Past, Present and Future of long baseline neutrino experiments. Apero SPP 16 September 2016 S.Bolognesi (CEA)

Transcription

1 Past, Present and Future of long baseline neutrino experiments Apero SPP 16 September 2016 S.Bolognesi (CEA)

2 Noble Prize 2015 BreakThrough Prize 2015

3 Neutrino oscillations Uαi are expressed in terms of 3 mixing angles (θ13, θ23, θ12) and a phase δcp neutrino oscillation probability also depends on mass differences: m2ij Long baseline neutrino accelerator experiments observe νµ νµ/e: m2 known at ~4%, θ ~ π/4 maximal mixing? Mass ordering unknown (θ13 and θ12, m221 measured with solar and reactor experiments) flavour pattern may indicate the symmetry beyond ν oscillation (door to New Physics!) precise measurement needed to test unitarity of PMNS matrix δ CP phase (unknown) parametrize the difference between ν and ν oscillation involved with matter-antimatter asymmetry in leptogenesis scenarios S.Bolognesi Apero Sept 2016 slide 3

4 NOVA T2K Same technology for near and far detector (14kTon): cells filled of scintillator oil Huge water cherenkov detector (50 kton) with optimal µ/e identification to distinguish νe, νµ Full tracking and particle reconstruction in near detectors (magnetized TPC!): measure precisely neutrino flux before oscillation µ clear ring fuzzy ring S.Bolognesi Apero Sept 2016 slide 4

5 A bit of (recent) history... SuperKamiokande 1996 today! 1998 Discovery of ν oscillation from zenith angle dependence of atmospheric νµ rate ( ) K2K Sudbury Neutrino Observatory (SNO) 1999 today! 2001 Solution of solar puzzle: νe / Σνα ~ 1/3 Need confirmation from accelerator experiment: high purity and tunable neutrino flux MINOS ( MINOS+) ( ) OPERA : 5 νµ ντ events obs. Beyond θ23 and m32: observation of νe apperance T2K ( today) to measure MH, longer baseline: NOVA started last year first results on δcp! S.Bolognesi Apero Sept 2016 slide 5

6 T2K data ν µ ν µ (disappearance) ν µ ν µ (disappearance) Large disappearance signal and clear oscillation shape (beyond counting experiment) Clear signal in antineutrino as well! ν µ ν e (appearance) ν µ ν e (appearance) 7.5 sigma observation of νe appearance Growing statistics of νe appearance: (~20% of final design statistics) S.Bolognesi Apero Sept 2016 slide 6

7 δcp and MH mainly from ν µ ν e / ν µ ν e Expected events as a function of δcp and MH: νe events Normal Hierarchy (NH) Inverted Hierarchy (IH) δ CP= π/2 (maximal CPV) δ CP= -π/2 (maximal CPV) ν events e δ CP= +/-π (CP conserved) δcp= 0 (CP conserved) S.Bolognesi Apero Sept 2016 slide 7

8 δcp and MH mainly from ν µ ν e / ν µ ν e 32 observed νe events Expected events as a function of δcp and MH: νe events Normal Hierarchy (NH) Inverted Hierarchy (IH) 4 observed νe events δ CP= π/2 (maximal CPV) δ CP= -π/2 (maximal CPV) ν events e δ CP= +/-π (CP conserved) δcp= 0 (CP conserved) Results favour maximal CP violation (and slightly favour NH) S.Bolognesi Apero Sept 2016 slide 8

9 First 90% limits on δcp!! Full joint fit of all data (νµ νµ/e and νµ νµ/e) with all proper statistical and systematic uncertainty included and exploiting also shape information : from reactor constraints Not Gaussian behaviour need to through toys to evaluate correct confidence interval Feldman-Cousins confidence interval: δcp = [-3.13, -0.39] NH at 90% CL [-2.09, -0.74] IH (NH slightly favoured) slide 9

