Neutrino Experiments with Reactors

Transcription

1 Neutrino Experiments with Reactors 1 Ed Blucher, Chicago Reactors as antineutrino sources Antineutrino detection Reines-Cowan experiment Oscillation Experiments Solar Δm 2 (KAMLAND) Atmospheric Δm 2 -- θ 13 (CHOOZ, Double-CHOOZ, Daya Bay) Conclusions Lecture 1 June 2008 NUFACT 08: Neutrino Summer School

2 Reactors as Antineutrino Sources 2 Reactors are copious, isotropic sources of ν e. ν e ν e ν e ν e β decay of neutron rich fission fragments and U and Pu fission

3 Example: 235 U fission 3 U n X + X 2n Stable nuclei with most likely A from 235 U fission: Zr 58 Ce Together, these have 98 p and 136 n, while fission fragments (X 1 +X 2 ) have 92 p and 142 n On average, 6 n have to decay to 6 p to reach stable matter ~ 200 MeV/fission and ~6 ν e / fission implies that 3GW th reactor ν 20 produces ~ 6 10 e / sec.

4 > 99.9% of ν are produced by fissions in 235 U, 238 U, 239 Pu, 241 Pu 4 Fissions per sec (fraction of total rate) Fission rates over single reactor fuel cycle 235 U 239 Pu Plutonium breeding over fuel cycle(~250 kg over fuel cycle) changes antineutrino rate (by 5-10%) and energy spectrum

5 Each reactor core is an extended antineutrino source: ~ 3 m in diameter and 4 m high. 5

6 Reactor refueling 1 month shutdown every months 1/3 of fuel assemblies are replaced and remaining fuel assemblies repositioned 6

7 ν Rate for 3-Core Palo Verde Plant 7

8 60.00 Detection of + Inverse β Decay: ν e + p e + n ν e Also possible: + ν e + d e + n+ n ν + e ν + e e e Neutrino Flux Cross Section # of Events Arbitrary Scale Enu (MeV) E ~ M + m M th n e+ p = MeV, so only ~1.5 ν e / fission can be detected. ~1 event per day per ton of LS per GW thermal at 1 km

9 Experiments detect coincidence between prompt e + and delayed neutron capture on hydrogen (or Cd, Gd, etc.) + ν e + p e + n 9 τ 200 μs p + n d + γ ( 2.2 MeV ) n γ 2200 kev ν e e + γ 511 kev kev 1.8 MeV γ 511 kev E E ) ν e + + E + ( M M + n n p m e + Including E from e + annihilation, E prompt =E ν 0.8 MeV

10 The First Detection of the Neutrino Reines and Cowan, Clyde Cowan Jr. Frederick Reines

11 1953 Experiment at Hanford 11

12 Detector from Hanford Experiment 12 γ γ 300 liters of liquid scintillator loaded with cadmium Signal was delayed coincidence between positron (2-5 MeV) and neutron capture on cadmium (2-7 MeV) High background (S/N~1/20) made the experiment inconclusive: 0.41 ± 0.20 events / minute

13 1956: Savannah River Experiment 13 Tanks I, II, and III were filled with liquid scintillator and instrumented with 5 PMTs. Target tanks (blue) were filled with water+cadmium chloride. Inverse β decay would produce two signals in neighboring tanks (I,II or II,III): - prompt signal from e+ annihilation producing two MeV γs - delayed signal from n capture on cadmium producing 9 MeV in γs

14 Data were recorded photographically from oscilloscope traces 14 I II III

15 Savannah River Experiment 15 Shielding: 4 ft of soaked sawdust Electronics trailer

16 By April 1956, a reactor-dependent signal had been observed. 16 Signal / reactor independent background ~ 3:1 In June of 1956, they sent a telegram to Pauli: A Science article reported that the observed cross section was within 5% of the cm 2 expected (although the predicted cross section had a 25% uncertainty). In 1959, following the discovery of parity violation in 1956, the theoretical cross section was increased by 2 to (10±1.7) cm 2 In 1960, Reines and Cowan reported a reanalysis of the experiment and quoted (12 7 σ = + ) 10 cm 4

17 17 Excellent account of Reines and Cowan experiments: R. G. Arms, Detecting the Neutrino, Physics in Perspective, 3, 314 (2001).

