Performance evaluation of a silicon strip detector for positrons/electrons from a pulsed a muon beam
T. Aoyagi, Y. Honda, H. Ikeda, M. Ikeno, K. Kawagoe, T. Kohriki, T. Kume, T. Mibe, K. Namba, S. Nishimura, N. Saito, O. Sasaki, N. Sato, Y. Sato, H. Sendai, K. Shimomura, S. Shirabe, M. Shoji, T. Suda, T. Suehara, T. Takatomi, M. Tanaka, J. Tojo, K. Tsukada, T. Uchida, T. Ushizawa, H. Wauke, T. Yamanaka, T. Yoshioka
PPrepared for submission to JINST
Performance evaluation of a silicon strip detector forpositrons/electrons from a pulsed a muon beam
T. Aoyagi a Y. Honda a H. Ikeda b M. Ikeno c K. Kawagoe f , g T. Kohriki c T. Kume e T. Mibe c K. Namba a S. Nishimura d N. Saito c O. Sasaki c N. Sato e Y. Sato c H. Sendai c K. Shimomura d S. Shirabe f , h M. Shoji c T. Suda a T. Suehara f T. Takatomi e M. Tanaka c J. Tojo f K. Tsukada a T. Uchida c T. Ushizawa i H. Wauke a T. Yamanaka g , T. Yoshioka g a Research Center for Electron Photon Science, Tohoku University,1-2-1 Mikamine Taihaku-ku, Sendai, Japan b Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency,3-1-1 Yoshinodai, Sagamihara, Japan c Institute of Particle and Nuclear Studies, High Energy Accelerator Research Organization,1-1 Oho, Tsukuba, Japan d Institute of Materials Structure Science, High Energy Accelerator Research Organization,1-1 Oho, Tsukuba, Japan e Mechanical Engineering Center, High Energy Accelerator Research Organization,1-1 Oho, Tsukuba, Japan f Department of Physics, Kyushu University,744 Motooka Nishi-ku, Fukuoka, Japan g Research Center for Advanced Particle Physics, Kyushu University,744 Motooka Nishi-ku, Fukuoka, Japan h Department of Physics, Tokyo Institute of Technology,2-12-1 Ookayama Meguro-ku, Tokyo, Japan i Department of Particle and Nuclear Physics, Graduate University for Advanced Studies,1-1 Oho, Tsukuba, Japan
E-mail: yamanaka_at_artsci.kyushu-u.ac.jp a r X i v : . [ phy s i c s . i n s - d e t ] M a y bstract: A high-intensity pulsed muon beam is becoming available at the at the Japan ProtonAccelerator Research Complex (J-PARC). Many experiments to study fundamental physics usingthis high-intensity muon beam are proposed. An experiment to measure the muon magnetic momentanomaly ( g −
2) and the muon electric dipole moment (EDM) is one of these experiments and itrequires a tracking detector for positrons from muon decay. Fine segmentation is required in adetector to tolerate the high rate of positrons. The time resolution is required to be much better thanthe muon anomalous spin precession period while a buffer depth of a front-end electronics needsto be much longer than the accelerated muon lifetime. Requirements of this detector also meetrequirements of a measurement of the muonium hyperfine structure interval at the J-PARC andanother experiment to measure the proton charge radius at Tohoku University. We have developeda single-sided silicon strip sensor with a 190 µ m pitch, a front-end electronics with a sampling rateof 200 MHz and a buffer memory depth of 8192, and a data acquisition system based on DAQ-Middleware for the J-PARC muon g − ontents g − A high-intensity pulsed muon beam at the order of 10 muons/s on average is becoming availableat the Muon Science Establishment (MUSE) located at the Material and Life Science ExperimentalFacility (MLF) in the Japan Proton Accelerator Research Complex (J-PARC) [1]. The repetitionrate of the beam is 25 Hz. The beam spill consists of two bunches separated by 600 ns with a bunchof 100-150 ns full width. The average intensity will be of the order of 10 muons/s when the 3 GeVproton synchrotron ring of the J-PARC is fully operated and the proton beam power of 1 MW isreached as designed, which is used for generating a muon beam. Short time operation of the protonbeam power of 1 MW was demonstrated in 2019 and it will be provided to beam users soon. Ina new muon beam line, the H-line, which is being constructed, another factor of improvement isexpected in the muon beam intensity [2]. Several experiments are planned at the H-line assumingthe beam intensity of 10 muons/s. A high-intensity muon beam is a key to search for new physicsbeyond the Standard Model in precise measurements of fundamental parameters of the muon withhigh statistics and is also useful for material science studies based on muon spin spectroscopy.– 1 – .2 Muon g − /EDM experiment at J-PARC A new experiment is proposed to measure the muon magnetic moment anomaly ( g −
