Quantum key distribution overcoming practical correlated intensity fluctuations

Introduction

Quantum key distribution¹ (QKD) stands at the forefront of secure communication protocols², as it enables two remote users, Alice and Bob, to share secret keys whose security is guaranteed by the principles of quantum mechanics^{3, 4, 5–6}. However, QKD security proofs typically extract secret bits from the raw data originating from single-photons, and on-demand high-quality single-photon sources at telecom wavelengths are not available yet. A popular solution to solve this pressing issue is the decoy-state method^{7, 8–9}, which provides the same secret key rate scaling as single-photon sources by means of using laser sources emitting phase-randomized weak coherent pulses (PRWCPs). Indeed, this technique is a standard tool in current QKD implementations^{10, 11, 12–13}.

An important breakthrough direction for advancing QKD is to increase its secret key rate, for which fast operating QKD systems are being developed^{10, 11, 12, 13, 14, 15, 16–17}. However, the implementation security problem¹⁸ brought by high-clock-rate decoy-state QKD systems cannot be ignored. Due to memory effects in the devices and the electronics that control them, high-clock-rate decoy-state QKD systems face a troublesome implementation security problem: intensity correlations^{16,19, 20, 21, 22, 23, 24–25}. This means that the intensity setting of any given round may influence the actual intensity emitted in subsequent rounds, resulting in a partial distinguishability of the intensity settings. This breaks a core assumption of the decoy-state method, posing an underestimated threat to the security of QKD^{24, 25–26}. To address this problem, various security analyses have been proposed^21,24,25. However, practical devices^{27, 28–29} often struggle to meet the criteria set by these analyses (such as the magnitude of the correlations) leading to low or even vanishing secret key rates.

Here, we solve this crucial limitation by proposing an approach that we name intensity-correlation-tolerant QKD, which is capable of mitigating the intensity correlations problem in QKD devices. By adding a local monitor, our protocol enables common devices to achieve notably higher secret key rates and longer transmission distances than previous solutions in the presence of this type of correlations. Importantly, we experimentally demonstrate the feasibility and effectiveness of our approach. This advancement is a significant step towards loophole-free and high-performance QKD.

Results