Sensitivity for detection of decay of dark matter particle using ICAL at INO
SSensitivity for detection of decay of dark matter particle usingICAL at INO
N. Dash ∗ , V. M. Datar † , and G. Majumder ‡ Nuclear Physics Division, Bhabha Atomic Research Centre, Mumbai - 400085, INDIA Homi Bhabha National Institute, Anushaktinagar, Mumbai - 400094, INDIA Tata Institute of Fundamental Research, Mumbai - 400005, INDIAOctober 6, 2018
Abstract
We report on the simulation studies on the possibility of dark matter particle (DMP) decaying intoleptonic modes. While not much is known about the properties of dark matter particles except throughtheir gravitational effect, it has been recently conjectured that the so called “anomalous Kolar Events”observed some decades ago may be due to the decay of unstable dark matter particles (M.V.N. Murthyand G.Rajasekaran, Pramana, , 609 (2014)). The aim of this study is to see if this conjecture can beverified at the proposed Iron Calorimeter (ICAL) detector at INO. We study the possible decay to leptonicmodes which may be seen in this detector with some modifications. For the purposes of simulation weassume that each channel saturates the decay width for the mass ranging from 1 −
50 GeV / c . The aimis not only to investigate the decay signatures, but also, more generally, to establish lower bounds on thelife time of DMP even if no such decay takes place. Index terms—
India-based Neutrino Observatory, Iron Calorimeter, Kolar Event, Dark Matter Particle,Life Time. ∗ [email protected] † [email protected] ‡ [email protected] a r X i v : . [ h e p - e x ] O c t Introduction
It is now established that the dark matter particles (DMPs) constitute 80 −
85% of all matter in the Universe,but we know very little about the properties of these particles. Even though they are non-luminous innature, their presence has been inferred through their gravitational effect. The presence of Dark matter wasinvestigated in detail by Vera Rubin using their kinematical study of the galaxies [1] during 1960 − Kolar events , may be due to the decay of unstable dark matter particles whose mass is inthe range of several GeV / c [10]. These events, discussed in Refs.[11, 12, 13] have neither been confirmednor shown to be spurious. They constitute 25% of the total recorded events over two decades at a depth of2.3 km. Immediately after the observation of these anomalous events, they were interpreted as the decays of ahypothetical heavy particle [14, 15, 16] with life time of the order of 10 − s and mass of around 2 − / c ,possibly produced in the interaction of neutrino or anti-neutrino with the rock surrounding the detector.However, no evidence for such a particle was found in other experiments including those using neutrinobeams [17, 18].While these anomalous events have evaded any conventional, standard model based, interpretation untilnow, the conjecture that the anomalous events are due to the decay of unstable dark matter opens up a2ew possibility. This may be established by future neutrino detectors, like the Iron Calorimeter (ICAL) atIndia-based Neutrino Observatory (INO) [19]. The sensitivity to dark matter decays may be enhanced byplacing detector elements on the walls and ceiling of the large cavern housing ICAL.In this paper we report on the simulation of the dark matter particle decaying into leptonic modes. Forthe purposes of simulation we assume that each channel saturates the width with the mass ranging from1 −
50 GeV / c . The aim is not only to investigate the conjecture, but also, more generally, to determine thelower limits on the life time of DMP by assuming their number to be with in a limited volume of the ICALcavern.The ICAL detector proposed to be placed in a cavern at INO which is much larger than that at KGF.It is planned to construct a 50 kton magnetised Iron Calorimeter under a rock cover of at least 1 km allaround. The main goal of the ICAL is to study atmospheric neutrinos ( ν µ ), in particular the mass hierarchyof neutrinos. However it can also be used for other purposes such as searching for magnetic monopole [20],to observe or set limits on the life time of DMP by considering their number in the ICAL cavern and alsofrom the annihilation or decay products of the DMP by looking at the centre of the solar system.The manuscript is organised as follows: in Sec.2, we briefly discuss the methods used to detect DMP. InSec.3, we focus on the proposed detector for DMP detection at ICAL cavern. In Sec.4 and 5, we present