Alan Sanders-Gutierrez, OscarLuna-Valenzuela, AnalilaPosada-Borbon, AlvaroSchoen, J. Christian...
11页
查看更多>>摘要：The thermal behavior of 38-atom mono-, bi-, and trimetallic clusters consisting of Cu, Ag, and Au atoms, is analyzed employing molecular dynamics simulations and DFT calculations for selected cluster compositions. Low-energy structures were singled out to perform NVT molecular dynamics simulations at several temperatures, using the Andersen thermostat for temperature control. The caloric curve is used to estimate the melting temperature and the specific heat. The pair distribution function g(r) of the solid and liquid-phase clusters is examined at different temperatures. When comparing the estimated melting points (T-m) among the monatomic clusters, the order becomes T-m(Cu38) > T-m(Ag38) > T-m(Au38). For bimetallic clusters, an increase of T-m is observed for Cu-Au compared to their monatomic counterparts, while the opposite occurs for Cu-Ag clusters. For trimetallic clusters, two low-energy isomers of the Cu36Ag1Au1 cluster are investigated. In this case, T-m is estimated to be 475 K, for the two isomers with the lowest-energy and second-to-lowest energy, respectively. For all the clusters studied, the pair distribution function g(r) shows that the first peak position is not shifted as an effect of temperature and its maximum value varies with composition, while the second peak essentially vanishes upon melting. The common-neighbor analysis (CNA) technique is used to analyze the local structural changes for the trimetallic clusters, again demonstrating a clear structural change upon melting. The HOMO-LUMO energy gap indicates that the trimetallic isomers' behavior is metallic, while the average binding energy show these clusters' energetic stability to be similar.
Schultz, Lane E.Afflerbach, BenjaminSzlufarska, IzabelaMorgan, Dane...
9页
查看更多>>摘要：We explore the use of characteristic temperatures derived from molecular dynamics to predict aspects of metallic Glass Forming Ability (GFA). Temperatures derived from cooling curves of self-diffusion, viscosity, and energy were used as features for machine learning models of GFA. Multiple target and model combinations with these features were explored. First, we use the logarithm of critical casting thickness, log(10)(D-max), as the target and trained regression models on 21 compositions. Application of 3-fold cross-validation on the 21 log(10)(D-max) alloys showed only weak correlation between the model predictions and the target values. Second, the GFA of alloys were quantified by melt-spinning or suction casting amorphization behavior, with alloys that showed crystalline phases after synthesis classified as Poor GFA and those with pure amorphous phases as Good GFA. Binary GFA classification was then modeled using decision tree-based methods (random forest and gradient boosting models) and were assessed with nested-cross validation. The maximum F1 score for the precision-recall with Good Glass Forming Ability as the positive class was 0.82 +/- 0.01 for the best model type. We also compared using simple functions of characteristic temperatures as features in place of the temperatures themselves and found no statistically significant difference in predictive abilities. Although the predictive ability of the models developed here are modest, this work demonstrates clearly that one can use molecular dynamics simulations and machine learning to predict metal glass forming ability.
查看更多>>摘要：Efficiency of two methods for structural analysis, the Common Neighbor Analysis (CNA) from the program OVITO and the Coordination Polyhedron Method (CPM), was compared when applied to detect all possible structural atoms in monoatomic test clusters. The analyzed clusters are in the form of spherical fragments of a pure or disturbed crystal structure or they represent Mackay icosahedra. It was observed that the number of detected structural atoms for both methods is initially similar but decreases rapidly with the increase of applied random translations of atoms. CNA gives higher number of detected structural atoms representing bulk-centered cubic structure, while CPM has significant advantage in detection of icosahedral units. CNA is unable to identify decahedral units; therefore, this method fails to visualize clearly any global symmetry in internal cluster structure. Attempts are made to understand the reasons for the poor performance of both methods in the mentioned aspects and some ways are proposed to improve them.
查看更多>>摘要：Shape effect on Metallic Nano particles has many applications in technology, engineering and industry. Molecular dynamic simulations have been performed to investigate size and shape effects of Iridium nanocluster on its surface energy (E-sur) and self-diffusion coefficient (D) as a function of temperature (T). Truncated octahedron (TO), octahedron (Oh), cubic (C), and face center cubic (FCC) of Iridium nanocluster were studied in this survey. At temperatures lower than 1000 K, E-sur order is as follows: TO > C > Oh > FCC while at temperatures higher than 1000 K it has a different order as TO > Oh > C > FCC. Self-diffusion coefficient increases versus temperatures as expected, but results show that it is roughly the same for all systems at T < 1800 K and T greater than 2500. At temperature range T (1800, 2500), self-diffusion coefficient order is as follows: TO approximate to Oh approximate to C < FCC. TO congruent to Oh congruent to C < FCCInterestingly, D-FCC is smaller than D for other shapes of Iridium nanocluster. Self-diffusion coefficient as a function of nanocluster size was calculated with molecular dynamics simulation. It is shown there is a peak for D at N = number of particles = 700 which corresponds to surface effect. MD results show that FCC has more solid like long-range interaction, but others have more liquid like order.
NSTL
Elsevier