Search for Muon Neutrino Disappearance in a Short-Baseline Accelerator Neutrino Beam
aa r X i v : . [ h e p - e x ] O c t Proceedings of the XXIX PHYSICS IN COLLISION 1
Search for Muon Neutrino Disappearance in a Short-Baseline AcceleratorNeutrino Beam
Yasuhiro Nakajima, for the SciBooNE Collaboration
Kyoto University Kitashirakawa-Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan
Abstract
We report a search for muon neutrino disappearance inthe ∆ m region of 0 . −
40 eV using data from both Sci-BooNE and MiniBooNE experiments. SciBooNE dataprovides a constraint on the neutrino flux, so that thesensitivity to ν µ disappearance with both detectors isbetter than with just MiniBooNE alone. The prelim-inary sensitivity for a joint ν µ disappearance search ispresented.
1. Introduction
Neutrino oscillations have been observed and con-firmed at mass splitting (∆ m ) of ∼ − eV and ∼ − eV , called the “solar” and “atmospheric” re-gions, respectively. The observed mixing is consistentwith three generations of neutrinos.However, the LSND experiment observed an excess of ν e in a ν µ beam, indicating a possible oscillation in the∆ m ∼ region [1]. To explain LSND with oscilla-tions requires more than three generations of neutrinosor other exotic physics beyond the Standard Model.To test the oscillation at ∆ m ∼ , the Mini-BooNE experiment recently made searches for both ν e appearance [2, 3] and ν µ disappearance [4] in this pa-rameter region. The experiment observed no signifi-cant ν e appearance signal and ruled out as being dueto 2-neutrino oscillations. However, the sensitivity ofMiniBooNE-only ν µ disappearance search was limitedby the large flux and neutrino interaction cross-sectionuncertainties.Here, we discuss an improved search for ν µ disappear-ance using data from both the SciBooNE [5] and theMiniBooNE experiments, where SciBooNE detector isused to constrain flux and cross-section uncertainties.
2. Experimental Setup
Fig. 1.
The setup of SciBooNE and MiniBooNE experiments.
The experiments use the Booster Neutrino Beam(BNB) at Fermilab [6]. The primary proton beam, ex-tracted with a kinetic energy of 8 GeV, strikes a 71 cmlong, 1 cm diameter beryllium target. The mesons, pri-marily π + , generated by the p − Be interactions are fo-cused with a magnetic horn and decay in the following50 m decay volume, producing an intense neutrino beamwith the peak energy of ∼ π − are focused and hence a predomi-nantly antineutrino beam is created. The SciBooNE detector [5] is located 100 m down-stream from the beryllium target.The detector complex consists of three sub-detectors:a fully active fine grained scintillator tracking detec-tor (SciBar), an electromagnetic calorimeter (EC) anda muon range detector (MRD).The SciBar detector consists of 14,336 extruded plasticscintillator strips (CH), each with dimension of 1 . × . ×
300 cm . The scintillators are arranged verticallyand horizontally to construct a 3 × × . detector.The detector itself is the neutrino target and its fiducialvolume is 10.6 tons.The EC is installed downstream of the SciBar, and ismade of scintillating fibers embedded in lead foil.The MRD is located downstream of the EC in order tomeasure the momentum of muons up to 1.2 GeV /c usingthe muon range. It consists of 12 layers of 2”-thick ironplates sandwiched between layers of 6 mm-thick plasticscintillator planes.The SciBooNE experiment ran from June 2007 untilAugust 2008, collecting a total of 2 . × Protonson Target (POT) for physics analysis; 0 . × POTin neutrino mode and 1 . × POT in antineutrinomode.
