Metal alloy anode materials with high specific capacity and low voltage have recently gained significant attention due to their excellent electrochemical performance and the ability to suppress dendrite growth.However...Metal alloy anode materials with high specific capacity and low voltage have recently gained significant attention due to their excellent electrochemical performance and the ability to suppress dendrite growth.However,experimental investigations of metal alloys can be time-consuming and expensive,often requiring extensive experimental design and effort.In this study,we developed a machine learning model based on the Crystal Graph Convolutional Neural Network(CGCNN)to screen alloy anode materials for seven battery systems,including lithium(Li),sodium(Na),potassium(K),zinc(Zn),magnesium(Mg),calcium(Ca),and aluminum(Al).We utilized data with tens of thousands of alloy materials from the Materials Project(MP)and Automatic FLOW for Materials Discovery(AFLOW)databases.Without any experimental voltage input,we identified over 30 alloy systems that have been experimentally validated with good precision.Additionally,we predicted over 100 alloy anodes with low potential and high specific capacity.We hope this work to spur further interest in employing advanced machine learning models for the design of battery materials.展开更多
The development of high-performance binders is a simple but effective approach to address the rapid capacity decay of high-capacity anodes caused by large volume change upon lithiation/delithiation.Herein,we demonstra...The development of high-performance binders is a simple but effective approach to address the rapid capacity decay of high-capacity anodes caused by large volume change upon lithiation/delithiation.Herein,we demonstrate a unique organic/inorganic hybrid binder system that enables an efficient in situ crosslinking of aqueous binders(e.g.,sodium alginate(SA)and carboxymethyl cellulose(CMC))by reacting with an inorganic crosslinker(sodium metaborate hydrate(SMH))upon vacuum drying.The resultant 3D interconnected networks endow the binders with strong adhesion and outstanding self-healing capability,which effectively improve the electrode integrity by preventing fracturing and exfoliation during cycling and facilitate Li^(+)ion transfer.SiO anodes fabricated from the commercial microsized powders with the SA/0.2SMH binder maintain 1470 mAh g^(-1)of specific capacity at 100 mA g^(-1)after 200 cycles,which is 5 times higher than that fabricated with SA binder alone(293 mAh g^(-1)).Nearly,no capacity loss was observed over 500 cycles when limiting discharge capacity at 1500 mAh g^(-1).The new binders also dramatically improved the performance of Fe_(2)O_(3),Fe_(3)O_(4),NiO,and Si electrodes,indicating the excellent applicability.This finding represents a novel strategy in developing high-performance aqueous binders and improves the prospect of using high-capacity anode materials in Li-ion batteries.展开更多
Aqueous zinc-ion batteries(ZIBs) are considered promising power sources for grid storage,but they face several issues,including dendrite growth,corrosion,hydrogen evolution,etc.,which are related to the Zn metal/liqui...Aqueous zinc-ion batteries(ZIBs) are considered promising power sources for grid storage,but they face several issues,including dendrite growth,corrosion,hydrogen evolution,etc.,which are related to the Zn metal/liquid electrolyte interface.To address these challenges,many researchers have focused on modifying the Zn anode with surface adsorption.However,the underlying mechanism between the Zn surface and adsorbed/protective molecules has not been thoroughly explored.In this study,we built a multiscale simulation platform that integrates state-of-art simulation methods to comprehensively investigate the adsorption process of amino acids on the Zn metal surface.Our major finding is that adsorption sites,adsorbate–surface angle,and average distance are critical parameters for the stability and strength of surface adsorption.Additionally,ab initio molecular dynamics reveal the kinetics of the surface adsorption and molecule reorientation processes.Specifically,it can be discovered that the amino acids prefer to align parallel to the Zn metal surface,leading to better surface protection against corrosion and preventing dendrite growth.These findings pave the way for an in-depth understanding of the surface adsorption process,as well as providing concrete design principles for stable Zn metal anodes.展开更多
In recent years,interdisciplinary research has become increasingly popular within the scientific community.The fields of materials science and chemistry have also gradually begun to apply the machine learning technolo...In recent years,interdisciplinary research has become increasingly popular within the scientific community.The fields of materials science and chemistry have also gradually begun to apply the machine learning technology developed by scientists from computer science.Graph neural networks(GNNs)are new machine learning models with powerful feature extraction,relationship inference,and compositional generalization capabilities.These advantages drive researchers to design computational models to accelerate material property prediction and new materials design,dramatically reducing the cost of traditional experimental methods.This review focuses on the principles and applications of the GNNs.The basic concepts and advantages of the GNNs are first introduced and compared to the traditional machine learning and neural networks.Then,the principles and highlights of seven classic GNN models,namely crystal graph convolutional neural networks,iCGCNN,Orbital Graph Convolutional Neural Network,MatErials Graph Network,Global Attention mechanism with Graph Neural Network,Atomistic Line Graph Neural Network,and BonDNet are discussed.Their connections and differences are also summarized.Finally,insights and prospects are provided for the rapid development of GNNs in materials science at the atomic scale.展开更多
基金supported by the National Key R&D Program of China(2023YFB2504000,YH)a start-up grant from Zhejiang University and the Fundamental Research Funds for the Central Universities(2021FZZX001,226-2024-00005)supported by Special Support Plan for High Level Talents in Zhejiang Province(2023R5231)。
文摘Metal alloy anode materials with high specific capacity and low voltage have recently gained significant attention due to their excellent electrochemical performance and the ability to suppress dendrite growth.However,experimental investigations of metal alloys can be time-consuming and expensive,often requiring extensive experimental design and effort.In this study,we developed a machine learning model based on the Crystal Graph Convolutional Neural Network(CGCNN)to screen alloy anode materials for seven battery systems,including lithium(Li),sodium(Na),potassium(K),zinc(Zn),magnesium(Mg),calcium(Ca),and aluminum(Al).We utilized data with tens of thousands of alloy materials from the Materials Project(MP)and Automatic FLOW for Materials Discovery(AFLOW)databases.Without any experimental voltage input,we identified over 30 alloy systems that have been experimentally validated with good precision.Additionally,we predicted over 100 alloy anodes with low potential and high specific capacity.We hope this work to spur further interest in employing advanced machine learning models for the design of battery materials.
