One Quadrature Is Enough: A Simpler Quantum Key Distribution System Hits 0.364 Mbit/s Over 12 km

First integrated experimental demo of 1D Gaussian-modulated CV-QKD with independent lasers, finite-size security, and full post-processing. Simpler transmitter, real key rates.

Imagine you're sending a secret message by hiding it in the wobble of a guitar string. Conventional quantum key distribution (CV-QKD) encodes secrets in both the amplitude and phase of a light wave — two independent wobble directions. That means your transmitter needs to precisely control both, calibrate both, and defend both against eavesdroppers. This paper asks: what if you only used one wobble direction and let the other serve as a security watchdog? The committed claim is straightforward: this is the first integrated experimental demonstration of unidimensional (1D) Gaussian-modulated CV-QKD that combines independent-laser coherent detection, practical information reconciliation, finite-size parameter estimation, and privacy amplification in a single end-to-end system. The distinction matters because previous 1D proposals either stayed theoretical or demonstrated only fragments of the chain — not the complete pipeline you'd need for a deployable system. On the performance ladder, the system delivers 1.421 Mbit/s under asymptotic security assumptions and 0.364 Mbit/s under the more realistic finite-size treatment, over 12 km of standard single-mode fiber. These numbers sit below the best reported 2D CV-QKD rates at comparable distances, which is expected — the whole point is trading peak throughput for transmitter simplicity. The paper does not name a specific 2D SOTA rate for direct comparison, which is a gap. The honest framing here is 'viable at reduced complexity,' not 'faster than existing systems.' Architecturally, this is a prepare-and-measure CV-QKD protocol in the Gaussian-modulation family, using a local local oscillator (LLO) scheme where transmitter and receiver lasers are independent — no shared optical reference, which removes a major practical vulnerability. The post-processing chain covers digital signal processing, multidimensional reconciliation, physicality-constrained covariance estimation, and Toeplitz-hashing privacy amplification. The security proof leverages the unused conjugate quadrature as a constraint on an eavesdropper's information, which is the key theoretical mechanism enabling the 1D simplification. Integrity is solid for an experimental demonstration. This is a real physical experiment over real fiber, not a simulation. The finite-size treatment is the more conservative (and honest) of the two reported rates, and the authors report both rather than burying the lower number. However, the benchmarking is self-referential — the baseline is essentially 'does 1D work at all?' rather than a head-to-head comparison with named 2D implementations at the same distance and detector setup. No code release is mentioned. The milestone question is about distance and rate scaling. At 12 km and 0.364 Mbit/s finite-size, the next meaningful threshold is demonstrating this architecture at metropolitan distances (25-50 km) with key rates still above zero under finite-size analysis. That's where the complexity advantage of 1D modulation would start to matter commercially — if you can build a cheaper, simpler transmitter that still delivers usable key rates at city-scale distances, you have a product argument. The obvious next experiment the authors didn't run: testing at longer distances (25+ km) and under higher channel loss to find where the 1D rate drops to zero versus a 2D system. The likely reason is (a) — this is a first demonstration establishing the baseline, and longer-distance results will come in a follow-up once the post-processing chain is optimized. The paper explicitly frames itself as establishing a 'performance-complexity trade-off baseline,' which signals they know the distance extension is next.