Consistency test of PMT SPE spectrum from dark-noise pulses and LED low-intensity light

  • Yu-ting Wei
  • Meng-yun Guan
  • Wei-xing Xiong
  • You-yu Gan
  • Chang-gen Yang
Original Paper


Photomultiplier tube (PMT) has been an essential instrument for low background experiments, and the accuracy of single photoelectron (SPE) calibration will directly affect the energy reconstruction of the detector. The SPE spectrum can be acquired from the charge distribution of dark-noise pulses or the low-intensity light signal triggered by LED. However, there are arguments on whether the two methods are consistent. In this paper, the Hamamatsu R5912-20mod PMT and the EMI-D642KNFL PMT have been tested adopting the two methods with signal waveform recording electronics. As comparison, traditional charge-to-digital conversion system is also used to test SPE spectrum. The results show that they are consistent when the PMTs are working at typical recommended high voltage (HV).


PMT SPE Dark noise LED Waveform 


PMT has been an essential instrument for experimental particle physics, especially for the low background experiments such as DarkSide [1] and XENON [2], which are all expecting the least material near the target volume. The accuracy of SPE calibration of PMT is essential to energy reconstruction, which will directly affect the results of the experiment. The commonly adopted method, measuring the charge distribution of LED-triggered signal [3], will introduce additional equipments in the sensitive area of the detector. Using the SPE spectrum of dark noise to calibrate the PMT needs no extra equipment which is better, especially for the dark matter direct searching experiments based on noble gases. According to Refs. [4, 5], the average SPE position and distribution spread have obvious differences.

According to Hamamatsu handbook [6], if the PMTs operate at best signal-to-noise ratio state, the dark noise is dominated by thermionic emission. In principle, when PMTs are working at typical recommended high voltage (HV), the dark noise of PMTs is mainly generated by thermionic electrons emitted from either the cathode or dynodes. The LED-triggered signals are generated either from the incident photon on the cathode or from noise, and the photon can be turned into electron on cathode according to quantum efficiency. So the calibration electrons, which are generated either from the incident photons or from the thermal electrons, can be amplified into electronic signal. Signal originally from one electron is the mainly component of SPE. Therefore, the two calibration methods should be consistent under the recommended PMT working state.
Fig. 1

Schematic plots of two SPE measuring methods. Left: method of measuring dark-noise charge distribution. Right: method of measuring LED-triggered charge spectrum. Dashed line shows the system using waveform recording electronics. Solid line shows the comparison test of QDC V965 and oscilloscope systems

Fig. 2

Here shows a typical waveform of a SPE signal of Hamamatsu PMT which is tested at 1500 V HV with LED trigger method when the gain is 4.09e8 and the mean of amplitude spectrum is 0.5 V

Fig. 3

Comparison of the SPE spectrum using two methods. Left is about Hamamatsu PMT tested at 1400 V and right is about EMI PMT tested at 1600 V. Blue line stands for the spectrum of LED-triggered SPE spectrum. Red one stands for the spectrum of dark noise (colour figure online)

The objective of this study is to test the consistency of two calibration methods using new electronic technology by obtaining the signal waveform and using the traditional charge-to-digital conversion (QDC) data acquisition system in comparison. In this article, the SPE spectrum evaluation tests of Hamamatsu R5912-20mod PMT and EMI D642KBFL PMT which is the same type of PMT adopted by ref. [4] are described with two methods namely dark-noise charge distribution and LED-triggered charge spectrum. In “Experimental setups and testing procedure” section, the general experimental facilities and testing procedure are introduced. The analytic method and derived PMT parameters are shown in “Waveform-recording date analysis and results” and “Comparison test of QDC and oscilloscope” sections. “Conclusions” section contains the conclusions of this article.

