betaFeedback

WorksPhysics and Astronomy

A superconducting-nanowire single-photon camera with 400,000 pixels

Abstract

For the last 50 years, superconducting detectors have offered exceptional sensitivity and speed for detecting faint electromagnetic signals in a wide range of applications. These detectors operate at very low temperatures and generate a minimum of excess noise, making them ideal for testing the non-local nature of reality, investigating dark matter, mapping the early universe, and performing quantum computation and communication. Despite their appealing properties, however, there are currently no large-scale superconducting cameras - even the largest demonstrations have never exceeded 20 thousand pixels. This is especially true for one of the most promising detector technologies, the superconducting nanowire single-photon detector (SNSPD). These detectors have been demonstrated with system detection efficiencies of 98.0%, sub-3-ps timing jitter, sensitivity from the ultraviolet (250nm) to the mid-infrared (10um), and dark count rates below 6.2e-6 counts per second (cps), but despite more than two decades of development they have never achieved an array size larger than a kilopixel. Here, we report on the implementation and characterization of a 400,000 pixel SNSPD camera, a factor of 400 improvement over the previous state-of-the-art. The array spanned an area 4x2.5 mm with a 5x5um resolution, reached unity quantum efficiency at wavelengths of 370 nm and 635 nm, counted at a rate of 1.1e5 cps, and had a dark count rate of 1e-4 cps per detector (corresponding to 0.13 cps over the whole array). The imaging area contains no ancillary circuitry and the architecture is scalable well beyond the current demonstration, paving the way for large-format superconducting cameras with 100% fill factors and near-unity detection efficiencies across a vast range of the electromagnetic spectrum.

For the last 50 years, superconducting detectors have offered exceptional sensitivity and speed for detecting faint electromagnetic signals in a wide range of applications. These detectors operate at very low temperatures and generate a minimum of excess noise, making them ideal for testing the non-local nature of reality [1, 2], investigating dark matter [3, 4], mapping the early universe [5, 6, 7], and performing quantum computation [8, 9, 10] and communication [11, 12, 13, 14]. Despite their appealing properties, however, there are currently no large-scale superconducting cameras – even the largest demonstrations have never exceeded 20 thousand pixels [15]. This is especially true for one of the most promising detector technologies, the superconducting nanowire single-photon detector (SNSPD) [16, 17, 18]. These detectors have been demonstrated with system detection efficiencies of 98.0% [19], sub-3-ps timing jitter [20], sensitivity from the ultraviolet (250 nm) [21] to the mid-infrared (>10 μm>10\,\mu\mathrm{m}) [22], and dark count rates below 6.2×10−66.2 \times10^{-6} counts per second (cps) [3], but despite more than two decades of development they have never achieved an array size larger than a kilopixel [23, [24]. Here, we report on the implementation and characterization of a 400,000 pixel SNSPD camera, a factor of 400 improvement over the previous state-of-the-art. The array spanned an area 4×2.54 \times2.5 mm with a 5×55 \times5 µm resolution, reached unity quantum efficiency at wavelengths of 370 nm and 635 nm, counted at a rate of 1.1×1051.1 \times10^{5} cps, and had a dark count rate of 1.0×10−41.0 \times10^{-4} cps per detector (corresponding to 0.13 cps over the whole array). The imaging area contains no ancillary circuitry and the architecture is scalable well beyond the current demonstration, paving the way for large-format superconducting cameras with 100% fill factors and near-unity detection efficiencies across a vast range of the electromagnetic spectrum.

Most superconducting sensors [25], such as microwave kinetic-inductance detectors (MKIDs) or hot-electron bolometers, produce continuously-valued outputs that lend themselves well to efficient readout schemes such as frequency multiplexing [26]. Superconducting nanowire single-photon detectors, however, are notoriously difficult to multiplex since they produce discrete pulses that are both low-amplitude and broadband. These pulses are typically read out individually using one microwave readout cable per detector, unlike CCD-type sensors which accumulate charge at each pixel and can be serially interrogated with a single readout line. This per-pulse readout process be advantageous when photon-counting as it means there is no read noise, but severely limits the number of detectors that can be read out due to cryogenic thermal load limitations on the readout wiring.

