In a laboratory at Paderborn University, a single laser pulse is split again and again until it becomes a thousand-odd fragments, each read by a superconducting detector that remembers how the light arrived. From those scattered clicks, the researchers can now reconstruct how many photons the original flash carried, reaching past nine thousand and, in some cases, pinning the count to within less than a single photon. It is a quiet step in humanity's long effort to see light not merely as present or absent, but as something countable, and it may one day help shape how quantum machines and sensi
Multiplexed Superconducting Detectors Count Up to 276 Photons with Sub-Photon Precision
In a laboratory at Paderborn University, a single laser pulse is split again and again until it becomes a thousand-odd fragments, each read by a superconducting detector that remembers how the light arrived. From those scattered clicks, the researchers can now reconstruct how many photons the original flash carried, reaching past nine thousand and, in some cases, pinning the count to within less than a single photon. It is a quiet step in humanity's long effort to see light not merely as present or absent, but as something countable, and it may one day help shape how quantum machines and sensitive cameras are built.
A single pulse of laser light enters a device at Paderborn University and is divided, step by step, until it has become 1,024 tiny pulses. Each carries only a few of the original photons, and each is read by a superconducting detector. From those readings the researchers can say how many photons the first pulse held, in some cases to within less than one photon, and their work, published in Optica, reaches from zero to more than 9,000.
Timon Schapeler explains that counting grows harder as numbers rise, and this approach stretches photon counting into the wide gap between single-photon measurement and high-sensitivity optical power meters. That gap exists because the usual tools each fall short in their own way. Avalanche photodiodes report only whether photons are present, while detectors built for large numbers carry enough electrical noise to blur the count. Superconducting nanowire detectors sit between these, prized for efficiency, timing and low noise, and the team asked how far their ability to resolve several photons could be pushed.
The answer was to combine that native resolution with multiplexing, spreading the light across many detector elements. A pulse is split into 128 time-separated pieces, then distributed across eight detectors, using low-loss optics that keep the photon information intact in each bin. A detector model built earlier with Mariia Sidorova of PTB describes every element, and calibrating each bin with it is what turned raw clicks into precise counts.
In tests with laser pulses averaging between one and 15,000 photons, the system counted pulses up to 276 photons with uncertainties below one photon. Tim Bartley notes that large-scale single-photon measurement is central to photonic quantum computing and imaging, and that the work offers lessons for scaling those machines and cameras. So far only laser light has been measured. The team next hopes to apply the method to bright quantum light, and expects that as the detectors improve, it will support precise sensing across many applications.
A single pulse of laser light goes into a device at Paderborn University in Germany and is divided, again and again, until it has become 1,024 tiny pulses. Each one carries only a few of the original photons, and each is read by a superconducting detector. From those readings, the researchers can say how many photons the original pulse contained, in some cases to within less than one photon.
The work, published in Optica, describes a system that can count from zero to more than 9,000 individual photons. Timon Schapeler of Paderborn said that counting gets harder as the numbers rise, and that the approach stretches photon counting into a range between single-photon measurements and high-sensitivity optical power meters.
That gap exists because existing tools each fail in a different way. Avalanche photodiodes, common single-photon detectors, typically report only whether photons are present or absent, not how many. Photodiodes and other detectors built for large photon numbers carry enough electrical noise to blur the count. Superconducting nanowire single-photon detectors, or SNSPDs, sit in between. They are known for high efficiency, very fine timing resolution and low noise, and it was recently shown that they can resolve a limited number of arriving photons using high-precision timing.
Schapeler said the team wanted to see how far that ability could be pushed. The answer was to pair the detectors' own photon-number resolution with multiplexing, which imitates number resolution by using many separate detector elements. Each technology had been studied alone, but the researchers say it had not been demonstrated before with both together at this scale, which they describe as an order of magnitude beyond earlier multiplexed single-photon counting systems.
In practice, one light pulse is split into 128 smaller pulses separated in time. Those are then spread across eight detectors, giving 1,024 pulses that share the original photons between them. Tim Bartley, also of Paderborn, said the system uses low-loss optical components to distribute light evenly across the 1,024 bins while preserving the photon-number information in each. Multiplexing, he said, widens the range of photon numbers the system can measure.
The other essential ingredient was a model of the detectors themselves. The team developed it earlier with Mariia Sidorova of PTB in Germany, and it describes each individual detector element in the system. Using it to characterize every bin is what let the group turn raw clicks into precise counts.
To test the system, they sent in laser pulses averaging between one and 15,000 photons and watched how the detector responded across that range. With the model's accuracy, they demonstrated counting of pulses up to 276 photons with uncertainties below one photon. The researchers say a conventional multiplexed system would need about 37,000 detector elements to match that precision, against their 1,024 bins.
Bartley said large-scale measurement of individual photons is a cornerstone of photonic quantum computing and single-photon-level imaging, and that the technology can teach researchers much about how photonic quantum computers and single-photon-sensitive cameras might be scaled.
So far the detectors have only measured laser pulses. The team plans to turn them on bright quantum states of light, which could reveal quantum effects at large scales. As SNSPD technology improves, the researchers expect the system could support precise, sensitive measurements across a range of sensing applications.
Envie de l'histoire complète ?
Lire l'original sur Phys.org ↗