2) and themuon electric dipole moment (EDM) at the J-PARC MUSE H-line [3]. The experiment uses adifferent method from the E821 experiment at Brookhaven National Laboratory [4] or the E989experiment at Fermilab [5] and aims to measure the muon g − . × − e · cm. In this experiment, amuon beam produced by a primary proton beam is slowed down and thermalized in a material. Thethermalized muon beam is accelerated to 300 MeV/ c to produce a low-emittance muon beam. Thismuon beam is injected into a storage ring using a novel 3D spiral injection method. The storagering is a 3 T magnetic resonance imaging-type solenoid with high field uniformity for the muonstorage region with an orbit diameter of 66 cm. A tracking detector is placed inside of the magnetand is required to reconstruct positron tracks from muon decay and to estimate the muon decaytime. The anomalous spin precession frequency of muons in this magnetic field is measured by themuon decay time distribution with higher-momentum positrons. The sensitivity to the muon spinbecomes maximum when positrons with momenta above 200 MeV/ c are counted. The MuSEUM experiment is another experiment proposed at the J-PARC MUSE H-line to measurethe muonium hyperfine structure interval and the ratio of muon to proton magnetic moments byusing the muonium spin resonance method at high magnetic field [6]. The number of positronsfrom decay of the muon beam after forming a muonium and experiencing a resonant microwave iscounted as a function of time in this experiment. A detector with fine segementation and a highsampling rate is required to tolerate positrons from a high-intensity muon beam of ∼ × muonsin one spill.Requirements for detectors in these experiments meet requirements of another experimentproposed at the Research Center for Electron Photon Science (ELPH), Tohoku University, whichwill measure the proton charge radius by low-energy electron scattering at the lowest momentumever achieved (ULQ2 experiment) [7]. In this experiment, electrons in the energy range of 20-60 MeV are scattered at a target, and the momentum and the scattering angle of scattered electronsneed to be measured. An electron spectrometer with the momentum resolution of 8 × − andthe scattering angle resolution of 5 mrad is developed. To achieve this resolution, a detector witha position resolution much better than 1 mm is required at a focal plane. The expected rate ofelectrons at the focal plane is 30 MHz in an area of 10 ×
10 cm . A positron tracking detector for the J-PARC muon g − c even inthe highest muon decay rate. To cover this volume with a reasonable cost, we choose silicon stripdetector technology. The maximum number of positrons from muon decay is expected to be six per1 ns and the maximum hit rate will reach 150 kHz/mm . The detector is required to tolerate thishit rate and to measure up to four minimum ionizing particle (MIP) charge to accommodate pile-up– 2 –f hits. The signal-to-noise ratio greater than 15 is required to reconstruct positron tracks in highefficiency. Since the anomalous spin precession period in the magnetic field of 3 T is about 2.1 µ s,the time resolution is required to be much better than this period. On the other hand, the data needsto be stored sufficiently longer than this period as well as the lifetime of muon with the momentumof 300 MeV/ c ( ∼ µ s). Another requirement comes from the pile-up of detector hits. If there arepile-up hits in the same sensor strip, a sensor signal waveform is distorted and the recorded timingis shifted. Since the rate of pile-up changes during the data taking period due to the lifetime ofmuon, it will bias the spin precession frequency measurement. This effect can be mitigated by areadout system with a small time walk. To constrain this effect, the time walk is required to be lessthan 1 ns over a signal range of 0.5-3 MIP charge.To satisfy these requirements, we have developed a single-sided silicon strip sensor and afront-end readout electronics optimized for the experiment. We have fabricated detector modulesconsisting of this sensor and the readout electronics controlling their qualities. Performance offabricated detector modules was evaluated at a laboratory and a beam test using the positron beamat the ELPH, Tohoku University. The design and the performance of the detector modules arepresented in this article. The detector module consists of a silicon strip sensor and two front-end readout boards.
