Jitter Calibration System and Method for Superconducting Nanowire Single Photon Detector

Abstract

The present invention discloses a jitter calibration system and method for a superconducting nanowire single photon detector, including an FPGA platform for generating a periodic pulse calibration signal, a radio frequency pulse amplifier, a lithium niobate electro-optic modulator and its bias voltage controller, a 1550nm 3kHz linewidth distributed feedback laser, a programmable optical attenuator, a three-paddle polarization controller, an SNSPD single photon detector, a multi-channel TDC module, and a host computer for processing data. The calibration method is simple and has a low cost. It can calibrate the correction curve of the relationship between the additional jitter of the SNSPD detected photons caused by pulse pile-up and the detection recovery time, and compensate for the additional jitter, effectively solving the problem of jitter deterioration of the SNSPD at high counting rates in high-speed detection applications.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless optical communication, and relates to a jitter calibration system and method for a superconducting nanowire single-photon detector. Background Art

[0002] Photon counting communication, as a high-sensitivity and low-power laser communication system, is very suitable for deep space link scenarios. However, with the continuous development of deep space exploration missions, the detection distance and the amount of transmitted data have further increased, and the requirement for the communication rate of the optical communication system is also getting higher and higher. Therefore, superconducting nanowire single-photon detectors (SNSPDs) with advantages such as high detection efficiency (close to 100%), low dark count (<10 cps), small time jitter (tens of picoseconds @ FWHM), and low blocking effect (recovery characteristic time of tens of nanoseconds) have gradually become a research hotspot in deep space optical communication.

[0003] For high-precision time measurement applications based on SNSPDs such as single-photon imaging, single-photon radar, quantum communication, and photon counting communication, the time-domain jitter of detected photons is an important factor limiting the timing resolution accuracy. Especially in high-speed photon counting communication applications, when the overall time-domain jitter distribution of detected photons spills over the slot duration, potential inter-slot count crosstalk will affect the accuracy of information recovery and communication synchronization. Therefore, its jitter performance participates in determining the upper limit of the slot frequency and thus affects the maximum data transmission rate. In a typical detected photon arrival time distribution, it mainly includes the edge jitter and pulse width of the pulse laser source at the transmitting end itself, the jitter of the time-to-digital converter (TDC) at the receiving end for reading the arrival time of photons, and the jitter of the SNSPD at the detection end. In high-speed applications, due to the characteristics of the detector itself, the SNSPD will experience a significant deterioration in the jitter performance of the TDC receiving the arrival time of photons triggered by a single threshold discriminator due to detection pulse pile-up and incomplete recovery of the superconducting current in the photon counting rate close to the saturation range (the additional jitter full width at half maximum usually reaches hundreds of picoseconds @ FWHM), and it will also introduce asymmetry in the photon arrival time distribution. As Figure 1 shown by the SNSPD detection pulse waveform and the current recovery situation represented by blue, the yellow time stamp represents the moment when the photon is responded to by the SNSPD, and the red and black time stamps represent the moments when the TDC is theoretically and actually triggered and reads the arriving photon respectively.

[0004] Currently, there are few solutions to this problem. The existing hardware method that uses a Constant Fraction Discriminator (CFD) instead of a single threshold trigger to read out at the front end of the TDC to overcome pileup jitter is not applicable to the expansion of multi-channel SNSPDs because of its complex structure and high cost, especially for the photon signal readout of large-scale SNSPD arrays. Therefore, there is an urgent need for a simple, low-cost, and effective software method to suppress the jitter deterioration problem of SNSPDs at high counting rates, so as to meet the application requirements of higher-precision time measurement. Summary of the Invention

[0005] The object of the present invention is to provide a jitter calibration system and method for a superconducting nanowire single-photon detector. The calibration method is simple and has a low cost. It can calibrate the correction curve of the relationship between the additional jitter of detected photons by SNSPD caused by pulse pileup and the detection recovery time, and compensate for the additional jitter, effectively solving the problem of jitter deterioration of SNSPDs at high counting rates in high-speed detection applications.

