A multiscale strategy for modeling polycrystalline TATB at extreme conditions : Extending the model with crystal plasticity and thermochemistry

When high-energy-density materials are subjected to thermal or mechanical aggression such as shock loadings, a coupling between the thermo-mechanical and chemical behavior is involved. The present work focuses on the development of MD informed mesoscopic models for both mechanical and thermochemical behavior of TATB.

A mesoscopic constitutive law has been proposed and validated on MD simulations, where two dominant deformation mechanisms were identified: a buckling instability and a non-symmorphic twinning (irreversible) transformation. The proposed continuum model contains two main features: an anisotropic non-linear hyper-elasticity (EOS dependent) and a phase field formalism for twinning. This model being limited to loadings involving low stresses, we present its extension with a crystal plasticity formalism. The later is calibrated such that the continuum model reproduces with a high fidelity the TATB single crystal flow surface previously computed at the atomic scale, which contains the entire spectrum of deformation mechanisms as well as their relative strengths and contributions.

From another perspective, a mid-term goal would be to extend this formalism with a thermochemistry model. We develop a MD-informed reactive model for TATB at the mesoscopic scale where the chemical behavior and thermal transport properties of the system are determined by the underlying microscopic reactive simulations. All-atom reactive MD simulations are conducted and a reduced-order chemical kinetics model is fitted on adiabatic simulations of single crystal chemical decomposition. An unsupervised learning method along with temperature evolutions, are used to calibrate a four components model for the thermochemistry kinetics of TATB. This analytical formulation, coupled to a diffusion equation is incorporated in a continuum formalism and the model is compared against one-to-one MD simulations of a 1D hotspot. A good agreement is found for both time and spatial evolutions of the temperature field. We finally discuss some perspectives about the coupling between thermochemistry and mechanical models, based on observations at the atomic scale.