High-performance superconducting nanowire single-photon detectors (SNSPDs) have facilitated numerous experiments and applications, particularly in the fields of modern quantum optics and quantum communication. Two kinds of optical coupling methods have thus far been developed for SNSPDs: one produces standard fiber-coupled SNSPDs in which the fibers vertically illuminate the meandered nanowires; the other produces waveguide-coupled SNSPDs in which nanowires are fabricated on the surface of a waveguide that guides photons, and the fibers are coupled to the waveguide. In this paper, we report on first experimental demonstration of a new type of SNSPD that is coupled with a microfiber (MF). Photons are guided by the MF and are evanescently absorbed by the nanowires of the SNSPD when the MF is placed on top of superconducting NbN nanowires. Room-temperature optical experiments indicated that this device has a coupling efficiency of up to 90% when a 1.3 μm-diameter MF is used for light with wavelength of 1550 nm. We were also able to demonstrate that our MF-coupled detector achieved system detection efficiencies of 50% and 20% at incident wavelengths of 1064 and 1550 nm, respectively, for a 2 μm-diameter MF at 2.2K. We expect that MF-coupled SNSPDs may show both high efficiency and broadband characteristics upon optimization and will be used for various novel applications, such as micro/nano-fiber optics.
© 2017 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Ever since superconducting nanowire single-photon detectors (SNSPDs or SSPDs) were demonstrated, they have attracted a considerable attention due to the detectors’ promising performance that includes high detection efficiency (DE), low dark count rate (DCR), broadband sensitivity, and low timing jitter . The system detection efficiency (SDE) has been improved to be over 90% with a wavelength of 1550 nm in a fiber-coupled system by embedding the SNSPD in optical cavities [2, 3]. However, the implementation of optical cavity usually reduces the bandwidth of SNSPDs. As a result, there are some recent works on increasing the bandwidth of SNSPD [4–6]. Furthermore, the size of the active area of meandered nanowires has to be compatible with the core size of an optical fiber such that good optical coupling can be achieved; this setup requires a long nanowire with a large kinetic inductance, thus resulting in a small counting rate and a large timing jitter, particularly for multi-mode fiber-coupled devices [7, 8].
Nevertheless, a particularly impressive waveguide-coupled SNSPD has been developed in recent years [9–12]. By placing an SNSPD atop an optical waveguide that has been fabricated on various substrate materials, photons propagating in the waveguide have been found to be evanescently absorbed by nanowires that are only a few tens of microns long. The on-chip DE of this device has been estimated to be over 90%, thus indicating that it has both high absorptance and high intrinsic DE. If the waveguide and the nanowire are correctly dimensioned, the absorptance of this device can reach over 90% for a wavelength range of more than 700 nm . Furthermore, the length of the nanowire is shorter than that of the fiber-coupled SNSPD, thus leading to a higher count rate and lower timing jitter (~18 ps) [9, 11]. However, unsatisfactory coupling losses between the optical fiber and the waveguide in such devices can decrease its SDE (~10%), thus limiting its applicability to quantum photonic integrated circuits (QPIC) .
Recently, we proposed a type of SNSPD that was coupled with microfibers (MFs), and simulations for this SNSPD indicated that its absorption would be as high as 90% for an array of 7 parallel nanowires, each of length 43 μm . Microfibers and nanofibers, tapered adiabatically from standard fibers, with diameters close to or smaller than the wavelength of the guided light were demonstrated to have low transmittance losses by Tong et al.  in 2003. MFs have been shown to have many interesting properties, such as strong evanescent fields, tight optical confinement, wideband abilities, and small masses; these properties are appealing for applications such as optical sensing and coupling . MFs also provide seamless connections with high transmittance from standard fibers ; therefore, MF-coupled SNSPDs can achieve both high SDEs and wideband compatibility.
In this paper, we first demonstrated MF-coupled SNSPDs experimentally. Room-temperature optical experiments found that they had absorptance of up to 90% for a 1.3 μm-diameter MF when light with an incident wavelength of 1550 nm was used. At 2.2 K, we were able to demonstrate that the MF-coupled detectors achieved maximum SDE values of up to 50% and 20% at wavelengths of 1064 and 1550 nm, respectively, with 2 μm-diameter MFs.
