Performance Evaluation of Electron Multiplier Tubes as a High-Intensity Muon Beam Monitor of Accelerator Neutrino Experiments Open Access

Abstract

Upgrade work towards increasing the beam intensity of the neutrino beamline at the Japan Proton Accelerator Research Complex (J-PARC) is underway. Monitoring tertiary muon beams is essential for stable operation of the beamline. Accordingly, we plan to replace the present muon monitor sensors with electron multiplier tubes (EMTs). We investigated the radiation tolerance and linearity response of EMTs using a 90 MeV electron beam. EMTs were irradiated with electrons up to 470 nC. EMTs show higher radiation tolerance than the Si sensors which are presently used as one of the muon monitor detectors for the Tokai-to-Kamioka (T2K) long-baseline neutrino experiment at J-PARC. The integrated charge yield decrease is found to be less than 8% after a beam irradiation equivalent to 132 days of operation at the future J-PARC beam power of 1.3 MW. The EMTs show linearity better than |$\pm$|5% up to the future beam intensity. The observed yield decrease is likely due to dynode deterioration, based on detailed investigation. The studies described here confirm that EMTs can be used as a high-intensity muon beam monitor. Owing to the reported results, we are proceeding with the installation in the J-PARC neutrino beamline.

1. Introduction

In the long-baseline neutrino experiment, it is essential to increase the number of observed neutrino statistics. At the Japan Proton Accelerator Research Complex (J-PARC) [1], the Tokai-to-Kamioka (T2K) experiment [2] is running and Hyper-Kamiokande [3] will run to search for CP violation in the lepton sector, and an upgrade to increase the intensity of proton beams for producing neutrino beams is underway. Neutrinos are produced along with muons from charged pions emitted by protons hitting a graphite target. Three magnetic horns are installed in the beamline to focus parallel charged pions in the direction of the T2K far detector, Super-Kamiokande [4]. Monitoring beam intensity and direction is crucial for maintaining safety at the facility as well as stable data taking and control of systematic uncertainties in the physics analysis. In T2K, for the purpose of monitoring the beam intensity and direction, a muon beam monitor (MUMON), composed of two arrays of silicon pin-photodiodes (Si) and ionization chambers (IC), has been used since the beginning of the operation [5,6]. Each sensor array has |$\rm 49\, (= 7 \times 7)$| sensors at 25 cm intervals and covers an area of |$\rm 150 \times 150\, cm^2$| with respect to the beam axis. MUMON measures the direction of the muon beam and hence the neutrino beam as well with a precision better than |$\rm 0.28\, mrad$|⁠. The resolution is better than |$\rm 3.0\, mm$|⁠, which is equivalent to |$\rm 0.025\, mrad$|⁠. It also monitors the stability of the beam intensity with a resolution better than 3%. To preserve the required precision, it is estimated from a Monte Carlo simulation study that the tolerance for nonlinearity is 5% [1]. As the proton beam power increases from |$\rm 510\, kW$| as of 2020 to |$\rm 1.3\, MW$| in 2028 [7], MUMON needs to be able to work with such high-intensity beams. This intensity increase is achieved by an increase in the repetition cycle and the number of protons contained per pulse. Eight bunches of protons, called a spill, are accelerated in the J-PARC Main Ring, and will be extracted every |$\rm 1.16\, s$| to be transported to the graphite target. The high-intensity neutrino and muon beams are made by increasing the intensity of the proton beam and the electromagnetic horn current. The muon flux at the MUMON is estimated from a Monte Carlo simulation to be |$1.09\times 10^{5}\, \rm muons/cm^2/(10^{12}\, protons\ on\ target)$| at a +250 kA horn current. The muon flux is estimated to be |$8.1\times 10^6\, \rm muons/cm^2/bunch$| and |$5.6\times 10^7\, \rm muons/cm^2/sec$| at the MUMON array center with a |$\rm 1.3\, MW$| proton beam power and a |$\rm +320\, kA$| horn current. In this study, the focus is both the response to the beam in one bunch and the tolerance to the integrated beam. Table 1 summarizes the beam upgrade planned at J-PARC.

