The Signal
Photon-Pairing Spectroscopy Reconstructs Signals Across 400 km of Fiber
A photon-pairing method reconstructed intensity spectra over a 400-kilometer fiber link, with a signal-to-noise ratio above 23.5 decibels at very low receiving power. The study reports that second-order photon correlations avoid interferogr
Quantum.world Team · Written with AI · fact-checked ·
A 400-kilometer fiber-link demonstration
A photon-pairing method has reconstructed intensity spectra across a 400-kilometer optical-fiber link, with a reported signal-to-noise ratio above 23.5 decibels. The Nature Communications study reports that result at an average receiving power of just 5.5 × 10⁻¹⁵ watts per frequency-comb line. The work, published October 31, addresses a problem that grows with distance: phase fluctuations can disrupt the interferograms used to rebuild spectra from photon-counting measurements.
The researchers instead use second-order correlations between photons. That approach, they write, removes the need to align interferometer phase while allowing longer integration times. The experiment demonstrates intensity-spectrum reconstruction in a distributed configuration; it does not, by itself, establish a field-deployed sensing network.
Correlations replace phase alignment
Photon-counting spectroscopy can work with extremely weak light, but conventional schemes build a spectrum by coherently accumulating interferograms. When the relative phase of interfering beams shifts, those measurements become difficult to combine, a particular challenge for sensing over distributed links. The authors’ method shifts the measurement to correlations between detected photons rather than relying on first-order interference.
“Here, we introduce a photon-pairing scheme that exploits second-order correlations rather than relying on first-order interference, thereby removing the need for phase alignment of interferograms while enabling long integration times.”
The authors — Jinping Lin, Lai Zhou and their co-authors — write.
The paper also reports a different scaling for signal-to-noise ratio. In the authors’ account, SNR increases linearly with photon count rate in their approach, rather than following the square-root dependence they describe for photon-counting techniques. That distinction matters at very low light levels, where collecting enough useful signal is a central constraint.
What the 400 km result establishes
The headline distance is not the only measure of the demonstration. The reported receiving power—5.5 × 10⁻¹⁵ watts per comb line—shows the method reconstructing spectra when little optical power reaches the receiver, while the SNR above 23.5 dB quantifies the reported quality of that reconstruction. The study also describes the scheme as resilient to fiber-induced fluctuations.
Those results point to a possible route around phase stability demands in long-haul dual-comb spectroscopy, where two optical frequency combs support spectral measurement. The authors identify frequency metrology, remote sensing and quantum networking as potential areas of relevance. The reported experiment, however, is specifically a 400-kilometer fiber-link demonstration of intensity spectra, not a test of every proposed application.
A practical trade-off for distributed sensing
For a distributed measurement system, avoiding phase alignment could simplify one demanding part of operating across a long link: maintaining coherence between interfering beams over time. Photon pairing does not eliminate the need for sensitive photon detection or establish performance under all environmental conditions; the paper’s result is evidence for the technique under the conditions tested. Its key claim is narrower and measurable: spectra reconstructed over 400 kilometers, at femtowatt-scale power per comb line, with an SNR exceeding 23.5 dB.
The paper appeared online in Nature Communications on October 31, 2026. Its central result is a specified link length and received power, paired with a measured SNR—not a claim that distributed spectroscopy has already moved into operational networks.