基金support from the National Outstanding Youth Foundation of China(52125104)the National Natural Science Foundation of China(52071285 and 51831009)+1 种基金the Fundamental Research Funds for the Central Universities(2021FZZX001-09)the National Youth Top-Notch Talent Support Program.Z.H.acknowledges support under the Australian Research Council's Future Fellowship(FT190100658).
文摘The development of high-performance binders is a simple but effective approach to address the rapid capacity decay of high-capacity anodes caused by large volume change upon lithiation/delithiation.Herein,we demonstrate a unique organic/inorganic hybrid binder system that enables an efficient in situ crosslinking of aqueous binders(e.g.,sodium alginate(SA)and carboxymethyl cellulose(CMC))by reacting with an inorganic crosslinker(sodium metaborate hydrate(SMH))upon vacuum drying.The resultant 3D interconnected networks endow the binders with strong adhesion and outstanding self-healing capability,which effectively improve the electrode integrity by preventing fracturing and exfoliation during cycling and facilitate Li^(+)ion transfer.SiO anodes fabricated from the commercial microsized powders with the SA/0.2SMH binder maintain 1470 mAh g^(-1)of specific capacity at 100 mA g^(-1)after 200 cycles,which is 5 times higher than that fabricated with SA binder alone(293 mAh g^(-1)).Nearly,no capacity loss was observed over 500 cycles when limiting discharge capacity at 1500 mAh g^(-1).The new binders also dramatically improved the performance of Fe_(2)O_(3),Fe_(3)O_(4),NiO,and Si electrodes,indicating the excellent applicability.This finding represents a novel strategy in developing high-performance aqueous binders and improves the prospect of using high-capacity anode materials in Li-ion batteries.
基金supported by the National Key R&D Program (2022YFB2502000)the Zhejiang Provincial Natural Science Foundation of China (LZ23B030003)the Fundamental Research Funds for the Central Universities (2021FZZX001-08,2021FZZX001-09)。
文摘Aqueous zinc-ion batteries(ZIBs) are considered promising power sources for grid storage,but they face several issues,including dendrite growth,corrosion,hydrogen evolution,etc.,which are related to the Zn metal/liquid electrolyte interface.To address these challenges,many researchers have focused on modifying the Zn anode with surface adsorption.However,the underlying mechanism between the Zn surface and adsorbed/protective molecules has not been thoroughly explored.In this study,we built a multiscale simulation platform that integrates state-of-art simulation methods to comprehensively investigate the adsorption process of amino acids on the Zn metal surface.Our major finding is that adsorption sites,adsorbate–surface angle,and average distance are critical parameters for the stability and strength of surface adsorption.Additionally,ab initio molecular dynamics reveal the kinetics of the surface adsorption and molecule reorientation processes.Specifically,it can be discovered that the amino acids prefer to align parallel to the Zn metal surface,leading to better surface protection against corrosion and preventing dendrite growth.These findings pave the way for an in-depth understanding of the surface adsorption process,as well as providing concrete design principles for stable Zn metal anodes.
文摘In recent years,interdisciplinary research has become increasingly popular within the scientific community.The fields of materials science and chemistry have also gradually begun to apply the machine learning technology developed by scientists from computer science.Graph neural networks(GNNs)are new machine learning models with powerful feature extraction,relationship inference,and compositional generalization capabilities.These advantages drive researchers to design computational models to accelerate material property prediction and new materials design,dramatically reducing the cost of traditional experimental methods.This review focuses on the principles and applications of the GNNs.The basic concepts and advantages of the GNNs are first introduced and compared to the traditional machine learning and neural networks.Then,the principles and highlights of seven classic GNN models,namely crystal graph convolutional neural networks,iCGCNN,Orbital Graph Convolutional Neural Network,MatErials Graph Network,Global Attention mechanism with Graph Neural Network,Atomistic Line Graph Neural Network,and BonDNet are discussed.Their connections and differences are also summarized.Finally,insights and prospects are provided for the rapid development of GNNs in materials science at the atomic scale.
