General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian.

The combination of deep learning and ab initio calculation has shown great promise in revolutionizing future scientific research, but how to design neural network models incorporating a priori knowledge and symmetry requirements is a key challenging subject. Here we propose an E(3)-equivariant deep-learning framework to represent density functional theory (DFT) Hamiltonian as a function of material structure, which can naturally preserve the Euclidean symmetry even in the presence of spin-orbit coupling. Our DeepH-E3 method enables efficient electronic structure calculation at ab initio accuracy by learning from DFT data of small-sized structures, making the routine study of large-scale supercells (>104 atoms) feasible. The method can reach sub-meV prediction accuracy at high training efficiency, showing state-of-the-art performance in our experiments. The work is not only of general significance to deep-learning method development but also creates opportunities for materials research, such as building a Moiré-twisted material database.

Controlled Growth of Wafer-Scale Transition Metal Dichalcogenides with a Vertical Composition Gradient for Artificial Synapses with High Linearity.

Artificial synapses are promising for dealing with large amounts of data computing. Great progress has been made recently in terms of improving the on/off current ratio, the number of states, and the energy efficiency of synapse devices. However, the nonlinear weight update behavior of a synapse caused by the uncertain direction of the conductive filament leads to complex weight modulation, which degrades the delivery accuracy of information. Here we propose a strategy to improve the weight update behavior of synapses using chemical-vapor-deposition-grown transition metal dichalcogenides (TMDCs) with a vertical composition gradient, where the sulfur concentration decreases gradually along the thickness direction of TMDCs and thus forms a certain direction of the conduction filament for synapse devices. It is worth noting that the devices show an excellent linear conductance of potentiation and depression with a high linearity of 0.994 (surpassing most state-of-the-art synapses), have a large number of states, and are able to fabricate synapse arrays with wafer-scale. Furthermore, the devices based on the TMDCs with the vertical composition gradient exhibit an asymmetric feature of potentiation and depression behaviors with high linearity and follow the simulated linear Leaky ReLU function, resulting in a high recognition accuracy of 94.73%, which overcomes the unreliability issue in the Sigmoid function due to the vanishing gradient phenomenon. This study not only provides a universal method to grow TMDCs with a vertical composition gradient but also contributes to exploring highly linear synapses toward neuromorphic computing.

Nanoscale Control of One-Dimensional Confined States in Strongly Correlated Homojunctions.

Construction of lateral junctions is essential to generate one-dimensional (1D) confined potentials that can effectively trap quasiparticles. A series of remarkable electronic phases in one dimension, such as Wigner crystallization, are expected to be realized in such junctions. Here, we demonstrate that we can precisely tune the 1D-confined potential with an in situ manipulation technique, thus providing a dynamic way to modify the correlated electronic states at the junctions. We confirm the existence of 1D-confined potential at the homojunction of two single-layer 1T-NbSe2 islands. Such potential is structurally sensitive and shows a nonmonotonic function of their interspacing. Moreover, there is a change of electronic properties from the correlated insulator to the generalized 1D Wigner crystallization while the confinement becomes strong. Our findings not only establish the capability to fabricate structures with dynamically tunable properties, but also pave the way toward more exotic correlated systems in low dimensions.

Deep-learning density functional theory Hamiltonian for efficient ab initio electronic-structure calculation.

The marriage of density functional theory (DFT) and deep-learning methods has the potential to revolutionize modern computational materials science. Here we develop a deep neural network approach to represent the DFT Hamiltonian (DeepH) of crystalline materials, aiming to bypass the computationally demanding self-consistent field iterations of DFT and substantially improve the efficiency of ab initio electronic-structure calculations. A general framework is proposed to deal with the large dimensionality and gauge (or rotation) covariance of the DFT Hamiltonian matrix by virtue of locality, and this is realized by a message-passing neural network for deep learning. High accuracy, high efficiency and good transferability of the DeepH method are generally demonstrated for various kinds of material system and physical property. The method provides a solution to the accuracy-efficiency dilemma of DFT and opens opportunities to explore large-scale material systems, as evidenced by a promising application in the study of twisted van der Waals materials.

Two-dimensional magnetic materials: structures, properties and external controls.

