1/f Noise in Synaptic Ferroelectric Tunnel Junction: Impact on Convolutional Neural Network

In recent years, neuromorphic computing has been rapidly developed to overcome the limitations of von Neumann architecture. In this regard, the demand for high-performance synaptic devices with high switching speeds, low power consumption, and multilevel conductance is increasing. Among the various synaptic devices, ferroelectric tunnel junctions (FTJs) are promising candidates. While previous studies have focused on improving reliability of FTJs to enhance the synaptic behavior, low-frequency noise (LFN) of FTJs has not been characterized and its impact on the learning accuracy in neuromorphic computing remains unknown. Herein, the LFN characteristics of FTJs fabricated on n- and p-type Si along with the impact of 1/f noise on the learning accuracy of convolutional neural networks (CNNs) are investigated. The results indicate that the FTJ on p-type Si exhibits a far lower 1/f noise than that on n-type Si. The FTJ on p-type Si exhibits a significantly higher learning accuracy (86.26%) than that on n-type Si (78.70%) owing to its low-noise properties. This study provides valuable insights into the LFN characteristics of FTJs and a solution to improve the performance of synaptic devices by significantly reducing the 1/f noise.

Automated summary

In recent years, the development of neuromorphic computing has faced the limitations of von Neumann architecture. Therefore, there is a growing demand for high-performance synaptic devices that possess high switching speeds, low power consumption, and multilevel conductance. Among various synaptic devices, ferroelectric tunnel junctions (FTJs) have emerged as promising candidates. While previous studies have focused on improving the reliability of FTJs to enhance synaptic behavior, the low-frequency noise (LFN) of FTJs and its impact on the learning accuracy in neuromorphic computing have not been thoroughly investigated. This study explores the LFN characteristics of FTJs fabricated on n- and p-type Si and evaluates the impact of 1/f noise on the learning accuracy of convolutional neural networks (CNNs). The results demonstrate that FTJs on p-type Si exhibit significantly lower 1/f noise than those on n-type Si. Consequently, the FTJs on p-type Si achieve a significantly higher learning accuracy (86.26%) compared to those on n-type Si (78.70%) due to their low-noise properties. This study provides valuable insights into the LFN characteristics of FTJs and offers a potential solution to enhance the performance of synaptic devices by drastically reducing 1/f noise.