Time resolution measurements of scintillating counters for a new NA62 trigger charged hodoscope

The results of time resolution measurements on cosmic muons with MWPC chambers as a tracking system are presented. Two options of dual SiPM readout considered: direct light collection from opposite cut corners of the scintillator and light collection using two oppositely directed bundles of WLS fibers. Two types of SiPMs are used: S10985-050C and MicroFB-60035. The best result achived using mean-time between two SiPMs. For the direct readout from 134 × 107 × 20 mm3 scintillator time resolution σ = 300±10 ps and maximum signal arrival time spread ΔT = 1.6±0.2 ns. Using WLS fiber readout option with 265 × 107 × 30 mm3 scintillator tile signal arrival time spread ΔT = 150±50 ps and time resolution σ = 370±10 ps.


Introduction
The main goal of NA62 [1] experiment is to study extremely rare kaon decay K + → π + νν with 10% precision. This decay is strongly supressed in Standard Model (SM) and could be calculated with high precision BR(K + → π + νν) = 7.81 · 10 −11 which makes it sensetive to New Physics beyond the Standard Model. Current experimental value of branching fraction BR(K + → π + νν) = 1.73 +1.15 −1.05 × 10 −10 [2] was obtained by the E787 and E949 experiments and has limited accuracy. NA62 experiment is designed to work with high intensity hadron beam (750 MHz) with 75 GeV/c momentum and ≈ 10 MHz expected rate of decayed kaons. To study so rare decays it is important to have good particle identification system and strong background suppression. One of the important detectors of NA62 trigger system is Charged-particle Hodoscope (CHOD)system of scintillation counters with pad structure covering area of 2140 mm around the beam pipe. CHOD is designed to define the apperture for charged particles for the level-0 trigger and to veto possible interactions in the RICH mirror plane (≈ 20% of radiation length). According to Monte-Carlo simulation the total rate of hodoscope expected to be ≈ 32-35 MHz with maximum 400-500 kHz rate per tile (see figure 1 left) in the beam pipe region. The acceptance of K + → π + νν decay per tile is presented on figure 1 right. It is important to note specific region marked with red line that could be excluded from trigger to significantly reduce the total rate with less than 1% signal loss.

New CHOD prototypes
New CHOD is scintillating hodoscope with a pad structure and total number of 148 counters. Each pad designed to have dual SiPM readout. Thus 296 electronic channels are required. The New CHOD consists of pads with size 268 × 108 mm 2 on periphery and 134 × 108 mm 2 at the center region. Pads with different size are used to keep balance between minimum total number of channels -1 - Figure 1. MC simulation of the total rate in hodoscope and K + → π + νν signal acceptance: left -total rate of New Charged Hodoscope per small tile 132.5 × 107 mm 2 ; right -acceptance of K + → π + νν decay in per mil per tile. and rate per channel less than 500 kHz. The sensitive area of hodoscope covers 2140 mm except the region close to the beam pipe ( 280 mm). Polymerized scintillators made in IHEP (Protvino) with similar to BC-408 properties are used. Two different options of light collection considered: direct light collection from opposite cut corners and light collection using WLS fibers.
Direct readout. For the direct readout the counter prototype consist of 133.5 × 107 × 20 mm 3 scintillator (wrapped by Tyvek) and two SensL MicroFB-60035 SiPMs with 6 × 6 mm 2 sensitive area as a photodetectors. The SiPMs were glued with optical contact to opposite cut corners of scintillator (refer to figure 2 left).
Fiber readout. The prototype with WLS fiber readout consist of 268 × 108 × 30 mm 3 scintillator (wrapped by Tyvek) with 18 grooves: 16 on the top face plus one on each of two side faces. The grooves are oriented perpendicular to the long (268 mm) side of the tile to minimize the effect of geometry on time resolution (refer to figure 2 right). BCF-92 WLS Fibers with 1 mm and 40 cm long are glued and grouped into two oppositely directed bundles with 9 fibers each. Distance between fibers in one bundle is ≈ 30 mm and each SiPM is connected to 8 fibers on the top face plus one from the side face. WLS fibers are polished on both ends and mirrored by aluminium foil tape on the side opposite to SiPM. As photodetectors Hamamatsu S10931-05P SiPMs with 3 × 3 mm 2 sensitive area are used.

