Dendrite-free Zn anode enabled by combining carbon nanoparticles hydrophobic layer with crystal face reconstruction toward high-performance Zn-ion battery.

Aqueous zinc ion batteries (ZIBs) have been considered promising energy storage systems due to their excellent electrochemical performance, environmental toxicity, high safety and low cost. However, uncontrolled dendrite growth and side reactions at the zinc anode have seriously hindered the development of ZIBs. Herein, we prepared the carbon nanoparticles layer coated zinc anode with (103) crystal plane preferential oriented crystal structure (denoted as C@RZn) by a facile one-step vapor deposition method. The preferential crystallographic orientation of (103) crystal plane promotes zinc deposition at a slight angle, effectively preventing the formation of Zn dendrites on the surface. In addition, the hydrophobic layer of carbon layer used as an inert physical barrier to prevent corrosion reaction and a buffer during volume changes, thus improving the reversibility of the zinc anode. As a result. the C@RZn anode achieves a stable cycle performance of more than 3000 h at 1 mA cm-2 with CE of 99.77 % at 5 mA cm-2. The full battery with C@RZn anode and Mn-doped V6O13 (MVO) cathode show stability for 5000 cycles at the current density of 5 A g-1. This work provides a new approach for the design of multifunctional interfaces for Zn anode.

Structural Design of π-d Conjugated TMx B3 N3 S6 (x=2, 3) Monolayer Toward Electrocatalytic Ammonia Synthesis.

Single-atom catalysts (SACs) have attracted wide attention to be acted as potential electrocatalysts for nitrogen reduction reaction (NRR). However, the coordination environment of the single transition metal (TM) atoms is essential to the catalytic activity for NRR. Herein, we proposed four types of 3-, 4-coordinated and π-d conjugated TMx B3 N3 S6 (x=2, 3, TM=Ti, V, Cr, Mn, Fe, Zr, Nb, Mo, Tc, Ru, Hf, Ta, W, Re and Os) monolayers for SACs. Based on density functional theory (DFT) calculations, I-TM2 B3 N3 S6 and III-TM3 B3 N3 S6 are the reasonable 3-coordinated and 4-coordinated structures screening by structure stable optimizations, respectively. Next, the structural configurations, electronic properties and catalytic performances of 30 kinds of the 3-coordinated I-TM2 B3 N3 S6 and 4-coordinated III-TM3 B3 N3 S6 monolayers with different single transition metal atoms were systematically investigated. The results reveal that B3 N3 S6 ligand is an ideal support for TM atoms due to existence of strong TM-S bonds. The 3-coordinated I-V2 B3 N3 S6 is the best SAC with the low limiting potential (UL ) of -0.01 V, excellent stability (Ef =-0.32 eV, Udiss =0.02 V) and remarkable selectivity characteristics. This work not only provides novel π-d conjugated SACs, but also gives theoretical insights into their catalytic activities and offers reference for experimental synthesis.

3D hierarchically fractal structure stabilized anode for achieving long-term cycle life of aqueous Zn-ion batteries.

The cycling stability of aqueous Zn-ion battery (AZIB) is a serious issue for their successful application, mainly due to the considerable growth of Zn dendrites and the existence of side effects during operation. Herein, the hierarchically three-dimensional (3D) fractal structure of the ZnO/Zn/CuxO@Cu (ZZCC) anode is prepared by a two-step process, where CuxO nanowires are prepared on Cu foam by thermal oxidation method and Zn layer and ZnO surface are formed by plating. This fractal structure increases the electrodynamic surfaces and reduces the local current density, which can regulate Zn plating and inhibit dendritic growth and side effects. Apparently, the symmetric ZZCC-based cell shows a long-term operation time of 3000 h at 1 mA cm-2 with 1 mAh cm-2, and an operation time of more than 1000 h with a discharge depth of 15.94%. Compared with the bare Zn foil anode, the AZIB assembled with the composite of Mn-doped vanadium oxide and reduced graphene oxide cathode and ZZCC anode (MnVO@rGO//ZZCC) exhibits significantly improved cyclability (i.e. with 88.5% capacity retention) and achieves a Coulomb efficiency of 99.4% at 2 A g-1. This hierarchically 3D structure strategy to design anodes with superior cyclic stability contributes to the next generation of secure energy.

Graph-guided deep hashing networks for similar patient retrieval.

