Double Calorimetry in Liquid Scintillator Detectors Marco Grassi - PowerPoint PPT Presentation

Double Calorimetry 
 in Liquid Scintillator Detectors Marco Grassi 
 APC - CNRS (Paris) in collaboration with: Stefano Dusini Anatael Cabrera Margherita Buizza Miao He Pedro Ochoa


 What Technique allowing redundancy for high precision calorimetry 
 within Liquid Scintillator detectors Why Upcoming high-resolution spectral measurements of neutrino interactions How Exploit two independent energy estimators 
 experiencing different systematic uncertainties (possibly implemented through independent detection systems) Disclaimer : limited time ▶︎ illustration rather than full explanation WIN 2017 M. Grassi 2

Motivation Calorimetry of (anti)neutrino interactions Example: θ 13 experiments Resolution dominated by photostatistics σ NST : residual issues in detector modeling 
 Double Chooz - Similar to other Exps after calibration (linearity, stability, uniformity) Next generation detector: 
 improve resolution (more than x2) σ NST no longer negligible σ ST ~ 7% σ NST ~ 2% Understating systematics is pivotal WIN 2017 M. Grassi 3

Two Calorimetry Observables in LS Detectors LS Detector ENERGY LIGHT Mean PMT Illumination DEPOSITION DETECTION λ = ⟨ N(PE) ⟩ / PMT PE/MeV C HARGE I NTEGRATION P HOTON C OUNTING λ ≲ 0.5 λ > 0.5 charge PE = hit PE = PMT gain PMT PMT gain linearity Different Single photoelectron gain = gain(PE)? Systematics threshold REDUNDANCY WIN 2017 M. Grassi 4

Calorimetry in Current LS Experiments Experiments typically implement one single observable PARTLY BECAUSE Deposited energy (signal signature) + detector geometry ▶︎ dynamic range DYB Only Charge 
 Integration Both observables D YNAMIC RANGE AT 1 M E V Why shall we go beyond this paradigm? WIN 2017 M. Grassi 5

Calorimetry in Future LS Experiments DETECTOR ENERGY M UST BE L ARGER TARGET MASS RESOLUTION 6%/ √ E KamLAND 1000 t Sizable difference in collected light detector center vs detector edge D. Chooz 30 t 8%/ √ E RENO 16 t M UST BE M ORE P RECISE Daya Bay 20 t 5%/ √ E Borexino 300 t Unprecedented light level 1200 pe/MeV 20000 t 3%/ √ E JUNO Both features • are highly expensive (civil engineering + photocathode density) • result in extreme detector dynamic range • reactor antineutrino detection yields λ ∈ [0.07,~50] in JUNO WIN 2017 M. Grassi 6

Deal with the detection of 1200 wild photoelectrons…

JUNO Calorimetry Light is not enough 1 MeV typical LS exp Energy σ NST ~ 2% Deposition σ NST needs to be controlled at better than 1% level Redundancy in systematics evaluation is pivotal WIN 2017 M. Grassi 8

Double Calorimetry: born within JUNO to better control / assess the resolution non-stochastic term

Double Calorimetry in Action: Energy Reconstruction E = f × PE f : calibration PE: raw detector response Uniformity Position dependent ACCOUNTED FOR USING Stability Time dependent Linearity Energy dependent WIN 2017 M. Grassi 10

Double Calorimetry in Action: Energy Reconstruction E = f × PE f : calibration PE: raw detector response Uniformity Position dependent ACCOUNTED FOR USING Stability Time dependent Linearity Energy dependent Limited dynamic range E [MeV] = f ABS x f U (r) x f S (t) x f L (PE) x PE Nowadays σ (E)/E 
 (eg θ 13 experiments) EVALUATED INDEPENDENTLY Wide dynamic range E [MeV] = f ABS, U , S , L (r, t, PE) × PE Demanding σ (E)/E Correlation among f terms might become relevant (degeneracy) EXAMPLE ▶︎ ▶︎ ▶︎ WIN 2017 M. Grassi 11

