High-speed phase-encoded quantum secure direct communication over 11.4 km heterogeneous free-space and fiber links

TL;DR

The paper demonstrates high-speed phase-encoded quantum secure direct communication over a hybrid 11.4 km link comprising free-space and fiber, addressing the historical challenge of phase stability in free-space channels. It employs a 1.25 GHz phase-modulated weak coherent-source and four-phase encoding implemented with a Faraday–Sagnac–Michelson interferometer within the STIKE protocol to realize quasi-QSDC across cross-medium links. The results show stable operation for nearly an hour with interference visibility close to 99% and QBER around a few percent, achieving a 4.22 kbps communication rate and enabling seamless cross-medium integration, with simulations suggesting feasibility beyond 30 km in satellite–ground scenarios. A cascaded-link model supports the practicality of cross-medium space–ground quantum networks, highlighting potential for chip-scale integration and compatibility with existing classical infrastructure.

Abstract

Robust quantum transmission is driving a new paradigm in space-ground quantum networking. Although phase encoding has been widely adopted in terrestrial fiber channels, it has long been considered unsuitable for free-space quantum communication. Here, we demonstrate phase-encoded quantum communication over 1400 m of urban free space. The system maintained stable operation for nearly one hour, achieving 99.07% interference visibility and an average quantum bit error rate of 2.38%. The free-space quantum states were directly coupled into the fiber and transmitted over an additional 10 km, confirming seamless interoperability across different media. We further show that turbulence-induced phase drifts between successive picosecond pulses can be effectively compensated. A cascaded-link model and numerical simulations indicate feasibility over free-space distances exceeding 30 km, underscoring the potential for satellite-to-ground quantum links. This work establishes the viability of phase encoding in free-space quantum networks, simplifying cross-medium integration and enabling compatibility with existing classical infrastructures.

Figures (6)

• Figure 1: Schematic of our experimental setup featuring a 1.4 km free-space atmospheric channel over a lake surface combined with a 10 km optical fiber link in Hefei, China. IM, intensity modulator; CIR, circulator; BS, beam splitter; FM, Faraday mirror; PBS, polarization beam splitter; FR, Faraday rotator; PM, phase modulator; ATT, attenuator; VOA, variable optical attenuator; SPD, single photon detector; WDM, wavelength division multiplexer; SL, synchronization laser; APD, avalanche photon diode.
• Figure 2: Interference visibility and QBER as functions of time. The average QBERs over the two experimental days were 3.61% and 2.38%, respectively. On the first day of operation the beam deviated from alignment at the 36th minute causing a link disconnection. The optical path was successfully recalibrated, which restored communication at the 45th minute.
• Figure 3: Transmission performance versus time. The average communication rates over the two days were 4.22 kbps and 3.90 kbps, the average key generation rates were 13.51 kbps and 30.42 kbps, the average key consumption rates were 21.18 kbps and 27.68 kbps, and the average key recovery percentages were 99.98% and 99.97%, respectively.
• Figure 4: Simulation (lines) and experimental (symbols) results for the rate. The triangular and circular markers indicate the experimental communication rate and key generation rate, respectively. The solid curve shows the relationship between rate and distance derived from our experimental parameters. The dotted curve is calculated using a detector efficiency of 80%, a telescope system conversion efficiency of $10^{-15.4/10}$, and a receiving telescope aperture of 1 m.
• Figure S1: Temporal variation of total loss in the 1.4 km free-space and 10 km fiber hybrid channel. The channel loss of the 10-km fiber is 2 dB, and the loss of its associated fiber optic adapter is 0.42 dB. The total channel loss is the sum of the free-space channel loss and the fiber channel loss (2.42 dB).