Continuous Variable Quantum Key Distribution over Long Distances

Quantum key distribution (QKD) is fundamentally different from most classical key distribution schemes, such as Diffie-Hellman key exchange, in the sense that no computational complexity assumption is required on the power of adversaries to prove its security. QKD relies on basic laws of quantum physics and it is proven that it can enable highly secure data communication. Such achievements, however, are facing technological problems that have to be resolved in order to provide a viable solution to a large group of customers. While there are discrete-variable QKD schemes, which rely on encoding data in discrete degrees of freedom, such as polarization of single photons, in this thesis, we focus on the continuous-variable QKD (CV-QKD) protocols, in which data is encoded on the quadratures of light. Currently, one of the major drawbacks of CV-QKD is its poor performance at long distances. Nevertheless, such a limitation in CV-QKD can be overcome with the assistance of quantum repeaters that rely on entanglement distillation via noiseless linear amplifiers (NLAs). Such systems can, in principle, offer large secret key rates over long distances. In this thesis, we aim to provide a realistic analysis of a CV-QKD protocol running over quantum scissors (QSs) as realistic NLAs. We will report the obstacles that one could face in realizing CV-QKD in such a scenario. A review of CV-QKD and QS-based NLAs will be given, based on which QS-assisted CV-QKD is proposed. We, particularly, focus on the modelling of the QSs' structure and their effect on the secret key rate aiming to find operational regimes where the performance of the QKD scheme is enhanced. This study paves the way for implementing long-distance CV-QKD protocols that rely on QS/NLA devices over CV quantum repeaters. In this thesis, we also consider and account for a realistic analysis of a CV-QKD protocol with non-Gaussian modulation, which is assisted by the means of QSs. We will show that, while we have to deal with similar obstacles as in the Gaussian modulation, we can potentially improve performance of the non-Gaussian modulation protocol. As an alternative approach to extend the secure distance of CV-QKD protocols, the last part of this thesis is devoted to presenting realistic threat models for satellite QKD, wherein we consider several eavesdropping scenarios by limiting eavesdroppers' access to the trusted ground and/or satellite stations. In such scenarios, the eavesdropper has only limited access to the sender and/or receiver stations. For example, we will explore the case where an eavesdropper can only receive an attenuated version of the transmitted signals. As well, we will focus on the case where Eve's signals would reach the receiver via a lossy channel inaccessible to the eavesdropper. We show that, in the case of both Gaussian and non-Gaussian protocols, this limitation would allow trusted parties to achieve higher key rates than what can be achieved when unrestricted eavesdropping is possible.