The system design diagram of the entire NFAD readout circuits and connections. All the components in the right panel are inside a deep freezer. The ones in the left panel are at room temperature. Subminiature version A (SMA) is the high frequency connector; fiber connector physical connection (FC/PC) is the single mode fiber connector; Hirose is the NFAD high voltage cable connector.
Testing setup for characterization of the NFAD using the time-correlated single-photon counting method. Details in the text.
(a) The experimental single-photon response recorded with a fast digital oscilloscope after the 40 dB gain amplification. (b) The SPICE model for the NFAD read out circuit. The dash line square highlights the NFAD equivalent circuit components; the testing points TPv1 and TPdv1 represent photon detection response output and intrinsic SPAD voltage drop, respectively; the voltage source VBR represents the breakdown voltage of the NFAD. (c) and (d) Results of the transient simulation of a single-photon responses, before the amplifier. (c) The simulated response for the readout circuits used in this work; (d) The simulated response with the conventional readout methods using a bias tee in Ref. 1. The model parameters for the bias tee in the the conventional method are simplified from the ones provided by its manufacturer.
Measurement of the darkcount rates for the two NFAD devices. Both are measured at 193 K with the same discriminator voltage settings.
DE vs darkcount measurements results for two NFAD detectors. The discriminator threshold voltage level is set to 100 mV, and the testing temperature is 193 K.
Verification of bucket detector model. The hollow squares are the experimental results. The solid lines are the simulation results based on modeling the detector as a bucket detector.
Effect of temperature on detection efficiency and darkcounts. The data points represent experimental measurements, and the lines are simple smooth fits.
Experimental setups used for testing the two NFAD 1 using SPDC. In dashed box (a), two free-running NFAD detectors. In dashed box (b), one NFAD is placed in one of photon arms, a gated commercial detector (IDQ) is at the other arm. Photon detections from the NFAD are used to trigger an id-Quantique commercial single-photon detector.
(a) The inferred detection efficiency of the NFAD, from a direct comparison with an id-201 commercial detector. The solid line (hollow circle) is the inferred DE using the SPDC scheme at 1530 nm; the dash line (cross circle) is the inferred DE corrected for stray light detection; the solid line (square) is the DE characterized by the WCP scheme at 1550 nm. (b) The darkcount observed in the presence (“apparent”) and absence (“intrinsic”) of stray light coupled into the optical fiber connecting to the detector when varying the bias voltage.
(a) Accumulated measurements of the transient detector response; this plot is the screen capture from a fast oscilloscope collected for 6 h; the saturated amplitude is about 300 mV. (b) The simulation for the transient voltage drop across an intrinsic SPAD inside the NFAD. Both plots share the same time scale of 5 . (c) Measurement of the detector deadtime and recovery time. The points are the measured photon detections in a 1 ns time resolution, and the solid line is the guide to eye. (d) Afterpulsing decay measurement and fitting. The random release of the trapped charges induced by photon detection events results in the afterpulsing probability decaying exponentially. The simulation curve is the solid line and the scattered dots are the accumulated measurement results for two temperatures.
(a) Afterpulsing under two threshold levels of the time tag unit input. The temperature is at 193 K. (b) Afterpulsing probability density at three distinct temperatures, while the darkcount rate was kept constant at about 100 CPS. All data are collected for 1000 s. The threshold was set to be 0.1 V. The reference clock rate, used to synchronize the time frame, is set to 10 kHz.
Afterpulsing parameter, κ, measurement at 193 K by WCP characterization setup for both two NFADs. (a) The afterpulsing coefficient as a function of the darkcount rate. (b) The afterpulsing coefficient as a function of the detection efficiency. The wavelength of probing WCP pulses is 1550 nm for all measurements.
Measurement of the detector jitter by Becker Hickl SPC-130 time-correlation card with a time resolution of 1.22 ps. (a) Experimental setup. (b) Measurements plots in logarithmic scale. (c) Normalized measurement plots in linear scale. The data are taken with 0.1 (solid lines) and 1.0 (dash line) photons per pulse for 1000 s, with different combinations of settings to the constant-fraction discriminator (CFD), the start signal for the time measurement. The function generator (AFG 3252), which produces the stop signal, is connected to the “Synch” input of the card. The detector output is connected to CFD input after an pulse inverter transformer. The average photon number per pulse, from 0.1 to 1, is controlled by changing the variable attenuator (VOA) settings.
Experimental setup for the investigation of high loss QKD. Photon pairs are produced in a PPLN waveguide, and are sent through a variable attenuator. They are then split up by a dichroic mirror (DM) and detected by the NFADs. PBS: polarizing beam splitter; attenuator: programmable optical attenuator; HWP: half wave plate; M: mirror; Col.: collimator.
The accumulated counting results collected by a 16-channel time tag unit with a time resolution of 156 ps, at three levels of attenuation settings, for (a) 62.4 dB and 24 h collection time; (b) 66.4 dB and 20 h collection time; (c) 70.4 dB and 24 h collection time. Solid blue lines are Gaussian distribution fittings, with each of them having σ= 0.25 ns, which is dominated by the timing resolution of the time tag unit. (d)–(f) Experimental results (left) in comparison with the simulation results (right). In each of them, all above three levels of channel loss are taken into account. (d) Accidentals-SNR: the ratio between the correlated and accidental counts. (e) The total correlated counts. (f) The total accidental background counts.
The estimated secret key rate for entanglement QKD versus the total channel loss. In the numerical simulation: for the cases of κ = 2, we use detector efficiency of 8% and 6%, respectively; and for the cases of κ = 1, two detector efficiencies are chosen to be 4% and 3%, respectively. The solid squares are based on actual measurements of the signal to noise for a given loss. The lines show secret key rates calculated for different source rates and afterpulsing coefficient. Two additional x axes show corresponding QKD distances according to the total channel loss, for standard SMF-28 single mode fiber and ultra-low loss fiber, SMF-28 ULL. These results indicate that entanglement-based QKD over 400 km is feasible.
Timing jitter measurement42 at 10 KHz repetition rate of pulse laser, 0.1 photon, and 1.0 photon on average per pulse. The measurements are accumulated for 1000 s.
Article metrics loading...
Full text loading...