Materials Data toward Machine Learning: Advances and Challenges.

Machine learning (ML) is believed to have enabled a paradigm shift in materials research, and in practice, ML has demonstrated its power in speeding up the cost-efficient discovery of new materials and autonomizing materials laboratories. In this Perspective, current research progress in materials data which are the backbones of ML are reviewed, focusing on high-throughput data generation, standardized data storage, and data representation. More importantly, the challenging issues in materials data that should be overcome to unlock the full potential of ML in materials research and development, including classic 5V (volume, velocity, variety, veracity, and value) issues, 3M (multicomponent, multiscale, and multistage) challenges, co-mining of experimental and computational data, and materials data toward transferable/explainable ML or causal ML, are discussed.

The Combination of Icotinib Hydrochloride and Fluzoparib Enhances the Radiosensitivity of Biliary Tract Cancer Cells.

BACKGROUND: Radiotherapy and chemotherapy are the main clinical treatments for biliary tract cancers (BTCs). Patients with advanced disease have a very poor prognosis, yet no molecular targets have been proven effective for the adjuvant therapy of BTCs. In this study, we aimed to explore the effect of combination treatment with icotinib hydrochloride (IH) and fluzoparib (FZ) on radiosensitivity and clarify its underlying mechanism in the HCCC-9810 and GBC-SD human BTC cell lines. METHODS: Cell proliferation was measured by Cell Counting Kit-8 (CCK-8) assay. The cell cycle distribution and apoptosis were analyzed by flow cytometry. The phosphorylation of EGFR and its downstream signaling molecules and the expression of RAD51 were measured by Western blot analysis. γ-H2AX foci in the cellular nuclei were visualized using immunofluorescence staining. A colony formation assay was performed to demonstrate cell radiosensitivity with IH and FZ combination treatment. RESULTS: In the HCCC-9810 and GBC-SD human BTC cell lines, combined treatment with IH and FZ with synergetic radiation significantly inhibited cell proliferation, redistributed the cell cycle, enhanced apoptosis and delayed DNA damage repair by suppressing activation of the EGFR signaling pathway and attenuating expression of the homologous recombination (HR) protein RAD51. CONCLUSION: This study demonstrates that combined treatment with IH and FZ may be an applicable therapy to enhance the radiosensitivity of BTCs and that RAD51 may serve as a biomarker for this combination treatment.

Tunable phase transitions and high photovoltaic performance of two-dimensional In2Ge2Te6 semiconductors.

Ultrathin semiconductors with great electrical and photovoltaic performance hold tremendous promise for fundamental research and applications in next-generation electronic devices. Here, we report new 2D direct-bandgap semiconductors, namely mono- and few-layer In2Ge2Te6, with a range of desired properties from ab initio simulations. We suggest that 2D In2Ge2Te6 samples should be highly stable and can be experimentally fabricated by mechanical exfoliation. They are predicted to exhibit extraordinary optical absorption and high photovoltaic conversion efficiency (≥31.8%), comparable to the most efficient single-junction GaAs solar cell. We reveal that, thanks to the presence of van Hove singularities in the band structure, unusual quantum-phase transitions could be induced in monolayers via electrostatic doping. Furthermore, taking bilayer In2Ge2Te6 as a prototypical system, we demonstrate the application of van der Waals pressure as a promising strategy to tune the electronic and stacking property of 2D crystals. Our work creates exciting opportunities to explore various quantum phases and atomic stacking, as well as potential applications of 2D In2Ge2Te6 in future nanoelectronics.

Composition-Gradient-Mediated Semiconductor-Metal Transition in Ternary Transition-Metal-Dichalcogenide Bilayers.

The semiconductor-metal transition (SMT) enables multiple applications of one single material, especially in modern devices. How to control it remains one of the most intriguing questions in material physics/chemistry, especially in two-dimensional layered materials. In this work, we report realization of SMT in MoS2-xOx bilayers, driven by the concentration gradient of the chalcogen atom across the van der Waals (vdW) gap of the disordered bilayers. Using the cluster expansion method, we determined that either semiconducting (stable) or metallic states (metastable) can be realized in MoS2-xOx bilayers with the same composition. Machine learning analysis revealed that the concentration gradient of the chalcogen atom across the vdW gap is the leading fingerprint of SMT, with structural distortion induced by atom mixing being a significant secondary factor. The electronic origin of the SMT is the broadening of the Mo dz2 and O pz bands, accompanied by the redistribution of the d electrons. This in-vdW-gap composition-gradient-driven SMT phenomenon also applies to MoSe2-xOx and MoTe2-xOx bilayers. The present work provides an alternative mechanism of SMT and demonstrates that the composition gradient across the vdW gap in the bilayer materials can be another degree of freedom to tune the band gaps without introducing extrinsic elements. Our findings will benefit the material design for small-scale and energy-efficient electronic devices.

Internal Polarization Modulation in Bi2 MoO6 for Photocatalytic Performance Enhancement under Visible-Light Illumination.

A built-in electric field from polarization inside polar photocatalysts could provide the driving force for photogenerated electrons and holes to move in opposite directions for better separation to improve their photocatalytic performance. The photocatalytic performance of a polar photocatalyst of Bi2 MoO6 has been enhanced through the precise control of its structure to increase internal polarization. DFT calculations predicted that a shortened crystal lattice parameter b in Bi2 MoO6 could induce larger internal polarization, which was achieved by the modulation of the pH of the reaction solution during a solvothermal synthetic process. A series of Bi2 MoO6 samples were created with reaction solutions of pH≈1, 4, and 8; the crystal lattice parameter b was found to decrease gradually with increasing solution pH. Accordingly, these Bi2 MoO6 samples demonstrated a gradually enhanced photocatalytic performance with decreasing crystal lattice parameter b, as demonstrated by the photocatalytic degradation of sulfamethoxazole/phenol and disinfection of Staphylococcus aureus bacteria under visible-light illumination due to improved photogenerated charge-carrier separation. This study demonstrates an innovative design strategy for materials to further enhance the photocatalytic performance of polar photocatalysts for a broad range of technical applications.

