We investigate the recovery of superconducting NbN-nanowire photon counters after detection of an optical pulse at a wavelength of 1550 nm, and present a model that quantitatively accounts for our observations. The reset time is found to be limited by the large kinetic inductance of these nanowires, which forces a tradeoff between counting rate and either detection efficiency or active area. Devices of usable size and high detection efficiency are found to have reset times orders of magnitude longer than their intrinsic photoresponse time. . Of particular interest would be a detector that combines ultrafast count rates (≥ GHz) with high single-photon detection efficiency at near-infrared wavelengths; however, current near-infrared photon-counting technologies such as avalanche photodiodes [6] and photomultiplier tubes [7] are limited to much lower count rates by long reset times.A promising detector technology was reported recently, in which ultrathin superconducting NbN wires are biased with a DC current I bias slightly below the critical value I C [8]. An incident photon of sufficient energy can produce a resistive "hotspot" which in turn disrupts the superconductivity across the wire, resulting in a voltage pulse. Observations of this photoresponse showed promise for high counting rates, with measured intrinsic response times as low as ∼30 ps [9], and counting rates in the GHz regime [10,11]. In this Letter, we present our own investigation into the counting-rate limitation of these devices, in which we directly observe the recovery of the detection efficiency as the device resets (after a detection event), and develop a quantitative model of this process. We find that detectors having both high detection efficiency and usable active area are limited to much lower count rates than studies of their intrinsic response time had suggested [9].We fabricated our nanowires using a newly developed process [12], on ultrathin (3 − 5 nm) NbN films [13]. We used several geometries, including straight nanowires having widths from 20−400 nm and lengths from 0.5−50 µm, as well as large-area "meander" structures [8,10] (e.g., Fig. 1(b)) having active-area aspect ratios from 1 − 50, fill factors from 25 − 50%, and sizes up to 10-µm square. The devices had critical temperatures T C ∼ 9 − 10 K, and critical current densities J C ∼ 2 − 5 × 10 10

We have fabricated and tested superconducting single-photon detectors and demonstrated detection efficiencies of 57% at 1550-nm wavelength and 67% at 1064 nm. In addition to the peak detection efficiency, a median detection efficiency of 47.7% was measured over 132 devices at 1550 nm. These measurements were made at 1.8K, with each device biased to 97.5% of its critical current. The high detection efficiencies resulted from the addition of an optical cavity and anti-reflection coating to a nanowire photodetector, creating an integrated nanoelectrophotonic device with enhanced performance relative to the original device. Here, the testing apparatus and the fabrication process are presented. The detection efficiency of devices before and after the addition of optical elements is also reported.

No abstract

We investigate the role of electrothermal feedback in the operation of superconducting nanowire singlephoton detectors ͑SNSPDs͒. It is found that the desired mode of operation for SNSPDs is only achieved if this feedback is unstable, which happens naturally through the slow electrical response associated with their relatively large kinetic inductance. If this response is sped up in an effort to increase the device count rate, the electrothermal feedback becomes stable and results in an effect known as latching, where the device is locked in a resistive state and can no longer detect photons. We present a set of experiments which elucidate this effect and a simple model which quantitatively explains the results. DOI: 10.1103/PhysRevB.79.100509 PACS number͑s͒: 85.25.Oj, 74.78.Ϫw, 85.60.Gz Superconducting nanowire single-photon detectors ͑SNSPDs͒ combine high speed, high detection efficiency ͑DE͒ over a wide range of wavelengths, and low dark counts. [1][2][3][4] Of particular importance is their high singlephoton timing resolution of ϳ30 ps, 4 which permits extremely high data rates in photon-counting communications applications. 5,6 Full use of this electrical bandwidth is limited, however, by the fact that the maximum count rates of these devices are much smaller ͑a few hundred MHz for 10 m 2 active area and decreasing as the area is increased 2 ͒, limited by their large kinetic inductance and the input impedance of the readout circuit. 2,7 To increase the count rate, therefore, one must either reduce the kinetic inductance ͑by using a smaller active area or different materials or substrates͒ or increase the load impedance. 7 However, either of these approaches causes the wire to "latch" into a stable resistive state where it no longer detects photons. 8 This effect arises when negative electrothermal feedback, which in normal operation allows the device to reset itself, is made fast enough that it becomes stable. We present experiments which probe the stability of this feedback, and we develop a model which quantitatively explains our observations. The operation of an SNSPD is illustrated in Fig. 1͑a͒. A nanowire ͑typically ϳ100 nm wide and 5nm thick͒ is biased with a dc current I 0 near its critical current I c . The nanowire has kinetic inductance L and is read out using a load impedance R L ͑typically a 50⍀ transmission line͒. When a photon is absorbed, a short ͑Ͻ100 nm long͒ normal domain is nucleated, giving the wire a resistance R n ͑t͒. This results in Joule heating which causes the normal domain ͑and consequently, R n ͒ to expand in time exponentially. The expansion is counteracted by negative electrothermal feedback from the load R L , which forms a current divider with R n , and diverts a current I L into the load ͑so that the current in the nanowire is reduced to I d ϵ I 0 − I L ͒, reducing the heating. However, in a correctly functioning device, this feedback is unstable: the inductive time constant is long enough so that before I L becomes appreciable, Joule heating has already increased R n , ...