Short answer: A transition-edge sensor is a superconducting detector that registers photons through a change in resistance, and it can tell how many photons arrived in a pulse. A 2005 Physical Review A paper reports fibre-coupled sensors with 88% efficiency at 1550 nm, a 2012 New Journal of Physics paper reports the first tomography of such a detector, and a 2022 Nature Photonics paper reports multiplexing them to resolve photon numbers between 0 and 100. These are laboratory results.
What the detector does
The 2012 paper describes transition-edge sensors as superconducting detectors in which impinging photons are detected through the change of resistance related to the transition from a superconducting to a non-superconducting state, caused by the deposited energy. The 2005 paper explains the operating principle for its tungsten sensors: they are quantum calorimeters, with the tungsten electrons acting as both absorber and thermometer, and a voltage bias keeps them on the edge of the superconducting-to-normal transition so that the steep dependence of resistance on temperature gives precise thermometry.
2005: efficiency without dark counts
The 2005 authors say high-efficiency detectors that can determine the number of photons in a pulse have applications including post-selection-based entanglement protocols for linear optics quantum computing, and experiments that close the detection and communication loopholes of Bell's inequalities. They report fibre-coupled, noise-free, photon-number-resolving transition-edge sensors with 88% efficiency at 1550 nm, and say efficiency could be made higher at any wavelength in the visible and near-infrared without a higher dark-count rate or degraded photon-number resolution.
2012: characterizing the detector
The 2012 authors present what they call the first experimental tomography of the positive operator-valued measure of a transition-edge sensor, which gives the probability of detecting n photons out of m impinging on it. They reconstruct the elements up to 11 detected photons and report demonstrating that the sensor is a linear detector with substantially no dark counts. They say precise characterization of photon-number-resolving detectors is a prerequisite for applications such as quantum metrology.
2022: resolving up to 100 photons
The 2022 Nature Photonics abstract says true photon-number-resolving detectors exist but are limited to resolving on the order of 10 photons, which is too small for several quantum-state generation methods based on heralded detection. The authors multiplex highly efficient transition-edge sensors to resolve photon numbers between 0 and 100, and use the system for a quantum random-number generator with no inherent bias. They say the method samples a coherent state in the photon-number basis and is robust against environmental noise, laser phase and amplitude fluctuations, loss and detector inefficiency, and eavesdropping.
Limits
These sensors work at superconducting temperatures, a point the three papers cover only through the operating principle. The random-number generator and the efficiency figures are laboratory demonstrations from the authors, and the abstracts' statements about applications are statements of intent.
FAQ
How does a transition-edge sensor detect photons?
Energy from absorbed photons changes the resistance of a superconducting film held at the edge of its transition, per the 2012 and 2005 papers.
How efficient were the 2005 sensors?
88% at 1550 nm, fibre-coupled, with the authors reporting no noise from dark counts.
What was new in 2012?
The first experimental tomography of a transition-edge sensor, showing a linear response with substantially no dark counts.
How many photons can be resolved?
The 2022 authors report resolving between 0 and 100 by multiplexing sensors, against about 10 for earlier true photon-number-resolving detectors.
What was the random-number generator?
The 2022 paper demonstrates one based on sampling a coherent state in the photon-number basis, which the authors say has no inherent bias.
All field notes