2. Device design and fabrication
We fabricated the SNSPDs on 6.5 nm-thick NbN films deposited on MgF2 (001) single crystal substrates (the detailed fabrication process is reported in ). Low-refractive-index MgF2 was used as a substrate to guarantee light transmission in the MF. The films were patterned into a meandered nanowire whose line width and pitch were both 100 nm; the patterning was done using electron beam lithography and reactive ion etching. The active area of the nanowires was designed into an array of parallel nanowires with four different parameters: eleven 50 μm-long lines, eleven 20 μm-long lines, three 250 μm-long lines and three 100 μm-long lines. To pattern the long nanowire without the stitching error, we set the writing field of electron beam lithography to 500 μm while keeping the acceleration voltage as 100 kV. The different design was created to compare our results with the simulated absorptance from MF to nanowire. The contact pads of the SNSPDs were designed not only to meet the requirements of MF coupling but also to reduce the accumulation of electrostatic charge on the non-conductive MgF2 substrates. We fabricated a 1.3 μm-diameter MF from a standard optical fiber by using a flame-heated taper drawing method . By using a low-refractive-index adhesive (LRIA) cladding, we affixed a 1.3 μm-diameter MF to the top of an SNSPD, as shown in Fig. 1.
To align the MF to the nanowire area, a microscope and a three-dimensional precision adjusting device were used to achieve micro–scale manipulation. We designed a nitrogen (N2) gas chamber that would meet the requirements for the curing of the LRIA and then the coupling process could be done within the chamber. The chamber is an assembly structure that can be disassembled into three sections: the base section, the right section with the adhesive port and the left section with the MF port, the schematic plot is shown in Fig. 2.
The coupling process will be described by six steps in Fig. 2. A sample of the device was placed at the center of the gas chamber, and the chamber was mounted onto the stage of the microscope such that the coupling process can be monitored through the microscope port of the chamber [Fig. 2(a)]. The MF was inserted to the chamber through an MF port by using a glass slide, and it was aligned to the nanowires by using the 3D precision adjusting device under a 50 × objective [Fig. 2(b)]. We used a fiber probe to dip the LRIA and dropped it onto the nanowire area through the adhesive port of the chamber [Fig. 2(c)]. N2 was added to the chamber through the N2 input port, and the adhesive was cured by UV light in a N2 atmosphere. Each port of the gas chamber was covered by preservative films such that a good N2 atmosphere could be maintained [Fig. 2(d)]. Other parts of the MF on the substrate should be covered by the adhesive such that low optical loss would be obtained. So we removed the right section in order to change 50 × objective to 5 × objective to add LRIA on the other parts of the MF [Fig. 2(e)], and the curing process of the adhesive was the same as in Fig. 2(d). The optical coupling process was then completed, both the right and left sections were removed from the base section and finally the MF was removed from the glass slide [Fig. 2(f)].
In the coupling process, the lateral center mismatch between MF and nanowires array can be well controlled within 0.5 μm with the aid of the microscope and alignment marks on chip. As the LRIA will cover the whole fiber taper part, the MF-nanowire hybrid structure is robust and isolated from the environment.
