Abstract
The intrinsic unpredictability of measurements in quantum mechanics can be used to produce genuine randomness. Here, we demonstrate a random number generator where the randomness is certified by quantum contextuality in connection with the Kochen-Specker theorem. In particular, we generate random numbers from measurements on a single trapped ion with three internal levels, and certify the generated randomness by showing a bound on the minimum entropy through observation of violation of the Klyachko-Can-Binicioglu-Shumovsky (KCBS) inequality. Concerning the test of the KCBS inequality, we close the detection efficiency loophole for the first time and make it relatively immune to the compatibility loophole. In our experiment, we generate $1 \times 10^5$ random numbers that are guaranteed to have $2.4 \times 10^4$ bits of minimum entropy with a 99% confidence level.
Introduction: The Challenge of Genuine Randomness
In modern cryptography and complex simulations, the requirement for "genuine" randomness is paramount. Traditional classical computers are deterministic and can only produce pseudo-random numbers—sequences that appear random but are ultimately predictable if the initial seed is known. Quantum mechanics provides a solution through its inherent indeterminacy. However, a significant challenge remains: how can a user verify that a device is truly utilizing quantum processes rather than outputting pre-recorded or classically generated noise?
The gold standard for such verification is "self-certification." While Bell inequality violations (based on entanglement) are the most famous method, they are experimentally demanding and slow. This research explores an alternative path: Quantum Contextuality. Based on the Kochen-Specker theorem, contextuality posits that the outcome of a quantum measurement depends on the "context"—the other compatible measurements performed alongside it. This property can be tested on a single three-level system (a qutrit), offering a more efficient route to certified randomness.
Methodology: The KCBS Test on a Trapped Ion
The experimental setup utilized a single $^{171}\text{Yb}^+$ ion confined in a radio-frequency trap. Three internal energy levels of the ion were selected to represent a qutrit. To certify the randomness, the researchers performed a series of five specific quantum measurements ($A_1$ through $A_5$) arranged in a "pentagram" configuration.
In a non-contextual classical framework, these measurements must satisfy the Klyachko-Can-Binicioglu-Shumovsky (KCBS) inequality. A violation of this inequality serves as a mathematical proof of quantum contextuality. A key advantage of using trapped ions is the high measurement efficiency. Unlike photonic experiments, which often suffer from photon loss (the "detection efficiency loophole"), ions can be measured with nearly 100% efficiency. Furthermore, the team used sequential microwave pulses and fluorescence detection to minimize the "compatibility loophole," ensuring measurement settings remained consistent across contexts.
Results and Entropy Certification
The experiment involved $1 \times 10^5$ trials. The primary metric for success was the violation of the KCBS inequality. The researchers observed a KCBS expectation value of $\hat{L} = 3.594 \pm 0.034$. Given that the classical limit is 3, this result represents a violation of 18 standard deviations, providing robust evidence of quantum contextuality.
Crucially, the degree of violation allows for the calculation of a lower bound on min-entropy. Min-entropy is a rigorous measure of unpredictability; a higher value indicates that the generated bits are more secure against an adversary. Following the 2018 corrigendum to the original paper, the researchers certified that the raw sequence of $1 \times 10^5$ numbers contained at least $2.4 \times 10^4$ bits of genuine quantum randomness with a 99% confidence level. This generation rate is approximately four orders of magnitude faster than previous methods relying on Bell-certified entanglement.
Discussion and Future Directions
This work demonstrates that contextuality is a powerful resource for practical quantum technologies. By moving from two-particle entanglement to single-particle contextuality, the researchers achieved significant gains in efficiency and speed.
However, certain limitations remain. The approach operates under an "honest provider" assumption, meaning it protects against technical noise or device failure but not necessarily against a malicious manufacturer who might pre-program the device. Additionally, while the compatibility loophole was addressed through rigorous experimental design, perfectly closing it in sequential measurements remains a theoretical challenge. Future developments in high-speed electronics and optimized pulse sequences are expected to further increase the generation rates, bringing certified quantum randomness closer to widespread practical application.
Summary updated