Fiber-Optic Quantum Random Number Generator

By Dr. Noopur JainReviewed by Louis CastelNov 26 2025

Researchers developed a fiber-optic quantum random number generator that uses single-photon path superposition and a validated stochastic model to guarantee high-quality quantum randomness. Operating in a narrow 0.3–0.4 photon-per-pulse range, the system delivers secure, standards-aligned randomness already used to drive a quantum key distribution setup. The article was published in the journal Scientific Reports.

Image Credit: kuruneko/Shutterstock.com

A Robust Fiber Path to Quantum Randomness

The foundation of the research lies in the principles of optical quantum randomness, where the probabilistic nature of photon detection at a beam splitter forms the core mechanism. Historically, optical implementations of QRNGs have utilized free-space optics, which, while effective, often suffer from environmental instability and complexity. Fiber optics, on the other hand, offer high stability, minimal environmental interference, and ease of integration into existing telecommunication networks. The core idea involves a single-photon source, such as a pulsed laser attenuated to the single-photon level, followed by a 50/50 fiber optic beam splitter that creates a spatial superposition of photon paths. Detection at the two output ports corresponds to XORing the quantum event into a random bit, detecting a photon at port one encodes a '0,' while detection at port two encodes a '1.' This configuration capitalizes optical components like fiber couplers, attenuators, and single-photon detectors, whose characteristics directly influence the quality and rate of randomness.

The Current Study

The experimental setup employs high-precision optical fiber components meticulously characterized prior to operation. It starts with a pulsed laser emitting at 640 nm, which is pulsed at rates up to 40 MHz. The laser beam is directed into a variable optical fiber attenuator (VOA), which allows the adjustment of photon flux to achieve the desired mean photon number per pulse, typically less than 1 photon. This level of attenuation ensures quantum behavior dominates and minimizes multi-photon events. The attenuated beam then enters a 50/50 fiber optic beam splitter, which divides the photon stream into two spatial paths. The optical fibers used are standard single-mode fibers with FC/APC connectors to minimize reflections and backscattering.

Post-splitting, each output fiber leads to a high-efficiency single-photon detector, specifically Laser Components COUNT-10C-FC modules. These detectors are designed for low dark counts and high detection efficiencies, crucial for high-quality randomness generation. To achieve a larger dynamic range of attenuation, additional fixed fiber attenuators are employed right before the beam splitter, allowing precise control over the photon flux. The entire optical setup is enclosed within a controlled environment to prevent temperature-induced drifts and mechanical vibrations.

The stochastic model developed incorporates the probability distribution of photons following a Poisson process, with parameters influenced by fiber losses, splitting ratios, detector efficiencies, dark counts, and afterpulsing effects. These models allow the prediction of both raw and post-processed randomness and help establish optimal operating parameters that maximize entropy and system efficiency. The optical losses are quantified by measuring the transmission through different fiber segments and components using standard optical power meters, though some losses inside fiber connections could not be precisely measured but are included in the probabilistic framework.

Results and Discussion

The optical implementation produced reliable, high-quality random bit streams, validated using classical statistical tools, including the NIST Statistical Test Suite, in line with standards such as BSI AIS 20/31 and NIST SP 800-90B. The measurements indicated a stable system where the probability of detecting a photon in either detector was nearly balanced and aligned with theoretical predictions. The fiber-based layout supports stable operating conditions, benefitting from well-characterized splitting ratios and attenuation.

One of the key optical parameters analyzed was the photon flux, directly manipulated through the variable optical attenuators. The experiments confirmed that the randomness quality significantly depends on maintaining the mean photon number below one. As optical losses and detector efficiencies influence the probability of photon detection, diligent calibration ensured high entropy per bit. The experimental data aligned closely with the stochastic model's forecasts, validating its accuracy in capturing the impact of fiber losses, dark counts, and afterpulsing.

Furthermore, the researchers observed that the optical system's stability allowed for consistent operation over extended periods, a vital feature for practical QRNG deployment. Despite some inherent losses, the fiber-based scheme achieved an operational efficiency of approximately 15%, translating into a random number generation rate of about 8 kb/s. While this rate is modest compared to electronic-based generators, the optical approach provides excellent security and immunity to parasitic electronic noise.

Conclusion

This study presents a practical and reliable fiber-optic quantum random number generator that leverages the spontaneous spatial superposition of single photons. The findings underscore how existing optical fiber infrastructure provides a strong foundation for secure and accessible quantum random number generation. Although current challenges, such as optical losses and limited detector efficiency, affect overall speed, ongoing improvements in fiber technology, detector capabilities, and system calibration are expected to drive significant gains. Taken together, this work reinforces the optical fiber-based approach as a viable path toward scalable, stable, and secure sources of quantum randomness, with key implications for future quantum communication and cryptographic systems.