Quantum Time Synchronization of Star Networks

Abstract

We extend the single source approach of Valencia et al in order to synchronize the clocks of an N user start network, connected both through fiber and in free space. Entangled photon pairs from a centralized SPDC source are distributed through a 1 by N splitter to four remote users arranged in a star topology. Using commercially available single photon detectors and time taggers, we achieve median time precision of 50 ps for atomic oscillators and 20 ps for GPS displayed oscillators in our Kalman models. Thus, we achieve three order of magnitude improvement over GPS alone. By monitoring the drift fo the correlation peaks over time, we also extract the frequency skew between users's local clocks to 35ps/s precision. From these measurements, e3ach user can compute its offset and drift relative to every other user, achieving full network synchronization without a central clock.

Figures (2)

• Figure 1: A central hub (Charlie) produces entangled photons and the photons are split up using an N splitter. Most likely, the photons leave separate ports of the splitter and two photons reach different users. In the figure, the pieces of the same photon pair are indicated by the same color. It should be noted, that the photon pairs are much further spaced in the time domain.
• Figure 2: A central hub produces entangled photon pairs and sends them to an N=8 splitter. In this case, four ends are terminated. The other four go to any of four detectors located between two stations. The two stations have separate clocks, but the detectors on a single station share the same time tagger and therefore clock. We use a single pulse for epoch alignment due to the difficulty in beginning the two separate TIA's recordings at the same time. In the field where this would happen continuously, epoch alignment is likely unnecessary.