theanalysis and results of the simulation study for DMP decay in the ICAL cavern. We conclude the paper inSec.6. The search for DMP has been carried out both directly and indirectly. In the latter category one looksfor the missing mass in events in particle colliders measured in a 4 π detector assuming that can then beascribed to the production of DMP. In the former category the search involves the detection of nuclearrecoils resulting from DMPs colliding with the atoms in a detector. These detectors consist of rare gasessuch as argon or xenon in the liquid or gas phase, scintillators (NaI(Tl), CaF etc.), semiconductors (HPGeor Si) or cryogenic bolometers. In space based experiments one looks for the products of the annihilation ordecay of the DMP. The experiments looking for high energy neutrinos, try to observe distinctive high energyneutrinos obtained from the annihilation or decay of the DMPs at the centre of the Sun, the Earth and theGalactic halo. Because the large number of the dark matter particle is involved there. Here we introduceyet another method to look for the possible decays of DMP. Since the DMP is present everywhere, a deepunderground neutrino detector should be able to detect the possible decay of DMP provided the mass of3MP is in a suitable range and its decay products are easily detectable. If no such decay is observed, oneshould be able to put lower bounds on the possible partial life times after sufficiently long exposure. Aspointed out in Ref.[10], based on the analysis of Kolar events, a large cavern located deep underground asin the case of neutrino detectors should be able to identify the DMP decay if the life time is around the ageof the Universe. Either way this provides a novel way of putting limits on the DMP decay or its detectionwhich has not been compared before.We consider the last scenario in this paper to detect DMP using the ICAL detector at INO with theassumption that it decays to standard model particles. If not we obtain appropriate limits. For convenience,in this simulation, it is assumed that DMP is a neutral scalar particle (Φ DM ) that decays to lepton pairsonly and is guided by the results from different satellite based experiments. They have a significant excess inpositron to electron ratio above 10 GeV / c but not in the anti-proton to proton ratio. This mode of decay isalso most suitable method for detecting DMP using the ICAL detector. The ICAL is a sampling calorimeterand is especially suitable for tracking muons which may arise from DMP decays to µ + µ − pairs. Therefore,in general, we look for the decay modes of the typeΦ DM → (cid:96) + + (cid:96) − ( (cid:96) = e , µ, τ ) . (2.1)In order to detect such events at INO, a detector configuration with ICAL as the central detector andsome additional detectors around it is proposed. This is described in the next section. Dark matter is believed to be present everywhere. The biggest cavern at INO is the ICAL cavern havingdimensions of ∼
132 m ×
26 m ×
32 m and a cavern volume ∼ cm . As mentioned in Ref.[10], thislarge volume will lead to an increase in the number of detected DMP decay events which is around ∼ / Yrbased on the analysis of Kolar events. So to detect all the decay products in the form of visible particles, insimulation we have placed detectors on the 4 walls of the ICAL cavern.The ICAL is a magnetized calorimeter consisting of 150 layers of iron plates as the absorber each havinga thickness of 5 . (0.34%). The gas mixture and the appropriate high voltage allow RPC to operate it in avalanche mode.The passage of a charged particle through the detector ionizes the gas medium which induces a signal andis collected by honey-comb patterned pick-up panels, which are placed orthogonally on either side of the4etector. They gives X, Y co-ordinates of a hit point with a position resolution of 3 cm and time resolutionof 1 ns. The iron plates are magnetized with an average field of 1 . ×
16 m ×
15 m of the ICALcavern. In principle, the remaining space of the ICAL cavern can be used for additional DMP detectorinstallations. One such simple configuration is proposed in this paper.As the ICAL detector will be located towards one end of the cavern, four scintillator detectors (SDs) aremounted close to the walls of the cavern. A schematic diagram of the detector is as shown in Fig. 1. Asshown in the figure surfaces 1, 2, 3 and 4 depict the scintillator detectors. Two of these detectors are placedalong the length of the cavern with dimension of 132 m × .
