Neutrino Oscillations at the T2K and Hyper-Kamiokande Experiments
Hyper-Kamiokande
Mark Hartz TRIUMF & Kavli IPMU, University of Tokyo CAP-PPD, June 9, 2020 Neutrino Oscillation Experiments
Over hundreds of km baselines, study change of neutrino flavor through oscillations
νe e ν1,ν2,ν3 W+ Mass states propagate with relative phases n p
Produce neutrinos as Interact as weak weak states eigenstates 2 General Purpose Experiments
• Far detectors are general purpose experiments: • Atmospheric neutrinos • Supernova neutrinos • Solar neutrinos • Nucleon decay searches • Dark matter searches Super-Kamiokande (Far Detector) • …. 3 Neutrino Oscillation Parameters
PMNS mixing matrix in standard 3-neutrino mixing framework:
Accessible through neutrino oscillations Majorana phases if neutrinos are s12 = sinθ12, etc. Majorana particles Three mixing angles θ12, θ13, θ23
δ, �21 and �31 may introduce new sources of CP violation
The flavor content of states oscillate as they traverse matter or vacuum:
Dependence on mass squared differences of mass states, distance and energy 4 Neutrino Sources and Measurements
Atmospheric Solar Reactor Accelerator
θ12 = 33.65º±0.80º θ13 = 8.49º±0.14º θ23 = 47.1º±1.6º
δcp = ?? New source of CP violation?
5 Mass Hierarchy (Ordering)
Preferred at ~3σ significance in global fits arXiv:1910.04688
32
21
• Sensitive to mass ordering through oscillation effect when neutrinos propagate through matter • Important interplay with neutrinoless double beta decay experiments
6 Long Baseline Neutrinos at T2K ND280 Near Detector
~500 members, 69 Institutes, 12 countries 7 T2K Overview
1 ) µ ν 2 sin 2θ23 = 1.0 → 0.5 µ sin22θ = 0.1
ν 13 2 -3 2
P( Δm32 = 2.4 × 10 eV 0.10 1 2 3 ) e NH, δCP = 0 IH, δCP = 0 ν NH, δCP = π/2 IH, δCP = π/2 →
µ 0.05 ν P(
0 1 2 3 1 OA 0.0° OA 2.0° OA 2.5° (A.U.) 0.5 µ 295km ν Φ
0 0 1 2 3 Eν (GeV) Off-axis beam
8 Neutrino Oscillations at T2K
1 sin2θ =0.5 0.8 23 2 2 -3 2 • Muon (anti)neutrino survival depends on sin (2θ23) Δm32=2.5×10 eV
Osc. Prob. 0.6 2 2 and Δm 32 sin 2θ13=0.085 0.4
0.2 νµ→νµ, νµ→νµ
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 • Electron (anti)neutrino appearance νµ→νe: NH, δcp = -π/2
νµ→νe: IH, δcp = -π/2 0.1 • sin2(θ23), sin2(2θ13) and Δm232 in leading term Osc. Prob. νµ→νe: NH, δcp = -π/2
νµ→νe: IH, δcp = -π/2 • Sub-leading dependence on δcp 0.05 • CP conservation at δcp=0,π
• Maximal CP violation at δcp=-π/2,π/2 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 1.5 6 2 21 (×10 ) (/cm /50 MeV/1×10 POT) • ν Mode, SK ν , No Osc. Matter effect → dependence on the mass hierarchy 1 µ Mode, SK , No Osc. ν νµ • Normal Hierarchy (NH): enhanced rate for neutrinos,
Neutrino Flux 0.5 decreased for antineutrinos
