Search

The ability to freeze and stabilize reaction intermediates in their metastable states and obtain their structural and chemical information with high spatial resolution would be very powerful to unravel the fundamentals in many important materials technologies such as catalysis and batteries. Here, we develop an electrified operando-freezing methodology for the first time to preserve these metastable states under electrochemical reaction conditions for cryogenic electron microscopy (cryo-EM) imaging and spectroscopy. Using Cu catalysts for CO2 reduction as a model system, we observe restructuring of the Cu catalyst in a CO2 atmosphere while the same catalyst remains intact in an air atmosphere at the nanometer scale. Furthermore, we discover the existence of single valance Cu (1+) state and C-O bonding at the electrified liquid-solid interface of the operando-frozen samples, which are key reaction intermediates that traditional ex situ measurements fail to detect. This work highlights our novel technique to study the local structure and chemistry of electrified liquid-solid interfaces, which has broad impact for many electrochemical reactions.

Summary One in three people worldwide does not have access to safe drinking water. Notably, heavy-metal ions (HMIs) are major water pollutants threatening human health because of their severe toxicity, even at trace levels. Efficient HMI detection thus plays a major defense against metal poisoning by enabling early pollution warning and efficient regulatory enforcement. However, it remains a formidable challenge to accurately detect these pollutants on site at ultratrace levels in a cost- or time-effective manner. Here, we introduce an efficient, portable sensor to concurrently quantify five different HMIs down to the sub-nanomolar level by sulfiding them on a superhydrophobic surface. Sulfidation serves as a colorimetric reaction while the superhydrophobic surface concentrates analytes for sensitive visual detection. Our superhydrophobic concentrator (SPOT) sensor can be made portable by being integrated with a smartphone application to quantify HMIs in

Summary Cathode electrolyte interphase (CEI), the intimate coating layer formed on the positive electrode, has been thought to be critical. However, many aspects of CEI remain unclear. This originates from the lack of effective tools to characterize structural and chemical properties of these sensitive interphases at nanoscale. Here, we develop a protocol to preserve the native state and directly visualize the interface on the positive electrode using cryogenic electron microscopy. We find that under normal operation conditions, there does not exist an intimate coating layer at the single-particle level in carbonate-based electrolyte. However, upon brief external electrical shorting, a solid-electrolyte interphase, which usually forms on anodes, could form on cathodes and be electrochemically converted into a stable, conformal CEI in situ. The conformal CEI helps improve Coulombic efficiency and overall capacity retention of the battery. This generates a different perspective of CEI in commercial carbonate-based electrolytes than previously understood.

Gas-phase heterogeneous catalysis is a process spatially constrained on the two-dimensional surface of a solid catalyst. Here, we introduce a new toolkit to open up the third dimension. We discovered that the activity of a solid catalyst can be dramatically promoted by covering its surface with a nanoscale-thin layer of liquid electrolyte while maintaining efficient delivery of gas reactants, a strategy we call three-phase catalysis. Introducing the liquid electrolyte converts the original surface catalytic reaction into an electrochemical pathway with mass transfer facilitated by free ions in a three-dimensional space. We chose the oxidation of formaldehyde as a model reaction and observed a 25000-times enhancement in the turnover frequency of Pt in three-phase catalysis as compared to conventional heterogeneous catalysis. We envision three-phase catalysis as a new dimension for catalyst design and anticipate its applications in more chemical reactions from pollution control to the petrochemical industry.

Significance Metallic lithium plating on the graphite anode is a predominant cause for capacity decays during the fast charging of lithium-ion batteries. This work studies the lithium-plating phenomenon in a previously neglected thermodynamic perspective, taking into account practical temperature distributions within batteries. We show that elevated temperatures could enhance the equilibrium potential of Li 0 /Li + , making local lithium plating more thermodynamically favorable. Furthermore, lithium-plating patterns are correlated with temperature heterogeneities, confirming the preferential lithium plating at high-temperature regions due to both kinetic and thermodynamic origins. These findings provide possible explanations of the heterogeneous lithium-plating morphology, deepen the understandings on the lithium plating phenomenon, and will guide future strategies to realize the extreme fast charging of lithium-ion batteries.