[0006] The technical solution for achieving the object of the present invention is as follows:

[0007] A jitter calibration device for a superconducting nanowire single-photon detector includes an FPGA platform. The FPGA platform is used to generate a periodic pulse calibration signal. By encoding two GT transceivers on the FPGA platform respectively, a first narrow pulse signal and another synchronous calibration signal are generated. The first narrow pulse signal undergoes electro-optic conversion through an optoelectronic conversion unit to generate a narrow pulse laser signal with the same repetition frequency and pulse width and a high modulation extinction ratio, which is used to trigger the detector to respond to the periodic photon-level signal after passing through an optical attenuator and a bias controller. The detection pulse signal generated by a low-noise RF amplifier is received and the timestamp is read out through multiple channels of a multi-channel TDC module. At the same time, the period of the other synchronous calibration signal is 64 times that of the first narrow pulse signal, and it is directly connected to the multi-channel TDC module and the timestamp is read out by an additional channel. The host computer processes the timestamps received by the multi-channel TDC module and generates a correction curve for the additional jitter.

[0008] In a preferred technical solution, the optoelectronic conversion unit includes a RF pulse amplifier, a lithium niobate electro-optic modulator, its bias voltage controller, and a 1550nm linewidth distributed feedback laser. The RF pulse amplifier is connected to one GT transceiver of the FPGA platform. The RF pulse amplifier is also connected to the lithium niobate electro-optic modulator. The lithium niobate electro-optic modulator is connected to the 1550nm linewidth distributed feedback laser and the modulator bias voltage controller.

[0009] In a preferred technical solution, the first narrow pulse signal is a narrow pulse signal with a repetition frequency of 1.25 GHz and a pulse width of 200 ps, and the other synchronous calibration signal is a synchronous calibration signal with a repetition frequency of 9.53125 MHz and a pulse width of 1.5 ns.

[0010] The present invention also discloses a jitter calibration method for a superconducting nanowire single photon detector, which is applied to the above-mentioned jitter calibration device. The jitter calibration method includes the following steps:

[0011] S01: Using a pulse signal with a transmission pulse width of 1.5 ns and a period of 51.2 ns of the GT transceiver on the FPGA platform as a synchronous calibration signal for the laser pulse, and collecting time stamps on the multi-channel TDC module with a trigger threshold of 460 mV;

[0012] S02: Interpolating 63 data points at equal intervals between adjacent time stamps of the synchronous calibration signal as a 1.25 GHz synchronous calibration signal with the same frequency as the laser pulse;

[0013] S03: There is only a fixed delay ΔT0 between the interpolated synchronous calibration signal and the periodic photon pulse signal. The arrival time of the SNSPD detection pulse triggered synchronously by using the interpolated synchronous calibration signal as the start signal is ΔT0 + ΔT', where ΔT' is caused by the additional jitter of pulse pile-up. At the same time, calculate the recovery time of each SNSPD detection pulse time stamp read out by the multi-channel TDC module relative to its previous time stamp as the recovery time T of this detection rec ;

[0014] S04: Increase the step size of T rec For the detection pulse time stamps in multiple recovery time intervals, generate histograms of the arrival time ΔT0 + ΔT' respectively, and perform Gaussian fitting on the time stamp distribution in each histogram to obtain the distribution mean μ and full width at half maximum of its arrival time. Take (μ0 - μ) as the ordinate, where μ0 is the stable value of μ, and T rec as the abscissa to obtain the calibration curve.

[0015] In a preferred technical solution, when generating the histogram in step S04, merge the time stamps with T rec ≤ 6.4 ns to generate a sub-histogram, and merge all the time stamps with T rec ≥ 100 ns to generate a sub-histogram.

[0016] In a preferred technical solution, calibrate the calibration curve according to the trigger threshold with the maximum average rising edge voltage slew rate at a counting rate of at least dozens of M cps / channel.

[0017] In the preferred technical solution, after obtaining the calibration curve, it further includes dynamically and real-time compensating the detection pulse timestamps collected under different light intensities based on the first-order time walk correction parameters of the lookup table.

[0018] In the preferred technical solution, the compensation method includes, for a photon timestamp collected by a multi-channel TDC module, finding the corresponding compensation time on the vertical axis according to its recovery time and the calibration curve of this channel of the detector, and then adding it to the timestamp of the multi-channel TDC module as the compensated timestamp.

