Every conventional optical filter operates the same way: it lets certain wavelengths through and blocks others, exploiting the material’s absorption or interference properties. Color filters, polarizers, and dichroic mirrors all rely on what light does as individual photons.
But light has properties that only emerge collectively. A stream of photons from a laser behaves differently from one emitted by a thermal source, not in color or brightness, but in how the photons cluster. Laser photons arrive at random intervals (a property called second-order coherence g⁽²⁾(0) = 1), while thermal light shows photon bunching (g⁽²⁾(0) = 2), and even stronger correlations appear in certain quantum sources. These differences encode the light’s quantum statistics, and until now, no material could distinguish them.
A team led by Chao You has changed that. Writing in Nature, they introduce quantum statistical plasmonic metacrystals, nanostructured gold surfaces that filter light based on its g⁽²⁾(0) value rather than its wavelength. The work was accompanied by a News & Views article by Sebastian Golat of King’s College London, who called it the first demonstration of a material that manipulates the collective quantum properties of light.
The metacrystal as a quantum sieve
The device consists of 100 gold nanoantennas (200 by 400 nanometers each) arranged on a 110-nanometer gold film on glass, spaced 1 micrometer apart, with coupling input gratings. The structure supports surface plasmon resonances, collective electron oscillations, whose near-field interactions between neighboring nanoantennas produce interference patterns that depend on how the arriving photons are correlated.
The key insight is borrowed from solid-state physics: just as a semiconductor has electronic band gaps that forbid certain electron energies, the plasmonic metacrystal creates forbidden statistical bands for certain values of g⁽²⁾(0). Light whose quantum statistics fall within an allowed band passes through essentially unchanged. Light in a forbidden band is filtered, its coherence is modified toward the nearest allowed state.
Golat, writing in the accompanying News & Views, noted that the concept was predicted theoretically by Mouloudakis and Lambropoulos in 2018 (Physical Review A, vol. 97, 053413), but You’s team provided the first experimental realization.
Watching the filter at work
The team tested the metacrystal with 13 different multiphoton sources spanning the range from coherent (g⁽²⁾(0) = 1) to superthermal (g⁽²⁾(0) = 3), generated by focusing a 780-nanometer continuous-wave laser on rotating ground glass and collecting the scattered light at various positions.
The results matched the theoretical predictions:
- Coherent light (g⁽²⁾(0) = 1) passed through an allowed band and emerged unchanged.
- Thermal light (g⁽²⁾(0) = 2) also fell within an allowed band and propagated freely.
- Superthermal light with g⁽²⁾(0) = 2.15 fell within a forbidden band and was filtered, emerging with g⁽²⁾(0) = 2.58, shifted toward the nearest allowed state.
- Intermediate states like g⁽²⁾(0) = 1.25 also encountered forbidden bands and were modified to g⁽²⁾(0) = 1.50.
The behavior is determined entirely by the geometry and arrangement of the nanoantennas. Aligned meta-atoms produce indistinguishable multiparticle interference (narrower bands), while differently oriented ones widen the bands. By controlling these parameters, the team demonstrated a deterministic route to engineering quantum statistical transport.
Beyond wavelength-based optics
The implications reach beyond fundamental physics. A passive, material-based filter that distinguishes quantum states of light could be integrated into on-chip photonic circuits for quantum information processing, replacing the sophisticated interferometry and conditional measurements currently required. The same approach could find applications in quantum metrology, where distinguishing non-classical states from classical backgrounds is essential.
The work also opens a broader question: if materials can be designed to respond to the quantum statistics of light rather than its classical properties, what other collective optical behaviors could be harnessed? Golat’s News & Views framed it as a new avenue for quantum optics, devices that control the bunching or anti-bunching behavior of photons as a routine material property, not a laboratory trick.
Caveats
The current demonstration operates at a single wavelength (780 nm) and requires cryogenic detection systems for the photon-number-resolving detectors used to characterize the output. Practical applications would require room-temperature operation and integration with fiber-optic or waveguide infrastructure. The metacrystal itself is a proof-of-concept, scaling the nanoantenna array to larger areas and different wavelength regimes remains to be shown.
Sources:
1. You, C. et al. “Quantum statistical plasmonic metacrystals.” Nature (2026). DOI: 10.1038/s41586-026-10782-3
2. Golat, S. “Optical filter sorts light by its ‘quantum statistics’.” Nature News & Views (15 July 2026). DOI: 10.1038/d41586-026-02038-x
3. Mouloudakis, G. & Lambropoulos, P. Phys. Rev. A 97, 053413 (2018). DOI: 10.1103/PhysRevA.97.053413