The Mu3e Experiment - Goal

Christian Graf 14.3.2016 The Mu3e Experiment A Calibration Scheme for the Mu3e Tile Detector Master Thesis Presentation for MPI Munich

The Mu3e Experiment - Goal p µ DECAY MODES Fraction (Γ i /Γ) Confidence level (MeV/c) e ν e ν µ 100% 53 e ν e ν µ γ [d] (1.4±0.4) % 53 e ν e ν µ e + e [e] (3.4±0.4) 10 5 53 Lepton Family number (LF )violatingmodes e ν e ν µ LF [f ] < 1.2 % 90% 53 e γ LF < 5.7 10 13 90% 53 e e + e LF < 1.0 10 12 10-16 90% 53 e 2γ LF < 7.2 10 11 90% 53 Measure the decay µ + e + e + e - with a sensitivity of 10-16 Heavily suppressed in the SM by ~10-50 2

Background µ DECAY MODES Fraction (Γ i /Γ) Con e ν e ν µ 100% e ν e ν µ γ [d] (1.4±0.4) % e ν e ν µ e + e [e] (3.4±0.4) 10 5 Total energy of eee should add up to mµ Irreducible background: Internal conversion decays - Only distinguishable via Emiss Excellent momentum resolution needed (~0.5 MeV) ν e - ν e + Accidental Background - Vertex resolution (< 200µm) - Time resolution (< 100ps) ν e + ν 3

Rate Recurl Stations Sensitivity Phase 1 1x10 8 2 10-15 Phase 2 2x10 9 4 10-16 Pixel Detector Fibre Detector Tile Detector

Rate Recurl Stations Sensitivity Phase 1 1x10 8 2 10-15 Phase 2 2x10 9 4 10-16 Pixel Detector Fibre Detector Tile Detector

Tile Detector 3600 tiles (56x60) per station Tiles: 6.5mm x 6.0mm x 5.0mm Light detected by Silicon Photomultipliers (SiPMs) Read-out by STiC, directly under tiles Test-beam results time-resolution: ~60ps 6

A Calibration Scheme for the Mu3e Tile Detector Track Tile Matching Time Calibration of the Tile Detector 7

Track Tile Matching Next step towards full reconstruction: Assign time-stamps to tracks 1) Propagate track as a helix 2) Calculate point of impact on tiles 3) Find best suited hit in a certain range 8

Time Resolution of Signal Decays 700 600 Time Difference of all matched tracks in a frame h_dt_ic_all h_dt_ic_all Entries 535094 Mean 0.06135 RMS 19.07 500 400 300 200 # Entries 800 700 Entries 30161 χ 2 / ndf 44.5 / 23 Constant 699 ± 9.8 Mean 2.33 ± 1.38 Sigma 123 ± 1.2 100 0-60 -40-20 0 20 40 60 dt [ns] 600 500 400 300 200 100 0-1500 -1000-500 0 500 1000 1500 Time Difference [ps] No vertex fit: ~70ps of additional time resolution 9

Time Calibration of Tile Detector Tiles are connected by cables with different lengths signals arrive at different times Time calibration needed of ~10ps in ~10min 1) LED system 2) Easier: use muon decays in normal operation 10

Events for Calibration Need events where we exactly know the time difference between several hits 1) Internal conversion decays: - BR: 3x10-5 takes too long! 2) Hit Cluster: ~ 50% of particles hit more than one tile + high rate - calibration just between neighbors -Position 1 Cluster Map [%] 1.1 1.5 60 50 0 1.2 4.3 39.3 100.0 40 30-1 1.4 3.4 11.5 11.2 20-2 10 11-3 -2-1 0 1 z-position

Cluster Calibration Calibrate neighbors and propagate through whole detector calibrate in z direction calibrate in φ-direction 12

Calibration Results # Entries 160 140 Entries 3360 2 χ / ndf 109 / 69 Constant 143 ± 3.2 Mean -10 ± 0.3 Sigma 18.2 ± 0.3 120 100 80 60 40 20 0-100 -80-60 -40-20 0 20 40 60 80 100 D [ps] σ(tiles) < 20ps in 50ms of simulated data 13

Summary Track tile matching Time calibration with hit clusters # Entries 800 700 Entries 30161 χ / ndf 44.5 / 23 2 Constant 699 ± 9.8 Mean 2.33 ± 1.38 Sigma 123 ± 1.2 600 500 400 300 200 100 0-1500 -1000-500 0 500 1000 1500 Time Difference [ps] Dead time measurement Two test beam campaigns at PSI Number of Det. Photons -- 2. Surface SigAlpha = 0.2 h2_detected Entries 1090 CALICE: Geant4 tile uniformity studies y [mm] 10 5 100 90 80 0 70 60-5 50-10 40-15 30-15 -10-5 0 5 10 15 x [mm] 14

Thank you

Backup

Why? Lepton flavor violation (LFV) in neutrino sector SM forbids charged lepton flavor violation (at tree level) new physics µ eee tests on loop and tree diagrams Loop diagram: SUSY, Little Higgs, Seesaw models, Leptoquarks Tree diagram: Z, LFV Higgs, Extra Dimensions 17

