Tong Wei, Wei Wei, Zhang Xiangzhe, Ding Shuimei, Lu Zheyi, Liu Liting, Li Wanying, Pan Chen, Kong Lingan, Wang Yiliu, Zhu Mengjian, Liang Shi-Jun, Miao Feng, Liu Yuan
Key Laboratory for Micro-Nano Optoelectronic Devices of Ministry of Education, School of Physics and Electronics, Hunan University, Changsha 410082, China.
National Laboratory of Solid State Microstructures, School of Physics, Collaborative Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China.
Nano Lett. 2023 Nov 8;23(21):9928-9935. doi: 10.1021/acs.nanolett.3c02888. Epub 2023 Oct 20.
Memristors have attracted considerable attention in the past decade, holding great promise for future neuromorphic computing. However, the intrinsic poor stability and large device variability remain key limitations for practical application. Here, we report a simple method to directly visualize the origin of poor stability. By mechanically removing the top electrodes of memristors operated at different states (such as SET or RESET), the memristive layer could be exposed and directly characterized through conductive atomic force microscopy, providing two-dimensional area information within memristors. Based on this technique, we observed the existence of multiple conducting filaments during the formation process and built up a physical model between filament numbers and the cycle-to-cycle variation. Furthermore, by improving the interface quality through the van der Waals top electrode, we could reduce the filament number down to a single filament during all switching cycles, leading to much controlled switching behavior and reliable device operation.
在过去十年中，忆阻器已引起了相当大的关注，对未来的神经形态计算有着巨大的前景。然而，其固有的稳定性差和器件变异性大仍然是实际应用的关键限制。在此，我们报告一种直接可视化稳定性差的根源的简单方法。通过机械去除处于不同状态（如设置或重置）下工作的忆阻器的顶部电极，可以暴露忆阻层并通过导电原子力显微镜直接对其进行表征，从而提供忆阻器内的二维区域信息。基于该技术，我们观察到在形成过程中存在多条导电细丝，并建立了细丝数量与逐周期变化之间的物理模型。此外，通过使用范德华顶部电极改善界面质量，我们可以在所有开关周期中将细丝数量减少到单根细丝，从而实现更可控的开关行为和可靠的器件操作。
