Quantum imaging applies non-classical properties of light to optical measurement and image formation. Its practical value is clearest when photon budgets are limited: for example, when imaging a delicate sample, measuring a weak absorption, or working at the single-photon level. By using sub-Poissonian photon statistics or correlations between beams, a measurement can reduce uncertainty compared with conventional methods using the same mean number of probe photons.
Why conventional optical measurements encounter noise limits
Optical imaging is commonly assessed through two distinct measures: resolution, or the ability to distinguish nearby features, and sensitivity, or the ability to detect a small change in a measured quantity. Diffraction constrains far-field spatial resolution, while photon statistics impose a limit on measurement sensitivity.
For a conventional probe with Poissonian photon-number statistics, the shot-noise limit (SNL) describes how uncertainty decreases as more photons are used. In simplified terms, the uncertainty scales as the inverse square root of the mean photon number. Increasing optical power can reduce this contribution, but it is not always practical. High power may introduce unwanted heating or scattering, and biological or other photosensitive samples may need to be measured with a low photon dose.
Reducing photon-number fluctuations below the Poissonian level offers another route. Light with variance smaller than its mean photon number is called sub-Poissonian. It is non-classical in the quantum-optics sense and can support more precise measurements without relying solely on a brighter probe.
How quantum imaging uses sub-Poissonian light
One example is estimating absorption. A probe passes through a sample, and the transmitted light is measured to infer how much was absorbed. With a Poissonian probe, photon-count fluctuations limit the precision. A probe with lower fluctuations can reduce this uncertainty. An ideal fixed-photon-number state, often called a Fock state, has no input photon-number variance; in practice, losses introduce additional fluctuations and weaken the advantage.
Producing a bright, single beam with useful sub-shot-noise statistics can be challenging. A widely used alternative is a twin-beam approach. The two beams have correlated photon numbers: one illuminates the sample as the probe, while the other serves as a reference for fluctuations in the source. Measuring the reference helps identify and subtract noise that would otherwise obscure a change in the probe.
Twin beams can be generated through spontaneous parametric down-conversion (SPDC). In a nonlinear optical medium, a pump photon can produce signal and idler photons that obey energy and momentum conservation. Although pair generation is probabilistic, the two output modes have strong photon-number correlations. Four-wave mixing in atomic vapours is another method described for generating correlated beams.
From correlated beams to wide-field imaging
The twin-beam strategy can extend beyond a single measurement point. In a multi-mode beam, many pairs of spatial modes are correlated. A spatially resolving detector, such as a camera, can record these modes in parallel. With the optical arrangement aligned so that corresponding correlated modes fall on matched pixels, the reference image can be used to compensate for fluctuations in the probe image on a pixel-by-pixel basis.
This approach has been demonstrated for wide-field sub-shot-noise imaging of weakly absorbing objects. It is conceptually related to conventional reference-beam techniques used in instruments such as spectrophotometers, where a second beam helps account for source or detector instability. The distinction is that classical correlated beams do not remove the uncorrelated shot-noise component, whereas quantum-correlated beams can reduce that contribution under suitable conditions.
Spatial resolution is an important design consideration. It depends on the size of the correlated regions, or coherence areas, at the object plane. Detector pixels must capture the correlated light adequately; if a pixel is too small, some photons in a correlated pair may not be measured together, reducing the noise advantage. The imaging system must therefore balance spatial detail against the need to collect correlated modes effectively.
What limits performance in real systems?
Quantum correlations are not immune to practical losses. Optical transmission, imperfect collection, and detector inefficiency all reduce the measured correlation. In a twin-beam system, losses add noise to the detected signal and reference, limiting how far the result can fall below the SNL. As a result, source quality alone does not determine performance: the full optical path and detection arrangement matter.
Single-photon sources offer another route to sub-Poissonian light. Their emission can exhibit antibunching, meaning photons are less likely to arrive together than they would from a classical source. However, real sources can have non-radiative emission losses, limited collection efficiency, and imperfect photon indistinguishability. Heralded photons generated in SPDC or four-wave mixing can provide a practical approximation, though some approaches require temporal selection to distinguish the heralded photon from background events.
Quantum imaging also does not automatically overcome every imaging constraint. Sub-shot-noise methods address measurement sensitivity; they do not, by themselves, remove the diffraction limit. Separate techniques and experimental arrangements are required to improve spatial resolution beyond conventional limits.
Implications for metrology and microscopy
For engineers and technical decision-makers, the key opportunity is measurement improvement where simply increasing illumination is not acceptable or cannot resolve the dominant noise problem. Potentially relevant areas include low-light microscopy, imaging of photosensitive samples, detector characterisation, and optical measurements at very low photon flux. Quantum correlations also have metrological applications in radiometry, including methods for calibrating photon detectors.
Successful implementation depends on matching the quantum-light source, optical layout, sample conditions, and detector to the measurement objective. Teams should distinguish sensitivity gains from resolution gains, account for losses across the system, and verify that pixel size and alignment preserve the relevant spatial correlations.
Conclusion
Quantum imaging with sub-Poissonian light uses controlled photon statistics and non-classical correlations to reduce measurement uncertainty when photon numbers are low. Twin-beam imaging is a practical strategy because a correlated reference can help suppress probe fluctuations across an image. Its benefit depends on managing loss, detection efficiency, and spatial-mode matching. For applications constrained by sample exposure or low light, these methods provide a route to more sensitive optical measurement, while remaining complementary to techniques that address diffraction and other sources of image degradation.