Having addressed the measurement side of QKD systems, another important vulnerability left is the QKD sources. Recall that in QKD, Alice needs to send random bits (e.g. 0101) encoded in the polarization of her photons. This often requires active modulators inside the sources that convert digital signals to photon polarization. However, such modulators are susceptible to side-channel information leakages or even attacks from Eavesdroppers (like a Trojan-Horse attack).
A clever way to remove these side-channels is to avoid the usage of active modulators altogether. Suppose one uses special light sources (e.g. based on photon interferometers) that naturally output various polarizations, he/she can then simply observe the signals and post-select the cases where the polarization happens to be in the desired states. This idea is called “passive” QKD (initially proposed by researchers in 2009), which can greatly improve the implementation security of QKD sources. However, for more than 10 years, passive QKD was limited to partial implementations (i.e. some components in the source still have to remain active).
We proposed the world’s first fully passive QKD source [1] that can perform the most popular BB84 QKD protocol without using any active modulators. We also extended it to more protocols such as MDI-QKD [2] and its variant Twin-Field QKD [3]. In the future, we hope to apply this passive QKD design to more potential protocols of interest, as well as look for a universal description for the design and security proof of passive protocols.
[1] W Wang, et al. "Fully passive quantum key distribution." Physical Review Letters 130.22 (2023): 220801.
[2] J Li, W Wang, HK Lo "Fully passive measurement-device-independent quantum key distribution." Physical Review Applied 21.6 (2024): 064056.
[3] W Wang, R Wang, HK Lo. "Fully-passive twin-field quantum key distribution." arXiv preprint arXiv:2304.12062 (2023).