Neutrino physics and astrophysics with the MACRO experiment at the Gran Sasso lab

Abstract

The results of the MACRO experiment on atmospheric neutrino oscillations are presented and discussed. The data concern different event topologies with average neutrino energies of ~ 4 and ~50 GeV. The Multiple Coulomb Scattering of muons in the MACRO absorbers was used to estimate the neutrino energy of each event of the higher energy sample. The angular distributions, the absolute fluxes and the L/E distribution strongly favour <FONT FACE=Symbol>n m</FONT> <FONT FACE=Symbol>® n t</FONT> oscillations with maximal mixing and deltam² = 0:0025 eV². Results are presented on the searches for astrophysical sources of high energy muon neutrinos, for bursts of electron antineutrinos from stellar gravitational collapses and on indirect searches for WIMPs from the Earth and from the Sun.

Neutrino physics and astrophysics with the MACRO experiment at the Gran Sasso lab

Giorgio Giacomelli^*

Dipartimento di Fisica dell'Università di Bologna and INFN, Sezione di Bologna, it I-40127 Bologna, Italy

ABSTRACT

The results of the MACRO experiment on atmospheric neutrino oscillations are presented and discussed. The data concern different event topologies with average neutrino energies of ~ 4 and ~50 GeV. The Multiple Coulomb Scattering of muons in the MACRO absorbers was used to estimate the neutrino energy of each event of the higher energy sample. The angular distributions, the absolute fluxes and the L/E distribution strongly favour n_m ® n_t oscillations with maximal mixing and Dm² = 0:0025 eV². Results are presented on the searches for astrophysical sources of high energy muon neutrinos, for bursts of electron antineutrinos from stellar gravitational collapses and on indirect searches for WIMPs from the Earth and from the Sun.

I Introduction

MACRO was a large multipurpose area underground experiment designed to search for rare events in the penetrating cosmic radiation. These included the study of atmospheric neutrinos and their oscillations, high energy (E_n 1 GeV) muon neutrino astronomy, indirect searches for WIMPs and search for low energy (E_n 7 MeV) stellar collapse neutrinos. The detector was placed in the Gran Sasso laboratory, located on the highway Rome-Teramo , 120 km east of Rome. The lab consists of three underground halls, each about 100 m long. It is at an altitude of 963 m above sea level, is well shielded from cosmic rays by a mean rock thickness of 3700 m.w.e.; the minimum is 3150 m.w.e: this defines the minimum muon energy at the surface at ~ 1.3 TeV in order to reach MACRO. The average residual muon energy and the muon flux are ~ 320 GeV and ~ 1 m^–2h^–1, respectively.

MACRO was composed of three sub-detectors: liquid scintillation counters, limited streamer tubes and nuclear track detectors. Each one of them could be used in stand-alone and in combined mode. A cross section of the detector is shown in Fig. 1. Notice the division in the lower and in the upper part (this was often referred to as the Attico); the inner part of the Attico was empty and lodged the electronics [1]. The mass of the lower MACRO was 4200 t, mainly in the form of boxes filled with crushed Gran Sasso rock. The detector had a modular structure: it was divided into six sections referred to as supermodules. Each active part of one supermodule had a size of 12.6×12 ×9 .3 m³ and had a separate mechanical structure and electronics readout. The full detector had global dimensions of 76.5 ×12 ×9.3 m³ and provided a total acceptance to an isotropic flux of particles of ~ 10000 m² sr. The total mass was 5300 t.

Data were taken from March 1989 till april 1994 with the detector under construction and from the middle of 1994 till the end of 2000 with the full detector.

II Atmospheric neutrino oscillations

If neutrinos have non-zero masses, one has to consider the weak flavour eigenstates n_e, n_m, n and the mass eigenstates n₁, n₂, n₃. The weak flavour eigenstates n_l are linear combinations of the mass eigenstates n_m through the elements of the mixing matrix U_lm:

In the simple case of only two flavour eigenstate neutrinos (n_m, n ) which oscillate with two mass eigenstates (n₂, n₃) one has

where q₂₃ is the mixing angle (q will be used in the following). In this case one may easily compute the following expression for the survival probability of a n_m beam

where Dm² = and L is the distance travelled by the muon neutrino from production to detection.

