Computational Approaches for Chemistry Under Extreme Conditions

High-fidelity modeling of hydrocarbon polymers is important for gaining fundamental understanding of the underlying material behavior as well as for designing high energy density (HED) experiments. In this chapter, we describe multi-scale modeling/simulation of hydrocarbon polymers done at Sandia and corresponding experiments on Sandia’s Z machine. For polymers, a combination of first-principles simulations using density functional theory (DFT) and atomistic simulations using classical molecular dynamics has proven to effectively model different aspects of the system and we will present both. Throughout, we find that the simulations are in qualitative and quantitative agreement with experiments, suggesting that the hierarchy of simulations can be used to increase our understanding of polymers under dynamic loading conditions.

High-nitrogen-content energetic compounds containing multiple N–N bonds are an attractive candidate for new generation of environmentally friendly, and more powerful energetic materials. High-N content translates into much higher heat of formation resulting in much larger energy output, detonation pressure, and velocity upon conversion to large amounts of non-toxic, strongly bonded \(\text {N}_{2}\) gas. This chapter describes recent advances in the computational discovery of a new family of polynitrogen pentazolate compounds using powerful first-principles evolutionary crystal structure prediction methods. After description of the methodology of the first-principles crystal structure prediction, several new high-nitrogen-content energetic compounds are described. In addition to providing information on structure and chemical composition, theory/simulations also suggests specific precursors, and experimental conditions that are required for experimental synthesis of such high-N pentazolate energetic materials. To aid in experimental detection of newly synthesized compounds, XRD patterns and corresponding Raman spectra are calculated for several candidate structures. The ultimate success was achieved in joint theoretical and experimental discovery of cesium pentazolate, which was synthesized by compressing and heating cesium azide \(\text {CsN}_{3}\) and \(\text {N}_{2}\) precursors in diamond anvil cell. This success story highlights the key role of first-principles structure prediction simulations in guiding experimental exploration of new high-N energetic materials.

Shock-induced phenomena in materials occur on timescales that while short may still be beyond the reach of traditional molecular dynamics simulations. The shock-induced chemistry of liquid benzene provides an excellent example of the importance of timescale in shock experiments; reactions are seen at about 13.3 GPa on microsecond timescales in plate impact experiments but it appears inert at up to 20 GPa over 100s of picoseconds during laser-driven shock experiments. We have studied the shock-induced chemistry of liquid benzene using a semiempirical reactive interatomic potential at timescales beyond those routinely accessible to traditional molecular dynamics simulations. We have applied replica-based accelerated molecular dynamics to this system because the initial chemical reactions themselves can be viewed as rare, state-to-state transitions that take place under thermal activation. Replica-based accelerated molecular dynamics enables us to parallelize the simulations in time with no loss of accuracy, provided that transitions (reactions) can be detected reliably. We have simulated the shocked chemical dynamics of benzene on timescales up to 7.7 ns with high parallel efficiency. The simulations show the formation dimers through Diels–Alder condensation. The dimers subsequently condense into larger polymeric structures, in good accord with experiments and quantum chemical data.

We present methods for the creation of semi-empirical quantum approaches and reactive force fields through force matching to quantum simulation data for materials under reactive conditions. Our methodologies overcome the extreme computational cost of standard Kohn–Sham Density Functional Theory (DFT) by mapping DFT computed simulation data onto functional forms with linear dependence on their parameters. This allows for quick parameterization of our models by avoiding the nonlinear fitting bottlenecks associated with most molecular dynamics model development. We illustrate our approach with two different systems: (i) determination of density functional tight binding models for aqueous glycine dimerization, and (ii) determination of the Chebyshev Interactional Model for Efficient Simulation (ChIMES) reactive force field for metallic liquid carbon. In each case, we observe that our approach is easy to parametrize and yields a model that is orders of magnitude faster than DFT while largely retaining its accuracy. Overall, our methods have potential use for studying complex long time and length scale chemical reactivity at extreme conditions, where there is a significant need for computationally efficient atomistic simulations methods to aid in the interpretation and design of experiments.

The old and challenging problem of dealing with the interaction between condensed matter systems and intense external electric fields are currently evolving in an impressive way. In fact, the growth of the computational resources allows for accurate first-principles numerical calculations showing unprecedented predictive power. We review the phenomenological evidence that has recently emerged from state-of-the-art ab initio molecular dynamics simulations in describing how static electric fields can be exploited to manipulate matter and possibly design novel compounds or materials, obtain new exotic properties, and achieve more efficient reaction yields. In particular, we show the microscopic behavior of simple molecular liquids (water, methanol, and simple mixtures), under the action of static and homogeneous electric fields, showing different shades of the effects produced by the application of the latter. In addition, ab initio molecular dynamics approaches are coupled with advanced free energy methods, that currently represents a unique technique for adequately treating, reproducing, and predicting both molecular mechanisms and chemical reaction networks triggered when matter is exposed to the action of intense electric fields.

