Grant D. Smith

A methodology for performing phase-field simulations of plating and stripping in the presence of a solid electrolyte interphase (SEI) is presented and applied to a lithium metal electrode. Material point method (MPM) simulations are performed assuming a homogeneous SEI layer and are compared with experiment. Results are consistent with experiment for two electrolytes and confirm the dominance of the SEI layer in determining cell impedance. Notably, in some instances, the SEI potential drop greatly affected activation overpotential, differing from the applied potential. To accurately depict stripping current vs potential for SEI-free systems, the model considered non-ideal electrolyte effects: concentration-dependent salt activity coefficient, salt diffusion coefficient, and electrolyte conductivity. Conversely, systems with SEI layers displayed minimal non-ideal concentration-related electrolyte effects due to impedance originating primarily from the SEI. In plating scenarios, low SEI salt concentration negated the need for non-ideal SEI effects. However, for stripping, non-ideal salt-concentration dependent SEI effects were crucial in reproducing experimental behavior, owing to high salt concentration at the electrode/SEI interface.

Quasi-static uniaxial tension and compression, three-point bending, and Hopkinson bar measurements have been carried out on low-porosity Al/Ni compacts (pressed powders) as well as compacts of the constituent Al and Ni. Concurrently, mesoscale material point method (MPM) simulations of fully resolved Al, Ni, and Al/Ni compact microstructures and engineering scale MPM simulations of homogenized Al/Ni compacts have been conducted. These simulations, in conjunction with the experimental measurements, have allowed us to parameterize mesoscale particle/particle contact models and engineering scale anisotropic damage mechanics models for the materials. The goal of this coordinated experimental and simulation study of Al/Ni and its constituent materials was to better understand their mechanical properties and gain insight into the nature of Al–Ni and Al–Al interfaces in Al/Ni compacts and their impact on mechanical properties of the compacts. We found that brittle tensile failure of Al/Ni compacts was dominated by the strength of the Al–Al interface while ductile compressive behavior of Al/Ni compacts was largely determined by the behavior of the Al phase. The properties of both Al–Al interface and Al phase appear to be altered by the presence of Ni. Finally, it was found that softening resulting from shear damage offset underlying strain hardening of the Al and Ni phases during compressive loading of the Al/Ni compacts.

Molecular dynamics simulations of reactions in Al/Ni layered systems have been carried out under isothermal conditions for a wide range of temperatures and several system sizes. An embedded atom method potential, known to reasonably reproduce the phase behavior of Al/Ni, was employed. Simulations revealed reaction mechanisms involving an initial fast process and much slower more complex longer-time reactions. The initial reaction process consists of diffusion of Ni from the pure solid Ni phase into the molten Al phase, resulting in the formation of an Al-rich Al/Ni liquid. The initial reaction ends when the Al/Ni liquid becomes saturated in Ni and solid Al/Ni phases begin to form at the interfaces between the pure solid Ni phase and the Al/Ni liquid. The growth of these solid phases is intrinsically slow compared to the formation of the liquid and is further slowed by the need for Ni to diffuse through the growing interfacial Al/Ni solid phases. Analysis of the initial Al/Ni liquid forming process indicates Fickian behavior with the Ni diffusion coefficient exhibiting Arrhenius temperature dependence. The longer-time slow reaction process(es) resulting in the growth of Al/Ni solid phases do not lend themselves to detailed numerical analysis because of the complex dependence of the Ni transport on the detailed nature of the interfacial layers.

Continuum level Material Point Method (MPM) simulations have been carried out on low porosity (around 90% or greater of theoretical maximum density) microstructures of Al/Ni granular composites (Al/Ni). The Al/Ni microstructures were subjected to uniaxial (strain) compressive loading and shock compression up to 25 GPa. The MPM model accounted for frictional heating between grains in addition to plastic work and compressional heating effects. The distributions of stresses and temperature in the composite materials were found to be highly heterogeneous due to the heterogeneous nature of the composite microstructure. The manner in which interfaces between grains were treated (sliding vs. non-sliding) was found to influence both mechanical and thermal responses to loading. Plastic deformation, mechanical work, and grain/grain frictional effects led to modest increases of mean temperature and local hotspots with maximum temperature not higher than 800–850 K for loadings investigated. For all scenarios investigated, heat generation was insufficient to cause local Al melting believed to be a precursor for shock initiation.

Molecular dynamics simulations of fluoroalkyl-derivatized imidazolium:bis(trifluoromethylsulfonyl)imide (TFSI) room temperature ionic liquids (FADI-RTILs) with cations of the structure 1-F(CF₂)_n(CH₂)₂-3-methyl imidazolium have been performed and compared with simulations of alkyl-derivatized 1-H(CH₂)n_{+ 2}-3-methyl imidazolium analogs (ADI-RTILs). Simulations yield RTIL densities, viscosities and ionic conductivities for the FADI-RTILs and ADI-RTILs in reasonably good agreement with experimental data. Partial fluorination results in a larger increase in density than would be anticipated based upon the density difference between perfluoralkane and alkane melts. Similarly, the slowing down in dynamics upon partial fluorination is greater than would be expected based upon the increase in cation volume. Examination of cation–cation, anion–anion and cation–anion centers-of-mass radial distribution functions reveal remarkably little influence of partial fluorination on the spherically averaged intermolecular structure of the RTILs. Similarly, simulations reveal little change in tail conformations and the extent of tail–tail aggregation upon partial fluorination. The interaction of the TFSI anion with the positively charged imidazolium ring hydrogen and nitrogen atoms is also little influenced by partial fluorination. However, the partially fluorinated alkyl tail exhibits increased interaction with the TFSI anion due to the electron withdrawing character of the fluorinated groups. We believe this strong tail–anion electrostatic interaction largely accounts for the higher than expected density and slower than expected dynamics in the FADI-RTILs.

Atomistic molecular dynamics simulations were performed on a low-molecular-weight nitroplasticized Estane® mixture representative of the binder used in PBX-9501. Pressure-volume-temperature (PVT) behavior over a wide range of pressures and temperatures above the order-disorder temperature (ODT) of Estane was determined and represented with the empirical Tait and Sun equations of state. The effect of temperature, pressure, and plasticization on transport properties of the mixture was also examined. A combination of molecular dynamics simulations and theoretical reptation models was used to predict the shear stress relaxation modulus G(t) of PBX-9501 binder at 473 K and 1 atm pressure. Data obtained from simulations of the model PBX-9501 binder presented here can be utilized to predict the temperature and pressure dependence of the shear stress relaxation modulus for temperatures above the ODT.