The presence of organic contaminants in wastewater poses considerable risks to the health of both humans and ecosystems. Although advanced oxidation processes that rely on highly reactive radicals to destroy organic contaminants are appealing treatment options, substantial energy and chemical inputs limit their practical applications. Here we demonstrate that Cu single atoms incorporated in graphitic carbon nitride can catalytically activate H2O2 to generate hydroxyl radicals at pH 7.0 without energy input, and show robust stability within a filtration device. We further design an electrolysis reactor for the on-site generation of H2O2 from air, water and renewable energy. Coupling the single-atom catalytic filter and the H2O2 electrolytic generator in tandem delivers a wastewater treatment system. These findings provide a promising path toward reducing the energy and chemical demands of advanced oxidation processes, as well as enabling their implementation in remote areas and isolated communities.

Speaker recognition is a crucial bio-identification technology, which is extensively used in our daily life. With the development of deep learning, convolutional neural networks (CNNs) are applied to speaker recognition tasks given their excellent performance. However, in real life, speaker recognition systems are frequently deployed on end-devices. Therefore, while obtaining recognition accuracy, the model of speaker recognition is expected to be as simple as possible. Inspired by 1-max pooling CNN and Gaussian mixture model-universal background model (GMM-UBM), this study proposes a one dimension convolutional neural networks (1D CNN) on the basis of original 2D CNN. The proposed model reduces the computational complexity of ResNet20 by 64% and the amount of parameters by 53%. In comparison with the original ResNet20 models, the recognition accuracy will be reduced by about one percent on the 15s data set. Then, on the basis of the 1D CNN, we propose a pyramid layer-folding pipeline structure and implement it on the Xilinx VC709 platform. According to the time-dimension partition, the proposed pyramid pipeline structure can process speech data of various lengths. Moreover, our accelerator is 5.1× faster on 3s dataset and 6.8× quicker on 15s dataset than those of the CPU platform.

Manipulating liquids with tunable shape and optical functionalities in real time is important for electroactive flow devices and optoelectronic devices, but remains a great challenge. Here, we demonstrate electrotunable liquid sulfur microdroplets in an electrochemical cell. We observe electrowetting and merging of sulfur droplets under different potentiostatic conditions, and successfully control these processes via selective design of sulfiphilic/sulfiphobic substrates. Moreover, we employ the electrowetting phenomena to create a microlens based on the liquid sulfur microdroplets and tune its characteristics in real time through changing the shape of the liquid microdroplets in a fast, repeatable, and controlled manner. These studies demonstrate a powerful in situ optical battery platform for unraveling the complex reaction mechanism of sulfur chemistries and for exploring the rich material properties of the liquid sulfur, which shed light on the applications of liquid sulfur droplets in devices such as microlenses, and potentially other electrotunable and optoelectronic devices., Manipulating liquids with tunable shape and optical functionalities in real time remains a great challenge. Here, the authors demonstrate electrotunable liquid sulfur microdroplets in an electrochemical cell and tune its characteristics in a fast, repeatable, and controlled manner.