0 0.5 1 1.5 2 E (GeV) ν 9 Neutrino Detection at Super-K SK MC Electron neutrino appearance signal:
- νe e
W+ Detected electron produces a shower: p n “fuzzy” ring Likelihood-based reconstruction SK MC development led by TRIUMF Muon neutrino survival signal:
- νμ µ W+ Detected muon n p produces a sharp ring
10 Using Near Detector Data
Neutrino-nucleus Interaction Model Neutrino Mode Flux atFlux ND280 Model 12 10 νµ p.o.t)
21 νµ 1011 νe νe 1010 /50MeV/10 2 109
8 Flux (/cm 10
0 1 2 3 4 5 6 7 8 9 10 FGD2 Eν (GeV)
Fit to ND280 data constrains neutrino flux ND280 Data parameters and interaction model parameters
ν-mode 2500 FGD1 FGD2 Data ν CCQE 2000 ν CC 2p-2h ν CC Res 1π 1500 ν CC Coh 1π Events/(100 MeV/c) 1000 ν CC Other ν NC modes
500 ν modes
0 1.2 1.1 1.0 0.9 0.8 TPC1 TPC2 TPC3 Data / Sim. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Reconstructed muon momentum (MeV/c) 11 PRELIMINARY T2K Data Collected
Future analysis: 33% increase in neutrino mode Analysis shown today: 3.1x1021 POT, 50%/50% statistics neutrino/antineutrino Accelerator has achieved 515 kW stable operation Data collected through 2018 in 2019 12 Fitted ND280 Data - Neutrino Mode
Before Fit After Fit
ν-mode ν-mode
2500 Data 2500 Data ν CCQE ν CCQE 2000 CCQE-like ν CC 2p-2h 2000 CCQE-like ν CC 2p-2h ν CC Res 1π ν CC Res 1π 1500 Events 1500 Events ν CC Coh 1π ν CC Coh 1π Events/(100 MeV/c) ν CC Other Events/(100 MeV/c) ν CC Other 1000 1000 ν NC modes ν NC modes
500 ν modes 500 ν modes
0 0 1.2 1.2 1.1 1.1 1.0 1.0 0.9 0.9 0.8 0.8 Data / Sim. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Data / Sim. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Reconstructed muon momentum (MeV/c) Reconstructed muon momentum (MeV/c)
PRELIMINARY PRELIMINARY Model-data agreement significantly improved by fit of model to data
13 What We Observe
Predictions for δ = -π/2 T2K Run 1-9 preliminary cp 24
22 20 18
2 16 sin θ23 = 0.50, 0.45, 0.55 2 -3 2 4 Δm32 = 2.45×10 eV /c 2 -3 2 4 14 Δm31 = -2.43×10 eV /c • Results largely consistent with δcp = -π/2 δ = π CP hypothesis 12 δCP = +π/2 δCP = 0 = - /2 10 δCP π • Observe 15 events in the single decay Data (stat. errors only) Antineutrino mode 1Re candidates 8 electron sample when 7 predicted 30 40 50 60 70 80 90 100 110 Neutrino mode 1Re candidates • Probability of fluctuation this large or larger in any of 5 samples is 7%
14 What We Observe
Predictions for δ = -π/2 T2K Run 1-9 preliminary cp 24 22 Neutrino 20 18
2 16 sin θ23 = 0.50, 0.45, 0.55 2 -3 2 4 Δm32 = 2.45×10 eV /c 2 -3 2 4 14 Δm31 = -2.43×10 eV /c • Results largely consistent with δcp = -π/2 δ = π CP hypothesis 12 δCP = +π/2 δCP = 0 = - /2 10 δCP π • Observe 15 events in the single decay Data (stat. errors only) Antineutrino mode 1Re candidates 8 Antineutrino electron sample when 7 predicted 30 40 50 60 70 80 90 100 110 Neutrino mode 1Re candidates • Probability of fluctuation this large or larger in any of 5 samples is 7%