10 NOVA δcp NOVA in agreement with T2K: favours maximal CPV and slightly favour NH NOVA has taken 6.05x1020 POT in ν mode (no ν data yet): νµ νe events (larger sensitivity to MH due to longer baseline than T2K) First combination of all data (T2K, NH CP conserved CPV maximal CP conserved Lisi et al. NEUTRINO 2016 CPV maximal CP conservation excluded at 2σ CP conserved NOVA, SK,...) S.Bolognesi Apero Sept 2016 slide 10

11 The other oscillation parameters (θ23, m232 ): mostly from ν µ and ν µ disappearance sin2θ23 enhance/suppress both ν µ and ν µ disappearance m232 regulate the position of the oscillation maximum as a function of the energy T2K data show maximal disappearance prefer maximal mixing: θ 23 = π/4 (sin2θ23=0.5) NOVA data excludes maximal mixing at 2.5σ T2K (NH) NOVA (NH) sin2θ m232 [10-3 ev2] ± 0.12 S.Bolognesi Apero Sept 2016 slide 11

12 Prospects to 2026 sensitivity CPV (True NH) sensitivity MH (True NH) NOVA T2K combination with final dataset (~2021) Request for new run of T2K beyond design statistics (7.8x1021 POT by) 20x1021 POT by 2026 good chances to observe CP violation at > 3σ by 2026 for a sizeable fraction of values large impact of systematics For definitive δ CP measurement need new generation of long baseline experiments: HyperKamiokande, DUNE change from 1% to 3% of neutrino cross-section systematics equivalent to a factor 2 in exposure! S.Bolognesi Apero Sept 2016 slide 12

13 Main systematics Let's take the νe sample at T2K total systematics before the ND constraints 11.9% total systematics after the ND constraints 5.41% specific to SuperKamiokande : 3.46% flux and ν cross-section : 4.17% flux 8.94% (before ND constr.) 3.64% (after ND constr.) Flux simulated (target and beamline with FLUKA/GEANT) and tuned to hadron scattering data in dedicated experiments (NA61) xsec 7.17% (before ND constr.) 5.12% (after ND constr.) dominated by nuclear effects which may give difference between νe/νµ and νe/νµ cross-section Xsec measured with limited precision on free nucleons in old bubble chamber experiments. In modern experiment ν interacts with target detectors of carbon, water or argon large nuclear effects not well known (very peculiar theoretical expertise) S.Bolognesi Apero Sept 2016 slide 13

14 Neutrino-nucleus interaction Crucial role of T2K Near Detector (ND280) : TPC (MicroMegas) developed at CEA for tracking, particle identification and momentum measurement Cross section of main T2K signal: + Charged Current Quasi-Elastic higher order corrections in nuclear target Model delevoped by Martini et al. (SphN) CCQE CCQE + multi-nucleon interactions ν interactions on carbon carbon target ν interactions on water water target S.Bolognesi Apero Sept 2016 slide 14

15 ND280 Upgrade for T2K Phase II T2K-II will require a 2% precision on the expected number of events at SK (5% today) to match the 400 νe appearance events We are currently studying an upgrade of the near detector ND280 comprising 4 additional TPCs and two new active targets (to be installed in 2020) expected efficiency Aim: acceptance over the full polar angle, with better tracking inside the target and lower proton threshold Workshop at CERN November 8-9th (open to all interested people!) S.Bolognesi Apero Sept 2016 slide 15

16 Summary First 90% CL exclusion of CP conservation: hint for maximal ν-ν asymmetry T2K and NOVA: agreement on δ CP while 2.5σ difference for θ 23 measurement Still mostly statistical limited Heavy work ahead: keep collecting data: NOVA, T2K-2 next generation of long baseline experiments (DUNE and HyperKamiokande) need to minimize the systematics for high statistics measurement: - precise measurements of ν-nucleus xsec and better theoretical nuclear modeling - upgrade of the T2K near detector under study S.Bolognesi Apero Sept 2016 slide 16

17 BACKUP slides

18 Non standard scenarios Sterile neutrinos: combination of MINOS, DayaBay and Bugey CPT violation in T2K by comparing disappearance νµ νµ and νµ νµ Limits on non-standard neutrino interactions from MINOS+ important to constrain to avoid degeneracies and biases with future precise δcp measurement! S.Bolognesi Apero Sept 2016 slide 16