With the rapid growth and widespread application of electronic health records (EHRs), similar patient retrieval has become an important task for downstream clinical decision support such as diagnostic reference, treatment planning, etc. However, the high dimensionality, large volume, and heterogeneity of EHRs pose challenges to the efficient and accurate retrieval of patients with similar medical conditions to the current case. Several previous studies have attempted to alleviate these issues by using hash coding techniques, improving retrieval efficiency but merely exploring underlying characteristics among instances to preserve retrieval accuracy. In this paper, drug categories of instances recorded in EHRs are regarded as the ground truth to determine the pairwise similarity, and we consider the abundant semantic information within such multi-labels and propose a novel framework named Graph-guided Deep Hashing Networks (GDHN). To capture correlation dependencies among the multi-labels, we first construct a label graph where each node represents a drug category, then a graph convolution network (GCN) is employed to derive the multi-label embedding of each instance. Thus, we can utilize the learned multi-label embeddings to guide the patient hashing process to obtain more informative and discriminative hash codes. Extensive experiments have been conducted on two datasets, including a real-world dataset concerning IgA nephropathy from Peking University First Hospital, and a publicly available dataset from MIMIC-III, compared with traditional hashing methods and state-of-the-art deep hashing methods using three evaluation metrics. The results demonstrate that GDHN outperforms the competitors at different hash code lengths, validating the superiority of our proposal.

CEHMR: Curriculum learning enhanced hierarchical multi-label classification for medication recommendation.

The medication recommendation (MR) or medication combination prediction task aims to predict effective prescriptions given accurate patient representations derived from electronic health records (EHRs), which contributes to improving the quality of clinical decision-making, especially for patients with multi-morbidity. Although in recent years deep learning technology has achieved great success in MR, the performance of current multi-label based MR solutions is unsatisfactory. They mainly focus on improving the patient representation module and modeling the medication label dependencies such as drug-drug interaction (DDI) correlation and co-occurrence relationship. However, the hierarchical dependency among medication labels and diversity of difficulty among MR training examples lack sufficient consideration. In this paper, we propose a framework of Curriculum learning Enhanced Hierarchical multi-label classification for MR (CEHMR). Motivated by the category hierarchy of medications which organizes standard medication codes in a hierarchical structure, we utilize it to provide more trustworthy prior knowledge for modeling label dependency. Specifically, we design a hierarchical multi-label classifier with a learnable gate fusion layer, to simultaneously capture the level-independent (local) and level-dependent (global) hierarchical information in the medication hierarchy. In addition, to overcome the diversity of training example difficulties, and progressively achieve a smoother training process, we introduce a bootstrap-based curriculum learning strategy. Hence, the example difficulty can be measured based on the predictive performance of the MR model, and then all training examples would be retrained from easy to hard under the guidance of a predefined training scheduler. Experiments on the real-world medical MIMIC-III database demonstrate that the proposed framework can achieve state-of-the-art performance compared with seven representative baselines, and extensive ablation studies validate the effectiveness of each component of CEHMR.

Interface engineering with porous graphene as deposition regulator of stable Zn metal anode for long-life Zn-ion capacitor.

The zinc (Zn) dendrite accumulation leads to poor Coulombic efficiency, continuously failing life and severe safety risks, which seriously impede the commercial application of Zn ion capacitors (ZICs). Herein, an interface engineering is proposed for the Zn metal anode to restrain the dendrite by using porous flame reduced graphene oxide (FRGO) as the ex-situ protective and regulated layer to induce the Zn crystal growth and restricts the side reactions. The FRGO possesses extensive nanoscale pores and zincophilic oxygen-containing functional groups, which can absorb Zn2+ and nucleate preferentially on the surface of FRGO, then induce the growth of Zn parallel to the graphene sheet by matching the basal (002) plane of metallic Zn to minimize lattice strain. As a result, it eliminates the tip effect and achieves the deposited Zn with a uniform and flat surface. Therefore, The FRGO on the Zn (FRGO@Zn) anode significantly reduces the nucleation overpotential and improves the cycling life during the plating/stripping process. Notably, FRGO@Zn based ZIC can achieve 91.0% capacity retention after more than 20,000 cycles at 5 A g-1, and its capacity and maximum energy density are 150.6 mAh g-1 and 118.8 Wh kg-1, respectively. This interface engineering of FRGO for the Zn metal anode has excellent application potential and theoretical guidance in the ZICs field.

Attentional disengagement effect based on relevant features.

