1. Neutrino Studies with the T2K P0D Detector Dmitriy Beznosko NN Group Stony Brook By: Dmitriy Beznosko Stony Brook University For the T2K Collaboration DPF2011

2. Outline T2K Overview ND280 Description P0D Detector ▫Physics goals ▫Design ▫MPPC photo detectors ▫Extruded Scintillator and WLS fiber ▫Cosmic ray tests, scanning tests ▫Calibration, efficiency ▫Analyses overview Summary 2

3. T2K Overview Designed for up to 0.75 MW, 30 GeV proton beam ~0.6GeV off-axis neutrino beam Data taking started at the end of 2009 A near detector (ND280) and far detector (SuperK) 3 T2K is Tokai to Kamioka long baseline neutrino oscillation experiment in Japan To measure  13 by e appearance from  beam 30 GeV PS 180MeV LINAC

4. T2K-ND280 4 On-axis Detector INGRID 37m Off-axis detector neutrino beam On – Axis Detector Beam monitoring Beam direction and profile Beam Intensity Off – Axis Detector Neutrino beam energy Spectrum Beam flux Backgrounds to e appearance measurement: Beam intrinsic e Cross Sections: Primarily NC  0 [2]

5. Off-axis sub-detectors ▫ P0D - Pi-Zero Detector ▫Tracker:  TPC - time projection chamber  FGD - fine grained detector ▫ECAL - electromagnetic calorimeter:  Downstream  P0D  Barrel ▫SMRD - Side Muon Range Detector (inside yoke) ▫UA1 magnet - ~0.2T 5 T2K ND280

6. Physics Goals – P0D P0D is designed to measure: ▫Intrinsic beam e ▫Cross-sections on H 2 O (for systematics reduction at SuperK detector)  NC  0  CC inclusive, including  CCQE  CC1  In this talk my analysis with one muon and one  + events are considered ▫CCp  + and CCn  + (as CC1  + ) ▫CC1  + resonant production isospin related to total CC1  cross-section ▫CC1  + /CCQE Data to MC prediction ratio is to be measured to remove beam flux uncertainties ▫To tune MC as input for oscillation analysis 6 SK ν e Event Selection See Glen Lopez's talk

7. T2K – P0D P0Dule: two perpendicular arrays of triangular scintillator bars + Pb or brass/water radiator. 134 vertical bars (Y, 2200 mm) 126 horiz. bars (X, 2340 mm) Each bar: Single coaxial hole 1mm wavelength-shifting (WLS) fiber (Kuraray multi-clad, S-35, J-type, doped with Y-11 (175 ppm) ) One end is mirrored The other end is optically coupled to a Hamamatsu MPPC for readout. Total of 10,400 channels 7 MPPC photo detector Scintillator bars TFB readout board coaxial cables

8. Neutrino Beam T2K-P0D Schematic Drawing A total of 40 P0Dules were used for P0D Total length of all P0Dules & targets assembled ~2.4m Ecals –7 P0Dules each, with 4mm Pb radiator Both Water Targets– 26 P0Dules, 25 water layers with 1.6mm brass sheets Fillable water bags for water-in and water-out measurements 8

9. MPPC Photosensor – Basic Characteristics For model S10361-050U ▫Chip size - 1.5 x 1.5 mm 2 ▫Active area - 1 x 1 mm 2 ▫Pixel size - 50 x 50  m 2 ▫Number of pixels – 400  (a sub-model S10363-050U for T2K ~667, non square) ▫Pixel effective size - 38.1×38.8  m 2 ▫ Geometric efficiency - 61.5% ▫PDE typ. ~40%-50% (400nm) ▫Gain typ. 7*10 5 ▫Time resolution – 220 ps ▫Temp. coeff. of bias voltage – 50 mV/ o C ▫Dark count rate ~270. 10 3 /sec 9 S10361-050U MPPC

10. Performance Measurements Independent pixels – PE separation Complex noise – dark rate, afterpulse, crosstalk 10 * “Study of Afterpulsing of MPPC with Waveform Analysis”, Hideyuki Oide et al. ADC Counts

