Pion multiplicity study update Seb Jones Department of Physics & Astronomy University College London December 16, 2019 S. Jones (UCL) DUNE LBL December 16, 2019 1 16
Recap Aim is to use gas TPC samples (seperated by pion multiplicty) along with NuWro fake data to see if the HPgTPC can be used to determine issues with our interaction model Gas TPC samples use pseudo-reconstruction based upon GEANT4 energy deposits NuWro fake data generated through reweighting process in CAFAna S. Jones (UCL) DUNE LBL December 16, 2019 2 16
Updates to samples / reconstruction Reconstructed vs. true neutrino energy for CC1 π Previously, saw some kind of 5 E ν, reco Events 4.5 bias within the gas true vs reco 4 10 3 distributions – suggested that 3.5 this may be caused by lack of 3 10 2 2.5 fiducial cut 2 Now a fiducial cut for the 1.5 10 1 samples – event vertex must 0.5 have r < 200 cm and 0 1 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 E ν, true / GeV | x | < 200 cm S. Jones (UCL) DUNE LBL December 16, 2019 3 16
Updates to samples / reconstruction Reconstructed vs. true neutrino energy for CC1 π Turns out that FV is not the 5 E ν, reco Events 4.5 cause of this 4 10 3 Peak in true - reco distribution 3.5 is at π ± mass 3 10 2 2.5 Caused by high energy pions 2 1.5 ( P reco > 1 . 5 GeV/c) being 10 1 mis-ID’d as protons 0.5 0 1 Leads to m π mistakenly being 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 E ν, true / GeV added back in S. Jones (UCL) DUNE LBL December 16, 2019 4 16
Updates to samples / reconstruction Reconstructed vs true π multiplicity in HPgTPC 4.5 10 5 N π, reco Additonally, now have samples 4 defined by reconstructed pion 3.5 10 4 3 multiplicity rather than true 2.5 10 3 Reconstruction requires 2 1.5 minimum track length 10 2 1 For π 0 s , require that decay 0.5 10 0 photons do not overlap with -0.5 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 one another 1 N π, true S. Jones (UCL) DUNE LBL December 16, 2019 5 16
Fitting efforts - FD only χ Oscillation parameters, FD only. 2 = 6.404 Value 3.5 3 2.5 2 1.5 1 0.5 0 dmsq32NHscaled th13 ssth23 delta(pi) rho Parameter Reproducing an FD only fit with full systematics leads to biased oscillation parameters but a low χ 2 – hard to detect that we have the wrong answer Red is true values of oscillation parameters while black shows the post-fit constraints S. Jones (UCL) DUNE LBL December 16, 2019 6 16
Fitting efforts - FD only σ -1.5 -0.5 difference introduced by the shift to NuWro Together with oscillation parameters they’re able to ‘absorb’ the Systematics remain close to their nominal values ( σ = 0) S. Jones (UCL) 0.5 1.5 -1 0 1 flux Nov17 0 flux Nov17 1 flux Nov17 2 flux Nov17 3 flux Nov17 4 flux Nov17 5 flux Nov17 6 flux Nov17 7 flux Nov17 8 flux Nov17 9 flux Nov17 10 flux Nov17 11 flux Nov17 12 flux Nov17 13 flux Nov17 14 flux Nov17 15 flux Nov17 16 flux Nov17 17 flux Nov17 18 flux Nov17 19 EnergyScaleFD UncorrFDTotSqrt UncorrFDTotInvSqrt UncorrFDHadSqrt UncorrFDHadInvSqrt UncorrFDMuSqrt UncorrFDMuInvSqrt UncorrFDNSqrt UncorrFDNInvSqrt UncorrFDEMSqrt UncorrFDEMInvSqrt EScaleMuLArFD ChargedHadUncorrFD NUncorrFD EMUncorrFD MuonResFD EMResFD Systematic constraints (FD only) ChargedHadResFD NResFD FDRecoNumuSyst FDRecoNueSyst FVNumuFD DUNE LBL FVNueFD RecoNCSyst FVNumuND MaCCQE VecFFCCQEshape MaCCRES MvCCRES MaNCRES MvNCRES Theta Delta2Npi AhtBY BhtBY CV1uBY CV2uBY FrCEx pi FrElas pi FrInel pi FrAbs pi FrPiProd pi FrCEx N FrElas N FrInel N FrAbs N FrPiProd N CCQEPauliSupViaKF E2p2h A nu E2p2h B nu E2p2h A nubar E2p2h B nubar NR nu n CC 2Pi NR nu n CC 3Pi NR nu p CC 2Pi NR nu p CC 3Pi NR nu np CC 1Pi NR nu n NC 1Pi NR nu n NC 2Pi NR nu n NC 3Pi NR nu p NC 1Pi December 16, 2019 NR nu p NC 2Pi NR nu p NC 3Pi NR nubar n CC 1Pi NR nubar n CC 2Pi NR nubar n CC 3Pi NR nubar p CC 1Pi NR nubar p CC 2Pi NR nubar p CC 3Pi NR nubar n NC 1Pi NR nubar n NC 2Pi NR nubar n NC 3Pi NR nubar p NC 1Pi NR nubar p NC 2Pi Parameter NR nubar p NC 3Pi BeRPA A BeRPA B BeRPA D C12ToAr40 2p2hScaling nu C12ToAr40 2p2hScaling nubar nuenuebar xsec ratio nuenumu xsec ratio 7 16