18 Oscillation Experiments with Reactors 18 Antineutrinos from reactors can be used to study neutrino oscillations with solar Δm 2 12 ~ ev 2 and atmospheric Δm 2 13 ~ ev 2 Mean antineutrino energy is 3.6 MeV. Therefore, only disappearance experiments are possible Δm L Δm L P( ν ) 1 sin 2 12 e νe θ13 sin cos θ13 sin 2θ12 sin, 4E 4E Δ m = m m where ij i j. Experiments look for non-1/r 2 behavior of antineutrino rate. Oscillation maxima for E ν =3.6 MeV: Δm 2 12 ~ ev 2 L ~ 60 km Δm 2 13 ~ ev 2 L ~ 1.8 km

19 Long history of neutrino experiments at reactors m KamLAND 6 m CHOOZ 1m Poltergeist

20 Normalization and spectral information 20 Predicted spectrum θ 13 =0 (from near detector) Observed spectrum (far detector) sin 2 2θ 13 =0.04 E ν (MeV) E ν (MeV) Counting analysis: Compare number of events in near and far detector Systematic uncertainties: relative normalization of near and far detectors relatively insensitive to energy calibration Energy spectrum analysis: Compare energy distribution in near and far detectors Systematic uncertainties: energy scale and linearity insensitive to relative efficiency of detectors

21 Issues affecting oscillation experiments 21 Knowledge of antineutrino flux and spectrum Detector acceptance Backgrounds: Uncorrelated backgrounds from random coincidences Reduced by limiting radioactive materials Directly measured from rates and random trigger setups Correlated backgrounds Neutrons that mimic the coincidence signal Cosmogenically produced isotopes that decay to a beta and neutron: 9 Li (τ 1/2 =178 ms) and 8 He (τ 1/2 =119 ms); associated with showering muons. Reduced by shielding (depth) and veto systems

22 22 In absence of a direct measurement, how well can antineutrino rate and flux be determined from reactor power? Recall that > 99.9% of ν are produced by fissions in 235 U, 239 Pu, 238 U, 241 Pu (90% from first two). Use direct measurements of electron spectrum from a thin layer of fissile material in a beam of thermal neutrons A. Schreckenbach et al. Phys. Lett B 160, 325 (1985); A. Hahn et al., Phys. Lett. B 218, 365 (1989). Must rely on calculation for 238 U. Total flux uncertainty is about 2-3%. Uncertainty can be checked with short-baseline experiments (Gösgen, Bugey)

23 β Spectrum for 235 U Fission Products 23

24 24 Positron Spectra from Gösgen Experiment The two curves are from fits to data and from predictions based on Schreckenbach et al.

25 25 Bugey3/ first principle calculation Bugey3/ best prediction (uses β spectra where possible and calculation for 238 U)

26 Solar Δm 2 : The KamLAND Experiment 26 Goal: Study oscillations at Δm 2 solar ~ ev 2 using nuclear reactors. Observed / expected flux for reactor experiments 2 Δm atm 2 Δm solar Predicted N obs /N exp for LMA solution before KamLAND (assuming θ 13 small)

27 ~100 km baseline requires very large detector and very large ν source (and very deep experimental site). Large concentrations of reactors in U.S., Europe, and Japan. 27

28 28 Kashiwazaki KamLAND uses the entire Japanese nuclear power industry as a long-baseline source KamLAND Takahama Ohi

29 29

30 Reactors contributing to antineutrino flux at KAMLAND Japan South Korea Site Dist (km) Cores (#) P therm (GW) Flux (cm -2 s -1 ) Rate noosc * (yr -1 kt -1 ) Kashiwazaki Ohi Takahama Tsuruga Hamaoka Mihama Sika Fukushima Fukushima Tokai Onagawa Simane Ikata Genkai Sendai Tomari Ulchin Yonggwang Kori Wolsong Total Nominal