Experimental setups and testing procedure

Method of dark noise and LED triggering using waveform recording electronics

As shown in Fig. 1 (solid line), the PMT is put into the dark box. A CAEN mod-N470 module is used to feed high voltage to PMT. The signal waveforms are recorded by LeCroy 610Zi oscilloscope using self-triggering mode for the dark-noise charge distribution measurement and external triggering mode for the LED triggering charge spectrum. The LED is driven by BNC 8010 pulse generator, and a synchronized signal is provided by the generator to give external triggering. The incident light intensity on the photocathode is tuned by changing the amplitude of the supply voltage to the LED. Less than 10% effective triggering rate is tuned to ensure that the acquired charge distribution is a SPE spectrum [7]. Each waveform of the PMTs has been recorded and been analyzed by off-line processing which includes peak searching and baseline calculation. The data taking time for one self-trigger run including 99,999 events is about 28 min and for one LED trigger run including 99,999 events is about 1 h. The Hamamatsu PMT is measured at 1200, 1300, 1400, 1500, 1600, 1700 and 1800 V high voltage. For EMI PMT, test is applied at 1600 V high voltage.
Table 1

Measuring result of PMTs SPE spectra test

PMT type

HV (V)


Threshold (mV)







\(6.10\pm 0.11\)

\(2.02\pm 0.12\)

Dark noise


\(6.19\pm 0.03\)

\(2.08\pm 0.03\)




\(14.23\pm 0.20\)

\(4.82\pm 0.20\)

Dark noise


\(14.23\pm 0.11\)

\(4.25\pm 0.36\)




\(32.20\pm 0.20\)

\(8.66\pm 0.18\)

Dark noise


\(31.91\pm 0.16\)

\(9.35\pm 0.15\)




\(65.49\pm 0.96\)

\(21.49\pm 1.31\)

Dark noise


\(67.58\pm 0.46\)

\(19.26\pm 0.95\)




\(122.3\pm 1.0\)

\(28.95\pm 3.08\)

Dark noise


\(120.7\pm 0.5\)

\(33.09\pm 1.99\)




\(245.1\pm 1.3\)

\(63.25\pm 3.80\)

Dark noise


\(245.6\pm 0.6\)

\(65.36\pm 1.71\)




\(442.2\pm 1.7\)

\(89.60\pm 3.80\)

Dark noise


\(437.4\pm 1.1\)

\(91.99\pm 2.48\)





\(2.80\pm 0.03\)

\(1.15\pm 0.05\)

Dark noise


\(2.80\pm 0.03\)

\(1.08\pm 0.03\)

Fig. 4

Gain, ER and P/V of the Hamamatsu PMT of different methods

Table 2

Result of the comparison test of QDC and waveform recording oscilloscope systems with the PMT working at 1200 and 1400 V

HV (V)

Test system

Test method


Difference (%)








\(7.92\pm 0.02\)


\(2.45\pm 0.03\)




Dark noise

\(7.90\pm 0.00\)


\(2.49\pm 0.04\)






\(8.91\pm 0.07\)


\(2.62\pm 0.15\)




Dark noise

\(8.71\pm 0.04\)


\(2.80\pm 0.05\)







\(40.09\pm 0.08\)


\(11.16\pm 0.45\)




Dark noise

\(39.10\pm 0.1\)


\(11.09\pm 0.28\)






\(43.25\pm 0.83\)


\(14.3\pm 1.6\)




Dark noise

\(42.62\pm 0.16\)


\(12.34\pm 0.57\)




Comparison test of QDC system and oscilloscope waveform recording

In order to test the influence of different testing system, an experiment was did to compare the QDC (CAEN V965) and oscilloscope data acquisition system. The Hamamatsu PMT is tested with the systems that are shown in Fig. 1 (solid line). Dark-noise test is shown in the left plot , and the original signal of PMT is sent to fan-in–fan-out to obtain two same signals: one is sent into the low threshold discriminator (LTD) of which the threshold is set as 3 mV and the output signal of LTD, with 110 ns width, is served as trigger of QDC or oscilloscope; the other signal is acquired by the oscilloscope or QDC after delay. LED trigger test is shown in the right plot, and a LED driven by a pulse generator is used as light source. A synchronized signal is used as trigger. Then the signal of PMT is sent to the QDC or oscilloscope after fan-in–fan-out.

