1. Systematic errors of reactor neutrino experiments and ideas about new detectors Yifang Wang Institute of High Energy Physics, Beijing Nov.28, 2003

2. How big sin 2 2  13 is ? J. Bahcall et al., hep-ph/0305159M. Maltoni et al., hep-ph/0309130 Sin 2 2  13 ~ 3%, Chooz limit: Sin 2 2  13 < 10%

3. Summary about systematic errors There are three main types of errors: reactor related(3%), background related(0.5-4%) and detector related(3%) Use two detectors, far/near to cancel reactor related errors completely and some of detector/background related errors Use movable detectors, near-far, to cancel all backgrounds and some of detector related errors Sufficient shielding to reduce backgrounds Can we do better than 1% ???

4. Possible arrangement 500m 1500m N f =P 1500  (A+B)/1500 2 N n =P 500  (A+B)/500 2 Reactor errors canceled out exactly

5. Another Possibility 1500m 300m N f =P 1500  (A+B)/1500 2 N nA =P 300  (A)/300 2 + P 1237  (B)/1237 2 N nB =P 300  (B)/300 2 + P 1237  (A)/1237 2 1200m ~0.1% @  A ~ 1.5% Reactor errors:

6. 100m is better than 500 m 500/1500 100/1500

9. Systematic errors of past reactor experiments

10. KamLAND 1000t scintillators Shielding: 3000 MWE/3m Water 180 km baseline Signal: ~0.5/day Eff. ~40% BK: corr.: ~0.001/day uncorr. ~0.01/day

11. Large error on fiducial volume

12. Systematic errors E>0.9 MeVE>2.6 MeV Total LS mass2.13 Fiducial mass ratio4.03 Energy threshold--2.13 Efficiency of cuts2.06 Live time0.07 Reactor power2.05 Fuel composition1.0 Time lag0.28  spectra 2.252.48 Cross section0.2 Total6.06.4

13. Experience gained Very good shielding Balloon not good  target mass not well defined Light transport in scintillator unknown  particularly bad for Large detectors  large error on position reconstruction Background from 8He/9Li Not good enough veto tracking system 3m water shielding gives a neutron reduction of >13*10 6 (high energy).

14. Systematic error comparison Chooz Palo VerdeKamLANDCancel ? Reactor power0.7 2.05 <0.1% Reactor fuel/  spectra 2.0 2.7 cross section 0.30.2 0 No. of protonsH/C ratio0.8 1.7 0 Mass--2.1 <0.1 EfficiencyEnergy cuts0.892.10.26 0.2 Position cuts0.323.5 0.2 Time cuts0.40. 0.1 P/Gd ratio1.0- 0 n multiplicity0.5- <0.1 backgroundcorrelated0.33.31.8 <0.1 uncorrelated0.31.80.1 <0.1 Trigger02.90 <0.1 livetime00.2 <0.1

15. Important point to have small systematic error Energy threshold less than 0.9 MeV Homogeneous detector Scintillator mass well determined Target scintillator all from one batch, mixing procedures well controlled Not too large detector Comprehensive calibration program Background well controlled  good shielding Be able to measure everything(Veto ineff., background, energy/position bias, …) A lot of unforeseen effects will occur when looking at 0.1% level

16. Further reduce systematic errors: multiple modules Many modules, 8t each, 100-200 8”PMT/module 1-2 at near, 4-8 at far, small enough for movable calibration Correlated error cancelled by far/near Uncorrelated error can be reduced Event rate: near: ~500-2000/day/module Far: ~40/day/module 100 days calibration at the near pit  0.2-0.5% statistical error Two reference modules 100 days, others ~ 10 days calibration Gd-scintillator oil

17. Advantages with multiple modules Smaller modules have less unknowns Multiple handling to control systematic error Easy construction Easy movable detector Scalable Easy to correct mistakes

18. Fe or low acticity concrete RPC water module

19. Structure of the module Three layers module structure I. target: Gd-loaded scintillator  -ray catcher: normal scintillator III. Buffer shielding: oil Advantages: –Well defined fiducial volume –No cut on position  small systematics Disadvantages: –Complicated mechanical structure –Light yield matching/energy bias ? Need careful MC study I II III

