With the rapid acceleration of the digital transformation era, the continued miniaturization of electronic devices has led to an unprecedented increase in device density and interconnect complexity, thereby driving the demand for high-speed data storage, processing, and communication within increasingly compact form factors [1,2]. At the same time, the rapid expansion of artificial intelligence (AI)-enabled space technologies is accelerating the adoption of cost-effective commercial off-the-shelf (COTS) electronic devices in radiation-rich environments [3]. In particular, neuromorphic computing, which integrates data storage and processing within a single device, has emerged as a promising solution for efficient AI computation in space systems [4]. However, in such applications, radiation-induced degradation or device failure can lead to critical system-level malfunctions, making device reliability under extreme environments a key requirement.
In this study, we systematically investigate the reliability of two-dimensional (2D) material-based resistive random-access memory (RRAM) devices for next-generation neuromorphic computing in extreme radiation environments. Owing to their atomically thin structures, 2D materials are expected to exhibit enhanced tolerance to radiation-induced defect formation and charge trapping, making them a promising platform for highly durable neuromorphic devices capable of operating in harsh environments. We select Te, MoS2, and h-BN as representative active-layer materials and comparatively examine the stability, degradation behavior, and key reliability issues of each material system under radiation exposure. In addition, we discuss possible material- and device- engineering strategies to mitigate radiation-induced degradation and improve operational stability.
References
[1] S. H. Shin. et al., Nano Letters, Mini Review, 25 (18), 7224–7233 (2025)
[2] H. H. Yoon. et al., npj 2d materials and applications, 9 (68), 1-11 (2025)
[3] Budroweit, J. et al., Electronics, 10(9), 1008 (2021)
[4] F. Ortiz et al., IEEE Transactions on Machine Learning in Communications and Networking, 2, 169-189 (2024)