Artem Samtsevych

We have developed Density Functional Tight Binding models for cerium and cerium oxide with generalized gradient approximation and hybrid functionals that accurately predict the electronic band structure of these materials. We show that determination of a many-body repulsion energy with the Chebyshev Interaction Model for Simulation correctly predicts cerium allotropic energetic ordering with a minimal training created from small-scale molecular dynamics calculations of a single phase only. Our approach determines DFTB models for both electronic and material property predictions that retain a high degree of transferability.

Density functional tight binding (DFTB) offers a computationally efficient alternative to ab initio methods, bridging the accuracy of density functional theory (DFT) and the speed of semiempirical models. The approximate nature of DFTB makes its reliability highly dependent on parameter quality. While recent advancements have significantly improved the parametrization of the so-called repulsive potential, the parametrization of the electronic part of the DFTB interaction remains relatively simplistic and underdeveloped. We present DFTB Slater-Koster Optimizer (DSKO), a novel framework that aims at producing accurate and transferable electronic parameter sets under rigorous physical constraints. Incorporating robust optimization algorithms and physics-informed loss functions, DSKO generates DFTB electronic parameters that yield electronic properties, such as density of states and band structures, closely matching DFT reference data. The versatility of DSKO facilitates the wide application of DFTB to materials science challenges, paving the way for routine high-fidelity semiempirical simulations.

Density Functional Tight-Binding (DFTB) offers a computationally efficient alternative to {\em ab initio} methods, bridging between the accuracy of DFT and the speed of semiempirical models. The approximate nature of DFTB makes its reliability highly dependent on parameter quality though. While recent advancements have significantly improved the parametrization of the so-called repulsive potential, the parametrization of the electronic part of the DFTB interaction remains relatively simplistic and underdeveloped. We present our in-house DFTB Slater-Koster Optimizer (DSKO), a novel framework that aims at producing accurate and transferable electronic parameter sets under rigorous physical constraints. Incorporating robust optimization algorithms and physics-informed loss functions, DSKO generates DFTB electronic parameters that yield electronic properties, such as density of states and band structures, closely matching DFT reference data. The versatility of DSKO facilitates the wide application of DFTB to materials science challenges, paving the way for routine high-fidelity semiempirical simulations.

The increasing popularity of machine learning (ML) approaches in computational modeling, most prominently ML interatomic potentials, opened possibilities that were unthinkable only a few years ago—structure and dynamics for systems up to many thousands of atoms at an ab initio level of accuracy. Strictly referring to ML interatomic potentials, however, a number of modeling applications are out of reach, specifically those that require explicit electronic structure. Hybrid (“gray box”) models based on, e.g., approximate, semi-empirical ab initio electronic structure with the aid of some ML components offer a convenient synthesis that allows us to treat all aspects of a certain physical system on the same footing without targeting a separate ML model for each property. Here, we compare one of these [Density Functional Tight Binding with a Gaussian Process Regression repulsive potential (GPrep-DFTB)] with its fully “black box” counterpart, the Gaussian approximation potential, by evaluating performance in terms of accuracy, extrapolation power, and data efficiency for the metallic Ru and oxide RuO2 systems, given exactly the same training set. The accuracy with respect to the training set or similar chemical motifs turns out to be comparable. GPrep-DFTB is, however, slightly more data efficient. The robustness of GPRep-DFTB in terms of extrapolation power is much less clear-cut for the binary system than for the pristine system, most likely due to imperfections in the electronic parametrization.

The discovery of new catalysts is one of the significant topics of computational chemistry as it has the potential to accelerate the adoption of renewable energy sources. Recently developed deep learning approaches such as graph neural networks (GNNs) open new opportunity to significantly extend scope for modelling novel high-performance catalysts. Nevertheless, the graph representation of particular crystal structure is not a straightforward task due to the ambiguous connectivity schemes and numerous embeddings of nodes and edges. Here we present embedding improvement for GNN that has been modified by Voronoi tesselation and is able to predict the energy of catalytic systems within Open Catalyst Project dataset. Enrichment of the graph was calculated via Voronoi tessellation and the corresponding contact solid angles and types (direct or indirect) were considered as features of edges and Voronoi volumes were used as node characteristics. The auxiliary approach was enriching node representation by intrinsic atomic properties (electronegativity, period and group position). Proposed modifications allowed us to improve the mean absolute error of the original model and the final error equals to 651 meV per atom on the Open Catalyst Project dataset and 6 meV per atom on the intermetallics dataset. Also, by consideration of additional dataset, we show that a sensible choice of data can decrease the error to values above physically-based 20 meV per atom threshold.

Polynitrogen molecules are attractive for high-energy-density materials due to energy stored in nitrogen–nitrogen bonds; however, it remains challenging to find energy-efficient synthetic routes and stabilization mechanisms for these compounds. Direct synthesis from molecular dinitrogen requires overcoming large activation barriers and the reaction products are prone to inherent inhomogeneity. Here we report the synthesis of planar N₆²⁻ hexazine dianions, stabilized in K₂N₆, from potassium azide (KN₃) on laser heating in a diamond anvil cell at pressures above 45 GPa. The resulting K₂N₆, which exhibits a metallic lustre, remains metastable down to 20 GPa. Synchrotron X-ray diffraction and Raman spectroscopy were used to identify this material, through good agreement with the theoretically predicted structural, vibrational and electronic properties for K₂N₆. The N₆²⁻ rings characterized here are likely to be present in other high-energy-density materials stabilized by pressure. Under 30 GPa, an unusual N₂^0.75−-containing compound with the formula K₃(N₂)₄ was formed instead.