The ultimate measurement?

Transcription

1 Reactor neutrinos θ13 The ultimate measurement? Pau Novella CNRS/APC 1

2 Overview Neutrino oscillations and the last mixing angle Reactor neutrinos as a probe to θ13 Unrevealing θ13 with reactor neutrino data Experimental results: critical view Towards the ultimate value of θ13 2

3 Neutrino Oscillations and the last mixing angle 3

4 In the beginning Pauli postulates 1956: reactor neutrinos detected 1990's: neutrino oscillations... Physics Beyond the Standard Model... Today Reactors play a major role again! 4

5 Neutrino mixing νe m1 νµ m2 ντ m3 Oscillation physics Atmospheric sector θ23 interference sector θ13, δ Solar sector θ12 5

6 Neutrino Oscillations If neutrinos are massive and have different masses... Oscillation parameters: ( 12, 13, 3), ( m221, m231), If θ13 small and m221 << m232 : amplitude P αβ =sin 2 2θ sin 2 frequency ( Δm2 L 4 E ν Posc ) 6

7 Neutrino Sources matm, atm msol, sol Δmij2 mi2-mj2 7

8 First Generation Of Experiments 8

9 Exploring the sectors... m2~2.4x10 3 ev2 ~ 45º m2atm atm) K2K, MINOS, Super K Solar oscillation: Phys. Rev. Lett. 100:221803, 2008 m2sol sol) KamLAND and solar data Interference sector: m2~7.6 x10 5 ev2 ~ 33º 13 Phys. Rev. Lett. 101:131802,2008 Atmospheric oscillation: Unrevealed until 2011!!! Measurement of cp Mass hierarchy, θ23 octant Design of next experiments 9

10 Interference sector as of 2010: SBL Reactor experiments Past reactor experiments... Phys. Rev. Lett. 90, (2003) Gratta et Al. Rev. Mod. Phys, 74, 2002 solar Measurement of reactor flux From CHOOZ: sin2(2 13) < 0.15, δ? 10

11 3 Global Analysis in 2010 Gonzalez Garcia et Al., JHEP 1004 (2010) 056 Global fit for 3 flavour scenario Preference for 13 0 First hint of 13 : sin2( 13) ~ G.L. Fogli et Al, hep/ph v2 11

12 First direct indications: MINOS Phys.Rev.Lett. 107 (2011) Appareance analysis: 13 from e 12

13 First direct indications: T2K First results on disappearance with 1.4x1020 pot e appearance: 6 events over background of 1.5 (2.5 ) NH (δ=0) sin2(2θ13)=0.11 and 90% C.L. IH (δ=0) sin2(2θ13)=0.14 and 90% C.L. The 5σ appearance result is expected by June 2013! Published in Phys. Rev. Lett. 107, (2011) 13

14 Reactor Neutrinos as a probe to θ13 14

15 Nuclear Reactors as a ν source e Neutrino flux: sum of all fission products from 235U, 238U, 239Pu and 241Pu Flux depends on fuel composition ( f(t) ): 1GWth -> 2 x 1020 /s ~200 MeV/fission ~6 /fission ILL spectra (reference last 25 years) 15

16 Reactor neutrino oscillation 16

17 13: Why reactor neutrinos? L ~ 1 km P( e x) In contrast to accelerator experiments... No parameter correlations Pure e beam Low energy No matter effects Cheap, as source exists High flux and large xsection 17

18 Detecting reactor neutrinos IBD: e + p e+ + n Th: 1.8 MeV. Disappearance! Target: scintillator + n catcher (Gd) Detector: PMTs Prompt signal (1 8 MeV) e- e+ e E spectrum p n Gd Delayed signal (30 s, 8 MeV) 18

19 Expected oscillation signal Deficit in the number of neutrinos Characteristic L/E pattern: θ13 1 km Toy MC Deficit! Energy dependent! 19

20 Setting up the experiment Reactor neutrinos: <E > ~ 4 MeV Solar sector Systematics! ~ 100 m Oscillation! ~ 1 km 20

21 Unrevealing θ13 with reactor neutrinos 21

22 Reactor neutrino experiments IBD detection in Gd doped scint. Multi detector setups 22

23 The Double Chooz Experiment Near Detector L = 400m 10m3 target 120 m.w.e Far Detector Chooz Reactors 4.27GWth x 2 cores Pioneered reactor experiments after CHOOZ: Experimental concept of using two detectors New detector structure: 4 layers detector Low background (S/N ~ 20, proven by reactor OFF) Stable Gd loaded LS developed L = 1050m 10m3 target 300m.w.e. April 2011 ~

24 Daya Bay and RENO Power Power Target (x2x4) 17.4 GW 20 tons Target 17.3 GW 16 tons Near Near (x2) Far m/ 260 mwe km/910 wme Far 290 m/130 wme 1.38 km/460 wme

