卤化物钙钛矿阻变存储器的超快阻变行为及其机理研究

Halide perovskite based resistive memory has attracted much attention in the memory field due to its simple process, the potential of high-speed operation and high-density integration. However, the resistive switching speed is generally limited to DC or microsecond pulses, and its switching characteristic and mechanism under ultrafast pulses (sub-nanosecond) are still unclear. In this project, the resistive memory with halide-perovskite and inert metal electrode is exploited with pico-second pulses. The ions migration induced conductive filament evolution are analyzed. By studying the resistive behavior of the device at different resistance speeds, the influence mechanism of physical phenomena such as electro-migration, thermal diffusion and thermochemical changes on resistive behavior are investigated. The shape of conductive filament, the electric field variation and temperature distribution can be established. By further characterizing the retention degradation because of ion transfer, the ultrafast resistive switching modulation method and the stability of the device can be optimized. It also provides an important scientific basis for elucidating the switching mechanism of perovskite-based resistive memory.

卤化物钙钛矿阻变存储器因具备制备工艺简单、可高速操作和高密度集成等优点，近年来成为存储器领域的关注热点。然而，目前对卤化物钙钛矿阻变存储器的研究普遍局限于直流或微秒级脉冲尺度，其在超高速条件下的阻变行为及其机理尚未明确。本项目提出以惰性金属为电极的卤化物钙钛矿材料体系构建高性能阻变存储器，拟重点研究卤化物钙钛矿阻变存储器的皮秒级超快阻变行为及其物理机制，获得离子迁移致导电通道动态变化的规律，并掌握超快阻变的性能调控方法。通过对器件进行不同阻变速度下阻变行为的研究，分析电致迁移、热致扩散以及热化学变化等物理现象对阻变行为的影响机理，初步获得导电通道的分布形态、电场变化和温度分布等多重物理效应的耦合规律。研究器件内部由于离子输运导致的保持特性退化机制，从而优化阻变性能的电学调控方法并提高器件稳定性。此项研究为卤化物钙钛矿阻变存储器的研发提供了理论依据和技术支持。

卤化物钙钛矿作为一种材料体系具有组分灵活可调、加工工艺简单、低温成膜、柔性集成等方面的优势。电场作用下卤素离子的快速迁移为研究卤素钙钛矿阻变器件提供了理论依据。本项目选用卤化物钙钛矿材料作为研究载体，研究了卤化物钙钛矿阻变器件在电场作用下的离子迁移特性以及高速阻变行为，构建了离子迁移形成导电通道的物理模型，实现了皮秒脉冲下模拟开关的阻变特性。这为高速存算一体的神经形态计算奠定了良好基础。研究了卤化物钙钛矿阻变器件的光电导调控特性，实现了全光控条件下其器件电导的连续可逆调制，可用来模拟光控突触权重的抑制和增强特性，并进一步验证了其光突触具有稳定图像和增强图像可识别度的能力。通过制备卤化物钙钛矿选通器件，获得了高非线性特性、超高速操作特性、超长循环特性和高柔性耐受性，建立了微观离子输运模型，构建了以其为核心器件的神经元电路并获得了良好的神经元活动功能。本项目的研究为卤化物钙钛矿阻变及选通器件的物理机制提供了理论依据，并为感存算一体的应用提供潜在解决方案。