IR- and visible- light single photon detection in superconducting $MgB_{2}$ nanowires

State-of-the-art Superconducting Nanowire Single Photon Detectors based on low-Tc materials reach 100% quantum efficiency. However, the response time is limited to>1-10 ns. Recently, it has been shown that due to a much lower kinetic inductance, a 100 ps response rate can be achieved in 120 $\mu$m-long $MgB_{2}$ nanowires. In this work, we demonstrate experimentally that such $MgB_{2}$ nanowires function as single-photon detectors for both visible ($\lambda$= 630 nm) and infrared ($\lambda$= 1550 nm) photons when biased close to the critical current, with a dark count rate of<10 cps. $MgB_{2}$ photodetectors over-perform NbN SNSPDs in speed by at least an order of magnitude for similar nanowire lengths. Such photodetectors offer a platform for single-photon detectors with a long thought-after combination of a large detector area and a response rate of up to 10 GHz with a single readout line.

Ability to detect single photons opens many interesting applications from quantum physics to space communication. 1 , 2 , 3 Recently, Superconducting Nanowire Single Photon Detectors (SNSPD or SSPD) 4 , 5 have shown a combination of both high efficiency and count rates up to 100 MHz in single photon detection from the visible to the mid-IR ranges. 6 , 7 , 8 Absorption of a photon in a superconducting nanowire (w∼100 nm wide) triggers a normal-state domain growth, which can (under favorable conditions) stretch across the entire nanowire width. 9 This causes a voltage drop of a magnitude sufficiently large to be registered with standard microwave amplifiers and pulse counters. The aforementioned favorable conditions include the presence of a dc bias current, I0, flowing through the nanowire close to the superconducting critical current, Ic (sometimes called a switch current). Ratios of I0/ Ic close to 1 are required in order to achieve single photon sensitivity and increase the quantum efficiency. 4 , 6 After the normal domain formation, the bias current is partly rerouted from the nanowire into the readout load, RL=50 Ω. The nanowire detector resets only after the normal domain has collapsed and the bias current has returned back into the nanowire. Whereas the former process is governed by electron-phonon cooling, the later process is mainly limited by the large kinetic inductance in the nanowire, 10 which is normally long in order to cover a large detection area. For the most common SNSPD materials, such as NbN and NbTiN, a kinetic inductance of Lk0∼90 pH/ has been reported, which results in a reset time of τ = Lk0 × l / (w × RL) ∼3-5 ns for ∼l = 100-150 µm long devices. 6 , 8 On the other hand, we recently demonstrated that in nanowires made from 5nm-thick MgB2 films a much lower kinetic inductance, Lk0∼1.5 pH/ , can be achieved. 11 Combined with a 12 ps electron relaxation time 12 , this has resulted in a reset time of ∼100ps for MgB2 devices as long as 120 µm. In this paper, we studied photon detection in MgB2 nanowires in both the visible and the laser-communication (1550 nm) ranges. We demonstrated that such detectors generate responses from 1-, 2, or 3-photon absorptions depending on the bias current. In a comparative experiment, MgB2 nanowire photodetectors demonstrate a response rate of at least an order of magnitude shorter than that in NbN SNSPDs.
For this study, MgB2 thin films were grown on 6H-SiC substrates using Hybrid Physical Chemical Vapor Deposition (HPCVD). Both the film growth process and the nanowire fabrication routines have been described in earlier publications. 11 , 13 , 14 MgB2 films were ∼5 nm-thick, as has been shown by Transmission Electron Microscopy, 11 with a critical temperature Tc of ∼32-34 K ( Fig.1a). All photon detection experiments were conducted in a cryogen-free probe station (base temperature 5 K) with an optical view port and a microwave ground-signal-ground (GSG) probe on a XYZ-moving stage, which provided both the dc biasing and the broadband readout.
