True Random Number Generator for Robust Data Security via Intrinsic Stochasticity in a 2D hBN Threshold Switching Memristor

Abstract

The explosive growth of Internet of Things (IoT) devices has demanded the urgency of robust data security to protect private information. The random numbers generated by pseudo-random number generators (PRNGs) play a crucial role in cryptographic algorithms. However, these sequences have become increasingly predictable due to advances in computing power and machine learning. Therefore, the development of true random number generators (TRNGs), which exploit the intrinsic hardware-level randomness as an entropy source, is important for future data security. In this study, we present TRNG circuits that harness the inherent stochasticity of a multilayer hBN volatile memristor. The fabricated device exhibits highly stable threshold switching with low set/hold voltages and a significantly high on/off ratio. We implement a spike generator by integrating the threshold switching memristor with passive components. Digitization with a fixed-reference comparator directly converts output spikes into bitstreams that successfully pass NIST randomness tests without postprocessing. The practical utility of the TRNG is demonstrated through XOR-based encryption and decryption of black-and-white and grayscale images, resulting in noise-like encrypted data and perfect recovery of the originals using identical keys. These results establish 2D hBN-based threshold switching memristors as compelling hardware entropy sources for secure, next-generation electronic systems.