Machine-Learning Prediction of CO Adsorption in Thiolated, Ag-Alloyed Au Nanoclusters.

We propose a machine-learning model, based on the random-forest method, to predict CO adsorption in thiolate protected nanoclusters. Two phases of feature selection and training, based initially on the Au25 nanocluster, are utilized in our model. One advantage to a machine-learning approach is that correlations in defined features disentangle relationships among the various structural parameters. For example, in Au25, we find that features based on the distribution of Ag atoms relative to the CO adsorption site are the most important in predicting adsorption energies. Our machine-learning model is easily extended to other Au-based nanoclusters, and we demonstrate predictions about CO adsorption on Ag-alloyed Au36 and Au133 nanoclusters.

Structural search for stable Mg-Ca alloys accelerated with a neural network interatomic model.

We have combined a neural network formalism with metaheuristic structural global search algorithms to systematically screen the Mg-Ca binary system for new (meta)stable alloys. The combination of these methods allows for an efficient exploration of the potential energy surface beyond the possibility of the traditional searches based on ab initio energy evaluations. The identified pool of low-enthalpy structures was complemented with special quasirandom structures (SQS) at different stoichiometries. In addition to the only Mg-Ca phase known to form under standard synthesis conditions, C14-Mg2Ca, the search has uncovered several candidate materials that could be synthesized under elevated temperatures or pressures. We show that the vibrational entropy lowers the relative free energy of several phases with magnesium kagome layers: C15 and C36 Laves structures at the 2 : 1 composition and an orthorhombic oS36 structure at the 7 : 2 composition. The estimated phase transition temperatures close to the melting point leave open the possibility of synthesizing the predicted materials at high temperatures. At high pressures up to 10 GPa, two new phases at the 1 : 1 and 3 : 1 Mg : Ca stoichiometries become thermodynamically stable and should form in multi-anvil experiments.

Firefly Algorithm Applied to Noncollinear Magnetic Phase Materials Prediction.

In most noncollinear crystal magnets, the number of metastable states is quite large and any calculation that tries to predict the ground state can fall into one of the possible metastable phases. In this work, we generalize the population based meta-heuristic firefly algorithm to the problem of the noncollinear magnetic phase ground state prediction within density functional theory (DFT). We extend the different steps in the firefly algorithm to this specific problem by using polarized constrained DFT calculations, whereby using Lagrange multipliers the directions of the atom magnetic moments remain fixed. By locking the directions of the magnetic moments at each search iteration, the method allows one to explore the entire Born-Oppenheimer energy surface of existing and physically plausible noncollinear configurations present in a crystal. We demonstrate that the number of minima can be large, which restrains the use of exhaustive searches.

A Machine-Driven Hunt for Global Reaction Coordinates of Azobenzene Photoisomerization.

Azobenzene is a very important system that is often studied for better understanding light-activated mechanical transformations via photoisomerization. The central C-N═N-C dihedral angle is widely recognized as the primary reaction coordinate for changing cis- to trans-azobenzene and vice versa. We report on a global reaction coordinate (containing all internal coordinates) to thoroughly describe the reaction mechanism for azobenzene photoisomerization. Our global reaction coordinate includes all of the internal coordinates of azobenzene contributing to the photoisomerization reaction coordinate. We quantify the contribution of each internal coordinate of azobenzene to the overall reaction mechanism. Finally, we provide a detailed mapping on how each significantly contributing internal coordinate changes throughout the energy profile (from trans to transition state and subsequently to cis). In our results, the central C-N═N-C dihedral remains the primary internal coordinate responsible for the reaction coordinate; however, we also conclude that the disputed inversion-assisted rotation is half as important to the overall reaction mechanism and the inversion-assisted rotation is driven by four adjacent dihedral angles C-C-N═N with very little change to the adjacent C-C-N angles.

Investigation of novel crystal structures of Bi-Sb binaries predicted using the minima hopping method.

Semi-conducting alloys BixSb1-x have emerged as a potential candidate for topological insulators and are well known for their novel thermoelectric properties. In this work, we present a systematic study of the low-energy phases of 35 different compositions of BixSb1-x (0 < x < 1) at zero temperature and zero pressure. We explore the potential energy surface of BixSb1-x as a function of Sb concentration by using the ab initio minima hopping structural search method. Even though Bi and Sb crystallize in the same R3[combining macron]m space group, our calculations indicate that BixSb1-x alloys can have several other thermodynamically stable crystal structures. In addition to the configurations on the convex hull, we find a large number of metastable structures which are dynamically stable. The electronic band structure calculations of several stable phases reveal the presence of strong spin-orbit interaction leading to the Rashba-Dresselhaus spin-splitting of bands which is of great interest for spintronics applications. We also find an orthorhombic structure of BiSb in the Imm2 space group which exhibits signatures of type-II Weyl semimetal. Additionally, we have studied the thermoelectric properties of the selected structures. Regarding thermoelectric properties, we find that the compositions which crystallize in the rhombohedral structure exhibit values of the Seebeck coefficient and the power factor similar to that of Bi2Te3 at room temperature, while the theoretical maximum figure of merit (ZeT) is smaller than that of Bi2Te3. We observe enhancement in the thermopower with the increase in the strength of the Rashba-Dresselhaus spin-splitting effect.

Firefly Algorithm for Structural Search.

The problem of computational structure prediction of materials is approached using the firefly (FF) algorithm. Starting from the chemical composition and optionally using prior knowledge of similar structures, the FF method is able to predict not only known stable structures but also a variety of novel competitive metastable structures. This article focuses on the strengths and limitations of the algorithm as a multimodal global searcher. The algorithm has been implemented in software package PyChemia ( https://github.com/MaterialsDiscovery/PyChemia ), an open source python library for materials analysis. We present applications of the method to van der Waals clusters and crystal structures. The FF method is shown to be competitive when compared to other population-based global searchers.