In visual search tasks, distractors similar to the target can attract our attention and affect the speed of attentional disengagement. The attentional disengagement refers to shifting attention away from stimuli that are not relevant to the task. Previous studies mainly focused on the attentional disengagement of one feature dimension. However, the mechanisms of different feature dimensions on attentional disengagement in single and conjunction visual search remain unclear. In the current study, we adopted the oculomotor disengagement paradigm and used saccade latency as an indicator to explore the effects of different feature dimensions of center stimuli on attentional disengagement. In both single and conjunction feature search tasks, participants began each search by fixating on a center stimulus that appeared simultaneously with search display but would not be the target. Participants were instructed to ensure the first saccade to the target location. In Experiments 1A (single feature search) and 1B (conjunction feature search), we found that the attentional disengagement was significantly delayed or accelerated when center stimuli shared color features with the target or salient distractor, but not in shape feature. Moreover, we found that the difference between the two feature dimensions might be caused by their different search difficulty (Experiment 1C). Therefore, in Experiment 2, we matched the difficulty of searching for color and shape tasks before exploring whether there were differences in the effects of different feature dimensions on attentional disengagement. However, the results in Experiment 2 were similar to those in Experiment 1A, indicating that the different effects of feature dimensions on attentional disengagement were caused by feature asymmetry. Therefore, in Experiment 3, we improved the salient discernibility of shape dimension and matched color search to it. The results showed that although the attentional disengagement was delayed in shape dimension, it was still smaller than that in color dimension. Our results supported that goal-oriented attention sets were the main cause of delayed attentional disengagement. By series of experiments, we found that the utilization of different feature dimensions was associated with task difficulty and the features asymmetry in both single and conjunction visual search.

A Synergistic Strategy of Organic Molecules Introduced a High Zn2+ Flux Solid Electrolyte Interphase for Stable Aqueous Zinc-Ion Batteries.

Aqueous rechargeable zinc-ion batteries (ARZIBs) are considered as attractive candidates for the next generation of high-safety and low-cost energy storage in large-scale power grids. However, challenges such as the dendrites and the corrosion on the zinc (Zn) surface result in short battery life and low reversibility of Zn plating/stripping. In this work, a method of preconditioning of a zinc anode in hybrid electrolytes (based on poly(ethylene glycol)-200 and H2O) to form a solid electrolyte interphase (SEI) that prevents anode corrosion and dendrites is proposed. Though surface composition analysis and density functional theory calculation, this SEI has dense organic and inorganic components due to the induction of organic molecules and anions and has rapid kinetic and high-throughput properties for the transport of zinc ions. As a result, the SEI-modified Zn anode can maintain a low-voltage hysteresis stable cycle for more than 1600 h in aqueous electrolyte. The anode also exhibits impressive reversibility with a high Coulomobic efficiency of 99.23% over 1300 cycles. Furthermore, the ARZIB encapsulated by this anode and Mn-doped V6O13 cathode enables an outstanding electrochemical stability (181.8 mAh g-1 after 800 cycles at room temperature, 102.2 mAh g-1 after 1000 cycles at -15 °C). This work provides an intriguing idea for the stability maintenance of the anode for ARZIBs or other metal-ion batteries.

Multi-level semantic fusion network for Chinese medical named entity recognition.

Medical named entity recognition (MNER) is a fundamental component of understanding the unstructured medical texts in electronic health records, and it has received widespread attention in both academia and industry. However, the previous approaches of MNER do not make full use of hierarchical semantics from morphology to syntactic relationships like word dependency. Furthermore, extracting entities from Chinese medical texts is a more complex task because it usually contains for example homophones or pictophonetic characters. In this paper, we propose a multi-level semantic fusion network for Chinese medical named entity recognition, which fuses semantic information on morphology, character, word and syntactic level. We take radical as morphology semantic, pinyin and character dictionary as character semantic, word dictionary as word semantic, and these semantic features are fused by BiLSTM to get the contextualized representation. Then we use a graph neural network to model word dependency as syntactic semantic to enhance the contextualized representation. The experimental results show the effectiveness of the proposed model on two public datasets and robustness in real-world scenarios.

Hierarchically nanostructured Zn0.76C0.24S@Co(OH)2 for high-performance hybrid supercapacitor.

It is a great challenge to achieve both high specific capacity and high energy density of supercapacitors by designing and constructing hybrid electrode materials through a simple but effective process. In this paper, we proposed a hierarchically nanostructured hybrid material combining Zn0.76Co0.24S (ZCS) nanoparticles and Co(OH)2 (CH) nanosheets using a two-step hydrothermal synthesis strategy. Synergistic effects between ZCS nanoparticles and CH nanosheets result in efficient ion transports during the charge-discharge process, thus achieving a good electrochemical performance of the supercapacitor. The synthesized ZCS@CH hybrid exhibits a high specific capacity of 1152.0 C g-1 at a current density of 0.5 A g-1 in 2 M KOH electrolyte. Its capacity retention rate is maintained at âˆ¼ 70.0% when the current density is changed from 1 A g-1 to 10 A g-1. A hybrid supercapacitor (HSC) assembled from ZCS@CH as the cathode and active carbon (AC) as the anode displays a capacitance of 155.7 F g-1 at 0.5 A g-1, with a remarkable cycling stability of 91.3% after 12,000cycles. Meanwhile, this HSC shows a high energy density of 62.5 Wh kg-1 at a power density of 425.0 W kg-1, proving that the developed ZCS@CH is a promising electrode material for energy storage applications.