[0019] The present invention further discloses a jitter calibration system for a superconducting nanowire single photon detector, which is applied to the above-mentioned jitter calibration device. The jitter calibration system includes:

[0020] A timestamp acquisition module, which uses a pulse signal with a transmission pulse width of 1.5 ns and a period of 51.2 ns of the GT transceiver on the FPGA platform as the synchronous calibration signal of the laser pulse to acquire timestamps on the multi-channel TDC module with a trigger threshold of 460 mV;

[0021] A synchronous calibration signal interpolation module, which equally inserts 63 data points at equal intervals between adjacent timestamps of the synchronous calibration signal as a 1.25 GHz synchronous calibration signal with the same frequency as the laser pulse;

[0022] A calculation module, there is only a fixed delay ΔT0 between the interpolated synchronous calibration signal and the periodic photon pulse signal. The arrival time of the SNSPD detection pulse triggered synchronously by using the interpolated synchronous calibration signal as the start signal is ΔT0 + ΔT', where ΔT' is caused by the additional jitter of pulse pile-up. At the same time, calculate the recovery time of each SNSPD detection pulse timestamp read by the multi-channel TDC module relative to its previous timestamp as the recovery time T of this detection rec ;

[0023] A calibration curve generation module, increasing the step size of T rec , generating histograms of the arrival times ΔT0 + ΔT' of the detection pulse timestamps in multiple recovery time intervals respectively, and performing Gaussian fitting on the timestamp distribution in each histogram to obtain the distribution mean μ and the full width at half maximum of its arrival time. Taking (μ0 - μ) as the ordinate, where μ0 is the stable value of μ, and T rec as the abscissa, to obtain the calibration curve.

[0024] Compared with the prior art, the present invention has the following remarkable advantages:

[0025] 1. There is no need to add an additional CFD before the TDC in the SNSPD detection signal readout part, but only a simple single-threshold comparator is required, which significantly saves the hardware cost and facilitates the expansion of multi-channel SNSPD.

[0026] 2. For the determined TDC comparator threshold, the calibrated jitter correction curve only requires the time interval between two adjacent photon detections of the same SNSPD as the input. This low-complexity correction algorithm is easy to implement in hardware platforms such as FPGAs in the form of look-up tables, and has good real-time potential.

[0027] 3. For the same SNSPD and the determined comparator threshold, the correction curve calibrated at a certain counting rate is also applicable to the working scenarios of other photon counting rates, which can ensure the robustness of the jitter optimization method under different signal photon intensities. Brief Description of the Drawings

[0028] Figure 1 Schematic diagram of the additional time jitter of photons triggered by a single TDC threshold caused by pulse pile-up at high counting rates of SNSPD;

[0029] Figure 2 Principle block diagram of the jitter calibration device of the superconducting nanowire single photon detector in this embodiment;

[0030] Figure 3 Schematic diagram of the original waveforms of two signals used for additional jitter calibration on the FPGA in this embodiment;

[0031] Figure 4 Interpolating 63 points within a 19.53125 MHz periodic signal as a 1.25 GHz SYNC signal;

[0032] Figure 5 SYNC and periodic laser pulses used for calibration after interpolation;

[0033] Figure 6 Schematic diagram of the generated histogram and Gaussian fitting;

[0034] Figure 7 Additional jitter correction curve of each pixel at a 50 mV TDC trigger threshold and a unit 25 Mcps count;

[0035] Figure 8 Comparison of the photon response timestamp distributions of the 1.25 GHz SYNC signal triggered by a single SNSPD before and after compensating for additional jitter at a counting rate of approximately 25 Mcps. Detailed Implementation Manner

[0036] The principle of the present invention is: to provide a correction curve for calibrating the relationship between the additional jitter of photons detected by SNSPD caused by pulse pile-up and the detection recovery time, and to compensate for the additional jitter. It can solve the problem of deteriorated jitter generated by SNSPD at high counting rates in high-speed detection applications.

[0037] Glossary of Terms:

[0038] 1. SNSPD (Superconducting Nanowire Single Photon Detector), superconducting nanowire single photon detector;

[0039] 2. TDC (Time - digital Converter), time - digital converter;

[0040] 3. FPGA (Field Programmable Gate Array), field programmable gate array;

[0041] 4. GT (Gigabyte Transceiver), gigabyte transceiver;

[0042] 5. FWHM (Full Width at Half Maxima), full width at half maxima;

[0043] 6. cps (Counts per second), counts per second;

[0044] 7. CFD (Constant Fraction Discriminator), constant fraction discriminator.