A silicon strip sensor developed for at the J-PARC muon g − × µ m. The activearea is 97.28 mm × µ mand a length of 48.365 mm. The size of strips is determined to constrain the hit rate less than2 MHz and the estimated maximum hit rate at the J-PARC muon g − + strip and a read out strip is measured to be 170 pF. Because ofdouble-metal structure, signals can be read out from the orthogonal direction as well. This enablesus to use the same design of the readout system for horizontal or vertical strips since the sensor shapeis square. By stacking two sensors with different strip directions, two-dimensional coordinates ofcharged particles passing through two sensors can be determined. To construct 40 detection planeswith two-dimentional coordinates for the J-PARC muon g − + strips are connected to a common bias ring via a polysilicon resister of the resistanceof 10 ± Ω . The bias ring is connected to ground via a bias pad on the surface. The positivebias voltage is applied from either the back plane of the sensor or pads on the surface. The fulldepletion voltage was measured to be from 65 V to 95 V depending on the sensor according to themeasurements by the manufacturer and sensors are usually operated with the bias voltage above120 V. The capacitance of one strip to the back plane was measured to be about 17 pF above the– 3 –ull depletion voltage. A typical total leakage current of the sensor is about 0.2 µ A with the biasvoltage above the full depletion voltage at room temperature.
Figure 1 . Silicon strip sensor used for the detector modules (left) and the magnified image at the corner ofthe sensor (right).
The signals from the silicon strip sensors are read out by front-end boards with application specificintegrated circuits (ASICs) named “SliT128A”, which is a prototype of the front-end ASIC forthe J-PARC muon g − µ mCMOS process and has a chip size of 9.0 mm × µ s( = × Figure 2 . SliT128A chip.
Unity gainbuffer
Test pulseSignal from sensor Reference voltage
Output(Analog waveform)
Control register
Voltage DAC
Pre-amp
Shaping-amp
Slow control signal
Threshold
Sampling clock(200 MHz)
D-type flip-flopx 128 Memory D CK Q Read clock(100 MHz)x 128 channels
Serializerx 4
128 bitx8192 depth
Write start signal
Readout start signalASIC status signal(write busy, readout busy) Serialized data of 32 chfor each line
Analog part Digital part
Comparator
Amp.
Amp.
Figure 3 . Schematic logic diagram of analog and digital circuits in a SliT128A chip.
The front-end readout board consists of four SliT128A chips, a Xilinx Artix-7 XC7A200Tfield-programmable gate array (FPGA), and a small form-factor pluggable (SFP) optical transceiveras shown in figure 4. The data from the SliT128A chips are processed with the FPGA. The startsignals for the writing and readout processes can be input to the front-end readout board throughthe connector with the Nuclear Instrumentation Module (NIM) standard logic. The data from fourSliT128A chips are transferred to the first in, first out (FIFO) buffer implemented inside the FPGA– 5 –imultaneously in parallel, when the start signal for the readout process is asserted. The data of 32channels with a depth of 8192 points comes per one readout line. The data is saved into the FIFOwith a format of 45 bits: 13 bits to represent hit timing and 32 bits to represent existence of hits foreach channel. The data can be compressed by the zero suppression algorithm: if all 32 channelshave no hits, the data is not stored to the FIFO to suppress the amount of the data. After completingthe data transfer from the four SliT128A chips to the FIFO, the data stored in the FIFO is seriallytransferred to a computer via an optical cable with Ethernet and SiTCP protocols [8]. PC FPGA (Xilinx
XC7A200T)
SliT128A
SliT128A
SliT128A
SliT128A
Signals from sensor through the FPC boardTriggers
Slow control signalSerial data
Front-end readout board • ASIC status signal • Serial data • Slow control signal • Clocks for sampling(200 MHz) and readout(100 MHz) • W rite/readout start signals (Optical cable) (LEMO cable) Figure 4 . Configuration of the front-end readout board.