It has previously been reported that there are concerns about Si and IC regarding their performance during the higher-intensity beam operation; Si yield was found to degrade after irradiation and IC showed a nonlinear behavior in relation to beam power [8]. Diamond sensors were investigated as a replacement candidate, but they do not show sufficient radiation tolerance [9]. As an alternative MUMON sensor candidate in the future, the first prototypes of electron multiplier tubes (EMTs) were investigated using the T2K muon beam [8]. This study revealed that the EMT time response is faster than Si and IC, and that they keep signal linearity up to a |$\rm 460\, kW$| beam power, which was the highest power achievable at that time. The signal yield was stable within |$\pm$|1% while a yield drift was observed immediately after the beam irradiation, which is referred to as initial instability in the present paper as one of the main study topics. These observations indicate the potentiality of EMTs as a new MUMON candidate. Linear response and radiation tolerance are required to use EMTs at the same position as Si in J-PARC for the next-generation MUMON. We therefore tested EMTs in detail using an electron beam at Research Center for Accelerator and Radioisotope Science (RARIS) of Tohoku University. In the beam irradiation test, we investigated the linearity, radiation tolerance, and initial instability of EMTs. In addition, we investigated the causes of the observed degradation due to radiation and potential causes for the initial instability.

The rest of this paper is structured as follows. After describing our EMT design in detail in Section 2, we show electron beam irradiation tests to evaluate the performance of EMTs in Section 3. In Section 4, the results of the beam irradiation tests are presented. In Section 5, we discuss temperature dependence and causes of radiation degradation, followed by a summary in Section 6.

2. EMTs

Photomultiplier tubes (PMTs) generate signal when photons enter a glass window with a photocathode. They can also generate signal when charged particles penetrate their electrode and secondary electrons are emitted. The performance of PMTs is degraded by radiation, mainly due to degradation of the transmittance of the glass window. Hence, their radiation tolerance as a charged particle detector using parts besides the glass window is expected to be high. PMTs were then tested as a candidate sensor, but they were not satisfactory for use as MUMON since the yields were continuously decreased and they showed a poor linearity [1]. Because the observed yield decrease seemed to be caused at the photocathode, we decided to replace it with aluminum to achieve higher radiation tolerance. HAMAMATSU R9880U series PMTs were modified with their photocathode deposited with aluminum to produce EMTs. The sensitive area of the aluminized cathode is |$50\, \rm mm^2$| and that of the dynode is about |$77\, \rm mm^2$|⁠. The secondary electrons produced by charged particles passing through the cathode or dynode are amplified in dynodes. The shape of the EMT dynode is a metal channel type with 10 stages.

The original bleeder circuit used for PMTs has an insulating cover made of polybutylene terephthalate (PBT) or polyacetate (POM), but we decided not to use the cover as shown in Fig. 1 to avoid potential degradation due to radiation. The bleeder circuit connected to EMTs consists of 11 resistors |$R_1 \!-\! R_{11}$| and 5 capacitors |$C_1 \!-\! C_5$|⁠, as shown in Fig. 2. Each resistance and capacitance value is optimized to maintain a good linearity up to large signal output; these values are listed in Table 2.

3. Electron beam irradiation test at RARIS

3.1. Experiment

Investigation items in the electron beam irradiation test include linearity, radiation tolerance, and initial instability of EMTs. The experiment was conducted four times between 2019 and 2022. In the first beam test in 2019, we mainly tested EMTs’ linearity. In the second beam test in 2020, we irradiated EMTs as well as Si sensors to compare their radiation tolerances. The third and fourth beam tests were performed in 2021 and 2022, respectively, to investigate specific features of EMTs and their individual difference in quality. In particular, we investigated temperature dependence of the EMT output.