To date, there have been many attempts to devise readout architectures that minimize the number of readout lines running from room temperature electronics to the detector chip. For example, one of the largest architectures to date used a time-of-flight measurement to measure photon position along the length of the detector and required only two readout lines [27]. Unfortunately, since the entire imager was made from a single long nanowire, this architecture is susceptible to fabrication yield issues, as a single discontinuity can compromise the entire imager. Additionally, the large inductance of the long nanowire will inevitably lead to very long reset times and limit the maximum count rate.

In row-column readout architectures, SNSPDs are arranged on an N×NN \times N grid [28], enabling readout of N2N^2 pixels using 2N2N readout lines. The detection event is defined as correlated voltage pulses appearing on both row and column lines within a predefined time window. The row-column readout architecture was used to create the largest SNSPD array to date, a 1,024-pixel imager, but due to diminishing signal-to-noise (SNR) ratio with increasing pixel count, this approach cannot scale to much larger sizes [29]. The thermal row-column scheme overcomes the SNR issue by keeping the row and column pixels electrically isolated and uses the local heat generated during the SNSPD detection process to generate correlated events on row and column channels [30].

More recently, the thermally-coupled imager (TCI) architecture used a combination of thermal coupling to a superconducting bus and time-multiplexing along the readout bus, to achieve kilopixel imaging capability [24]. This scheme used an asymmetric thermal coupling process to transmit pulses on a row or column detector to a readout bus, significantly reducing inter-detector crosstalk and enabling much greater scalability. The SNSPD array demonstrated in this work combines a thermal row-column sensor element with TCI readout in order to achieve its large degree of multiplexing. Additionally, the perpendicular orientation of the two nanowire layers enables polarization-insensitive optical cavity designs.

The array, shown in Figure 1, was comprised of 800 row and 500 column detectors and was operated at a temperature of 0.8 K. The row and column detectors were made by patterning 1.11.1-μm\mu\mathrm{m}-wide wires from two independent 44-nm-thick WSi films with a TcT_{\mathrm{c}} of 3.4 K and kinetic inductance of 250 pH/sq. Row and column detectors were spaced 5 μ\mum apart and arranged in a grid, giving a resolution of 5×55 \times5 μ\mum as shown in Figure 1c. There were four buses used for readout, requiring a total of 8 microwave coax lines, and the row and column detectors were interleaved between them as shown in Figure 1c. Each row or column detector was connected to a resistive thermal coupler, which contained a small heater that is thermally coupled to (but electrically isolated from) the readout bus by a thin, electrically insulating dielectric spacer.

Image not extracted

Figure 1. (a) Imaging with the 800×500800 \times500 array at 370 nm370\,\mathrm{nm}. Raw time-delay data from the buses are shown as individual dots in red, and binned 2D histogram data is shown in black and white. (b) Count rate as a function of bias current for various wavelengths of light as well as dark counts. (c) False-color scanning electron micrograph of the lower-right corner of the array, highlighting the interleaved row and column detectors. (inset, lower left) Schematic diagram showing detector-to-bus connectivity. (inset, lower right) Closeup showing 1.1 μm1.1\,\mu\mathrm{m} detector width and effective 5×5 μm5 \times5\,\mu\mathrm{m} pixel size.

The readout buses were made from the same layer of material as the column detectors and so have the same basic properties. The majority of the bus length had a width of 8 μm8\,\mu\mathrm{m} that, in combination with the 5050-nm-thick SiO2_2 dielectric spacer and ground plane, formed a microstrip transmission line with approximately 50 Ω50\,\Omega impedance. During operation, these sections of the bus contained a relatively low current density and were thus not directly photosensitive, allowing the readout bus to avoid generating false counts from stray photons. Below each heating element, the bus briefly narrowed to a 1.51.5-µm-wide constriction with high current density, creating a heater-tron like device [31]. The high current density ensured that every time the heater was activated during a detection event, it created a corresponding hotspot on the bus.