Schematics of the experimental set-up are shown in Fig.2a. A 630nm-line cw laser and a 1550nmline 80 fs pulsed laser (100MHz repetition rate) flood illuminated the samples through a set of neutral density filters (visible or IR), a quartz sealing window of the probe station and an IR filter (transparent for wavelengths <2µm) at the 60 K heat shield. At 5K, the critical currents were ∼97µA and ∼399µA for a 35nm-wide and 100nm-wide MgB2 nanowires (Fig.1b). For a direct comparison of the response rate of MgB2 and NbN nanowire detectors, we fabricated a set of 100nm-wide NbN nanowire detectors, 115 µm and 460 µm long. NbN films were grown on C-cut sapphire substrates using dc magnetron sputtering at a substrate temperature of 800 °C. Other details could be found in the supplementary material. The NbN film was ∼5nm thick, as estimated from the deposition rate. The critical temperature was ∼10 K (Fig.1a) and the room temperature sheet resistance was ∼300 Ω/ , which corresponded fairly well to previously published NbN SNSPDs. Forms of current-voltage ( I(V)) curves for NbN samples were quite similar to those for MgB2 samples, with a distinct switching behavior at Ic , without latching. However, the critical current density in MgB2 samples showed a factor of 10-12 times higher than that shown in NbN samples (Fig.1b). All samples were integrated into coplanar waveguide contacts suitable for the 100µm pitch GSG-probe of the cryo-station. At room temperature, both microwave readout and dc biasing were arranged (Fig.2a) though a bias-T (Picosecond 5547, 12kHz-15GHz) followed by a low noise amplifier (MITEQ 3D-0010025, 0.5-5 GHz) and either a real time oscilloscope (Keysight Infiniium 54854 DSO, 4GHz, 20 GSa/s) or a pulse counter (HP 53131A 225).
Nanowires were biased with a Yokogawa 7651 programmable dc source. a) have to be very short, which requires the nanowire to be integrated with a photonic waveguide in order to provide good optical coupling. 15,16 Alternatively, several NbN nanowires could be connected in parallel, which leads to a reduced total inductance and hence to a shorter voltage decay time. 17 However, reduction of the nanowire length (aiming for a faster reset) has a negative effect on the maximum utilizable bias current for which NbN SNSPDs still do not latch into the permanent normal state (Ilatch/Ic could be as low as ∼0.7). 15 Interestingly, despite a very low inductance, MgB2 nanowire detectors did not show latching for bias currents up to Ic . This fact can be explained by a factor of 4 times faster electron energy relaxation in MgB2 (12 ps) 12 than in NbN (50ps) 18 thin films. Fast electron relaxation leads to rapid normal domain collapse and makes a fast current return from the load back into the nanowire much less critical. 19 For an initial estimate of the MgB2 photon detector response jitter, we utilized statistics for response pulse intervals collected within a 10 µs time period. Although the oscilloscope trigger was not locked to the pulsed laser, the response interval jitter is still only ∼50 ps, which corresponds to a ∼25 ps jitter of the rising edge (Fig.3a). This value is an upper limit for MgB2 nanowire detectors, because a jitter less than that in NbN could be expected in MgB2 considering a factor of 33 lower kinetic inductance. 20 Direct demonstration of single photon detection capability requires a 100% photon coupling and a known quantum efficiency. Considering that the former is a challenging engineering task and the latter would require complex modelling, the single photon detection mode is often verified by statistical analysis of the photon counts per photon flux 4 or inter-counts intervals. 21 We chose the first method, utilizing initially a cw laser (630nm). The photon count rate N vs the laser attenuation F is shown in Fig. 3b at several bias currents for a 90 nm-wide and 40 µm-long nanowire. For bias currents I0 ≥396 µA, the photon count is linearly proportional to the photon flux, N ∝ F. For smaller bias currents we observed that N ∝ F m , where m>1, which indicates that multi-photon detection mode becomes dominant. These results show that, in contrast to NbN nanowire detectors, the single photon detection mode is limited in MgB2 nanowires to a narrow bias current range in the vicinity of Ic (Fig.4a). Physical modeling using accurate material parameters, is required in order to obtain a full understanding of photon detection in MgB2 thin film nanowires. In this moment, we note that the aforementioned fact is probably caused by a smaller size of the normal domain (compared to the nanowire width), formed by the photon-deposited energy. Therefore, a higher dc current (vs Ic) is required to extend the normal domain to the edges of the nanowire. The dark count rate in the studied MgB2 nanowire detectors was very small, <10 cps at the highest utilized bias current (398µA), despite the fact that the samples were not shielded from either the electrical or magnetic field interference. At this proximity to Ic , the dark count rates in NbN SNSPDs often reach values of >10 3 -10 5 cps. 6  In conclusion, in a comparative experiment we demonstrated that reset time in 120 µm-long MgB2 nanowire photon detectors is ∼100 ps, i.e. much shorter than that of their NbN counterparts.
Despite a low inductance, MgB2 photo detectors self-reset (i.e. they do not latch) at bias currents up to Ic . Our results show that MgB2 nanowire photon detectors are capable of single photon detection for both visible and communication (1550 nm) spectral ranges, combined with a low dark count rate, hence forming a technological platform for sensitive and high speed photon detectors with a large cross-section and a low jitter.  (Fig.3b)