[0045] Example 1:

[0046] As Figure 2 shown, a jitter calibration device for a superconducting nanowire single photon detector includes an FPGA platform for generating a periodic pulse calibration signal, a radio frequency pulse amplifier, a lithium niobate electro - optic modulator and its bias voltage controller, a 1550 nm 3 kHz linewidth distributed feedback laser, a programmable optical attenuator, a three - paddle polarization controller, an SNSPD single photon detector, a multi - channel TDC module, and a host computer for processing data.

[0047] First, at the transmitting end, by encoding two GT transceivers on the FPGA platform respectively, a narrow pulse signal with a repetition frequency of 1.25 GHz and a pulse width of 200 ps and another synchronous calibration signal (SYNC) with a repetition frequency of 19.53125 MHz and a pulse width of 1.5 ns are generated. The first signal is used for electro-optic conversion through a radio frequency pulse amplifier, a lithium niobate electro-optic modulator, its bias voltage controller, and a 1550 nm 3 kHz linewidth distributed feedback laser to generate a narrow pulse laser signal with a high modulation extinction ratio of the same repetition frequency and pulse width. After appropriate optical attenuation and bias controller, it is connected to a 16-pixel SNSPD system to trigger the detector's response to the periodic photon-level signal. The detection pulse signal in the order of hundreds of mV generated by the readout electronics system supporting the detector is received by 16 channels of a multi-channel TDC module triggered by a single comparator threshold and the timestamps are read out. At the same time, the SYNC signal generated by the other GT transceiver has a period 64 times that of the previous signal, which is directly connected to the multi-channel TDC module and an additional timestamp is read out. The host computer processes the timestamps received by the multi-channel TDC module and generates a correction curve with additional jitter according to the method described below.

[0048] Among them, the radio frequency pulse amplifier, the lithium niobate electro-optic modulator, its bias voltage controller, and the 1550 nm linewidth distributed feedback laser as the optoelectronic conversion unit are a preferred embodiment. Of course, it can also be an optoelectronic conversion unit with other structures, which is not limited here. The radio frequency pulse amplifier of the optoelectronic conversion unit is connected to one GT transceiver of the FPGA platform. The radio frequency pulse amplifier is also connected to the lithium niobate electro-optic modulator. The lithium niobate electro-optic modulator is connected to the 1550 nm linewidth distributed feedback laser and the modulator bias voltage controller.

[0049] Among them, the SNSPD system includes a 16-pixel niobium nitride SNSPD array, its supporting readout electronics, and a cryogenic refrigeration system.

[0050] Among them, the readout electronics includes a DC bias system. The DC bias system is connected to the 16-pixel niobium nitride SNSPD array and a low-noise radio frequency amplifier through a bias tee. The 16-pixel niobium nitride SNSPD array is connected to a quasi-constant voltage bias circuit. The 16-pixel niobium nitride SNSPD array is connected to the low-noise radio frequency amplifier after passing through a 2.4K cryogenic refrigeration and vacuum system.

[0051] Preferably, there are preferably two programmable optical attenuators.

[0052] Figures 3 - 5 The original simulation waveforms of the two signals used in the calibration process of the correction curve in the FPGA, as well as the waveforms of the SYNC electrical pulse and the optical pulse triggering the SNSPD collected by a high-bandwidth oscilloscope, are given.

