First Anti-Neutrino Oscillation Results from the T2K Experiment Alfons Weber University of Oxford STFC/RAL

Measurements Overview Introduction The Experiments Summary Beam Near Detectors Super-Kamiokande Measurements Disappearing muon anti-neutrinos Appearing electron anti-neutrinos Other measurements Summary Sep 2015 Alfons Weber

Neutrino Mixing The PMNS Matrix Pontecorvo-Maki- Nakagawa-Sakata Assume that neutrinos do have mass: mass eigenstates  weak interaction eigenstates Analogue to CKM-Matrix in quark sector! Normal Hierarchy Inverted Hierarchy weak “flavour eigenstates” Mass eigenstates m1, m2, m3 Unitary mixing matrix: 3 mixing angles & complex phases Sep 2015 Alfons Weber

Who is Who beamνμ disappearance Solar & reactor ν–less double beta decay νe appearance in νμ beam or νe disappearance in reactor experiments Sep 2015 Alfons Weber

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 Sep 2015 Alfons Weber

Muon Neutrino Disappearance Monte Carlo (MINOS Experiment) (Input parameters: sin22θ = 1.0, Δm2 = 3.35x10-3 eV2 ) spectrum ratio νμ spectrum Unoscillated Oscillated Sep 2015 Alfons Weber

A Little Bit of Math Sep 2015 Alfons Weber

Collaboration Sep 2015 Alfons Weber

T2K Experiment Sep 2015 Alfons Weber

T2K experimental overview The oscillation probability for numu disappearnce, is sinusoidal, and dpeneds on L and E, the distance the neutrinos travel and their energy. In order to maximinze this probability, this argument would need to be of order 1), and also to observe the lepton through charged current interactions, then I need to go hundreds of km. Long baseline experiments are driven by intense neutrino sources, and are tuned to a beam energy relevant for oscillation physics. J-PARC accelerator Near detectors Off-axis: ND280 On-axis: INGRID Far detector Sep 2015 Alfons Weber

Design Principle 0º Sep 2015 Alfons Weber

Neutrino Interactions W Charged Current Quasi-Elastic (CCQE) ν p n Neutrino energy from lepton momentum and angle: e- or - 2 body kinematics and assumes the target nucleon is at rest The beam travels essentially a fixed distance L, and we probe the oscillation dip around the relevant region in energy. T2K’s peak flux energy is ~0.6 GeV, so the dominant interaction is charged current quasi-elastic interactions (CCQE). I’ll talk later about why this is an oversimplification, but for now the neutrino interactons on a bound neutron, and produces a muon or an electron in the detector. The kinematics of the lepton are used to infer the neutirno energy, assuming simple 2 body inematics. Some relevant background processes are those which mimic CCQE, such as single pion production. In this case there is a lepton and a pion in thef inal sate. When the pion is absorbed or not reconstructed, then this event appears “CCQE-like” but is actually due to a higher energy event. This smears out our determination of the oscillation probability. Neutral current events can also be a background to the nue appearnce (NCpi0) or numu disappearance analysis (Ncpi+) W ν N N’ Δ π CCπ NCπ Additional significant processes: CCQE-like multinucleon interaction Charged current single pion production (CCπ) Neutral current single pion production (NCπ) Z ν N N’ Δ π e- or  - Sep 2015 Alfons Weber

Beamline MC Hadronic interactions Pions/Kaons - use CERN NA61/SHINE pion measurement (large acceptance: >95% coverage of ν parent pions) Pions outside NA61 acceptance, other interaction use FLUKA simulation Secondary interactions outside the target based on experimentally measured cross-sections GEANT3 transport simulation used downstream of target Sep 2015 Alfons Weber

NA61/SHINE @ CERN ~13m ~10m Hadron(π, K) yield measurement 30 GeV proton High-acceptance ToFs and spectrometers Measurement program 2 cm thin target - 4%λI T2K Replica target ~13m ~10m Sep 2015 Alfons Weber

