1. New results from T2K A.Minamino (Kyoto Univ.) on behalf of the T2K collaboration KEK seminar Feb. 18, 2014 1

2. Neutrino oscillation 2

3. Neutrino Mixing θ 13 is now precisely known, and relatively large Long-baseline experiments (T2K & NO A) can constrain δ CP However, the large uncertainty on θ 23 is limiting the information that can be extracted from ν e appearance measurements Precise measurements of all the mixing angles will be needed to maximize sensitivity to CP violation Note: c ij = cos(θ ij ), s ij = sin(θ ij ) “ Atmospheric ν ” sin 2 2θ 23 > 0.95 (90% C.L.) “ Solar/reactor ν ” sin 2 2θ 12 = 0.857±0.024 “ Reactor/Acc. ν ” sin 2 2θ 13 = 0.098±0.013 Majorana phases; Not yet observed Flavor States Mass States 3

4. Unknown sign of Δm 2 32 4 Can be determined from matter effects, as is our knowledge that Δm 2 21 >0 from solar neutrinos [NH] Normal hierarchy 3 1 2 1 2 3  m atm 2  m sol 2 [IH] Inverted hierarchy e  

5.  disappearance 5  disappearance probability T2K: L = 295 km, E peaks at ~ 0.6 GeV -> sin 2  solar ~ 0, sin2  atm ~ 0 Leading-term Next-to-leading Probability depends on sin 2 2  13 sin 2  23 to second order -> Therefore now fit for sin 2  23 -> Can be used to resolve the θ 23 octant with known sin 2 2θ 13

6. e appearance 6 ±25% anti- Comparison of /anti-  run enhances sensitivity to  (T2K will have anti- test run in early 2014.) e appearance probability  CP can be determined only from only run if  13 is known from the reactor experiments. e disappearance probability (Reactor exp.: Baseline ~1km, E ~ 3MeV) Simple 2 flavor oscillation formula is valid at L~1km w/ no matter effect sin 2 2  13 =0.1  /2  /2   /2  /2 

7. The T2K experiment 7

8. The T2K Collaboration 8 Canada TRIUMF U. Alberta U. B. Columbia U. Regina U. Toronto U. Victoria U. Winnipeg York U. France CEA Saclay IPN Lyon LLR E. Poly. LPNHE Paris Germany Aachen U. Poland IFJ PAN, Cracow NCBJ, Warsaw U. Silesia, Katowice U. Warsaw Warsaw U. T. Wroklaw U. Russia INR U. Sheffield U. Warwick USA Boston U. Colorado S. U. Duke U. Louisiana S. U. Stony Brook U. U. C. Irvine U. Colorado U. Pittsburgh U. Rochester U. Washington Spain IFAE, Barcelona IFIC, Valencia Switzerland ETH Zurich U. Bern U. Geneva United Kingdom Imperial C. London Lancaster U. Oxford U. Queen Mary U. L. STFC/Daresbury STFC/RAL U. Liverpool ~500 members, 59 Institutes, 11 countries Italy INFN, U. Bari INFN, U. Napoli INFN, U. Padova INFN, U. Roma Japan ICRR Kamioka ICRR RCCN Kavli IPMU KEK Kobe U. Kyoto U. Miyagi U. Edu. Osaka City U. Okayama U. Tokyo Metropolitan U. U. Tokyo

9. The T2K Experiment Searches for neutrino oscillations in a high purity ν μ beam – Intrinsic beam e from , K decays ~1% The neutrinos travel 295 km to the Super-K detector – ν e appearance (sensitive to θ 13 & δ CP ) – ν μ disappearance (sensitive to θ 23 & Δm 2 32 ) 295 km Super-K Detector J-PARC Accelerator ν 30 9 30 GeV Synchrotron

10. 10 30 GeV protons hit 90 cm graphite target – Profile/Intensity from SSEMs *1 +OTR *2 /CTs *3 Three magnetic horns focus positive hadrons –  from  + decay – (small) e contamination from  and K decay 2.5 degree off-axis beam – Intense, low energy narrow-band beam – Peak E tuned for oscillation max. (~0.6 GeV) – Reduce BG from high energy tail – First application to LBL experiment T2K Neutrino Beam (96 m) *1 SSEM: Segmented Secondary Emission Profile Monitor, *2 OTR: Optical Transition Radiation monitor *3 CT: Current Transformer