Since the discovery of intrinsic ferromagnetism in atomically thin Cr2Gr2Te6 and CrI3 in 2017, research on two-dimensional (2D) magnetic materials has become a highlighted topic. Based on 2D magnetic materials and their heterostructures, exotic physical phenomena at the atomically thin limit have been discovered, such as the quantum anomalous Hall effect, magneto-electric multiferroics, and magnon valleytronics. Furthermore, magnetism in these ultrathin magnets can be effectively controlled by external perturbations, such as electric field, strain, doping, chemical functionalization, and stacking engineering. These attributes make 2D magnets ideal platforms for fundamental research and promising candidates for various spintronic applications. This review aims at providing an overview of the structures, properties, and external controls of 2D magnets, as well as the challenges and potential opportunities in this field.

Pressure-Dependent Intermediate Magnetic Phase in Thin Fe3GeTe2 Flakes.

We investigated the evolution of ferromagnetism in layered Fe3GeTe2 flakes under different pressures and temperatures using in situ magnetic circular dichroism (MCD) spectroscopy. We found that the rectangular shape of the hysteresis loop under an out-of-plane magnetic field sweep can be sustained below 7 GPa. Above that pressure, an intermediate state appears in the low-temperature region signaled by an 8-shaped skewed hysteresis loop. Meanwhile, the coercive field and Curie temperature decrease with increasing pressures, implying the decrease of the exchange interaction and the magneto-crystalline anisotropy under pressures. The intermediate phase has a labyrinthine domain structure, which is attributed to the increase of the ratio of exchange interaction to magneto-crystalline anisotropy based on Jagla's theory. Moreover, our calculations reveal a weak structural transition around 6 GPa that corresponds to a significant change in the FeI-FeI bond length, which has strong influences on magnetic interaction. Detailed analysis on exchange interaction and magneto-crystalline anisotropy with pressure shows a consistent trend with experiments.

Electric Field-Modulated Magnetic Phase Transition in van der Waals CrI3 Bilayers.

Two-dimensional van der Waals (vdW) magnetic materials are well-recognized milestones toward nanostructured spintronics. An interesting example is CrI3; its magnetic states can be modulated electrically, allowing spintronics applications that are highly compatible with electronics technologies. Here, we report the electric field alone induces the interlayer antiferromagnetic-to-ferromagnetic (AFM-to-FM) phase transition in CrI3 bilayers with critical field as low as 0.12 V/Å. The AFM-FM energy difference ΔE increases with electric field and is closely related to the field-induced on-site energy difference defined as the splitting between the electronic states of the two vdW layers. Our tight-binding model fits closely with ΔE as a function of electric field and gives a consistent estimation for orbital hopping, exchange splitting, and crystal field splitting. Furthermore, a CrI3-based spin field-effect device is suggested with the spin current switched on and off solely by the electric field. These findings not only reveal the physics underlying the transition but also provide guidelines for future discovery and design.

Chemical Vapor Deposition Grown Large-Scale Atomically Thin Platinum Diselenide with Semimetal-Semiconductor Transition.

Among two-dimensional (2D) transition-metal dichalcogenides (TMDCs), platinum diselenide (PtSe2) stands in a distinct place due to its fancy transition from type-II Dirac semimetal to semiconductor with a thickness variation from bulk to monolayer (1 ML) and the related versatile applications especially in mid-infrared detectors. However, achieving atomically thin PtSe2 is still a challenging issue. Herein, we have designed a facile chemical vapor deposition (CVD) method to achieve the synthesis of atomically thin 1T-PtSe2 on an electrode material of Au foil. Thanks to the high crystalline quality, we have confirmed the complete transition from semimetal to semiconductor from trilayer (3 ML) to 1 ML 1T-PtSe2. More importantly, we have found that such atomically thin 1T-PtSe2 can serve as perfect electrocatalysts, featured with a record high hydrogen evolution reaction (HER) efficiency (comparable to traditional Pt catalyst). Our work is helpful toward the large-scale synthesis, exotic physical property exploration, and intriguing application development of atomically thin TMDCs.

Simultaneous Production and Functionalization of Boron Nitride Nanosheets by Sugar-Assisted Mechanochemical Exfoliation.