Self-grooming is an innate, cephalo-caudal progression of body cleaning behaviors that are found in normal rodents but exhibit repetitive and stereotyped patterns in several mouse models, such as autism spectrum disorders (ASDs). It is also recognized as a marker of stress and anxiety. Mice with Shank3B gene knockout (KO) exhibit typical ASD-like behavioral abnormalities, including stereotyped self-grooming and increased levels of anxiety. However, the exact relationship between anxiety and stereotyped self-grooming in certain types of animal models is not clear. We selected three animal models with high anxiety to compare their self-grooming parameters. First, we confirmed that Shank3B KO mice (ASD model), acute restraint stress mouse model (stress model), and chronic inflammatory pain mouse model (pain model) all showed increased anxiety levels in the open field test (OFT) and elevated plus maze (EPM). We found that only the ASD model and the stress model produced increased total grooming duration. The pain model only exhibited an increasing trend of mean self-grooming duration. We used the grooming analysis algorithm to examine the self-grooming microstructure and assess the cephalo-caudal progression of grooming behavior. The results showed distinct self-grooming microstructures in these three models. The anxiolytic drug diazepam relieved the anxiety level and the total time of grooming in the ASD and stress models. The grooming microstructure was not restored in Shank3B KO mice but was partially relieved in the stress model, which suggested that anxiety aggravated stereotyped self-grooming duration but not the grooming microstructure in the ASD mouse model. Our results indicated that stereotyped behavior and anxiety may be shared by separate, but interacting, neural circuits in distinct disease models, which may be useful to understand the mechanisms and develop potential treatments for stereotyped behaviors and anxiety.

Background: To evaluate the feasibility, efficacy, and safety of minimally invasive pedicle screw (MIPS) fixation, including the fractured vertebra, combined with percutaneous vertebroplasty (PVP) for the treatment of acute thoracolumbar osteoporotic compression fracture in middle-age and elderly individuals.Methods: Between January 2016 and August 2019, a total of 30 patients, with a mean age of 69.4 years (range, 58–75 years), who experienced thoracic or lumbar fracture without neurological deficits, underwent the MIPS procedure combined with PVP. Preoperative and postoperative pain were assessed using a visual analog scale (VAS) and Oswestry Disability Index (ODI). Cobb angles and anterior column height were measured on lateral radiographs before surgery and at 3 days, 1, 3, and 6 months, and 1 and 2 years at final follow-up after surgery.Results: All patients underwent surgery successfully, with a mean follow-up of 18.2 ± 5.7 months (range, 12–45 months). Mean preoperative VAS score decreased from 7.3±2.2 to 1.4±0.3 at the final follow-up (pppConclusions: MIPS, including the fractured vertebra, combined with PVP, was a reliable and safe procedure, with satisfactory clinical and radiological results for the treatment of thoracolumbar osteoporotic compression fracture in patients without neurological deficits.

The presence of organic contaminants in wastewater poses considerable risks to the health of both humans and ecosystems. Although advanced oxidation processes that rely on highly reactive radicals to destroy organic contaminants are appealing treatment options, substantial energy and chemical inputs limit their practical applications. Here we demonstrate that Cu single atoms incorporated in graphitic carbon nitride can catalytically activate H(2)O(2) to generate hydroxyl radicals at pH 7.0 without energy input, and show robust stability within a filtration device. We further design an electrolysis reactor for the on-site generation of H(2)O(2) from air, water and renewable energy. Coupling the single-atom catalytic filter and the H(2)O(2) electrolytic generator in tandem delivers a wastewater treatment system. These findings provide a promising path toward reducing the energy and chemical demands of advanced oxidation processes, as well as enabling their implementation in remote areas and isolated communities.

Summary Effectively regulating heat flow between the human body and its environment not only increases thermal comfort but also presents a novel and potentially cost-effective approach to reducing building energy consumption. Infrared property-engineered textiles have been shown to passively regulate radiative heat dissipation for effective cooling and warming of the human body. However, a lack of dyes that can tune the textile color without compromising the infrared properties remains a major impediment to textile commercialization. Here, we report a new strategy utilizing inorganic nanoparticles as a coloring component for scalable brightly colored, infrared-transparent textiles. The as-fabricated composite textiles not only show a high infrared transparency of ∼80% and a passive cooling effect of ∼1.6°C–1.8°C but also exhibit intense visible colors with good stability against washing. This facile coloration approach will promote the commercialization of radiative cooling textiles in temperature-regulating wearable applications for effective energy savings.