14
T2K Results δcp Nature 580 (2020) 7803, 339-344 T2K Run 1-9 )
13 0.034 T2K + Reactors θ ( • Include constraint on θ13 from reactor experiments 2 0.032 T2K Only Reactor sin 0.03 0.028 NH: [−3.41, −0.03] 0.026 1σ interval 3σ range for δcp 0.024 IH: [−2.54, −0.32] 0.022 0.02 ) 0.65 23 −3 −2 −1 0 68.27%1 CL2 3 θ
( • CP conserving value δcp = -π is still included in 3σ 2 99.73% CL δCP 0.6 sin interval 0.55 • Both CP conserving values (δcp = 0,-π) are disfavored 0.5 at 2σ 0.45 • Normal mass hierarchy preferred with posterior 0.4 NO−3 −2 −1 0 1 2 3 probability of 0.89 δ IO CP −3 −2 −1 0 1 2 3 15 δCP T2K Results - Atmospheric Parameters
T2K Run 1-9 Preliminary ×10−3 2.8 ) Normal - 68CL
-4 Normal - 90CL c
2 Best fit Inverted - 68CL 2.7 Inverted - 90CL
2.6 (IO) (eV 2 13 2.5 m ∆ 2.4 (NO),
2 32 2.3 m ∆ 2.2 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 2 sin (θ23)
2 • T2K measures sin θ23=0.53+0.03-0.04 • Remains consistent with 45º at the 1 sigma level
16 T2K Results - Atmospheric Parameters
NOvA: Phys. Rev. Lett. 123 (2019) no.15, 151803T2K Run 1-9 Preliminary ×10−3 2.8 ) Normal - 68CL
-4 Normal - 90CL c
2 Best fit Inverted - 68CL 2.7 Inverted - 90CL
2.6 (IO) (eV 2 13 2.5 m ∆ 2.4 (NO),
2 32 2.3 m ∆ 2.2 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 2 sin (θ23)
2 • T2K measures sin θ23=0.53+0.03-0.04 • Remains consistent with 45º at the 1 sigma level
16 Hyper-K In Canada
Hyper-Kamiokande
• Hyper-K Canada group formed in 2018 • Supported by NSERC project grant • Currently 11 faculty from 8 institutes - looking to grow
• Hyper-K is now an IPP project 17 Hyper-K In Canada
Hyper-K Canada Collaboration
Hyper-Kamiokande
• Hyper-K Canada group formed in 2018 • Supported by NSERC project grant • Currently 11 faculty from 8 institutes - looking to grow
• Hyper-K is now an IPP project 17 Hyper-Kamiokande
• Fiducial mass is 8x larger than Super-Kamiokande • Neutrino beam from J-PARC will be 2.5 times more intense (1.3 MW proton beam) • New photon detectors and near detectors • 20x the rate of long baseline neutrinos than the T2K experiment • Not just accelerator-based neutrino experiment: • Proton decay searches • Supernova neutrino detection • Atmospheric neutrino detection Hyper-K approved in January 2020 Construction phase has started! • Solar neutrino detection Start of operation planned in 2027 • Dark matter search Hyper-K Design Report: arXiv:1805.04163 • … 18 Hyper-K Construction Access tunnel construction to start in FY2021 Access tunnel entrance
Construction of the entrance yard is proceeding!