31 A limited range of baselines contribute to the flux of reactor antineutrinos at Kamioka 31 Korean reactors: 3.4±0.3% L ~ 180 km Rest of the world +JP research reactors: 1.1±0.5% Japanese spent fuel: 0.04±0.02%

32 KAMLAND Detector 32

33 Looking up at containment vessel 33

34 Antineutrino signature: coincidence between prompt e + and delayed neutron capture on hydrogen + ν e + p e + n 34 τ 200 μs p + n d + γ ( 2.2 MeV ) n γ 2200 kev ν e e + γ 511 kev kev 1.8 MeV γ 511 kev E E ) ν e + + E + ( M M + n n p m e + Including E from e + annihilation, E prompt =E ν 0.8 MeV

35 Anti-Neutrino Candidate 35 (colour is time) Prompt Signal E = 3.20 MeV Δt = 111 ms ΔR = 34 cm Delayed Signal E = 2.22 MeV

36 Delayed vs. Prompt Energy for ν e Candidates 36 γ from n 12 C

37 Energy Calibration with Sources 37 ΔE/E ~ 7.5% / E, Light Yield: 260 p.e./mev

38 Test of Position Reconstruction Along Vertical Axis 38

39 Detector Performance 39 σ E ~7.5%/ E( MeV ) σ R ~20 cm

40 Event Selection Requirements 40 Fiducial volume: R < 5 m Time correlation: 0.5 μsec < Δt < 660 μsec Vertex correlation: ΔR< 1.6 m Delayed energy: 1.8 MeV < E delay < 2.6 MeV Prompt energy: E prompt > 2.6 MeV Muon veto: 2 msec veto after any muon + 2-sec veto following a showering muon + reject events within 2 sec and 3 m of muon tracks

41 Estimated Systematic Uncertainties 41 % Total LS mass 2.1 Fiducial mass ratio 4.1 Energy threshold 2.1 Selection cuts 2.1 Live time 0.07 Reactor power 2.0 Fuel composition 1.0 Time lag 0.28 ν e spectra 2.5 Cross section 0.2 Total systematic error 6.4 %

42 Observed Event Rates with E prompt > 2.6 MeV (Data collected from March--October 2002) 42 Observed Expected Total Background 54 events 86.8 ± 5.6 events 1 ± 1 events accidental ± Li/ 8 He 0.94 ± 0.85 fast neutron < 0.5 Inconsistent with 1/R 2 flux dependence at % c.l.

43 N obs N N expected bckg = 0.611± 0.085( stat) ± 0.041( syst) 43

44 Oscillation Effect in Rate and Energy Spectrum 44

45 reactor neutrinos geo neutrinos accidentals Events/0.425 MeV MeV (analysis) KamLAND data no oscillation best-fit oscillation sin 2 2θ = 1.0 Δm 2 = 6.9 x 10-5 ev 2 Energy spectrum consistent with oscillations at 93% c.l., but also consistent with no oscillation shape at 53% c.l. 10 Need more data Prompt Energy (MeV)

46 Allowed Values of Δm 2 and sin 2 2θ 46 Best fit : Δm 2 = 6.9 x 10-5 ev 2 sin 2 2 θ = 1.0

47 In 2005, with more data, clear evidence for spectral distortion (Data collected from March 2002 to January 2004) 47 Expected w/o oscillations: 365 ± 24 (syst) events Observed: 258 events No oscillation shape only consistent with observed spectrum at 0.5% level.

48 Observed ν spectrum / expectation with no oscillations 48 Shape-only analysis gives: Δm 2 =(8.0 ± 0.5) 10 5 ev 2

49 49

50 2 Allowed Values for Δm 12 and θ12 50 Best fit (assuming CPT): Δ m = (7.9 ) 10 ev and tan θ =