ABSTRACT We present a photospike‐based true random number generator (PS‐TRNG) that exploits the intrinsic randomness from light–matter interaction and stochastic charge trapping. By integrating copper vanadate (CuV 2 O 6 ) nanostructures with a tin dioxide quantum dot (SnO 2 QD) layer, the device induces probabilistic trapping–de‐trapping dynamics, producing random photospike currents under optical pulse trains. The spike currents show high entropy and enable multi‐level random number generation beyond binary, providing ternary outputs with near‐ideal statistics (33.30% uniformity, 33.28% inter‐Hamming distance) and full success in all 15 NIST tests. We further develop an image authenticity verification system by integrating the PS‐TRNG with a mobile platform and custom‐designed circuit board, enabling hardware‐based detection of unauthorized image modifications. The random numbers are embedded as a hidden layer within the image data without degrading visual quality, enabling detection of unauthorized modifications. The system can successfully identify image modifications, even those involving highly sophisticated manipulations generated by artificial intelligence (AI)‐based image editing tools. The device maintains stable operation over 2 million cycles and remains reliable even after more than 460 days, demonstrating its long‐term stability.
ABSTRACT We present a photospike‐based true random number generator (PS‐TRNG) that exploits the intrinsic randomness from light–matter interaction and stochastic charge trapping. By integrating copper vanadate (CuV 2 O 6 ) nanostructures with a tin dioxide quantum dot (SnO 2 QD) layer, the device induces probabilistic trapping–de‐trapping dynamics, producing random photospike currents under optical pulse trains. The spike currents show high entropy and enable multi‐level random number generation beyond binary, providing ternary outputs with near‐ideal statistics (33.30% uniformity, 33.28% inter‐Hamming distance) and full success in all 15 NIST tests. We further develop an image authenticity verification system by integrating the PS‐TRNG with a mobile platform and custom‐designed circuit board, enabling hardware‐based detection of unauthorized image modifications. The random numbers are embedded as a hidden layer within the image data without degrading visual quality, enabling detection of unauthorized modifications. The system can successfully identify image modifications, even those involving highly sophisticated manipulations generated by artificial intelligence (AI)‐based image editing tools. The device maintains stable operation over 2 million cycles and remains reliable even after more than 460 days, demonstrating its long‐term stability.
The pursuit of hardware-based security solutions has highlighted the true random number generator (TRNG). Various physical phenomena, from noise generation to quantum physics complexities, have been explored for random number generation. The arc discharge light-induced TRNG (ALTRNG) is introduced, featuring wavelength-dependent photocurrent generation and arc discharge irradiation. A bipolar photo-responsive photodetector (BPPD) differentiates "1" and "0" states, producing highly random bits validated by the National Institute of Standards and Technology (NIST) 15 tests. The BPPD's response to deep-ultraviolet (DUV) and blue light enables distinct photocurrent generation under arc discharge illumination. The ALTRNG generates true random signals, yielding bit streams with unpredictability, uniform distribution, and stability. With a readout circuit achieving 2-kbps, wireless random number transmission is demonstrated, highlighting potential for secure password systems and artificial X-ray image generation.
This paper presents a high-throughput true random number generator (TRNG) based on optical speckle patterns generated by illuminating natural plant leaves as biological physical unclonable functions (PUFs). A coherent He–Ne laser produces stochastic far-field speckle fields that are captured by a CMOS camera and converted into random bitstreams using a clearly defined post-processing pipeline. The full extraction procedure— including binary quantization, Von Neumann debiasing, and statistical balancing—is formalized in Algorithm 1 to support reproducibility. Experimental validation is performed using a 38-s dataset (1100 frames), and long-term stability is demonstrated through a 2-hour continuous recording. All evaluated segments show near-ideal entropy (0.9999991 bits/bit), balanced bit distributions, and zero-mean autocorrelation, confirming the robustness of the biological scattering medium over extended durations. After debiasing, the system achieves an effective throughput of approximately 75 Mbit/s, aligning with requirements for cryptographic and security-critical applications. The optical setup is compact and compatible with laser-diode sources and embedded imaging modules, enabling future development of portable, low-power speckle-based TRNG systems.
