arXiv:0808.0392v1 [hep-ex] 4 Aug 2008 etioOclainRslsfo IO n MiniBooNE and MINOS from Results Oscillation 2 1 al htnurn hsc a nee h el of realm this the from entered clear has is measurements. physics precision It neutrino con- final parameter. that biggest The given table the a [5]). with to from experimental experiment tribution (taken available the I all shows Table to in column shown fit is global data a from eters France. in experiment [4] the aeetbihdtesaleso h hr iigan- mixing third the gle, of reactors smallness nuclear the at established have measurements Finally, [3]. Observatory Neutrino (SNO) Sudbury oscil- the from neutrino 1960s, comes solar the a lations of as measurements After early precise as most . starting the oscillations experiments of in list measured long km. 250 been of also distance have mea- a beam at neutrino accelerator sured an using confirmed [2] result experiment this K2K The oscillations. neutrino r eae yauiaytasomto,tePMNS eigenstates the of transformation, sets unitary two a matrix: The by related neu- identical. are the not of are eigenstates interaction non- trino the and from and mass and non-zero that states fact neutrino the the of from masses follow degenerate They nomenon. Introduction 1. omnyprmtie ytrertto angles rotation three by matrix, PMNS parametrised the in commonly entries the constrained and tions eedn nnurn ao,eeg n eihan- characteristic zenith the to and pointing atmosphere energy [1], flavor, upper gle neutrino the in on disappear- produced dependent the neutrinos measured of Japan ance in experiment perK) phase complex a and oN:BotrNurn Experiment Neutrino Search Booster Oscillation BooNE: Neutrino Injector Main MINOS: h ol nweg nnurn silto param- oscillation neutrino on knowledge world The etiooclain r o eletbihdphe- established well a now are oscillations Neutrino eea xeiet aeosre etiooscilla- neutrino observed have experiments Several θ 13 h etlmto hsprmtrcmsfrom comes parameter this on limit best The . aaees eitouetocretnurn oscillation re neutrino a current and two oscillations introduce neutrino we to introduction parameters, brief a After UK Laboratory, Appleton Rutherford STFC Raufer M. Tobias ae rcs esrmnso h silto parameters oscillation the of measurements precise makes erve eeteprmna eut rmbt experiment both from results ex experimental LSND recent the by review reported We region parameter the in oscillations | ν α i δ = h ue-aikne(Su- Super-Kamiokande The . X i lvrPyisadC ilto ofrne api 2008 Taipei, Conference, Violation CP and Physics Flavor U αi ∗ | ν i i L/E intr of signature (1) θ ij eatoscoet h ore hsi oprdto study to compared order is in oscillations. detector This far neutrino the in- source. at neutrino measurement the of the to number large close the spec- teractions of energy use and making composition trum, detec- beam near the The measures tor respectively. target situ- production detectors trino two of utilises distances It at ated Injec- line. Main beam the (NuMI) at tor” “Neutrinos ’s at periment experiment MINOS The 2. interactions. neutrino current neutral measuring anomaly”. “LSND so-called this test to cally in given results section. previous experimental the the with incompatible ..SeieNeutrinos Sterile 1.1. E xeiet sn h it fthe of width the using experiments LEP al lblfi etiooclainparameters oscillation neutrino fit Global I Table sarsl yteLN xeiet[]wihclaimed which [7] experiment LSND for the evidence by result a is non-interacting neutrinos. of There sterile number called the precision. customarily on high neutrinos, limit at no three however is with consistent is [6], iwo h ol nweg fnurn oscillation neutrino of knowledge world the of view eiet hc ol eur e etiostate. neutrino new a require would which periment, | xeiet,MINOS experiments, ∆ IO saln-aeienurn silto ex- oscillation neutrino long-baseline a is MINOS by models neutrino sterile investigates also MINOS specifi- designed was experiment MiniBooNE The h ubro cienurns esrdb the by measured neutrinos, active of number The n oiainfrtesac o trl neutrinos sterile for search the for motivation One n iea ulo nftr measurements. future on outlook an give and s sin sin | sin ∆ Parameter ∆ m m m 2 2 2 32 2 (2 21 2 θ θ 32 2 | 13 12 n sin and θ | 23 ) ν µ 2 (2 (2 → (7 θ . . 4 23 6 ν ± etfit Best ∼ 0 ± .MnBoEtssneutrino tests MiniBooNE ). e < 0 1 . . 0 320 siltoswt assplitting mass a with oscillations mand km 1 0 50 n MiniBooNE and 0 . . 15) . 2) 5a 3 at 05 ± ± × × 0 0 . 10 ± . 063 10 023 1 − σ − σ ∼ 5 3 eV 3 mfo h neu- the from km 735 eV 2 2 2 MINOS . Experiment KamLAND CHOOZ Katm SK MINOS Z SNO 0 resonance 1 2 Flavor Physics and CP Violation Conference, Taipei, 2008