A sensor is glued on a 1.6 mm-thick glass epoxy circuit board which has a hole in the active area ofthe sensor. The positive bias voltage is applied from pads on the sensor surface, which is connectedto the back plane of the sensor. The bias pads are connected to ground. To suppress parasiticcapacitance, the bare SliT128A chips are directly placed on the front-end readout board and arewire bonded using 25 µ m-diameter aluminum (99% Al with 1% Si) wires. A sensor and SliT128Achips are connected by two flexible printed circuit (FPC) boards. One FPC board connects 512channels of a sensor to four SliT128A chips. Wire bonding is used to connect a sensor with an FPCboard and SliT128A chips with an FPC board.A glass epoxy circuit board with a sensor and two readout boards are glued on the same 3 mm-thick aluminum plate, which has a hole in the active area of the sensor. Two types of modules wereconstructed; one with strips in vertical and the other with strips in horizontal direction (figure 5).By stacking two types of modules successively, two-dimensional coordinates of charged particlespassing through two modules can be determined. The data acquisition (DAQ) system is constructed based on DAQ-Middleware [9, 10], which isa software framework for network distributed DAQ software. The framework consists of DAQ-– 6 – igure 5 . Detector modules with vertical (top) and with horizontal (bottom) strips.
Components and a DAQ-Operator. The DAQ-Component is a base unit that can run fully indepen-dently of each other. Users can develop a data acquisition path by connecting DAQ-Components.The DAQ-Operator controls DAQ-Components by sending the signals to each DAQ-Component,such as a start and a stop. Figure 6 shows a structure of the DAQ system. There are five typesof the DAQ-Components in the DAQ system for this module: reader, merger, dispatcher, logger,and monitor. Data from one front-end readout board is received by a reader component. The dataacquisition paths for each front-end readout board are merged by the merger component and themerged path is split into two by the dispatcher component. One path is connected with the storagecomponent and the data is saved in the hard disk drive. The other path is connected with the monitorcomponent and a part of the data is analyzed to monitor the status of the module and the DAQsystem itself. The maximum data rate from a front-end readout board is about 20 MB/s withoutthe zero suppression algorithm. It takes about 2.6 ms for the data transfer from four SliT128Achips to the FIFOs in the FPGA and it takes about 5.3 ms per front-end readout board without thezero-suppression algorithm for the data transfer from the FIFOs in the FPGA to a personal computer.One detector module can be operated up to a repetition rate of 75 Hz without the zero-suppressionalgorithm and it satisfies the requirement of the muon beam rate at the J-PARC.– 7 – erger
Reader
Dispatcher MonitorDAQ PCReadout boards LoggerDAQOperator
Figure 6 . Structure of the DAQ system based on the DAQ-Middleware.
Basic performance of silicon strip sensors was measured by the manufacturer before shipment. Thetotal leakage current and capacitance of the sensor were measured as a function of the bias voltage.Any strip defects (e.g., a shortage of AC-DC strips, or an isolation of strips) were recorded andonly the sensors whose defective strip ratios were less than 5% were delivered. The full depletionvoltage and the statistics of bad strips for 187 sensors measured by the manufacturer are shown infigure 7. The total leakage current and capacitance were also measured in our laboratory (figure 8)and they were consistent with the measurements by the manufacturer. The total leakage current wasalso measured at each step of module assembly to find any damage that appeared during assembly.