MnCo2O4/Ni3S4 nanocomposite for hybrid supercapacitor with superior energy density and long-term cycling stability.

MnCo2O4 is regarded as a good electrode material for supercapacitor due to its high specific capacity and good structural stability. However, its poor electrical conductivity limits its wide-range applications. To solve this issue, we integrated the MnCo2O4 with Ni3S4, which has a good electrical conductivity, and synthesized a MnCo2O4/Ni3S4 nanocomposite using a two-step hydrothermal process. Comparing with individual MnCo2O4 and Ni3S4, the MnCo2O4/Ni3S4 nanocomposite showed a higher specific capacity and a better cycling stability as the electrode for the supercapacitor. The specific capacity value of the MnCo2O4/Ni3S4 electrode was 904.7 C g-1 at 1 A g-1 with a potential window of 0-0.55 V. A hybrid supercapacitor (HSC), assembled using MnCo2O4/Ni3S4 and active carbon as the cathode and anode, respectively, showed a capacitance of 116.4 F g-1 at 1 A g-1, and a high energy density of 50.7 Wh kg-1 at 405.8 W kg-1. Long-term electrochemical stability tests showed an obvious increase of the HSC's capacitance after 5500 charge/discharge cycles, reached a maximum value of ∼162.7% of its initial value after 25,000 cycles, and then remained a stable value up to 64,000 cycles. Simultaneously, its energy density was increased to 54.2 Wh kg-1 at 380.3 W kg-1 after 64,000 cycles.

Enhanced electrochemical performance of CuCo2S4/carbon nanotubes composite as electrode material for supercapacitors.

CuCo2S4 is regarded as a promising electrode material for supercapacitor, but has inferior conductivity and poor cyclic stability which restrict its wide-range applications. In this work, hierarchically hybrid composite of CuCo2S4/carbon nanotubes (CNTs) was synthesized using a facile hydrothermal and sulfuration process. The embedded CNTs in the CuCo2S4 matrix provided numerous effective paths for electron transfer and ion diffusion, and thus promoted the faradaic reactions of the CuCo2S4 electrode in the energy storage processes. The CuCo2S4/CNTs-3.2% electrode exhibited a significantly increased specific capacitance of 557.5 F g-1 compared with those of the pristine CuCo2S4 electrode (373.4 F g-1) and CuO/Co3O4/CNTs-3.2% electrode (356.5 F g-1) at a current density of 1 A g-1. An asymmetric supercapacitor (ASC) was assembled using the CuCo2S4/CNTs-3.2% as the positive electrode and the active carbon as the negative electrode, which exhibited an energy density of 23.2 Wh kg-1 at a power density of 402.7 W kg-1. Moreover, the residual specific capacitance of this ASC device retained 85.7% of its original value after tested for 10,000 cycles, indicating its excellent cycle stability.

Ultrafast Response/Recovery and High Selectivity of the H2S Gas Sensor Based on α-Fe2O3 Nano-Ellipsoids from One-Step Hydrothermal Synthesis.

Ultrafast response/recovery and high selectivity of gas sensors are critical for real-time and online monitoring of hazardous gases. In this work, α-Fe2O3 nano-ellipsoids were synthesized using a facile one-step hydrothermal method and investigated as highly sensitive H2S-sensing materials. The nano-ellipsoids have an average long-axis diameter of 275 nm and an average short-axis diameter of 125 nm. H2S gas sensors fabricated using the α-Fe2O3 nano-ellipsoids showed excellent H2S-sensing performance at an optimum working temperature of 260 °C. The response and recovery times were 0.8 s/2.2 s for H2S gas with a concentration of 50 ppm, which are much faster than those of H2S gas sensors reported in the literature. The α-Fe2O3 nano-ellipsoid-based sensors also showed high selectivity to H2S compared to other commonly investigated gases including NH3, CO, NO2, H2, CH2Cl2, and ethanol. In addition, the sensors exhibited high-response values to different concentrations of H2S with a detection limit as low as 100 ppb, as well as excellent repeatability and long-term stability.