Atomic scale defects critically limit performance of semiconductor materials. To improve materials, defect effects and defect formation mechanisms must be understood. In this paper, we demonstrate multiple examples where molecular dynamics simulations have effectively addressed these issues that were not well addressed in prior experiments. In the first case, we report our recent progress on modelling graphene growth, where we found that defects in graphene are created around periphery of islands throughout graphene growth, not just in regions where graphene islands impinge as believed previously. In the second case, we report our recent progress on modelling TlBr, where we discovered that under an electric field, edge dislocations in TlBr migrate in both slip and climb directions. The climb motion ejects extensive vacancies that can cause the rapid aging of the material seen in experiments. In the third case, we discovered that the growth of InGaN films on (0001) surfaces suffers from a serious polymorphism problem that creates enormous amounts of defects. Growth on surfaces, on the other hand, results in single crystalline wurtzite films without any of these defects. In the fourth case, we first used simulations to derive dislocation energies that do not possess any noticeable statistical errors, and then used these error-free methods to discover possible misuse of misfit dislocation theory in past thin film studies. Finally, we highlight the significance of molecular dynamics simulations in reducing defects in the design space of nanostructures.
Austenitic stainless steels (Fe-Cr-Ni) are resistant to hydrogen embrittlement but have not been studied using molecular dynamics simulations due to the lack of an Fe-Cr-Ni-H interatomic potential. Herein we describe our recent progress towards molecular dynamics studies of hydrogen effects in Fe-Cr-Ni stainless steels. We first describe our Fe-Cr-Ni-H interatomic potential and demonstrate its characteristics relevant to mechanical properties. We then demonstrate that our potential can be used in molecular dynamics simulations to derive Arrhenius equation of hydrogen diffusion and to reveal twinning and phase transformation deformation mechanisms in stainless steels.
Fe-Ni-Cr stainless-steels are important structural materials because of their superior strength and corrosion resistance. Atomistic studies of mechanical properties of stainless-steels, however, have been limited by the lack of high-fidelity interatomic potentials. Here in this paper, using density functional theory as a guide, we have developed a new Fe-Ni-Cr embedded atom method potential. We demonstrate that our potential enables stable molecular dynamics simulations of stainless-steel alloys at high temperatures, accurately reproduces the stacking fault energy—known to strongly influence the mode of plastic deformation (e.g., twinning vs. dislocation glide vs. cross-slip)—of these alloys over a range of compositions, and gives reasonable elastic constants, energies, and volumes for various compositions. The latter are pertinent for determining short-range order and solute strengthening effects. Our results suggest that our potential is suitable for studying mechanical properties of austenitic and ferritic stainless-steels which have vast implementation in the scientific and industrial communities.
Solid-state hydrogen storage materials undergo complex phase transformations whose behavior are collectively determined by thermodynamic (e.g., Gibbs free energy), mechanical (e.g., lattice and elastic constants), and mass transport (e.g., diffusivity) properties. These properties depend on the reaction conditions and evolve continuously during (de)hydrogenation. Thus, they are difficult to measure in experiments. Because of this, past progress to improve solid-state hydrogen storage materials has been prolonged. Using PdHx as a representative example for interstitial metal hydride, we have recently applied molecular dynamics simulations to quantify hydrogen diffusion in the entire reaction space of temperature and composition. Here, we have further applied molecular dynamics simulations to obtain well-converged expressions for lattice constants, Gibbs free energies, and elastic constants of PdHx at various stages of the reaction. Our studies confirm significant dependence of elastic constants on temperature and composition. Specifically, a new dynamic effect of hydrogen diffusion on elastic constants is discovered and discussed.
Solid-state hydrogen storage materials undergo complex phase transformations whose kinetics is often limited by hydrogen diffusion. Among metal hydrides, palladium hydride undergoes a diffusional phase transformation upon hydrogen uptake, during which the hydrogen diffusivity varies with hydrogen composition and temperature. Here we perform robust statistically-averaged molecular dynamics simulations to obtain a well-converged analytical expression for hydrogen diffusivity in bulk palladium that is valid throughout all stages of the reaction. Our studies confirm significant dependence of the diffusivity on composition and temperature that elucidate key trends in the available experimental measurements. Whereas at low hydrogen compositions, a single process dominates, at high hydrogen compositions, diffusion is found to exhibit behavior consistent with multiple hopping barriers. Further analysis, supported by nudged elastic band computations, suggests that the multi-barrier diffusion can be interpreted as two distinct mechanisms corresponding to hydrogen-rich and hydrogen-poor local environments.