19 NOVA T2K comparison: nue appearance 32 events observed (background XX) 4 events observed (background XX)

20 NOVA T2K comparison: ν µ disappearance ν µ ν µ (disappearance) ν µ ν µ (disappearance) ν µ ν µ (disappearance) NOVA ν T2K ν T2K ν Expected w/o oscillations 473 ± ± ± 10 Best fit Observed T2K: agreement between ν and ν data No clear suspect T2K-NOVA difference is maybe just a statistical fluctuation? S.Bolognesi Apero Sept 2016 slide 12

21

22

23 How does it work? SUPERKAMIOKANDE Signal: (anti)ν µ (anti)ν e oscillation µ clear ring fuzzy ring Lepton momentum and angle neutrino energy Select events with no outgoing pions (1 ring) (Quasi-Elastic interactions) νn l-p (outgoing nucleon undetected) Backgrounds: Outer volume with outward facing PMT to veto external background PMT timing to select beam bunches and reconstruct vertex position in fiducial volume ν interactions from beam: intrinsic νe component in the beam pions: π +/- undetected and π0 γγ e-like ring + γ undetected ν oscillations: intrinsic ν component in the beam No magnetic field no charge measurement (ν/ν) R&D: Gd doping to tag neutrons to distinguish: νn l-p from νp-> l+n HYPERKAMIOKANDE: Working to improve PMTs and on Gd doping. Electronics and calibration system very similar to SuperK

24 From SuperK to HyperK Total volume 50 kton 990 kton Fiducial volume 22.5 kton 560 kton Tanks 1 cylindrical 41.4m (h) x 39.3m (d) inner detector 2 egg-shape tanks 48m (w) x 50m (h) x 250m (l) Photocoverage 40% 20% Sensor efficiency (Collection x Quantum eff.) 18% (22x80%) 29% (30x95%) PMTs outer detector Tanks and PMT design under discussion: minimize risk due to pressure on PMTs (avoid cascade implosion as in SK 2001 incident) minimize cost (volume vs #PMTs) need PMT R&D (next slide)

25 R&D on PMTs Optimization should include pressure resistance possible to put protective cover need precise control of glass quality Response to single photoelectron: time resolution charge resolution Integrated system of inner and outer PMTs under study (solve problems of pressure and in-water electronics) 3' PMTs for inner detector large PMT for outer detector veto

26 Gadolinium doping νp l+n n get captured in Gd with emission of few γ ~8MeV for beam neutrino physics: ν vs ν separation, but also useful to enhance sensitivity to SuperNova ν and proton decay R&D studies (eg, WATCHMAN) as reactor monitoring EGADS: 200 ton scale model of SuperK fully operative in Kamioka mine 1y time Gd concentration Neutron capture time tested with Am/Be source: data-mc perfect agreement All the trick is about keeping water pure and transparent without loosing Gd (dedicated filtration system) SuperKamiokande will run with loaded Gd in next years! S.Bolognesi (CEA,Saclay) 8 IFD Torino December 2015

27 Liquid Argon technology Ionizing particle in LAr 2 measurements: charge from ionization tracking and calorimetry scintillation light trigger and t 0 DUNE: staged approach with 4 modules of ~10kTon fiducial mass each (drift time third coordinate for non-beam events) 4 x (60m x 12m x 12m) µ track momentum from range (or from multiple scattering if not contained) PID from de/dx Very good electron/γ ID and π0 reconstruction Calorimetric energy from total collected charge (+ light) (ICARUS)

28 Result of years of R&D S.Bolognesi (CEA,Saclay) 10 IFD Torino December 2015

29 Single-phase VS Double-phase Double Phase charge readout Very long charge drift path diffusion and attenuation anode wires cathode LIQUID SiPM ionizing particle e- γ high signal/noise LEM thanks to avalanche multiplication in gas anode Single Phase charge readout limited to short drift distances: 4 drift regions of 3.6m each γ GAS LIQUID secondary scintillation eionizing particle scintillation primary scintillation γ transparent cathod PMTs anode cathode anode cathode anode drift direction for single phase 3. 6 anode drift direction for double phase x 4 m anode wires m cathode IFD Torino December 2015

