Optimizing Continuous-Variable Quantum Key Distribution with Phase-Shift Keying Modulation and Postselection

Abstract

A numerical security proof technique is used to analyse the security of continuous-variable quantum key distribution (CV-QKD) protocols with phase-shift keying modulation against collective attacks in the asymptotic limit. We argue why it is sufficient to consider protocols with a maximum number of eight signal states and analyse different postselection strategies for protocols with four (QPSK) and eight (8PSK) signal states for untrusted ideal and trusted nonideal detectors. We introduce a \emph{cross-shaped} postselection strategy, and show that both cross-shaped and radial and angular postselection clearly outperform a radial postselection scheme (and no postselection) for QPSK protocols. For all strategies studied, we provide analytical results for the operators defining the respective regions in phase space. We outline several use-cases of postselection, which can easily be introduced in the data processing of both new and existing CV-QKD systems: Motivated by the high computational effort for error-correction, we studied the case when a large fraction of the raw key is eliminated by postselection and observed that this can be achieved while increasing the secure key rate. Postselection can also be used to partially compensate the disadvantage of QPSK protocols over 8PSK protocols for high transmission distances, while being experimentally less demanding. Finally, we highlight that postselection can be used to reduce the key rate gap between scenarios with trusted and untrusted detectors while relying on less assumptions on Eve's power.