The application of theory and computation to understand reactivity at high pressures is beset by several challenges: (1) the nontrivial changes in electronic structure that take place during the reaction, (2) the many possible initial configurations of reacting species, and (3) the simulation timescales needed for reaction events to occur. In this chapter, we will discuss two methods for meeting these challenges. The development of accurate molecular mechanics force fields is needed to sample initial configurations of reactants. This chapter provides a perspective on the functional forms and parameterization strategies of modern force fields. In particular, we highlight the ForceBalance parameterization method for optimizing force fields systematically and reproducibly using a free and open-source code. The ab initio nanoreactor is a new simulation method for rapidly discovering new reaction pathways from first-principles molecular dynamics. The main components of the nanoreactor approach include an external time-dependent potential that induces high-velocity molecular collisions, a trajectory analysis and visualization tool for identifying and extracting individual reaction events, and a reaction path optimization workflow for estimating the reaction energies and barrier heights from a reaction event.

Rocket engines, gas turbines, HCCI engines, and other such combustion devices frequently exceed the critical pressure of the fuel or the oxidizer. Modeling of combustion processes at high-pressure operating condition is required to determine the reaction rates based on which chemical kinetic models are developed. The current need is to focus on the transfer from low pressure to high-pressure conditions as this can have a significant effect on the chemistry as well as the reaction rates. The ReaxFF reactive force field method is a computationally feasible method used to study the combustion kinetics of fuels and fuel mixtures at supercritical condition. In this chapter, ReaxFF-MD is used to investigate the effect of a highly reactive fuel on the properties of a less reactive fuel at different levels of concentration, temperature, and density/pressure. The activation energies, based on Arrhenius-type rate laws, are compared with those from Continuum simulations and the limitations of the latter has been elaborated on. The study reveals a pressure/temperature regime and mixing conditions, where simple first-order kinetics-based Arrhenius-type relations cannot be applied. The reason can be attributed to different initial reaction mechanisms and product distributions of the two fuels considered. These results indicate how ReaxFF-based molecular dynamics simulations can provide significant atomistic insights on the combustion properties of fuel mixtures at supercritical conditions, where experiments are difficult to perform.

The fast loading rates associated with shockwaves in solids make molecular dynamics (MD) a particularly well-suited tool for their study. This chapter focuses on recent methods to study shock-induced chemistry using all-atom reactive MD and coarse-grained simulations and their application. We describe insight on the formation of hot spots formed following the shock-induced collapse of pores and their transition to a deflagration wave in high energy density materials obtained from large-scale MD simulations using the reactive force field ReaxFF. Experimental validation of such simulations is critical to assess the predictive capabilities of these methods to describe new materials and show how to extract observables from the simulations that can be directly contrasted with experiments. Such direct comparisons are not just critical for validation but also contribute to the interpretation of the experimental results. We also describe coarse-grained simulations to study the possibility and effectiveness of shock-induced, endothermic, volume-collapsing reactions; these simulations quantify how the various characteristics of the chemical reactions attenuate the propagating shockwave and provide key information to experimentalists designing and synthesizing such materials.

Complex chemical processes such as those found in combustion, the decomposition of energetic materials, and the chemistry of planetary interiors, are typically studied at the atomistic level using molecular dynamics (MD) simulations. A nascent but growing trend in many areas of science and technology is to consider a data-driven approach to studying complex processes, and molecular dynamics simulations, especially at high temperatures and pressures, are a prime example of an area ripe for disruption with this approach. MD simulations are expensive, but each simulation generates a wealth of data. In this chapter, we discuss a statistical learning framework for extracting information about the underlying chemical reactions observed in MD data, and using it to build a fast kinetic Monte Carlo (KMC) model of the corresponding chemical reaction network. We will show our KMC models can not only extrapolate the behavior of the chemical system by as much as an order of magnitude in time but can also be used to study the dynamics of entirely different chemical trajectories. We will also discuss a new and efficient data-driven algorithm for reducing our learned KMC models using L1-regularization. This allows us to reduce complex chemical reaction networks consisting of thousands of reactions in a matter of minutes.

This chapter describes efforts to enable multiscale modeling of energetic material response to insult through a concurrent hierarchical multiscale framework. As a demonstration, a quantum-derived, particle-based coarse-grain model of an energetic material is used to provide part of the constitutive response in a finite element multiphysics simulation. Bottom-up coarse-grain models of hexahydro-1,3,5-trinitro-s-triazine (RDX) and the methods used to perform reactive simulations at the microscale will be described. Simulations demonstrating microstructure-dependent initiation are also presented. Research opportunities addressing the remaining challenges related to detonation are discussed.