Summary Electrocatalytic oxygen evolution and reduction reactions play a central role in clean energy technologies. Despite recent efforts to achieve fast gas reactant delivery to the reaction interface, efficient gas product evolution from the catalyst/electrolyte interface remains challenging. Inspired by the mammalian breathing process, here we developed an efficient electrocatalytic system to enable ample gas-solid-liquid three-phase contact lines and bidirectional gas pathways for evolution and consumption. During the oxygen evolution reaction, the newly formed O2 molecules quickly diffuse to the gas phase, waiving the bubble formation energy in the electrolyte. A record low overpotential of 190 mV at 10 mA⋅cm−2 was achieved using Au/NiFeOx catalysts. During the oxygen reduction reaction, O2 gas can transport to the catalyst/electrolyte interface, overcoming low O2 solubility in water and leading to ∼25-fold higher current densities for Ag/Pt bilayer nanoparticle catalysts. This breathing-mimicking design demonstrates efficient three-phase catalysis with a minimal catalyst thickness.

Recently, the face detection and alignment is so popular and widely used in many research and application fields. Many superior face detection algorithms such as multi-task cascade convolutional network have been presented. However, it has difficulty in predicting faces among the big scale images in real time due to its three stages cascade architecture with less optimization. In this paper, we propose a full GPU-based batch multi-task cascade convolutional network which is carefully designed and optimized in each step to gain a superior speed performance. In addition, we present a novel parallel memory allocation strategy, which further enables our algorithm to support the batch operation, so that the system throughput increases significantly. In the experiment, our method achieves up to 300fps, over 600% speedup with an equal accuracy over the state-of-the-art methods on the face detection benchmarks.

Skeleton-based Graph Convolutional Networks (GCNs) models for action recognition have achieved excellent prediction accuracy in the field. However, limited by large model and computation complexity, GCNs for action recognition like 2s-AGCN have insufficient power-efficiency and throughput on GPU. Thus, the demand of model reduction and hardware acceleration for low-power GCNs action recognition application becomes continuously higher. To address challenges above, this paper proposes a runtime sparse feature compress accelerator with hybrid pruning method: RFC-HyPGCN. First, this method skips both graph and spatial convolution workloads by reorganizing the multiplication order. Following spatial convolution workloads channel-pruning dataflow, a coarse-grained pruning method on temporal filters is designed, together with sampling-like fine-grained pruning on time dimension. Later, we come up with an architecture where all convolutional layers are mapped on chip to pursue high throughput. To further reduce storage resource utilization, online sparse feature compress format is put forward. Features are divided and encoded into several banks according to presented format, then bank storage is split into depth-variable mini-banks. Furthermore, this work applies quantization, input-skipping and intra-PE dynamic data scheduling to accelerate the model. In experiments, proposed pruning method is conducted on 2s-AGCN, acquiring 3.0x-8.4x model compression ratio and 73.20\% graph-skipping efficiency with balancing weight pruning. Implemented on Xilinx XCKU-115 FPGA, the proposed architecture has the peak performance of 1142 GOP/s and achieves up to 9.19x and 3.91x speedup over high-end GPU NVIDIA 2080Ti and NVIDIA V100, respectively. Compared with latest accelerator for action recognition GCNs models, our design reaches 22.9x speedup and 28.93\% improvement on DSP efficiency., Comment: 8 pages, 2021 IEEE 32nd International Conference on Application-specific Systems, Architectures and Processors (ASAP)

The increasing demand for next-generation energy storage systems necessitates the development of high-performance lithium batteries

