Ultra-high-rate detection of entangled photon pairs

TL;DR

This work tackles the bottleneck of detector dead time in high-rate entangled-photon experiments by using a 16-pixel SNSPD array with SFQ readout, driven by a 5-GHz sequential time-bin photon-pair source. It demonstrates coincidence rates exceeding 3 Mcps in both two-photon interference and CHSH inequality tests, while maintaining substantial entanglement visibility (71.4% at high rate) and a CHSH value of $S=2.05$ at multi-Mcps rates. The results establish the first time-bin entangled-photon-pair detection at multi-Mcps rates, illustrating a viable path toward high-speed, scalable photonic quantum information processing. The study discusses practical enhancements (jitter reduction, CFD integration, higher clock rates, and improved sources) and the potential for real-time FPGA-based feedback to enable advanced quantum protocols.

Abstract

The high-rate detection of entangled photons is essential for advancing photonic quantum information processing. Although several experimental demonstrations have been reported, the achievable coincidence rates have so far remained limited. One of the main bottlenecks arises from the dead time of single-photon detectors, which constrains coincidence detection at high photon-pair generation rates. In this work, we employ 16-pixel superconducting nanowire single-photon detectors (SNSPDs) to mitigate the impact of detector dead time. Consequently, we achieve coincidence rates exceeding 3 million counts per second (Mcps) in two-photon interference and CHSH inequality experiments using 5-GHz clocked sequential time-bin entangled photon pair source. To the best of our knowledge, this is the first demonstration of multi-Mcps coincidence detection of entangled photons, paving the way for high-speed entangled-photon-based quantum information processing.

Figures (10)

• Figure 1: Coincide probability per pulse as a function of average photon pair per pulse. $D$ is the dead time divided by the interval of the pulses.
• Figure 2: Experimentally obtained coincidence-to-accidental ratio (CAR) plotted as a function of the estimated average number of correlated photon pairs per pulse, $\mu$. The dashed line indicates the theoretical CAR assuming negligible dark counts.
• Figure 3: Two-photon interference fringes and corresponding histogram of single-photon detection events in the signal channel at a low average photon-pair rate ($\mu = 0.001$). The measured visibility was 96.6%, and the peak coincidence and signal count rates were 47 kcps and 532 kcps, respectively.
• Figure 4: Two-photon interference fringes and corresponding histogram of single-photon detection events in the signal channel at a high average photon-pair rate ($\mu = 0.094$). The visibility was 71.4%, with a peak coincidence rate of 3.3 Mcps and a signal count rate of 47 Mcps.
• Figure 5: Measured visibility and coincidence rate as functions of the average number of photon pairs per pulse. As the photon-pair generation rate increases, the visibility decreases due to increased timing jitter in the single-photon detectors. Despite this degradation, a coincidence rate exceeding 3 Mcps was achieved with visibility above 70.7%, sufficient for violating Bell's inequality.