11. Near Detectors Located 280 m downstream of the target Measure unoscillated neutrinos 11 1.5m ~10m Beam center INGRID @ on-axis (0°)ND280 @ 2.5° off-axis 16 identical modules (14 in cross) Iron/scintillator layers Monitor beam profile/rate Tracker (FGDs *1 + TPCs *2 ) in a 0.2 T magnet Principal target is plastic scinti. in FGDs Measures flux/spectrum *1 FGD: Fine-Grained Detector, *2 TPC: Time Projection Chamber

12. Super-K (Far) Detector 50 kton Water Cherenkov detector – 22.5 kton Fiducial volume Good performance for sub-GeV neutrinos – Good e/  separation from ring shape topology T2K recorded events – All interactions in ±500  sec around arrival time 12 ID ODOD ● Data − MC w/ oscillation e-like  -like CCQE muon CCQE electron Atmospheric ν

13. T2K measurement 13

14. T2K Data Set (until May 8, 2013) Total delivered beam: 6.57x10 20 Proton on Target (POT) – ~8% of T2K goal Dead time fraction in T2K Run-4 due to T2K beam line failure ~ 3% Super-K running efficiency > 99% over run period 14 Great Earthquake (Mar. 11, 2011) Recovering facility

15. Beam stability Neutrino rate per POT is stable to 0.7% over run period Neutrino beam direction is stable < 1mrad over run period 15 Stability of ν interaction rate normalized by # of protons (INGRID) Stability of ν beam direction (INGRID) Note: Dataset includes 0.21x10 20 POT with 250 -> 205kA horn operation (13% flux reduction at peak)

16. What’s new?  disappearance analysis – Results using a data set of 6.57×10 20 POT e appearance analysis – Constraint on the CP violating phase  CP by combining our e appearance results with  13 measurements by reactor experiments T2K future sensitivity study 16

17. Neutrino oscillation analysis 17

18. Neutrino oscillation analysis principle 18 Super-K prediction with systematics flux prediction Hadron production (NA61@CERN,…) Systematics Hadron production Proton/ beam monitoring cross section Generator: NEUT Systematics External data (MiniBooNE,  scattering exp., …) Super-K performance Systematics Atmospheric Cosmic ray  ND280 measurement Constrain strongly-correlated systematics between ND280/SK (Reduce abs. “flux × XSEC” error) Super-K measurement Compare

19. flux prediction 19

20. Flux prediction External hadron production measurement – CERN-NA61: Same proton beam energy/target material as T2K T2K – Proton beam monitoring Profile on target from SSEMs *1, OTR *2 Intensity from CTs *3 – Alignment of and current in horns – The neutrino beam direction 1 mrad direction shift -> ~3% energy shift at peak 20 *1 SSEM: Segmented Secondary Emission Profile Monitor, *2 OTR: Optical Transition Radiation monitor *3 CT: Current Transformer, *4 FLUKA: Hadron production simulator NA61/FLUKA *4  + production P  (GeV/c)

21. Flux and Uncertainties 21 Super-K flux ND280 flux overlaid (not stacked) Super-K  uncertainty Super-K flux has 10-15% uncertainties from 0.1 to 5 GeV Near (ND280) and Far (Super-K) fluxes are not identical, but highly correlated overlaid (not stacked)

22. cross section and its uncertainty 22

23. Neutrino Interactions in T2K CC(Charged-Current) quasi elastic (CCQE) – + n ->  - + p CC (resonance) single  (CC-1  ) – + n(p) ->  - +  + + n(p) DIS(Deep Inelastic Scattering) – + N ->  - + m  +/-/0 + N’ CC coherent  – + A ->  - +  + + A NC (Neutral-Current) – copious process (NC-1  0, …) + Nuclear Effects 23 σ/E(10 -38 cm 2 /GeV) E ν (GeV) NEUT model a main BG for e appearance analysis T2K: E peak at ~ 0.6 GeV

