New quantum light test cuts photon measurements by 20%

Physicists at the University of Twente have improved one of quantum physics' oldest and most widely used experiments — by adding a single extra particle of light. The new method, published in Physical Review Letters, could help speed up the construction of future quantum computers.

The experiment in question is called the Hong-Ou-Mandel test, and it has been the standard way to check the quality of individual light particles, or photons, since 1987. It works by firing two photons at a device called a beam splitter — essentially a half-silvered mirror. If the two photons are truly identical, quantum physics forces them to exit the device together through the same side. If they're not perfectly identical, they sometimes come out on opposite sides. By repeating the experiment thousands of times, physicists can measure exactly how well-matched their photons are.

That precision matters because photonic quantum computers work by manipulating individual photons rather than electrical signals. For that technology to function, every photon involved has to be nearly indistinguishable from the others — and in real labs, they never start out that way.

Improving the photon measurement benchmark

The trouble is that checking photons two at a time doesn't scale well. A machine using ten photons already has 45 different pairs to test. The Twente team, led by PhD candidate Stefan van den Hoven, found a way around that bottleneck: add a second beam splitter so a third photon can join in.

"Hong-Ou-Mandel is one of the first experiments you learn about in quantum optics," Van den Hoven said. "It's usually also the first checkpoint quantum researchers use in their experiments."

By letting three photons interfere at once instead of two, each measurement yields more information about how well the photons match. The team calculated that their approach needs about 20% fewer repeated measurements to reach the same precision as the old two-photon method. Using a statistical tool called the Fisher information matrix, they also proved that their three-photon setup is the best possible version of the test — one that actually beats even a flawless, noise-free version of the traditional experiment.

Gamble without losing

Normally, adding more photons to a quantum experiment also increases the risk that some get lost along the way, which would usually ruin the measurement. Here, that isn't a problem: if a photon does go missing mid-experiment, the setup automatically falls back to behaving exactly like the traditional two-photon test.

"You can gamble without losing," Van den Hoven said.

For now, the technique is mainly useful to quantum optics laboratories fine-tuning their photon sources. But the researchers say it could eventually serve as a calibration standard, or even be built directly into large-scale photonic quantum computers as a built-in quality check — something that becomes more important as these machines are scaled up to handle hundreds of photons at once. Whether the same trick works just as well with four or five photons remains an open question for future research.