High energy primary cosmic rays interact in the upper atmosphere producing pions and Kaons, which by decay yield muons and muon neutrinos, p ® mn_m; further decays of the muons lead to electron and muon neutrinos, m ® n_m n_l l. The neutrinos are produced in a spherical shell at about 10-20 km from the earth surface. Upgoing n_m's may interact in the rock below MACRO or inside its lower part leading to upgoing muons, n_mN ® m^+– +....

Upward going muons are identified by the streamer tube system (for tracking) and the scintillator system (for time-of-flight measurement). A rejection factor of at least 10⁷ is needed in order to separate upgoing muons from the background due to the downgoing muons. Fig. 1 shows sketches of the different neutrino event topologies analyzed: Upthroughgoing muons, Upsemicontained (also called Internal Upgoing muons, IU), Upgoing Stopping muons (UGS), Internal Downgoing muons (ID). The numbers of events measured and expected for the three topologies are given in Table . All data samples deviate from Monte Carlo expectations; the deviations point to neutrino oscillations.

II.1 Upthroughgoing muons

The upthroughgoing muons come from n_m interactions in the rock below the detector; the n_m's have an average energy _n ~ 50 GeV [2,3]. The data refer to a lifetime of 6.16 years (full detector equivalent). The data deviate in absolute value and in shape from Monte Carlo (MC) predictions, see Fig. 2. We studied a large number of possible effects that could affect our measurements: no significant systematic problems exist in the detector or in the data analyses.

Figure 2

The measured data have been compared with MC simulations. For the upthroughgoing muon simulation, the neutrino flux computed by the Bartol group was used [4]. The cross sections for the neutrino interactions were calculated using the deep inelastic parton distributions of ref. [5]. The muon propagation to the detector was done using the energy loss calculation in standard rock. The total systematic uncertainty on the expected muon flux, obtained adding in quadrature the errors from neutrino flux, cross section and muon propagation, is estimated to be 17 %. This uncertainty is mainly a scale error; the error on the shape of the angular distribution is ~ 5 %.

Figure 3 shows the allowed region for n_m ® n oscillations in the sin²2q–Dm² plane, computed according to ref. [6] for the upthroughgoing muon events; our region is compared with those obtained by the SuperKamiokande (SK) [7] and Soudan 2 [8] experiments.

II.2 Matter effects. n_m ® n

versus n_m ® n_sterile

Matter effects due to the difference between the weak interaction effective potential for muon neutrinos with respect to sterile neutrinos would produce a different total number and a different zenith distribution of upthroughgoing muons [9]. In Fig. 4 the measured ratio between the events with –1 < cos Q < –0.7 and the events with –0.4 < cos Q < 0 is shown as a black point. In this ratio most of the theoretical uncertainties on neutrino flux and cross section cancel. The remaining theoretical error combined with the experimental error is estimated to be 7%. We measured 305 events with –1 < cos Q < –0.7 and 206 with –0.4 < cos Q < 0 ; the ratio is R = 1.48 ±0.13_stat ±0.10_sys. For Dm² = 2.5·10^–3 eV² and maximal mixing, the expected value of the ratio for n_m ® n is R = 1.72 while for n_m ® n_s is R_sterile = 2.16. The maximum probabilities P_best to find a value of R and of R_sterile smaller than the expected ones are 9.4 % and 0.06 % respectively. The ratio of the maximum probabilities is = 157, so that n_m ® n_s oscillations are disfavoured at 99% c.l. compared to the n_m ® n channel [3,9].