11. MPPC Large Scale Deployment Each sensor tested before installation ▫Found 83 ‘bad’ (strongly deviated from expected characteristics) from ~11k shipment Additional testing after installation ▫Using dark noise ▫During scanning and after superp0dule assembly ▫Found 14 inoperable  2 had surface damage After transit to Japan ▫2 more sensors bad ▫After 1.5 years of operation ~17 bad channels total (~0.16%) ~12 damaged during installation Bias spread between sensors ▫Manuf. supplied biasing voltage values vs. breakdown voltage 11

12. Extruded Scintillator Bars 12 Scintillator emission spectrum PMT response radioactive source position along the bulk triangular shape scintillator without WLS fiber, yielding attenuation lengths of ~10.2. D. Beznosko, A. Dyshkant, C.K. Jung, C. McGrew, A. Pla-Dalmau, V. Rykalin, "MRS Photodiode Coupling with Extruded Scintillator via Y7 and Y11 WLS Fibers", FERMILAB-FN-0796, Feb 2007

13. WLS fiber 13 The PMT output for various WLS fibers (with and without optical grease for each fiber tested). Absorption and emission spectra for WLS fibers (bottom) Response of Y11 and Y7 WLS fibers vs. distance from photodetector. Y11 was selected for P0D. Y11 attenuation lengths: 35.5  0.3 cm and 463.4  1.2 cm D. Beznosko, A. Dyshkant, C.K. Jung, C. McGrew, A. Pla-Dalmau, V. Rykalin, "MRS Photodiode Coupling with Extruded Scintillator via Y7 and Y11 WLS Fibers", FERMILAB-FN-0796, Feb 2007

14. MPPC – Rad. Source Scanning 14 With 0.7  Cu 60 Co source Average response of all bars with source position at 0,0 is plotted (no attenuation correction needed) ~13% rms channels response spread using manufacturer supplied biasing voltage values ADC Counts ~13% RMS ADC Counts mm

15. c 15 DSECAL TPC3 FGD2 TPC2 FGD1 TPC1 P0D Barrel - P0D ECALs SMRD detector   ND280 Detector with Magnet Open

16. Neutrino Event Display in ND280 Neutrino event in P0D. ▫2 tracks contained in P0D 16 P0D TPC1TPC2TPC3 FGD1FGD2

17. Neutrino Event Display in ND280 ‘Sand’ muons going through P0D/TPC1/FGD1/TPC2/FGD2/DSECAL 17 P0D TPC1TPC2TPC3 FGD1FGD2DSEcal

18. 18 Single channel response and all channels spread (~12% rms) p0=16.7±0.33, p1= 0.774±0.043, p3=0.355±0.027 chi2/dof 1.08 Cosmics Data calibration – channels response and fiber light attenuation ADC Counts mm

19. P0D Hits efficiency 19 Track reconstruction hit efficiency for 1-38 P0Dules Values for first and last P0Dule are biased by selection Using ‘sand’ muon long tracks Theoretical probability of 0 hits for 15 bad channels, no double-bad ones (15/10400)*0.27=~0.039% not all edge/side bars are included in data sample Double hits Single hits >2 hits Missed hits Relative response in custom units P0Dule # Efficiency, % Total efficiency = 99.965 % ▫% of 1 hits = 27.475 ▫% of 2 hits = 68.367 ▫% >2 hits = 4.122 % of 0 hits = 0.035 Double hit Single hit

20. P0D Performance plots 20 P0D neutrino beam data ▫XZ distrib. of vertices ▫XY distrib. of vertices ▫Integral of vertices vs Protons on Target Direction to beam axis

21.  0 Analysis Overview NC  0 production contributes one of the largest sources of background to the e appearance search ▫Use measurement of NC  0 in the P0D to constrain systematic uncertainty in the NC  0 background expectation at SK Strategy: Measure NC  0 production on water: ▫ Require neutrino interaction in water target ▫ Perform separate measurements of water in/water out ▫ Make statistical subtraction to get water in cross section Reconstructing  0 's ▫Require vertex in fiducial volume ▫Require no tracks identified as muon-like ▫Reconstruct NC  0 events from EM showers produced by decay photons ▫Identify  0 's using invariant mass 21