Flux Prediction Flux prediction external or in-situ measurements p-C data alignment and off-axis angle π± , K± production from NA61/SHINE The distribution of momentum and production angle for secondary pi+ that produce a neutrino (directly or through daughter particles) in the 250 kA flux.  The phase space of the NA61 2009 (new) and 2007 (old) thin target data are also shown by the region outlined with black lines.  The reduced coverage at low momentum and high angle for 2009 are due to the fact tehat the preliminary 2009 data only use the TOF-dE/dx analysis.  In that region, the 2007 data is used for tuning. NA61 has thick target data to be analyzed. Modified BMPT prameterization http://www.t2k.org/docs/plots/033/fluxplots2015/offcialplots2015 <10% tuning applied to all channel Uncertainties are comparable for neutrino or antineutrino mode operation (10-15%) Sep 2015 Alfons Weber

Flux Prediction II µ- µ+ π- π+ νµ νµ Neutrino mode operation Antineutrino mode operation FLUKA/Geant3-based neutrino beam simulation (PRD 87, 012001) Significant neutrino component to antineutrino mode beam Increases in event rate due to lower antineutrino cross section “wrong sign” component “Intrinsic” ~0.5% electron (anti)neutrino component The beam travels essentially a fixed distance L, and we probe the oscillation dip around the relevant region in energy. T2K’s peak flux energy is ~0.6 GeV, so the dominant interaction is charged current quasi-elastic interactions (CCQE). I’ll talk later about why this is an oversimplification, but for now the neutrino interactons on a bound neutron, and produces a muon or an electron in the detector. The kinematics of the lepton are used to infer the neutirno energy, assuming simple 2 body inematics. Some relevant background processes are those which mimic CCQE, such as single pion production. In this case there is a lepton and a pion in thef inal sate. When the pion is absorbed or not reconstructed, then this event appears “CCQE-like” but is actually due to a higher energy event. This smears out our determination of the oscillation probability. Neutral current events can also be a background to the nue appearnce (NCpi0) or numu disappearance analysis (Ncpi+) µ- µ+ π- π+ νµ νµ Sep 2015 Alfons Weber

Data Taking Maximum beam power is 371kW! Stable operation at ~345kW achieved! 11.0x1020 (total) = 7.0x1020 (ν) + 4.0x1020 (ν-bar) Partial POT for ND280, so confirm beam stability Full antineutrino up to June 2015 Protons on Target (POT) for the antineutrino analyses today: Run 5c+6 datasets for far detector, Super-K: 4.0 x 1020 POT Run 5c datasets for off-axis near detector, ND280: 4.3 x 1019 POT Sep 2015 Alfons Weber

Near detectors Muon monitor INGRID (On-axis detector ) spill-by-spill monitoring INGRID (On-axis detector ) measures beam intensity/direction 1 mrad precision in 1 day ND280 (same off-axis angle as SK) Detailed flux measurement Exclusive cross-section measurements ν beam Sep 2015 Alfons Weber

Beam Stability neutrino beam profile measured with INGRID Maximum beam power is 371kW! Stable operation at ~345kW achieved! 10.3x1020 (total) = 7.0x1020 (ν) + 3.3x1020 (ν-bar) Partial POT for ND280, how does the rest of the running look? neutrino beam profile measured with INGRID scintillator/iron tracking detectors (0-0.9 degrees off-axis) POT normalized event rate stable to better than 1% Beam direction is stable to within 1 mrad corresponds to a 2% shift to peak in off-axis neutrino energy Sep 2015 Alfons Weber

Analysis Strategy Flux Model ND280 Detector Model NA61/SHINE Data ND280 Data INGRID/Beam monitor Data ND data reduces flux and cross-section uncertainties ND280 Fit External Cross-section Data Cross-section Model Oscillation Fit Super-K Data Super-K Detector Model Oscillation Parameters Sep 2015 Alfons Weber

Reducing Uncertainties Neutrino flux prediction Neutrino cross section model Far Detector selection, efficiency Neutrino flux prediction Neutrino cross section model Near Detector selection, efficiency N_{FD}\sim\Phi(E_\nu)\sigma(E_\nu)\epsilon_{FD}P(\nu_\mu\rightarrow\nu_e) N_{ND}\sim\Phi(E_\nu)\sigma(E_\nu)\epsilon_{ND} Sep 2015 Alfons Weber