LAr ND samples ND LAr ν µ FHC 0 < y reco < 0 . 1 ND LAr ν µ FHC 0 . 1 < y reco < 0 . 2 × 10 3 × 10 3 Events Events ND LAr ν µ FHC 0 < y reco < 0 . 1 ND LAr ν µ FHC 0 . 1 < y reco < 0 . 2 ND LAr ν µ FHC (NuWro) 0 < y reco < 0 . 1 2000 2500 ND LAr ν µ FHC (NuWro) 0 . 1 < y reco < 0 . 2 2000 1500 1500 1000 1000 500 500 0 2 4 6 8 10 0 2 4 6 8 10 E ν, reco ( GeV ) E ν, reco ( GeV ) Shown here are the 1D projections of the ND analysis bins with both the nominal sample and with the NuWro reweights At low values of y reco relatively small discrepancies between GENIE and NuWro S. Jones (UCL) DUNE LBL December 16, 2019 8 16
LAr ND samples ND LAr ν µ FHC 0 . 2 < y reco < 0 . 3 ND LAr ν µ FHC 0 . 3 < y reco < 0 . 4 × 10 3 × 10 3 Events Events 1500 1200 ND LAr ν µ FHC 0 . 2 < y reco < 0 . 3 ND LAr ν µ FHC 0 . 3 < y reco < 0 . 4 ND LAr ν µ FHC (NuWro) 0 . 2 < y reco < 0 . 3 ND LAr ν µ FHC (NuWro) 0 . 3 < y reco < 0 . 4 1000 1000 800 600 500 400 200 0 2 4 6 8 10 0 2 4 6 8 10 E ν, reco ( GeV ) E ν, reco ( GeV ) Begin to see greater discrepancies as we move to more inelastic interactions S. Jones (UCL) DUNE LBL December 16, 2019 9 16
LAr ND samples ND LAr ν µ FHC 0 . 4 < y reco < 0 . 6 ND LAr ν µ FHC 0 . 6 < y reco < 1 × 10 3 × 10 3 1400 Events Events 3000 1200 ND LAr ν µ FHC 0 . 6 < y reco < 1 ND LAr ν µ FHC 0 . 4 < y reco < 0 . 6 1000 ND LAr ν µ FHC (NuWro) 0 . 6 < y reco < 1 ND LAr ν µ FHC (NuWro) 0 . 4 < y reco < 0 . 6 2000 800 600 1000 400 200 0 0 0 2 4 6 8 10 0 2 4 6 8 10 E ν, reco ( GeV ) E ν, reco ( GeV ) Begin to see greater discrepancies as we move to more inelastic interactions S. Jones (UCL) DUNE LBL December 16, 2019 10 16
Comparison with exclusive samples in HPgTPC – CC1 π Lead π energy, CC 1 π × 10 3 Events 300 Lead π energy, CC 1 π 200 Lead π energy, CC 1 π (NuWro) 100 0 0 0.5 1 1.5 2 2.5 3 E π, lead ( GeV ) For events with a low pion multiplicity, leading π energy distribution looks quite similar across generators S. Jones (UCL) DUNE LBL December 16, 2019 11 16
Comparison with exclusive samples in HPgTPC – CC2 π Lead π energy, CC 2 π Sub-leading π energy, CC 2 π × 10 3 250 Events 60000 Events Lead π energy, CC 2 π 200 Lead π energy, CC 2 π (NuWro) Sub-leading π energy, CC 2 π 40000 150 Sub-leading π energy, CC 2 π (NuWro) 100 20000 50 0 0 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 E π, sub − lead ( GeV ) E π, lead ( GeV ) For higher 2 π events we see a significant divergence between the GENIE and NuWro samples S. Jones (UCL) DUNE LBL December 16, 2019 12 16
Comparison with exclusive samples in HPgTPC – CC3 π Lead π energy, CC 3 π Sub-leading π energy, CC 3 π 80000 25000 Events Events Lead π energy, CC 3 π 20000 60000 Lead π energy, CC 3 π (NuWro) Sub-leading π energy, CC 3 π 15000 Sub − leading π energy , CC 3 π ( NuWro ) 40000 10000 20000 5000 0 0 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 E π, sub − lead ( GeV ) E π, lead ( GeV ) Similarly, for CC 3 π events there is a significant divergence in the distributions of both the leading and sub-leading pion energy S. Jones (UCL) DUNE LBL December 16, 2019 13 16
Comparison with exclusive samples in HPgTPC – CC > 3 π Lead π energy, CC > 3 π Sub-leading π energy, CC > 3 π 30000 14000 Events Events Sub-leading π energy, CC > 3 π 12000 25000 Sub-leading π energy, CC > 3 π (NuWro) 10000 20000 8000 15000 6000 10000 4000 Lead π energy, CC > 3 π 5000 2000 Lead π energy, CC > 3 π (NuWro) 0 0 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 E π, sub − lead ( GeV ) E π, lead ( GeV ) For events at very high multiplicity the divergences appear to once again not be so obvious However, one can see that by using just some of the exclusive HPgTPC samples it would quickly become obvious that there was an issue with out interaction model if our data was like NuWro S. Jones (UCL) DUNE LBL December 16, 2019 14 16
Backup Backup S. Jones (UCL) DUNE LBL December 16, 2019 15 16
Recommend
More recommend
