Reference-frame-independent quantum key distribution with passive decoy states

Reference-frame-independent quantum key distribution (RFI-QKD) enables the generation of secure keys when the reference frame suffers from unknown and slow drifting. To combat the potential photon-number-splitting attacks in RFI-QKD, the decoy-state method should be adopted. Generally, the decoy states are prepared by actively modulating different photon numbers of weak coherent pulses, which, however, may leak the side-channel information due to the imperfect modulation. To address this security loophole, we propose to implement the improved RFI-QKD protocol [Opt. Lett.47, 4219 (2022)10.1364/OL.470558], which bounds the eavesdropper's information analytically and is less sensitive to statistical fluctuations, with passive decoy states. By interfering two weak coherent pulses on a beam splitter and then directing one arm of the outputs into the single-photon avalanche detector (SPAD), different photon-number distributions of pulses can be prepared according to the response status of SPAD. To make this scheme more practical, we consider the inevitable afterpulses of the SPADs for both the transmitter and receiver, and construct an afterpulse-compatible analysis model. Simulation results show that our scheme can achieve good performance in practical scenarios.