Abstract
In continuous-variable quantum information processing, detectors are necessarily coarse grained and of finite range. We discuss how, especially, the latter feature is a bug and may easily lead to overoptimistic estimates of entanglement and of security when missed data outside the detector range are ignored. We demonstrate that a straightforward worst-case analysis is sufficient to avoid false positive statements about entanglement. We show that, for our worst-case analysis, entropic separability or security criteria are much superior to variance-based criteria.
- Received 20 February 2013
DOI:https://doi.org/10.1103/PhysRevA.88.042326
©2013 American Physical Society