04 m ×
32 m. The third one is placed inYZ plane with area of cross-section 26 m ×
32 m, having thickness of 4 cm. The fourth one is above theICAL surface at a height of 17 m from the top surface of the ICAL. This can also be used as a muon vetofor ICAL. In the simulation only one layer of detector is used for each surface. So no energy measurementis available from these detectors. On the other hand they provide the signals for identifying back-to-backleptonic decays of DMPs.ICAL uses around 30 ,
000 RPCs of dimensions 2 m × ,
000 of them would be needed for a single layer lining oneach plane. This is 10 times less than those required by ICAL.
The simulation is carried out in two regions separately. These two regions are specified by the air regioni.e. the gap between the ICAL and the scintillator planes and another one is inside the ICAL detector. Asmentioned in Sec.2, three decay channels of DMP are considered one at a time, with 100% branching ratio(B) for each channel. These are Φ DM → e + + e − (B = 1) , (4.1)Φ DM → µ + + µ − (B = 1) , (4.2)and Φ DM → τ + + τ − (B = 1) . (4.3)If a DMP decays to µ + µ − channel, it can be identified unambiguously in the ICAL provided its energyis 0.5 GeV or more, since there are two muon tracks back to back, which can be easily distinguished fromother events. But for the other two channels i.e. e and τ , some uncertainties arise. As τ is the heaviest5 xz32 meter 132 meter 26 meterSurface 1 Surface 2Surface 3Surface 40, 0, 0 m
16 m48 m
Figure 1: Schematic view of DMP detector with side detectors and the Iron Calorimeter.lepton with a short life time ( ∼ . × − sec), its decay products include hadrons, muon, electron andtheir corresponding neutrinos. As the ICAL detector will use 56 mm thick iron plates it will be difficultto separate electrons from π decays. Due to the non-relativistic speed of DMP which is almost at rest, itsdecay will be isotropic. For a given mass of the DMP the energies of the daughter particles are obtained bytwo body kinematics. Assuming the mass of the decay particle and anti-particle pair to be m, the mass ofthe DMP to be M the momentum of each of the daughter particles, p and p , is given by |−→ p | = |−→ p | = (cid:114) M − m . (4.4)Hence the DMP mass is taken as input to the simulation instead of the daughter particle energy. The massof the DMP, decaying into µ and e pairs, is varied from 1 GeV / c to 50 GeV / c in 1 GeV / c steps. As the τ mass is 1 . / c , the DMP mass is varied from 6 GeV / c to 50 GeV / c in this decay channel. In thesimulation two daughter particles start from a single vertex in opposite directions and with momenta givenin Eq.(4.4). To obtain isotropic flux during the decay process the zenith angle ( cosθ ) is smeared from 0 to6 and the azimuthal angle ( φ ) by 2 π . The charge of one of the daughter particles in a decay event is chosenrandomly and the other daughter particle has the opposite charge. The High Energy Physics (HEP) simulation tool-kit GEANT4 [21], is used to do the simulation for DMPdecay in the air region. In the simulation, the defined detector geometry is same as mentioned in Sec.2. Theevents are generated in the air gap i.e. between the side detectors and the ICAL detector. In a fraction ofthe simulated events the trajectories of the daughter particles are such that at least one of them enters theICAL detector and its partner hits the side detector. Thus it will be possible to measure the energy of atleast one decay product.In case the DMP decays to a pair of muons, one of them will give rise to a clean track in the ICALdetector and the other one will have hits in the side detector in the opposite direction. The backgroundfor such events will be the cosmic ray muon or a muon produced due to the interaction of neutrino withthe rock and detection in the ICAL detector. But this can be eliminated by using the timing informationfrom the detector. The genuine events are selected by considering the reconstructed momentum within ± τ pair, either it will have a bunch of hits or a clean muon track depending on itsdecay products in the detector. In this case the events produced due to the interaction of the neutrino with7 DMP Mass[GeV/c0 10 20 30 40 50 R ec on s t r u c t i on E ff i c i e n cy DMP Mass[GeV/c0 10 20 30 40 50 R ec on s t r u c t i on E ff i c i e n cy - m + m ] DMP Mass[GeV/c0 10 20 30 40 50 D e t ec t i on E ff i c i e n cy DMP Mass[GeV/c0 10 20 30 40 50 D e t ec t i on E ff i c i e n cy - m + m - t + t - e + e Figure 2: Left Panel: The reconstruction efficiency for DMP decays to µ + µ − . Right Panel: The detectionefficiency for DMP decays to e, µ , and τ channel in air region of the ICAL cavern as marked by differentmarkers inside the plot.the rock matter will be act as a background. So if