2.1. The NuMI beam in the LE beam configuration. The results reported here use a data sample of 2.5 × 1020 POT. The NuMI beam line uses 120 GeV protons from the Main Injector accelerator incident on a segmented The three analyses have many steps on common. graphite target where they produce mainly π- and K- Firstly, they all rely on a good understanding of mesons. These secondary particles are focused by two the details of the neutrino beam. The major un- parabolic horns and subsequently decay in a 675m known here is the production of hadrons in the NuMI long helium-filled decay volume to produce neutri- target. A detailed Monte Carlo simulation of the nos. The remaining hadrons are absorbed in a water- beam line components based on Fluka05 reproduces cooled beam dump at the end of the decay volume the data well but not perfectly. To further improve while muons produced in the decay are absorbed in data/Monte Carlo agreement, the hadron production the rock separating the near detector hall from the is parametrised as a function of xF and pt and a simul- hadron absorber. A small fraction of muons further taneous fit is performed to the near detector spectra in decay before being absorbed, producing electron and various different beam configurations. Individual neu- anti-muon neutrinos. The resulting neutrino beam trino interactions are reweighted according to the xF has the following composition: 92.9% νµ, 5.8% νµ, and pT of the parent pion (or kaon). Figure 1 shows 1.2% νe and 0.1% νe for the low energy (LE) beam near detector data and Monte Carlo for two different configuration. beam configurations before and after the tuning pro- The focusing peak and thus the neutrino energy cedure. A total of seven different beam configurations spectrum of the NuMI beam can be varied by moving the target with respect to the magnetic horns. This feature is unique to the NuMI beam and is used in all MINOS analyses in order to reduce systematic uncer- MINOS Near Detector tainties related to hadron production in the neutrino 2550 Low energy beam (x2) production target. High energy beam 2040 POT Tuned MC 16 2.2. The MINOS detectors Fluka05 MC 1530 The MINOS detectors are functionally identical steel/scintillator sampling calorimeters equipped with 1020 a magnetic field to allow charge sign determination of muon tracks. The far detector has a mass of 5.4kt and Events / GeV 10 is situated at a depth of 2130 m.w.e. in the Soudan Un- 105 derground Laboratory. The smaller near detector is ∼1 kt and is situated 100 m below ground at Fermilab. 0 The detectors are constructed from 2.5cm thick steel 0 5 10 15 20 30 50 planes mounted with a 1 cm thick layer of plastic scin- Reconstructed neutrino energy (GeV) tillator segmented into ∼4 cm wide strips. The scin- tillation light is collected by wavelength shifting fibres Figure 1: Near Detector data and Monte Carlo for the low and readout by multi-anode photomultiplier tubes. and high energy beam configurations. The grey lines show The detectors are calibrated using single particle re- the untuned Monte Carlo simulation based on Fluka05. sponse data collected with a smaller 12 t calibration The black lines show the Monte Carlo after hadron produc- detector at the CERN PS test-beam and a light in- tion tuning is applied. The agreement is much improved. jection system tracking PMT gain variations. Inter- detector calibration of the energy response is achieved using cosmic ray muons. were used in the fit. The thus improved hadron pro- duction model is then used in the prediction of the far detector neutrino energy spectrum. 3. MINOS data analysis Another common feature is the “blind analysis” technique. This involves not looking at the far detec- In this document, we present results from two dif- tor data until the complete analysis procedure – e.g. ferent analyses and give a status report on a third one. selection cuts, extrapolation, fitting, etc. – is defined. They can be classified by the main type of interaction The near detector data is not sensitive to oscillations of interest: νµ charged current, neutral current and νe at the atmospheric scale and was therefore accessible charged current. At the time of writing of this docu- throughout the development of the analyses, thus al- ment, MINOS has collected an integrated 4.8 × 1020 lowing to understand low level detector effects in the protons-on-target (POT), most of which was recorded data. Flavor Physics and CP Violation Conference, Taipei, 2008 3

