Measurement-Device-Independent Quantum Key Distribution (MDI-QKD) enhances security by removing vulnerabilities associated with detector side channels. Real-world implementations of measurement device independent (MDI) QKD face practical challenges such as channel asymmetry and physical imperfections, which degrade the visibility of Hong-Ou-Mandel (HOM) interference, an essential factor in determining the secure key rate. In this work, we evaluate the performance of differential phase encoded (DPE) MDI-QKD under realistic, asymmetric conditions. We analyse the effects of polarisation and pulse-width mismatches caused by channel asymmetry and identify tolerable limits for secure key generation. We introduce a plug-and-play DPE-MDI-QKD architecture that mitigates polarisation fluctuations using Faraday mirrors and also compensates for pulse-width mismatches. Furthermore, we propose an improved sifting scheme that raises the sifting efficiency from
94to
32, improving the secure key generation rates by 50% and making it especially suitable for plug-and-play systems. These results offer practical insights into enhancing the robustness and efficiency of DPE-MDI-QKD, enabling more secure quantum communication over imperfect and asymmetric quantum channels.
This paper proposes a measurement-device-independent quantum key distribution (MDI-QKD) scheme based on differential phase encoding. The differential phase shift MDI-QKD (DPS-MDI-QKD) couples the advantages of DPS-QKD with that of MDI-QKD. The proposed scheme offers resistance against photon number splitting attack and phase fluctuations as well as immunity against detector side-channel vulnerabilities. The design proposed in this paper uses weak coherent pulses in a superposition of three orthogonal states, corresponding to one of three distinct paths in a delay-line interferometer. The classical bit information is encoded in the phase difference between pulses traversing successive paths. This 3-pulse superposition offers enhanced security compared to using a train of pulses by decreasing the learning rate of an eavesdropper and unmasking her presence with an increased error rate upon application of intercept and resend attack and beamsplitter attack. The proposed scheme employs phase locking of the sources of the two trusted parties so as to maintain the coherence between their optical signal, and uses a beamsplitter (BS) at the untrusted node (Charlie) to extract the key information from the phase encoded signals.