Detectors were grouped into sections of 50 rows or 50 columns, and each section was biased independently. For each section, the bias-current was resistively distributed among the 50 detectors, sharing a common source and ground. Figure 2 shows the circuit diagram for a single section comprising 50 column SNSPDs. When a photon was absorbed by a detector, a hot-spot formed on the detector with a peak value of a few kΩ\Omega. This resistance diverted the bias current out of the detector and into the heating element of the thermal coupler. The resulting phonons generated within the thermal coupler locally destroyed the superconducting state on the readout bus, creating two opposite-polarity voltage pulses that traveled down the readout bus as shown at the top and inset of Figure 2. Owing to the large kinetic inductance of the bus material, these microwave pulses propagated at 0.36%0.36\% of cc (1.10×1061.10 \times10^{6} m/s), allowing relatively short lengths of bus (400 μm400\ \mu\mathrm{m}) to separate adjacent detectors while maintaining excellent distinguishability.

Image not extracted

Figure 2. (a) Circuit diagram of a bus and one section of 50 detectors with ancillary readout components. SNSPDs are shown in the gray boxes, and all other components are placed outside of the imaging area. A photon that arrives at time t0t_0 has its location determined by a time-of-flight readout process based on the time-of-arrival difference t2−t1t_2-t_1. (b) Oscilloscope traces from a photon detection, showing the arrival of positive (green) and negative (red) pulses at times t1t_1 and t2t_2 respectively.

By measuring the time-difference between the arrival of these pulses at the ends of the buses, one can compute the location of the detection event and the corresponding row or column. As shown in Figure 2a, for a photon arriving at time t0t_0, the emitted positive pulse will arrive at the left readout at time t1=t0+τ1t_1=t_0+\tau_1, and the negative pulse at t2=t0+τ2t_2=t_0+\tau_2. Applying this process to both row and column readout, we were able to precisely identify the location where each photon was absorbed. Given arrival times of t1t_1 and t2t_2 from a column bus and t3t_3 and t4t_4 from a row bus, we calculated the differential time-delays as Δtcol=t1−t2\Delta t_{\mathrm{col}} = t_1 - t_2, Δtrow=t3−t4\Delta t_{\mathrm{row}} = t_3 - t_4. Similarly, the photon time of arrival can be calculated from t0=(t1+t2−τb)/2t_0 = (t_1 + t_2 - \tau_b)/2 where τb\tau_b is the time-length of the bus. To ensure that these pulses were generated by the same single-photon detection event, we imposed the four-fold coincidence condition that the arrival time of all four pulses were within 100 ns of each other.

To capture the image shown in Figure 1a, we patterned a metallic mask on a glass substrate, placed it directly on top of the array, and flood illuminated the entire mask. We then recorded the pulse timings coming out of the readout buses with a timetagger, extracted the four-fold coincidence events, and computed Δtcol\Delta t_{\mathrm{col}} and Δtrow\Delta t_{\mathrm{row}}. No further post processing of the pulse data was performed and no 4-way coincidences were discarded. The raw differential-delay data shown in red corresponds closely to the black and white binned histogram image, excepting the regularly-spaced gaps in the differential-delay data that correspond to extra bus delay between adjacent detector sections.

Figure 1b shows photon count rate measurements for multiple wavelengths for one of the sections as a function of bias current. As evident from the figure, our detectors have large plateau regions for wavelengths 370 nm and 635 nm, indicating unity quantum efficiency over a wide range of bias current. At 370 nm, the wavelength for which this array was targeted, a bias current bias of 40 μ\muA yielded unity quantum efficiency and at this bias current we recorded 5 dark counts over a measurement period of 1,000 s. Since this measurement occurred over a section of 50 detectors, this corresponded to a dark count rate of 1±0.45×10−41 \pm0.45 \times10^{-4} cps per detector. In a setup that was better shielded optically, this value would likely be even lower, as the dark count curve in Figure 1b suggests that the source of dark counts at this bias current were likely stray blackbody photons [32].