3.1. νµ charged current analysis oscillations into sterile neutrinos, i.e. neutrino states with no interactions. The charged current disappearance analysis is the Neutral current events in MINOS are selected with main analysis of MINOS. Measuring the charged cur- a set of simple cuts using event shape variables. The rent energy spectrum in both detectors, MINOS mea- main distinguishing feature between charged current 2 sures the parameters |∆m32|and and neutral current events is the presence or lack of a 2 sin (2θ23). The analysis reported here is an update of muon in the final state leading to a reconstructed track our initial results published in [8]. in the detector. Short events, events with no track and Charged current candidate events are selected by events where the reconstructed track is shorter than ◦ demanding a reconstructed track within 53 of the the hadronic shower are classified as neutral current. beam direction with a vertex inside the fiducial vol- In the near detector, additional cuts on spacial sep- ume and in time with the beam spill. To further reject aration and timing are applied to make sure the event neutral current background events a likelihood-ratio is well reconstructed and not a remnant of another discriminant is constructed from six variables, reflect- bigger event in the vicinity. ing event topology, energy loss, inelasticity, etc. A In the far detector, an additional set of cuts remov- more detailed discussion of the charged current event ing cosmic rays and low energy noise is applied. The selection and extrapolation to the far detector for this resulting neutral current reconstructed energy spec- analysis can be found in [9]. trum is shown in Figure 3. Figure 2 shows the selected charged current energy spectrum corresponding to an exposure of 2.5 × 1020 POT. The points with error bars show the far de- 50 20 tector data and the black and red lines show un- MINOS Preliminary: 2.46 × 10 POT oscillated and oscillated Monte Carlo predictions re- Far Detector Data 40 θ spectively. The best-fit values for the oscillation 13 = 0 θ = 0.21, δ = 3π/2 30 13 MINOS Preliminary CC Background 140 20 ∆m2 = 2.38 × 10-3eV2 Un-Oscillated 2 120 Events/ GeV sin (2θ ) = 1 Best Fit 10 32 NC 100 Data 0 80 0 5 10 15 20 25 30

Events / GeV E (GeV) 60 vis

40 Figure 3: Far Detector reconstructed energy spectrum for 20 selected neutral current events. The data is shown as

20-30 GeV 30-50 GeV > 50 GeV black points with error bars, the Monte Carlo is shown 0 in blue and red for two extremes of the third angle θ13. 0 5 10 15 20 The charged current background is shown in pink. Reconstructed Neutrino Energy (GeV) The Monte Carlo expectation for this spectrum de- Figure 2: Reconstructed far detector neutrino energy spec- trum. The black markers with error bars show the data, pends on the value of the third mixing angle θ13 since the black and red lines show unoscillated and best-fit νe interactions have a high probability of being classi- Monte Carlo predictions. The neutral current background fied as neutral currents in this analysis. We therefore is shown in blue. show two Monte Carlo lines, one for θ13 = 0 (red) and one for appearance at the CHOOZ 2 +0.20 −3 2 limit (blue). The width of the red band reflects the parameters are: |∆m32|= 2.38−0.16 × 10 eV and 2 systematic uncertainties. For reconstructed energies sin (2θ23)= 1.00−0.08. The errors given combine both statistical and systematic errors. The latter were in- below 3GeV, a deficit with respect to the Monte Carlo cluded in the fit via nuisance parameters. expectation for θ13 = 0 of 1.15 σ is observed. The data is thus consistent with no active neutrino disappear- ance. 3.2. Neutral current analysis While the above discussion is model independent, it is also instructive to compare the data to a particu- If neutrino oscillations only involve the usual ac- lar model involving mixing between active and sterile tive neutrino flavours, the neutral current event rate neutrinos. We use a simple model introducing a sin- remains unchanged. The measurement of neutral cur- gle state and retaining one-mass-scale rent interactions therefore allows the investigation of dominance. Assuming no mixing of the first and sec- 4 Flavor Physics and CP Violation Conference, Taipei, 2008

