Comparison of the measured atmospheric muon flux with Monte Carlo simulations for the first KM3NeT detection units
CComparison of the measured atmospheric muonflux with Monte Carlo simulations for the firstKM3NeT detection units
The KM3NeT Collaboration ‡* ‡ [email protected] The KM3NeT Collaboration has successfully deployed its first detection units in the Mediter-ranean Sea in December 2015 (ARCA) and September 2017 (ORCA). The sample of datacollected between September 2016 and March 2017 has been used to measure the atmosphericmuon flux at two different depths under the sea level: about 3.5 km with ARCA and about 2.5 kmwith ORCA. The atmospheric muon flux represents an abundant signal for a neutrino telescopeand can be used to test the reliability of the Monte Carlo simulation chain. In this work, themeasurements are compared to Monte Carlo simulations based on MUPAGE and CORSIKAcodes. Measured events and simulated events are treated using the same approach, making thecomparison reliable.
Corresponding authors:
Piotr Kalaczynski †1 , Rosa Coniglione National Centre for Nuclear Research, 00-681 Warsaw, Poland Instituto Nazionale Di Fisica Nucleare - Laboratori Nazionali del Sud, Via S. Sofia 62 - 95123Catania, Italy36th International Cosmic Ray Conference 201925.07.2019 - 1.08.2019Madison, Wisconsin, USA * for collaboration list see PoS(ICRC2019)1177 † Presenter. © Copyright owned by the author(s) under the terms of the Creative CommonsAttribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND 4.0). http://pos.sissa.it/ a r X i v : . [ h e p - e x ] J a n easured and simulated atmospheric muons for the first KM3NeT detectors Piotr Kalaczynski
1. Introduction
KM3NeT is a research infrastructure being constructed at the bottom of the MediterraneanSea. It consists of two neutrino detectors: ARCA (Astroparticle Research with Cosmics in theAbyss) located off-shore Portopalo di Capo Passero, Sicily, Italy, at a depth of 3500 m and ORCA(Oscillation Research with Cosmics in the Abyss) off-shore Toulon, France, at a depth of 2450 m.The main purpose of the ARCA telescope is the detection of TeV-PeV neutrinos from astro-physical sources or in coincidence with other high energy events, for example gravitational waves,gamma ray bursts, blazar flares. The ORCA telescope is designed to study the neutrino mass hier-archy (NMH), using oscillations of the atmospheric neutrinos in the GeV range [1].Both KM3NeT detectors are based on the same technology [1]: vertically aligned detectionunits (DUs), each carrying 18 digital optical modules (DOMs). Each DOM contains 31 3-inchphotomultiplier tubes (PMTs), calibration and positioning instrumentation and readout electronicsboards. The difference between ARCA and ORCA is the horizontal (90 m and 20 m respectively)and vertical (36 m and 9 m respectively) spacing between the DOMs, which is optimised to searchfor high- and low-energy neutrinos, respectively. The first DUs are already taking data at both theARCA and ORCA sites. Atmospheric muons represent an important background to the physicsanalyses in a neutrino telescope, but they are also a useful tool to test the performance of thedetector and validity of the simulations. In this paper a comparison of the first data collected so farby ARCA and ORCA to the Monte Carlo (MC) simulations is presented.