II.3 n_m energy estimates by MCS of upthroughgoing muons

The oscillation probability is a function of the ratio L/E_n, Eq. 3. Even if a precise measurement of E_n is not possible, MC simulations have shown a correlation between E_n and the residual muon energy E_m. E_n was thus estimated for each event by measuring the muon energy E_m, via Multiple Coulomb Scattering (MCS) of the iduced muon in the absorbers. The r.m.s. of the lateral displacement for a muon crossing the apparatus on the vertical is s_MCS 10 cm/ E_m ( GeV). The muon energy E_m can be measured up to a saturation point, occurring when s_MCS is comparable with the detector space resolution. A first analysis was made studying the deflection of upthroughgoing muons with the streamer tubes in digital mode''. This method could reach a spatial resolution of ~ 1 cm which implies a maximum measurable energy of 10 GeV [10]. As the interesting energy region for atmospheric neutrino oscillations spans from ~ 1 GeV to tens of GeV, it is important to improve the spatial resolution of the detector to push the saturation point as high as possible. For this purpose, a second analysis was performed with the streamer tubes used in "drift mode''. To check the electronics and the feasibility of the analysis, two tests were performed at the CERN PS-T9 and SPS-X7 beams. The space resolution achieved is 3 mm , a factor 3.5 better than in the first analysis. For each muon, seven MCS sensitive variables were given in input to a Neural Network (NN) previously trained to estimate the muon energy with MC events of known energy crossing the detector at different zenith angles. The method allowed to separate the upthroughgoing muons in 4 subsamples corresponding to average neutrino energies of 12, 20, 50 and 102 GeV, respectively. The comparison of the 4 zenith angle distributions with the predictions of the no oscillations MC shows a disagreement at low energies (where there is a deficit of vertical events), while the agreement is restored at the higher neutrino energies. The distribution of the ratios R = (Data/MC_noosc) obtained by this analysis is plotted in Fig. 5 as a function of log₁₀(L/E_n) [3,11]. The black points with error bars are the data; the vertical extent of the shaded areas represents the uncertainties on the MC predictions for n_m ® n oscillations with maximal mixing and Dm² = 2.5·10^-3 eV² . The horizontal dashed line is the expectation without oscillations. The black square point was obtained from the low energy IU sample.

II.4 Low energy data

The Internal Upgoing (IU) muons come from n_m interactions in the lower apparatus; for these events two scintillation counters are intercepted; the T.o.F. is applied to identify upward going muons. The average parent neutrino energy for these events is 4.2 GeV. For neutrino oscillations one expects a reduction by about a factor of two in the flux, without any distortion in the shape of the angular distribution. This is what is observed in Fig. 6a, [12].

The upstopping muons (UGS) are due to external n_m interactions yielding upgoing muons stopping in the detector. The semicontained downgoing muons (ID) are due to n_m– induced downgoing muon tracks with vertex in the lower MACRO (Fig. 1). The two types of events are identified by topological criteria; the lack of time information prevents to distinguish the two sub-samples. An almost equal number of UGS and ID events is expected. In case of oscillations, the flux of the UGS muons should be reduced by 50%; no reduction is instead expected for the semicontained downgoing events (coming from neutrinos with path lengths of ~ 20 km); therefore one expects a global reduction of 25%.

The number of events and the angular distributions are compared with MC predictions in Table 1 and Figs. 6a,b. The data show a uniform deficit for the whole angular distribution with respect to predictions, ~ 50% for IU, 75% for ID + UGS; there is good agreement with the predictions based on neutrino oscillations with the parameters obtained from the upthroughgoing muons. The average value of the double ratio R = (Data/MC)_IU/(Data/MC)_ID+UGS over the measured zenith angle range is R 0.77 ±0.07 ; the error includes statistical and theoretical uncertainties; R = 1 is expected in case of no oscillations.

III Search for Astrophysical Sources of High Energy Muon Neutrinos

High energy n_m are expected to come from several galactic and extragalactic sources. Neutrino production requires astrophysical accelerators of charged particles and some kind of astrophysical beam dumps.

The excellent angular resolution of our detector allowed a sensitive search for upgoing muons produced by neutrinos coming from celestial point sources, with a negligible atmospheric neutrino background. An excess of events was searched for around the positions of known sources in 3^° (half width) angular bins. This value was chosen so as to take into account the angular smearing due to multiple muon scattering in the rock below the detector and by the energy-integrated angular distribution of the scattered muon, with respect to the neutrino direction.