ond mass eigenstates with the sterile state and assum- event is then reconstructed again in order to deter- ing that the first and fourth mass eigenstate are nearly mine, how often such an event is misclassified as an degenerate leads to the following oscillation probabil- electromagnetic shower. Using this method, a discrep- ities: ancy of 20 % is found, with the Monte Carlo overesti- mating the background. A second method compares 2 2 2 2 Pνµ→νµ = 1 − 4|Uµ3| (1 − |Uµ3| ) sin (1.27∆m32L/E) data taken with and without the focusing horns being 2 2 2 2 powered. This allows a deconvolution of the differ- Pνµ→νe = 4|Uµ3| |Ue3| sin (1.27∆m32L/E) 2 2 2 2 ent background components since the composition is → Pνµ νs = 4|Uµ3| |Us3| sin (1.27∆m32L/E) quite different in the two samples.

Pνµ→ντ = 1 − Pνµ→νµ − Pνµ→νe − Pνµ→νs . (2) The results obtained using the two different meth- ods are consistent, justifying an ad-hoc correction The charged and neutral current energy spectra are which is applied to the Monte Carlo prediction. The fitted simultaneously to this 4-flavour model under background estimates are extrapolated to the far de- two different scenarios. The first scenario assumes no tector where the sensitivity to measuring θ13 is evalu- 2 νe appearance, i.e. |Ue3| = 0, the second assumes νe ated. These sensitivities for three different exposures 2 appearance at the CHOOZ limit, |Ue3| = 0.04. The are shown in Figure 4. results are:

2 +0.18 2 +0.16 2 MINOS Projected 90% Exclusion Region |Us3| =0.14−0.13 |Uµ3| =0.50−0.15 (|Ue3 | = 0 ) 2 2 +0.20 2 +0.18 2 |Us3| =0.21−0.12 |Uµ3| =0.48−0.12 (|Ue3 | = 0 .04 ) CHOOZ 90% CL 3.25x1020 POT ∆m2 > 0 (2008, 10%) 1.5 32 20 ∆ 2 Systematic errors were included in the fit as nuisance 3.25x10 POT m32 < 0 (2008, 10%) 20 ∆ 2 parameters; the errors quoted above thus include both 6.5x10 POT m32 > 0 (+ ~1 year, 5%) ) 20 ∆ 2