Three different electron beam intensities were adopted depending on the purpose, as shown in Table 3. We mainly used the low- and high-intensity beams for multiple different purposes. The low-intensity beam was adjusted to about |$1\, \rm pC$| in a macro pulse so that one pulse is equivalent to the muon yield at MUMON for one bunch of the J-PARC proton beam, which is common in the present and future beams. The high-intensity beam corresponds to the maximum available intensity of the facility, i.e. a 20 nC pulse with 7 Hz, and was used to expose EMTs to a large amount of radiation, which is suitable for radiation tolerance investigation. Another beam setting, the middle intensity, was used only for initial instability investigation. As with the high-intensity beam, the purpose is to expose EMTs to a large amount of radiation, but the amount of electrons in one pulse in the middle-intensity beam is smaller than in the high-intensity beam. The high-intensity beam operation is aimed at exposing the sensor to charged particles equivalent to an accumulated several hundred days’ operation of the J-PARC beam line. The initial instability test using a medium-intensity beam is sufficient for an accumulated 14-day operation of the J-PARC beam line, so a middle-intensity beam was arranged to avoid overexposure. In order to reduce the effect of electron scattering, the energy of the electron beam was set to the highest available one at the facility, 90 MeV. Several types of detectors were used to measure the irradiation amount and the profile for each beam intensity. Our setup was similar but with some optimization for each beam test, and Fig. 3 shows the one from the fourth beam test. In the first beam test, the beam operation was stopped for a few minutes each time the intensity was switched in order to change the setting in the beamline, which might change the vacuum state of the accelerator and cause the beam intensity to become unstable. Therefore, in the second beam test, we introduced remote controls for some devices, and in the third and fourth beam tests, all settings were controlled remotely.

In order to quantify the amount of irradiation of EMTs, the intensity and profile of the electron beam during the high-intensity operation were measured using a Current Transformer (CT) [10] and an Optical Transition Radiation monitor (OTR) [11], respectively. In the low-intensity operation, we measured the beam intensity and profile using a Si pin-photodiode (ref-Si) of the same type as used in MUMON and a 64-channel silicon pin-photodiode array (Si-array). The ref-Si was used as a reference to normalize the EMT signal. The detectors used in the tests are summarized in Table 4. To minimize the exposure, the Si sensor and Si-array were moved to a position away from the beam axis using movable stages during the high-intensity operations. Figure 4 schematically shows the locations of the OTR, CT, ref-Si, Si-array, and EMT. In the fourth beam test, a heater and a thermometer were installed to measure the temperature dependence of the EMT response. The CT, EMT, and Si signals were acquired with an 8-channel 250-MHz sampling 14-bit 2-Vpp Flash Analog to Digital Converter (FADC) (CAEN DT5725). The Si-array signal was acquired using two 32-channel 62.5-MHz sampling 12-bit 2-Vpp FADCs (CAEN DT5740).

3.2. High- and middle-intensity beams

During the high-intensity beam operation, the EMT signal line and ground were connected by a 1|$\, \rm \Omega$| resistor to suppress the voltage drop caused by too much current. The CT has a ring core whose inner and outer diameters are |$79\, \rm mm$| and |$51\, \rm mm$|⁠, respectively, and width is |$25\, \rm mm$|⁠. The material is FINEMET (Hitachi Metals, Ltd.) with a magnetic permeability of |$18500\, \rm H/m$|⁠. The number of turns of the coil is 1 which makes the CT response differential. This was decided upon to get a sufficiently large pulse height. The frequency-dependent response was measured in advance and was used to reconstruct the beam longitudinal distribution from the CT output waveform. The beam flux was calculated by integrating the area of the reconstructed waveform. Typical raw and reconstructed waveforms are shown in Fig. 5. As can be seen around the 2|$\sim$|3 |$\rm \mu$|s and 4|$\sim$|7 |$\rm \mu$|s segments of the reconstructed waveform, the baseline distortion remained and reconstruction is not as good as desired. This is because of the low sampling rate compared to the beam pulse. The flux of each beam pulse was calculated from the product of the height and width of the reconstructed waveform.