25 Detector technology Daya Bay IBD: e + p e+ + n RENO Double Chooz PMTs Gamma catcher Target: scin + Gd 25

26 Detector Design: Double Chooz Far Detector operating since early 2011 Scint. 26

27 Detector Design: DB and RENO Daya Bay: Muon IV: Water pool (no scint.) Muon OV: RPCs RENO: No Outer Muon veto IV: water (no scint.) 27

28 Neutrino Selection Prompt signal energy cut Delayed signal energy cut T between prompt delayed Multiplicity cut 28

29 Backgrounds Tagged by OV and IV related + radiactivity Uncorrelated: Radioactivity + neutron like sigan Correlated: Fast neutrons: p recoil + n capture Stopping : + Michel electron cosmogenic isotopes (9Li): n decay 9 8 Li He p n n Th U Background measurements on site 29

30 Experimental results on θ13: Critical view 30

31 Double Chooz first results First results on θ13 from reactor experiments 100 days of data, FD only, Nov 2011 Rate + Shape analysis DC released new results on 2012 Smaller systematic in detector response Larger background reduction OV+ dedicated showering muon veto 31

32 Summary of 2012 results 2 integration periods! sin2(2θ13) days arxiv 32

33 Summary on 7.7 cyh. Minakata a r u cc A s v Is this the end of the road? T.Schwetz n o i s i Prec 33

34 Missing piece: L/E analysis Spectral shape fit is a must to measure θ13: Compatibility with θ13-driven oscillation Rate analysis: any deficit interpreted as θ13 Background model may bias the value of θ13 So far, only in Double Chooz: Spectral fit to θ13 and backgrounds Consistent with θ13 oscillation 34

35 Missing piece: L/E analysis Double Chooz R+S analysis Nobs/Nexp: MC short E/L: no rise Daya Bay RENO R-only analysis R-only analysis Healthy shape Unique shape (?) Longer BL: rise θ13? All experiments show a feature around 6 MeV (?) Rate-Only: this deficit impacts the θ13 value... 35

36 Missing piece: backgrounds Once correlated systematics are canceled (multi-detector setup)... Backgrounds are one of the main systematics sources Need an accurate Rate+Shape knowledge Fast-n and cosmogenics might bias θ13 So far, only in Double Chooz: 9 Li Accidentals R+S background analysis Up to five independent x-checks Fast neutrons Shape: limited by stats 36

37 Accidentals Random coincidences of β decay + n like coincidences in off time window Double Chooz Very well known: no impact on θ13 δbg/signal ~0 S = signal rate B = background rate 1 FAR DET. Rate (d ) δb/b (%) B/S (%) δb/s (%) S(d-1) DC 0.261± % 0.6% 0 45 DB 3.30± % 4.7% 0 70 RENO 0.68± % 0.9%

38 Fast neutrons and stop-µ Fast n: proton recoil+ n capture Stop µ: muon + Michel e (DC, RENO?) Double Chooz IBD Red: Best-fit Spectrum Grey: Tagged IV-OV events DB and RENO: Flat distribution, extrap. from E>12 possible bias up to 25% FD Rate (d 1) δb/b (%) B/S (%) δb/s (%) DC 0.67± % 1.5% 0.4% DB 0.04± % 0% 0% RENO 0.97.±0.06 6% 1.3% 0% If slope, bias on θ13! DC: Fit to IV and OV tagged events (<12MeV!) Best fit: slope, consistent with flat θ13 Fit: pull for rate consistent 38

39 Cosmogenics Spallation products from : β-n emitters (9Li,8He) DC, DB, RENO: Rate estimated from time distribution w.r.t to last muon Double Chooz DC: rate+shape in θ13 fit Shape: MC(KamLAND) + DC data Consistent fit pull, error reduced 1 FAR DET. Rate (d ) δb/b (%) B/S (%) δb/s (%) DC 1.25± % 2.7% 1.2% DC fit 1.00± % 2.2% 0.6% DB 0.16± % 0.2% 0.2% RENO 2.59± % 3.5% 1.0% Large δbg/signal: The most important background 39

40 Total background % 20 db/b FAR DET. Rate (d ) δb/b (%) B/S (%) B/S db/s DC Fit Daya Bay RENO δb/s (%) DC 2.2± DC fit 1.9± DB* 3.7± RENO 4.2± *DB: extra bkg of 0.2±0.2 events/day from Am-C calibration source Day Bay: impressive δb/s (almost no cosmogenics and fast n background) DC: best B/S and unique in providing x-checks: Pulls in a shape-constrained θ13 fit Two integration periods (2R-1R) in θ13 fit Fit of the Observed vs expected rate and reactor-off data 40