π 9.5x10 POT m > 0 (+ ~2 years, 5%) statistical and systematic uncertainties. The values 32 2 2 ( 1 for |Uµ3| are consistent with the parameter sin (2θ23) δ reported in section 3.1. More details on the analysis of neutral current in- 0.5 |∆ m2 | = 2.4x10-3 eV2 teractions in MINOS can be found in [10]. 32 2 θ sin (2 23) = 1.0 0 ν 0.1 0.2 0.3 0.4 0.5 3.3. e appearance analysis 2 θ PRELIMINARY sin (2 13) The νe appearance analysis investigates the sub- dominant oscillation channel νµ → νe which is sensi- Figure 4: Sensitivity to θ13 for three different exposures. 20 tive to the third angle in the PMNS matrix, θ13. The For our current data set of 3.25 × 10 the sensitivity for non-observation of the reverse process at the CHOOZ both the normal (blue) and inverted (red) mass hierar- reactor experiment [4] provides an upper limit for this chies are shown. A 10% systematic error is included in all parameter. A potential observation of νe events in predictions. With increasing statistics, improvements over the MINOS far detector beyond the inherent level in the CHOOZ bound can be made. the NuMI beam would provide evidence for a non-zero value below the CHOOZ limit. The data for this analysis are still blinded and final νe events are characterised by an electromagnetic studies are being performed, including the investiga- shower in the detector in addition to the hadronic ac- tion of data “sidebands”. Results from this analysis tivity caused by the struck nucleus. The success of this are eagerly awaited and are expected later this year. analysis therefore relies on the ability to identify elec- tromagnetic showers in a large background of hadronic activity. Due to the relatively coarse segmentation of 4. The MiniBooNE experiment the MINOS detectors, this is a very difficult task. νe candidates are selected using a neural net- MiniBooNE is a short-baseline neutrino oscillation work technique. Since these techniques rely on good experiment using Fermilab’s booster neutrino beam data/Monte Carlo agreement, two independent data line. With a source-detector distance of ∼0.5 km and driven methods using the near detector have been de- a peak neutrino energy of 0.8 GeV it is specifically veloped in order to improve the modeling of the back- designed to test the parameter space of the LSND grounds. This is possible since the baseline at the near claim discussed in section 1.1. detector is too short for oscillations to develop. MiniBooNE uses a conventional neutrino beam sim- One method takes well understood, well recon- ilar to the NuMI beam described above. The booster’s structed charged current events and removes the muon 8 GeV protons impinge on a beryllium target at a rate track, leaving a sole hadronic shower. This remnant of up to 4Hz. The secondary mesons are focused by Flavor Physics and CP Violation Conference, Taipei, 2008 5 a single pulsed magnetic focusing horn and allowed to decay in a 50m-long decay region. The detector is sep- 4.0 MiniBooNE data (stat. error)y arated from the end of the decay volume by ∼500m 3.5 expected background (syst. error)µ ν of earth. 3.0 µ The MiniBooNE detector consists of a spherical vol- νµ background 2.5 ν ume filled with 800 tons of mineral oil (CH2), acting e background both as a scintillator and Cerenkov medium. The light 2.0 produced in the detector is read out by 1280 inward 1.5 facing 8-inch photomultiplier tubes (PMTs) provid- events / MeV ing a 10% photo-cathode coverage, viewing a target 1.0 volume 575 cm in diameter. The target volume is sur- 0.5 rounded by a 35 cm thick, optically isolated veto re- 300 500 700 900 1100 1300 1500 3000 gion viewed by 240 PMTs. reconstructed Eν (MeV)

Figure 5: Reconstructed neutrino energy spectrum from 5. MiniBooNE oscillation analysis MiniBooNE for data (black dots) and Monte Carlo (ma- genta line). The data is shown with statistical error bars and the Monte Carlo shows the expected non-oscillation The MiniBooNE analysis was conducted in a background including systematic errors. A low energy blinded fashion. Only Monte Carlo simulation and threshold of 475 MeV was applied to the analysis. data insensitive to νµ → νe oscillations was used to develop the analysis procedure. As in the MINOS experiment, hadron production hypothesis. The disagreement was found to come uncertainties in the target are an important contrib- from the lowest energy events. It was therefore de- utor to the systematic error for MiniBooNE analyses. cided to move the energy threshold for the analysis to MiniBooNE uses a Sanford-Wang parametrisation to 475 MeV, while still showing the lower energy events the pion data from the HARP experiment [11] and to the public. The resulting visible energy spectrum a fit to the world kaon production data in the range for data and Monte Carlo is shown in Figure 5. of 10–24 GeV to improve on this uncertainty. The quoted uncertainties are 17 % for the pion and 30 % Above the energy threshold, the data does not show a significant excess and is consistent with no for the kaon flux. The predicted kaon flux is checked 2 using off-axis muon counters and high energy events electron-neutrino appearance at the 1 eV scale. Mini- BooNE is therefore able to exclude a region in the in the MiniBooNE detector. 2 2 To select electron-like neutrino interactions in ∆m –sin 2θ parameter space. Figure 6 shows how MiniBooNE, events are first divided into so-called the MiniBooNE result compares with previous exper- subevents, i.e. sets of PMT hits which are clustered iments Bugey, KARMEN and LSND. It can be seen within ∼100 ns. Only subevents in time with the beam that MiniBooNE rules out large parts of the LSND spill are selected. Subevents are required to have favoured region (shown in blue). Assuming that neu- more than 200 hits in the tank, less than 6 hits in trinos and antineutrinos oscillate in the same way and the veto region and a reconstructed vertex position using a two-neutrino approximation, LSND and Mini- with R < 500 cm, eliminating cosmic ray muons and BooNE are incompatible at 98 % C.L. Several potential sources for the reported excess of associated decay electrons. νe charged current event candidates a required to have exactly one subevent. events below the 474 MeV threshold have been dis- After these initial selection cuts, two visible en- cussed and are currently being investigated by the col- ergy dependent likelihood ratio discriminants are con- laboration. At the time of writing, no new results on structed in order to further reduce the backgrounds. this issue were available. The first one separates muon-like from electron-like Cerenkov rings. The second discriminant separates the remaining events into electron-like events and 6. Summary events from π0-decay. For the complete selection, the Monte Carlo predicts an efficiency for νe charged cur- rent quasi-elastic events of 20.3±0.9%. Further details Neutrino oscillations have been introduced and ev- on the event selection can be found in [12]. idence and current world knowledge on oscillation The MiniBooNE collaboration decided to open their parameters from several experiments have been re- blinded data set in stages, revealing more and more ported. We reported the latest neutrino oscillation re- details about the data while still blind to the oscil- sults from the MINOS and MiniBooNE experiments, lation result. During this unblinding procedure, it detailing aspects of the MINOS νµ charged current, was found that data and Monte Carlo expectation re- neutral current and νe appearance analysis as well as sulted in a bad χ2 for both the oscillation and null the MiniBooNE two-flavour oscillation analysis. 6 Flavor Physics and CP Violation Conference, Taipei, 2008