Correlation Among Calibration Terms (Illustration) Deploy 1MeV calibration source at different positions (simulation) R 0 m 6 m 8.5 m N(pe) Events TRUTH 1000 1600 TRUTH 800 1500 600 1400 400 1300 200 1200 0 1100 Events 1000 1000 800 600 400 200 0 800 900 1000 1100 1200 1300 1400 1500 1600 N(pe) TRUTH : “Genuine” detector non-uniformity (geometry + LS attenuation) WIN 2017 M. Grassi 12

Correlation Among Calibration Terms (Illustration) Deploy 1MeV calibration source at different positions (simulation) R 0 m 6 m 8.5 m N(pe) Events TRUTH 1000 1600 TRUTH 800 RECO 1500 600 1400 400 1300 200 1200 0 1100 Events 1000 RECO 1000 800 Ratio 600 1 0.98 400 0.96 200 0.94 0.92 0 0 1 2 3 4 5 6 7 8 9 10 800 900 1000 1100 1200 1300 1400 1500 1600 Radius [m] N(pe) RECO : Introducing a 1% bias for each detected pe Residual charge non-linearity shows up as additional non-uniformity WIN 2017 M. Grassi 13

Correlation Outcome Use response map derived at 1MeV Events / 26 4 10 TRUTH Reconstruct 2.2 MeV gamma line 
 RECO from n captures on H 3 10 (uniformly distributed in the detector) 2 10 Actual resolution worse than intrinsic resolution 10 σ 2NON-STOCH is dominant 1 2000 2500 3000 N(pe) Experimental Challenge Understand the source of additional resolution (& distortion) How to break down systematic uncertainty budget? WIN 2017 M. Grassi 14

Double Calorimetry in JUNO (Large & Small PMTs) 18,000 PMTs (20” diameter) → Large-PMT system ( LPMT ) 25,000 PMTs (3” diameter) → Small-PMT system ( SPMT ) WIN 2017 M. Grassi 15

Double Calorimetry in JUNO Large PMTs (LPMT) Small PMTs (SPMT) CALIBRATION 75% photocoverage 3% photocoverage 1200 PE/MeV 50 PE/MeV PE = charge / gain PE = hits SPMT in photon counting regime across all dynamic range (energy & position) DYB WIN 2017 M. Grassi 16

Breakdown of the Non-Stochastic Resolution Term N(pe) Look at calibration data IDEAL 160 using SPMT RECO SPMT 150 140 130 120 110 100 Ratio 1 0.98 0.96 0.94 0.92 0 1 2 3 4 5 6 7 8 9 10 Radius [m] WIN 2017 M. Grassi 17

Breakdown of the Non-Stochastic Resolution Term N(pe) Look at calibration data IDEAL 160 using SPMT RECO SPMT 150 RECO LPMT Photon Counting Regime: 
 140 Negligible charge non-linearity 130 Compared to LPMT 120 110 SPMT provide a good reference 
 to understand LPMT response 100 Ratio 1 0.98 0.96 0.94 0.92 0 1 2 3 4 5 6 7 8 9 10 Radius [m] WIN 2017 M. Grassi 18

Breakdown of the Non-Stochastic Resolution Term N(pe) Look at calibration data IDEAL 160 using SPMT RECO SPMT 150 RECO LPMT Photon Counting Regime: 
 140 Negligible charge non-linearity 130 Compared to LPMT 120 110 SPMT provide a good reference 
 to understand LPMT response 100 Ratio 1 Ratio LPMT/SPMT “ ” 0.98 Calibration Data 0.96 Extra resolution due to 0.94 unaccounted charge non-linearity 0.92 0 1 2 3 4 5 6 7 8 9 10 Radius [m] SPMT: resolve otherwise unresolvable response degeneracy WIN 2017 M. Grassi 19

Summary & Conclusions 1.7 Normalized Response Three examples of 
 Uniformity (simulation) SPMT 1.6 1 MeV Double Calorimetry in action 1.5 2 MeV 1.4 LPMT 4 MeV 1.3 Detector uniformity map 
 1.2 valid at different energies 1.1 1 0 2 4 6 8 10 Detector Radius [m] Reliable measurement of detector DYB Linearity light non-linearity (LS quenching) Break correlation among calibration terms Redundancy: key ingredient to achieve high-precision calorimetry WIN 2017 M. Grassi 20