After an initial screening we observed that some detectors were defective, producing anomalously high dark count rates, overwhelming the bus, and obfuscating true counts. To resolve this, we applied a measure-and-prune process to eliminate the faulty detectors. We first determined exactly which detectors were anomalous by flood illuminating the array and collecting 1D histograms for each bus. We then identified any row or column detectors that produced excessive counts and disconnected them by etching through their bias wiring, effectively turning them off. Within a single iteration of this pruning process, we were able to effectively remove all of the defective detectors (58 defective detectors out of 1300), resulting in the smooth Gaussian-spot shape of the histogram shown in Figure 3a. Identifying the offending detectors was straightforward – the full-width half-max of the peaks was 62 ps and the separation of the peaks had a median value of 362 ps, giving a greater than 6 FWHM distinguishability between neighboring rows or columns. We further examined the 2D uniformity of the array by flood illuminating it and examining the output of the buses. As shown in Figure 3b, the photon counts are tightly clustered in ∼62\sim62-ps-diameter circles, indicating that the readout process works to accurately position both the photon row and column simultaneously.

Image not extracted

Figure 3. (a) Histogram of the pulse differential time-delays Δt=t1−t2\Delta t = t_1 - t_2 from the north bus during flood illumination with a Gaussian spot. All 400 detectors resolved clearly, with gaps indicating detectors that were pruned. (a, inset) Zoomed region showing that counts from adjacent detectors are easily resolvable, and no counts were generated by a pruned detector. (b) Plot of raw trowt_{\mathrm{row}} and tcolt_{\mathrm{col}} time-delays when flood illuminated at 370 nm. (c) Zoomed subsection of the array with 25×2525 \times25 detectors. (d) Histogram of time-delays for a 2×22 \times2 detector subset with 10 ps bin size showing clear distinguishability between adjacent detectors.

One potential area for concern was that detected photons may trigger the detectors but not produce corresponding outputs on the buses, resulting in lost detection events. We characterized this photon-to-bus conversion efficiency with test structures, and found the minimum energy required to create a hotspot on the bus was 8.1×10−178.1 \times10^{-17} J. In the array, the detection process deposited 12LsIbias=1.2×10−15\frac{1}{2} L_{\mathrm{s}} I_{\mathrm{bias}} = 1.2 \times10^{-15} J of energy on the heater, greatly exceeding the requirements of the thermal coupler. Correspondingly, during operation we observed large operating margins for the detectors and buses, indicating good uniformity and sensitivity of the thermal coupling process. When the detectors were biased between 30-55 μ\muA, any IbusI_{\mathrm{bus}} between 15-40 μ\muA produced a photon-to-bus conversion efficiency of approximately unity. We note that while the thermal coupling efficiency was high, an area for improvement will be the increasing the density of the layout for high fill-factors – in the this work, only 13.7% of all photons detected resulted in a four-fold coincidence, due to the relatively large spacing between adjacent detectors.

We additionally measured the timing performance of the array. Figure 4a shows the count rate from one of the 50-detector sections versus optical attenuation measured at 635 nm and with the detectors biased at 45 μA45\,\mu\mathrm{A} per detector. We observed that the photon count rate was proportional to the photon flux, indicating the array was operating in the single-photon regime [16] [33]. The count rate was limited primarily by the speed of the bus – after detection, the current in the bus dropped to zero for a period of time τdead\tau_{\mathrm{dead}}, and any other photon detections during this time would be lost. For our device this dead time was 9.16 μs. Similar to single-pixel SNSPDs, once count-rates approached 1/τdead1/\tau_{\mathrm{dead}}, photons could still be lost due to blocking loss as evidenced by the 3 dB compression point at 1.09×1051.09 \times10^{5} cps. To increase the maximum readout rate, additional buses may be added trivially. For a fixed array size, increasing the total number of buses by a factor of BB leads to a BB-fold increase in the maximum readout rate per bus and thus a B2B^{2}-fold increase in the maximum readout rate of the entire array.