22. e Analysis Overview Measure beam e contamination Selection Criteria: ▫Vertex in fiducial volume  Water Target in Z, 25 cm from X/Y sides. For a future water in/out subtraction ▫Single 3D Track ▫Width based PID  Selecting wide tracks rejects muons and charged pions ▫Kinematic Cut - theta 1.5 GeV., to measure high energy neutrinos from Kaon decays in secondary beam first  lower E_nu to measure pion component. 22

23. CC1  + /CCQE Analysis Overview As input to MC tuning for oscillation analysis CC1  + /CCQE ratio removes flux uncertainty and reduces systematics Selection Criteria: ▫Using P0D contained 1, 2 and 3-track events ▫Vertex in fiducial volume (Water Target in Z, 25cm from X/Y sides) ▫Use Track length and angle-to-Z axis cut to obtain very pure CCQE sample from 1 track events  For CCQE baseline direct comparison  For energy spectrum comparison between data and MC files Fit combined MC CCQE, CC1  + and ‘other’ samples to data and determine the ratio 23

24. Summary P0D detector was built and commissioned successfully During run periods, performance is within expected specifications and is stable A range of physics measurements with P0D-only contained events or together with tracker (e.g. CC inclusive analysis) are in progress Water-in water-out measurements to be performed when larger dataset is accumulated 24

25. T2K Overview T2K is Tokai to Kamioka long baseline neutrino oscillation experiment in Japan It includes the first off-axis beamline at J-PARC, a near detector (ND280) and far detector (SuperK) A primary goal is to measure  13 neutrino oscillation parameter by e appearance from  beam The Near Detector contains an on-axis beam monitor and an off-axis detector. ▫ND280 off-axis detector consists of several sub- detectors, all in ~0.2T magnetic field  Accurate momentum measurements  Charge discrimination 25

26. Physics Goals – P0D Main backgrounds to the major goal of T2K affecting  13 by means of e appearance from  beam:  0 production by  where the  0 is mis-ID as e -  e events from the primary beam contamination P0D is designed to measures the following cross-sections on H2O ▫  0 ▫ e ▫CC inclusive  CCQE  CC1  For single pion, CC interactions cross section ~ 3  NC ▫more events detected. ▫Reconstructing CC events is simpler than NC – early test of detector performance. For my analysis, events with one muon and one  + are considered ▫CCp  + and CCn  + (commonly called CC1  + ) ▫CC1  + resonant production isospin related to total CC1  cross-section CC1  + /CCQE Data to MC prediction ratio is to remove beam flux uncertainties ▫To tune MC as input for oscillation analysis 26

27. T2K – Neutrino beamline 27 The T2K Target: Graphite Target in Ti-alloy capsule Focusing horn  From left to right: extracted proton beam, beam target and focusing horn system (toroidal B field), decay volume (filled with Helium), beam dump (not shown), muon monitor, ND280, 2km detector (proposed only), SuperK. [2] 90 m Decay Volume 110m

28. Pixel Recovery -double exponent fit: Short time const - ~19ns Long time const - ~85ns After-pulsing probability (short) - ~6.1% (long) - ~5.9% Cross-talk - ~5.7% 28 Characteristics measurements * “Study of Afterpulsing of MPPC with Waveform Analysis”, Hideyuki Oide et al. Dark Noise Counts vs. Threshold for several bias values

29. MPPC MC simulation Software simulation of electronics components that converts the energy deposited by a particle in the scintillator (from GEANT4) into the PE output of the photosensor + simulates readout electronics effects. Below is the output from used for testing code that was intended for use with cosmic rays simulation and was designed to process small to intermediate signals (output <~300 PE). For larger signals, the simulated value increasingly deviates from the supposed value due to simplifications used. This is now properly handled by ElecSim package in ND280 software. 29 MPPC simulation output for 8000 runs – dark noise MPPC simulation output for 8000 runs – low energy

30. Part of increase is due to cross- talk, dark noise and short after- pulse increase rate. Further increase is limited by narrow gate (~30ns). For this sensor, bias range is ~69.1V to ~70.9V ~100x amplifier is used 30 Single PE Amplitude vs. Bias Signal in PE vs. Bias Characteristics measurements

31. Multi-pixel family – Sensor-fiber alignment 31 Signal amplitude versus the position of the 0.8mm clear fiber for SiPM sensor. Signal amplitude versus the position of the 0.94mm clear fiber for MRS sensor.