The author would like to thank the MiniBooNE collaboration and Rustem Ospanov for the material 102 provided. Furthermore, we would like to thank the organisers of FPCP08 for the help provided. MiniBooNE 90% C.L. KARMEN2 90% C.L. 10 Bugey 90% C.L. ) 4

/c References 2 1 | (eV 2

m [1] Y. Ashie et al. [Super-Kamiokande Collabora- ∆ | tion], Phys. Rev. D 71 112005 (2005). [2] M. H. Ahn et al. [K2K Collaboration], Phys. Rev. 10-1 D 74 072003 (2006). LSND 90% C.L. [3] Q. R. Ahmad et al. [SNO Collaboration], Phys. LSND 99% C.L. Rev. Lett. 89, 011301 (2002) -2 [4] M. Apollonio et al. [CHOOZ Collaboration], Eur. 10 -3 -2 -1 10 10 10 1 Phys. J. C 27, 331 (2003) 2 θ sin (2 ) [5] M. Maltoni, T. Schwetz, M. A. Tortola and J. W. F. Valle, arXiv:hep-ph/0405172v6. Figure 6: The MiniBooNE 90% exclusion curve is shown [6] ALEPH Collaboration and DELPHI Collabora- as a solid black line along with the curves from the KAR- tion and L3 Collaboration and OPAL Collabora- MEN and Bugey experiments (dashed). The LSND al- 427 lowed regions are shown in blue. MiniBooNE and LSND tion, Phys. Rept. , 257 (2006) are incompatible at 98% C.L. [7] A. Aguilar et al. [LSND Collaboration], Phys. Rev. D 64, 112007 (2001) [8] D. G. Michael et al. [MINOS Collaboration], Acknowledgments Phys. Rev. Lett. 97, 191801 (2006) [9] P. Adamson et al. [MINOS Collaboration], The MINOS work reported here was supported arXiv:0708.1495 [hep-ex]. by the US DOE; the UK STFC; the US NSF; the [10] P. Adamson et al. [MINOS Collaboration], State and University of Minnesota; the University of arXiv:0807.2424 [hep-ex]. Athens, Greece and Brazil’s FAPESP and CNPq. MI- [11] M. G. Catanesi et al. [HARP Collaboration], NOS would like to thank the Minnesota Department Nucl. Phys. B 732, 1 (2006) of Natural Resources, the crew of the Soudan Under- [12] A. A. Aguilar-Arevalo et al. [The MiniBooNE ground Laboratory, and the staff of Fermilab for their Collaboration], Phys. Rev. Lett. 98, 231801 contributions to this effort. (2007)