νμ ND280 Detector dE/dx (TPC: data) 0.2 T magnet Recycled from UA1@ CERN Recycled from NOMAD @ CERN Plastic scintillator detectors: Fine Grained Detector (FGD) 1.6 ton fiducial mass for analysis π0 detector (P0D) ECals and SMRD Time projection chambers (TPC) better than 10% dE/dx resolution 10% p resolution at 1GeV/c Micromegas from CERN Example: neutrino candidate in antineutrino mode Muon-like track TPC ECAL FGD νμ FGD2 TPC3 TPC2 TPC1 FGD1 dE/dx (TPC: data) Sep 2015 Alfons Weber

neutrino selection, neutrino mode samples ND Event Samples (I) CC1π CCOther neutrino selection, neutrino mode samples CC0π Select CC νμ candidates based on interactions with μ- Select highest momentum track with negative charge, and PID consistent with a muon Event samples provide information on flux, cross section model Separated based on presence of charged pion in final state (CC0π, CC1π, CC Other) Pions identified using track multiplicity, dE/dx in TPCs, photons in ECALs Dominant detector systematics are pion re-interactions, MC statistics and out of fiducial volume backgrounds Example: neutrino candidate in antineutrino mode Sep 2015 Alfons Weber

ND Event Samples (II) Select CC νμ candidates based on interactions with μ+ highest momentum track, positive charge, and PID consistent with muon Two sub-samples based on track multiplicity: CC1-Track, CC>1 Track Complementary selection of neutrino candidates in antineutrino mode CC1Track: antineutrino selection, antineutrino mode CC inclusive: neutrino selection, antineutrino mode Dominant detector systematics are pion re-interactions, MC statistics and out of fiducial volume backgrounds http://www.t2k.org/docs/plots/033/nd280-numu-plots-tn-212 http://www.t2k.org/docs/plots/033/nubar/plots1/cc1trk_reac_mom_niwg_0var.png http://www.t2k.org/docs/plots/033/nd280-numubkg-measurement/NuMuBkgWithNIWG Sep 2015 Alfons Weber

Near Detector Fit (I) Far detector expectations tuned with fit to ND samples (p,θ) (Anti-)Neutrino fluxes highly correlated between ND and FD Cross sections highly correlated Significant reduction to overall uncertainties CC0π: neutrino selection, neutrino mode Dominant detector systematics are pion re-interactions, MC statistics and out of fiducial volume backgrounds http://www.t2k.org/docs/plots/033/banff-plots C1Track: Antineutrino selection, antineutrino mode CC1π: neutrino selection, neutrino mode Sep 2015 Alfons Weber

Near Detector Fit (II) Predicted flux at FD is generally increased Some cross-section parameters are significantly different to prior values In general error on parameters is decreased PRELIMINARY νμ-flux in νμ-mode beam PRELIMINARY Sep 2015 Alfons Weber

Super-K Detector Located in Mozumi mine Water Cherenkov detector 2700 m.w.e. overburden Water Cherenkov detector 22.5 kt fiducial mass Inner detector 11000 20-inch PMTs Outer veto 1900 8-inch PMTs New DAQ system No deadtime Excellent μ/e separation Probability to reconstruct μ as e ~1% Sep 2015 Alfons Weber

Particle Identification μ-like MC e-like MC Sep 2015 Alfons Weber

PID (II) Sep 2015 Alfons Weber

Previous T2K Measurements 90% CL constrain on δCP First measurement νe appearance (7.3 σ) Measurement of θ13 (w/ & w/o reactor constrain sin2θ13=0.095±0.001) Open Questions Mass Hierarchy CP Phase δCP World-leading measurement of θ23 Significant measurement of Δm223 Abe, K. et al, Physical Review D 91.7 (2015): 072010 Sep 2015 Alfons Weber

T2K Oscillation Analysis Muon antineutrino disappearance Fit for and Use separate parameters for neutrino interactions Other oscillation parameters fixed to T2K neutrino data and PDG2014 Test of NSI or CPT theorem Electron antineutrino appearance Search for presence of appearance with antineutrinos Necessary step toward future CPV searches \bar{\theta}_{23} Sep 2015 Alfons Weber

T2K Search for Muon neutrinos Sep 2015 Alfons Weber

 Selection (I) FD Event selection Muon-like PID Fully contained in fiducial volume 1 ring ≤1 decay electron(s) pμ > 200 MeV Beam direction Fiducial volume boundary Events during run 5 Events during run 6 Out of fiducial volume events Sep 2015 Alfons Weber

 Selection (II) Sep 2015 Alfons Weber

Predicted Event Rates Predominantly antineutrino interactions, but significant components from other channels Expect 34.6 events w/ oscillation Expect 103.6 events w/o oscillation νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Add selection? Sep 2015 Alfons Weber

