Table Of Content

University of Bologna Report DFUB 01/01 MACRO AND THE ATMOSPHERIC NEUTRINO PROBLEM M. SPURIOfor the MACRO collaboration a Dipartimento di Fisica dell’Università and INFN, 40126 Bologna, Italy e-mail: [email protected] 1 0 AfterabriefpresentationoftheMACROdetectorwediscusstheupdateddataon 0 atmospheric muon neutrinos, and the interpretation in terms of neutrino oscilla- 2 tions. n a J 1 Introduction 5 MACRO is a multipurpose underground detector designed to search for rare 1 events in the cosmic radiation. It performs measurements in areas of astro- 1 physics,nuclear,particleandcosmicrayphysics. Inthispaperweshalldiscuss v the measurement of the atmospheric muon neutrino flux in the energy region 9 from a few GeV up to a few TeV, and the interpretation of the MACRO re- 1 sults in terms of neutrino oscillations. The detector has global dimensions of 0 1 12×9.3×76.5 m3 and providesa total acceptance to anisotropic flux of par- 0 ticles of ∼ 10,000m2sr. The total mass is ≃ 5300 t. A cross section of the 1 detector is shown in Fig. 1a1. It has three sub-detectors: liquid scintillation 0 counters, limited streamer tubes and nuclear track detectors. The mean rock x/ depthofthe overburdenis≃3700m.w.e. Theaverageresidualenergyandthe e muon flux at the lab depth are ∼310 GeV and ∼1 m−2 h−1, respectively. - p e 2 MACRO as atmospheric ν detector h µ : v Primarycosmic raysproducein the upper atmospherepions andkaons,which i decay, π → µν, K → µν; and µ → eν ν . Fig. 1a shows the three different X e µ topologies of neutrino events detected by MACRO: up throughgoing muons2, r a semicontained upgoing muons (IU) and up stopping muons + semicontained downgoing muons (UGS + ID)3. The parent ν energy spectra for the three µ event topologies were computed with Monte Carlo (MC) methods, and presented in Fig. 1b. The number of events measured and expected for the three topologies is given in Table 1. Sources of background and systematic effects were studied in detail in3,4 and found to be negligible. The neutrino energies involvedinalleventtopologiesaresuchthattheµ fluxisup-downsymmetric. ν High energy sample. The up throughgoing muons come from ν ’s in- µ teractingintherockbelowthe detector;theyareidentifiedusingthestreamer tube system(fortracking)andthe scintillatorsystem(for time-of-flight(ToF) aInvitedtalkattheThirdInternational WorkshoponNewWorldsinAstroparticlePhysics 1-3September 2000,UniversityoftheAlgarve. Faro,Portugal 1 Figure1: a)Eventtopologiesinducedbyνµ interactionsinandaroundMACRO.The blackboxesarescintillator hits. Thetrackdirectionandversusaremeasuredbythe streamertubesandtimeofflight. b)Neutrinoenergydistributionforthethreeevent topologies detected by MACRO Events Predictions R = Data/MC selected No oscil. Up throughgoing 723 989 0.731±0.028 st ±0.044 ±0.124 sys th Internal Up 135 245 0.55 ±0.04 st ±0.06 ±0.14 sys th UpG Stop + 229 329 0.70 ±0.04 st In Down ±0.07 ±0.18 sys th Table 1: Event summary for the MACRO atmospheric neutrino flux analyses. The data correspond to a lifetime of ∼ 5 y with the full detector, after background corrections[2,3]. TheratiosR=Data/MCarerelativetoMCexpectationsassuming no oscillations. measurement). A rejectionfactor ofatleast105 isneeded inorderto separate the upgoing muons from the large background of down-going muons2. The ν median energy is E ∼ 50 GeV, and muons with E > 1 GeV cross the µ ν µ whole detector. Fig. 2a shows the zenith angle distribution of the measured flux ofupthroughgoingmuons. The datahavebeencomparedwithMC simu- lations2; theexpectedangulardistributionsfornooscillationsisshowninFig. 