An ultrathin and flexible memristive nanodevice with analog resistance switching behavior emulating synaptic plasticity and neuromorphic functions is the building block to developing high-density flexible artificial neural networks with stacking tree-dimensional architecture. In order to biorealistically implement multiple synaptic functions with different time-evolution characteristics, the memristive devices are expected to perform resistive switching and memory properties with wide time scale from volatile to permanent states. That means the memristive device should be the second-order memristor with the secondary short-term state-variable. So far, the memristive devices with neuromorphic functions have been realized in the thin films of transition-metal oxides. Single-layer or few-layers two-dimensional transition metal dichalcogenide materials (TMDs) have shown many advantages over those oxide thin films, including ultrathin thickness (less than 5 nm), excellent mechanical flexibility, appealing physical and chemical properties, which renders two-dimensional TMDs extremely promising for flexible ultrathin transparent second-order memristors. The inherent learning and memory properties with wide time scale from volatile to permanent states are expected to be achieved in the ultrathin TMDs-based second-order memristive devices through precisely and optimally controlling the defects and microstructures of two-dimensional TMDs materials, selection and preparation of electrodes and parameters of external electric stimuli. Then, various cell-level and network-level synaptic and psychological functions will be biorealistically implemented in the flexible ultrathin two-dimensional TMDs materials-based memristive devices through engineering the pre- and post-synapse spikes and designing neuromorphic circuits. Furthermore, by using in-situ TEM, Kelvin probe microscopy (KFM), angle-resolved XPS, the resistive switching mechanisms and the physical nature of the secondary short-term state-variable in the ultrathin two-dimensional TMDs second-order memristor will be systematically investigated. This proposed project will provide experimental and theoretical guidance for the application of two-dimensional TMDs materials as the flexible ultrathin artificial synapses.
为构建柔性高密度三维堆垛神经形态芯片,需要忆阻突触器件具有超薄和柔性的特点。同时,为从本质上模拟突触不同时间特性的多种可塑性,需要忆阻器的电阻态从易失性到非易失性长时间跨度的连续可调,即要求其为具有短时程状态变量的二阶忆阻器。与传统金属氧化物忆阻材料相比,单层或少层二维过渡金属硫化物(TMDs)具有超薄(<5 nm)的厚度和优异的柔韧性,在构建小尺寸柔性透明二阶忆阻器时具有明显优势。本项目拟以超薄二维TMDs材料为研究对象,通过对二维材料微观结构和缺陷的精确调控,结合电极材料的优化选择制备以及外加电激励参数的控制,实现具有从易失性到非易失性长时间跨度连续可调电阻态的电阻转变性能,模拟突触的多种可塑性。在此基础上,采用原位TEM、开尔文探针显微镜KFM、角度分辨的XPS等技术深入研究TMDs的阻变机制和其短时程状态变量的物理本质。为制备超薄二维TMDs二阶忆阻突触器件提供实验和理论依据。
忆阻器凭借其简单的三明治结构和丰富的阻变动态特性,成为构建新一代神经网络的革命性智能纳米器件。为模拟具有不同时程特性的突触可塑性和神经元特性,需要忆阻器的电阻态从易失性到非易失性长时间跨度的连续可调,即要求其为具有短时程状态变量的二阶忆阻器。同时,为适应新一代电子产品柔性、透明、轻薄、便携等方面的进一步需求,二维材料,尤其是过渡金属硫系化合物(Transition Metal Dichalcogenides, TMDs)为制备超薄(阻变层厚度小于5 nm)柔性透明忆阻器件提供了可能。在本项目的资助下,我们以多种单层或少层二维TMD材料,如MoS2、WSe2和MoTe2等为研究对象,通过等离子体处理和热氧化等不同的后处理工艺,精确调节二维材料的微观缺陷;系统研究了不同金属电极以及石墨烯电极对二维材料忆阻器阻变性能的影响规律,在单层MoS2、少层WSe2-xOY和WO3-x/WSe2-y异质结中获得了具有多时程特性的模拟型阻变,实现了短时可塑性、长时可塑性和双脉冲易化等具有不同时程的多种光电突触可塑性。同时,采用柔性PET衬底,研制出基于Graphene/WSe2-xOY Graphene的柔性全二维忆阻突触器件,并系统评估了其弯折稳定性,阐明了其阻变的微观机制。本项目的研究成果为研制基于二维材料的高密度忆阻突触阵列和具有多功能的神经元电路提供了理论和实验基础。
{{i.achievement_title}}
数据更新时间:2023-05-31
路基土水分传感器室内标定方法与影响因素分析
拥堵路网交通流均衡分配模型
基于ESO的DGVSCMG双框架伺服系统不匹配 扰动抑制
基于二维材料的自旋-轨道矩研究进展
适用于带中段并联电抗器的电缆线路的参数识别纵联保护新原理
相转变调控二维过渡金属硫化物基柔性阻变存储器阻变效应及其阻变机理研究
BTC忆阻突触器件及其类脑学习认知功能研究
氧化物忆阻材料与人工神经突触器件
基于氧化物薄膜的瞬态忆阻器件制备及其瞬态特性、阻变机理研究