20. Preliminary results: 100PMT/8t Energy resolution

21. position resolution Position bias Cubic Cylindrical

22. Cylinder with reflection CaseTotal PE  E/E (%)  E/E (%)(fit)  R (cm) No reflection, = 11m 41914.27.711 No reflection, = 7 m 38814.97.911 Reflection 90%, = 11m 7195.3 7.117 Reflection 90%, = 7m 6465.87.017 8MeV, Light yield 7000/MeV, QE=20%, QE D1 =60%, Buffer=50cm, 100 8” PMT

23. Background - correlated Cosmic-muon-induced neutrons: –B/S < 0.005  1/day @ ~1km –Can be measured by veto tagging, accuracy<20% –Veto rate < 1KHz, 2-3 layers RPC(1600-2400 m 2 ) ? –Methods: Overburden > 100 MWE Active Veto, ineff. < 0.5%, known <0.2% –Three scenarios: 100 MWE300 MWE1000 MWE muon rate/m 2 (Hz) 4 0.4 0.02 n rate in rock/m 3 (/day) 11000 1600 160 reduction required (10 6 ) 9.2 1.4 0.14 Shielding (water equivalent) (m) 2.5m 2.1m 1.5m

24. Other correlated backgrounds  -neutron instable isotopes from cosmic  – 8 He/ 9 Li, Br(n) = 12%/48%, 9 Li dominant –Production rates = f  ·N A ·  ·Br – Depth > 300 MWE, best 1000 MWE 100 MWE300 MWE1000 MWE Average E  (GeV) 36 64 160 muon rate/m 2 (Hz) 4 0.4 0.02 Cross section (  b) 0.61 0.94 1.86 8 He/ 9 Li (1/day/module) 3.4 0.53 0.053

25. Background - Uncorrelated B/S < 0.05  < 8/day @ far site Can be measured by swap method, precision ~  B/s=2.5%/day single rate @ 0.9MeV < 50Hz 2· R  · R n ·  < 0.04/day/module Methods: –Low activity glass for PMT, > 0.5m oil shielding (dominant!) –3 MWE shielding, low activity sand/aggregate or Fe ? –Rn concentration < 20 Bq/m 3, N 2 flushing ? –(U, Th, K) in Scintillator < 10 -13 g/g, clean Gd –All mechanical structure made of low activity materials –Calibration gadget made of clean materials such as Teflon, … –Clean everywhere, no dust, no …

26. Radiopurity of some materials 232 Th [ppb] 238 U [ppb] 40 K [ppb] 60 Co [mBq/kg] Packard PPO<4<2<0.3- Carbon fiber 56568383 <0.9 Chinese glass(3) 140  2057  5108  6 Schott glass 8246<43 47  73  0.9 SuperK glass 470  50480  5080  10 KamLAND concrete 2630  501280  301470  20 Kamioka dust 1800  100620  402100  30 Mitsui SS<2 2  6 - 20  2 Welded SS(1) 3.2  0.52.1  0.32.4  0.214  2 Lemo cable5-30<1 3  6 - Nylon tube<2<0.7<0.15 Kevlar braid washed<2<0.7 3.9  0.2

27. Our estimate of systematic errors per module Reactor < 0.1% Background < 0.2% Energy cut ~ 0.2% Position cut ~ 0.2% Time cut < 0.1% Livetime ~ 0.1% Other unexpected <0.2% Total < 0.5%

28. Budget for detector(6 module) Unit price($)Quantity Total ($) PMT 1500(100 – 200)*6 900K – 1800K Scintillator 10/kg 8000*6 480K Buffer oil/scintillator 2/kg 8000*6 96K Outer Tank 10000 6 60K Inner Tank 10000 6 60K Electronics/HV 400(100-200)*6 240K – 480K RPC 150/m 2 800-1200 120K – 180K RPC electronics 30/ch 4000-6000 120K – 180K Mechanics+shielding 500K 1 Triger + Online/offline 100K 1 Contingency 500K 1 Total 3176K – 4436K

29. R&D efforts Low radioactive glass PMT Low radioactive Gd-loaded liquid scintillator –Development –Mixing/purification –Monitor/Aging study Mechanics of shielding structure A lab for Low radioactive materials Mass production of glass RPC

30. Summary Measuring Sin 2 2  13 is a very important experiment, an unique opportunities for us Systematic error can be controlled to ~0.5%. Daya bay power plant is a good choice for the experiment