Waveform-recording date analysis and results

The LeCroy610Zi oscilloscope is operated at 2.5 GHz sampling rate, 500ns time window. A typical pulse of Hamamatsu PMT, which is at 1500 V HV with LED trigger method, is shown in Fig. 2. The charge distributions, acquired by using the two methods, are plotted on one canvas, as shown in Fig. 3. Left is Hamamatsu PMT tested at 1400 V and right is EMI PMT tested at 1600 V. Good consistency can be found from the comparison.

The spectra are fitted by a Gaussian plus an exponential function. The fitting parameters are listed in Table 1. Mean and sigma of the SPE spectra are two parameters representing the peak position and the spread of Gaussian function, respectively. The dispersion of the means is about 3% within an acceptable range (a 3% error is assumed according to the measurement system). Figure 4 shows the comparison of gain varying with the applied HV of Hamamatsu PMT using the two methods either the dark-noise pulses or the LED low-intensity light. It indicates that in the thermionic emission noise-dominated region the consistency of the two methods is pretty well. Figure 4 also shows the P/V and energy resolution (ER). It shows that the dark-noise pulses method could describe the SPE reasonably in the aspect of P/V and ER.

Comparison test of QDC and oscilloscope

Fig. 5

Comparison of SPE spectrum tested with QDC and the oscilloscope systems. Red line and blue line represent the spectra using LED triggering and dark-noise methods by oscilloscope system. Black line and green line represent the spectra using LED triggering and dark-noise methods by QDC system (colour figure online)

Table 2 shows the result of the comparison test of QDC and the oscilloscope systems. Data of the mean, ER and P/V well illustrate that these two systems are consistent within an acceptable range. Due to the different HV equipments used, the applied HV is different with the same HV setting value, resulting in different SPE responses between Tables 1 and 2. Figure 5 shows the comparison of SPE spectra obtained by QDC and oscilloscope system by multiplying a factor of 0.90 between the QDC spectrum and oscilloscope spectrum. The factor of 0.90 is because that there is a difference of expressing 1 pC by QDC and oscilloscope which would be ascribed to systematic error. Spectrum comparison illustrates that the test results are consistent between the QDC and oscilloscope at PMT gain, ER, and P/V within error range.


Our experiment confirms that the SPE calibration is consistent between using dark-noise pulses and using LED low-intensity light when the PMT working under recommended state. Therefore, we can calibrate the PMT by dark-noise online even replacing the commonly used LED low-intensity light method. In these ways, it will introduce less material near the target volume and calibration can be achieved directly which could be useful for the low background experiment.


  1. 1.
    P. Agnes et al., [DarkSide-50], First results from the DarkSide-50 dark matter experiment at Laboratori Nazionali del Gran Sasso. Phys. Lett. B 743, 456 (2015). arXiv:1410.0653
  2. 2.
    E. Aprile et al., First dark matter search results from the XENON1T experiment, Phys. Rev. Lett. (2017, in press). arXiv:1705.06655v2
  3. 3.
    The DEAP Collaboration, In-situ characterization methods for the Hamamatsu R5912 photomultiplier tubes used in the DEAP-3600 experiment (2017). arXiv:1705.10183v1 [physics.ins-det]
  4. 4.
    X. Meng et al., Measuring and study for photoelectron spectrum of the photomultiplier. N&D 25, 6 (2005)Google Scholar
  5. 5.
    R. Dossi et al., Methods for precise photoelectron counting with photomultipliers. Nucl. Instrum. Methods A451, 623 (2000)ADSCrossRefGoogle Scholar
  6. 6.
  7. 7.
    E.H. Bellamy et al., Absolute calibration and monitoring of a spectrometric channel using a photomultiplier. Nucl. Instrum. Methods A339, 468–476 (1994)ADSCrossRefGoogle Scholar

Copyright information

© Institute of High Energy Physics, Chinese Academy of Sciences; China Nuclear Electronics and Nuclear Detection Society and Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Key Laboratory of Particle Astrophysics, Institute of High Energy PhysicsChinese Academy of ScienceBeijingChina
  2. 2.School of PhysicsUniversity of Chinese Academy of ScienceBeijingChina

Personalised recommendations