Robust time-averaged molecular dynamics has been developed to calculate finiteerature elastic constants of a single crystal. We find that when the averaging time exceeds a certain threshold, the statistical errors in the calculated elastic constants become very small. We applied this method to compare the elastic constants of Pd and PdH0.6 at representative low (10 K) and high (500 K) temperatures. The values predicted for Pd match reasonably well with ultrasonic experimental data at both temperatures. In contrast, the predicted elastic constants for PdH0.6 only match well with ultrasonic data at 10 K; whereas, at 500 K, the predicted values are significantly lower. We hypothesize that at 500 K, the facile hydrogen diffusion in PdH0.6 alters the speed of sound, resulting in significantly reduced values of predicted elastic constants as compared to the ultrasonic experimental data. Literature mechanical testing experiments seem to support this hypothesis.
In atomistic simulations, diffusion energy barriers are usually calculated for each atomic jump path using a nudged elastic band method. Practical materials often involve thousands of distinct atomic jump paths that are not known a priori. Hence, it is often preferred to determine an overall diffusion energy barrier and an overall pre-exponential factor from the Arrhenius equation constructed through molecular dynamics simulations of mean square displacement of the diffusion species at different temperatures. This approach has been well established for interstitial diffusion, but not for substitutional diffusion at the same confidence. Using In 0.1 Ga 0.9 N as an example, we have identified conditions where molecular dynamics simulations can be used to calculate highly converged Arrhenius plots for substitutional alloys. As a result, this may enable many complex diffusion problems to be easily and reliably studied in the future using molecular dynamics, provided that moderate computing resources are available.
TlBr has the properties to become the leading radiation detection semiconductor. It has not yet been deployed due to a short lifetime of only hours to weeks. While the rapid structural deteriorations must come from ionic conduction under operating electrical fields, detailed aging mechanisms have not been understood. As a result, progress to extend lifetime has been limited despite extensive studies in the past. We have developed new atomistic simulation capabilities to enable study of ionic conduction under electrical fields. Our combined simulations and experiments indicate that dislocations in TlBr climb under electrical fields. This climb is the root cause for structural deterioration. Hence, we discovered new strengthening methods to reduce aging. Our new atomistic simulation approach can have broader impact on other Sandia programs including battery research. Our project results in 4 publications, a new invention, new LAMMPS capabilities, solution to mission relevant materials, and numerous presentations.
This paper uses molecular dynamics simulations to study surface and interface properties of PdHx that are relevant to hydrogen storage applications. In particular, surface energies, interfacial energies, surface diffusivities, and surface segregations are all determined as a function of temperature and composition. During the course of the calculations, we demonstrated robust molecular dynamics methods that can result in highly converged finite temperature properties. Challenging examples include accurate calculations of hydrogen surface diffusivities that account for all possible atomic jump mechanisms, and constructions of surface segregation composition profiles that have negligible statistical errors. Our robust calculations reveal that the Arrhenius plots of hydrogen surface diffusion is ideally linear at low compositions, and becomes nonlinear at high compositions. The fundamental cause for this behavior has been identified. This nonlinear surface diffusion behavioe is also in good agreement with available experimental data for bulk diffusion. The implication of our calculated properties on hydrogen storage application discussed.
Molecular dynamics simulations have been used to analyse microstructures of CdTe films grown on {112} surfaces of zinc-blende CdS. Interestingly, CdTe films grow in 331 orientations as opposed to 112 epitaxial orientations. At the CdTe-{331}/CdS-{112} interface, however, there exists an axis that is parallel to the 110 orientation of both CdS and CdTe. It is the direction orthogonal to this 110 that becomes different, being 116 for CdTe and 111 for CdS, respectively. Missing CdTe-{110} planes are found along the 110 axis, suggesting that the misfit strain is released by the conventional misfit dislocation mechanism along this axis. In the orthogonal axis, the misfit strain is found to be more effectively released by the new grain orientation mechanism. Our finding is supported by literature experimental observations of the change of growth direction when Cd0.96Zn0.04Te films are deposited on GaAs. Analyses of energetics clearly demonstrate the cause for the formation of the new orientation, and the insights gained from our studies can help understand the grain structures experimentally observed in lattice mismatched systems.
CdS homoepitaxy growth was performed by molecular dynamics using different substrate orientations and structures in order to analyze the CdS crystallinity. As anticipated from thermodynamics of homoepitaxy, highly crystalline films with only point defects were obtained on substrates with rectangular surface geometries, including [112¯] zinc blende (ZB), [101¯0] wurtzite (WZ), [112¯0] WZ, [110] ZB, [010] ZB, and [1101110] ZB. In contrast, films grown on substrates with hexagonal surface geometries, corresponding to the [0001] WZ and [111] ZB growth directions, showed structures with a large number of defects including; anti-sites, vacancies, stacking faults, twinning, and polytypism. WZ and ZB transitions and grain boundaries are identified using a lattice identification algorithm and represented graphically in a structural map. A dislocation analysis was performed to detect, identify, and quantify linear defects within the atomistic data. Systematic simulations using different temperatures, deposition rates, and substrate polarities were perform to analyze the trends of dislocation densities on [0001] WZ direction and showed persistent polytypism. The polytypism observed in the films grown on the substrates with hexagonal surface geometry is attributed to the similar formation energies of the WZ and ZB phases.