2. Simulation chain
The KM3NeT simulation chain shown in Figure 1 has a modular structure. The generation ofthe atmospheric muon bundles is performed using two different codes: MUPAGE [2] and COR-SIKA [3]. MUPAGE is a fast, parametric simulation code developed for the ANTARES experiment[4] that generates muon bundles, induced by cosmic rays impinging the Earth atmosphere, at differ-ent undersea depths and different zenith angles, on the surface of the active volume of the detector.CORSIKA (COsmic Ray SImulations for KAscade) is a software package that simulates the inter-actions of the primary cosmic ray (CR) nuclei in the upper layers of the atmosphere and follows thedevelopment of the shower to a specified observation level (in this work it is the sea level). Theirenergy spectrum is simulated following a power law, E − γ , where the spectral index γ is chosenby the user. The CORSIKA package is flexible and allows the user to choose between differentmodels to describe the primary interactions and the composition of the CR primary flux.A dedicated software, called "Corant", has been developed to convert the CORSIKA outputformat to the KM3NeT standard format of events and to evaluate the weights to be applied toeach event, according to the model of the CR composition chosen by the user. The propagation ofthe muon bundles to the active volume of the detector is performed using a 3-dimensional muonpropagator contained in the gSeaGen [6] library, a GENIE-based [5] package designed mainly tosimulate neutrino interactions. After the muon bundle propagation, the Cherenkov photons emittedalong the path of muons and the probability of their detection by the DOMs are simulated with acustom application developed for KM3NeT using multidimensional interpolation tables [7], called"JSirene". The next step is the creation of a simulated data stream, similar to the data stream coming2 easured and simulated atmospheric muons for the first KM3NeT detectors Piotr Kalaczynski
Figure 1: Schematic drawing of the KM3NeT atmospheric muon simulation chain. Programs arecolour-coded, corresponding to the physics event generator for which they are used. The MUPAGEchain is marked in red and the CORSIKA chain in blue. The software is grouped according to thesimulation stage it belongs to.from the detector. This is done with the addition of the environmental optical background, due tothe bioluminescence and to the K decay, with the simulation of the detector response, taking intoaccount the front-end electronics behaviour. The same trigger algorithms used for real data are usedto identify possible interesting events in the simulated sample [8]. Finally, the simulated eventsare processed with the same reconstruction programs used for the real data stream. At this stagethe simulated MC events can be easily compared to the calibrated data. A track reconstructionalgorithm is applied both to data and MC events. In this work only a track reconstruction code("JGandalf" [9]) is considered, which is specifically designed to reconstruct atmospheric muonsand muons induced by neutrinos.
3. Data vs MC comparisons
A sample of data collected with 2 DUs of the ARCA detector (ARCA2) from the 23th ofDecember 2016 to the 2nd of March 2017 and with 1 DU of the ORCA detector (ORCA1) fromthe 28th of September 2017 to the 13th of December 2017 has been reconstructed with the trackreconstruction algorithm "JGandalf" (see Figure 1).A sample of muon bundles with energy larger than 10 GeV and multiplicity up to 100 trackswas simulated with MUPAGE. The equivalent livetime is close to the considered detector livetime,3 easured and simulated atmospheric muons for the first KM3NeT detectors
Piotr Kalaczynski
Figure 2: Rate of atmospheric muons as a function of the reconstructed zenith angle for data andMC simulation for the ARCA2 detector. Most of the upgoing events (all MC events) are downgoingevents that are badly reconstructed as upgoing. No quality cuts have been applied to remove thebadly reconstructed tracks.about 20 days for ARCA2 and 23 days for ORCA1. The muon bundle energy is evaluated at thesurface of the detector sensitive volume. The events processed as described in Section 2 are shownin Figures 2, 3 and 4 with the red marks. Only statistical errors are shown in the plots.A total number of 2 . · showers have been simulated with CORSIKA, using SIBYLL-2.3cto describe the high-energy hadronic interaction model [10]. Five different species of nuclei wereconsidered: p , He , C , O and Fe with energies between 10 and 10 GeV and an energy spectrum E − primary . For the muon bundles generated with CORSIKA, the simulation procedure presentedin Section 2 has been followed and events were reconstructed with the same set of programs asfor MUPAGE and data events. At the end of the reconstruction, the events have been weightedaccording to the composition model described in [11]. The results of the CORSIKA simulationare shown as blue markers in Figures 2, 3 and 4. The errors shown are calculated as ∆ x = (cid:113) ∑ w i ,where w i are the weights of each event. In all plots no systematic uncertainties on the simulationare considered.The preliminary results discussed in this paper show that the considered MC samples can re-produce the data with a good level of agreement. The upgoing muon contribution in the zenithangle plots (Figures 2 and 3) is due to badly reconstructed downgoing muons. In this work, noquality cuts or selections have been applied to remove the poorly reconstructed tracks. Notice thatexpected systematic uncertainties, which have not been considered here, are much larger than thestatistical errors. The good data/MC agreement is not surprising as both approaches, the parame-4 easured and simulated atmospheric muons for the first KM3NeT detectors Piotr Kalaczynski
Figure 3: Rate of atmospheric muons as a function of the reconstructed zenith angle for data andMC simulation for the ORCA1 detector. Most of the upgoing events (all MC events) are downgoingevents that are badly reconstructed as upgoing. No quality cuts have been applied to remove thebadly reconstructed tracks.Figure 4: Rate of atmospheric muons as a function of the reconstructed energy for data and MCsimulation for the ARCA2 detector. More details on the energy reconstruction can be found in [9].5 easured and simulated atmospheric muons for the first KM3NeT detectors
Piotr Kalaczynski terized simulation with MUPAGE and the full simulation with CORSIKA, have been successfullyused in ANTARES [12], and confirm that the simulation chain set up for KM3NeT analyses isreliable and under control.