To calculate the initial optical loss of the device, we coupled an MF to a blank MgF2 substrate that does not contain any nanowires. Given that MgF2 forms typical rutile-like tetragonal crystals, the refractive index of MgF2 varies with regard to the direction of incident light. The refractive indices of MgF2 to light orthogonal to the optical axis (no) and along the optical axis (ne) were measured to be 1.3709 and 1.3823, respectively, by using prism coupler instruments. The refractive index of LRIA (nadh) is 1.3692. The modes with dominant horizontal and vertical electric field components were defined as being the TE and TM modes. Assuming the optical axis is along the Y-axis, the simulated profiles of the TE and TM modes for different MF diameters are shown in Fig. 3 (the simulation method is reported in an earlier paper ). The TM mode was found to experience a large amount of energy leakage into the MgF2 substrate for fiber diameters equal to and below 1.2 μm; this phenomenon was due to the larger difference between the refractive indices of ne and nadh. However, the TE mode was found to have better optical confinement than the TM mode. To avoid optical leakage into the substrate, we chose a diameter of 1.3 μm for the MF because both the TE and TM modes of an MF of this size exhibit good optical confinement abilities. To measure the transmittance of the TE and TM modes in a 1.3 μm diameter-MF on a blank MgF2 substrate, a squeeze-type polarizer (Thorlabs: PLC-900) was used to control the polarization of the light input into the MF. The residual light output from the MF was monitored by a power meter. When the output power reached the minimum value, the light in the MF was regarded as the TE mode, while the maximum value corresponded to the TM mode. For a blank MgF2 (001) substrate, the optical axis is along the Y-axis, and the transmittances of the TE and TM modes were found to be 90% and 84%, respectively; these values corresponded to optical losses of 0.46 dB and 0.76 dB, respectively.
We measured the transmitted power of the MFs during a coupling process conducted at room temperature (300K) and calculated the absorptance of the nanowires; the results are shown in Table. 1. P0 is the measured output power of a free-standing MF in the air that is not coupled to the nanowires.
We found that after the MFs were coupled to the nanowires, their output optical powers became dependent on the polarization of the incident light because of the difference between the absorptance of the TE and TM modes. Given that the TE mode has a larger absorptance than the TM mode , the maximum output power (minimal absorption of the SNSPD) was recorded as Pmax and was recognized as being the transmitted power of the TM mode, whereas the minimum output power (maximal absorption of the SNSPD) was recorded as Pmin and was recognized as being the transmitted power of the TE mode.
The measured absorptance, ηm, of nanowires of the TE mode can be represented by the following equation:
The numerical simulation was conducted with a 2D finite element method using COMSOL Multiphysics . The effective refractive index of propagation modes in the MF, neff, was simulated, and the imaginary part of neff represents the absorption of the nanowires. The theoretical absorptance, ηtheo, could then be calculated as follows:Table.1 indicates that the deviations of all the measured absorptance values, ηm, from the calculated absorption values, ηtheo, were within 10%; this result indicates the good controllability of the experiment. The negative deviations of the absorptions for fibers #12, #14, and #15, were attributed to MF–nanowire misalignments; meanwhile, the positive deviation for fiber #16 can be explained by the fact that its optical loss was greater than 0.46 dB because its Pmax value was smaller than that of the other fibers. The maximum ηm was found for fiber #12, and it was measured to be 89.9% for an active area that consisted of eleven 50 μm-long lines.
In the low-temperature measurements of the MF-coupled SNSPD, we packaged the devices in copper blocks and mounted them on the cold head of the two-stage Gifford-McMahon (G-M) refrigerator with a working temperature of 2.200±0.005K. A few MF-coupled SNSPDs have experienced several thermal cycling between 2K and 300 K. No evident difference was observed. A supercontinuum laser (NKT: EXB-3) was used as the light source with a wavelength of 532–1550 nm. The power of the laser was attenuated to be a single photon per pulse level by using two variable attenuators. The light was then fed into the G–M refrigerator through the MF coupled to the nanowires and the residual light output from the MF was then monitored by a power meter to estimate the absorptance of the nanowires at low temperature. The polarization of the light was adjusted using a polarization controller to achieve the TE mode for the MF (and the maximum SDE for the MF-coupled SNSPD). The MF-coupled SNSPD was biased by a quasi-constant-voltage source through a Bias-Tee, and the voltage pulses generated by the MF-coupled SNSPD were transmitted through the Bias-Tee before being amplified using a low-noise amplifier. The amplified signal was then fed into a photon counter or an oscilloscope to characterize the performance of the MF-coupled SNSPD.
