Figure 1

(Color online) Orientation of different detector polarizations. Coincidences $(1,1^{\prime })$ and $(2,2^{\prime })$ are used for key generation, while coincidences between any of (3,4,5,6) and any of $(1^{\prime },2^{\prime },3^{\prime },4^{\prime })$ will be used for testing a CHSH inequality for various settings.Reuse & Permissions

Figure 2

(Color online) Experimental setup. Polarization-entangled photon pairs are generated via parametric down conversion pumped by a laser diode (LD) in a nonlinear optical crystal (BBO) with walk-off compensation (WP, CC) into single mode optical fibers (SMF). A free-space optical channel for one detector set (Bob) is realized using small telescopes on both sides (ST, RT). Spectral filtering is implemented with a pinhole (PH) which is imaged completely onto the active area of the photodetectors with a relay lens (R) through a spectral filter (F). Both parties perform polarization measurements in bases randomly chosen by beam splitters (B1–B3), and defined by properly oriented wave plates (H1–H3) in front of polarizing beam splitters (PBS) and photon counting detectors. Photoevents are registered separately with time stamp units (TU) connected to two PC linked via a classical channel.Reuse & Permissions

Figure 3

(Color online) Experimental results in a key distribution experiment implementing an E91 protocol. (a) shows total (upper trace) and accidental (lower trace) detected coincidence rates between Alice and Bob, (b) the error ration in the raw key, (c) the degree of violation of a CHSH-type Bell inequality, (d) the final key rate after error correction and privacy amplification. The experiment was terminated by a storm misaligning a telescope of the optical link at 5 am.Reuse & Permissions