Due to their extraordinary properties, boron nitride nanosheets (BNNSs) have great promise for many applications. However, the difficulty of their efficient preparation and their poor dispersibility in liquids are the current factors that limit this. A simple yet efficient sugar-assisted mechanochemical exfoliation (SAMCE) method is developed here to simultaneously achieve their exfoliation and functionalization. This method has a high actual exfoliation yield of 87.3%, and the resultant BNNSs are covalently grafted with sugar (sucrose) molecules, and are well dispersed in both water and organic liquids. A new mechanical force-induced exfoliation and chemical grafting mechanism is proposed based on experimental and density functional theory investigations. Thanks to the good dispersibility of the nanosheets, flexible and transparent BNNS/poly(vinyl alcohol) (PVA) composite films with multifunctionality is fabricated. Compared to pure PVA films, the composite films have a remarkably improved tensile strength and thermal dissipation capability. Noteworthy, they are flame retardant and can effectively block light from the deep blue to the UV region. This SAMCE production method has proven to be highly efficient, green, low cost, and scalable, and is extended to the exfoliation and functionalization of other two-dimensional (2D) materials including MoS2 , WS2 , and graphite.

Biotemplating Growth of Nepenthes-like N-Doped Graphene as a Bifunctional Polysulfide Scavenger for Li-S Batteries.

The practical application of lithium-sulfur (Li-S) batteries is hindered by their poor cycling stabilities that primarily stem from the "shuttle" of dissolved lithium polysulfides. Here, we develop a nepenthes-like N-doped hierarchical graphene (NHG)-based separator to realize an efficient polysulfide scavenger for Li-S batteries. The 3D textural porous NHG architectures are realized by our designed biotemplating chemical vapor deposition (CVD) approach via the employment of naturally abundant diatomite as the growth substrate. Benefiting from the high surface area, devious inner-channel structure, and abundant nitrogen doping of CVD-grown NHG frameworks, the derived separator favorably synergizes bifunctionality of physical confinement and chemical immobilization toward polysulfides, accompanied by smooth lithium ion diffusions. Accordingly, the batteries with the NHG-based separator delivers an initial capacity of 868 mAh g-1 with an average capacity decay of only 0.067% per cycle at 2 C for 800 cycles. A capacity of 805 mAh g-1 can further be achieved at a high sulfur loading of ∼7.2 mg cm-2. The present study demonstrates the potential in constructing high-energy and long-life Li-S batteries upon separator modification.

Vertical 1T-TaS2 Synthesis on Nanoporous Gold for High-Performance Electrocatalytic Applications.

2D metallic TaS2 is acting as an ideal platform for exploring fundamental physical issues (superconductivity, charge-density wave, etc.) and for engineering novel applications in energy-related fields. The batch synthesis of high-quality TaS2 nanosheets with a specific phase is crucial for such issues. Herein, the successful synthesis of novel vertically oriented 1T-TaS2 nanosheets on nanoporous gold substrates is reported, via a facile chemical vapor deposition route. By virtue of the abundant edge sites and excellent electrical transport property, such vertical 1T-TaS2 is employed as high-efficiency electrocatalysts in the hydrogen evolution reaction, featured with rather low Tafel slopes ≈67-82 mV dec-1 and an ultrahigh exchange current density ≈67.61 µA cm-2 . The influence of phase states of 1T- and 2H-TaS2 on the catalytic activity is also discussed with the combination of density functional theory calculations. This work hereby provides fundamental insights into the controllable syntheses and electrocatalytic applications of vertical 1T-TaS2 nanosheets achieved through the substrate engineering.

Half-Metallicity in Co-Doped WSe2 Nanoribbons.

The recent development of two-dimensional transition-metal dichalcogenides in electronics and optoelelectronics has triggered the exploration in spintronics, with high demand in search for half-metallicity in these systems. Here, through density functional theory (DFT) calculations, we predict robust half-metallic behaviors in Co-edge-doped WSe2 nanoribbons (NRs). With electrons partially occupying the antibonding state consisting of Co 3dyz and Se 4pz orbitals, the system becomes spin-polarized due to the defect-state-induced Stoner effect and the strong exchange splitting eventually gives rise to the half-metallicity. The half-metal gap reaches 0.15 eV on the DFT generalized gradient approximation level and increases significantly to 0.67 eV using hybrid functional. Furthermore, we find that the half-metallicity sustains even under large external strain and relatively low edge doping concentration, which promises the potential of such Co-edge-doped WSe2 NRs in spintronics applications.