Molecular dynamics (MD) simulations and experimental evaporation were applied to study the growth of evaporated (Cu)ZnTe on mono- and polycrystalline CdTe. The simulated structures show polytypism and polycrystallinity, including texturing and grain boundaries, diffusion, and other phenomena in excellent qualitative agreement with experimental atomic probe tomography, transmission electron microscope, and secondary ion mass spectrometry. Results show formation of Cu clusters in nonstoichiometric growths even at early stages of deposition. Results also show significantly faster diffusion along defected regions (uncorrelated CdTe grain boundaries) as compared with more highly crystalline areas (high-symmetry grain boundaries and pristine regions). Activation energies and pre-exponential factors of Cu, Zn, and Te diffusion were determined using simulation. The MD model captures crystal growth phenomena with a high degree of fidelity.
The growth dynamics and evolution of intrinsic stacking faults, lamellar, and double positioning twin grain boundaries were explored using molecular dynamics simulations during the growth of CdTe homoepitaxy and CdTe/CdS heteroepitaxy. Initial substrate structures were created containing either stacking fault or one type of twin grain boundary, and films were subsequently deposited to study the evolution of the underlying defect. Results show that during homoepitaxy the film growth was epitaxial and the substrate's defects propagated into the epilayer, except for the stacking fault case where the defect disappeared after the film thickness increased. In contrast, films grown on heteroepitaxy conditions formed misfit dislocations and grew with a small angle tilt (within ∼5°) of the underlying substrate's orientation to alleviate the lattice mismatch. Grain boundary proliferation was observed in the lamellar and double positioning twin cases. Our study indicates that it is possible to influence the propagation of high symmetry planar defects by selecting a suitable substrate defect configuration, thereby controlling the film defect morphology.
Al-Based Al-Cu alloys have a very high strength to density ratio, and are therefore important materials for transportation systems including vehicles and aircrafts. These alloys also appear to have a high resistance to hydrogen embrittlement, and as a result, are being explored for hydrogen related applications. To enable fundamental studies of mechanical behavior of Al-Cu alloys under hydrogen environments, we have developed an Al-Cu-H bond-order potential according to the formalism implemented in the molecular dynamics code LAMMPS. Our potential not only fits well to properties of a variety of elemental and compound configurations (with coordination varying from 1 to 12) including small clusters, bulk lattices, defects, and surfaces, but also passes stringent molecular dynamics simulation tests that sample chaotic configurations. Careful studies verified that this Al-Cu-H potential predicts structural property trends close to experimental results and quantum-mechanical calculations; in addition, it properly captures Al-Cu, Al-H, and Cu-H phase diagrams and enables simulations of H2 dissociation, chemisorption, and absorption on Al-Cu surfaces.
A Stillinger-Weber potential is computationally very efficient for molecular dynamics simulations. Despite its simple mathematical form, the Stillinger-Weber potential can be easily parameterized to ensure that crystal structures with tetrahedral bond angles (e.g., diamond-cubic, zinc-blende, and wurtzite) are stable and have the lowest energy. As a result, the Stillinger-Weber potential has been widely used to study a variety of semiconductor elements and alloys. When studying an A-B binary system, however, the Stillinger-Weber potential is associated with two major drawbacks. First, it significantly overestimates the elastic constants of elements A and B, limiting its use for systems involving both compounds and elements (e.g., an A/AB multilayer). Second, it prescribes equal energy for zinc-blende and wurtzite crystals, limiting its use for compounds with large stacking fault energies. Here in this paper, we utilize the polymorphic potential style recently implemented in LAMMPS to develop a modified Stillinger-Weber potential for InGaN that overcomes these two problems.