Air-filtering masks, also known as respirators, protect wearers from inhaling fine particulate matter (PM2.5) in polluted air, as well as airborne pathogens during a pandemic, such as the ongoing COVID-19 pandemic. Fibrous medium, used as the filtration layer, is the most essential component of an air-filtering mask. This article presents an overview of the development of fibrous media for air filtration. We first synthesize the literature on several key factors that affect the filtration performance of fibrous media. We then concentrate on two major techniques for fabricating fibrous media, namely, meltblown and electrospinning. In addition, we underscore the importance of electret filters by reviewing various methods for imparting electrostatic charge on fibrous media. Finally, this article concludes with a perspective on the emerging research opportunities amid the COVID-19 crisis., Considering that air pollution will continue to be a major health concern in newly industrialized countries and the ongoing COVID-19 pandemic may create lasting cultural shifts around the world to increase mask usages, the pursuit of more efficient and cost-effective masks will be a cornerstone to maintaining public health. This review provides a comprehensive picture of the historic development of air-filtering masks, identifies remaining knowledge gaps, and envisions emerging research opportunities amid the pressing challenges of air pollution and COVID-19.

Recently, Deep Neural Networks (DNNs) have been widely used in natural language processing. However, DNNs are often computation-intensive and memory-expensive. Therefore, deploying DNNs in the real world is very difficult. In order to solve this problem, we proposed a network model based on the dilate gated convolutional neural network, which is very hardware-friendly. We further expanded the word representations and depth of the network to improve the performance of the model. We replaced the Sigmoid function to make it more friendly for hardware computation without loss, and we quantized the network weights and activations to compress the network size. We then proposed the first FPGA (Field Programmable Gate Array)-based event detection accelerator based on the proposed model. The accelerator significantly reduced the latency with the fully pipelined architecture. We implemented the accelerator on the Xilinx XCKU115 FPGA. The experimental results show that our model obtains the highest F1-score of 84.6% in the ACE 2005 corpus. Meanwhile, the accelerator achieved 95.2 giga operations (GOP)/s and 13.4 GOPS/W in performance and energy efficiency, which is 17/158 times higher than the Graphics Processing Unit (GPU).

Fast-charging is considered as one of the most desired features needed for lithium-ion batteries to accelerate the mainstream adoption of electric vehicles. However, current battery charging protocols mainly consist of conservative rate steps to avoid potential hazardous lithium plating and its associated parasitic reactions. A highly sensitive onboard detection method could enable battery fast-charging without reaching the lithium plating regime. Here, we demonstrate a novel differential pressure sensing method to precisely detect the lithium plating event. By measuring the real-time change of cell pressure per unit of charge (dP/dQ) and comparing it with the threshold defined by the maximum of dP/dQ during lithium-ion intercalation into the negative electrode, the onset of lithium plating before its extensive growth can be detected with high precision. In addition, we show that by integrating this differential pressure sensing into the battery management system (BMS), a dynamic self-regulated charging protocol can be realized to effectively extinguish the lithium plating triggered by low temperature (0 °C) while the conventional static charging protocol leads to catastrophic lithium plating at the same condition. We propose that differential pressure sensing could serve as an early nondestructive diagnosis method to guide the development of fast-charging battery technologies.

Supercooled liquid sulfur provides higher capacity, faster kinetics, and better cycling life compared to solid sulfur., In lithium-sulfur (Li-S) chemistry, the electrically/ionically insulating nature of sulfur and Li2S leads to sluggish electron/ion transfer kinetics for sulfur species conversion. Sulfur and Li2S are recognized as solid at room temperature, and solid-liquid phase transitions are the limiting steps in Li-S batteries. Here, we visualize the distinct sulfur growth behaviors on Al, carbon, Ni current collectors and demonstrate that (i) liquid sulfur generated on Ni provides higher reversible capacity, faster kinetics, and better cycling life compared to solid sulfur; and (ii) Ni facilitates the phase transition (e.g., Li2S decomposition). Accordingly, light-weight, 3D Ni-based current collector is designed to control the deposition and catalytic conversion of sulfur species toward high-performance Li-S batteries. This work provides insights on the critical role of the current collector in determining the physical state of sulfur and elucidates the correlation between sulfur state and battery performance, which will advance electrode designs in high-energy Li-S batteries.

