Session M47: Computational Design and Discovery of Novel Materials V

Structure prediction of epitaxial inorganic interfaces by lattice and surface matching with Ogre

We present a new version of the Ogre open source Python package with the capability to perform structure prediction of epitaxial inorganic interfaces by lattice and surface matching. In the lattice matching step, a scan over combinations of substrate and film Miller indices is performed to identify the domain-matched interfaces with the lowest mismatch. Subsequently, surface matching is conducted by Bayesian optimization to find the optimal interfacial distance and in-plane registry between the substrate and the film. For the objective function, a geometric score function is proposed based on the overlap and empty space between atomic spheres at the interface. The score function reproduces the results of density functional theory (DFT) at a fraction of the computational cost. The optimized interfaces are pre-ranked using a score function based on the similarity of the atomic environment at the interface to the bulk environment. Final ranking of the top candidate structures is performed with DFT. Ogre streamlines DFT calculations of interface energies and electronic properties by automating the construction of interface models. The application of Ogre is demonstrated for two interfaces of interest for quantum computing and spintronics, Al/InAs and Fe/InSb.   

Crystal structure prediction of energetic materials and a twisted arene with Genarris and GAtor

A molecular crystal structure prediction (CSP) workflow, based on the random structure generator, Genarris, and the genetic algorithm (GA), GAtor, is applied to the energetic materials 2,4,6-trinitrobenzene-1,3,5-triamine (TATB) and 2,4,6-trinitrobenzene-1,3-diamine (DATB), and the chiral arene, 4,5-dimethylphenanthrene. The experimental structures of all three materials are successfully generated multiple times by both Genarris and GAtor, and ranked as the most stable structures by dispersion-inclusive density functional theory (DFT) methods. For 4,5-dimethylphenanthrene the evolutionary niching feature of GAtor helps find the experimental structure by penalizing the fitness of over-sampled regions and steering the GA to an under-explored basin. For DATB, a putative structure with a sheet packing motif, which is associated with reduced sensitivity, is found to be very close in energy to the experimental structure and could be a viable polymorph. Principal component analysis of atom-centered symmetry functions is used to compare the crystal structure landscapes of TATB and DATB. Genarris and GAtor exhibit robust performance for diverse targets with varied intermolecular interactions. This work demonstrates the potential of including CSP as a part of the energetic materials development process.   

Design principles for the discovery of non-cubic polymorphous networks

Symmetry breaking resulting in the formation of a distribution of local structural and/or spin motifs (i.e., polymorphous network) has been experimentally observed and theoretically described for a number of cubic ABX₃ perovskites (e.g., CsPbI₃, CsSnI₃, BaTiO₃, and PbTiO₃), demonstrating that properties of such compounds cannot be predicted accurately with density functional theory (DFT) using high-symmetry Pm-3m monomorphous unit cells. Herein, we show that structural and/or spin symmetry breaking is not limited to cubic ABX3 systems but exists for a range of paraelastic, paraelectric, and paramagnetic non-cubic compounds (e.g., GeTe, Bi₂O₃, EuTiO₃, and YNiO₃). Importantly, such potential polymorphous compounds can be predicted successfully based on the analysis of group/subgroup crystal symmetries and their relative energetics. Using such group/subgroup relations and DFT, we form an extensive list of realistic symmetry-broken polymorphous compounds ready for experimental validation. We also demonstrate that accounting for symmetry breaking can allow describing correctly the electronic structure of potential polymorphous compounds without including any dynamic correlation.   

Ab initio modeling of MAB phase solid solutions

MAB phases are layered ternary or quaternary compounds, where M stands for a transition metal element. A indicates a group III-A or IV-A element, and B is boron. MAB structures are composed of M-B sublattices interleaved by A-atom mono- or bilayers. MAB phases are technologically important materials exhibiting both metallic and ceramic properties. MAB phases crystallize in different crystal structures depending on the composition. Such diversity of crystal structures strongly suggests that MAB phases' properties are tunable through the crystal structure and compositional control. Tuning the chemical composition offers an additional degree of freedom in the quest for stable materials with physical and chemical properties appealing for both fundamental research and technological applications. In this respect, we investigated the phase stability of MAB phase solutions by mixing two different transition metal elements using, density functional calculations, cluster expansion, and special quasirandom structure approach. We elucidated the mixing thermodynamics of different transition metal atoms in MAB solid alloy solutions. We determined the alloying behaviors, namely (1) an ordering tendency to form ordered structures, (2) a clustering tendency to form phase separation, and (3) a mixing tendency to form disordered alloy structures. We discussed the material properties of the predicted MAB solid solutions.   