Image not extracted

Figure 4. (a) Count rate versus optical attenuation for a section of detectors biased at 45 μA45\,\mu\mathrm{A} per detector. The dashed purple line shows a slope of 1, with deviations from that line at higher rates indicating blocking loss. (b) System jitter of a 50-detector section. Detection delay was calculated as the time elapsed between the optical pulse being generated and detection event recorded.

Figure 4b shows the timing jitter of one of the array sections. To obtain this data a single section was biased and illuminated with a 1550 nm mode-locked laser with 100 fs optical pulse width. The distribution of detection delays had the FWHM of 336 ps and standard deviation of 245.4 ps. This relatively large jitter can be primarily attributed to the longitudinal geometric jitter of our 5 mm long detectors, and could be significantly reduced by converting to a differential-style readout [34].

In this work, we demonstrated the largest superconducting photon-counting camera to date by a factor of 20, using an efficient, time-resolved architecture that is both scalable and tolerant of fabrication defects. The architecture allows for the imaging area to contain only detectors, allowing future arrays to reach 100% fill factors with high system detection effiencies. While the work here demonstrated classical imaging at the single-photon level, we expect future large-format implementations to greatly enable applications in quantum imaging[35] such as sub-Rayleigh super-resolution imaging[36] or sub-shot-noise imaging[37]. These applications typically require the detection of quantum correlations in both time and space, and depend heavily on high efficiencies and low background noise.

Methods

The chip was fabricated with 6 layers: two WSi detectors layers, two Nb wiring layers, a resistor layer, and a dielectric spacer layer. First, we sputtered a thin (4 nm) WSi film on a 75 mm Si wafer with 150 nm of thermal oxide. The film was in-situ capped with 2 nm of aSi to avoid oxidation. Then column detectors, series inductors and the readout bus were patterned using SPR660 and optical lithography with an i-line (365 nm) stepper. Finally these patterns were etched with an inductively-coupled plasma (ICP) etcher using an AR:SF6_6 etch. Next, a liftoff pattern was created using optical lithography and 90 nm thick Nb wires with 10 nm of PdAu capping were sputtered to create a low inductance bottom Nb wiring layer.

At this point, it is important to planarize the surface before depositing row detectors. In our tests, it was observed that if this planarization step is skipped, the topography of the column detectors results in a step-like edges on the row detectors which reduces the critical current of row detectors, degrading their performance. The planarization can be achieved by using a dilute divinylsiloxane-benzocyclobutene (DVS-BCB) polymer. This BCB polymer was applied to the wafer, and spun dry at 4000 rpm. The wafer was later baked at 250 C in a nitrogen furnace for 60 min, to cure the BCB. Later, 25 nm of SiO2_2 was sputtered on the wafer with 4 nm of amorphous silicon sputtered in-situ as a capping layer. Finally, the wafer was patterned and etched with a CHF3_3 and O2_2 chemistry to create vias through the dielectric layer. The PdAu deposited on the bottom Nb wiring layer acts as a non-oxidizing etch stop, allowing us to overetch and create 1 × 1 µm vias.

Next, 4 nm of WSi film was sputtered again, patterned and etched to produce row detectors and the heaters on the thermal coupler. Later, 30 nm of PdAu was sputtered with a liftoff process to create all of the resistors needed for the device operation (see Figure 2). Finally, 200 nm of Nb was sputtered and patterned with a liftoff process to create the top Nb wiring layer and the GND plane for readout buses. 2 nm of Ti was used as an adhesion layer for the Nb and PdAu layers.

After this we performed the final etch to disconnect the ground from the readout bus, which was intentionally left connected to the ground plane. Skipping this step usually causes antenna-like RF-pickup, leading to damage of the readout bus in our devices. To improve the yield and reproducibility of lithographically defined patterns, the detectors were designed as images containing 10 detectors, and the stepper was used to repeat this pattern forming the full array. The readout circuits and readout bus were designed to connect to 50 detectors each. The overall pattern was generated by rotating the wafer inside the stepper and repeating these images, hence both column and row detectors and their readout circuitry are produced with the same reticles.