Audio event recognition is applied in many novel application areas. Opposing the deep CNN, 1-max pooling CNN is a simple, but efficient CNN architecture for robust audio event recognition. This study proposes a parallel architecture to accelerate robust audio event recognition. To implement this in hardware, we evaluate the precision of 1-max pooling CNN model and propose an approximate algorithm to replace the complex calculation in spectral image feature (SIF) extraction. We then propose a scalable parallel structure of SIF extraction and 1-max pooling CNN. The SIF extraction unit has eight parallelisms and the 1-Max Pooling CNN accelerator has 40 processor elements (PEs) in our implementation. The entire system is implemented on the Xilinx VC709 board. The average performance of our FPGA accelerator is 675.7 fps under 100 MHz working frequency, which is about 31.9 × speed-up compare with CPU. We further implement a small-scale FPGA array with four Xilinx FPGA for robust audio event recognition. To communicate between the four FPGA and the host, we design a route protocol based on source route algorithm.

Dust filtration is a crucial process for industrial waste gas treatment. Great efforts have been devoted to improve the performance of dust filtration filters both in industrial and fundamental research. Conventional air-filtering materials are limited by three key issues: (1) Low filtration efficiency, especially for particulate matter (PM) below 1 μm; (2) large air pressure drops across the filter, which require a high energy input to overcome; and (3) safety hazards such as dust explosions and fires. Here, we have developed a “smart” multifunctional material which can capture PM with high efficiency and an extremely low pressure drop, while possessing a flame retardant design. This multifunctionality is achieved through a core–shell nanofiber design with the polar polymer Nylon-6 as the shell and the flame retardant triphenyl phosphate (TPP) as the core. At 80% optical transmittance, the multifunctional materials showed capture efficiency of 99.00% for PM2.5 and >99.50% for PM10–2.5, with a pressure drop of only 0.25 kPa (0.2% of atmospheric pressure) at a flow rate of 0.5 m s–1. Moreover, during direct ignition tests, the multifunctional materials showed extraordinary flame retardation; the self-extinguishing time of the filtrate-contaminated filter is nearly instantaneous (0 s/g) compared to 150 s/g for unmodified Nylon-6., A “smart” air filter which can efficiently capture tiny particulate matters while possessing a flame retardant property is developed.

Aerosol-induced haze problem has become a serious environmental concern. Filtration is widely applied to remove aerosols from gas streams. Despite classical filtration theories, the nanoscale capture and evolution of aerosols is not yet clearly understood. Here we report an in situ investigation on the nanoscale capture and evolution of aerosols on polyimide nanofibers. We discovered different capture and evolution behaviors among three types of aerosols: wetting liquid droplets, nonwetting liquid droplets, and solid particles. The wetting droplets had small contact angles and could move, coalesce, and form axisymmetric conformations on polyimide nanofibers. In contrast, the nonwetting droplets had a large contact angle on polyimide nanofibers and formed nonaxisymmetric conformations. Different from the liquid droplets, the solid particles could not move along the nanofibers and formed dendritic structures. This study provides an important insight for obtaining a deep understanding of the nanoscale capture and evolution of aerosols and benefits future design and development of advanced filters.

Face masks are widely used to filter airborne pollutants, especially when particulate matter (PM) pollution has become a serious concern to public health. Here, the concept of thermal management is introduced into face masks for the first time to enhance the thermal comfort of the user. A system of nanofiber on nanoporous polyethylene (fiber/nanoPE) is developed where the nanofibers with strong PM adhesion ensure high PM capture efficiency (99.6% for PM2.5) with low pressure drop and the nanoPE substrate with high-infrared (IR) transparency (92.1%, weighted based on human body radiation) results in effective radiative cooling. We further demonstrate that by coating nanoPE with a layer of Ag, the fiber/Ag/nanoPE mask shows a high IR reflectance (87.0%) and can be used for warming purposes. These multifunctional face mask designs can be explored for both outdoor and indoor applications to protect people from PM pollutants and simultaneously achieve personal thermal comfort.