32. MRS photo sensors were tested in 9T at FNAL A number of orientations w.r.t. B field tested LED and bias circuit output of B field 32 Multi-pixel family – B-field effect Scintillator UV LED Splitter WLS fiber Front view Light Input via Fiber Power Supply MRS sensor To oscilloscope 10k 

33. Multi-pixel family – B-field effect Using Agilent [12] Infiniium 54832D MSO oscilloscope Two runs at different bias, also measured output area and rise time qT measurement is done directly after magnet quench ▫More in backup slides 33

34. Multi-pixel family – B-field effect 34 Output area (total charge) and output rise time were measured as well Rise time as seen after 4m coax cable Using Agilent [12] Infiniium 54832D MSO oscilloscope Two runs at different bias qT measurement is done directly after magnet quench

35. 35 Multi-pixel family – Irradiation effects – 1Mrad Ratio of noise measurements vs. threshold at set bias before and after: any major changes would indicate damage to the internal cell structure of the sensor. Signal detection: change would indicate that the sensor’s surface has been affected by the radiation. Noise ratios vs. bias: any changes would indicate change in the range of the bias voltage.

36. MPPC – P0D prototype cosmic test Cosmic rays test done on prototype module Using WLS fibers with blackened ends Without Mirror – measured average light yield:  25 cm from MPPC: 19.8±0.5 PE/MeV  205 cm from MPPC: 8.7±0.2 PE/MeV ▫With Mirror estim. (assuming 0.8 reflectivity)  25 cm from MPPC: 23.8 PE/MeV  205 cm from MPPC: 15.7 PE/MeV 36 P0Dule prototype with WLS fibers P0Dule prototype construction

37. JPARC: Upstream Ecal being readied for cosmic run 37 Using P0D UPSTREAM ECAL only At JPARC facility (June 09) Averaged light yield over full bar length - ~20PE/MeV

38. Event Selection criteria Data totalMC totalMC cc1piMC ccqe 1 light track751172386 9462 (13.1%) 46378 (64.1%) 2 tracks: with at least one light track 208020578 6617 (32.2%) 7776 (37.8%) 2 light tracks3053588 1388 (38.7%) 1295 (36.1%) 3 tracks: with at least one light track 2652693 850 (31.5%) 110 (4.1%) with at least two light tracks 50616 283 (45.9%) 18 (2.9%) 38 Data - 1.346e+20 POT Light track – PID 8; Generator Level reaction codes used

39. 2 and 3 track events - Fiducial volume 39 To check data behavior around accepted fv of 25cm from P0D edge (circled) For 3track: any 3track to 1 light track ratio for statistics MCDataMCData 0.0465280.038731 std:0.0004290.000411 0.0472380.039193 std diff:-1.7E-05 0.0475240.039487 0.047260.039505 0.0473770.03989 0.0478470.039604 0.0477090.039913 MCData MCData 0.0943830.108221 std: 0.0003810.000345 0.0944770.108095 std diff: -3.7E-05 0.0951720.107661 0.0947760.107523 0.0954050.108099 0.0951620.107722 0.0950130.107292

40. Data to MC comparison - Single Light Track # events normalized Light track = PID 8 40 chi^2/dof 1.42 chi^2/dof 3.43 As the angle recon is typically not as good as track length -will use (longer) length histograms for fit later

41. Data to MC comparison Three Tracks (at least 2 light) track length # events normalized Light track = PID 8 41 chi^2/dof 0.93 chi^2/dof 0.71 chi^2/dof 0.65

42. Track multiplicity # of events normalized, ~3.5% difference at 1track events Bin 0 is 2d tracks, excluded from normalization 42

43. “Raw” ratio comparison This is 3track/1track ratio (‘raw’ result) Results presented using Data and MC processing 4B (NEUT only) Stat. errors using Poisson Systematics for data also include: energy spectrum diff. (~5%), vertex resolution (~3%) and fv-contained track mid-id (~5% for 1track and ~9% for 3track) Syst. for MC from “Neutrino Interaction Uncertainties” by Ryan Terri, T2K NIWG ▫Here using linear approx. of 30% E/GeV with 34%max value 43 Data 0.00665 ± 0.00094 (stat) ± 0.00083 (syst) Neut 0.00851 ± 0.00034 (stat) ± 0.00268 (syst)