Optimizing Continuous-Wave-Pumped Entanglement-based QKD in Noisy Environments

Quantum key distribution (QKD) has emerged as a promising solution to protect current cryptographic systems against the threat of quantum computers. As QKD transitions from laboratories to real-world applications, its implementation under various environmental conditions has become a pressing challenge. Major obstacles to practical QKD implementation are the loss of photons in the transmission media and the presence of extreme noise, which can severely limit long-range transmission. In this paper, we investigate the impact of extreme noise on QKD system parameters, including timing jitter, rate-dependent timing shifts, changes in effective detector dead time, and rate-dependent detection efficiency. Contrary to manufacturers’ specifications, which assume these parameters to be constant, we demonstrate that these parameters exhibit significant variations in extreme noise conditions. We show that changes in these parameters play a key role in determining system performance in noisy environments. To address these nonidealities, we develop a model that adapts to detector-dependent timing distortions and recovery effects. In particular, our model is independent of source parameters and can be implemented using data from the detection unit. Our results show that the model enables reliable characterization and optimization of QKD performance under strong noise.

💡 Research Summary

The paper investigates how extreme optical noise affects continuous‑wave (CW) pumped entanglement‑based quantum key distribution (QKD) systems, focusing on the non‑linear, rate‑dependent behavior of single‑photon avalanche diodes (APDs). Using a 405 nm CW laser to pump a PPKTP crystal, the authors generate degenerate 810 nm photon pairs via spontaneous parametric down‑conversion (SPDC). One detection channel is deliberately mixed with broadband thermal light from a tungsten‑halogen source, creating a highly asymmetric noise environment while keeping the hardware identical in both arms.

Measurements with identical silicon APDs (Excelitas SPCM‑A QRH‑14‑FC) reveal four distinct noise‑induced effects once the detection rate exceeds ~1 Mcps: (1) a systematic shift of the coincidence peak (timing shift), (2) broadening of the peak (increase in full‑width at half‑maximum, FWHM), (3) a rising baseline due to accidental coincidences, and (4) a reduction of peak amplitude reflecting rate‑dependent efficiency loss. The authors attribute the timing shift to dead‑time‑induced truncation of inter‑arrival times, which skews the detector’s temporal response and consequently displaces the convolution peak of the two detectors. The FWHM growth follows a saturating recovery model with a characteristic rate (R_0 = 1.43\times10^8) Hz, indicating a common underlying recovery mechanism for both shift and broadening.

To capture these dynamics, the paper proposes a detector‑centric model that incorporates individual timing functions (W_i, W_j), dead time (t_d), and a rate‑dependent width coefficient. The core expression (Eqs. 7‑8) separates genuine coincidences from accidental ones and allows the extraction of detector parameters directly from time‑tag data, without any dependence on source characteristics. This model is then extended to a full two‑basis BBM92‑type QKD protocol, enabling analytical optimization of the coincidence window (\tau_{cw}).

Simulation results demonstrate that, even when the noise photon flux reaches 10 Mcps, selecting an optimal window (≤ 12 ns) keeps the quantum bit error rate (QBER) below 5 % while maintaining a secure key rate (SKR) above 0.5 kbps. Thus, the approach substantially mitigates the performance degradation typically observed with silicon APDs under harsh noise, offering a cost‑effective alternative to superconducting nanowire detectors for free‑space or satellite‑ground links.

Beyond performance gains, the model provides a real‑time monitoring and correction framework for detector parameters, paving the way for automated QKD network management in environments where spectral filtering alone is insufficient. In summary, the work delivers both experimental evidence and a robust theoretical toolset for optimizing CW‑pumped entanglement‑based QKD in noisy real‑world conditions, marking a significant step toward practical quantum‑secure communications.