Soil contamination by heavy metals constitutes an important environmental problem, whereas field applicability of existing remediation technologies has encountered numerous obstacles, such as long operation time, high chemical cost, large energy consumption, secondary pollution, and soil degradation. Here we report the design and demonstration of a remediation method based on a concept of asymmetrical alternating current electrochemistry that achieves high degrees of contaminant removal for different heavy metals (copper, lead, cadmium) at different initial concentrations (from 100 to 10,000 ppm), all reaching corresponding regulation levels for residential scenario after rational treatment time (from 30 min to 6 h). No excessive nutrient loss in treated soil is observed and no secondary toxic product is produced. Long-term experiment and plant assay show the high sustainability of the method and its feasibility for agricultural use., Soil pollution by heavy metals is a problem of global concern, requiring the development of remediation technologies. Here the authors report a method based on asymmetrical alternating current electrochemistry, which enables recycling of soil washing chemicals and eliminates secondary pollution.

Treatment of heavy-metal pollution in both point-of-use water and industrial wastewater is critical in protecting human health and the environment. Current methods for heavy-metal treatment in both sources have limitations. For point-of-use water, current methods usually suffer from limited capacity and difficulties in spontaneously removing multiple heavy metals. For industrial wastewater, current methods greatly reduce the value of heavy metal by precipitating them as sludge which requires further treatment. Here we developed an electrochemical method that can treat both low-concentration and high-concentration heavy-metal pollution using either direct current (DC) or alternating current (AC) electrodeposition with graphene-oxide-modified carbon felt electrode (CF-GO). The graphene oxide provides a high density of surface functional groups to assist the electrodeposition. The electrodeposition method showed 2 orders of magnitude higher capacity (>29 g heavy metal for 1 g of graphene oxide) compared with traditional adsorption methods. For low levels of heavy-metal pollution in point-of-use water, DC electrodeposition with a CF-GO electrode can reduce single heavy-metal ion pollution (Cu, Cd, and Pb) as well as multiple ion mixtures to below safe water drinking levels. This method can tolerate a much wider range of heavy-metal pollution in point-of-use water than traditional adsorption methods. For high-level pollution in industrial wastewater, AC electrodeposition can recover >99.9% heavy-metal ions. By tuning the AC frequency and voltage, the electrodeposition method can further selectively recover Cu, Cd, and Pb separately, which adds values to the heavy-metal removal process.

Fast-charging and high-energy-density batteries pose significant safety concerns due to high rates of heat generation. Understanding how localized high temperatures affect the battery is critical but remains challenging, mainly due to the difficulty of probing battery internal temperature with high spatial resolution. Here we introduce a method to induce and sense localized high temperature inside a lithium battery using micro-Raman spectroscopy. We discover that temperature hotspots can induce significant lithium metal growth as compared to the surrounding lower temperature area due to the locally enhanced surface exchange current density. More importantly, localized high temperature can be one of the factors to cause battery internal shorting, which further elevates the temperature and increases the risk of thermal runaway. This work provides important insights on the effects of heterogeneous temperatures within batteries and aids the development of safer batteries, thermal management schemes, and diagnostic tools., Operation of lithium batteries at high, non-uniform temperatures can lead to safety issues, but the effects of localized high temperatures are difficult to probe. Here the authors use micro-Raman spectroscopy to show that local-temperature hotspots can induce lithium metal growth and trigger circuit shorting.

Lithium (Li) metal has long been considered the "holy grail" of battery anode chemistry but is plagued by low efficiency and poor safety due to its high chemical reactivity and large volume fluctuation, respectively. Here we introduce a new host of wrinkled graphene cage (WGC) for Li metal. Different from recently reported amorphous carbon spheres, WGC show highly improved mechanical stability, better Li ion conductivity, and excellent solid electrolyte interphase (SEI) for continuous robust Li metal protection. At low areal capacities, Li metal is preferentially deposited inside the graphene cage. Cryogenic electron microscopy characterization shows that a uniform and stable SEI forms on the WGC surface that can shield the Li metal from direct exposure to electrolyte. With increased areal capacities, Li metal is plated densely and homogeneously into the outer pore spaces between graphene cages with no dendrite growth or volume change. As a result, a high Coulombic efficiency (CE) of ∼98.0% was achieved under 0.5 mA/cm