µ Decay - Signal & Background µ DECAY MODES Fraction (Γ i /Γ) Con e ν e ν µ 100% e ν e ν µ γ [d] (1.4±0.4) % e ν e ν µ e + e [e] (3.4±0.4) 10 5 Total energy of eee should add up to mµ Internal conversion background Only distinguishable via Emiss Excellent momentum resolution needed 18

Accidental Background µ DECAY MODES Fraction (Γ i /Γ) Con e ν e ν µ 100% e ν e ν µ γ [d] (1.4±0.4) % e ν e ν µ e + e [e] (3.4±0.4) 10 5 ν e - ν e + Overlay of: Michel decays, radiative decays and internal conversion decays ν e + ν with processes as: Bhabha, compton or misreconstr. Dominant: Bhabha + 1 Michel requires: - momentum resolution - vertex resolution (< 200µm) - time resolution (< 100ps) bhabha 19

Tracking Pixel tracker for mom. resolution Multiple scattering (MS) dominates low material budget HV-MAPS technology Use recurler & high acceptance small, long tube design B 50 MeV/c 25 MeV/c 12 MeV/c 20

The Detector: Phase Ia Rate 2x10 7 Sensitivity 10-14 40cm Minimal configuration 4 layers of HV-MAPS (thinned to ~50µm) Glued on kapton foil, self-supporting Momentum resolution: 0.5 MeV 21

The Detector: Phase Ib Rate 1x10 8 Sensitivity 10-15 Adding sci. fibres in central station + recurl stations with sci. tiles Combined with Silicon Photomultipliers Time resolution fibres: ~1ns, tiles: ~100ps Limited by statistics 3-5 layers of 250µm fibres 22

The Detector: Phase II Rate 2x10 9 Sensitivity 10-16 Higher rate at a possible new beam-line Size matters: add two more recurl stations 23

Simulation Full detector simulation in Geant4: Optimize geometry Analysis with truth data: e. g. Background estimation Reconstruction development Vertex fit Track fitting Track-Tile matching Track-Fibre matching 24

Track Fitting No hardware trigger, read out everything Reconstruct every event on an online filter farm has to be fast! Multiple scattering fit Describe track as a sequence of triplets MS in middle hit of triplet Track = weighted mean of triplets 25

Tile Detector Studies rate [MHz] 0.6 0.5 h_total Phase II h_total Entries 172678 Mean -131.7 RMS 467 0.4 0.3 0.2 0.1 0-1000 -800-600 -400-200 0 200 400 600 800 1000 z position [mm] 26

Tile Detector Studies Phase II phi position 50 40 h_hitmap_track h_hitmap_track 1 Entries 64 Mean x -21.64 0.9 Mean y 28.25 RMS x 33.71 0.8 RMS y 16.75 0.7 phi position 50 40 h_hitmap_digi h_hitmap_digi Entries 161 Mean x -22.94 45 Mean y 27.37 RMS x 42.85 40 RMS y 16.41 35 0.6 30 30 0.5 30 25 20 0.4 0.3 20 20 15 10 0.2 10 10 0.1 5 0-100 -50 0 50 100 z position 0 0-100 -50 0 50 100 z position 0 27

Kinematics of Particles Hitting the Tile Detector # Entries / 1000 5 4 3 2 1 Entries 257318 10 Entries 257318 # Entries / 1000 8 6 4 2 0 0 10 20 30 40 50 60 Transverse Momentum [MeV] 0 0 10 20 30 40 50 60 z-momentum [MeV] 28

Track Tile Matching Resolution # Entries 800 700 600 2 χ / ndf 771.9 / 371 Constant 698.7 ± 3.8 Mean 2.164 ± 0.271 Sigma 64.42 ± 0.22 500 400 300 200 100 0-400 -200 0 200 400 Time Difference [ps] Phase 2 29

Calibration Precision over n Calibration Uncertainty [ps] 60 55 50 45 40 35 30 χ 2 / ndf 0.647 / 4 p0 135 ± 1.83 p1-4.05 ± 0.523 25 20 15 10 20 30 40 50 Calibration Time [ms] 30

Beam Infrastructure Cockcroft-Walton: accelerates hydrogen atoms to 870 kev 31

Beam Infrastructure Injector 2: 72 MeV Cyclotron 32

Beam Infrastructure Ring Cyclotron: 590 MeV 33

Beam Infrastructure Target E (rotating): 40mm of graphite generates π + -> µ + 34

Beam Infrastructure SINQ new beam line for phase 2 35

Misplaced SiPM y [mm] 15 10 Ratio of Det. Photons: SiPM 1mm deeper h2_detected Entries 951 1.4 1.3 y [mm] 15 10 Ratio of Det. Photons: SiPM 1mm higher h2_detected Entries 700 1.4 1.3 5 1.2 5 1.2 1.1 1.1 0 1 0 1-5 0.9-5 0.9-10 0.8 0.7-10 0.8 0.7-15 -15-10 -5 0 5 10 15 x [mm] 0.6-15 -15-10 -5 0 5 10 15 x [mm] 0.6 36

Thinner Tile y [mm] Ratio of detected photons: 2.8mm / 3.0mm thickness 15 10 h2_detected Entries 967 1.1 5 1.05 0 1-5 -10 0.95-15 -15-10 -5 0 5 10 15 x [mm] 0.9 37