Toshiba creates Quantum Random Number Generation ready for mass manufacture
- Toshiba develops chip quantum random number generator ready for mass production, bridging the gap between lab implementation and mass commercial deployment.
- Quantum chips manufactured using standard semiconductor processes and packaged ready for serial assembly.
- Represents a significant advance towards mass deployment of quantum random number generators.
Cambridge, UK, 15th May 2024: Toshiba Europe Ltd today announced it has developed a fully on-chip quantum random number generator ready for serial assembly. This advance will enable the mass manufacture of this quantum security technology, bringing its application to a much wider range of real-life scenarios.
Random numbers are essential for a wide variety of applications from computer simulations to lotteries. With the ever-growing threat that quantum computers pose to asymmetric cryptography, it has become of critical importance to find methods to generate random numbers securely and at a high rate. Quantum random number generators (QRNGs) which exploit the intrinsic probabilistic nature of quantum processes to generate unpredictable random numbers have shown great promise in solving this problem.
Proof-of-principle implementations of numerous potential QRNG architectures have already been demonstrated. However, to be a viable alternative to existing random number generators, unpredictability cannot be the only figure of merit. The ease of mass-manufacture, hardware integration and generation rate must also be considered.
Manufacturing advances
For the use of QRNGs to become widespread, the gap between one-off lab implementations and mass-manufacturability must be closed. This requires implementations that are compatible with mass production. It is also important that the size, weight, and power consumption are optimised in the design process. Chip-based solutions alone promise to fulfil all these requirements; therefore, their development is essential to enabling mass market applications, which will be integral to the realisation of a quantum-ready economy.
Toshiba has developed techniques to fully integrate the optical entropy core (OEC) of its QRNG into miniature semiconductor chips. These chips have no optical inputs or outputs which makes them easier to embed in a hardware system as they do not require alignment of a fibre with the chip. Additionally, they are much smaller, lighter and consume less power than their fibre optic counterparts. Significantly, they can be fabricated in large numbers in parallel on the same semiconductor wafer using standard techniques used within the semiconductor industry.
In addition, Toshiba has developed techniques to encase its OEC in a package compatible with surface-mount technology, which is ubiquitous in the high-tech electronic printed circuit assembly industry. This advance is key in enabling serial production of Toshiba’s QRNGs.
Andrew Shields, Head of Quantum Technology at Toshiba Europe, remarked: “Quantum processes are entirely unpredictable and therefore ideally suited for generating high-quality random numbers. Implementing the design in a semiconductor chip allows the mass manufacture of these quantum random number generators, with many potential applications in cryptography, numerical modelling and gaming.”
This work was partially funded by InnovateUK through the project AQuRAND. Full details of the advancement have been published in the scientific journal, Nature Electronics.
Technical Summary
QRNG research has so far largely been confined to one-off laboratory experiments controlled by high end lab equipment, making it unsuitable for widespread deployment. The post-processing, which is key to the security of QRNGs, is rarely performed live.
The purpose of this work was to develop a fully-fledged QRNG comprising an OEC encased in a standard quad flat no-lead package, assembled on a printed circuit board. This PCB features not only all the electronics necessary to provide the driving signals for the OEC and measure its output, but also those required to post-process this output live, stably generating unpredictable, uniformly distributed random numbers.
To ensure the security of its final output, the QRNG performs health tests on the OEC’s output to verify that it continuously operates as expected as well as calculating the secure generation rate that it can achieve in real time. If this rate drops, the QRNG can automatically adjust the post-processing to ensure that the final output remains unpredictable. To showcase the reproducibility and stability of this QRNG technology, the performance of eight devices was benchmarked for over a week of continuous gigahertz operation, including one device deployed in a quantum key distribution system for 38 days. Despite operating in uncontrolled environments, the physical randomness output by the OECs in these QRNGs featured minimal variations, resulting in steady and robust random number generation at 2 Gbps. Such high rate, stable randomness generation is key for cryptographic applications such as quantum key distribution.