Full depletion voltage
Entries 187Mean 80.64Std Dev 6.9850 20 40 60 80 100 120 140Full depletion voltage [V]0102030405060708090 N u m be r o f s en s o r s Full depletion voltage
Entries 187Mean 80.64Std Dev 6.985
Bad strips
Entries 187Mean 0.06417Std Dev 0.3207 N u m be r o f s en s o r s Bad strips
Entries 187Mean 0.06417Std Dev 0.3207
Figure 7 . The full depletion voltage (left) and the statistics of bad strips per sensor (right) for 187 deliveredsensors measured by Hamamatsu Photonics K. K. – 8 –
50 100 150 200Bias voltage [V]00.020.040.060.080.1 A ] m T o t a l l ea k age c u rr en t [ Sensor No.59Sensor No.104Sensor No.175Sensor No.176 ] [ / n F / ( t o t a l c apa c i t an c e ) Sensor No.59Sensor No.104Sensor No.175Sensor No.176
Figure 8 . The total leakage current (left) and capacitance (right) as a function of the bias voltage of sensors,which were used in detector module assembly. The total leakage current value is corrected to the measurementat 20 ◦ C. Basic performance of SliT128A chips was measured before assembly to the readout board. Aprobe card was used to deliver signals from a SliT128A chip and power supplies to a SliT128Achip. Quality of SliT128A chips were categorized into a rank of A, B and C. Chips without anydefects are categorized as a rank A. Chips with a noise level greater than 1250 electrons on averageor more than five noisy (comparator output is always high) or dead (no comparator output) channelsare categorized as a rank B. Chips with severe defects such as malfunction of comparator output inmore than 32 channels are categorized as a rank C. The ratio of quality of SliT128A measured for115 chips is shown in figure 9. SliT128A chips without any defects (rank A) were used for moduleassembly. The same tests were performed after connecting SliT128A chips to a readout board andafter connecting a sensor to SliT128A chips. The readout boards without any defective SliT128Achips were used for module assembly.
A (26 %)B (59 %) C (16 %)
Figure 9 . The ratio of quality of SliT128A measured for 115 chips.
The pull test was performed on wires bonded on spare parts of detector modules prior to– 9 –ractical assembly. The bonding parameters were optimized to satisfy the pull strength greater than60 mN at any bonding positions (typically, greater than 80 mN). The same bonding parameterswere used at practical assembly.
Before the test with a positron beam, the noise, the time over threshold (ToT), and the time walk of thedetector module were evaluated using test pulses generated by a function generator. The noise canbe evaluated from the threshold scan with different injected test pulse charges. A complementaryerror function is fitted to the counting efficiency as a function of the threshold voltage. The signalamplitude corresponds to the threshold at 50% efficiency and the noise level corresponds to thedifference of thresholds at efficiency of 50% and 84.3%. Figure 10 shows examples of the countingefficiency curves of one readout channel. Figure 11 is the equivalent noise distribution of all readoutchannels in two detector modules. The mean of the equivalent noise charge (ENC) is measured to be725 electrons with a detector capacitance of the silicon strip sensor of 17 pF. For the actual detectorto be used at the J-PARC muon g − ∼ g − The performance of the detector modules was evaluated using a positron beam at the ELPH, TohokuUniversity [11]. The momentum-analyzed positron beam with the momentum of 730 MeV/ c wasproduced from 1.3 GeV electrons by a series of reactions of Bremsstrahlung and pair production.The setup of the beam test is shown in figure 14. Two detector modules whose strip directions areorthogonal were put perpendicular to the beam axis. The module with horizontal strips (Module 1)is downstream of the beam line and the module with vertical strips (Module 2) upstream. Eventswere triggered by the coincidence of two scintillation detectors sandwiching the silicon strip detector– 10 – hreshold DAC [LSB] - - - C oun t E ff i c i en cy Figure 10 . Examples of the counting efficiency of one readout channel as a function of the threshold withinjected charges of 1.92 fC (open black circles), 2.40 fC (open red triangles), and 2.88 fC (open greensquares), respectively. The threshold is shown in the unit of the least significant bit (LSB) of the DAC. / ndf c – – – Noise [electrons] N u m be r o f c hanne l s / ndf c – – – c – – – Figure 11 . Equivalent noise charge distribution of all readout channels in two detector modules. A Gaussianfunction is fitted to data. – 11 – njected test pulse charge [fC] T i m e o v e r t h r e s ho l d [ n s ] ToT
Entries 2041Mean 186.4Std Dev 31.82
Time over threhsold [ns] N u m be r o f c hanne l s ToT
Entries 2041Mean 186.4Std Dev 31.82
Figure 12 . ToT as a function of the injected test pulse charge of typical 16 readout channels (left) and theToT distribution with an injected test pulse charge of 3.84 fC for all readout channels in two detector modules(right).