the DMP vertex is near to the SD, then the backgroundarises from neutrino induced events. This can be identified and eliminated by using time of flight methodas shown in Fig. 3. It is obtained by measuring the distance between hit coordinates in the two SDs andtheir corresponding time difference. The cosmic muons originate from outside the ICAL cavern while thosethat arise from DMP decay are generated inside the ICAL cavern. By excluding a small region near theSD obtained from the simulation for this particular channel we exclude the cosmic muon events. Thiscorresponds to a region close to the edges of the region defined by the 2 delimiting lines in Fig. 3. A similarprocedure can be applied for the other two channels.The detection efficiencies are also obtained for DMP decays to e + e − and τ + τ − mode. They are obtainedby using the minimum number of hits in the ICAL detector and is represented in the right panel of theFig. 2. Even though it is difficult to identify e ± from hadrons, which mainly lead to showers, the detectionprobability grows with increasing mass of DMP and saturates beyond ∼
20 GeV / c . The detection efficiencyfor µ ± channel obtained in similar way is also shown in the same plot.8 istance[m]0 50 100 d t[ n sec ] -400-2000200400 distance[m]0 50 100 d t[ n sec ] -400-2000200400 m DMP
Figure 3: The distance between hit points verses time difference.
In this case all the events are generated inside the ICAL detector within a fiducial volume of 40 m ×
14 m ×
12 m.Inside the ICAL it is possible to measure the energy of two muons and hence the invariant mass.For DMP decays to µ + µ − the timing and trajectory information can be used to separate them frombackground like cosmic ray muon and neutrino. The Monte-Carlo technique is used to simulate DMPdecays, uniformly distributed with the fiducial volume of ICAL, and track the decay muons. For each energyand theta bin the momentum resolution and direction resolution are used separately for µ + and µ − frommuon look-up table. Figure 4 shows the detection efficiency for DMP decays to µ + µ − channel inside theICAL.The GEANT4 simulation tool-kit is used to do the simulation for DMP decays to τ + τ − channel. As τ decays to other particles in two- and three-body decay modes on a time scale of ∼ psec, it is easier to dothe simulation in GEANT4 instead of a simple Monte-Carlo simulation. Inside the ICAL the backgroundfor such events will be arise from neutral current (NC) interactions induced by neutrinos. However the maindifference between them is that in a NC event all the particles will be in one direction, whereas in DMPdecay the decay products will be on either side of a single vertex resulting in an increase in the time withhit position in both the directions. 9 DMP Mass[GeV/c0 10 20 30 40 50 R ec on s t r u c t i on E ff i c i e n cy DMP Mass[GeV/c0 10 20 30 40 50 R ec on s t r u c t i on E ff i c i e n cy - m + m Figure 4: The detection efficiency inside the ICAL for the DMP decays to µ + µ − pair. In the above two cases, we forced the particles to be generated in their respective regions. But to get thedetector acceptance, the events are generated uniformly whole over the ICAL cavern i.e. including the ICALdetector and the air region. Only the DMP decays to µ ± channel are used and the simulation is carriedout using GEANT4. Figure 5 shows the detection efficiency for 5 different situations. Bands with differentmarkers represent 5 different cases. The classification is based on their type of detection. The efficiency isobtained by taking the ratio between the numbers of events with hit in the respective detector to the totalnumber of simulated events.The cases II and IV are sensitive to DMP decay detection. The Case III is also sensitive if absorbers areplaced between SD and is relevant for low mass of DMP. The obtained efficiency in different regions from Sec.4 is used to estimate the life time of a DMP for a finitenumber of observed events using Frequentist Method [22]. If ρ (GeV/cm ) is the local dark matter density,V (cm ) is the detection volume, (cid:15) is the detection efficiency, B is the branching ratio, M (GeV/c ) is theDMP mass and R (Yr − ) is the rate of decay, then the life time T (Yr) of the DMP is given byT = ρ V (cid:15) BM R . (5.1)10 DMP Mass[GeV/c0 10 20 30 40 50 E ff i c i e n cy DMP Mass[GeV/c0 10 20 30 40 50 E ff i c i e n cy CASE ICASE IICASE IIICASE IVCASE V
Figure 5: The detector acceptance by considering the DMP decays to µ pair in the whole ICAL cavern. case I: Not detected, case
II: 2 or 1 in ICAL not in SD, case
III: 2 in SD and not in ICAL, case
IV: 1 inICAL and 2nd one in SD, and case
V: 1 in SD and another one is not detected.Here and for further calculation the local dark matter density is taken as 0 .
