Measuring Neutrino Oscillations with the T2K Experiment Alfons Weber University of Oxford STFC/RAL Dec-2011

Overview Introduction The Hardware –Beam –Near Detector Complex –Far Detector (SuperKamiokande) Analysis & Results –Muon neutrino disappearance –Electron neutrino appearance –Cross section systematics Outlook and Conclusion Dec

Neutrino Mixing The PMNS Matrix Assume that neutrinos do have mass: –mass eigenstates  weak interaction eigenstates –Analogue to CKM-Matrix in quark sector! weak “flavour eigenstates” Mass eigenstates m 1, m 2, m 3 Unitary mixing matrix: 3 mixing angles & complex phases Pontecorvo-Maki- Nakagawa-Sakata Dec

Neutrino Oscillations If mass and weak eigenstates are different: –Neutrino is produced in weak eigenstate –It travels a distance L as a mass eigenstate –It will be detected in a (possibly) different weak eigenstate Dec

T2K Experiment Dec

T2K Collaboration International collaboration (~500 members, 59 institutes, 12 countries ） Dec

Super-K Detector Located in Mozumi mine –2700 m.w.e. overburden Fiducial mass –22.5 kt Water Cherenkov detector –Inner detector ~ inch PMTs –Outer detector (veto) ~ inch PMTs New DAQ system –100% livetime Excellent μ /e separation Dec

Particle Identification in SK e -- 8

Neutrino interactions Below ~1 GeV CCQE dominates –Clear 1-ring signature in Super-K –Simple kinematics → calculate E ν At higher energies resonance/DIS important –Background to CCQE signals –Want narrow-band beam 1 st oscillation maximum NC1π CC1π Dec

Design Principle Main feature : off-axis beam to reduce high-energy tail –Narrow-band beam around oscillation maximum –Feed-down from miss-reconstructed DIS/resonance events at SK into analysis region reduced. π→μν 118 m off-axis (2.5°) (30 GeV from MR synchrotron) 0º0º Dec

Beamline Dec

Beamline MC Flux estimated from beamline MC with inputs from data Primary beam data from beamline monitors Hadronic interactions –Pions - use CERN NA61/SHINE pion measurement (large acceptance: >95% coverage of ν parent pions) –Pions outside NA61 acceptance, other interaction (inc. kaons) based on FLUKA simulation –Secondary interactions outside the target based on experimentally measured cross-sections GEANT3 transport simulation used downstream of target Dec

NA61/SHINE Goal: measure hadron(π, K) yield distribution in 30 GeV p + C inelastic interaction High-acceptance ToFs and spectrometers 2cm thin target - 4%λ I π + analysis: – dE/dx only analysis for low momenta – dE/dx+ToF selection for high momenta ~13m ~10m Dec

Predicted Flux ν μ fluxes in analysis region dominated by pion decays –Kaons important in tail ν e flux in analysis region dominated by muons –From decay chain –Primary pions modelled with NA61 data νμνμ νμνμ νeνe Dec

Data Sample All good physics data taken to date was used for this analysis: –Run 1 – 3.23x10 19 p.o.t with 50kW beam, Jan '10 – Jun '10 –Run 2 – 11.08x10^19 p.o.t. with 145kW beam, Nov '10 – Mar '11 Total 1.43x10 20 p.o.t. – 2% of T2K final goal accumulated # of protons proton per pulse Run 1 Run 2 50kW 145kW Dec

Near detectors Muon monitor –spill-spill monitoring On-axis detector ( INGRID ) measures beam intensity/direction – 1 mrad precision in 1 day –Corresponds to 10 MeV shift in peak E ND280 detector at same off-axis angle as SK –Detailed flux measurement –Exclusive cross-section measurements ν beam Dec

MUMON & INGRID INGRID –Daily measurement pointing Stability MUMON –Spill-by-spill stability Dec

ND280 detector 0.2 T magnet –Re-commissioned UA1/NOMAD Plastic scintillator detectors –Fine Grained Detector (FGD) –π 0 detector (P0D) –ECals and SMRD Time projection chambers (TPC) –<10% dE/dx resolution –10% momentum 1 GeV/c Analysis use total ν μ CC interaction rate in FGDs only νμνμ FGD2 TPC3 TPC2 TPC1 FGD1 dE/dx (TPC: data) Dec

ND280 Input Measure inclusive ν μ -CC event rate in near detector –Total event rate, no shape measurement for this analysis –Events with vertex in FGD and muon-like track in TPC selected –Achieve purity of 90% ν μ -CC Good agreement of data with beam MC + neutrino interaction generator, without any tuning to ND data. Dec

Dec

Far Detector Events Selection Fixed before looking at data Requirements –fully contained –consistent with expected arrival time –1 ring –muon like Compare energy spectrum with expectation –Beam simulation –Near detector measurement Dec