The OTR monitor was used since the third beam test as a specialized profile monitor for the high-intensity beam. An aluminum foil was set at 45 degrees to the beam axis as the OTR generation medium (see Fig. 6), and images were taken with a CMOS camera (model number CS165MU, lens: FL-BC7528-9M) positioned 30|$\, \rm cm$| away. Figure 7 shows a typical beam profile taken by the OTR during the high-intensity beam operation. The irradiation efficiency is defined as the ratio of electrons that hit the EMT to the electrons in the beam. The electron beam size of RARIS is about 1 mm, which is smaller than the sensitive area of EMTs. The muon beam size of the J-PARC neutrino beamline is a few hundred cm and then expected to irradiate EMTs almost uniformly. We designed an experiment to achieve a uniform irradiation on EMTs by changing their position with a movable stage during the high-intensity beam operation. In the second beam test, EMTs were irradiated at nine points in a grid pattern, and in the third beam test, EMTs were irradiated at seven points in a hexagonal arrangement, as schematically shown in Fig. 8. From the third beam test, we installed the OTR to measure the high-intensity beam profile, and calculated the irradiation efficiency which is defined as the fraction of beam particles that hit the EMT’s sensitive area. The irradiation efficiency at the third beam test was found to be 70%. However, it was also found that the irradiation density differed by a factor of 3 depending on the position on the EMT. In the fourth beam test, in order to improve the uniformity, each EMT was irradiated while being continuously scanned over a range of |$16\times 16\, \rm mm^2$|⁠, as shown in the right-hand side of Fig. 8. The resultant irradiation efficiency was about 28%, and the nonuniformity was about 3% or less. By multiplying the irradiation efficiency by the beam flux measured by the CT, the total exposure on each EMT was determined.

3.3. Low-intensity beam

The ref-Si was used to measure the intensity of the low-intensity beam. An attenuator with a factor of 1/79.4 was applied. A low-pass filter was inserted in the bins line to ensure stable voltage supply for the Si sensors. When measuring the EMT yield, we acquired the ref-Si signal at the same time, and by taking the ratio, we can accurately measure the EMT yield regardless of shot-to-shot variations in the electron beam. The applied voltage to EMTs was |$-450\, \rm V$| for the signal check. Typical ref-Si and EMT waveforms are shown in Fig. 9.

The ref-Si was calibrated using the CT. Since the low-intensity beam is too weak to be detected by the CT, the calibration was performed using a beam that is several times more intense than the low-intensity beam. Figure 10 shows the relation between the ref-Si signal yield and beam intensity measured by the CT.

The Si-array was used to measure the profile of the low-intensity beam. The Si-array has 64 |$(= 8\times 8)$| silicon pin-photodiode sensors whose size is |$\rm 2.5\, mm$| at |$\rm 3\, mm$| intervals and can monitor a |$23.5 \times 23.5\, \rm mm^2$| area. The variation in gain for each channel was measured in advance using an LED. We calculated the signal charge for each channel of the Si-array and reconstructed a 2D profile as shown in Fig. 11. By fitting this with a 2D Gaussian, the position and size of the beam are obtained. The fluctuation in the position of the low-intensity beam is approximately |$1\, \rm mm$| in the horizontal direction and |$4\, \rm mm$| in the vertical direction, and a typical beam width is 1|$\sim$|6 mm, which is smaller than the EMT cathode area.