24. Cross-section Model: CCQE Signal reaction for T2K energies – Elastic kinematics allow us to measure neutrino energy from e/  T2K is currently using a very simple model – Nucleon form factors from e - scattering and νD scattering – Model of nucleus is Fermi gas ~20% diff. between Data/Model – Approach: add effective parameters (M A, normalization) with uncertainties based on external data sets (MiniBooNE,…) 24 (Phys. Rev. D81 092005, 2010) Data/Model(M A =1.03) (MiniBooNE  CCQE sample)

25. ND280 measurement/fit 25

26. ND280 Event Categories 26 Exclusive samples based on # of final state charged  s 26 Charged current (CC) 0  (CCQE 64%) CCQECCRESCCDIS FGD TPC μ-μ- μ-μ- π+π+ μ-μ- CC 1   (CCRES 40%) CC Other (CCDIS 68%)

27. ND280 measurement/fit 27 fit CC 1   CC 0  CC Other Fit p  -cos   distributions Constrain strongly-correlated syst. between ND280/SK MC nominal MC fit Data MC nominal MC fit Data MC nominal MC fit Data

28. Flux/Cross-Section uncertainties after ND280 constraint ND280 constraint reduces both flux and cross- section model uncertainties individually 28 Cross-section parameters Super-K  flux

29. Uncertainty of # of SK e events before/after ND280 constraint Flux and cross-section parameters are anti-correlated after ND280 constraint becauseND280 constrains only the observable event rate 29 sin 2 2  13 = 0.1 w/o ND280 constraint w/ ND280 constraint Flux/XSEC (ND280 constraint) 25.9%2.9% Other XSEC7.5% Super-K +FSI3.5% Total27.2%8.8% Systematic uncertainties on # of e candidate events SK flux ( , , e, e ) XSEC Correlation matrix of systematics

30. CC0π sample ND280 ν e Measurement Interactions in FGD and particle ID in TPC Major background: photons from π 0 decays Fit CC0π, CC1π+CCother and γ sideband 30 γ sample fit prefers scale factor of 0.77±0.02(stat) CC1π + + CCother sample Intrinsic beam e background prediction is validated!

31.  disappearance analysis 31

32. Particle Identification at SK  scattering is minimal Rings with sharp edges  0 γ from π 0 decays shower and look like electrons Multiple fuzzy rings 32 MC e Electromagnetic shower Rings are “fuzzy”

33.  event selection Fully-contained fiducial volume (FCFV) event Single-ring  -like event Reconstructed momentum > 200 MeV/c # of decay electron ≤ 1 33  -like e-like 120 events in 6.57x10 20 POT

34. SystematicsUncertainties Flux/XSEC (ND280 constraint)2.7% Other XSEC4.9% Super-K +FSI5.6% Total8.1% Event category# of events  CCQE 77.93  CCnonQE 40.78 e CC 0.35 NC All6.78 Total125.85 Predicted # of events & syst. errors 34 Systematic uncertainties of # of events * (sin 2  23,  m 2 32 )=(0.5, 2.4×10 -3 eV 2 ) Predicted # of events (sin 2  23,  m 2 32 )=(0.5, 2.4×10 -3 eV 2 ) N SK per bin w/ error Reconst. E (GeV) w/o ND280 constraint w/ ND280 constraint (sin 2  23,  m 2 32 )=(0.5, 2.4×10 -3 eV 2 ) * Binding energy/SK energy scale are some of the dominant uncertainties affecting T2K  m 2 32 precision, but they don’t appear in the left table of # of events since they don’t affect overall normalization.

35. Multi-Nucleon Systematic Uncertainty Lively discussion motivated by CCQE cross section inconsistency between MiniBooNE/other experiment Not incorporated directly into analysis – But we have a large systematic uncertainty (100%) on decays of  resonances w/ prompt  absorption (“  -less  -decay”). It has similar impact on neutrino energy reconstruction as a 100% uncertainty in the multi-nucleon interaction model (Nieves model) – Dedicated MC study shows the impact on oscillation analysis is small relative to our current statistical error. 35

36. Oscillation Likelihood Fits Search for Oscillation parameters which maximize L – Observables: # of events, E rec distribution 36 # of events E rec dist.Syst. Osci. param. Neutrino energy from elastic kinematics * E b is mean binding energy. L = L norm × L shape × L syst × L osci  (P ,   )    Note: sin 2  13, sin 2  12,  m 2 21 are constrained by PDG2012.  CP is unconstrained.