[0053] The degree of obvious jitter deterioration of SNSPD at high counting rates (generally dozens of mega counts per second (Mcps)) is closely related to the recovery time of the detector pixels (the time interval between two adjacent photon detections). Especially in the range of several nanoseconds to dozens of nanoseconds, the additional jitter will vary violently with the size of the recovery time. Therefore, the purpose of selecting laser pulses with extremely short periods to trigger SNSPD is to make the subsequent calibrated recovery time step small enough (0.8 ns) to accurately depict the calibration curve, that is, except for the noise counts, the adjacent detection events must be the responses to two laser pulses with an interval that is an integer multiple of 800 ps. At the same time, we use another TXP2 of the GT transceiver on the transmitting end FPGA to send a pulse signal with a pulse width of 1.5 ns and a period of 51.2 ns as the SYNC signal of the laser pulse to collect time stamps at a trigger threshold of 460 mV on the multi-channel TDC module, so as to adapt to the minimum signal pulse width input of 350 ps of the time-to-digital converter TDC and the readout dead time requirement of 1.5 ns. Then, 63 data points are equally interpolated between adjacent time stamps of the SYNC signal as the 1.25 GHz SYNC signal with the same frequency as the laser pulse, as Figure 4 shown. If the influence of the photon arrival time distribution within the 200 ps broadening range of the laser pulse is not considered, then we can consider that there is only a fixed delay ΔT0 between the interpolated SYNC signal and the periodic photon pulse signal, as Figure 5 shown. The arrival time of the SNSPD detection pulse triggered by synchronizing with the interpolated SYNC as the start signal is ΔT0 + ΔT' (ΔT' is caused by the additional jitter of pulse pile-up). At the same time, calculate the recovery time of each SNSPD detection pulse time stamp read by the TDC relative to its previous time stamp (the interval corresponding to the two black actualtriggers in Figure 1 ) and approximately use it as the recovery time T of this detection rec . The basis for this approximation is that the additional time jitter caused by pile-up is generally in the range of hundreds of picoseconds, and the change in the additional jitter compensation value caused by the change of the recovery time by hundreds of picoseconds on the calibration curve can be ignored.

[0054] Subsequently, 800 ps can be used as the minimum step size for the increase of T rec . For each recovery time interval of T rec ∈(6.4, 7.2), (7.2, 8.0), …, (99.2, 100.0) ns, histograms of the arrival time ΔT0 + ΔT' of the detection pulse time stamps are generated respectively, and the Gaussian fitting is performed on the time stamp distribution within each histogram to obtain the distribution mean μ and the full width at half maximum FWHM of its arrival time, as Figure 6 shown. T recThe reason for the lower limit of the statistical interval is that at recovery times below 6.4 ns, the number of probe pulse timestamps in the 800 - ps recovery - time sub - interval is too small to effectively perform a Gaussian fit on the arrival - time distribution of the timestamps within the sub - interval. Therefore, timestamps with T rec ≤6.4 ns are combined to generate a single sub - histogram; the reason for the upper limit is that the impact of pulse - pileup jitter corresponding to recovery times above 100 ns basically disappears. Therefore, timestamps with T rec ≥100 ns are also all combined to generate a single sub - histogram.

[0055] From Figure 6 it can be seen that as T rec increases, the mean μ of the Gaussian fit also continuously shifts to the right and stabilizes at the value μ0 as the impact of pulse - pileup disappears. Then, the offset of the mean μ of the sub - histogram at a smaller T rec relative to μ0 can be considered as the deterministic part of the additional jitter that can be compensated; while the FWHM (Full Width at Half Maximum) of the fit deteriorates at smaller T rec values. This is due to the decrease in the pulse amplitude caused by insufficient nanowire current resulting from too short a probe - pulse interval (as shown by the second SNSPD probe pulse in Figure 1 ), which reduces the voltage slew rate at the TDC threshold trigger, and will introduce more severe jitter under the same voltage noise. This part is the non - deterministic part of the additional jitter that cannot be compensated. After obtaining μ0, the mean μ obtained from the Gaussian fit of the sub - histogram at different T rec values minus μ0 is used as the ordinate, and T rec (here with a step of 800 ps) is used as the abscissa to obtain the calibration curve.

[0056] According to the above method, Figure 7 the additional - jitter calibration curve of (μ0 - μ) versus T rec is given for all pixels of the SNSPD array in a working scenario with a single - channel count rate of approximately 25 Mcps as an example. The additional jitter below 20 ns is mainly limited by the combined effect of pulse - fall - edge pileup and amplitude reduction caused by incomplete recovery of the nanowire current, and mainly shows an asymmetric forward tail in the detection - pulse time - jitter distribution at high count rates; while the additional jitter around 22 ns is due to the combined effect of negative overshoot at the falling edge and amplitude reduction caused by incomplete recovery of the nanowire current, which is the main cause of the backward tail. As T rec continues to increase beyond the time - interval range of the crosstalk between the front and back detection pulses, the impact of the additional jitter completely disappears.

