WorksPhysics and Astronomy
Search for Light Neutral Bosons in the TREK/E36 Experiment at J-PARC
Abstract
The Standard Model (SM) represents our best description of the subatomic world and it has been very successful in explaining how elementary particles interact under the influence of the fundamental forces. Despite its far reaching success in describing the building blocks of matter, the SM is still incomplete; falling short to explain dark matter, baryogenesis, neutrino masses and much more. The E36 experiment conducted at J-PARC in Japan, allows for sensitivity to search for light gauge bosons, in the muonic decay channel. Such bosons could be associated with dark matter or explain established muon-related anomalies such as the muon value, and the proton radius puzzle. A scintillating fiber target was used to stop a beam of positively charged mesons. The products were detected with a large-acceptance toroidal spectrometer capable of tracking charged particles with high resolution, combined with a large solid angle CsI(TI) photon detector and particle ID systems. A realistic simulation was employed to search for these rare decays in the mass range of 20100 MeV/. Preliminary results of the upper limits for the branching ratio extracted at 95% CL, will be discussed.
Light Neutral Boson Search with the TREK/E36 Detector Apparatus
TREK/E36 was conducted at the K1.1BR kaon beamline in the hadron hall facility at J-PARC using a stopped beam in conjunction with a 12-sector superconducting toroidal spectrometer [13, 14]. Figure 1 shows the end and side view of the E36 detector apparatus. Details of the experimental method and detector configuration have been described in Ref. [15].

Figure 1. TREK/E36 apparatus. General end and side views of the detector system.
The incoming was tagged with a Fitch Čerenkov detector [13] in order to distinguish it from a , before being slowed down by a BeO degrader and stopped in the active target (TGT). Charged decay products were tracked by using the spiral fiber tracker (SFT) [16] which surrounded the target, and three multi-wire proportional chamber (C2, C3, C4) in each spectrometer sector. The C2, and C3/C4 were placed at the entrance and exit of the magnet gaps, respectively. To suppress mis-identification, redundancy was introduced into the particle identification detectors which was provided by time of flight detectors (TOF1, TOF2), aerogel Čerenkov, and lead glass Čerenkov counters (PGC). A highly segmented and large acceptance CsI(Tl) calorimeter barrel which consisted of 768 crystals and covered about 70% of [17] was used to identify and correct for structure dependent (SD) background events and also to search for light neutral bosons . The search necessitated that a is tracked in the gap, whose primary vertex lies within the fiducial volume of the target (good gap event) and two-clusters in the CsI calorimeter to detect the pair, along with at least 3 TOF1 counters that registered a charged particle hit.
Analysis and Preliminary Results
The aforementioned signal search required a TOF1 multiplicity of of at least 3, implying that three charged particles have passed through the TOF1 counters. Specifically the is tracked in the magnet gap and the form clusters in the CsI(Tl) calorimeter, and an invariant mass distribution was reconstructed from the clusters. A Geant4 Monte Carlo was used to generate and reconstruct mass , whose reconstructed widths, corresponding to , were used as the search window in the spectrum, shown in Figure 2 [18].
Figure 2. MC reconstructed invariant mass vs. generated mass. The horizontal bars represent a search window for a given mass. [18].
This analysis considered masses in the range of 20-110 MeV/, in incremental step-sizes of 5 MeV/. Several cut conditions in addition to the TOF1 multiplicity were employed in order to suppress the reducible background:
i. a momentum endpoint cut as a function of the generated
ii. a correlated angle cut between fired TOF1 counter and corresponding CsI(Tl) cluster
iii. a muon mass squared cut for the charged particle in the gap.
A polynomial fit to the resulting invariant mass data can be seen in Figure 4 below. The bottom panel of Figure 3 shows the residual, which is defined as the relative difference between data and fit, normalized to the data.
Figure 3. Invariant mass distribution with fit to data. The normalized residual is shown in the bottom plot.
Figure 4. Upper limit extraction (a) as a function of reconstructed invariant mass . Data (blue) and theory comparisons (b) on particles with scalar and pseudoscalar (squares), and polar and axial vector (stars) couplings. Particles with both polar and axial vector couplings appear to be excluded for masses below 80 MeV/.
The fit function from Figure 3 and the total number of stopped kaons were used to extract the upper limits. The total number of stopped in the fiducial region of the active target was computed as follows
where is the number of tracked data candidates satisfying the decay momentum, is the acceptance fraction of the muons evaluated with the MC simulations and is the nominal branching ratio of the decay mode. Upper limits at 95% CL on the branching ratio of for each mass corresponding to a limit for not observing the was computed for each aforementioned mass and corresponding search window of the invariant spectrum using the following relation
where is the number of candidate events within a given search window after applying all cuts to suppress the reducible background, is the fractional acceptance of the signal process , , and is the livetime fraction for candidate events. The upper limit as a function of the invariant mass is shown in Figure 6 (a). With of the data analyzed, the upper limit was found to be at .
Experimental limits for the Carlson model for new light neutral bosons with polar (scalar) and axial vector (pseudoscalar) couplings are . Whereas Figure 6 (b), shows the comparison between data and theory prediction of particles with scalar and pseudoscalar (green), and polar and axial vector (orange) couplings [7]. An particle with both polar and axial vector couplings, appears to be already excluded for masses below 80 MeV/ with only of the data analyzed. However, an with scalar and pseudoscalar couplings is not yet excluded. With further development in reducible background suppression, and more data analysis, we are hoping to enhance the sensitivity and experimental reach. Presently PID cuts to suppress mis-identification have not been applied. These cuts would further reduce the reducible background, and would enhance our sensitivity to particles with scalar and pseudoscalar couplings.
Acknowledgements
This work has been supported in part by NSERC and NRC in Canada, the KAKEN-HI Grants-in-Aid for Scientific Research in Japan, the DOE and NSF in the US, and in Russia by the Ministry of Science and Technology.
Prepared by LLNL under Contract DE-AC52–07NA23744.
References
- [7]Carlson C E and Rislow B C 2012 Phys. Rev. D86 035013 (Preprint 1206.3587)
- [13]Abe M et al. 2003 Nucl. Instrum. Meth. A506 60–91 (Preprint hep-ex/0302001)
- [14]Abe M et al. 2006 Phys. Rev. D73 072005
- [15]Ito H et al. 2022 Phys. Lett. B 826 136913 (Preprint 2107.03583)
- [16]Mineev O et al. 2017 Nucl. Instrum. Meth. A847 136–141 (Preprint 1607.00219)
- [17]Dementev D V et al. 2000 Nucl. Instrum. Meth. A440 151–171
- [18]Dongwi H D 2020 Search for Light Neutral Bosons in the TREK/E36 Experiment at J-PARC PhD dissertation Hampton University