The chip was fabricated as a 1000×10001000 \times1000 array, and during testing we identified a uniform section of 800 column detectors and 500 row detectors, which we designated as the active detection area. This demonstrates the flexibility of the architecture, since the unbiased sections of the array do not waste any bandwidth on the readout buses. The same technique can be straightforwardly applied to allow selection of an active area on-the-fly.

Circuit parameter values

In this implementation, the values for the circuit elements were optimized using test structures as well as SPICE simulations. The section biasing resistor RbiasR_{\mathrm{bias}} was 1000 Ω\Omega, the detector biasing resistor RsR_{\mathrm{s}} was 80 Ω\Omega, the shunting resistor RshR_{\mathrm{sh}} was 16 Ω\Omega, the thermal coupler series resistor RtcR_{\mathrm{tc}} was 16 Ω\Omega, the series inductor LsL_{\mathrm{s}} was 1.25 μH\mu\mathrm{H}, and the inductance of the SNSPD LsnspdL_{\mathrm{snspd}} was 1.14 μH\mu\mathrm{H}. The biasing resistor RsR_{\mathrm{s}} needed to be sufficiently large to ensure that current was distributed uniformly to each detector – if made too small, parasitic resistances in the wiring or vias can cause uneven distribution. However, larger values also contribute to the static power consumption of the array and so its value cannot be too large. RshR_{\mathrm{sh}} was chosen to be relatively small, as it allows the current to be shunted to ground after a detector clicked. By choosing an appropriately small value, the detectors are sufficiently shunted such that latching is impossible, simplifying the operation of the array. Additionally, a small value of RshR_{\mathrm{sh}} guarantees that voltages generated during the detection process are shunted before they reach the communal bias line, minimizing crosstalk between neighboring detectors in a section.

A large value of LsL_{\mathrm{s}} serves to both minimize crosstalk voltages on the bias line during detection, and acts as an energy reservoir with which to power the heating element on the thermal coupler. Too-large values, however, result in wasted space on the chip as well as generate excess heating of the bus that can negatively affect reset times. While smaller LsnspdL_{\mathrm{snspd}} values would be preferred to minimize geometric jitter and reset time in the detectors, for this design we required least 5 mm-long detectors to span the imaging area.

Data availability

Unprocessed time tags and data used to produce images and graphs shown in Figure 1,Figure 3, and Figure 4 are available from the corresponding author on reasonable request.

Code availability

Code used to process the time tags and produce images shown in Figure 1 and Figure 4 is available from the corresponding author on reasonable request.

Acknowledgements

The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation thereon. Part of this research was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA (Contract No. 80NM0018D0004). A.N.M. was supported in part by NASA APRA via Grant No. NNH17ZDA001N. Support for this work was provided in part by the DARPA DSO Invisible Headlights program. Certain equipment, instruments, software, or materials, commercial or non-commercial, are identified in this paper in order to specify the experimental procedure adequately. Such identification is not intended to imply recommendation or endorsement of any product or service by NIST, nor is it intended to imply that the materials or equipment identified are necessarily the best available for the purpose. This research was funded by NIST (https://ror.org/05xpvk416), University of Colorado Boulder (https://ror.org/02ttsq026), and the Jet Propulsion Laboratory (https://ror.org/027k65916).

Author contributions

ANM, BK, and BGO conceptualized these experiments. Fabrication of devices was done by BGO. Measurements were performed by BGO, DSR and ANM. Analysis and interpretation of the data was done by BGO, SWN, MDS, BK, and JA. ANM directed and supervised this work. BGO and ANM prepared the manuscript with input from all co-authors.