[0057] The calibration of the correction curve theoretically only relates to the SNSPD parameters, the inherent characteristics of the readout circuit, and the pulse edge trigger threshold used by the TDC. The correction curve calibrated at a certain counting rate is applicable to the additional jitter compensation in different photon counting rate scenarios. However, it is best to calibrate the correction curve according to the trigger threshold with the maximum average rising edge voltage slew rate near the counting rate level in the actual detector working scenario, and this counting rate should be at least dozens of Mcps / channel, so that Figure 6 the smaller T in rec the corresponding number of sub-histogram samples increases, so that it also contains enough samples for Gaussian fitting and generates an accurate estimate of the mean μ. For high-speed real-time applications, it is only necessary to calculate the difference between adjacent timestamps on the same channel (the recovery time experienced by the detector of this channel before this detection), and then, under the condition that the trigger threshold and detector parameters are fixed, the detection pulse timestamps collected under different light intensities can be dynamically and real-time compensated through the first-order time walk correction parameters based on the look-up table (that is, for a photon timestamp collected by a multi-channel TDC module, according to its recovery time and the correction curve of this channel of the detector, find the corresponding compensation time on the vertical axis, and then add it to the timestamp of the multi-channel TDC module as the compensated timestamp, and a look-up table structure can be adopted). Therefore, it has great engineering application value.

[0058] Figure 8 The time-domain distribution histograms of SNSPD detected photons before and after applying jitter compensation at a relatively high counting rate of 25 Mcps for a single channel are given (shown in red and blue respectively, and the photon timestamp count in the inset is on a logarithmic axis).

[0059] The above embodiments are the preferred embodiments of the present invention, but the embodiments of the present invention are not limited by the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement methods and are all included in the protection scope of the present invention.

Claims

1. A jitter calibration device for a superconducting nanowire single-photon detector, characterized in that: The invention comprises an FPGA platform, wherein the FPGA platform is used to generate a periodic pulse calibration signal. The two GT transceivers on the FPGA platform are respectively encoded to generate a first narrow pulse signal and another synchronous calibration signal. The first narrow pulse signal is subjected to electro-optical conversion by a photoelectric conversion unit to generate a high-modulation extinction ratio narrow pulse laser signal with the same repetition frequency and pulse width as the first narrow pulse signal, which is connected to an SNSPD system after passing through an optical attenuation and bias controller to trigger the detector to respond to a periodic photon-level signal. The detection pulse signal generated by a low-noise radio frequency amplifier is received via multiple channels of a multi-channel time-to-digital converter module and a timestamp is read out. Meanwhile, the other synchronous calibration signal has a period of 64 times that of the first narrow pulse signal, is directly connected to the multi-channel time-to-digital converter module, and a timestamp is read out by an additional channel. The host computer processes the timestamps of each channel received by the multi-channel time-to-digital converter module and generates a correction curve for additional jitter.

2. The jitter calibration device for superconducting nanowire single-photon detector according to claim 1, characterized in that: The photoelectric conversion unit includes a radio frequency pulse amplifier, a lithium niobate electro-optic modulator and its bias voltage controller, and a 1550nm narrow linewidth distributed feedback laser. The radio frequency pulse amplifier is connected to a GT transceiver of the FPGA platform. The radio frequency pulse amplifier is also connected to the lithium niobate electro-optic modulator. The lithium niobate electro-optic modulator is connected to the 1550nm narrow linewidth distributed feedback laser and the modulator bias voltage controller.

3. The jitter calibration device for superconducting nanowire single-photon detector according to claim 1, characterized in that: The first narrow pulse signal is a narrow pulse signal with a repetition rate of 1.25 GHz and a pulse width of 200 ps, ​​and the other synchronous calibration signal is a synchronous calibration signal with a repetition rate of 9.53125 MHz and a pulse width of 1.5 ns.

