Raising the Rate of Single-Photon Detection
JQI researchers have devised and demonstrated a novel solution to a growing problem in quantum optics: the limited detection rate of single-photon detectors. Those devices, which are increasingly in demand for applications such as quantum key distribution and metrology, require a brief recovery interval – called “deadtime” – after each detection. During that period, the device cannot record the arrival of a photon. Although the typical deadtime interval is on the order of microseconds, the interruption necessarily restricts data transfer speeds.
“Detection rates have become a major bottleneck in developing high-speed quantum communication and other quantum-enabled technologies,” says Sergey Polyakov of JQI and the National Institute of Standards and Technology (NIST). So he and JQI Fellow Alan Migdall, in collaboration with a group headed by Ivo Degiovanni from INRIM, Italy, set out to solve the problem.
Their solution, published earlier this year in the Journal of Modern Optics, allows photon counting at significantly higher rates than is otherwise possible with existing arrangements by “multiplexing” a pool of detectors so that it operates as a single unit. An electronic circuit keeps track of which detectors have fired most recently, and activates switches which route each arriving photon to a detector that is ready. The active multiplexing system has a much higher throughput than passive arrays of multiple detectors, and the degree of improvement is non-linear. That is, it scales faster than the number of detectors. For example, when the researchers went from one to two detectors in the comparison test, the detection rate rose by a factor of 2.2 in the multiplexed system.
The JQI scheme should be able to accommodate greatly increased speeds. “At high count rates,” the researchers write (see reference publication), “many of the detectors may fire within a short period of time and be dead. But as long as the first detector recovers to its live state before the last detector fires, the whole arrangement will still be live and ready to register an incoming photon.”
In addition, the scientists determined that their multiplexing scheme did not adversely affect either of the two other major sources of error in single-photon detectors: dark counts and afterpulses.
Dark counts are false signals caused by thermally generated charges created within the detector. In conventional multi-detector systems, the number of them increases linearly with the number of detectors. But in the JQI multiplex system, the dark counts stay the same irrespective of number of detectors. Afterpulses, in which the electric pulse produced by an arriving photon is followed by a spurious pulse, actually decreased with the JQI scheme. The reason is that the high-speed detectors used in the experiment (called single-photon avalanche detectors, or SPADs) are less prone to afterpulses the longer they rest between firings, and the multiplex scheme maximizes their rest time.
In addition, the researchers found that there are other, previously unappreciated, factors that limit the detection rate increase and refined a theoretical model of the actively multiplexed scheme to include these effects. Particularly, they found that detectors that must operate in a gated mode such as indium gallium arsenide avalanche photodiodes, exhibit a deadtime associated with their gate circuitry, independent of whether the detector fires or not, which should be treated separately from the deadtime due to an actual photodetection. A better understanding of this effect allows further improvements of the multiplexing algorithm. The team expects to publish experimental proof of these further improvements shortly.