In quantum optics, certain boundaries have long been treated as fundamental dividers separating classical and quantum regimes. If you shine classical light into a Hanbury Brown and Twiss (HBT) setup or a Hong-Ou-Mandel (HOM) interferometer, you hit hard walls: second-order interference visibility cannot exceed 50%, and the second-order degree of coherence, $g^{(2)}(0)$, for thermal light is strictly capped at 2. To surpass these limits, conventional wisdom dictated a shift to non-classical states, such as entangled photon pairs or single-photon Fock states.
However, a cohesive body of research demonstrates that these boundaries are not absolute properties of classical light itself, but rather constraints imposed by standard measurement geometries and statistical assumptions. By re-engineering the statistical fluctuations and path configurations of classical light, researchers are breaking through these ceilings, achieving quantum-level imaging and sensing performance using entirely classical architectures.
The 50% Visibility Barrier and Indistinguishability
Why does classical light face a 50% visibility limit in second-order interference? In a typical two-photon interference experiment, the correlation signal depends on whether the paths taken by the photons to reach the detectors are indistinguishable. For classical light, random phase and amplitude fluctuations inherently wash out the contrast of these overlapping paths, capping the interference visibility at 50%. This limitation directly impacts applications like Hong-Ou-Mandel sensing, where using attenuated classical light as a reference restricts the maximum achievable visibility to 50%, limiting the sensitivity of the measurement Fluorescence lifetime Hong-Ou-Mandel sensing.
However, Jianbin Liu et al. reframed this constraint by showing that indistinguishability is a property of the measurement apparatus rather than an intrinsic property of the photons themselves. In Two-photon interference with non-identical photons, the authors demonstrated that even non-identical photons with different wavelengths can exhibit second-order interference, provided the detection system is physically incapable of resolving their differences.
The authors extended this principle in The second-order interference of two independent single-mode He–Ne lasers, proving that two completely independent classical lasers can produce second-order spatial and temporal interference patterns. While these setups still respect the 50% classical visibility threshold, they prove that quantum-like path interference can be engineered using independent classical sources. Furthermore, in The second-order interference between laser and thermal light, the authors showed that mixing coherent laser light and pseudothermal light can produce two-photon anticorrelation—a phenomenon often considered a hallmark of quantum light—within specific dual-source configurations.
Standard Thermal Light: [Laser] ---> [Single Ground Glass] --------> g²(0) = 2 (Gaussian Limit)
Cascaded Superbunching: [Laser] ---> [Ground Glass 1] ---> [Ground Glass 2] ---> g²(0) > 2 (Super-Gaussian)
The Statistical Ceiling of Thermal Light
To break past these visibility limits entirely, one must manipulate the statistical properties of the light source. The degree of second-order coherence, $g^{(2)}(\tau)$, measures the probability of detecting two photons simultaneously compared to a completely random (Poissonian) distribution. For standard thermal or chaotic light, the intensity fluctuations follow Gaussian statistics, which mathematically dictates that $g^{(2)}(0) = 2$. This means photons are twice as likely to arrive in pairs (bunching) than they would be in a purely coherent laser state ($g^{(2)}(0) = 1$).
This statistical ceiling directly limits correlation-based imaging. In classical ghost imaging, an image is reconstructed by correlating the intensity fluctuations between two separate optical paths. When using standard thermal light, the contrast of the reconstructed image is strictly bounded by the source's intensity fluctuations, restricting the maximum image visibility to 33% High visibility temporal ghost imaging with classical light. To bridge the performance gap with quantum ghost imaging—which achieves near-perfect contrast using entangled photons—it is necessary to make classical light "clumpier," entering the regime of superbunching where $g^{(2)}(0) > 2$.
Mechanisms for Engineering Superbunching
To exceed a $g^{(2)}(0)$ of 2 without relying on quantum states or non-linear optical materials, researchers have introduced several linear optical techniques to manipulate classical intensity statistics.
Cascaded Scattering
One elegant approach is to pass a laser beam through multiple rotating ground glasses in series. A single rotating ground glass randomizes the phase of a laser, creating standard pseudothermal light governed by Gaussian statistics ($g^{(2)}(0) = 2$). However, cascading these glasses compounds the intensity fluctuations, generating super-Gaussian statistics. Fuli Li et al. demonstrated this in Superbunching pseudothermal light, showing that the statistical limits of natural thermal light could be bypassed using simple laboratory components. Using a similar cascaded setup, Bin Bai et al. achieved a second-order coherence value of 7.10 in Photon superbunching of classical light in the Hanbury Brown–Twiss interferometer, proving that high-order optical correlations do not require exotic materials.
Intensity Modulation
Rather than relying solely on multiple physical scattering stages, modulating the intensity of the initial laser beam before it strikes a rotating ground glass provides a more streamlined route to superbunching. This mechanism was explored by Xuexing Zhang et al. in Superbunching pseudothermal light with intensity modulated laser light and rotating groundglass and further optimized by Jianbin Liu et al. in Simple and efficient way to generate superbunching pseudothermal light. By pre-conditioning the laser's temporal profile, the authors unlocked significantly higher coherence levels with less optical loss than cascaded ground glasses.
Path Multiplicity
Superbunching can also be induced by altering the measurement architecture rather than the source itself. Sheng Luo et al. demonstrated that the traditional $g^{(2)}=2$ limit for chaotic light is a constraint of standard geometries rather than an inherent property of the light. In Two-photon superbunching effect of broadband chaotic light at the femtosecond timescale based on a cascaded Michelson interferometer, the authors used a cascaded Michelson interferometer to increase the multiplicity of indistinguishable optical paths. This geometric complexity successfully pushed $g^{(2)}(0)$ to 2.42 for true broadband chaotic light at a 95-femtosecond scale.
Unlocking High-Visibility Imaging
The practical payoff of breaking these coherence limits is the dramatic enhancement of correlation-based technologies. By generating classical light with extreme bunching properties, the fundamental visibility constraints of classical imaging dissolve.
For instance, Jianbin Liu et al. utilized superbunched light in High visibility temporal ghost imaging with classical light to shatter the 33% visibility ceiling, achieving high-contrast images previously thought to be the exclusive domain of quantum entanglement.
This concept scales to higher-order correlations as well. Sheng Luo et al. extended this to three-photon correlations in Experimental observation of three-photon superbunching with classical light in a linear system, achieving a third-order coherence degree ($g^{(3)}$) of 22.55—nearly four times higher than standard thermal light. This massive increase in correlation contrast directly translates to higher-visibility ghost imaging, paving the way for high-fidelity, quantum-inspired sensing using accessible, classical architectures.
Key Foundational Work
The second-order interference of two independent single-mode He–Ne lasers (2015) — contributed 4 concept elements to this direction.
The second-order interference between laser and thermal light (2014) — contributed 3 concept elements to this direction.
Two-photon interference with non-identical photons (2015) — contributed 2 concept elements to this direction.