4. A jitter calibration method for a superconducting nanowire single-photon detector, characterized in that: Applied to the jitter calibration device according to any one of claims 1 to 3, the jitter calibration method comprises the following steps: S01: Use the pulse signal with a pulse width of 1.5ns and a period of 51.2ns sent by the GT transceiver on the FPGA platform as the synchronous calibration signal of the laser pulse to collect the timestamp with a trigger threshold of 460mV on the multi-channel time-to-digital converter module; S02: 63 data points are interpolated at equal intervals between adjacent timestamps of the synchronous calibration signal as a 1.25 GHz synchronous calibration signal with the same frequency as the laser pulse; S03: There is only a fixed delay ΔT0 between the interpolated synchronous calibration signal and the periodic photon-level signal. The arrival time of the SNSPD detection pulse synchronously triggered by the interpolated synchronous calibration signal as the start signal is ΔT0+ΔT', where ΔT' is caused by the additional jitter of pulse accumulation. At the same time, the recovery time of each SNSPD detection pulse timestamp read out by the multi-channel time-to-digital converter module relative to its previous timestamp is calculated as the recovery time T of this detection. rec ; S04: Increase T rec The step size is , and the histograms of the arrival time ΔT0+ΔT' are generated for the detection pulse timestamps in multiple recovery time intervals. The Gaussian fitting is performed on the timestamp distribution in each histogram to obtain the distribution mean μ and half-maximum full width of its arrival time. (μ0-μ) is used as the ordinate, μ0 is the stable value of μ, T rec As the abscissa, a calibration curve is obtained.

5. The jitter calibration method of the superconducting nanowire single-photon detector according to claim 4, characterized in that: When the histogram is generated in step S04, T rec The timestamps ≤6.4ns are merged to generate a sub-histogram. rec All timestamps ≥100ns are merged to generate a sub-histogram.

6. The jitter calibration method of the superconducting nanowire single-photon detector according to claim 4, characterized in that: When the count rate is at least tens of megacps / channel, the correction curve is calibrated according to the trigger threshold with the maximum average rising edge voltage slew rate.

7. The jitter calibration method of the superconducting nanowire single-photon detector according to claim 4, characterized in that: After the correction curve is obtained, the method also includes dynamically and real-timely compensating the detection pulse time stamps collected under different light intensities based on the first-order time walk correction parameters of the lookup table.

8. The jitter calibration method of the superconducting nanowire single-photon detector according to claim 7, characterized in that: The compensation method includes, for a photon timestamp collected by a multi-channel time-to-digital converter module, finding the corresponding compensation time on the vertical axis according to its recovery time and the calibration curve of the corresponding channel of the detector, and then adding it to the timestamp of the multi-channel time-to-digital converter module as the compensated timestamp.

9. A jitter calibration system for a superconducting nanowire single-photon detector, characterized in that: Applicable to the jitter calibration device according to any one of claims 1 to 3, the jitter calibration system comprises: The timestamp acquisition module uses the pulse signal with a pulse width of 1.5ns and a period of 51.2ns sent by the GT transceiver on the FPGA platform as the synchronous calibration signal of the laser pulse to acquire the timestamp with a trigger threshold of 460mV on the multi-channel time-to-digital converter module; The synchronous calibration signal interpolation module interpolates 63 data points of adjacent time stamps of the synchronous calibration signal at equal intervals as a 1.25 GHz synchronous calibration signal with the same frequency as the laser pulse; In the calculation module, there is only a fixed delay ΔT0 between the interpolated synchronous calibration signal and the periodic photon-level signal. The arrival time of the SNSPD detection pulse synchronously triggered by the interpolated synchronous calibration signal as the start signal is ΔT0+ΔT', where ΔT' is caused by the additional jitter of pulse accumulation. At the same time, the recovery time of each SNSPD detection pulse timestamp read out by the multi-channel time-to-digital converter module relative to its previous timestamp is calculated as the recovery time T of this detection. rec ; Calibration curve generation module, add T rec The step size is , and the histograms of the arrival time ΔT0+ΔT' are generated for the detection pulse timestamps in multiple recovery time intervals. The Gaussian fitting is performed on the timestamp distribution in each histogram to obtain the distribution mean μ and half-maximum full width of its arrival time. (μ0-μ) is used as the ordinate, μ0 is the stable value of μ, T rec As the abscissa, a calibration curve is obtained.

Images