Table of Contents
Fetching ...

Controlling the recovery time of the superconducting nanowire single-photon detector with a voltage-controlled cryogenic tunable resistor

Hui Wang, Nikita Dmitrievic Orlov, Niels Noordzij, Thomas Descamps, J. W. Niels Los, Val Zwiller, Iman Esmaeil Zadeh

Abstract

Superconducting nanowire single-photon detectors (SNSPD), owing to their unique performance, are currently the standard detector in most demanding single-photon experiments. One important metric for any single-photon detector is the deadtime (or recovery time), defined as the minimum temporal separation between consecutive detection events. In SNSPDs, the recovery time is more subtle, as the detection efficiency does not abruptly drop to zero when the temporal separation between detection events gets smaller, instead, it increases gradually as the SNSPD current recovers. SNSPD's recovery time is dominated by its kinetic inductance, the readout impedance, and the degree of saturation of internal efficiency. Decreasing the kinetic inductance or increasing the readout impedance can accelerate the recovery process. Significant reduction of the SNSPD recovery time, by, for example, adding a series resistor in the readout circuitry, is possible but can lead to detector latching which hinders further detector operation or enforces underbiasing and hence a reduction in detection efficiency. Previous research has demonstrated passive resistive networks for the reduction of recovery time that rely on trial and error to find the appropriate resistance values. Here we show, using a novel, cryogenically compatible, and tunable resistor technology, one can find the optimized impedance values, delivering fast SNSPD recovery time, while maintaining maximum internal detection efficiency. Here we show an increase of more than 2 folds in both maximum achievable detection rates and the achievable detection efficiency at high photon fluxes, demonstrating detection rates as high as 120 Mcps with no loss of internal detection efficiency.

Controlling the recovery time of the superconducting nanowire single-photon detector with a voltage-controlled cryogenic tunable resistor

Abstract

Superconducting nanowire single-photon detectors (SNSPD), owing to their unique performance, are currently the standard detector in most demanding single-photon experiments. One important metric for any single-photon detector is the deadtime (or recovery time), defined as the minimum temporal separation between consecutive detection events. In SNSPDs, the recovery time is more subtle, as the detection efficiency does not abruptly drop to zero when the temporal separation between detection events gets smaller, instead, it increases gradually as the SNSPD current recovers. SNSPD's recovery time is dominated by its kinetic inductance, the readout impedance, and the degree of saturation of internal efficiency. Decreasing the kinetic inductance or increasing the readout impedance can accelerate the recovery process. Significant reduction of the SNSPD recovery time, by, for example, adding a series resistor in the readout circuitry, is possible but can lead to detector latching which hinders further detector operation or enforces underbiasing and hence a reduction in detection efficiency. Previous research has demonstrated passive resistive networks for the reduction of recovery time that rely on trial and error to find the appropriate resistance values. Here we show, using a novel, cryogenically compatible, and tunable resistor technology, one can find the optimized impedance values, delivering fast SNSPD recovery time, while maintaining maximum internal detection efficiency. Here we show an increase of more than 2 folds in both maximum achievable detection rates and the achievable detection efficiency at high photon fluxes, demonstrating detection rates as high as 120 Mcps with no loss of internal detection efficiency.

Paper Structure

This paper contains 4 sections, 4 figures.

Figures (4)

  • Figure 1: (a) Schematic representation of electrical equivalent model of a SNSPD with a resistor $R_\mathrm{s}$ in series. (b) Simulated output voltage profile with $R_\mathrm{s}=$0Ω, 150Ω and 350Ω, respectively. (c) The evolution of the temperature profile of the superconducting nanowire with $R_\mathrm{s}=$0Ω, 150Ω and 350Ω, respectively. Details of the simulation parameters can be found in supplementary material section II.
  • Figure 2: (a) Illustration of a cryogenic tunable resistor $R_\mathrm{s}$ in series with a SNSPD and its electronic readout circuit. (b) Representative optical microscopy image of a cryogenic tunable resistor. (c) Characterization of the channel resistance in a cryogenic tunable resistor as a function of the channel bias current and the heater current. The channel has a width of 1 and the heater has a width of 100.
  • Figure 3: (a) Amplified output waveforms of the SNSPD (curves are averaged over 100 sweeps). (b) Recovery of the detection probability after one detection event with various $I_\mathrm{h}$ for the wavelength of 1548. The filled dots represent the experimental data, and the solid ones are the predicted efficiency recovery curves derived from internal detection efficiency curves and output pulses. (c) Normalized internal detection efficiency curves with various heater currents $I_\mathrm{h}$, obtained for the wavelength of 1548. (d) Comparison of the output pulse recovery times width $\tau_\mathrm{out}$ (measured at $1/e$ of the pulse amplitude), the experimental recovery time $\tau_\mathrm{rec,exp}$, and the derived recovery time $\tau_\mathrm{rec,der}$. $\tau_\mathrm{rec,exp}$ and $\tau_\mathrm{rec,der}$ are obtained by fitting the data with the sigmoid function $y=A(1+\mathrm{erf}(S(x-x_\mathrm{0})))$.
  • Figure 4: Normalized detection efficiency as a function of detection count rates at various heater currents $I_\mathrm{h}$ with a repetition rate of (a) $f_\mathrm{rep}=$50 and (b) $f_\mathrm{rep}=$120. For each laser repetition rate, the best detector performance can be achieved by tuning the heater current. Measurements at $I_\mathrm{h}=$64, which consistently delivers near-optimal or optimal performances across a broad range of count rates, are highlighted with triangles for a better display.