The MiniBooNE detector [7] is located 440 m down-stream from the SciBooNE detector. The detector is a 12m diameter spherical tank filled with 800 tons of min-eral oil (CH ). The MiniBooNE experiment has beentaking beam data since 2002, including the SciBooNEand MiniBooNE joint-run period. The collected numberof POT after data quality cut in the neutrino mode is5 . × in addition to the data from the joint-runperiod. ν µ Disappearance Analysis
In this paper, we report only the neutrino data ( ν µ → ν x ) disappearance analysis. We search for muon neu-trino disappearance by comparing neutrino fluxes at Sci-BooNE and MiniBooNE detectors.The analysis is performed in the following three steps:(1) Neutrino flux measurement at SciBooNE, (2) Fluxextrapolation to MiniBooNE, and (3) Oscillation fit.At each step, systematic errors are estimated andpropagated to the final prediction. The majority of theflux and cross-section uncertainties cancels since the neu-trino interaction target in both detectors is effectivelycarbon, and the two detectors are on the same beamline.We describe these steps in detail in the following sec-tions. c (cid:13) For the spectrum analysis at SciBooNE, we use inclu-sive ν µ charged current (CC) interactions, whose signa-ture is long muon tracks. First, we identify muons by se-lecting the longest track with energy deposit consistentwith a minimum-ionizing particle. Second, we requirethe vertex of the track to be within the SciBar fiducialvolume. The events are further divided into two sub-samples based on the location of the muon track endpoints: a “SciBar-stopped” sample containing muonsthat have stopped inside the SciBar detector and a“MRD-stopped” sample with muons that have stoppedin the MRD. These two samples each contain approxi-mately 14k and 20k events with mean energies of 0.8 and1.1 GeV, respectively. Spectrum Fitting
The neutrino spectrum at SciBooNE is extracted byfitting muon momentum ( P µ ) and muon angle ( θ µ ) dis-tributions from each sample.We prepare MC templates for P µ and θ µ distributionsfor several true neutrino energy ( E ν ) regions. The E ν regions are divided by 250 MeV up to 1.25 GeV, and asingle region is assigned for E ν > .
25 GeV. Then, thescale factors for each E ν region are determined to mini-mize the χ between data and MC. Figure 2. shows thefit result. The systematic errors from SciBooNE detectorresponse and neutrino cross-section models are estimatedand shown in the plot.Figure 3. is the P µ and θ µ distributions of SciBooNE’sMRD-stopped sample, after applying scale factors ob-tained by the spectrum fitting. We confirm the MC dis-tributions agrees well to data after fitting. (GeV) n E0 0.5 1 1.5 2 2.500.20.40.60.811.21.41.61.82
Preliminary
Fig. 2.
Scale factors obtained by SciBooNE spectrum fitting.The error bars show the sum of SciBooNE statistical andsystematic uncertainties.
We select events in MiniBooNE by requiring singlemuon and its decay electron. Neutrino energy is recon-structed from muon kinematics by assuming CC QuasiElastic (CCQE) interaction ( ν µ n → µ − p ): E Recν = 2( M n − E B ) E µ − ( E B − M n E B + ∆ M + M µ )2[( M n − E B ) − E µ + p µ cos θ µ ] , Sat Oct 17 15:44:00 2009
Entries 20227 (GeV) m P0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 202004006008001000120014001600
Entries 20227
Entries 20227
DataOtherDirtNCCC other p CC coherent p CC resonance CCQE
Recostructed Muon Momentum
Sat Oct 17 15:44:01 2009
Entries 20227 (deg) mq Entries 20227
Entries 20227
DataOtherDirtNCCC other p CC coherent p CC resonance CCQE
Recostructed Muon Angle
PreliminaryPreliminary
Fig. 3.
Distribution of reconstructed muon momentum (top)and muon angle (bottom) for the MRD-stopped sample. Thedots show the data, and histograms show the MC predictionwith the contributions from neutrino interaction modes. TheMC distributions are tuned by the E ν scale factors obtainedby the spectrum fit. where ∆ M = M n − M p ; M indicate the muon, proton,or neutron mass with appropriate subscripts; E B is thenucleon binding energy; E µ is the reconstructed muonenergy. MiniBooNE E Recν prediction
To predict the E Recν distribution at MiniBooNE, weextrapolate the measured SciBooNE flux to MiniBooNEin two steps.First, we apply MiniBooNE/SciBooNE flux ratio tomake a prediction of the true neutrino energy distribu-tion at MiniBooNE. Then, we smear the true neutrinoenergy prediction to the reconstructed neutrino energy.Systematic uncertainties for the flux ratio is estimatedby varying the cross-section and flux models. Addition-ally, the uncertainties of the smearing function, whichconvert true E ν to E Recν , is estimated by varying thecross-section models.Finally, we add MiniBooNE detector response error tothe E Recν prediction.The predicted MiniBooNE reconstructed neutrino en-ergy distribution and its systematic uncertainties areshown in the Figure 4..