References

  1. [1]Giustina, M. et al. Significant-Loophole-Free Test of Bell’s Theorem with Entangled Photons. Physical Review Letters 115, 250401. ISSN: 0031-9007. https://link.aps.org/doi/10.1103/PhysRevLett.115.250401 (Dec. 2015).DOI
  2. [2]Shalm, L. K. et al. Strong Loophole-Free Test of Local Realism. Phys. Rev. Lett. 115, 250402. https://link.aps.org/doi/10.1103/PhysRevLett.115.250402 (25 Dec. 2015).DOI
  3. [3]Chiles, J. et al. New Constraints on Dark Photon Dark Matter with Superconducting Nanowire Detectors in an Optical Haloscope. Phys. Rev. Lett. 128, 231802. https://link.aps.org/doi/10.1103/PhysRevLett.128.231802 (23 June 2022).DOI
  4. [4]Dixit, A. V. et al. Searching for Dark Matter with a Superconducting Qubit. en. Physical Review Letters 126, 141302. ISSN: 0031-9007, 1079-7114. https://link.aps.org/doi/10.1103/PhysRevLett.126.141302 (2023) (Apr. 2021).DOI
  5. [5]Riechers, D. A. et al. A dust-obscured massive maximum-starburst galaxy at a redshift of 6.34. en. Nature 496, 329–333 (Apr. 2013).DOI
  6. [6]de Bernardis P et al. A flat Universe from high-resolution maps of the cosmic microwave background radiation. en. Nature 404, 955–959 (Apr. 2000).
  7. [7]Nones, C. et al. High-impedance NbSi TES sensors for studying the cosmic microwave background radiation. Astronomy & Astrophysics 548, A17. ISSN: 0004-6361, 1432-0746. http://www.aanda.org/10.1051/0004-6361/201218834 (2023) (Dec. 2012).DOI
  8. [8]Arrazola, J. M. et al. Quantum circuits with many photons on a programmable nanophotonic chip. en. Nature 591, 54–60 (Mar. 2021).DOI
  9. [9]Todaro, S. L. et al. State Readout of a Trapped Ion Qubit Using a Trap-Integrated Superconducting Photon Detector. en. Physical Review Letters 126, 010501. ISSN: 0031- 9007, 1079-7114. https://link.aps.org/doi/10.1103/PhysRevLett.126.010501 (2023) (Jan. 2021).
  10. [10]Madsen, L. S. et al. Quantum computational advantage with a programmable photonic processor. en. Nature 606, 75–81. ISSN: 0028-0836, 1476-4687. https://www.nature.com/articles/s41586-022-04725-x (2023) (June 2022).DOI
  11. [11]Boaron, A. et al. Secure Quantum Key Distribution over 421 km of Optical Fiber. Phys. Rev. Lett. 121, 190502 (2018).
  12. [12]Wang, S. et al. Twin-field quantum key distribution over 830-km fibre. en. Nat. Photonics 16, 154–161 (Jan. 2022).
  13. [13]Grünenfelder, F. et al. Fast single-photon detectors and real-time key distillation enable high secret-key-rate quantum key distribution systems. en. Nat. Photonics 17, 422–426 (Mar. 2023).DOI
  14. [14]Li, W. et al. High-rate quantum key distribution exceeding 110 Mb s−1. en. Nat. Photonics 17, 416–421 (Mar. 2023).DOI
  15. [15]Walter, A. B. et al. The MKID Exoplanet Camera for Subaru SCExAO. Publications of the Astronomical Society of the Pacific 132, 125005. ISSN: 1538-3873. https://iopscience.iop.org/article/10.1088/1538-3873/abc60f (Nov. 2020).DOI
  16. [16]Gol’tsman, G. N. et al. Picosecond superconducting single-photon optical detector. Applied Physics Letters 79, 705–707. ISSN: 0003-6951. eprint: https://pubs.aip.org/aip/apl/article-pdf/79/6/705/8782545/705_1_online.pdf. https://doi.org/10.1063/1.1388868 (Aug. 2001).
  17. [17]Hadfield, R. H. Single-photon detectors for optical quantum information applications. Nat. Photonics 3, 696–705 (Dec. 2009).
  18. [18]Marsili, F. et al. Detecting single infrared photons with 93 % system efficiency. 7 (2013).