39 GeV / cm [22]. The limit onthe life time is obtained by considering the upper limit for the 0 observed events with 90% confidence levelfor 0 background in 1 year of detector running time using Eq.(5.1). The obtained limit for different regionswith mass is shown in the left panel of Fig. 6 by different symbols. With increase in the mass of DMP, thelimit on the life time of the DMP is decreasing due to the decrease in number density in a fixed volume.From Super-Kamiokande (SK) [8] results there is a stringent limit on the life time of the dark matter particledecaying to µ ± pair obtained from ν µ signal starting with mass 10 GeV / c . For 10 GeV / c DMP mass thelife time is of the order of 10 Gyr as obtained by SK. The ICAL can also put bound on the DMP life timein the similar way as followed by the cherenkov detectors and will be comparable to them.Alternatively, using the higher limit of DMP life time (2 Gyr) from the above result and the local darkmatter density, the number of expected events due to the decay of DMP is obtained separately for eventssimulated in air region between cavern wall and ICAL detector, inside the ICAL detector and the whole ofthe ICAL cavern by combining the first two results. The expression for expected number of events (N ex )due to the decay of DMP is given by N ex = ρ V (cid:15) BM T . (5.2)The right panel of the Fig. 6 shows the number of expected events for 10 years of counting for the 3 cases11entioned above. ] DMP Mass[GeV/c0 10 20 30 40 50 L i f e T i m e [ G y r ] In ICALIn AirICAL Cavern ] DMP Mass[GeV/c0 10 20 30 40 50 ex N In ICALIn AirICAL Cavern
Figure 6: Left Panel: The lower limit in the life time of DMP verses its mass for µ + µ − decay channel. RightPanel: The number of expected events due to the DMP decays to µ ± with life time 2 Gyr at ICAL cavernfor 10 yrs of detector running period. If the reinterpretation of the Kolar events as being due to the decay of dark matter particles is correct,it should be possible to observe them with much larger numbers in the ICAL cavern at the India-basedNeutrino Observatory. Due to the larger volume of the ICAL cavern and the size of the ICAL detector, theexpected number of such events is doubled compared to those observed at Kolar Gold Fields. In additionto addressing the issues towards understanding of the anomalous Kolar events, a non-observation of suchevents at INO will be able to place bounds on the life time of the dark matter particles with mass between1 GeV / c and 50 GeV / c . In another way it is estimating limit on the life time of the DMP with mass below10 GeV / c , however Super-Kamiokande has put limit from 10 GeV / c to few TeV. Acknowledgements
We would like to thank Prof. M.V.N. Murthy for his critical comments and suggestions. We also wish tothank Prof. D. Indumathi and Prof. G.Rajasekaran for their interest in the work reported in this paper.12e are also grateful to Prof. Pijushpani Bhattacharjee and Prof. Nita Sinha for their valuable commentsand suggestion during the preparation of the manuscript. Our thanks are also to the INO collaborators fortheir invaluable support.