A proof-technique-independent framework for detector imperfections in QKD

Abstract

The security of Quantum Key Distribution (QKD) protocols is theoretically established using idealised device models. However, the physical implementations upon which practical security relies inevitably deviate from these ideals. This thesis develops a rigorous and versatile framework to address a subclass of such deviations: detector imperfections.

This framework, termed ’noise channels’, is independent of security proof technique. This approach recasts imperfections as a quantum channel preceding an idealised measurement process. By granting the eavesdropper control over this channel, the security analysis is simplified to an ideal scenario, with the effects of the imperfections mathematically contained within a well-defined parameter. The utility and versatility of the framework are demonstrated through applying it to the postselection technique, and for phase error estimation. The application to phase error estimation is an improvement over past analyses which either assumed qubit detection setups, IID attacks, or required hardware modifications. We observe a remarkably high tolerance to imperfections when using the postselection technique.

Finally, we extend the framework to address cross-round correlations, providing a methodology to prove security against detector memory effects such as afterpulsing and dead times. This work thus establishes a structured and powerful toolkit for analysing detector imperfections in practical QKD systems, unifying their treatment across different security proof techniques and advancing the development of robust implementation security.