We test the oscillation hypothesis assuming the mixingbetween 2 neutrino flavors; ν µ and ν x . The ν µ → ν x disappearance probability is given as P ( ν µ → ν x ) = sin θ sin (1 . m L/E ) , where θ is the mixing angle, ∆ m [eV ] is the mass split-ting between 2 flavors, L[km] is the distance traveled andE[GeV] is the neutrino energy. (GeV) n Reconstructed E0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.80500010000150002000025000
Total err.Flux + X-sec. err.MiniBooNE det. err.
Preliminary
Fig. 4.
Predicted MiniBooNE reconstructed neutrino energydistribution. MiniBooNE detector error, flux and cross-sec-tion uncertainty, and the total systematic uncertainty are sep-arately shown.
We fit the MiniBooNE E Recν distribution to find thebest fit parameter minimizing the χ value: χ = bins X i,j ( N datai − N pi ) M − ij ( N dataj − N pj ) , where i, j denote E Recν bins, N datai,j and N pi,j denote ob-served and predicted number of events at each bin, re-spectively, and M ij represents statistical and systematicuncertainties for the final E Recν prediction.Then we define the allowed region by ∆ χ = χ ( true ) − χ ( best ) values, where χ ( true ) is the χ atthe oscillation prediction being tested, and χ ( best ) isthe smallest χ value across the (∆ m , sin θ ) plane.To obtain the confidence level at each oscillation pa-rameter point (∆ m , sin θ ), we use Feldman-Cousins’method [8]. In this method, 1000 “fake-data” predic-tions are formed, using random draws of the statisticaland systematic uncertainties and some underlying oscil-lation hypothesis. Then, each fake-data is fit to obtainthe relation between the ∆ χ values and the correspond-ing probabilities. This process is repeated for each pairof (∆ m , sin θ ) true oscillation parameter being tested. Expected Limit
The sensitivity is defined as the average of limits ob-tained from fake experiments with null underlying oscil-lation.Figure 5. shows the 90% CL. sensitivity for the ν µ dis-appearance. The expected ± σ band is also shown inthe plot. The expected sensitivity directly supersedesthe MiniBooNE only ν µ disappearance result, as sub-stantial flux and cross section uncertainties have beenreduced.
4. Summary and Prospects
We present SciBooNE-MiniBooNE joint analysis of asearch for ν µ disappearance in a accelerator neutrinobeam. The analysis is sensitive to the oscillation at the∆ m region of 0 . −
40 eV . The sensitivity to ν µ disap-pearance has been improved relative to the MiniBooNEshape-only analysis, with results to be released soon. Inaddition, a joint anti-neutrino oscillation analysis will beperformed using the anti-neutrino data set. q sin -2 -1
10 1 ] [ e V m D -1 CDHS 90% CLCCFR 90% CLMiniBooNE only 90% CL sensitivitySciBooNE + MiniBooNE 90% CL expected s – SciBooNE + MiniBooNE 90% CL
Preliminary
Fig. 5.
The expected sensitivity for ν µ disappearance. Thedotted curve shows the 90% CL limits from CDHS [9] andCCFR [10] experiments. The thin solid curve is the Mini-BooNE-only 90% CL sensitivity. The thick solid curve andthe filled region are the 90% CL sensitivity and ± σ bandfrom SciBooNE-MiniBooNE joint analysis, respectively.
5. Acknowledgements
SciBooNE collaboration gratefully acknowledges thesupport from various grants and contracts from the De-partment of Energy (U.S.), the National Science Foun-dation (U.S.), the MEXT (Japan), the INFN (Italy) theMinistry of Education and Science and CSIC (Spain),and the STFC (UK). We thank MiniBooNE collab-oration for various informations and simulation out-puts. The author was supported by Japan Society forthe Promotion of Science, and by the Grant-in-Aid forthe Global COE Program “The Next Generation ofPhysics, Spun from Universality and Emergence” fromthe MEXT of Japan.