  19. [19]Reddy, D. V., Nerem, R. R., Nam, S. W., Mirin, R. P. & Verma, V. B. Superconducting nanowire single-photon detectors with 98% system detection efficiency at 1550 nm. Optica 7, 1649–1653. https://opg.optica.org/optica/abstract.cfm?URI=optica-7-12-1649 (Dec. 2020).
  20. [20]Korzh, B. et al. Demonstration of sub-3 ps temporal resolution with a superconducting nanowire single-photon detector. Nature Photonics 14, 250–255. https://doi.org/10.1038/s41566-020-0589-x (4 2020).
  21. [21]Wollman, E. E. et al. UV superconducting nanowire single-photon detectors with high efficiency, low noise, and 4 K operating temperature. Opt. Express 25, 26792 (2017).DOI
  22. [22]Verma, V. B. et al. Single-photon detection in the mid-infrared up to 10 μm wavelength using tungsten silicide superconducting nanowire detectors. APL Photonics 6, 056101 (2021).
  23. [24]McCaughan, A. N. et al. The thermally coupled imager: A scalable readout architecture for superconducting nanowire single photon detectors. Applied Physics Letters 121, 102602. eprint: https://doi.org/10.1063/5.0102154. https://doi.org/10.1063/5.0102154 (2022).
  24. [25]Morozov, D. V., Casaburi, A. & Hadfield, R. H. Superconducting photon detectors. Contemporary Physics 62, 69–91. eprint: https://doi.org/10.1080/00107514.2022.2043596. https://doi.org/10.1080/00107514.2022.2043596 (2021).
  25. [26]Day, P. K. P., LeDuc, H. G. H., Mazin, B. A., Vayonakis, A. & Zmuidzinas, J. A broadband superconducting detector suitable for use in large arrays. Nature 425, 817–821. ISSN: 0028-0836. http://www.nature.com/nature/journal/v425/n6960/abs/nature02037.html%20http://dx.doi.org/10.1038/nature02037%20http: //www.nature.com/nature/journal/v425/n6960/full/nature02037.html (Oct. 2003).
  26. [27]Zhao, Q.-Y. et al. Single-photon imager based on a superconducting nanowire delay line. Nature Photonics 11, 247–251. https://doi.org/10.1038/nphoton.2017.35 (4 2017).
  27. [28]Verma, V. B. et al. A four-pixel single-photon pulse-position array fabricated from WSi superconducting nanowire single-photon detectors. Applied Physics Letters 104, 051115. eprint: https://doi.org/10.1063/1.4864075. https://doi.org/10.1063/1.4864075 (2014).
  28. [29]Wollman, E. E. et al. Kilopixel array of superconducting nanowire single-photon detectors. Opt. Express 27, 35279–35289. https://opg.optica.org/oe/abstract.cfm?URI=oe-27-24-35279 (Nov. 2019).
  29. [30]Allmaras, J. P. et al. Demonstration of a Thermally Coupled Row-Column SNSPD Imaging Array. Nano Letters 20. PMID: 32091221, 2163–2168. eprint: https://doi.org/10.1021/acs.nanolett.0c00246. https://doi.org/10.1021/acs.nanolett.0c00246 (2020).
  30. [31]McCaughan, A. N. et al. A superconducting thermal switch with ultrahigh impedance for interfacing superconductors to semiconductors. Nature Electronics 2, 451–456. eprint: https://doi.org/10.1038/s41928-019-0300-8. https://doi.org/10.1038/s41928-019-0300-8 (2019).
  31. [32]Yamashita, T. et al. Origin of intrinsic dark count in superconducting nanowire single-photon detectors. Applied Physics Letters 99. 161105. ISSN: 0003-6951. eprint: https://pubs.aip.org/aip/apl/article-pdf/doi/10.1063/1.3652908/14458078/161105\_1\_online.pdf. https://doi.org/10.1063/1.3652908 (Oct. 2011).

Paper details

vlnc:W_bt7hn6e5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents