First neutrino event detection with nuclear emulsion at J-PARC neutrino beamline

ν e interactions, as a possible interpretation of the existence of a sterile neutrino, has been observed in such an energy region. The nuclear emulsion technique can measure all the ﬁnal state particles with low energy threshold for a variety of targets (Fe, C, H 2 O, and so on). Its sub- μ m position resolution allows measurements of the ν e cross-section with good electron/gamma separation capability. We started a new experiment at J-PARC to study sub-multi-GeV neutrino interactions by introducing the nuclear emulsion technique. The J-PARC T60 experiment has been implemented as a ﬁrst step in such a project. Systematic neutrino event analysis with full scanning data in the nuclear emulsion detector was performed for the ﬁrst time. The ﬁrst neutrino event detection and its analysis are described in this paper.


Introduction
Currently, many experimental neutrino oscillation projects are running, or planned, to search for CP violation, and/or to probe the neutrino mass hierarchy [1][2][3][4][5][6][7][8][9]. Precise cross-section measurements of neutrino-nucleus interactions in the sub-multi-GeV region are important to reduce the systematic uncertainty, particularly for the neutrino mixing angle θ 23 , in these experiments. Furthermore, one possible explanation of the MiniBooNE anomaly [10] in this energy region is a result of the existence PTEP 2017, 063C02 T. Fukuda et al. of the so-called sterile neutrino indicated by LSND [11]; precise cross-section measurements are thus also important in verifying the MiniBooNE anomaly.
Nuclear emulsion is a three-dimensional solid tracking detector with sub-micron positional resolution. Thanks to its high spatial resolution, nuclear emulsion technology contributed to the discovery of the pion [12], of the charmed particle in cosmic rays [13], the direct observation of ν τ [14], the discovery of the ν τ appearance in a ν μ beam [15], and so on. This high spatial resolution allows measurements of all the final state particles with a low energy threshold. Nuclear emulsion has 4π track detection capability [16,17]. This is also very useful to investigate the event topology in lowenergy neutrino interactions. Furthermore, good e/γ separation capability allows measurement of ν e charged current interactions with a strong suppression of the background from ν μ neutral current interactions with π 0 production. There is much flexibility for target material selection because the detector, a so-called emulsion cloud chamber (ECC), is constructed as a sandwich structure of thin nuclear emulsion films and the target materials.
The J-PARC T60 experiment [18,19] was proposed to study the feasibility and check the detector performance of nuclear emulsion at the J-PARC neutrino beamline. The first results of beam exposure in 2015 will be presented in this paper.
Accelerator-based neutrino experiments with nuclear emulsion were established from the end of 1970s [20][21][22]. The experimental concept is that of a hybrid analysis with the emulsion and the electronic detectors, a so-called "emulsion-counter hybrid experiment." Only a small area of emulsions that is predicted by the electronic detectors has been analyzed, or only a small number of tracks that are predicted have been traced back because of the limitation of the track readout speed in the emulsions. A high-speed scanning system for the emulsion has been developed [23][24][25][26][27][28], and track readout for the entire area in the emulsion has recently become available. This allows one to analyze the emulsions independently of the prediction from the electronic detectors. In this paper, the first systematic neutrino event analysis from scanning all the data in the nuclear emulsion will also be described.

Neutrino detector 2.1. Nuclear emulsion films
The nuclear emulsion gel was produced in the facility at Nagoya University. In this emulsion gel production, high-sensitivity emulsion gel, containing 55% AgBr crystals by volume ratio, was produced [29]. Two emulsion layers, each 50 μm thick, were formed on both faces of a 180 μm-thick polystyrene base. The size of sheets was 100 mm × 125 mm. The long-term variation of the emulsion sensitivity and background noise was investigated using cut pieces of the emulsion sheets. The emulsion sensitivity was measured as grain density (GD), the number of grains per 100 μm along a minimum ionizing particle (MIP) trajectory. The noise of the emulsion was measured as fog density (FD), the number of randomly distributed grains per 1000 μm 3 in the emulsion. If an emulsion is left undeveloped for a long time after track recording, the silver atoms tend to oxidize into silver ions, owing to the presence of water and oxygen in the material. Some of the latent image cores are destroyed and fewer grains are produced in the development process. This effect is called fading. Variations of GD and FD over 133 days are shown in Fig. 1 under four different temperature and humidity conditions. We found that GD and FD remain at safe levels even after 133 days at 23 • C.

Emulsion Cloud Chamber
The ECC is constituted by alternatively stacking 41 emulsion films and 40 stainless steel plates (SUS304, 500 μm thickness), as shown in Fig. 2. These are inserted in a paper box and then vacuumpacked with a light-shield aluminum laminated bag in 40% relative humidity conditions. The total target mass is approximately 2.2 kg (1.2 kg inside the fiducial volume).

Emulsion shifter
Precise tracking detectors with timing information, e.g. a silicon microstrip detector or scintillating fiber detector, have been applied to connect tracks between the main emulsion detector and the electronic detectors, which have muon-identification capability in the conventional emulsion counter hybrid experiment. In the present work, an emulsion device with timing information played  this role. The emulsion detector intrinsically has no timing information. Therefore, the timing information is produced from the emulsion tracks by track coincidence between the films on a mechanically controlled moving stage [30,31], the so-called "emulsion shifter." An emulsion shifter ( Fig. 3) was reused from a balloon experiment with nuclear emulsion [32]. This equipment gives a time-stamp to the emulsion track by using a clock-based multi-stage emulsion shifter technique. Three stages play the roles of the hour, minute, and second hands of a watch, respectively. Emulsion tracks that are coincident on all three stages obtain a time-stamp corresponding to each stage position. The timing resolution of this emulsion shifter is approximately 7.9 s, which was investigated using tracks passing through the emulsion shifter and electronic detector located at the back of the emulsion detectors. The details for this equipment and the analysis are described in Refs. [32,33].

Interactive Neutrino Grid (INGRID)
INGRID is one of the near detectors for the T2K experiment [34]. The tracking device is a 1 cmthickness plastic scintillator (5 cm width) read out with wavelength-shifting fibers. The constitution of the detector is a sandwich structure of the tracking planes and 6.5 cm-thickness iron blocks. The detector was assembled from a total of nine iron layers (58.5 cm) and 11 vertical and horizontal tracking planes. The size of a module is 1.2 m × 1.2 m × 0.9 m. INGRID consists of 16 identical modules arranged in horizontal and vertical arrays around the beam center. The original purpose of INGRID was to measure the neutrino beam direction and verify the number of neutrino interactions per protons on target. In the J-PARC T60 experiment, INGRID is utilized to identify the muons from neutrino-nucleus interactions in the ECC. As shown in Fig. 4, the emulsion shifter is placed between the ECC and INGRID. The particles that cross the emulsion shifter are supplied with timestamps. The tracks are then matched between the emulsion detector and INGRID using the timing information.

Installation at J-PARC
The emulsion detectors and the assembly frame were prepared at Toho University. These detectors were transported to J-PARC before the installation. The emulsion detectors were installed in front of an INGRID module adjacent to the on-axis module of the Neutrino Monitor Building at J-PARC on 14 January 2015, as shown in Fig. 5. The operation of the emulsion shifter has been monitored by verifying the stage reproducibility, which was sent to Kobe University with a monitoring laptop PC.

Neutrino beam exposure
The neutrino beam exposure started on 15 January 2015. Unfortunately, beam exposure was stopped on 16 January, then restarted from 25 February and finished on 1 April. The beam exposure was performed with the anti-neutrino beam mode after the restart of beam exposure. The accumulated number of protons on target is approximately 13.8 × 10 19 . The ratio of neutrino to anti-neutrino interactions is ∼ 1 : 3 because of the contamination of neutrinos in the anti-neutrino beam. The total number of neutrino or anti-neutrino events is estimated as approximately 44 events (fiducial volume: 24 events, including two events at the region of the emulsion layer). After beam exposure, the stages of the emulsion shifter were moved to a reference position and reference cosmic-ray 5   tracks were accumulated for one week. The emulsion detectors were uninstalled on 8 April and all films were immediately developed at Toho University. The temperature at the experimental site was approximately 22 • C during the exposure, as shown in Fig. 6. Therefore, the tracks that were recorded in the emulsion films were kept in good condition, as shown in Fig. 1. In operation, the emulsion shifter performed well [33].

Event analysis 4.1. Scanning
The entire areas of all the emulsion films for the experiment (∼1.2 m 2 for the two emulsion layers of all the emulsion films, including six films for the emulsion shifter) were scanned using a high-speed automated microscope system, Hyper Track Selector (HTS) [28] (Fig. 7). This system was developed at Nagoya University to quickly analyze particle tracks in emulsion films, and allows tracks to be read out at a scanning speed of ∼5000 cm 2 h −1 . HTS takes 16 tomographic images in an emulsion layer and recognizes a series of grains on a straight line as a track by comparing these images. The   slope acceptance of HTS was set at | tan θ| < 1.7 (|θ | 60 • ), where θ is the track angle with respect to the perpendicular of the emulsion film (the z-axis).

Analysis in ECC 4.2.1. Track reconstruction
First, track segments, so-called "micro tracks," were detected on each layer of the emulsion films (Fig. 8). The positions (x, y), slopes (tan θ x , tan θ y ), number of hits in an emulsion layer, pulse height (PH), number of pixels of constructed track data, and volume pulse height (VPH) of the micro tracks were measured by HTS. The number of grains belonging to the track (blackness) is well correlated with the PH and the VPH. As shown in Fig. 8, the tracks connecting the positions of micro tracks on both sides of the plastic base are called "base tracks." Sub-MeV or few-MeV electrons from the environment are recorded in the emulsion because the nuclear emulsion records all tracks from immediately after production until development with no dead time. Therefore, the automated scanning system recognizes "fake" tracks made of low-energy electrons and/or random noise grains (fog). The chance coincidences of these fake tracks become the background of base tracks. However, such background base tracks can be rejected by evaluating their blackness (VPH) and linearity (the angle difference between base track and micro tracks) [35].  The positions (x, y) and slopes (tan θ x , tan θ y ) of the base tracks were measured with respect to the microscope's coordinate system. After connecting the base tracks among the emulsion films, the rotation, slant, parallel translation, and distance between every pair of adjacent emulsion films are adjusted, so that the differences in the track positions and slopes between the two films are minimized after the reconstruction procedure. These connection elements of base tracks are called "linklets." All base tracks in all films are reconstructed by tracking linklets, as a chain. Finally, the track density in a film is ∼4000 tracks cm −2 . The track efficiency after reconstruction was evaluated at each measured angle for all plates, as shown in Fig. 10. In the small-angle region, the track efficiency is approximately 96%. However, in the large-angle region, the loss of efficiency is visible.

Proton identification
The momentum of charged particles can be estimated in the ECC by measuring their multiple Coulomb scattering (MCS) [36]. When a particle of charge z, momentum p, and velocity βc traverses a material of depth x and radiation length X 0 , the deviation of the distribution of the scattering angle is given by:  The GD is nearly proportional to its energy loss dE/dx in nuclear emulsion [37]. The average number of hit layers (PH) or the average number of pixels associated with the reconstructed track (VPH) in scanned track data corresponds to the GD of tracks [38]. In Fig. 11, the measured momentum and the average VPH of each track that is connected to more than 10 segments is shown. The proton-like tracks are found in the low-momentum and high-dE/dx region.

Track selection
The reconstructed tracks are selected to search for neutrino event vertices as shown in the following steps: (1) Fiducial area selection: The acceptance of the average track angle is less than | tan θ | = 1.6 and the searched position area is set at x: 10-120 mm and y: 6-76 mm to avoid the bad film condition area (thin thickness of emulsion layer). The plate IDs of the most upstream and downstream plates in the neutrino beam direction are PL41 and PL01, respectively. The number of tracks and the number of tracks that just start and connect to downstream plates after fiducial selections at each plate are shown in Fig. 12(a). As shown by the red line in Fig. 12(a), there are many tracks upstream of the detector that are identified as penetrating tracks (cosmic ray muons or muons produced from outside of the detector). from the most upstream plate to its upstream plate using their position and angle is out of the fiducial area. The number of tracks and the number of starting tracks after veto selections at each plate is shown in Fig. 12(b). The angle distribution of penetrate and edge out tracks is shown in Fig. 13(a). This is consistent with cosmic rays from the sky. Figure 13  selected to avoid accidentally disconnecting tracks in the large-angle region and low reliability tracks. Finally, 41087 tracks are selected as the candidate tracks from neutrino interactions. The number of tracks and the number of starting tracks after these selections at each plate are shown in Fig. 12(c). Additionally, the track angle distribution after this selection is shown in Fig. 13(c).

Event selection
The event selection is performed with the following procedure. The events are then categorized into types (a)-(h), as shown in Fig. 14.  Single track vertices are mainly considered to be disconnecting of low-energy cosmic ray muons or from electron/positron pair production, as shown in Fig. 13(c). (3) Track blackness: "Two track" vertices are categorized according to track blackness. Black tracks are defined as tracks whose average VPH is over 120, and minimum ionizing particle (MIP) tracks are defined as tracks whose average VPH is less than 120, as shown in Fig. 11. When two-track vertices are reconstructed by only black tracks, they belong to Category (b). In Category (    According to the comparison between real and background data, an excess of small opening angle is found in the real data. They are almost parallel tracks, so are explained as electron/positron pair production from gamma-rays. Gamma-ray candidate events are selected by using their opening angle and minimum distance of vertex tracks as shown in Fig. 16(c). Thus, 125 vertices are selected as gamma-ray events and the estimated number of background is 0.5, as shown in Fig. 16(d). Gamma-ray events are stored in Category (g), and the other vertices are stored in Category (h).   On the other hand, there may be some neutrino events in Category (d) because the number of vertex signals is 5.2 (= 16 − 10.8). The near vertices in Category (h) are accidental chance coincidence of misconnected cosmic-ray muons or from an electron/positron pair. It is considered that they are the same group as Category (c).

Gamma-ray events
In this section, Category (g), or gamma-ray candidate events, will be discussed. Only forward-going gamma-rays are reconstructed in this category. The angle distribution of gamma-ray candidate events 14  is shown in Fig. 17(a). The incident direction of tracks from gamma-rays is the forward and upper side, and it is consistent with the direction of cosmic-rays [ Fig. 17(b)]. Additionally, the number of converted gamma-rays is high in the upstream region of the detector, as shown in Fig. 18. This is evidence that the gamma-rays come from outside the detector. The detected electron/positron pairs from gamma-rays in the ECC are shown in Fig. 19. It is considered that cosmic gamma-rays from the backward and upper side directions make the single-track vertex in Category (c) and background vertices.

Black track vertex events
In this section, a discussion of the 18 vertices found that are reconstructed by only black tracks in Category (b) will be presented. The chance coincidence background is estimated as 0.5 events. Therefore, almost all the events are real vertices. These events are considered as neutron interactions 15   or part of neutrino neutral current interactions with no MIP tracks. Additional short or backward black tracks attached to these vertices were also searched for. As shown in Table 1, new black tracks were found in more than half of the events. The event feature is shown in Fig. 20. The emission angle and range distributions are shown in Fig. 21 5  0  3  2  13  iron  2  1  0  3  14  iron  3  2  0  4  23  iron  2  2  0  5  26  iron  1  7  0  6  26  iron  1  3  1  7  27  emulsion  2  2  1  8  28  emulsion  3  7  0  9 30 criteria, the existence of two black tracks that are emitted to the forward direction is required. Thus, this becomes one of the biases to drop neutron events from the backward direction.

Neutrino candidate events
The other converged tracks, for example low-energy tracks, tracks emitted to the backward direction, or electron/positron pairs from vertices, are searched for in the 12 vertices found in Categories (a), (e), and (f). The detected particle momenta are then measured with MCS and the timing information is obtained with the emulsion shifter to connect the INGRID and identify muons [33]. In this section, the event summary is shown in Table 2 and then some of the detailed event features are described. The position distribution of 12 vertices is shown in Fig. 22. The impact parameter (IP) and emission angle for each MIP and black tracks is also shown in Fig. 23.

Event 7
The vertex is in the emulsion region at plate 27 and its vertex point is (x, y) = (96034.4 μm, 74184.9 μm) in the ECC coordinate system (the origin (0.0 μm, 0.0 μm) is the corner at PL01). One electron/positron pair from a gamma-ray in Category (g) points to this vertex point. The detailed track information is listed in Table 3. There are two MIP tracks and two black tracks in the vertex [ Fig. 24(a)]. Also, this event was created at the emulsion region, so we could see more detail 18   with a microscope. As shown in Fig. 24  ranges in INGRID. The momentum of the muon candidate track is also enough to pass through INGRID.

Event 3
The vertex is the steel region upstream of plate 14, and its vertex point is (x, y) = (93105.6 μm, 48182.9 μm) in the ECC coordinate system. There are three MIP tracks and one black track at the vertex. The details of the event tracks are listed in Table 4. This event is also matched between the emulsion detector and INGRID with the timing information, as shown in Figs. 28 and 29. The measured track momenta are consistent with the measured range in INGRID. Unfortunately, since two tracks penetrate through INGRID, it is difficult to decide which is the muon track from the information from INGRID.

Improvement of the systematic analysis
Further neutrino events are going to be searched for in Categories (c), (d), and (h). It is important to reject many fake starting tracks to find new neutrino events. To do that, the energy threshold to connect tracks between emulsion films will be lowered. Another method is to reconstruct the electron/positron pairs of gamma-rays and neutron-interacted vertices from the backward direction. Then, the search for additional much larger-angle black tracks will also be effective in identifying the interaction point with large-angle track scanning [16,17]. Furthermore, when some tracks reach the emulsion shifter, it will be possible to take the coincidence with timing information for background rejection.

Future experimental plan
In 2016, an anti-neutrino beam was exposed to a 60 kg steel target, and ECC and film scanning was done. Currently, the neutrino event analysis for this experiment is in progress to investigate the detector performance with higher statistics. Approximately 3000 anti-neutrino interaction events are 21/23 PTEP 2017, 063C02 T. Fukuda et al. expected to accumulate in this ECC. Moreover, not only anti-muon neutrino events but also antielectron neutrino events will be detected. Furthermore, the neutrino analysis will proceed in a few-kg water-target emulsion chamber that is assembled with vacuum-packed emulsion films and frametype spacers to fill the water as a target. The status and results of these analyses will be reported in the future. Then, a large-scale water-target emulsion chamber will be constructed to study the neutrino-water interactions in detail.

Conclusion
We proposed a new experimental project with a nuclear emulsion detector to study low-energy neutrino-nucleus interactions in detail for future neutrino oscillation physics experiments. The J-PARC T60 experiment has been implemented as a first step in this program to check the feasibility and confirm the neutrino event analysis. In this work, systematic neutrino event analysis with full ECC scanning data has been developed, and the first neutrino event detection has been successfully demonstrated. Moreover, hybrid analysis with INGRID has been performed for the first time.
We will polish the systematic neutrino analysis chain, develop the water-target emulsion chamber, and steadily expand the detector scale to reach the desired physics output.