Key points are not available for this paper at this time.
Abstract True random number generators (TRNGs), which create cryptographically secure random bitstreams, hold great promise in addressing security concerns regarding hardware, communication, and authentication in the Internet of Things (IoT) realm. Recently, TRNGs based on nanoscale materials have gained considerable attention for avoiding conventional and predictable hardware circuitry designs that can be vulnerable to machine learning (ML) attacks. In this article, a low‐power and low‐cost TRNG developed by exploiting stochastic ferroelectric polarization switching in 2D ferroelectric CuInP 2 S 6 (CIPS)‐based capacitive structures, is reported. The stochasticity arises from the probabilistic switching of independent electrical dipoles. The TRNG exhibits enhanced stochastic variability with near‐ideal entropy, uniformity, uniqueness, Hamming distance, and independence from autocorrelation variations. Its unclonability is systematically examined using device‐to‐device variations. The generated cryptographic bitstreams pass the National Institute of Standards and Technology (NIST) randomness tests. This nanoscale CIPS‐based TRNG is circuit‐integrable and exhibits potential for hardware security in edge devices with advanced data encryption.
Kang et al. (Tue,) studied this question.