β-octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (β-HMX)

Abstract

Material models for single-crystal β-HMX are systematically examined in the context of continuum pore-collapse simulations. Continuum predictions using five different isotropic material models are compared head-to-head with molecular dynamics (MD) predictions for a 50 nm cylindrical pore in β-HMX subject to a range of shock strengths. Shock waves were generated using a reverse-ballistic configuration, propagating along [010] in the MD simulations. The continuum models are improved hierarchically, drawing on temperature- and pressure-dependent MD-derived material parameters. This procedure reveals the sensitivity of the continuum predictions of pore collapse to the underlying thermophysical models. The study culminates in an MD-calibrated isotropic rate- and temperature-dependent strength model, which includes appropriate submodels for the temperature-dependent melting point of β-HMX [M. P. Kroonblawd and R. A. Austin, Mech. Mater. 152, 103644 (2021)], pressure-dependent shear modulus [A. Pereverzev and T. Sewell, Crystals 10, 1123 (2020)], and temperature-dependent specific heat, that produces continuum pore-collapse results similar to those predicted by MD. The resulting MD-informed model should improve the fidelity of simulations to predict the detonation initiation of HMX-based energetic materials containing micrometer-scale pores.

PARC: Physics-aware recurrent convolutional neural networks to assimilate meso scale reactive mechanics of energetic materials

“…The design loop for such a materials-by-design framework can be accelerated by using predictive multiscale simulations to obtain structure-property-performance (SPP) linkages for general EMs. Highly resolved molecular (15)(16)(17) or continuum simulations (18)(19)(20) are usually necessary for accurate modeling of the energy localization phenomena in complex microstructures. However, such direct numerical simulations (DNS) can be computationally intensive, and extracting useful quantities of interest (QoI) to inform the design may present challenges (2,21).…”

Strongly Anisotropic Thermomechanical Response to Shock Wave Loading in Oriented Samples of the Triclinic Molecular Crystal 1,3,5-Triamino-2,4,6-trinitrobenzene

“…By contrast, in many single-crystal shock experiments, the impact planes correspond to lower-index surfaces. − Therefore, understanding the anisotropic single-crystal shock response along a generalized direction is critical for connecting atomistic simulations to experiments. It is also highly important for the reliable formulation and parameterization of continuum mesoscale models which increasingly are being used to simulate the single-crystal and grain-scale response of pressed powder aggregates and polymer composites. − The recently reported generalized crystal-cutting method (GCCM) enables the construction, in principle, of two-dimensionally (2D) periodic cells suitable for MD simulations of shock wave propagation normal to any desired plane in crystals with an arbitrary space group and asymmetric unit.…”

Predicting Melt Curves of Energetic Materials Using Molecular Models

“…Finally, accurate prediction of T m P ð Þ is crucial when attempting to inform mesoscale or continuum scale simulations of phase coexistence behavior derived from molecular models, as it has been shown that different models for T m P ð Þ can qualitatively change the behavior observed in pore collapse studies of energetic materials [1]. The methodologies and analysis shown in this work have been streamlined and partially automated, thus, applying these approaches to other atomistic or coarse-grain models of energetic crystals to determine thermodynamically rigorous predictions of phase coexistence behavior is now straightforwardly attainable.…”

“…The number of grid spacings used in the long-range electrostatic calculations was set to 5 and 6 for RDX and HMX, respectively. The differences in the long-range electrostatic solver settings for RDX and HMX were not motivated by physical justification, but rather the desire to be consistent with previous studies [1,17].…”

“…Knowledge of the melting behavior of energetic materials is crucial for modeling the diffusion of mass and heat at a given state point since some physical properties do not vary continuously across a first-order phase transition. Continuum-level simulations must be informed of material properties, such as the melting temperature (T m ), from either experimental data or the predictions of finer-scale simulations [1,2]. Experimental determination of solid-liquid coexistence can be challenging for a variety of reasons, which include overheating, chemical reactivity, sample purity, as well as difficulties associated with generating high pressure environments.…”