Injected test pulse charge [fC] T i m i ng [ n s ] Time walk
Entries 2041Mean 17.21Std Dev 3.558
Time walk [ns] N u m be r o f c hanne l s Time walk
Entries 2041Mean 17.21Std Dev 3.558
Figure 13 . Leading edge timing of the comparator as a function of the injected test pulse charge of typical16 readout channels (left) and the distribution of the time walk over a signal range of 1.92-11.52 fC for allreadout channels in two detector modules (right). modules. The size of the scintillator is 800 mm ×
800 mm × µ s,which is longer than buffer length of 40.96 µ s. This is to avoid injecting the start signal for thewriting in the state of writing process. If the FPGAs detect the coincidence trigger signal, theystop the continuous start signal for the writing process and send the start signal for the readoutprocess. To reduce the trigger rate lower than the acceptable rate of the DAQ system, the triggerrate is prescaled to about 90 Hz. Almost only one positron hit the detector modules in one event inthis condition. A beam profiling monitor (BPM) consisting of two layers of 14 scintillating fibers– 12 –as installed to determine the incident position of positrons. The fiber is made from 3 mm squarecross-section scintillator, and the incident position is determined with a 3 mm resolution. Thecenter-of-the-beam position is set at the lower left of the modules as shown in figure 14 (right). Figure 14 . Perspective view of the detector setup in the beam test (left) and the view from the beam axis(right).
Figure 15 shows the charge distribution of the beam signal measured by two silicon strip detectormodules. The most probable value (MPV) of the charge distribution is 3 . ± .
002 fC, which isconsistent with the simulated MIP charge of 3.84 fC within an uncertainty on the charge calibration(about 2%). The signal-to-noise ratio is derived to be 32 . .
19 fC (about 30% of a MIP charge) is measured to be99 . ± . onstant 1.854e+02 – – – Signal charge [fC] N u m be r o f h i t s Constant 1.854e+02 – – – Figure 15 . Charge distribution of the positron beam measured by two detector modules. A Landaudistribution is fitted to data.
Threshold [fC] D e t e c t i on e ff i c i en cy Module 1Module 2
Figure 16 . The detection efficiency of two detector modules as a function of the threshold. – 14 –he time resolution is measured to be 3.4 ns (the Module 1 and 2) as the standard deviation in thecore region with a long tail distribution at the bias voltage of 120 V. At the bias voltage of 200 V,the time resolution in the core region is improved to be 3.2 ns and 3.1 ns as the standard deviationfor the Module 1 and the Module 2, respectively, and there is a small distribution in the tail region.
