适用于模拟神经突触功能的电阻可渐变调控忆阻器及其机理研究

Aiming at the problems of resistance fluctuation memristor resistance can be regulated continuously changes in synaptic function simulation when the big, bad repetition. Directional active particles using fast ion conductors in the solid electrolyte in the electric field under the control of the transfer gradient type form the memristor function layer and carrier in the nonlinear electric field controllable transmission which, as a method of memristors resistance gradient control. Theoretical calculation and experimental study of the combination, to carry out the memristor material and device design optimization study, nano scale memristor material and fabrication of electrical performance test system and the characteristics of technology research, control methods, control of the resistance characteristics, resistance mechanism of gradient. The key to solve the AgInSbTe memristor material components, performance optimization and device fabrication in scientific issues; to establish a solid electrolyte memristor resistance gradient control method, promulgates between the memristor control signal and the response of conductance of the rules, to develop a high performance memristor mechanism clear, gradual and controllable resistance. To provide theoretical and technical support for the simulation of neural synaptic function; electronic devices based on neural synaptic function is to achieve the nanotechnology, biotechnology, information technology, a combination of cognitive science, an important breakthrough in intelligent machine development has a biological cognitive characteristics, of great significance to study the original device of our country.

针对阻值可连续调控变化的忆阻器在模拟神经突触功能时存在的阻值波动性大、重复性差等问题。利用快离子导体固体电解质中的活性粒子在电场调控下的定向迁移形成的梯度型忆阻功能层和载流子在其中的非线性电场可控传输，作为忆阻器阻值渐变调控的方法。理论计算和实验研究相结合，开展忆阻器材料及器件的设计优化研究、纳米尺度忆阻器材料及器件的制备工艺研究、电学性能测试系统及特性调控方法研究、阻值渐变特性调控研究、阻值渐变机理研究。解决AgInSbTe忆阻材料组份、性能优化及器件制备中的关键基础科学问题；建立固体电解质忆阻器阻值渐变调控方法,揭示忆阻器调控信号与响应电导之间的规律，研制出机理清晰、阻值渐变可控的高性能忆阻器。为模拟神经突触功能提供理论和技术支撑；类神经突触功能的电子器件研究是实现将纳米科学、生物技术、信息技术、认知科学相结合，发展具有生物认知特征的智能机器的重要突破口，对我国原创性器件研究意义重大。

针对阻值可连续调控变化的忆阻器在模拟神经突触功能时存在的阻值波动性大、重复性差 等问题，对忆阻器新材料、模拟设计、制备工艺、测试表征、器件性能、机理分析和可能的应用等进行了深入的研究。解决了AgInSbTe忆阻材料组分、性能优选及其器件制备中的关键基础科学问题；针对忆阻器特性测试，提出了忆阻器表征测试方法，包括忆阻器直流特性、交流特性以及脉冲特性测试等，搭建了忆阻器常规测试系统和高速擦写测试系统；建立固体电解质忆阻器阻值渐变直流、脉冲、极性、等调控方法,研制出了类神经突触功能的阻值渐变忆阻器，并揭示忆阻器调控信号与阻值渐变之间的规律，为模拟神经突触功能提供理论和技术支撑；探索了阻值渐变的忆阻器在模式识别网络中的可能应用。