41 Total DC background Observed vs Expected Candidates Both Reactors On Both 1 Reactor Off Both Reactors Off Very unlikely in DB an RENO! Data: not background subtracted 41

42 Total DC background Observed vs Expected Candidates 2011 Reactor-off data: 0.84 day Direct total BKG measurement: BKG rate = 2.2 events/day Consistent with estimation: 2011 Reactor-off data! 2.2 ±0.6 event/day Best fit Nexp = 0: 2.9 ± 1.1 event/day Independent BG measurement! 42

43 Total DC background Reactor-Off data 2011 and 2012 reactor-off data samples: 7.53 days arxiv: IBD selection in first DC publication IBD selection in second DC publication 43

44 The Ultimate value of 13 Backgrounds Systematics Predictions 44

45 Current limiting systematics 1.0% background 45

46 Rate Systematics: FD only % Without canceling the correlated systematics... Normalization uncertainties Reactor Flux Efficiency Background Total Double Chooz Daya Bay Error (%) DC DB RENO Reactor Efficiency BKG TOTAL RENO Double Chooz: the best one-detector experiment most accurate knowledge on reactor fluxes The smallest detection systematics 46

47 Rate Systematics: FD + ND Canceling the correlated systematics with the ND... * DC estimates Total Error DC (0.3%) DC (0.1%) Daya Bay Reactor DC* DB RENO Efficiency BKG TOTAL Daya Bay: so far the best multi-detector experiment RENO Error (%) Limited by the uncorrelated reactor systematics Double Chooz: expected competitive with DB Room for improvement in background uncertainty 47

48 The dominant systematic DC prediction: 0.1% 48

49 Uncorrelated flux uncertainty Daya Bay RENO Each ND sees different reactors DB geometry: 0.8%/ 6 = 0.3% (?) RENO: 0.9% Limiting systematic in Daya Bay and RENO Double Chooz ND sees 1 virtual reactor isoflux Error: 0.1% (under study) 49

50 Towards the ultimate Ɵ13value Ultimate = systematics limited, reliable background model, and R+S fit Daya Bay: the most precise result so far Precision limited by uncor. flux uncertainty (0.8%) Also limited ways to test the background model Oscillation shape results not yet available RENO: good precision, but debatable numbers and results Precision limited by uncor. flux uncertainty (0.9%) Also limited by backgrounds (1%): No OV, no scint. IV (can improve?) Double Chooz: shape results, precise background model (can improve!) The best FD-only experiment: good prospects for FD+ND Not limited by uncorrelated flux uncertainty ND not yet available: long time to get enough stats 50

51 Improving DC backgrounds More reactor Off-Off data? Cosmogenics: Getting more statistics Fitting 9Li R+S in θ13 fit Correlated background (FN/SM) Getting larger stats sample Fully exploiting the IV tagging (FN) Use of the outer muon veto (SM) δb/s ~ 0.3% 51

52 The ultimate θ13 value? Assume negligible stats uncertainty Bkackground sys: 0.3% Long time to get stats! unofficial predictions. Rate Only! 52

53 The ultimate θ13 value? ntify a u q o t t l u ut dific b, n o i s i c e ase pr e r c n i l l i w lysis Shape ana Background sys: 0.5% x2 improvement Background sys: 0.5% x3 improvement DB/RENO: Little room for improvement! unofficial estimates. Rate Only! 53

54 Beyond the standard θ13 analysis 54

55 Double Chooz n-h analysis Neutrino selection based on n captures in H, instead of Gd (F. Suekane) Provides x2 signal statistics (NT + GC) Data sample completely independent Systematic uncertainty very different Excellent x-check of Gd analysis result Better constrain on θ13 by combining H+Gd?. Delayed IBD signal below 3.5 MeV Extended T cut Dominant background: Accidentals But uncertainty very small!. 55

56 Double Chooz: n-h results Paper in preparation 56

57 All in all... 57

58 Summary Reactor experiments have proven θ13>0 Daya Bay, RENO and Double Chooz DB: >5σ, DC: shape analysis, backgrounds Is this the end of the road? Accuracy vs precision: beyond the # of σ The most accurate θ13: Oscillation shape analysis Improved background model Reduced uncorr. reactor flux sys? 58

59 Summary (II) The Ultimate measurement still to come! DB: powerful setup and great performance 8 different detectors, high fluxes, small backgrounds, Limited by reactor flux uncertainties RENO: large exposure, debatable numbers Limited by flux uncertainty (background also?) Difficult to predict its evolution Double Chooz: can be competitive with DB Very precise knowledge of the background, reactor-off Simple baseline: L/E, negligible flux systematic (ND+FD) n-h analysis: improve precision on θ13? No Near Detector yet... so need to wait! 59

60 Thank you! Photo: Lola Garrido 60