Systematics PRELIMINARY The near detector significantly reduces the systematic uncertainty in the predicted event rate at far detector Systematic Without ND With ND Flux and Cross-section Common to ND280/SK 9.2% 3.4% Super-K Only Multi-nucleon effect on oxygen 9.5% All Super-K Only 10.0% All 13.0% 10.1% Final State Interaction/Secondary Interaction at Super-K 2.1% Super-K Detector 3.8% Total 14.4% 11.6% Explain plot: “Prefit” = MC prediction without ND280 constraint, “Postfit” = MC prediction with ND280 constraint Note this is all for numubar (nuebar similar magnitude) ND Measurements on water not jet included Sep 2015 Alfons Weber

Antineutrino Disappearance νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Add selection? Likelihood based estimation of oscillation parameters Binned in reconstructed neutrino energy Other oscillation parameters fixed to T2K neutrino data and PDG2014 34 events observed 103.6 expected (w/o osc.) Sep 2015 Alfons Weber

Comparison with MINOS Results compatible with MINOS combined beam+atm νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Results compatible with MINOS combined beam+atm P. Adamson et al., Phys. Rev. Lett. 110 (2013) 25, 251801 Sep 2015 Alfons Weber

Comparison with Neutrino Result νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Sep 2015 Alfons Weber

T2K Search for Electron neutrinos Sep 2015 Alfons Weber

Identifying e Sep 2015 Alfons Weber

Anti-neutrino e-like selection e Events FD Event Selection Electron-like PID In fiducial volume 1 ring No decay electrons pe > 100 MeV Eν < 1250 MeV Pass π0-rejection Anti-neutrino e-like selection 3 events Vertex X (cm) Vertex Y (cm) PRELIMINARY Beam direction Fiducial volume boundary Events during run 5 Events during run 6 Out of fiducial volume events Sep 2015 Alfons Weber

νe Selection Sep 2015 Alfons Weber

Appearance Analysis β scales green Normal hierarchy Inverted hierarchy νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Test of no e appearance hypothesis:  = 0 or 1 Sep 2015 Alfons Weber

Data does not favor or disfavor e appearance Rate Measurement Generate an ensemble of test experiments with β=0 (no e appearance) p-value: fraction of test experiments that have as many or more candidate events as T2K data Sensitivity: mean p-value for an ensemble of fake data experiments with β=1 Rate only p-value Likelihood ratio 3 evts 0.26 0.9 Rate only p-value Mean p-value 0.20 Likelihood ratio: L(β=0)/L(β=1) is close to 1 νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy Data does not favor or disfavor e appearance Sep 2015 Alfons Weber

Rate & Shape 4646 Background only Signal only νe uncertainty is largest—but these events are rare ๏ Neutral current uncertainty dominates at low energy -2 \Delta \textrm{ln} \mathcal{L} = -2 \textrm{ln} \frac{\mathcal{L}_{marg}(\beta=0)}{\mathcal{L}_{marg}(\beta=1)} Sep 2015 Alfons Weber 4646

Rate & Shape P-values from data Likelihood ratio No evidence for -2Δln(Lp-θ) = -1.16 P-values from data Likelihood ratio No evidence for β = 1 over β = 0 -2Δln(LEREC)= 0.16 Shape term P-value Erec 0.16 Lepton P-θ 0.34 PRELIMINARY Shape term B10 Erec 1.1 Lepton P-θ 0.6 Distribution of test statistic for β = 0 using Lepton P-θ shape information Sep 2015 Alfons Weber

T2K Potential So far, 14% of T2K design POT taken -mode: 6.9 x 1020 POT; -mode: 4.0 x 1020 POT Final sensitivity may be sufficient to find indication for CPV and/or octant (combined with others) Sep 2015 Alfons Weber

