Abstract: We present our studies on ultrafast superconducting single-photon detectors (SSPDs) based on ultrathin NbN nanostructures. Our SSPDs are patterned by electron beam lithography from 4-nm thick NbN film into meander-shaped strips covering square area of 10/spl times/10 /spl mu/m/sup 2/. The advances in the fabrication technology allowed us to produce highly uniform 100-120-nm-wide strips with meander filling factor close to 0.6. The detectors exploit a combined detection mechanism, where upon a single-photon absorption, an avalanche of excited hot electrons and the biasing supercurrent, jointly produce a picosecond voltage transient response across the superconducting nanostrip. The SSPDs are typically operated at 4.2 K, but they have shown that their sensitivity in the infrared radiation range can be significantly improved by lowering the operating temperature from 4.2 K to 2 K. When operated at 2 K, the SSPD quantum efficiency (QE) for visible light photons reaches 30-40%, which is the saturation value limited by optical absorption of our 4-nm-thick NbN film. For 1.55 /spl mu/m photons, QE was /spl sim/20% and decreases exponentially with the increase of the optical wavelength, but even at the wavelength of 6 /spl mu/m the detector remains sensitive to single photons and exhibits QE of about 10/sup -2/%. The dark (false) count rate at 2 K is as low as 2 /spl times/ 10/sup -4/ s/sup -1/, what makes our detector essentially a background-limited sensor. The very low dark-count rate results in the noise equivalent power (NEP) as low as 10/sup -18/ WHz/sup -1/2/ for the mid-infrared range (6 /spl mu/m). Further improvement of the SSPD performance in the mid-infrared range can be obtained by substituting NbN for the other, lower-T/sub c/ superconductors with the narrow superconducting gap and low quasiparticle diffusivity. The use of such materials will shift the cutoff wavelength towards the values even longer than 6 /spl mu/m.

Abstract: Absorption of an infrared photon in an ultrathin film (such as 10-nm NbN) creates a localized nonequilibrium hotspot on the submicron length scale and sub-ns time scale. If a strip /spl sim/1 /spl mu/m wide is biased in the middle of the superconducting transition, this hotspot will lead to a resistance pulse with amplitude proportional to the energy of the incident photon. This resistance pulse, in turn, can be converted to a current pulse and inductively coupled to a SQUID amplifier with a digitized output, operating at 4 K or above. A preliminary design analysis indicates that this data can be processed on-chip, using ultrafast RSFQ digital circuits, to obtain a sensitive infrared detector for wavelengths up to 10 /spl mu/m and beyond, with bandwidth of 1 GHz, that counts individual photons and measures their energy with 25 meV resolution. This proposed device combines the speed of a hot-electron bolometer with the single-photon-counting ability of a transition-edge microcalorimeter, to obtain an infrared detector with sensitivity, speed, and spectral selectivity that are unmatched by any alternative technology.

Abstract: We developed the model of the internal phonon bottleneck to describe the energy exchange between the acoustically soft ultrathin metal film and acoustically rigid substrate. Discriminating phonons in the film into two groups, escaping and nonescaping, we show that electrons and nonescaping phonons may form a unified subsystem, which is cooled down only due to interactions with escaping phonons, either due to direct phonon conversion or indirect sequential interaction with an electronic system. Using an amplitude-modulated absorption of the sub-THz radiation technique, we studied electron-phonon relaxation in ultrathin disordered films of tungsten silicide. We found an experimental proof of the internal phonon bottleneck. The experiment and simulation based on the proposed model agree well, resulting in τe−ph∼140–190 ps at TC=3.4K, supporting the results of earlier measurements by independent techniques.