Nue Cross Section Measurement Mark Hartz Kavli IPMU (WPI), - PowerPoint PPT Presentation

Nue Cross Section Measurement Mark Hartz Kavli IPMU (WPI), University of Tokyo and TRIUMF


 
 
 
 Goals • Using the intrinsic electron (anti)neutrino component, measure the ratios: • The purity improves with larger off-axis angle: 
 ν e Flux 
 ν μ Flux O ff -axis angle (º) Ratio ν e / ν μ 0.3-0.9 GeV 0.3-5.0 GeV 2.5 1.24E+15 2.46E+17 0.507% 3.0 1.14E+15 1.90E+17 0.600% 3.5 1.00E+15 1.47E+17 0.679% 4.0 8.65E+14 1.14E+17 0.760% • We can use linear combinations to match the fluxes: 6 × 10 Graph Graph (Arb. Norm.) Entries Entries 0 0 8000 Mean 0 Mean 0 Fitted Coefficient RMS 0 RMS 0 0.09 PRISM (2.5-4.0 ) ν ν ° e 7000 0.08 ν PRISM ν Linear Combo. 6000 µ 0.07 5000 0.06 σ × 4000 0.05 Φ 0.04 3000 0.03 2000 0.02 1000 0.01 0 0 0.5 1 1.5 2 2.5 3 1 1.5 2 2.5 3 3.5 4 E (GeV) Off-axis Angle (degrees) ν 2 NuPRISM Reco

Showing Today • Updated electron neutrino selection in neutrino mode • Re-optimized ToWall, DWall and PID likelihood ratio cuts • Re-calculate systematic and statistical error estimates • Discussion on how to proceed 3 NuPRISM Reco

Monte Carlo • T2K 13bv3.2 neutrino flux model • NEUT 5.1.4.2 with RFG, MA=1.2 GeV • WCSim simulation of the NuPRISM inner detector • ID radius = 3 m • ID height = 10 m • 8 inch diameter PMTs with 40% photo-coverage • Center of ID positioned at 2.5, 2.92, 3.34 and 3.76 degrees off-axis • 2.28e21 POT simulated at each position • fiTQun single ring and pi0 reconstruction of simulated events 4 NuPRISM Reco

Event Pre-selection • Before optimizing fiducial volume and particle ID cuts • Cuts on true quantities: • No OD information, so cut muons with mom (MeV/c)>2.0*ToWall (cm) • No multi-ring reconstruction, so cut events with >1 ring with “true” visible energy above 30 MeV • Reconstructed quantities: • fq1rpcflg[0][1]==0 (contained event) • fqnse==0 (no decay electron) • fq1rmom[0][1]>100. (momentum>100 MeV/c) • Preliminary PID cuts for fiducial volume optimization for Yoshida-san (next slide) 5 NuPRISM Reco

PID Pre-selection T. Yoshida p<375: L ratio >-30+0.3*p mass<133 && 
 p>375: L ratio >270-0.34*p L ratio <120-mass*0.9 6 NuPRISM Reco

Procedure • Apply the pre-selection on the previous two slides • Optimize the dWall and toWall cuts using S/sqrt(S+B+ σ syst2 )* • Signal = CC interactions of nu_e • BGND = CC interactions of nu_mu + NC interactions • Remove the PID cuts and apply the optimized dWall and toWall cuts • Redo the optimization of the PID cuts • Redo the optimization of the fiducial volume cuts • Evaluate the signal statistics and purity with the optimized cuts *Formula includes background systematic error in the denominator 7 NuPRISM Reco

dWall/toWall Distributions -CC ν -CC ν NC e µ 300 300 300 toWall (cm) Events 70 toWall (cm) Events toWall (cm) Events 40 60 250 60 250 250 35 50 30 50 200 200 200 40 25 40 150 150 150 20 30 30 15 100 100 100 20 20 10 50 50 50 10 10 5 0 0 0 0 0 0 0 50 100 150 200 250 300 0 50 100 150 200 250 300 0 50 100 150 200 250 300 dWall (cm) dWall (cm) dWall (cm) • Events with toWall<100 cm are generally not reconstructed as contained events • NC background events are at low dWall (second photon has small toWall?) • CC-numu background events are at low toWall (poor sampling of ring) 8 NuPRISM Reco

dWall/toWall Cut Optimization Cut Optimization 300 S+B toWall Cut (cm) 40 250 S/ 35 200 30 25 150 20 100 15 50 10 5 0 0 50 100 150 200 250 300 dWall Cut (cm) Optimized on S/sqrt(S+B+(0.05*B NC ) 2 +(0.05*B CC ) 2 ) • Most optimal point in cut space is dWall>135 cm and toWall>280 cm • Old optimization was dWall>150 cm and toWall>260 cm • Will re-optimize after optimizing the PID cuts • 9 NuPRISM Reco

L e /L mu Cut Optimization 1000 24 S+B Cut 22 ratio S/ 20 500 L 18 16 0 14 12 10 -500 Electrons 8 Muons 6 -1000 0 200 400 600 800 1000 p (MeV/c) e Apply all cuts except for L e /L mu cut • Optimized on S/sqrt(S+B cc +(0.05*B CC ) 2 ) for each bin in electron momentum • Cut lines are: • p<325: L ratio >80+0.277*p p>325: L ratio >252-0.252*p 10 NuPRISM Reco