Future Systematics Nuclear effects can enhanced “CCQE” cross section and change reconstructed energy CCQE interaction simulated on single nucleon (1p1h) But contribution from correlated pair of nucleons (2p2h) J. Nieves, I. Ruiz Simo, and M. J. Vicente Vacas, PRC 83 045501 (2011) M. Martini, M. Ericson, G. Chanfray, and J. Marteau, PRC 80 065501 (2009) Most oscillation experiments employ the use of a near detector to determine the iniital rate of charged current interactions, and T2K is no different. 280 m away from the neutrino production point, we use a set of offaxis tracking detectors called ND280 which sit within the UA1 magnet. Neutrinos interact on scintillator tracking detector, and then are reconstructed thrioguh surrounding TPCs. The muon momentum is determined from the curvatire in the field. Additional tracks and timing information are used to indicate if there is a pion in the final state, and thus separate the CC numu candidates into CC0pi (which is CCQE enhanced) and CC1pi, and all other CC interactions. 295km later, after oscillation has occurred, we observe numu and nue interactions at Super-Kamiokande, a 50kton water Cherenkov detector. CCQE interactions produce just the lepton above cherenkov threshold, so we select single ring interactions and use the characteristics of the ring to determine if the particle was electron like or muon like and thus a muon neutrino or an electron neutrino. Interactions which aren’t CCQE are vetoed with the presence of one and only one ring, and decay electron tagging, there are also additional selection cuts to remove neutral current interactions which can mimic nue candites. The electron or muon momentum and angle with respect to the beam direction is determined from the ring, and this is used to determine the energy of the neutrino. T2K collab PRL 112, 181801 (2014) Picture by M. Martini Sep 2015 Alfons Weber

Cross Section Systematics ND280 flux Eν - EνQE smearing for Eν=0.8 GeV SK oscillated flux Eν - EνQE smearing for Eν=0.8 GeV What is the effect of this n the oscilation analyssi? Well, if the entire cross section was just increased, clearly that would cancel in the event rate at the near and far detector. The issue is that a shift in the underlying true neutrino energy distribution is not really observable at the near detector. This plot shows the ND flux, with a enu -> ereco smearing. The low side tail of this is due to multinucleon events feeding down. The effect of this feed down at the far detector is different– in th eoscillated flux these events fill in the peak and so affect the measurement of th23. Sep 2015 Alfons Weber

CC 0π Final State Solid: Martini et al Dashed: Nieves et al Martini and Ericson, Phys.Rev. C90 (2014) 2, 025501 measured cross-section as a function of muon momentum for the first muon costheta bin compared to Monte Carlo models. Black dots are results from data with shape (systematics and statistical) uncertainties; gray band is the normalization uncertainty due to flux; red solid line is the prediction from Martini model; red dotted line is the prediction from Nieves model. New measurement muon kinematics for muon, muon + proton, both with no pion in final state from ND280 off-axis beam Sep 2015 Alfons Weber

T2K Cross Sections Cross section measurements Target Reported in Detector  CC inclusive CH PRD 87, 092003 (2013) ND280, Tracker  CCQE Accepted by PRD e CC inclusive PRL 113, 241803 (2014)  NC π0 CH/Water Publication in progress ND280, P0D  NC elastic Water PRD 90, 072012 (2014) SK CH/Fe PRD 90, 052010 (2014) INGRID PRD 91, 112002 (2015)  CC coherent  CCπ+  CC0π What we did on T2K was to test how unsimulated multinucleon interactions could potentially affect the analysis. We generated fake data ynder flux, detector and cross section uncertainties. For each of these fake data sets, we took two cases, with and without a 2p2h multinucleon model present, and fit that for oscillation. We can see the effect of this unsimulated process on our determination of theta23. Theere is an increased uncertainty associated with this process at the level of 3%, and in the case of the larger Martini model, also a shift to the mean indicating a bias in the analysis assuming this model is the correct one. While these uncertainties are within our current systematic uncertainty, they would be considered significant relative to other uncertainties. As we move toward precision measurements, this will become more and more important. Sep 2015 Alfons Weber

Thank You! CERN contributions to T2K UA1/NOMAD magnet And infrastructure Production and testing of micromegas TS/DEM group NA61/SHINE experiment Hadron production CERN-KEK cooperation on superconducting magnets For neutrino beam line Test beams For ND detector components Sep 2015 Alfons Weber

Summary & Conclusion T2K made its first measurement using an anti-neutrino beam Clear measurement of muon anti-neutrino disappearance No clear signal for electron neutrino appearance Statistics very low Anti-neutrino measurements are statistically limited And will be for foreseeable future More anti-neutrino running is planned Much more data expected May have first indication for CP-Violation Sep 2015 Alfons Weber

Back-up