Parallel On-Chip Physical Random Number Generator Based on Self-Chaotic Dynamics of Free-Running Broad-Area VCSEL Array

Random numbers, as a cornerstone in the interconnected digital world, are used in secure cryptographic protocols for commercial transactions, computing, and communications. Instead of the traditional deterministic pseudorandom numbers, physical random number generation (RNG) is currently being investigated by leveraging the chaotic dynamics of semiconductor lasers for improved security, speed, and compactness. However, those RNG approaches suffer from discrete and expensive components with limited scalability due to the enormous footprint imposed by the edge-emitting configuration, which increases the cost and impedes practical use in integrated devices. Herein, we demonstrated a parallel chip-scale RNG by first harnessing the self-chaotic dynamics of free-running broad-area vertical-cavity surface-emitting lasers (BA-VCSELs). The intense mode interaction within the broad-area cavity provides a robust foundation for ultrafast dynamics, allowing for high-security and high-speed RNG. Comparative analysis with a small-area quasi-single-mode VCSEL (QSM-VCSEL) confirms the efficacy of achieving high-speed RNG with hundreds of Gb/s from a single BA-VCSEL channel and 2 Tb/s from four channels as a proof-of-concept device. Given the easy fabrication and high scalability of VCSELs, this finding opens avenues for low-cost, massively parallel high-speed RNG chips with photodetector integration, unveiling opportunities for fields demanding unprecedented RNG rates and high levels of cybersecurity.