High-throughput screening has become one of the major strategies for the discovery of novel functional materials. However, its effectiveness is severely limited by the lack of sufficient and diverse materials in current materials repositories such as the open quantum materials database (OQMD). Recent progress in deep learning have enabled generative strategies that learn implicit chemical rules for creating hypothetical materials with new compositions and structures. However, current materials generative models have difficulty in generating structurally diverse, chemically valid, and stable materials. Here we propose CubicGAN, a generative adversarial network (GAN) based deep neural network model for large scale generative design of novel cubic materials. When trained on 375 749 ternary materials from the OQMD database, the authors show that the model is able to not only rediscover most of the currently known cubic materials but also generate hypothetical materials of new structure prototypes. A total of 506 such materials have been verified by phonon dispersion calculation. Considering the importance of cubic materials in wide applications such as solar panels, the GAN model provides a promising approach to significantly expand existing materials repositories, enabling the discovery of new functional materials via screening. The new crystal structures discovered are freely accessible at www.carolinamatdb.org.

Two-dimensional (2D) materials have emerged as promising functional materials with many applications such as semiconductors and photovoltaics because of their unique optoelectronic properties. Although several thousand 2D materials have been screened in existing materials databases, discovering new 2D materials remains challenging. Herein, we propose a deep learning generative model for composition generation combined with a random forest-based 2D materials classifier to discover new hypothetical 2D materials. Furthermore, a template-based element substitution structure prediction approach is developed to predict the crystal structures of a subset of the newly predicted hypothetical formulas, which allows us to confirm their structure stability using DFT calculations. So far, we have discovered 267 489 new potential 2D materials compositions, where 1485 probability scores are more then 0.95. Among them, we have predicted 101 crystal structures and confirmed 92 2D/layered materials by DFT formation energy calculation. Our results show that generative machine learning models provide an effective way to explore the vast chemical design space for new 2D materials discovery.

The surface termination of MXenes greatly determines the electrochemical properties and ion kinetics on their surfaces. So far, hydroxyl-, oxygen-, and fluorine-terminated MXenes have been widely studied for energy storage applications. Recently, sulfur-functionalized MXene structures, which possess low diffusion barriers, have been proposed as candidate materials to enhance battery performance. We performed first-principles calculations on the structural, stability, electrochemical, and ion dynamic properties of Li-adsorbed sulfur-functionalized groups 3B, 4B, 5B, and 6B transition-metal (M)-based MXenes (i.e., M2CS2 with M = Sc, Ti, Zr, Hf, V, Nb, Ta, Cr, Mo, and W). We performed phonon calculations, which indicated that all of the above M2CS2 MXenes, except for Sc, are dynamically stable at T = 0 K. The ground-state structure of each M2CS2 monolayer depends on the type of M atom. For instance, while sulfur prefers to sit at the FCC site on Ti2CS2, it occupies the HCP site of Cr-based MXene. We determined the Li adsorption configurations at different concentrations using the cluster expansion method. The highest maximum open-circuit voltages were computed for the group 4B element (i.e., Ti, Zr, and Hf)-based M2CS2, which are larger than 2.1 V, while their average voltages are approximately 1 V. The maximum voltage for the group 6B element (i.e., Cr, Mo, W)-based M2CS2 is less than 1 V, and the average voltage is less than 0.71 V. We found that S functionalization is helpful for capacity improvements over the O-terminated MXenes. In this respect, the computed storage gravimetric capacity may reach up to 417.4 mAh/g for Ti2CS2 and 404.5 mAh/g for V2CS2. Ta-, Cr-, Mo-, and W-based M2CS2 MXenes show very low capacities, which are less than 100 mAh/g. The Li surface diffusion energy barriers for all of the considered MXenes are less than 0.22 eV, which is favorable for high charging and discharging rates. Finally, ab initio molecular dynamic simulations performed at 400 K and bond-length analysis with respect to Li concentration verify that selected promising systems are robust against thermally induced perturbations that may induce structural transformations or distortions and undesirable Li release.

Crystal structure prediction is now playing an increasingly important role in the discovery of new materials or crystal engineering.