The pointing accuracy of the detector was checked with the observation of the shadow of primary cosmic rays (and thus on downgoing muons) by the Moon and the Sun [13]. It was also checked with the study of the correlations between muons detected by MACRO and extensive air showers detected by the EASTOP experiment located above the Gran Sasso massif [14].

In a total lifetime of 6.16 y we obtained a total of 1356 events, see Fig. 7a [3, 13]. The 90% c.l. upper limits on the muon fluxes from specific celestial sources lay in the range 10^–15–10^–14 cm^–2 s^–1 , see Fig. 7b. The solid line is our sensitivity vs declination. Notice that we have two cases, GX339-4 (a = 255.71^o , d = -48.79^o ) and Cir X-1 (a = 230.17^o , d = -57.17^o ), with 7 events: in Fig. 7b they are considered as background, therefore the upper flux limits are higher; but they could also be indications of signals [3, 15]. We searched for time coincidences of our upgoing muons with sources of g-ray bursts as given in the BATSE 3B and 4B catalogues, for the period from April 91 to December 2000 [15]. No statistically significant time correlation was found.

We have also searched for a diffuse astrophysical neutrino flux, for which we establish a flux upper limit at the level of 1.5·10^–14 cm^–2 s^–1 [16].

IV Indirect Searches for WIMPs

Weakly Interacting Massive Particles (WIMPs) could be part of the galactic dark matter; they could be intercepted by celestial bodies, slowed down and trapped in their centers, where WIMPs and anti-WIMPs could annihilate and yield n_m ® upthroughgoing muons. The annihilations in these celestial bodies would yield neutrinos of GeV- TeV energy, in small angular windows from their centers. For the Earth we have chosen a 15^o cone around the vertical: we find 863 events. The MC expectation for atmospheric n_m without oscillations gives a larger number of events. We set a conservative flux upper limit assuming that the measured number of events equals the expected ones. We obtain the 90 % c.l. limits for the upgoing muon flux shown in Fig. 8a (it varies from 0.8 to 0.5 10^–14 cm^–2 s^–1 ) [17, 3]. If the WIMPs are identified with the smallest neutralino mass, our limits may be used to constrain the neutralino mass, following the model of ref. [18], Fig. 8a.

A similar procedure was used to search for muon neutrinos from the Sun, using 10 search cones from 3^o to 30^o . In the absence of statistically significant excesses the muon flux upper limits are at the level of 1.5–2·10^–14 cm^–2 s^–1 . The limits are shown in Fig. 8b as a function of the WIMP (neutralino) mass.

V Neutrinos from Stellar Gravitational Collapses

A stellar gravitational collapse (GC) of the core of a massive star is expected to produce a large burst of all types of neutrinos and antineutrinos with energies of 7-30 MeV and with a duration of < 10 s . The

_e's can be detected via the process _e + p ® n + e⁺ in the liquid scintillator. About 100¸150 _e events should be detected in our 580 t scintillator for a stellar collapse at the center of our Galaxy.

We used two electronic systems to search for bursts of

_e's from stellar gravitational collapses. The first system was based on the dedicated PHRASE trigger, the second one was based on the ERP general trigger. Both had energy thresholds of ~ 7 MeV and recorded pulse shape, charge and timing informations. Immediately after a trigger, the PHRASE system lowered its threshold to 1 MeV, for a duration of 800ms , in order to detect (with a 25 % efficiency) the 2.2 MeV g released in the reaction n + p ® d + g_{2.2 MeV} induced by the neutron produced in the primary process.

A redundant supernovae alarm system was in operation, alerting immediately the physicists on shift. We defined a general procedure to alert the physics and astrophysics communities in case of an interesting alarm. Finally, a procedure to link the various supernova observatories around the world was set up [3, 19].

The effective MACRO active mass was ~ 580 t; the lifetime fraction was 97.5 % ; we kept always ruming at least 1-2 supermodules. No stellar gravitational collapses were observed in our Galaxy from the beginning of 1989 to the end of 2000 [3][19].

