The laws of quantum physics can be used to secure communications between two distant parties in a scheme called quantum key distribution (QKD), even against a technologically unlimited eavesdropper. While the theoretical security of QKD has been proved rigorously, current implementations of QKD are generally insecure. In particular, mathematical models of devices, such as detectors, do not accurately describe their reallife behaviour. Such seemingly insignificant discrepancies can compromise the security of the entire scheme, especially as novel detector technologies are being developed with little regard for potential vulnerabilities.;In this thesis, we study how detector imperfections can impact the security of QKD and how to overcome such technological limitations. (Abstract shortened by UMI.). |