As technologies such as artificial intelligence, autonomous driving, the Internet of Things, wearable electronics, and edge computing continue to spread in the era of the Fourth Industrial Revolution, the importance of hardware security for the safe storage, transmission, and processing of large volumes of data has grown substantially. One of the key components of such security systems is the true random number generator (TRNG), which produces unpredictable random numbers for cryptographic use. In recent years, research on TRNGs has increasingly moved beyond conventional CMOS-based approaches toward semiconductor devices built from emerging materials. These material-based TRNGs offer several advantages, including high integration density, low power consumption, compact form factors, and strong suitability for next-generation edge and IoT environments, because they can directly exploit the intrinsic stochasticity of the device itself as an entropy source. In this review, recent studies on TRNGs based on emerging material-based semiconductor devices are examined from the perspectives of entropy sources, device structures, randomness validation, and wearable/flexible extensions. By bringing together the key physical mechanisms, device platforms, evaluation criteria, and prospects for wearable and flexible electronics in edge and IoT environments, this review aims to provide a useful framework for future research on hardware security devices.
The true random number generator (TRNG) is an invaluable resource for many applications. Present TRNG methods all require specialised and expensive hardware, which are not suitable for personal use. A portable TRNG configuration is proposed, which is simply based on the camera of a commercial smartphone. No other auxiliary equipment except the mobile phone itself was needed, which offers a promising solution for personal encryption applications. With the flash lamp of the camera and the thumb of the operator, quasi‐uniform illumination was obtained, setting the image sensor at a working point where the shot noise of the photocurrent dominates the overall characteristics of the sensor output noise. Random bits were first produced according to the shot noise and then entropy enhanced by vector–matrix multiplication. A sequence of 10 9 bits was produced and its randomness has been proved by the National Institute of Standards and Technology (NIST) tests.
We present an integrated photonic quantum random number generator combining photon-path selection and time-delay encoding methods to efficiently produce high-quality random bits. True random number generators, particularly quantum-based random number generators, play a critical role in secure communications, cryptographic protocols, and simulations, as they provide genuine unpredictability derived from fundamental quantum processes. This study presents the development of an optical random number generator utilizing a photonic chip, resulting in a very stable, compact, and high key-to-signal ratio device. Our setup employs an attenuated pulsed laser source combined with a photonic circuit featuring a multi-mode interference, whose outputs are monitored by single-photon avalanche diodes. For photon-path encoding, a von Neumann extractor effectively reduces output bias by discarding correlated detection events. In addition, we measure photon arrival times (time-delay encoding) using a novel bin-merging and modulo-based encoding strategy to counteract hardware-induced biases, achieving near-ideal bit uniformity. Experimental results show that combining both encoding techniques doubles the bit-generation rate compared to path-encoding only, achieving 0.49 bits per signal while simultaneously meeting stringent randomness criteria validated by passing all tests in the NIST statistical test suite.
Unpredictability and entropy are key factors governing the success of cryptographic security, and true random numbers play a pivotal role in this space. While software-driven random numbers are easy and efficient to generate, their deterministic nature creates a vulnerability that could compromise the gate, barring entry to a sensitive system. Here, we propose using measurement randomness arising from physical and chemical aspects of transport (Layer 1), hardware characteristics (Layer 2), event features and extraction (Layer 3; facilitated by the HyperXtract 2.0), and cryptographic whitening via the secure hash algorithm (SHA-256; Layer 4) to create a high-entropy TRNG system. We demonstrate a host of unique experimental conditions (e.g., electrolyte chemistry, applied voltage, analyte type, transport direction, temperature, and amplifier type) across these four layers to generate ~480,000 256-bit long keys with 100% uniqueness in a Pearson-correlation-based statistical entropy harvesting method and most other methods reporting >99.9% unique keys in most instances. Furthermore, the randomness of these keys was evaluated using the NIST SP800-22b statistical test suite and yielded a success rate of ~99% at a 1% significance level. This study thus showcases the utility of solid-state nanopore sensing for generating high-entropy TRNGs and the potential as a robust, non-deterministic source for nextgeneration cybersecurity engineering.