4. Results

4.1. Linearity

In order to evaluate the linearity of EMTs, we irradiated a similar flux as from the J-PARC neutrino beamline operation and compared the yield with Si sensors (see Fig. 12). According to the previous studies [5], Si sensors keep their linearity within |$\pm 1.9\%$| up to |$3.5\, \rm pC/cm^2$|⁠. For the linearity measurement, instead of setting certain beam powers, which requires special care in beam control and hence is difficult, we made use of pulse-to-pulse variations of the low-intensity beam. The expected muon charge per pulse from the J-PARC neutrino beam at the 1.3 MW operation is 1.0 pC for each EMT. The beam irradiation test result demonstrates that the EMT response is linear within |$\pm$|5% for the range between 0.2 and 1.0|$\, \rm pC/pulse$|⁠, the region of interest in the present study.

4.2. Radiation tolerance

In order to confirm whether the EMT can be used for a long period at the J-PARC neutrino beamline, we irradiated EMTs with a high-intensity electron beam equivalent to several years of J-PARC neutrino beam operation, and measured the relative integrated charge yield of EMTs. As shown in Fig. 13, the EMT integrated charge signal is decreased, while linearity keeps within 5% after electron irradiation at |$\sim$||$\rm 450\, nC/cm^{2}$| that corresponds to |$\sim$|580 days’ operation at J-PARC in 2028. Figure 14 shows the irradiation results obtained from the second to fourth beam tests. For comparison, the Si sensor results are also shown. The radiation tolerance of the EMT is overwhelmingly better than that of the Si, and even after 700 days of operation at the future power of 1.3 MW, the integrated charge signal decrease is less than 18%. It was found that there are individual differences in EMTs. Some EMTs only degrade by 5% after 550 days of irradiation at 1.3 MW.

The typical operating period for the beam in a year anticipated at Hyper-Kamiokande is 132 days, and therefore we require that the sensors withstand at least 132 days of continuous operation. If the Si is operated for 132 days at the future beam condition, the integrated charge yield would decrease by 25%, and in addition, the leakage current would increase, making it impossible to apply high voltage. On the other hand, the integrated charge signal decrease of the EMT is expected to be less than 8% even after 132 days. Some EMTs hardly even degraded after 132 days. The degradation rate was found to be slower after more than 200 days of use. Therefore, by calibrating regularly, it can be used for several years without replacement.

Systematic errors on the result in Fig. 14 are estimated and shown in Table 5. The uncertainty on the EMT yield normalized by ref-Si arises from the variations in beam position because the sensitive areas of the EMT and Si are different. The beam profiles measured by Si-array were used to estimate the fluctuation of the beam entering the sensitive areas. We estimated the systematic uncertainty of the CT charge calculation method. This is the uncertainty that arises when converting the irradiation amount to the number of operation days in the J-PARC neutrino beamline. The MUMON sensors are exposed to muons with an average energy around |$\rm 3\, GeV$|⁠, while the beam is composed of |$90\, \rm MeV$| electrons in this beam test. Simulations with FLUKA [12,13] show that the energy deposit to an EMT is different by about 3%. Therefore, when converting the exposure to the EMT in this beam test to that of the J-PARC neutrino beam operation, this 3% is taken as a systematic uncertainty. For the low-intensity beam operation, the effect of particle difference is cancelled out because the evaluation is based on the EMT yield over the Si yield.

4.3. Signal stability

In measurements at the MUMON site in the past, an integrated charge yield decrease by 2.6% was observed about a week after the start of irradiation, and then the yield became stabilized, as reported in Ref. [8]. In order to confirm the stability of EMTs at the beginning of the operation in T2K, we repeated the exposure process to produce an exposure equivalent to several days of T2K operation using the middle-intensity setting (see Table 3) and measured the EMT response before irradiation with the high-intensity beam. The measurement results are shown in Fig. 15. The result indicates that EMTs have individual differences in performance, which is part of our items for study in the beam test. In this measurement, the variation in the integrated charge yield during the middle-intensity irradiation was less than 2%. This will be further discussed in Section 5.1.