Energy Spectrum Fit to muon neutrino energy spectrum –Main results without fitting systematics –alternative analysis fitting systematic parameters –Identical sensitivity Result Dec

Dec

A Little Bit of Math Dec

Appearance analysis Basic idea –Apply selection criteria to T2K far detector data to isolate ν e -CCQE events –Compare with expected number of background events → measure appearance probability Backgrounds – Intrinsic ν e contamination : μ, K decays in decay volume – NC- π 0 interactions of ν μ missed or merged gamma-rays → single e-like ring detected γ (lost) γ (e-like ring) Dec

Event Rate Events at T2K far detector – Signal events, depending on ν μ flux and oscillation parameters – ν e background, depending on intrinsic beam ν e flux – ν μ background, depending on ν μ flux Inputs to event number estimation – ND280 ν μ -CC event rate – Beam MC predictions for near and far detector rates, F/N ratio from MC Measured ND280 event rate Dec

Total BG Estimate Final estimates –multiply MC predictions by data/MC ratio of event rate in ND: Calculate total for 1.43 x p.o.t., θ 13 =0: ν μ CC BG is insignificant Dec

Systematic errors Number of expected events at T2K far detector Evaluate systematics for each term smaller uncertainty on signal acceptance compare to BG rejection Dec

Selection (1) Fiducial volume cut FC events, R<1490cm FC events, |Z|<1610cm Only consider fully-contained (FC) events. Distance between recon vertex and wall >200cm. Dec

Selection (2) Single e-like ring e-like μ-like (sin 2 2θ 13 = 0.1) Dec

Selection(3) Visible energy >100MeV No decay electron (sin 2 2θ 13 = 0.1) Dec

Selection (4) “P0LFit” invariant mass cut Reconstructed nu energy <1250MeV (sin 2 2θ 13 = 0.1) Dec

Final Selection 6 candidate ν e events after all cuts Expected BG of 1.5±0.3 events for sin 2 2 θ 13 =0 Dec

Angular distribution (sin 2 2θ 13 = 0.1) Dec

Vertex distribution Most events clustered at high R in upstream part of FV –No corresponding excess in OD or outside FV - no evidence of any plausible background contamination Difficult to meaningfully calculate probability of this distribution after the fact –K.S. test on the R 2 distribution yields a p-value of 3% –Next analysis will define procedure to check this distribution before looking at the data Outside FV beam Dec

Final Result Dec-2011 p-value of 0.7% for null hypothesis Equivalent to 2.5 σ rejection of sin 2 2 θ 13 =0 6 candidate ν e events after all cuts Expected BG of 1.5+/-0.3 events for sin 2 2 θ 13 =0 36

Oscillation fit Contour calculated using Feldman-Cousins method: –Ensure correct coverage by performing many toy experiments at each point in ( θ 13, δ CP ) space and finding δχ 2 crit (θ 13, δ CP ) s.t. (68%, 90%) of toy experiments have δχ 2 < δχ 2 crit at true (θ 13,δ CP ) –Exclude points in ( θ 13,δ CP ) space where δχ 2 (θ 13, δ CP ; data ) > δχ 2 crit (θ 13, δ CP ) – Systematics included in toy experiment generation Dec-2011 d=0 sin22th23= 1 37

Comparison Overlay of MINOS, T2K and DC allowed region (NOT a combined fit) * arXiv: * DoubleChooz (68%CL) DC  68%CL  Dec

Combined Fit Dec From T. Schwetz

Dec

Flux Uncertainties Dominated by hadron production modelling –Pion production: NA61 bin-by-bin systematics –Kaon and secondary nucleon production Compare FLUKA with data –Secondary interaction cross-section compare FLUKA/GCALOR with data (NA61 thin-target) Dec-2011 External data used T. Eichten et al., Nucl. Phys. B 44, 333 (1972) J. V. Allaby et al., CERN Some cancellation between ND280 and SK 41

Far Detector Uncertainties Large uncertainty in NC- π 0 rejection Evaluated using hybrid data/MC sample –one data electron –one simulated gamma. Systematic error calculated from difference between this and pure MC sample. Other uncertainties constrained using large ν e -enriched atmospheric data Dec

Cross Sections Uncertainty Methods to estimate cross-section uncertainties –Data/MC comparison –Comparison of physics models –Tweaking model parameters Evaluate resulting error on F/N ratio by –varying cross-sections Important processes: –Far detector NC- π 0, ν e - CCQE –Near detector ν μ -CC inclusive Dominant error is pion FSI Dec

Total Systematics Dec-2011 Smaller for θ 13 ≠ 0 due to small uncertainties on signal 44

Summary T2K had a successful initial run –2% of the expected data set Beam flux errors are small –Cross sections measurements will be possible Analysis of neutrino oscillations –Disappearance measurement –Evidence for non-zero θ 13 T2K only: 2.5 σ Reactor and LBL: >3 σ Watch this space! Dec