Based on the properties of the LRIA (Luvantix, SPC-373 AP) used in our MF-coupled SNSPD, we knew that its refractive index would increase during the cooling of the device. We measured the optical losses as a function of temperature for MFs with different diameters coupled to blank MgF2 substrates, and compared the measured optical losses with the stimulation results to determine the temperature variation of the refractive index of the LRIA. The refractive index of LRIA was found to change from 1.37 (300 K) to 1.41 (2.2K), which resulted in a significant optical loss (~25 dB) measured at 2.2 K. This loss is ~300 times larger than the loss observed at room temperature. As a result, the SDE at this temperature was found to be less than 0.1% for the MF-coupled SNSPD at a wavelength of 1550 nm. Thus, a larger MF diameter was used (2 μm instead of 1.3 μm) to reduce the low-temperature optical loss of the SNSPD to be 3.3 dB for a wavelength of 1550 nm. However, this setup resulted in the room-temperature measured absorptance, ηm, of the SNSPD with an active area consisting of eleven 50 μm lines, decreasing to ~80% due to the larger diameter of the fiber. At temperatures of 2.2 K, meanwhile, for a total optical loss of 3.3 dB, the low-temperature absorptance, ηm, of the SNSPD was measured to be around 60%. The low-temperature optical loss can be reduced by using a LRIA with lower refractive index at 300K to achieve the designed refractive index (1.37) at 2.2K. We noticed that a new LRIA (Polymers, MY-133-EA) with a refractive index of 1.333 at 300K, which will be used in the future research. It is noted here that, the temperature dependence of the LRIA refractive index is around −10−4/K, and our customized cryostat based on G-M cryocooler exhibited a temperature fluctuation of 5 mK, so the change of refractive index due to the temperature fluctuation is estimated to be 5 × 10−7, which can be neglected.
3. Device characterization
We measured SDE and DCR of the device (see the optical image in Fig. 4(a)) as a function of the normalized bias current, Ib/Isw, where Isw is the maximum bias current for an SNSPD switching from the superconducting state to the resistive state [Fig. 4(b)]. The SDE is defined by the detected counts subtracting the dark counts divided by the incident photon number (~106 per second). This batch of devices used the same parameters (2-μm-diameter MF and eleven 50-μm nanowires). Here we present the performances of our best devices. The device 2# shows an SDE of 18.2% at a DCR of 100 Hz and an unsaturated maximum SDE of 22.5% which can be seen in Fig. 4(b). If we fit an SDE–Ib curve with an empirical sigmoid function , the saturated SDE is estimated to be 28%. The current dependence of the DCR was found to have two different zones that are dominated by thermal radiation (low bias zone) and the intrinsic DCR (high bias zone); this finding is the same result that is found for traditional fiber-coupled SNSPDs that do not use any filters . The DCR was found to be around 100 Hz at a normalized bias current of 0.9, in close agreement with previous reports .
To estimate the spectrum sensitivity of the MF-coupled SNSPDs, we measured their SDEs for wavelengths (λ) that ranged from 532 nm to 1550 nm. Figure 5(a) shows the SDE of device 2# as a function of Ib for λ = 532, 850, 1064, 1250, 1310, and 1550 nm. We observed that the SDE decreases as a function of the wavelength between 1064 nm and 1550 nm. On the shorter wavelength range, the SDE drops also drastically and a saturated plateau was found at 1064 nm; the maximum SDE value of 50% was obtained at this saturated plateau. Further decrease in the wavelength caused the SDE to decrease drastically; this observation can be explained by the weak evanescent field at the surface of the MF when the wavelength becomes less than half of the diameter of the fiber. When the wavelength is 850 nm, abnormal unsaturated SDE behavior was obtained, which needs further investigation in the future. Simulation showed that the maximum SDE value is near 700 nm for an MF with a diameter of 1.3 μm. It indicates that MF-coupled SNSPDs may have a unique merit with high SDE for both visible and near-infrared wavelengths, i.e., a broadband SNSPD with high SDE.