37. sin 2  23 [NH] ([IH]) 0.514 (0.511)  m 2 32 [NH] (  m 2 13 [IH]) 2.51 (2.48) N exp [NH] ([IH]) 121.41 (121.39) 37 Results of  disappearance analysis [NH] Normal hierarchy, [IH] Inverted hierarchy Best fit ×10 -3 Note: N obs = 120 events Data Best fit No oscillation Data Best fit Depth: sin 2 2  23 Position:  m 2 32 2 flavor approximation

38. 38 Previous T2K result PRL 111, 211803 (2013) Feldman-Cousins 2D confidence regions T2K new Results of  disappearance analysis [NH] Great improvement from the previous T2K result! T2K favors maximal mixing [NH] Normal hierarchy, [IH] Inverted hierarchy F&C 1D intervals  23 [NH] [42.6°, 48.9°][40.9°, 50.7°]  23 [IH] [42.5°, 48.8°][40.8°, 50.5°]

39. Results of  disappearance analysis 39 T2K measures  23 with the world-leading precision! Comparison w/ other experiments [NH] Normal hierarchy, [IH] Inverted hierarchy

40. e appearance analysis T2K collaboration, PRL 112, 061802 (2014) 40

41. e event selection Fully-contained fiducial volume (FCFV) event Single-ring e-like event E visible > 100 MeV # of decay electron = 0 0 < E rec < 1250 MeV  0 cut 28 events in 6.57x10 20 POT 41  0 cut select  0 mass (MeV/c 2 )  -like e-like

42. Oscillation Likelihood Fits Search for Oscillation parameters which maximize L – Observables: # of events, (p e,  e ) distribution 42 L = L norm × L shape × L syst × L osci # of events (p e,  e ) dist. Syst. Osci. param. (p e,  e ) distribution sin 2 2  13 =0.0 (BG only) sin 2 2  13 =0.1 ( e CCQE dominant)

43. 43 Results of e appearance analysis [NH] Normal hierarchy, [IH] Inverted hierarchy Note: sin 2  23,  m 2 32 are constrained by previous T2KRun 1-3  result (PRL 111, 211803 (2013)) Best fit for [NH] ([IH]) (Assuming  CP =0) sin 2 2  13 = 0.136 (0.166 ) +0.044+0.044 -0.033-0.033 +0.051+0.051 -0.042-0.042 Significance of non-zero  13 yields 7.3  Discovery of e appearance! Electron p-  distribution

44. 44 Results of e appearance analysis [NH] Reactor 1σ range (PDG2012) [IH] [NH] Normal hierarchy, [IH] Inverted hierarchy Reactor 1σ range (PDG2012) Allowed region of sin 2 2  13 for each value of  CP Note These are 1D contours for values of  CP, not 2D contours in  CP -  13 space

45. 45 Results of e appearance analysis Combination of T2K + Reactor ( sin 2 2  13 =0.098±0.013 from PDG2012 ) 90% C.L. limits by Feldman-Cousins 90% C.L. excluded region [NH]: 0.19  ~ 0.80  [IH]: -0.04  ~ 1.03  [NH] Normal hierarchy, [IH] Inverted hierarchy Best fit is found at very interesting point,  CP ~ -  /2. If it is true, severe competition w/ NO A. Very important to increase statistics ASAP. [NH] [IH]

46. 46 Our paper on the e appearance analyses (PRL 112, 061802 (2014)) was selected for a “Viewpoint in Physics” of APS.

47. Future sensitivity study 47 e appearance and  disappearance combined fit Realistic (shape-dependent) systematic errors Errors are assumed to be fully correlated between /anti-

48. 48 Koseki-san’s slides @ T2K collaboration meeting, Sep, 2013 Expected POT is estimated based on the information.