2a as a dashed line, and for a ν → ν oscillated flux with sin22θ = 1 and µ τ ∆m2 = 0.0025 eV2, as a solid line. The total theoretical uncertainty on the expected muon flux, adding in quadrature the errorsfrom neutrino flux, cross section and muon propagation, is 17 %; it is mainly a scale error that does not change the shape of the angular distribution. The ratio of the observed 2 8 2) V –13–2–1–10 cm s sr) 765 D2m (e1100––12 Up 1 through flux ( 4 10–3 -going mg 3 n oi ardg 2 10–4 Low energies w Up 1 10–5 0 – 1 – 0.8 – 0.6 – 0.4 – 0.2 0 0 0.2 0.4 0.6 0.8 1 cos q sin2 2q Figure2: (a)Measuredflux(points)oftheupthroughgoingmuonsplottedvs. zenith angle Θ. The solid line is the fit to an oscillated muon flux, obtaining maximum mixing and ∆m2 = 0.0025 eV2. (b) Confidence regions for νµ → ντ oscillations at the90 % CL, obtained from MACRO high energy [2] and low energy [3] data. number of events to the expectation without oscillations is given in Table 1. Low energy samples. The upgoing semicontained muons (IU, in Fig. 1a) come from ν interactions inside the lower apparatus. Since two scintilla- µ tion counters are intercepted, the ToF is applied to identify the upward going muons. The average parent neutrino energy for these events is 4.2 GeV. The upgoing stopping muons (UGS) are due to external ν interactions yielding µ upgoing muon tracks stopping in the detector; the semicontained downgoing muons (ID) are due to ν induced downgoingtrackswith vertices in the lower µ MACRO. The events were selected by means of topological criteria; the lack of time information prevents to distinguish the two sub samples, for which an almost equal number of events is expected. The average neutrino energy for theseeventsis≃3.5GeV. Thenumberofeventsandtheangulardistributions are compared with the predictions3 in Fig. 3 and Table 1. 3 Interpretation in terms of neutrino oscillations We interpreted the reduction on the detected number of events (Table 1) and the deformationofthezenithangledistributionofthe up-throughgoingevents (Fig. 2a) as a consequence of ν disappearance. µ The weak flavor eigenstates (ν , ν , ν ) are relevant for the production of e µ τ atmosphericneutrinos,viaπ →µν ,K→µν andµ→eν ν decays. Formas- µ µ e µ sive neutrinos one has to consider the mass eigenstates (ν1, ν2, ν3) which are relevantfor propagation. The flavoreigenstatesare linear combinationsof the 3 mass eigenstates, ν = P U ν . In the simplest two flavors oscillations l m=1 lm m scenario,onlyonemixingangleθ is involved,andthe rateofν disappearance µ is given by P(ν →ν )≃1−sin22θsin2(1.27∆m2L/E ) (1) µ µ ν 3 Figure3: Measured(points)andexpectednumberoflowenergymuonneutrinoevents versus zenith angle. IU: up semicontained. ID+UGS: up stopping plus down semi- containedmuons. Theshadowedregionsarethepredictionswithoutoscillations; the dotted lines are the predictions assuming neutrino oscillations with the parameters obtained from the upthroughgoing sample. ∆m2 = m2 − m2 , L = distance (in km) from the decay point to the ν3 ν2 interaction point, E is the neutrino energy, in GeV. ν This simple relation could be modified when neutrinos propagate with a flavor-dependentinteractionthroughmatter5. Thiscaseisimportant,because the matter effect could help to discriminate between ν oscillation in different µ neutrino channels. The matter effect in the Earth could enhance (in a range of values of the oscillations parameters) the ν disappearance effect through µ ”resonances” (MSW effect). This happens only for ν → ν and for ν → ν µ e µ s channels (ν = sterile neutrino), while for ν → ν there is no matter effect s µ τ (ν and ν have the same weak potential). µ τ In the scenario described by eq. 1, relatively fewer up throughgoing muons are expected near the vertical (cosΘ = −1) than near the horizon- tal (cosΘ=0), due to the longer path length of neutrinos from production to observation, in a wide range of ∆m2 values. The probabilities to obtain the number of events in Table 1 and the angular distribution observed in Fig. 2a havebeencalculatedvs.oscillationparametervalues. Forν →ν oscillations µ τ the maximum probability is 57 %, for ∆m2 = 0.0025 eV2, sin22θ = 1 (com- bination from the shape of the angular distribution and the reduction in the number of events). Fig. 2b shows the 90 % CL regions for ν →ν . µ τ The probability for no-oscillations is 0.4%. The maximum probability for ν → ν oscillations is 15%. Another way to discriminate between the ν µ s µ oscillation in ν or in ν is to study the angular distribution in two bins. The τ s statistical significance is higher6 with respect to the 10-bins fit of the angular distribution. From this ratio, the ν → ν hypothesis is disfavored at level of µ s 2.5σ compared to ν →ν 5. µ τ 4 From our results on the high energy sample, a ∼ 50% reduction in the flux of the up stopping events (UGS) as well of semicontained upgoing muons (IU) is expected, with no distortion in the shape of the angular distribution. No reductionis insteadexpected for the semicontaineddowngoingevents (ID, comingfromneutrinoswithpathlengthsof∼15km). Thisiswhatisobserved in Fig. 3. The allowed region in the ∆m2, sin22θ plane is shown in Fig. 2b. 4 Up throughgoing muon energy separation The up throughgoing muon sample is induced by atmospheric ν in a wide µ energy range (for 90% of events 5∼< E ∼<700 GeV, and 2∼< E ∼<300 GeV, see ν µ Fig. 1b). If the observed event reduction is due to neutrino oscillations, the first minimum of eq. 1 occurs at E ≃ 10−30 GeV (for neutrinos from the ν nadir direction and oscillation parameters in the range of Fig. 2b). Thus, the ν disappearance mechanismworks on (relatively)lower energy neutrinos. As µ a consequence, we expect (from MC) a ∼ 50% reduction of the up throughgoing muons with E ∼<10 GeV, and only a ∼ 20% reduction for muons with µ E ∼>10 GeV. µ Muons traversing MACRO are deflected by many small-angle scatters, the bulk of which are Coulomb scatterings from the nuclei and the atomic electrons (multiple Coulomb scattering). As an effect, relatively low energy up throughgoing muons are (slightly) deviated from a straight-line direction inside the detector. We evaluated that this deviation is of the order of few centimeters, up to E ∼10 GeV, and it exceed the intrinsic spatial resolution µ of the MACRO streamer tube cells. A selection cut, based on the deviation of the muon sagitta from the straight line when the muon cross the whole detector, was defined. Each up throughgoing muon is classified as: Low= muon with a deviation from the straight line (from MC, lower energy muons with median ELow = 11 GeV); µ High=muonwithnomeasurabledeviation(muonswithmedianenergyEHigh = µ 52GeV). Theremainingeventswithambiguousdeviation,andwhichdoesnot satisfy the high or the low energy cuts, were classified as Medium. From MC, they have an intermediate energy EMedium = 33 GeV. The same cuts were µ then applied to the simulated neutrino-induced events, and the results compared. To avoidsystematic differences from data and MC, a detailed check of the detector simulation was performed. We used a large sample of measured and simulated downward going atmospheric muons, and we required that the distributions, used to define the selection cut, match together. The (prelimi- nary)resultsarepresentedinFig. 4. Itisincludedalsothepointcorresponding tothepartiallycontainedIUevents. Only87%ofIUeventshavebeeninduced 5 C M A / T A D Figure 4: Ratio (detected events/expected events) vs. median muon energy for the wholesampleof723upthroughgoingmuons(All),andforthesubsamplewithlarge (Low, 79 events) and small (High, 122 events) deviation from a straight line. It is alsoincludedtheratiofortheeventswith intermediateenergy(Medium,49events). The point IU corresponds to the 135 partially contained upgoing events. For all samples,theνµ pathlengthis∼104 km. Thepointsareplottedwithequaldistance in abscissa, to guide the eye. The full line (data/MC=1) is theexpectation without neutrinooscillations, forwhicha17%totaluncertaintyisassociated (25%fortheIU point). The dashed lines are the expected reductions in case of νµ →ντ oscillations with maximum mixing and three values of ∆m2. by ν CC interactions, the remaining by ν and NC interactions. µ e Fig. 4 shows that lower energy events are more suppressed than higher energy ones. In a similar way, Fig. 2a shows that neutrinos with a longer path length are more suppressed; both effect are well explained by neutrino oscillation (eq. 1), with the oscillation parameter presented in Fig. 2b. IacknowledgeallthecolleaguesoftheMACROcollaboration,inparticular D. Bakari and Y. Becherini. References [1] S. Ahlen et al., Nucl. Instr. Meth. A324(1993)337. [2] M. Ambrosio et al., Phys. Lett. B434(1998)451; Phys. Lett. B357(1995)481. [3] M. Ambrosio et al., Phys. Lett. B478(2000)5; M. Spurio, Nucl.Phys.Proc.Suppl.85:37-43,2000(hep-ex/9908066). [4] M. Ambrosio et al., Astropart. Phys. 9(1998)105. [5] F. Ronga, Nucl.Phys.Proc.Suppl.87:135,2000(hep-ex/0001058),and references therein. [6] P. Lipari and M.Lusignoli, Phys. Rev. D57, 3842(1998). 6