4. Conclusions
Using the data taken with the first KM3NeT detection units, it is possible to compare the mea-sured and the expected muon rate at two different sites. The good level of agreement confirms thereliability of the simulation procedure defined so far within KM3NeT. Sharing the same technologyand detection medium (sea water), the ARCA and ORCA detectors offer a unique opportunity toprobe the depth dependence of the atmospheric muon flux without applying systematic uncertain-ties related to comparison between different types of experiments [8].Once the simulation chain is well established, next steps foresee the increase of the simulatedsamples, particularly at very high energies, and the study of the impact of the physics parameteres.This includes the description of the hadronic interaction models in some regions of the phase spacethat cannot be explored with accelerators, of the chemical composition of the primary CRs andof the optical properties of deep-sea water. A detailed study of the detection and reconstructionefficiency can be also performed.
References [1] The KM3NeT Collaboration: S. Adrian-Martinez et al.,
Letter of Intent for KM3NeT 2.0 , Journal ofPhysics G: Nuclear and Particle Physics, 43 (2016) 084001.[2] G. Carminati et al.,
MUPAGE: a fast atmospheric MUon GEnerator for neutrino telescopes based onPArametric formulas , Comput. Phys. Commun. 179, 915 (2008).[3] D. Heck, J. Knapp, J.N. Capdevielle, G. Schatz, T. Thouw,
CORSIKA: A Monte Carlo code tosimulate extensive air showers , FZKA-6019 (1998)[4] M. Ageron et al.,
ANTARES : The first undersea neutrino telescope , Nucl. Instrum. Meth. A 656(2011) 11.[5] C. Andreopoulos, Ch. Barry, S. Dytman, H. Gallagher, T. Golan, R. Hatcher, G. Perdue, J. Yarba,
TheGENIE Neutrino Monte Carlo Generator: Physics and User Manual , arXiv:1510.05494 (2015)[6] C. Distefano, for the ANTARES and KM3NeT Collaborations, gSeaGen: a GENIE-based code forneutrino telescopes , arXiv:1602.00501 (2016)[7] M. de Jong,
Multi-dimensional interpolations in C++ , arXiv:1907.02597 (2019)[8] The KM3NeT Collaboration: M. Ageron et al.,
Dependence of atmospheric muon rate on seawaterdepth measured with the first KM3NeT detection units , arXiv:1906.02704 (2019)[9] K. Melis, A. Heijboer, M. De Jong on behalf of the KM3NeT Collaboration,
KM3NeT/ARCA EventReconstruction Algorithms , PoS(ICRC2017)950 (2017)[10] F. Riehn, H. P. Dembinski, R. Engel, A. Fedynitch, T. K. Gaisser, T. Stanev,
The hadronic interactionmodel SIBYLL 2.3c and Feynman scaling , PoS ICRC2017 301 (2017)[11] T. K. Gaisser, T. Stanev, S. Tilav,
Cosmic Ray Energy Spectrum from Measurements of Air Showers ,Front.Phys. 8 (2013) 748 easured and simulated atmospheric muons for the first KM3NeT detectors Piotr Kalaczynski[12] J.A. Aguilar et al.,
Zenith distribution and flux of atmospheric muons measured with the 5-lineANTARES detector , Astropart. Phys. 34 (2010) 179.