49. sin 2  23 /  m 2 32 1  Precision vs. POT 49 50% POT + 50% POT anti- Solid Lines: no sys. err. Red Dashed: with conservative projected sys. err. (~7%, ~14% anti- ) Statistical limit of 1  precision at full POT sin 2  23 (  23 ): ~0.045 (~2.6°)  m 2 32 : ~4×10 -5 eV 2 Assuming true: sin 2 2  13 =0.1,  CP =0°, sin 2  23 =0.5,  m 2 32 =2.4×10 -3 eV 2, [NH]  13 constrained by  (sin 2 2  13 ) = 0.005 [NH] Normal hierarchy, [IH] Inverted hierarchy now ~2016 now ~2016 Precisions are increased drastically over the next few years.

50. Appearance 90% C.L. Sensitivity 50 [NH] Normal hierarchy, [IH] Inverted hierarchy 7.8×10 21 POT (50% POT + 50% POT anti- ) Solid Lines: no sys. err., Dashed: with 2012 sys. err. (~10% e, ~13%  ) Case study (1): True  CP = 0° Case study (2): True  CP = -  Assuming true: sin 2 2  13 =0.1, sin 2  23 =0.5,  m 2 32 =2.4×10 -3 eV 2, [NH] T2K w/ Reactor  (sin 2 2  13 ) = 0.005 T2K only

51. Sensitivity for Resolving sin  CP ≠0 51 7.8×10 21 POT (50% POT + 50% POT anti- ) True [NH] True [NH] True [IH] True [IH] No sys. err.w/ 2012 sys. err. (~10% e, ~13%  ) Assuming true: sin 2 2  13 =0.1,  m 2 32 =2.4×10 -3 eV 2  13 constrained by  (sin 2 2  13 ) = 0.005 [NH] Normal hierarchy, [IH] Inverted hierarchy

52. 52 T2K + NO A Sensitivity for Resolving sin  CP ≠0 Assuming 5% (10%) normalization uncertainty on signal (background) Assuming true: sin 2 2  13 =0.1,  m 2 32 =2.4×10 -3 eV 2,  13 constrained by  (sin 2 2  13 ) = 0.005 [NH] Normal hierarchy, [IH] Inverted hierarchy Region where sin  =0 can be excluded by 90% C.L. solid(dash): w/o (w/) systematics NO A T2K Both T2K/NO A -> full POT (50% POT + 50% POT anti- ) Shown in [NH] case. Sensitivity to resolve sin  =0

53. 53 T2K + NO A Sensitivity to Mass Hierarchy Both T2K/NO A -> full POT (50% POT + 50% POT anti- ) Shown in [NH] case. Assuming true: sin 2 2  13 =0.1,  m 2 32 =2.4×10 -3 eV 2,  13 constrained by  (sin 2 2  13 ) = 0.005 Red: T2K alone, Blue: NO A alone, Black: T2K + NO A [NH] Normal hierarchy, [IH] Inverted hierarchy Region where MH can be distinguished by 90% C.L. Sensitivity to resolve MH solid(dash): w/o (w/) syst. NO A

54. Summary  disappearance results – We have measured  23 with the world-leading precision – New result favors maximal mixing e appearance results – We have constrain the CP violating phase  CP by combining our e appearance results with the reactor measurements – Best fit is found at very interesting point,  CP ~ -  /2. If it is true, severe competition w/ NO A. Very important to increase statistics ASAP. Future sensitivity study – T2K can constrain  CP – Combined analysis w/ NO A enhances the sensitivities to  CP and the mass hierarchy – Achievement of 750 kW beam operation is essential 54

55. Backup 55

56. A case study on expected POT projection 56 Linac updade 400 MeV High rep. rate (1.28 sec cycle)

57. Reconstructed μ “Vertex” Accuracy μ Momentum/Range (~1 m < Range < 5 m) ~0.7 GeV/c Decay-e Momentum μ vertex pμpμ Decay-e Vertex ΔθΔθ Δ θ (°) Minimum angular binning ΔMean < 2.4% ΔMean < 5 cm Validation of new algorithm with Stopping μ in Super-Kamiokande Data/MC agreement within systematic uncertainties 57