30 Charge signal We = 23.6 ev mip produces ~ 100k e- per cm 60k e- after recombination dirft velocity ~mm/µs ( total drift time ~10 ms) Very long drift path diffusion and attachment DUNE double phase single phase diffusion ~few mm with kv/cm ( pitch readout few mm) O2 pollution captures ionization electrons charge attenuation new pump double phase single phase ( impurity ~20 ppt O2 needed) ICARUS 2% of DUNE 1 module mass S.Bolognesi (CEA,Saclay) 12 DUNE

31 Charge readout plane (CRP) Single Phase no gain 3 views uniform CRP design Double Phase stable gain of 20 on 10x10cm LEM 2 views (x,y) of equal quality to scale up: CRP segmented in 50x50cm modules

32 Many other challenges scintillation light: single phase: first test of wavelenght shifting bars to SiPM integrated with a TPC double phase: standard PMTs (with coating), high voltage on large surfaces: cathode-anode V ~few hundreds V (double phase) ~180 V (single phase) large number of channels electronics in gas accessible only in double phase design calibration and uniformity (eg: flattening of cathode and of charge readout plane, E field between different modules of charge readout...) software for automatic reconstruction huge amount of info (efficient zero suppression) ICARUS: LAr TPC as calorimeter fully omogeneus with very low threshold very good resolution and detailed tracking inside shower potential to improve shower models! S.Bolognesi (CEA,Saclay) 14 Low energy electrons: σ(e)/e = 11%/ E(MeV)+2% Electromagnetic showers: σ(e)/e = 3%/ E(GeV) Hadron shower (pure LAr): σ(e)/e 30%/ E(GeV) IFD Torino December 2015

33 Water Cherenkov vs Liquid Argon Hyperkamiokande much more sensitive to CP violation while DUNE much more sensitive to Mass Herarchy (see backup). But sensitivities depend on assumed beam power, detector mass and on baseline. Comparison of technologies: LIQUID ARGON WATER CHERENKOV well known and solid technology very large mass (~MTon) info only about particles above Cherenkov threshold model dependent assumptions to reconstruct Eν size limited by drift length (~40KTon) full reconstruction of tracks and showers down to very low threshold, very good particle ID precise Eν shape accessible and needed for good sensitivity no need of precise Eν shape: mainly a counting experiment S.Bolognesi (CEA,Saclay) successfull R&D first very large scale realization need to reach very good control on detector calibration/uniformity and on neutrino interaction modelling 15 IFD Torino December 2015

34 Sensitivities Assuming 1MW beam HK 3 years (1MTon): CPV measured at 3s(5s) for 75% (60%) of dcp values DUNE 10 years (40 kton): CPV measured at 3s (5s) for >50% (~25%) of dcp values DUNE 10 years: definitive determination of MH HK 10 years: wrong MH excluded at 3s

35 Future experiments: ν e and ν xsec T2K: T2K uncertainties HK In future (HK, DUNE) large samples of 4 ν species the uncorrelated uncertainties are relevant HK needed uncertainty to have negligible impact on dcp: ν e-ν e uncorrelated 1-2% 5% ± 1% 5% ± 2% For DUNE assumed: uncorrelated ν µ - ν µ 5% and ν e - ν e 2% (shape of νµ itself may be more important for DUNE: shape analysis and spanning over different xsec) 5% ± 3% DUNE equivalent to factor 2 in exposure! DUNE and HK NuFact 2015 We are interested to ν e appeareance and δ CP from ν ν comparison but in ND we mostly measure νµ cross-sections. 18/21

36 Moving to larger energies... T2K flux 13/21 DUNE

37 Moving to larger energies... T2K flux 14/21 DUNE

38 Moving to larger energies... 15/21 Need to control well all different xsec, each process has very different detector acceptance T2K flux DUNE