Hydrogen diffusion impacts the performance of solid-state hydrogen storage materials and contributes to the embrittlement of structural materials under hydrogen-containing environments. In atomistic simulations, the diffusion energy barriers are usually calculated using molecular statics simulations where a nudged elastic band method is used to constrain a path connecting the two end points of an atomic jump. This approach requires prior knowledge of the "end points". For alloy and defective systems, the number of possible atomic jumps with respect to local atomic configurations is tremendous. Even when these jumps can be exhaustively studied, it is still unclear how they can be combined to give an overall diffusion behavior seen in experiments. Here we describe the use of molecular dynamics simulations to determine the overall diffusion energy barrier from the Arrhenius equation. This method does not require information about atomic jumps, and it has additional advantages, such as the ability to incorporate finite temperature effects and to determine the pre-exponential factor. As a test case for a generic method, we focus on hydrogen diffusion in bulk aluminum. We find that the challenge of this method is the statistical variation of the results. However, highly converged energy barriers can be achieved by an appropriate set of temperatures, output time intervals (for tracking hydrogen positions), and a long total simulation time. Our results help elucidate the inconsistencies of the experimental diffusion data published in the literature. The robust approach developed here may also open up future molecular dynamics simulations to rapidly study diffusion properties of complex material systems in multidimensional spaces involving composition and defects.
We report on the thermodynamic properties of binary compound mixtures of model groups II-VI semiconductors. We use the recently introduced Stillinger-Weber Hamiltonian to model binary mixtures of CdTe and CdSe. We use molecular dynamics simulations to calculate the volume and enthalpy of mixing as a function of mole fraction. The lattice parameter of the mixture closely follows Vegard's law: a linear relation. This implies that the excess volume is a cubic function of mole fraction. A connection is made with hard sphere models of mixed fcc and zincblende structures. The potential energy exhibits a positive deviation from ideal soluton behaviour; the excess enthalpy is nearly independent of temperatures studied (300 and 533 K) and is well described by a simple cubic function of the mole fraction. Using a regular solution approach (combining non-ideal behaviour for the enthalpy with ideal solution behaviour for the entropy of mixing), we arrive at the Gibbs free energy of the mixture. The Gibbs free energy results indicate that the CdTe and CdSe mixtures exhibit phase separation. The upper consolute temperature is found to be 335 K. Finally, we provide the surface energy as a function of composition. It roughly follows ideal solution theory, but with a negative deviation (negative excess surface energy). This indicates that alloying increases the stability, even for nano-particles.
TlBr crystals have superior radiation detection properties; however, their properties degrade in the range of hours to weeks when an operating electrical field is applied. To account for this rapid degradation using the widely-accepted vacancy migration mechanism, the vacancy concentration must be orders of magnitude higher than any conventional estimates. The present work has incorporated a new analytical variable charge model in molecular dynamics (MD) simulations to examine the structural changes of materials under electrical fields. Our simulations indicate that dislocations in TlBr move under electrical fields. This discovery can lead to new understanding of TlBr aging mechanisms under external fields.
Carbon nanostructures, such as nanotubes and graphene, are of considerable interest due to their unique mechanical and electrical properties. The materials exhibit extremely high strength and conductivity when defects created during synthesis are minimized. Atomistic modeling is one technique for high resolution studies of defect formation and mitigation. To enable simulations of the mechanical behavior and growth mechanisms of C nanostructures, a high-fidelity analytical bond-order potential for the C is needed. To generate inputs for developing such a potential, we performed quantum mechanical calculations of various C structures.
In this project we developed t he atomistic models needed to predict how graphene grows when carbon is deposited on metal and semiconductor surfaces. We first calculated energies of many carbon configurations using first principles electronic structure calculations and then used these energies to construct an empirical bond order potentials that enable s comprehensive molecular dynamics simulation of growth. We validated our approach by comparing our predictions to experiments of graphene growth on Ir, Cu and Ge. The robustness of ou r understanding of graphene growth will enable high quality graphene to be grown on novel substrates which will expand the number of potential types of graphene electronic devices.
II-VI quantum dots, such as CdSe and CdTe, are attractive as downconversion materials for solid-state lighting, because of their narrow linewidth, tunable emission. However, for these materials to have acceptable quantum yields (QYs) requires that they be coated with a II-VI shell material whose valence band offset serves to confine the hole to the core. Confinement prevents the hole from accessing surface traps that lead to nonradiative decay of the exciton. Examples of such hole-confined core/shell QDs include CdTe/CdSe and CdSe/CdS. Unfortunately, the shell can also cause problems due to lattice mismatch, which ranges from 4-6% for systems of interest. This lattice mismatch can create significant interface energies at the heterojunction and places the core under radial compression and the shell under tangential tension. At elevated temperatures (~240°C) interfacial diffusion can relax these stresses, as can surface reconstruction, which can expose the core, creating hole traps. But such high temperatures favor the hexagonal Wurtzite structure, which has lower QY than the cubic zinc blende structure, which can be synthesized at lower temperatures, ~140°C. In the absence of alloying the core/shell structure can become metastable, or even unstable, if the shell is too thick. This can cause result in an irregular shell or even island growth. But if the shell is too thin thermallyactivated transport of the hole to surface traps can occur. In our LDRD we have developed a fundamental atomistic modeling capability, based on Stillinger-Weber and Bond-Order potentials we developed for the entire II-VI class. These pseudo-potentials have enabled us to conduct large-scale atomistic simulations that have led to the computation of phase diagrams of II-VI QDs. These phase diagrams demonstrate that at elevated temperatures the zinc blende phase of CdTe with CdSe grown on it epitaxially becomes thermodynamically unstable due to alloying. This is accompanied by a loss of hole confinement and a severe drop in the QY and emission lifetime, which is confirmed experimentally for the zinc blende core/shell QDs prepared at low temperatures. These QDs have QYs as high as 95%, which makes them very attractive for lighting. Finally, to address strain relaxation in these materials we developed a model for misfit dislocation formation that we have validated through atomistic simulations.