Crystal structure prediction from bulk (oxynitrides) to interfaces (SnO2/CdTe)

In recent discovery efforts for mixed-cation ternary nitrides and oxynitrides [JCP 154, 234706 (2021)], we successfully utilized the kinetically limited minimization (KLM) approach. Of increasing interest and importance is the prediction of interface structures, a rather multifaceted problem: Epitaxial interfaces between isostructural materials can often be readily constructed without structure prediction. Otherwise, interface structures are often obtained by splicing two freestanding surfaces with subsequent geometry optimization. However, the general interface problem remains very challenging. We adopted the KLM approach to slab geometry and applied it to SnO2/CdTe interfaces without and with the ubiquitous CdCl2 treatment in CdTe photovoltaics. We find that the lowest energy SnO2/CdTe interface is not a surface splice but has instead a fractional first CdTe atomic layer that forms the connection between the two rather dissimilar materials. This interface is highly defective, both structurally and electronically. The CdCl2 addition results in a thin interlayer which strongly reduces the interface energy, improves the atomic connectivity, and dramatically reduces the defect density of states in the CdTe band gap.   

Novel high-pressure phases in the binary transition metal chalcogenide systems of Nb, Mo, Se and S

We explore and compare the phase diagrams of the binary Mo-S, Mo-Se, Nb-Se and Nb-S systems from ambient pressures up to 250 GPa using ab initio evolutionary crystal structure prediction. We investigate new stable compositions and phases, especially in the high-pressure regime, and discuss their electronic, vibrational, and superconducting properties as a function of the elemental composition.

This systematic study provides detailed insights on the phase diagrams of binary transition metal chalcogenide systems in a wide pressure range, representing a reference work for future investigations.   

Coordinate-free representation of crystals for accelerated materials discovery using machine learning.

Machine-learning (ML) is emerging as a useful technique to create very efficient computational models of the relationships between the structure and properties of a material. The ability to uncover such relations is highly dependent on the choice of representation of the material. Many representations rely either on the description of the composition, which fails to capture the uniqueness of the material (i.e., does not differentiate polymorphs) or on the accurate spatial description via atomic coordinates, which are not readily available when screening hypothetical materials. In this talk, we present a crystal symmetry-based coordinate-free representation¹. It describes the structure via the occupied Wyckoff positions, which captures enough spatial information, but coarse-grains the precise placements of the atoms into an enumerable search space. This descriptor allows us to train an ML model with an exceptionally high hit ratio for finding new stable materials. We demonstrate this by showing that our model identifies 1,558 materials below the known convex hull of previously calculated materials from just 5,675 ab-initio calculations.

1.  arXiv:2106.11132v2 [cond-mat.mtrl-sci]    

Boron-doped alumina for smart lubricants: Theory and experiment

We have carried out atomistic modeling of ceramic alumina (Al₂O₃) doped with boron oxide (B₂O₃) using density functional theory (DFT). Alumina exhibits excellent material properties, including a high melting point, hardness, and corrosion resistance. Alumina also possesses good tribological properties that can be enhanced with the addition of solid lubricants. In this study we investigated structural changes of alumina with respect to boron weight percentages (wt%) between 5% to 20% in order to understand the role of boron, introducing tribological properties to a mechanically favorable structure. Increasing amounts of boron oxide doping were added by substitution to an alumina unit cell and optimized using DFT. The most energetically favorable structures at each boron wt% were identified and were directly compared to experimental x-ray diffraction (XRD) patterns, finding agreement between theoretical and experimental patterns for stable aluminum borate phase (9Al₂O₃ 2B₂O₃). Further investigation into theoretically produced structures can then be carried out to determine properties favorable to lubrication.   

Thermophysical Investigation of Sodium-Indium Alloy

This study explores the mixing nature of sodium-indium liquid alloy at 713 K, 850 K, 950 K and 1050 K temperatures. It uses quasi-chemical approximation for the thermodynamic analysis of concentration dependent mixing behaviors of the alloy under the assumption of Na₃In complex. It compares the obtained theoretical results with the experimental and results of Redlich-Kister equation for the validity. The researchers concentrate on the viscosity and surface tension of the alloy under the modelling equations as suggested by Kaptay and improved derivation of Butler equation respectively. This paper focuses on the interaction energy parameters among neighboring atoms of the alloy. It observes that the alloy is moderately interacting and ordering nature at the lower concentration of sodium. The theoretical results of the thermodynamic properties are nearly in agreement with the corresponding experimental data as well as results obtained by R-K equation at 713 K.  The study shows that the ordering behavior of the alloy decreases with the increase in temperature. Both viscosity and surface tension have a decreasing tendency with an increase in temperature.

Keywords: Liquid alloys, thermodynamic properties, R-K equation, energy parameters