5. Discussion

In this section, we will discuss the following two points. One is temperature dependence as a potential cause of the initial instability observed at J-PARC. The other is the cause of the integrated charge signal decrease of up to 18% by the irradiation, as shown by the red points in Fig. 14.

5.1. Temperature dependence

As described in Section 4.3, initial instability was not observed as clearly as in the previous J-PARC test [8]. During the test at J-PARC, EMTs were installed at a location where the temperature is not controlled and can vary as much as |$\rm 15\, K$| on a weekly timescale depending on the accelerator operation status. Therefore, we measured the temperature dependence of the EMT integrated charge signal in the fourth beam test. Unlike the test location at J-PARC in Ref. [8], the actual MUMON sensor location is surrounded by the enclosure and temperature is controlled at approximately |$34^\circ \rm C$|⁠, so the temperature dependence measurements were performed including |$34^\circ \rm C$|⁠.

Although the EMT cathode is made of vapor-deposited aluminum, it is sensitive to photons to some extent, so we evaluated using an LED at first. The EMT was placed in a constant temperature bath and the yield was measured at different temperatures. Since the LED response also changes depending on the temperature, we corrected the LED response based on taking the standard specification “NSPW500GS-K1” of Nichia Chemical Industries, Ltd. as typical temperature characteristics. At the same time, we also measured the Si, which is supposed to have small signal temperature dependence,¹ to confirm the correction of the temperature characteristics of the LED. Figure 16 shows the integrated charge yield at different temperatures. The temperature dependence coefficient of the EMT yield is 0.16–0.18%/|$^\circ \rm C$|⁠, which is insufficient to explain the integrated charge signal fluctuations around the MUMON test site.

We measured the temperature dependence of the EMT integrated charge signal also in the fourth beam test. The temperature was varied by applying hot air to the EMT using a heater and fan, and a platinum resistance thermometer was used to measure the temperature at the EMT location. Figure 17 shows the measured temperature dependence, and the temperature change coefficient in the range of 30|$^\circ \rm C$| or higher is |$0.6\%/^\circ \rm C$|⁠. This is larger than that measured using an LED.

Among the measurements using the LED and that of the electron beam irradiation test, the latter measures the signal generated by penetrating charged particles. The reason for the different temperature dependence is unknown. We consider the latter case to estimate the effect at J-PARC because the signal generation mechanism is the same. Then, the yield instability of 2%–3% corresponds to a variation of about |$4^\circ \rm C$| and can explain the observed initial instability at J-PARC. The temperature inside the temperature-controlled enclosure, which is the actual environment of the MUMON sensor location, is controlled so that the temperature change is less than |$\pm 0.7^\circ \rm C$| [14]. Therefore, the temperature dependence of the EMT is not an issue even if it exists.

5.2. Cause of yield decrease

The radiation tolerance of EMTs is found to be much better than that of the presently used Si, but an integrated charge signal drop up to |$18\%$| was observed when EMTs were irradiated with |$700\times 10^3\, \rm nC$| electrons. If the cause of the degradation is identified, EMTs could be used for longer by taking countermeasures. In order to find out where the degradation originates, we investigated each of the four components of the EMT: the last dynode, bleeder circuit, aluminum cathode, and entire dynodes. Each location is schematically shown in Fig. 18.

5.2.1. Last dynode

With amplification of electrons, a large number of secondary electrons collide with the final stage of the dynode. In order to verify the hypothesis that the integrated charge signal is reduced due to degradation of the last dynode, we conducted an exposure test with and without high-voltage application in the third beam test. Figure 19 shows the result from this test. Even when high voltage was not applied during exposure, similar degradation was observed. This result indicates that degradation in the later stage of the dynodes is not a main cause of radiation damage.

5.2.2. Bleeder circuit

We irradiated only the bleeder circuit with a high-intensity beam and then attached it to an EMT to measure the change in yield. The result is summarized in Table 6. We observed a tendency for the integrated charge yield to slightly increase rather than decrease. We measured the resistance and capacitance of the circuits before and after irradiation using a multimeter. There is no significant difference in the measured values and they are consistent with the catalog values.