L pi0 /L e Cut Optimization 1000 Electrons S+B Cut 40 Pi0 ratio 35 S/ L 30 25 500 20 15 10 5 0 0 50 100 150 200 2 m (MeV/c ) 0 π Apply all cuts except for L e /L mu cut • Optimized on S/sqrt(S+B NC +(0.05*B NC ) 2 ) for each bin in electron momentum • Cut lines are: • mass<90 && L ratio <130-0.95*mass 11 NuPRISM Reco

dWall/toWall Cut Re-optimization Cut Optimization 300 S+B toWall Cut (cm) 50 45 250 S/ 40 200 35 30 150 25 100 20 15 50 10 5 0 0 50 100 150 200 250 300 dWall Cut (cm) Optimized on S/sqrt(S+B+(0.05*B NC ) 2 +(0.05*B CC ) 2 ) • Change cut slightly dWall>135 cm and toWall>275 cm • 12 NuPRISM Reco

Background and Signal Error Estimate • The NCpi0 and CC-numu backgrounds should be constrained by high statistics samples in NuPRISM • Assume we can achieve a 5% uncertainty on the background prediction • The NCgamma background cannot be directly constrained. Assume a 50% error coming mainly from the cross section and flux models • Assume we can achieve a 1% error on the modeling of the signal efficiency 13 NuPRISM Reco

POT Weighted Signal (Old Cuts/MC Stats) • Weighting to 1.5e21 neutrino mode POT for each off-axis position between 2.5 and 4.0 degrees 1-Ring e Candidates Purity for E rec <1.2 GeV = 73% Events/(200 MeV) Nue Signal for E rec <1.2 GeV =3184 -CC ν e 1500 0 NC π NC γ Expect improvements: -CC ν 1000 µ - OD veto on exiting particles - More optimization of PMT 
 500 granularity - Larger ID size? 0 0 500 1000 1500 2000 E (MeV) rec 14 NuPRISM Reco

Total Error Size (Old Cuts/MC Stats) 200 MeV bins 1 GeV bins 0.25 0.25 Fractional Error Fractional Error Total Total Statistical (S-B) 0.2 0.2 Statistical (S-B) Background Systematics Background Systematics Signal Efficiency 0.15 0.15 Signal Efficiency Flux ( / ) ν ν µ e Flux ( / ) ν ν e µ 0.1 0.1 0.05 0.05 0 0 0 500 1000 1500 2000 0 500 1000 1500 2000 E (MeV) rec E (MeV) rec • The estimated error for 200 MeV wide bins is shown on the left • For an estimate of the error on the overall rate, I divided into two bins of 1 GeV (right) • Current estimate is ~5% error in the sub-GeV region 15 NuPRISM Reco

POT Weighted Signal (New Cuts/MC Stats) • Weighting to 1.5e21 neutrino mode POT for each off-axis position between 2.5 and 4.0 degrees 1-Ring e Candidates Events/(200 MeV) 1500 Purity for E rec <1.2 GeV = 71(73)% -CC ν e Nue Signal for E rec <1.2 GeV =3501(3184) 0 NC π NC γ 1000 -CC ν µ 500 0 0 500 1000 1500 2000 20 inch PMT Results E (MeV) rec 16 NuPRISM Reco

Old vs New Total Error Size New Old 0.25 0.25 Fractional Error Fractional Error Total Total 0.2 0.2 Statistical (S-B) Statistical (S-B) Background Systematics 0.15 Background Systematics 0.15 Signal Efficiency Flux ( / ) ν ν Signal Efficiency e µ 0.1 Flux ( / ) 0.1 ν ν e µ 0.05 0.05 0 0 0 500 1000 1500 2000 0 500 1000 1500 2000 E (MeV) E (MeV) rec rec • The total uncertainty below 1 GeV is only slightly reduced with cut optimization on large MC statistics • Configuration changes are likely needed to get any significant improvments 17 NuPRISM Reco

What’s Next • For the T2K phase II studies, run some 2.5-4.0 degree MC with 4 m ID radius • For the summer PAC meeting, run MC for the full off-axis angle range with the following configurations: • 8 inch PMT, 3 m radius ID (2.5-4.0 degrees already exist) • 8 inch PMT, 4 m radius ID • 3 inch PMT, 4 m radius ID • 3 inch PMT, 3 m radius ID • Do full error propagation once MC is available for the full off-axis angle range • Antineutrino analysis if there is time 18 NuPRISM Reco

Extra Slides 19 NuPRISM Reco

Flux Error Estimate Flux error estimates were previously calculated by selecting Nue candidates based on truth information Since the error will be applied after the background subtraction, we can apply the same error to our reconstructed events / Flux Error σ σ ν ν e µ 0.15 Fractional Error Nominal Flux Error Saved as covariance matrix 1/2 Had. Int. Error 0.1 for application to the 
 signal after background 
 subtraction 0.05 0 0 0.5 1 1.5 2 E (GeV) ν 20 NuPRISM Reco

Cut Optimization • Start with ln(L e /L μ ) cut set at >150 • Just eyeballed. Numu background is small compared to NC background. • Optimize DWall and ToWall cuts on the NCpi0 background and Signal with events having pi0mass<60 MeV/c2 • Optimize cut line in ln(L π 0 /L e ) plane 21 NuPRISM Reco

DWall/ToWall Optimization S/ S+B sig sig Entries Entries 117800 117800 300 Mean x Mean x nan nan Mean y Mean y nan nan RMS x RMS x nan nan DWall (cm) RMS y RMS y nan nan 2.5 250 2 200 1.5 150 100 1 50 0.5 0 0 50 100 150 200 250 300 ToWall (cm) • Set cut at DWall>150 cm, ToWall>260 cm • May be able to improve the optimization with more stats (some of the structure is probably from statistical fluctuations) 22 NuPRISM Reco