2D Intrinsic Ferromagnets from van der Waals Antiferromagnets.

Intrinsically ferromagnetic 2D semiconductors are essential and highly sought for nanoscale spintronics, but they can only be obtained from ferromagnetic bulk crystals, while the possibility to create 2D intrinsic ferromagnets from bulk antiferromagnets remains unknown. Herein on the basis of ab initio calculations, we demonstrate this feasibility with the discovery of intrinsic ferromagnetism in an emerging class of single-layer 2D semiconductors CrOX (CrOCl and CrOBr monolayers), which show robust ferromagnetic ordering, large spin polarization, and high Curie temperature. These 2D crystals promise great dynamical and thermal stabilities as well as easy experimental fabrication from their bulk antiferromagnets. The Curie temperature of 2D CrOCl is 160 K, which exceeds the record (155 K) of the most-studied dilute magnetic GaMnAs materials, and could be further enhanced by appropriate strains. Our study offers an alternative promising way to create 2D intrinsic ferromagnets from their antiferromagnetic bulk counterparts and also renders 2D CrOX monolayers great platform for future spintronics.

Adsorption and diffusion of hydrogen and oxygen in FCC-Co: a first-principles study.

Hydrogen and oxygen play an important role in the hydrogen embrittlement and oxidation of novel Co-based alloys with γ/γ' microstructure. In this study, the adsorption of hydrogen and oxygen atoms on the FCC-Co(111) surface and their diffusion behavior from the surface into the sub-layers and bulk have been investigated by means of first-principles calculations. It is observed that hydrogen and oxygen atoms prefer to adsorb on the fcc and hcp (threefold hollow) sites, respectively. The hydrogen atom can penetrate into the first and second sub-layers energetically, while it is not feasible for the oxygen atom as diffusion from the surface into the first sub-layer is more difficult. It is found that the calculated diffusion coefficients of hydrogen are in good agreement with the available experimental data. Finally, we briefly discuss the changes in total magnetic moment along the Oct-Tet-Oct diffusion path and the associated electronic structures. The present work is helpful to provide comprehensive guidance for the development and applications of novel Co-based alloys.

Metastable Stacking-Polymorphism in Ge2Sb2Te5.

Metastable rocksalt structured Ge2Sb2Te5 is the most widely used phase-change material for data storage, yet the atomic arrangements of which are still under debate. In this work, we have proposed metastable stacking-polymorphism in cubic Ge2Sb2Te5 based on first-principles calculations. Our results show that cubic Ge2Sb2Te5 is actually polymorphic, varying from randomly distributed vacancies to highly ordered vacancy layers; consequently, the electrical property varies between metallic and semiconducting. These different atomic stackings of cubic Ge2Sb2Te5 can be obtained at different experimental synthetic conditions. The concept of stacking-polymorphic Ge2Sb2Te5 provides important fundamentals for metastable Ge2Sb2Te5 and is useful for tuning the performance of the phase-change materials.

The effect of electron localization on the electronic structure and migration barrier of oxygen vacancies in rutile.

By applying the on-site Coulomb interaction (Hubbard term U) to the Ti d orbital, the influence of electron localization on the electronic structure as well as the transport of oxygen vacancies (VO) in rutile was investigated. With U = 4.5 eV, the positions of defect states in the bandgap were correctly reproduced. The unbonded electrons generated by taking out one neutral oxygen atom are spin parallel and mainly localized on the Ti atoms near VO, giving rise to a magnetic moment of 2 µB, in agreement with the experimental finding. With regard to the migration barrier of VO, surprisingly, we found that U = 4.5 eV only changed the value of the energy barrier by ±0.15 eV, depending on the diffusion path. The most probable diffusion path (along [110]) is the same as that calculated by using the traditional GGA functional. To validate the GGA + U method itself, a hybrid functional with a smaller supercell was used, and the trend of the more probable diffusion path was not changed. In this regard, the traditional GGA functional might still be reliable in the study of intrinsic-defect transportation in rutile. Analyzing the atomic distortion and density of states of the transition states for different diffusion paths, we found that the anisotropy of the diffusion could be rationalized according to the various atomic relaxations and the different positions of the valence bands relative to the Fermi level of the transition states.

Trapping of interstitial defects: filling the gap between the experimental measurements and DFT calculations.

We show that any impurity will slow the diffusion of oxygen in Nb. Using a first-principles plane-wave pseudopotential method and the supercell model, we calculated the interaction energies between substitutional atoms (SA) (X = Ti, V, Ta, Zr, and Hf) and interstitial oxygen in a Nb matrix. All impurities act as traps for oxygen: undersized SA (Ti and V) have strongest binding at the nearest octahedral interstice, while for oversized SA (Zr and Hf), the strongest trapping site is the second-nearest octahedral interstice. We evaluated the diffusion coefficients of O in the Nb-X alloys using kinetic Monte Carlo (KMC) modeling based in the transition state theory, using our calculated oxygen migration energies. From this, the effective (average) X-O interaction energies were extracted using the Oriani model (Oriani 1970 Acta Metall. 18 147-57). The effective X-O interaction energies are close to the strongest interaction energies between X and O obtained from the direct supercell calculations. The phenomenological effective diffusion barrier obtained from the KMC modeling is close to the energy difference between the most stable configuration and the highest saddle point along the diffusion path. Both results demonstrate that the weaker trapping site has negligible influence on the diffusion of O.