Figure 5(b) shows the wavelength dependence of the saturated SDE and theoretical absorptance values for wavelengths of 1064–1550 nm. When the SDE is saturated, the probability to detect the absorbed photon (intrinsic detection efficiency or quantum efficiency) is usually considered to be 100%, so we use saturated SDE in Fig. 5(b) to make a direct comparison with the theoretical absorptance. The results for λ = 532–850 nm are not given because the mode of the guided light becomes multi-mode at these wavelengths in standard optical fibers, and these high-order modes will be cut off when the light propagates into the MF ; this phenomenon will cause great losses in the fiber and hinders the achievement of the theoretical absorptance calculated by the simulation. The experimental results in Fig. 5(b) follow the trend observed for the theoretical wavelength dependence of the absorptance, and the loss was 3.8 ± 0.8 dB; this finding is consistent with the estimated low-temperature optical loss (3.3 dB) discussed above. Based on calculation , the fluctuation of the SDE related to the misalignment between the MF and the nanowires is less than 2%. It indicates that there are other factors that affect the SDE of the MF-coupled SNSPD, which need to be explored in the future research.
In conclusion, we successfully fabricated MF-coupled SNSPDs. The room-temperature optical experiments we conducted indicated that the MF–nanowire absorptance of these devices reached 90% when a 1.3 μm-diameter MF was used with 1550 nm wavelength incident light. The SDE values for the TE mode obtained for 2 μm-diameter MF-coupled SNSPDs were 50% and 20% at wavelengths of 1064 and 1550 nm, respectively. Optimized MF-coupled SNSPDs are expected to achieve both high SDEs and wideband compatibility compared with fiber-coupled SNSPDs and waveguide-coupled SNSPDs. We have determined that several steps could be taken to further improve the performance of MF-coupled SNSPDs: (1) the fabrication process of SNSPDs could be optimized to improve their intrinsic DEs to a saturated value of 100%; (2) the optical losses could be reduced and improvements to the MF–nanowire absorptance values could be made by using an LRIA that has a suitable refractive index at low temperatures and by optimizing the diameter of the MF used; this process would enable the MF-coupled SNSPDs to exhibit both high SDE values and ultra-broadband capabilities; (3) the parameters of the devices could be further optimized by reducing the length of the nanowire to decrease the recovery time and then increase the count rates of the SNSPDs. In addition, the stimulation results in  indicates that the SDE for the TM mode of a MF-coupled SNSPD can be much smaller than for the TE mode because of an extremely low absorptance. This phenomenon has been observed in our experiments, which makes it interesting to develop SNSPD with high polarization sensitivity.
National Key R&D Program of China (2017YFA0304000); National Natural Science Foundation of China (61671438); Science and Technology Commission of Shanghai Municipality (16JC1400402).
References and links
1. G. N. Gol’tsman, O. Okunev, G. Chulkova, A. Lipatov, A. Semenov, K. Smirnov, B. Voronov, A. Dzardanov, C. Williams, and R. Sobolewski, “Picosecond superconducting single-photon optical detector,” Appl. Phys. Lett. 79, 705 (2001).
2. F. Marsili, V. B. Verma, J. A. Stern, S. Harrington, A. E. Lita, T. Gerrits, I. Vayshenker, B. Baek, M. D. Shaw, R. P. Mirin, and S. W. Nam, “Detecting single infrared photons with 93% system efficiency,” Nat. Photonics 7, 210–214 (2013).
3. W. J. Zhang, L. X. You, H. Li, J. Huang, C. L. Lv, L. Zhang, X. Y. Liu, J. J. Wu, Z. Wang, and X. M. Xie, “NbN superconducting nanowire single photon detector with efficiency over 90% at 1550 nm wavelength operational at compact cryocooler temperature,” Sci. China Phys. Mech. 60, 120314 (2017).
4. L. Redaelli, G. Bulgarini, S. Dobrovolskiy, S. N. Dorenbos, V. Zwiller, E. Monroy, and J. M. Gérard, “Design of broadband high-efficiency superconducting-nanowire single photon detectors,” Supercond. Sci. Technol. 29, 065016 (2016).
5. Y. Wang, H. Li, L. You, C. Lv, J. Huang, W. Zhang, L. Zhang, X. Liu, Z. Wang, and X. Xie, “Broadband near-infrared superconducting nanowire single-photon detector with efficiency over 50%,” IEEE Trans. Appl. Supercond. 27, 1–4 (2017).