5.2.3. Aluminum cathode

We investigated the possibility that the probability of electron emission from the aluminum-deposited cathode surface was worsened by irradiation. We prepared a special bleeder circuit in which the resistance between the cathode and the first dynode (⁠|$R_1$| in Fig. 2) was set to |$0\, \Omega$|⁠. When this bleeder circuit is used, when a charged particle passes through, the initial electron is not generated at the cathode but at the first dynode. Therefore, if the cathode degradation is the main cause of the integrated charge signal decrease, the special circuit would prevent the integrated charge yield from decreasing. Table 7 shows the yield change after the high-intensity beam irradiation for each condition. There are variations in the result, but yield reduction is observed for one of the tests with the special bleeder circuit.

5.2.4. Entire dynodes

Since the three hypotheses discussed above have been found to be not likely the major cause of degradation, we suspect that the degradation of dynodes would be the main cause. The state of the secondary electron emitting agent (alkali-metal antimony) coated on the dynodes might have been deteriorated due to irradiation, and so the secondary electron amplification factor would have decreased. In Fig. 14, EMTs that decreased by only a few percent can be explained by the aluminum cathode, while it is not enough to explain the EMT that decreased by 18%, hence a remaining possibility suggests the entire dynodes. We prepared EMTs without the secondary electron emitting agent and conducted a test. However, the gain of such EMTs was found to be very low even when applying a voltage of |$\rm -1000\, V$|⁠: about 0.5% compared to a usual EMT with an applied voltage of |$\rm -450\, V$|⁠. The radiation tolerance of that EMT was not measured, as no signal was seen in the low-intensity setting of the beam irradiation test. The signal is too small to accurately measure the muon beam and we need an alternative approach to suppress the integrated charge signal decrease if it is due to degradation of the dynodes.

6. Summary

We have been developing EMTs as a new sensor for the muon monitor in the future J-PARC neutrino beamline operation. We measured the linearity and radiation tolerance of EMTs using the electron beam at RARIS. The EMTs show linearity better than |$\pm 5\%$| up to the future beam intensity. The integrated charge yield decreases by 8% over the equivalent of 132 days of the 1.3 MW beam operation, indicating that EMTs have much higher radiation tolerance than the currently used Si sensors. After over 200 days of use with the 1.3 MW operation at J-PARC, the degradation becomes rather mild, so with careful calibration, it may be possible to keep using EMTs without replacement. In the beam test, the cause of the degradation was investigated in detail. The aluminium cathode, bleeder circuits, and last dynodes are found to be not likely the main cause. The most suspicious cause is degradation of the entire dynodes. We also investigated the temperature dependence of the EMT response. The measured temperature dependence of the EMT integrated charge yield indicates that the initial instability previously observed in the prototype test at J-PARC is due to the variation of temperature at the test location, but this should not be an issue as the actual muon monitor location will be temperature controlled. This study confirms that EMTs retain a good linearity response and high radiation tolerance compared to the present Si sensors. Based on the reported results, we concluded that EMTs can be used as a new MUMON detector and are currently preparing for their actual installation in the J-PARC neutrino beamline.

Acknowledgements

We thank Hamamatsu Photonics K. K. for the cooperation in producing the EMTs. This study was performed using facilities of Research Center for Accelerator and Radioisotope Science, Tohoku University (Proposal No. 2915, 2943, 2977, 3007). The authors are grateful to the J-PARC accelerator group for supplying a stable beam.

Funding

This work was partially supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers JP17J06141, JP16H06288, JP21K03591 and the U.S.-Japan Science and Technology Cooperation Program in High Energy Physics. This work was supported by Japan Science and Technology Agency (JST), the establishment of university fellowships towards the creation of science technology innovation, Grant Number JPMJFS2138.

Footnotes

References

Author notes