Neutron discrimination using waveform information. Yasuyuki - PowerPoint PPT Presentation

1 Neutron discrimination using waveform information. Yasuyuki Sugiyama � (Yamanaka Taku Lab.) Year-end annual report. � 2015 Dec. 25(Fri)

2 d T J-PARC KOTO Exp. O ν K calorimeter measures the s ν the K 0 L → π 0 νν deca • search for 
 CsI calorimeter KOTODetector (B.R.= 2.4x10 -11 @S.M.) 
 FB NCC MB CV CC04 CC05 CC06 BHCV of 576 pure CsI crystals. @J-PARC • Strategy π 0 → 2 γ measures K L • Detect by π 0 ν ν ν measuring 2 γ on CsI pure CsI Calorimeter. BCV LCV OEV CC03 HINEMOS • Require P t π 0 >0 
 500 [MeV/c] 450 with no other particle. :Signal 
 400 t P 0 350 � reconstructed Box • First physics run in 
 300 250 2013 May. 200 150 100 • More physics run in 2015 50 Spring, Autumn. 0 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 0 reconstructed vertexZ [mm] �

つくりなお 3 KOTO CsI Calorimeter measures π 0 ν • measure γ to reconstruct from K L decay. � pure CsI • Use 2716 Un-doped CsI crystals from KTeV � • Waveform is recorded by 125MHz ADC with Bessel Filter Pulse Height [ADC counts] 950 500mm 900 (27X 0 ) 850 800 750 700 650 600 550 2.5cm □ 500 5cm □ 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Time[Clock=8ns] CsI Crystal waveform recorded by ADC KOTO CsI Calorimeter

4 日金曜日 月 Opened the signal box 年 Expected Observed 1 0.36± 0.16 Neutron Background • Single neutron (1~2GeV) hits on CsI Calorimeter may cause Background events. � • considered to be dominant in the first physics run. � • Need a new method to distinguish neutrons from photons. Analysis Result for 2013 Physics Run • N obs =1, N exp =0.36±0.16 500 4 [MeV/c] 450 3.5 400 3 350 t P 2.5 CsI neutron 300 250 2 neutron 200 1.5 150 CsI 1 100 0.5 50 0 0 2000 3000 4000 5000 6000 31 Rec. z [mm] from K.Shiomi’s talk in JPS 2014 Autumn

5 CsI Calorimeter CsI Calorimeter How to reject Neutron Background? • Current Cut Methods use the transverse shower shape. � • Longitudinal characteristics or Timing structure of Hadron Shower may be new point of view to reject the background. � • These characteristics may affect pulse shape. Y Y Smaller Size 
 Larger Size 
 of Cluster of Cluster X X Gamma Neutron Z Shower exists 
 Shower only in Upstream also in Downstream

6 Fitting waveform with Asymmetric gaussian. • Pulse shape of CsI calorimeter has Gaussian-shape due to Bessel Filter in ADC module. � • Fit waveform with “asymmetric Gaussian” to get information about the change of pulse shape. � • Larger tail component will make “a” bigger. Pulse fitted by Asymmetric Gaussian Pulse Height [ADC counts] 950 900 Asymmetric Gaussian: 850 800 − ( t − t 0 ) 2 ✓ ◆ 750 A ( t ) = | A | exp , 2 σ ( t ) 2 700 650 σ ( t ) = σ 0 + a ( t − t 0 ) 600 550 500 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Time [Clock = 8ns]

7 Procedure • Check the difference of fit parameters from Gamma’s waveform � • Make Template of waveform parameters for Gamma events � • Compare acquired fit parameters ( p fit ) with Template( p exp (H) ). � fit 
 template • Calculate χ 2 value for γ and π 0 result ! 2 crystal p fit − p exp χ 2 1 ( H i ) X X i i γ NDF = 2 N Crystal σ p i ( H i ) p = σ 0 ,a i Precision H i : Pulseheight of i th crystal χ 2 χ 2 χ 2 NDF = 1 X π 0 γ i γ NDF γ i 2 γ i

8 Fit parameter Template • Make templates of fit parameter from K L ->3 π 0 events taken in 2013 Physics Run. � • Get Profile (Mean values) p exp (H) and Its RMS σ p (H). � • Use TSpline for interpolation between bins. h_asym_height_0 0.3 Asymmetric Parameter 350 0.25 300 0.2 σ p (H) 250 0.15 p exp (H) 200 0.1 150 0.05 100 0 -0.05 50 -0.1 0 1.5 2 2.5 3 3.5 4 Log10(Pulse Height)[ADC count]

9 0 in KL->3 π 0 events χ 2/NDF for each γ , π • Calculate χ 2/NDF for each γ clusters on the CsI Calorimeter and π 0 . � • γ ’s χ 2/NDF distributes around 1 and has small Energy dependence. � • π 0 ’s χ 2/NDF distributes around 1. asym chi2 for gamma vs GammaE Pi0_AsymChi2 {Iteration$<3} 5 h1 h1 4 10 � 50000 Entries Entries 631302 631302 /NDF for Threshold 4.5 Mean Mean 0.9621 0.9621 RMS RMS 0.3378 0.3378 4 40000 3 10 3.5 2 � 3 30000 2.5 2 10 Cut 2 20000 1.5 5% 
 10 1 10000 loss 0.5 0 0 1 0 500 1000 1500 2000 2500 3000 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 Gamma Energy [MeV] 2/NDF for � � X: χ 2 /NDF of π 0 small E dependence