58. Future sensitivity study e appearance and  disappearance combined fit Realistic (shape-dependent) systematic errors – Errors are assumed to be fully correlated between /anti- 58 *2012 errors were used for this study ** anti- systematic errors are treated as equivalent to errors with an additional 10% normalization uncertainty

59. 59 Sensitivity for Resolving  23 octant 7.8×10 21 POT (50% POT + 50% POT anti- ) Case study: True sin 2  23 = 0.4 Solid Lines: no sys. err., Dashed: with 2012 sys. err. (~10% e, ~13%  ) w/ Reactor  (sin 2 2  13 ) = 0.005 Assuming true: sin 2 2  13 =0.1,  CP =0°,  m 2 32 =2.4×10 -3 eV 2, [NH] 90% C.L. Sensitivity Assuming true: sin 2 2  13 =0.1,  m 2 32 =2.4×10 -3 eV 2, [NH]  13 constrained by  (sin 2 2  13 ) = 0.005 [NH] [IH] [NH] Normal hierarchy, [IH] Inverted hierarchy Resolved nearly resolved!

60. Improved Super-K Reconstruction Algorithm Each hit PMT gives charge and time information For a given event topology hypothesis, it is possible to produce a charge and time PDF for each PMT – Based on MiniBooNE likelihood model (NIM A608, 206 (2009)) Event hypotheses are distinguished by best-fit likelihoods, e.g., electron vs muon or π 0 60 Light Yield Integral over track length PMT solid angle Water attenuation PMT angular response electron s (cm) cos θ Cherenkov light emission profile muon

61. Enhanced π 0 Rejection New reconstruction algorithm can use mass of the π 0 hypothesis and best-fit likelihood ratio of e - and π 0 Cut removes 70% more π 0 background than previous method for a 2% added loss of signal efficiency select 61 Likelihood ratio vs  0 Mass BG   0 +X Signal e CCQE

62. e event selection (1) Fully-contained fiducial volume (FCFV) event Single-ring e-like event E visible > 100 MeV # of decay electron = 0 0 < E rec < 1250 MeV  0 cut 28 events in 6.57x10 20 POT 62  -like e-like

63. e event selection (2) Fully-contained fiducial volume (FCFV) event Single-ring e-like event E visible > 100 MeV # of decay electron = 0 0 < E rec < 1250 MeV  0 cut 28 events in 6.57x10 20 POT 63  0 cut select  0 mass (MeV/c 2 )

64. e appearance analysis Predicted # of events & syst. errors 64 Systematics sin 2 2  13 =0sin 2 2  13 =0.1 Flux/XSEC (ND280 constraint)4.82.9 Other XSEC6.87.5 Super-K +FSI7.33.5 Total11.18.8 Systematic uncertainties of # of events (%) Predicted # of events Distribution of Predicted # of events sin 2 2  13 =0.0 sin 2 2  13 =0.1 Syst. fluctuation Stat. fluctuation Syst. fluctuation Stat. fluctuation

65.  disappearance 65  disappearance probability in vacuum T2K: L = 295 km, E peaks at ~ 0.6 GeV -> sin 2  solar ~ 0, sin2  atm ~ 0 Leading-term Next-to-leading  disapp. probability depends on sin 2 2  13 sin 2  23 to second order -> Can be used in combination with known sin 2 2θ 13 to resolve the θ 23 octant

66. e appearance 66 25% anti- for P(  -> e )  -> -  a -> -a  13 (Leading term) Matter effect CPV Solar CPC Comparison of /anti-  run enhances sensitivity to  or  can be determined if  13 is known from other measure.

67. 67 Results of  disappearance analyses Feldman-Cousins 1D confidence intervals [NH] Normal hierarchy, [IH] Inverted hierarchy 68% CL90% CL  23 [NH] [27.3°, 34.6°][25.3°, 36.7°]  23 [IH] [27.1°, 34.5°][25.3°, 36.6°]

68. e disappearance (Reactor) Reactor experiments (Double Chooz, Daya Bay, RENO) – Baseline ~1km, E ~ 3MeV 68 Simple 2 flavor oscillation formula is valid at L~1km with no matter effect