Test setup
The test setup is schematically presented on figure 3 and consist of four ordinary scintillation counters S 1 -S 4 with EMI 9813B PMT, four planes of MWPC with 120 × 120 mm 2 sensitive area -two pairs of X and Y planes, time setting counter C Fast and light tight box with a prototype.  The coincidence of four counters S 1 -S 4 generates a triggering signal for cosmic muon. Two counters are above (S 1 and S 2 with 100 × 160 mm 2 sensitive area) and two below (S 3 -S 4 with 95 × 90 mm 2 sensitive area) the prototype. MWPC with delay line readout and 4 mm step between wires generates two signals for each event. The time difference between two signals from one plane defines the coordinate of cosmic muon. To reconstruct the coordinates specific calibration events are generated with 1 Hz rate, firing three specific wires.
Time setting Cherenkov counter C Fast is a 40 mm×45 mm cylinder made of Plexiglas viewed by XP2020 PMT. Time resolution of C Fast counter is 175 ± 10 ps. The prototype is placed in the light tight box located between two blocks of X and Y planes of MWPC.
-3 -Electronics. The signal from each of four scintillating counter S 1 -S 4 comes to passive splitter. One part goes to ADC module through delay cable. The second part of signal goes to Constant Fraction Discriminator (CFD ORTEC 934 QUAD) with two ouput NIM signals: one output signal used by coincidence module and the second signal goes to TDC. The coincidence of signals from four scintillating counters S 1 -S 4 generates a triggering signal to start ADC and TDC modules and DAQ readout proccess. Signals from Fast Cherenkov counter C Fast and both SiPMs installed on CHOD prototype go to ADC and TDC modules through CFD (CAEN N253). Multihit TDC CAEN Mod.V1290N allows to select different combinations of time setting counters during offline analysis.

Results
For the time analysis intervals between signal from C Fast and signals from SiPMs measured by TDC were fitted with Gaussian. Fit parameters MEAN and SIGMA were used for further analysis as signal arrival time and time resolution respectively. Amplitude spectrum was fitted with Landau function and fit parameter MPV (peak value) in ADC channels was recalculated to the approximate number of photoelectrons using a calibration with LED. Linear track parametrization was used to retrieve the coordinate of cosmic muon on the prototype.
To compare results of measurements with WLS fibers and direct readout options the minimum GATE width parameter is introduced: GATE = ∆T + 5 × σ , where ∆T is a maximum signal arrival time spread while the cosmic muon hit point is moved over the prototype area, σ is intrinsic local time resolution. Within such a gate ≈ 98% of signals are contained. Different algorithm options of time measurements are considered: mean-time of the signals from SiPMs, first and last signals.

Fiber readout
For the WLS Fiber readout cartesian coordinates with two axis are used: along and across fibers.
Across fibers: on figures 4, 5 (left) the amplitude in number of detected photoelectrons and signal arrival time vs coordinate across fibers are presented. Both dependencies have oscillating character with a period ≈ 30 mm -the distance between fibers in one bundle. Amplitude dynamic range defined as A max /A min is about 1.10.
Along fibers: signal arrival time vs coordinate along the fibers is presented on figure 5 (right). The amplitudes vs coordinates along fibers for one SiPM presented on figure 4 (right). Signal arrival time depends on muon coordinate along fiber as 4.0 ± 1.5 ns/m and is close to the expected light speed in plastic with refractive index of n = 1.58. Maximum signal arrival time spread estimated as as 400 ± 150 ps. Amplitude of SiPM is decreasing with increasing distance to SiPM as ≈ 0.03 p.e./mm.
The final results for WLS Fiber readout are presented on the table 1. The best time resolution with dual SiPM readout achived using mean-time signal and is σ = 370 ± 10 ps.

Direct readout
For the direct readout there is a prefered direction -the diagonal between opposite corners with SiPMs. In this case the polar coordinate system is used with parameters R -distance from cosmic muon coordinate in prototype to SiPM and φ -angle to the diagonal between SiPMs.    For the direct readout the amplitude and signal arrival time strongly depend on the distance between cosmic muon hit point and SiPM. On figure 6 the amplitude vs coordinate position is presented using cartesian (left) and polar (right) coordinate systems. Amplitude reaches maximum close to the active SiPM (point [0, 0]) and has nearly the same minimum values at other corners of scintillator including the opposite corner with second SiPM [point 107,133]. The dynamic range is estimated as A max /A min ≈ 3. The signal arrival time versus position is presented on figure 7 for single SiPM readout and using mean-time of two SiPMs. The maximum signal arrival time spread for single SiPM is 2.7 ± 0.2 ns and is decreased to 1.6 ± 0.2 using mean-time of two SiPMs. Another possible option is to use the First or the Last signal. The final results are presented on table 2.