6. T. Yamashita, K. Waki, S. Miki, R. A. Kirkwood, R. H. Hadfield, and H. Terai, “Superconducting nanowire single-photon detectors with non-periodic dielectric multilayers,” Sci. Rep. 6, 35240 (2016). [PubMed]
7. K. V. Smirnov, A. V. Divochiy, Y. B. Vakhtomin, M. V. Sidorova, U. V. Karpova, P. V. Morozov, V. A. Seleznev, A. N. Zotova, and D. Y. Vodolazov, “Rise time of voltage pulses in NbN superconducting single photon detectors,” Appl. Phys. Lett. 109, 705–3040 (2016).
8. A. J. Kerman, E. A. Dauler, W. E. Keicher, J. K. W. Yang, K. K. Berggren, G. Gol’tsman, and B. Voronov, “Kinetic-inductance-limited reset time of superconducting nanowire photon counters,” Appl. Phys. Lett. 88, 111116 (2006).
9. W. H. Pernice, C. Schuck, O. Minaeva, M. Li, G. N. Goltsman, A. V. Sergienko, and H. X. Tang, “High-speed and high-efficiency travelling wave single-photon detectors embedded in nanophotonic circuits,” Nat. Commun. 3, 1325 (2012). [PubMed]
10. K. Waki, T. Yamashita, S. i. Inoue, S. Miki, H. Terai, R. Ikuta, T. Yamamoto, and N. Imoto, “Fabrication and characterization of superconducting nanowire single-photon detectors on Si waveguide,” IEEE Trans. Appl. Supercond. 25, 1–4 (2015).
11. A. Vetter, S. Ferrari, P. Rath, R. Alaee, O. Kahl, V. Kovalyuk, S. Diewald, G. N. Goltsman, A. Korneev, C. Rockstuhl, and W. H. P. Pernice, “Cavity-enhanced and ultrafast superconducting single-photon detectors,” Nano Lett. 16(11), 7085–7092 (2016). [PubMed]
12. M. K. Akhlaghi, E. Schelew, and J. F. Young, “Waveguide integrated superconducting single-photon detectors implemented as near-perfect absorbers of coherent radiation,” Nat. Commun. 6, 8233 (2015). [PubMed]
13. Y. Xu, J. Wu, W. Fang, L. You, and L. Tong, “Microfiber coupled superconducting nanowire single-photon detectors,” Opt. Commun. 405, 48–52 (2017).
14. L. Tong, R. R. Gattass, J. B. Ashcom, S. He, J. Lou, M. Shen, I. Maxwell, and E. Mazur, “Subwavelength-diameter silica wires for low-loss optical wave guiding,” Nature 426(6968), 816–819 (2003). [PubMed]
15. L. M. Tong, F. Zi, X. Guo, and J. Y. Lou, “Optical microfibers and nanofibers: a tutorial,” Opt. Commun. 285, 4641–4647 (2012).
16. J. E. Hoffman, S. Ravets, J. A. Grover, P. Solano, P. R. Kordell, J. D. Wong-Campos, L. A. Orozco, and S. L. Rolston, “Ultrahigh transmission optical nanofibers,” AIP Adv. 4, 067124 (2014).
17. J. J. Wu, L. X. You, L. Zhang, W. J. Zhang, H. Li, X. Y. Liu, H. Zhou, Z. Wang, X. M. Xie, and Y. X. Xu, “NbN superconducting nanowire single-photon detector fabricated on MgF2 substrate,” Supercond. Sci. Technol. 29, 065011 (2016).
18. T. Yamashita, S. Miki, H. Terai, and Z. Wang, “Low-filling-factor superconducting single photon detector with high system detection efficiency,” Opt. Express 21(22), 27177–27184 (2013). [PubMed]
19. T. Yamashita, S. Miki, W. Qiu, M. Fujiwara, M. Sasaki, and Z. Wang, “Performances of fiber-coupled superconducting nanowire single-photon detectors measured at ultralow temperature,” IEEE Trans. Appl. Supercond. 21, 336–339 (2011).
20. M. C. Frawley, A. Petcu-Colan, V. G. Truong, and S. N. Chormaic, “Higher order mode propagation in an optical nanofiber,” Opt. Commun. 285, 4648–4654 (2012).