10 Analysis for Neutron-rich data • Data taken with Inserting Aluminum target during 2015 April/May Run � • Only neutrons can arrive at Calorimeter from Aluminum target. � • Analyze events with 2 clusters on the CsI Calorimeter � • Apply all cuts used in 1st Physics Run (2013) Analysis 
 except for Transverse Cluster Shape Cut. All cut except for 
 Transverse Pulse Shape Cut CsI calorimeter h_PtZ_cut h_PtZ_cut 500 7 Pt[MeV/c] Entries Entries 1915 1915 FB NCC MB CV CC Mean x Mean x 4810 4810 450 Mean y Mean y 221.9 221.9 6 RMS x RMS x 438.2 438.2 400 RMS y RMS y 64.09 64.09 0 5 350 � 300 4 250 3 200 150 2 t=10mm 
 100 Al Target 1 50 BCV LCV OEV CC03 HINEMOS 0 0 3000 3500 4000 4500 5000 5500 6000 0 reconstructed Z vertex[mm] �

75% 11 0 with gamma template χ 2/NDF of Rec. π • neutron-like events in Al target data have big χ 2/NDF. � • K L ->3 π 0 data distribute around 1. � • if we reject events with χ 2/NDF>1.5, 
 75% of neutron-like events can be rejected with 5% loss for π 0 ->2 γ χ 2/NDF( γ ) Chi2/NDF for Pi0 with gamma template Chi2/NDF for Pi0 with gamma template h_chi2_pi_gamma 180 5 � Entries 5523 Entries Entries 3830 3830 2/NDF for 14 Mean x Mean x 511.3 511.3 4.5 160 Mean 1.974 Mean y Mean y 1.822 1.822 RMS x RMS x 268.6 268.6 RMS 0.897 4 12 140 RMS y RMS y 1.009 1.009 Al target with 
 3.5 � 120 10 3 all cut except 
 100 8 2.5 80 for ClusterShape 2 6 60 1.5 4 40 1 2 20 0.5 0 0 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 500 1000 1500 2000 2500 3000 χ 2/NDF Energy[MeV] � X:Energy[MeV]

12 Likelihood ratio signal MC (from downstream) ← background (from upstream) signal → backward π 0 Template of neutron-like events • Make template of fitting parameters using Al target run data only with Veto Cut. � • Use only a half of data to make template 
 and use another half of data to estimate performance. � • Calculate Chi2/NDF as same as the case with 6gamma template. � • Calculate Likelihood for 6gamma template and Z0Al template and determine which template is more closer to data. par . ch . γ ijk | s exp . � � � Gaussian( s obs . L n = ijkn , σ k ( e ij )) , i =0 , 1 j k =0 , 1

13 Pulse shape Difference in each crystal • Template difference become larger 
 in Pulse Height >~500count (~50MeV). � • χ 2/NDF difference in KL->3pi0 data 
 becomes larger for E deposit > 50MeV/crystal. χ 2/NDF(Al target template) 
 Template for CsI Ch.800 - χ 2/NDF(3pi0 template) asym.par GamClusCsi_AsymChi2_Al-GamClusCsi_AsymChi2:GamClusCsiE {GamClusCsiE>3} h1 h1 4 Entries Entries 653839 653839 Mean x Mean x 32.14 32.14 Mean y -0.009284 Mean y -0.009284 Al target 
 3 4 10 RMS x RMS x 60.63 60.63 RMS y RMS y 0.6891 0.6891 2 3pi0 3 10 1 0 2 10 -1 -2 10 -3 1 log10(PulseHeight) 0 50 100 150 200 250 300 350 400 450 500 Energy in CsI crystal[MeV]

14 Likelihood ratio backward π 0 signal MC (from downstream) ← background (from upstream) signal → Chi2/NDF for limited number of channel • Calculate chi2/NDF and Likelihood ratio only with information from channels with Edep >50MeV. � • If #(CsI with Edep>50MeV) is 0, use Chi2 and Likelihood of the CsI Crystal with Maximum Edep. fit 
 template result ! 2 crystal p fit − p exp χ 2 1 ( H i ) X X i i γ NDF = 2 N Crystal σ p i ( H i ) p = σ 0 ,a i Precision H i : Pulseheight of i th crystal χ 2 χ 2 χ 2 NDF = 1 X π 0 γ i γ NDF γ i 2 γ i par . ch . γ ijk | s exp . � � � Gaussian( s obs . L n = ijkn , σ k ( e ij )) , i =0 , 1 j k =0 , 1

15 Al target data broad narrow narrow broad KL->3π 0 data χ 2/NDF distribution with 50MeV threshold/Crystal Chi2/NDF for Pi0 with gamma template Chi2/NDF for Pi0 with gamma template Chi2/NDF for Pi0 with gamma template Chi2/NDF for Pi0 with gamma template 3 10 × h_chi2_pi_gamma h_chi2_pi_gamma 80 Entries 5523 Entries 1.158191e+07 Mean 2.49 500 Mean 0.8097 70 RMS 1.217 RMS 0.5906 60 400 3 π 0 
 50 300 40 template 30 200 20 100 10 0 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Chi2/NDF for Pi0 with Z0Al template Chi2/NDF for Pi0 with Z0Al template Chi2/NDF for Pi0 with Z0Al template Chi2/NDF for Pi0 with Z0Al template 3 10 h_chi2_pi_al × h_chi2_pi_al 350 Entries 1.158191e+07 Entries 5523 Mean 1.287 250 300 Mean 0.9017 RMS 0.8414 RMS 0.5917 250 200 Al target 
 200 150 template 150 100 100 50 50 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5