Theoretical Insight into the Band Alignment at High-κ Oxide XO2/Diamond (X = Hf and Zr) Interfaces with a SiO2 Interlayer for MOS Devices.

Diamond has become a promising candidate for high-power devices based on its ultrawide bandgap and excellent thermoelectric properties, where an appropriate gate dielectric has been a bottleneck hindering the development of diamond devices. Herein, we have systematically investigated the structural arrangement and electronic properties of diamond/high-κ oxide (HfO2, ZrO2) heterojunctions by first-principles calculations with a SiO2 interlayer. Charge analysis reveals that the C-Si bonding interface attracts a large amount of charge concentrated at the diamond interface, indicating the potential for the formation of a 2D hole gas (2DHG). The diamond/HfO2 and diamond/ZrO2 heterostructures exhibit similar "Type II" band alignments with VBOs of 2.47 and 2.21 eV, respectively, which is consistent with experimental predictions. The introduction of a SiO2 dielectric layer into the diamond/SiO2/high-κ stacks exhibits the typical "Type I″ straddling band offsets (BOs). In addition, the wide bandgap SiO2 interlayer keeps the valence band maximum (VBM) and conduction band minimum (CBM) in the stacks away from those of diamond, effectively confining the electrons and holes in MOS devices. This work exhibits the potential of SiO2/high-κ oxide gate dielectrics for diamond devices and provides theoretical insights into the rational design of high-quality gate dielectrics for diamond-based MOS device applications.

Machine Learning Paves the Way for High Entropy Compounds Exploration: Challenges, Progress, and Outlook.

Machine learning (ML) has emerged as a powerful tool in the research field of high entropy compounds (HECs), which have gained worldwide attention due to their vast compositional space and abundant regulatability. However, the complex structure space of HEC poses challenges to traditional experimental and computational approaches, necessitating the adoption of machine learning. Microscopically, machine learning can model the Hamiltonian of the HEC system, enabling atomic-level property investigations, while macroscopically, it can analyze macroscopic material characteristics such as hardness, melting point, and ductility. Various machine learning algorithms, both traditional methods and deep neural networks, can be employed in HEC research. Comprehensive and accurate data collection, feature engineering, and model training and selection through cross-validation are crucial for establishing excellent ML models. ML also holds promise in analyzing phase structures and stability, constructing potentials in simulations, and facilitating the design of functional materials. Although some domains, such as magnetic and device materials, still require further exploration, machine learning's potential in HEC research is substantial. Consequently, machine learning has become an indispensable tool in understanding and exploiting the capabilities of HEC, serving as the foundation for the new paradigm of Artificial-intelligence-assisted material exploration.

Active Learning the High-Dimensional Transferable Hubbard U and V Parameters in the DFT + U + V Scheme.

Density functional theory (DFT) is a powerful quantum mechanical computational tool to perform electronic structure calculations for materials. Few DFT methods can ensure accuracy and efficiency simultaneously. DFT + U + V is an alternative effective approach to overcome this drawback. However, the accuracy sensitively depends on the self-consistent estimation of the high-dimensional onsite and intersite Hubbard interaction U and V terms. We propose Bayesian optimization using a dropout (BOD) algorithm, one type of active learning method, to optimize U and V terms. The DFT + U + V with U/V obtained by BOD can produce improved electronic properties for diverse bulk materials of comparable quality to the hybrid functionals with lower computational cost compared to the linear response approach. Note that the band gaps calculated by BOD are somewhat different from that of hybrid functionals by simply applying the same U/V parameters as in the case of surface slabs and interfaces, which suggests that the transferability of U/V from the bulk models to slabs and interfaces is not as well as expected. BOD is extended to calculate the U/V parameters for slabs and interfaces and reach similar results as bulk solids. Moreover, we find that the U/V are reasonably transferable between surface slabs and interfaces with different thicknesses under various effects of quantum confinement, which contributes to fast access to the electronic properties of large-scale systems with higher accuracy.