Below-shot-noise capacity in phase estimation using nonlinear interferometers

In the world of high-precision measurement, light is often treated as a stream of tiny, independent particles. When scientists use these particles to measure something as delicate as the phase of a wave—a property that describes the wave's position in its cycle—they are usually limited by a fundamental barrier known as the shot-noise limit. This limit arises because the arrival of photons is inherently random, like raindrops hitting a roof; even with a steady stream, the count fluctuates slightly from moment to moment. These fluctuations create a background fuzziness that obscures the true signal. For decades, researchers have sought ways to push past this barrier, hoping to achieve a level of precision that scales not with the square root of the number of particles used, but with the number itself. This would allow for measurements of unprecedented sharpness, useful for everything from detecting gravitational waves to imaging biological tissues without damaging them.

To achieve this, scientists have turned to a special class of devices called nonlinear interferometers. Unlike standard interferometers that split and recombine light beams using simple mirrors and glass plates, these advanced machines use crystals that can generate new light from existing light. Specifically, they use a process where a single high-energy photon splits into two lower-energy twins, known as a signal and an idler. These twins are born with a deep, quantum mechanical connection, or correlation, that standard light does not possess. By manipulating these linked pairs, researchers can theoretically create a state of light that is far more orderly than random rain, potentially allowing them to sense changes in the environment with a sensitivity that defies the usual limits of nature.

A team of researchers has now taken a close, practical look at three different ways to build these quantum-enhanced sensors, comparing how well they perform when real-world imperfections are introduced. The study focuses on configurations that rely on simple, easy-to-build intensity measurements—counting how many photons arrive at a detector—rather than complex, delicate setups that require sophisticated equipment. The researchers examined a design known as the Yurke interferometer, which uses two crystals to amplify the light in a symmetric loop; a Mandel interferometer, which uses a different arrangement where only one path is amplified; and a hybrid version that smoothly transitions between the two. Their goal was to determine which design remains the most reliable when light is lost along the way, a common problem in any real experiment.

The investigation began by considering an ideal world where no light is lost. In this perfect scenario, the Yurke design showed a remarkable potential. When the light passes through the system and is detected at a specific point of cancellation, the fluctuations in the signal can be suppressed so effectively that the precision improves dramatically as more photons are used. This behavior, known as Heisenberg scaling, represents the ultimate limit of quantum measurement. However, the researchers found that this advantage is incredibly fragile. The moment even a small amount of light is lost in the system, the delicate quantum order begins to unravel. In the Yurke setup, if the loss is not perfectly balanced between the two paths of the interferometer, the sensitivity does not just drop; it actually gets worse as more photons are added, eventually performing worse than a standard, non-quantum measurement.

In contrast, the Mandel configuration, which looks different on the inside, proved to be far more robust against these imperfections. In this setup, the researchers found that by measuring the difference in brightness between two output ports in an unseeded, gain-balanced configuration, they could achieve the highest phase sensitivity. This approach reaches, but does not surpass, the shot-noise limit with respect to the number of photons probing the sample, even in the presence of loss. While this method does not reach the ultimate, dream-like precision of the Heisenberg limit in the presence of loss, it consistently provides a stable level of sensitivity that scales predictably with the number of photons. The study demonstrated that in the high-gain regime, where large numbers of photons are generated to probe a sample, the Mandel design with differential detection outperforms the Yurke design whenever there is any imbalance in the losses, which is almost always the case in real experiments.

The researchers also explored a middle ground, a hybrid scheme that allows one to tune the device from the Yurke style to the Mandel style. They observed that as the device is adjusted away from the Yurke configuration, the sensitivity transitions smoothly from the fragile, high-precision regime to the stable, shot-noise-limited regime. This transition highlights a critical trade-off: the Yurke design offers the promise of superior precision, but only under conditions that are nearly impossible to maintain perfectly in a laboratory. The Mandel design, while offering a lower ceiling for precision, provides a floor that is much harder to fall through.

The findings suggest that for practical applications in quantum sensing and imaging, where samples often absorb light or where optical components are not perfect, the Mandel configuration with differential detection is the superior choice. It offers a robust solution that can be implemented with standard intensity detectors, avoiding the need for complex coincidence counting or homodyne detection. While the Yurke interferometer remains a fascinating example of quantum potential, its sensitivity to loss makes it less suitable for real-world scenarios where light cannot be perfectly preserved. The study concludes that by choosing the right architecture and detection method, scientists can build quantum sensors that are not just theoretically interesting, but genuinely useful for measuring the world with greater clarity than ever before.