Recently, machine learning, especially deep learning, has been a core algorithm to be widely used in many fields such as natural language processing, speech recognition, object recognition, and so on. At the same time, another trend is that more and more applications are moved to wearable and mobile devices. However, traditional deep learning methods such as convolutional neural network (CNN) and its variants consume a lot of memory resources. In this case, these powerful deep learning methods are difficult to apply on mobile memory-limited platforms. In order to solve this problem, we present a novel memory-management strategy called mmCNN in this paper. With the help of this method, we can easily deploy a trained large-size CNN on any memory size platform such as GPU, FPGA, or memory-limited mobile devices. In our experiments, we run a feed-forward CNN process in some extremely small memory sizes (as low as 5 MB) on a GPU platform. The result shows that our method saves more than 98% memory compared to a traditional CNN algorithm and further saves more than 90% compared to the state-of-the-art related work “vDNNs” (virtualized deep neural networks). Our work in this paper improves the computing scalability of lightweight applications and breaks the memory bottleneck of using deep learning method on memory-limited devices.

With the development of the Internet, the volume of information is expanding rapidly, and the complex information makes it particularly important to extract information quickly and intelligently. Event extraction algorithm is such a kind of fast and accurate information acquisition algorithm. With the development of deep learning in recent years, deep learning also shines in the field of time extraction. CEE-DGCNN network is an efficient and accurate event extraction algorithm for Chinese text. At present, the acceleration research on deep learning in the academic community mainly focuses on GPU, FPGA and other platforms. In fact, compared with these platforms, CPU is more versatile. Considering that the algorithm should have universality and universality in industrial application, this paper makes an in-depth analysis of CEE-DGCNN algorithm to study its acceleration on CPU. Experiments were carried out on x86 and ARM architectures respectively, and good acceleration effects were achieved.

Shape control has proven to be a powerful and versatile means of tailoring the properties of Bi2Se3 nanostructures for a wide variety of applications. Here, three different Bi2Se3 nanostructures, i.e., spiral-type nanoplates, smooth nanoplates, and dendritic nanostructures, were prepared by manipulating the supersaturation level in the synthetic system. This mechanism study indicated that, at low supersaturation, defects in the crystal growth could cause a step edge upon which Bi2Se3 particles were added continuously, leading to the formation of spiral-type nanoplates. At intermediate supersaturation, the aggregation of amorphous Bi2Se3 particles and subsequent recrystallization resulted in the formation of smooth nanoplates. Furthermore, at high supersaturation, polycrystalline Bi2Se3 cores formed initially, on which anisotropic growth of Bi2Se3 occurred. This work not only advances our understanding of the growth mechanism but also offers a new approach to control the morphology of Bi2Se3 nanostructures.

Tea (Camellia sinensis (L.) O. Kuntze, Theaceae) flowers possess many physiological functions and have been used in traditional medicines for deodorization, skin care, cough suppressant and expectorant in China. However, there is little information about its effect on nonalcoholic fatty liver disease (NAFLD). In this study, an oleic acid-induced HepG2 cell model of steatosis was established, and the anti-NAFLD effects of tea flowers and the related mechanisms were investigated. Three tea flower extracts, 40% TFE, 80% TFE and TFRE couldn’t prevent triglyceride (TG) accumulation in oleic acid-treated HepG2 cells (p > 0.05), but significantly decreased the TG level in lipid-overloaded HepG2 cells after 48 h treatment (p 0.05), but up-regulated the mRNA level of carnitine palmitoyl-CoA transferase (CPT) (p < 0.05). Moreover, tea flower extracts could significantly reduce intracellular reactive oxygen species (ROS) level (p < 0.05). These results indicated tea flowers may be a potential natural resourse for the treatment of NAFLD.