19 Proton Decay Discovery Potential
• Hyper-K excels in the p→e+π0 • Hyper-K is competitive p→νK+ channel, very high efficiency channel, very high efficiency
• Largest fiducial mass • DUNE has potential for better efficiency since kaon is visible 20 Supernova Burst Neutrino
• Inverse beta decay and neutrino-electron scattering channels • Neutrino-electron scattering introduces pointing capability • 54k-90k events for 10 kpc distant supernova • 1.0-1.3 degree accuracy for 10 kpc distant • ~10 neutrino events for supernova in supernova Andromeda 21 CP Violation at Hyper-K
arXiv:1805.04163
2058 events 1906 events
• Recall that T2K and NOvA are observing 10’s of candidate events
• Hyper-K will observe ~2000 electron neutrino and electron antineutrino candidates each
• 3% statistical error on the CP violation measurement is achieved
• Controlling systematic errors is critical: T2K’s current errors are ~6% 22 Oscillation Measurements arXiv:1805.04163 Known Mass Hierarchy
• CP violation discovery for: • With atmospheric neutrino data, achieve • 76% of values at 3σ >4σ rejection of the wrong mass hierarchy
• 57% of values at 5σ 23 Systematic Error Reduction with IWCD
• Intermediate detector for Hyper-K
• Located about 1 km from neutrino source
• 600 ton water Cherenkov detector
• Position can be moved to different off-axis angles
• Loading with Gd to enhance neutron detection
• Using new high resolution multi- PMT modules inspired by Approved Hyper-K project includes IWCD KM3NeT Stage-1 approval at J-PARC as E61 • Project led by Canadian institutes https://j-parc.jp/researcher/Hadron/en/pac_1507/pdf/P61_2015-5.pdf 24 Multi-PMT (mPMT) Photosensor
• 19 3-inch diameter PMTs integrated in module with high voltage and readout electronics • 3-inch PMTs developed with Hamamatsu to achieve 1.7 ns FWHM timing resolution • Improved spatial and timing resolution compared to 20-inch PMTs is necessary for detector of IWCD size • Considered as a photodetector for Hyper-K detector as well 25 Machine Learning
• Improve reconstruction of indiscernible event topologies:
• e/γ separation possible with improved granularity of mPMT in IWCD µ • Multi-ring: directionality of high-energy e γ atmospheric ν, nucleon decay
• Neutron tagging
• First application of ResNet (type of CNN) looks promising
• Investigating several architectures:
• Graph CNN, PointNet, GAN, UNET
26 Photogrammetry Calibration
• Fiducial volume of IWCD must be known with a bias of <1 cm
• The position of all mPMTs and calibration sources must be precisely measured
• Positions can change after water filling, so need in-situ measurement
• Photogrammetry:
• Fixed cameras and remote operated submersible take pictures of the tank interior
• Software able to build an accurate 3-D model of the detector
27 Water Cherenkov Test Experiment
• Aim for unprecedented precision to reconstruct high energy events in a water Cherenkov detector of IWCD size • Need platform to test the hardware and validate the calibration, modeling and reconstruction techniques • Operate detector with 4 m diameter in test beam line with incident particle fluxes of known type and moment
Location: CERN East Area Proposal: CERN-SPSC-2020-005
Planned operation in 2022 28 EMPHATIC Experiment EMPHAT C • Table top hadron production experiment - improve neutrino flux simulation
• Unique application of technologies to hadron production measurements
• Silicon strip tracking layers
• Halbach array permanent magnet
• Aerogel ring imaging Cherenkov detector for PID
• Operating in Fermilab MTEST beam line
• 2018 - Pilot Run
• 2020 - First Physics run with 100 mrad acceptance
• 2021 - Second physics run with 400 mrad acceptance 29 Conclusions
• T2K is providing world leading measurements of neutrino oscillations parameters
• Next phase: move to precisions measurements at Hyper-Kamiokande
• Program includes neutrino oscillations, proton decay, supernova neutrinos, dark matter and more
• Construction has started and planned start of operation in 2027
• Many Canadian efforts for Hyper-K focussed on control of systematic errors for precision measurements
• Still room for new collaborators. Come join us!
30 Thank You
31 Mass Ordering Preference
5000 99.73% Credible Interval 4000 95.45% Credible Interval • One of our analyses uses Markov Chain 90% Credible Interval 68% Credible Interval Monte Carlo to fit oscillation parameters 3000 T2K Run 1-9d preliminary
• Perform Bayesian statistical inference 2000
• Posterior probabilities and credible 1000 intervals 0 Posterior probability density −0.003 −0.002 −0.001 0 0.001 0.002 0.003 • Start with equal prior probability of 2 2 Δ m32 (eV ) normal and inverted hierarchy • Normal hierarchy is preferred with posterior probability of 0.89
32 T2K Systematic Errors
• Uncertainty on the relative rate of electron neutrino and electron antineutrino interactions • This is a purely theoretical estimate, no measurement • Uncertainty on how nuclear effects impact inference of the neutrino energy 33