VI Conclusions

The MACRO detector took data from 1989 to the end of the year 2000 and it obtained important results in all the items listed in the proposal, in particular:

• Atmospheric neutrino oscillations. Analyses of different event topologies at different energies, the exploitation of the muon Coulomb multiple scattering in the detector give strong support to the hypothesis of n_m ® n oscillations with Dm² = 0.0025 eV² and maximal mixing. We are presently making the final global analysis of all our data.

• High energy muon neutrino astronomy. MACRO has been highly competitive with other underground experiments thanks to its large acceptance and good angular accuracy. It has been limited by its lifetime and the size of the detector.

• Search for bursts of

  
  _e from stellar gravitational collapses. MACRO was sensitive to supernovae events in our Galaxy, it started the SN WATCH system, and for a certain time it was the only detector in operation.

• Sensitive indirect searches have been carried out for possible Dark Matter candidates like the neutralinos, looking for upgoing muons from the center of the Earth and from the Sun.

It may be pointed out that MACRO also obtained the best flux upper limits for GUT Magnetic Monopoles over the widest b range [20].

Acknowledgments

I thank the members of the MACRO Collaboration for their cooperation; in particular I acknowledge the discussions and the help from the members of the Bologna group, in particular Drs. Y. Becherini, M. Giorgini and M. Spurio. I thank Ms Anastasia Casoni for typing and correcting the manuscript.

Reference

[1] S. Ahlen et al., Nucl. Instr. Meth. Phys. Res. A 324, 337 (1993); M. Ambrosio et al, Nucl. Instr. Meth. Phys. Res. A 486, 663 (2002); hep-ex/0207043.

[2] S. Ahlen et al., Phys. Lett. B 357, 481 (1995); Phys. Lett. B 434, 451 (1998).

[3] M. Ambrosio et al., hep-ex/0206027.

[4] V. Agrawal et al., Phys. Rev. D 53, 1314 (1996).

[5] M. Gluck et al., Z. Phys. C 67, 433 (1995).

[6] G. Feldman and R. Cousins, Phys. Rev. D 57, 3873 (1998).

[7] SK Coll., Y. Fukuda et al., Phys. Rev. Lett. 81, 1562 (1998); Phys. Rev. Lett. 85, 3999 (2000); hep-ex/0105023.

[8] Soudan 2 Coll., W.W.M. Allison et al., Phys. Lett. B 449, 137 (1999).

[9] M. Ambrosio et al., Phys. Lett. B 517, 59 (2001).

[10] D. Bakari et al., hep-ex/01050871.

[11] M. Ambrosio et al., Nucl. Instr. Meth. Phys. Res. A ??????, hep-phys/0203018.

[12] M. Ambrosio et al., Phys. Lett. B 478, 5 (2000); hep-ex/01120021.

[13] M. Ambrosio et al., Phys. Rev. D 59, 012003 (1999).

[14] EASTOP-MACRO Coll., M. Aglietta et al., Phys. Lett. B 337, 376 (1994).

[15] M. Ambrosio et al., Astrophys. J. 546, 1038 (2001).

[16] M. Ambrosio et al., astro-phys/0203181.

[17] M. Ambrosio et al., Phys. Rev. D 60, 082002 (1999).

[18] A. Bottino et al., Proceedings of the Ringberg Euroconf. New Trends in Neutrino Physics, Ringberg Castle, Germany, 1998, edited by B. Kniel, World Scientific, Singapore.

[19] M. Ambrosio et al., Astroparticle Phys. 8, 123 (1998).

[20] M. Ambrosio et al., Eur. Phys. J. C 25, 511 (2002) hep-ex/0207020; Eur. Phys. J. C 26, 163 (2002), hep-ex/0207024.

Received on 1 October, 2002

*

MACRO Collaboration. See Ref. 1 for a list of MACRO Collaborators and Institutions.

Publication Dates

• Publication in this collection
  25 Aug 2003
• Date of issue
  June 2003

History

• Accepted
  01 Oct 2002
• Received
  01 Oct 2002