Sensor HV = 200 V / ndf c – – - Sigma 0.035 – t [ns] D - - - N o r m a li z ed nu m be r o f h i t s Sensor HV = 200 V / ndf c – – - Sigma 0.035 – Sensor HV = 120 V / ndf c – – – Sensor HV = 120 V / ndf c – – – Sensor HV = 200 V / ndf c – – - Sigma 0.032 – t [ns] D - - - N o r m a li z ed nu m be r o f h i t s Sensor HV = 200 V / ndf c – – - Sigma 0.032 – Sensor HV = 120 V / ndf c – – – Sensor HV = 120 V / ndf c – – – Figure 17 . Distributions of the time difference ( ∆ t ) between the detector modules and trigger for the Module 1(left) and the Module 2 (right). The detector modules are operated with the bias voltage of 120 V (solid redcircles with a solid curve) and 200 V (open blue circles with a dashed curve). The timing correction is appliedbased on the data with test pulses. The ∆ t value is shifted to adjust the peak position of the distribution atthe bias voltage of 200 V to 0 ns. A Gaussian function is fitted to data in a core region. The detector for positrons/electrons from a pulsed muon beam was developed aiming at the mea-surements of the muon g − , to have thesignal-to-noise ratio greater than 15 and time resolution much better than 2 µ s while it needs tostore data in a period more than 10 µ s. The time walk over a signal range of 0.5-3 MIP charge isrequired to be less than 1 ns to constrain a bias on the time measurement.The detector uses a silicon strip sensor with a 190 µ m strip pitch and a front-end electronicswith a sampling rate of 200 MHz and a buffer memory having a depth of 8192 points. The dataof one detector module in a period of 40.96 µ s (=8192 points × ∼ g − . ± .
1% at a threshold of 1.19 fC. The time resolution is measured to be 3.2 ns as thestandard deviation in the core of the distribution when the bias voltage to a sensor is 200 V. Theseresults satisfy the requirements of the detector. The time walk is measured to be 17.2 ns on averageand it is larger than the requirement. This will be solved by the next version of the front-end ASIC.
Acknowledgments
The authors would like to thank the KEK and the J-PARC muon section staffs for their strongsupport, the Open Source Consortium of Instrumentation (Open-It) of KEK for their support onthe electronics design, and Simon Eidelman at Budker Institute of Nuclear Physics and NovosibirskState University for his diligent proofreading of this article. This work is supported by JSPSKAKENHI Grants No. JP15H05742 and JP16H06340.
References [1] Y. Miyake et al.,
Current status of the J-PARC muon facility, MUSE , J. Phys.: Conf. Ser. (2014)012061.[2] N. Kawamura et al.,
New concept for a large-acceptance general-purpose muon beamline , Prog.Thoer. Exp. Phys. (2018) 113G01.[3] M. Abe et al.,
A new approach for measuring the muon anomalous magnetic moment and electricdipole moment , Prog. Theor. Exp. Phys. (2019) 053C02.[4] G. W. Bennett et al. (Muon ( g −
2) Collaboration),
Final report of the E821 muon anomalousmagnetic moment measurement at BNL , Phys. Rev. D (2006) 072003.[5] J. Grange et al. (E989 Collaboration), Muon ( g − ) Technical Desgin Report , arXiv:1501.06858 .[6] K. Shimomura, Muonium in J-PARC; from fundamental to application , Hyperfine Interact. (2015) 89–95.[7] T. Suda et al.,
Measurement of Proton Charge Radius by Low-Energy Electron Scattering , J. ParticleAccelerator Society of Japan (2018) 52–59.[8] T. Uchida, Hadware-Based TCP Processor for Gibabit Ethernet , IEEE Transactions on NuclearScience (2008) 1631–1637.[9] Y. Yasu et al., A Data Acquisition Middleware , in ,(Batavia, IL, USA), pp. 1–3, 2007. DOI.[10] Y. Yasu et al.,
Development of DAQ-Middleware , J. Phys. Conf. Ser. (2010) 022025.[11] T. Ishikawa et al.,
A detailed test of a BSO calorimeter with 100-800 MeV positrons , Nucl. Instrum.and Meth. A (2012) 348–360.(2012) 348–360.