基于金属卟啉分子调控离子迁移的高性能忆阻器及其在神经形态模拟方面的应用

The stimulus-response mode of the memristor resembles the synaptic plasticity of the basic connection unit of biological neural network, making it possible to realize the novel brain-inspired functions. And this strategy has aroused the academic researching focus. As for the construction of memristors, the currently prevalently-studied devices are mainly the oxide-based memristors which are based on the migration of oxygen ions. However, the ionic migration process under electric field is extremely sensitive to the device structure and composition, and has great random characteristics, which results in the rough memristive response with low signal-to-noise ratio characteristics. Differently, the biological synaptic activity changes continuously as a function of the accumulative stimulating signals. Thus the rough and fluctuant memristive behaviors have inverse influences on the realization of synaptic functions. In view of this, this project intends to exquisitely regulate the ionic movements using the organic molecule—metalloporphyrin with a specific functional site. In combination with the results of theoretical calculations, we propose a strategy for constructing a metalloporphyrin/alumina-based memristor. In this device, the alumina layer plays the role of oxygen ion source, and the coordination centers of the metalloporphyrin molecules inside the organic active layer can regularly modulate the ionic transport, which can reduce the randomness of ionic transport and obtain the smooth memristive responses. Furthermore, we will use different metalloporphyrin molecules with different coordinate functional sites to further regulate the multiple ionic dynamics, for realizing the ideal platforms with a view to molecule-regulating the advanced memristive characteristics. Importantly, this can provide the device platforms to simulate diverse synaptic plasticity for comprehensive brain-inspired functions.

忆阻器的刺激响应模式可与神经网络的连接单元—突触的可塑性比拟，使得用其实现类脑功能成为可能，引起学术界的重点关注。关于忆阻器的构筑，当前研究的器件主要为基于氧离子迁移机制的氧化物忆阻器，然而电场驱动的氧离子迁移过程对器件结构与成分极其敏感，有很大的随机性特点，造成忆阻响应较为粗糙、信噪比低，而生物突触活动是随累积刺激信号逐渐变化的，因此粗糙波动的忆阻行为不利于突触功能的实现。鉴于此，本项目拟利用具有特定功能位点的有机分子—金属卟啉来精致地调控离子迁移。结合理论计算的结果，我们提出构建金属卟啉/氧化铝叠层结构忆阻器的策略，以氧化铝为离子源，通过有机功能层中的金属卟啉分子配位调控氧离子迁移，降低离子运动无序性，以获取平滑渐变的忆阻响应；并通过具有不同配位功能位点的金属卟啉分子进一步调控多样化的离子动态过程，实现分子调控器件导电态可塑性的目的，为模拟多样性的突触可塑性提供器件平台。

忆阻器的刺激响应模式可与神经网络的连接单元—突触的可塑性比拟，使得用其实现类脑功能成为可能，引起了学术界的重点关注。关于忆阻器的构筑，当前研究的器件主要为基于氧离子迁移机制的氧化物忆阻器，然而电场驱动的氧离子迁移过程对器件结构与成分极其敏感，有很大的随机性特点，造成忆阻响应较为粗糙、信噪比低，而生物突触活动是随累积刺激信号逐渐变化的，因此粗糙波动的忆阻行为不利于突触功能的实现。鉴于此，我们制备了金属卟啉基忆阻器ITO/MTPP/AlO3-x/Al，该离子型忆阻器在多圈周期性电压扫描下具有平滑渐变的忆阻响应特征，其忆阻行为源自于氧化物功能层AlO3-x中氧元素在有机半导体功能层中卟啉分子的配位下可逆的传输迁移过程，我们通过STEM-EDX确认了外电场激发作用下MTPP功能层中氧迁移过程，通过XPS确认了金属卟啉分子和氧之间的配位作用。MTPP功能层中氧的迁移堆积与累积输入的电信号之间存在正相关的关系，基于此我们模拟了基于Hebb规则的突触可塑性，例如短时程记忆、长时程记忆、经验学习等学习记忆规律。由于金属配位中心与氧之间的配位键需要一定的外加电压才能使之断裂，而在小电压下，氧只会在局部范围内调整方位重新分布，造成功能层内逐渐增强的极化和逐渐减弱的器件响应电流，也就是说金属卟啉忆阻器在不同大小的信号下有不同的响应模式，基于此我们可以模拟视觉系统的信号过滤功能和神经系统的习惯化和敏感化现象。