Commentary: New developments in single photon detection in the short wavelength infrared regime

Abstract

Many new and exciting fields of scientific research rely on the efficient detection of single photons in the infrared regime. To give just two examples, quantum key distribution has the potential to offer verifiably secure methods of generating cryptographic keys and active imaging techniques permit the creation of detailed 3D models of distant objects. Currently, many of these technologies use photons in the wavelength range 400 to 1000 nm due to the ready availability of relatively efficient single photon detectors for use at these wavelengths (recent surveys may be found in Refs. [1] and [2]). However, extending the wavelength range to 1.3 µm and 1.55 µm would result in many benefits, for example compatibility with lowloss optical fiber spectral regions and atmospheric transmission windows, as well as infrared gas absorption features, notably those of greenhouse gases. Let us briefly consider the current status of the potential single photon detection technologies available for this spectral region. Currently, there are three main types of detector technology in widespread use; photomultipliers, semiconductor avalanche photodiodes and superconducting detectors. Within these different technologies, there are different material systems and micro-structure designs leading to a wide spread in key performance characteristics for similar devices. One figure of merit for such detectors is the noise equivalent power (NEP), defined as the signal power required to attain a unity signal-to-noise ratio within a one second integration time. Photomultiplier tubes (PMTs) operate by the cascade generation of a stream of electrons from an initial photo-generated electron. A photon incident on a photocathode will, with a probability less than unity, trigger the emission of an electron. This electron is accelerated towards an anode and, on collision, triggers the release of a number of secondary electrons. This process is repeated through a series of anodes, multiplying the total number of electrons each time, until a detectable current pulse is generated. These devices have large photocathode areas (typically of the order of a few mm 2 ) and gain which depends on dynode geometry and bias. PMTs typically operate with photons at lower wavelengths than those considered in this commentary but devices with an InGaAs photocathode such as the those produced by Hamamatsu have demonstrated an NEP ~ 4 × 10