Cross-section measurements at the NOvA near detector Linda - PowerPoint PPT Presentation

Sandbox Studio, Chicago Cross-section measurements at the NOvA near detector Linda Cremonesi for the NOvA Collaboration

Outline • Overview of the NOvA beam, detector and simulation • Inclusive measurements • Pion production measurements • NC coherent π 0 results • Summary and outlook L. Cremonesi “Cross-sections at NOvA ND” 2

Beam at NOvA 10 • NOvA detectors are 14 mrad off the NuMI beam axis. 8 • narrow 2-GeV spectrum = 0 mrad θ [GeV] 6 = 14.6 mrad (NO A) θ ν • small flux shape uncertainties 
 ν E 4 (hadron production uncertainties are mostly normalisation effect) 2 • 95% pure ν μ beam 0 0 10 20 30 40 E [GeV] π NO A Simulation ν CC / 6E20 POT / kton / 0.1 GeV FLUKA11 On-Axis On-Axis 25 14.6 mrad Off-Axis (NO A) ν 14.6 mrad Off-Axis (NO A) ν 20 15 10 5 µ ν 6 10 0 0 5 10 15 E [GeV] ν L. Cremonesi “Cross-sections at NOvA ND” 3

Beam at NOvA T2K + MicroBooNE + NOvA + MINERvA NOvA Simulation T2K (Fe) PRD 90, 052010 (2014) CDHS, ZP C35, 443 (1987) 6 10 × GGM-SPS, PL 104B, 235 (1981) T2K (CH) PRD 90, 052010 (2014) / GeV) 1.6 GGM-PS, PL 84B (1979) T2K (C), PRD 87, 092003 (2013) 0.25 IHEP-ITEP, SJNP 30, 527 (1979) Events/(8.09E+20 POT) ArgoNeuT PRD 89, 112003 (2014) CC Res ν IHEP-JINR, ZP C70, 39 (1996) ArgoNeuT, PRL 108, 161802 (2012) µ 1.4 MINOS, PRD 81, 072002 (2010) CC DIS ν ANL, PRD 19, 2521 (1979) NOMAD, PLB 660, 19 (2008) µ 2 BEBC, ZP C2, 187 (1979) CC QE ν cm NuTeV, PRD 74, 012008 (2006) 0.2 1.2 µ BNL, PRD 25, 617 (1982) SciBooNE, PRD 83, 012005 (2011) CC MEC ν CCFR (1997 Seligman Thesis) SKAT, PL 81B, 255 (1979) µ -38 CC Coh ν 1 µ (10 0.15 NC - N X ν → µ 0.8 µ ν / E 0.1 0.6 CC σ 0.4 0.05 + N X ν → µ 0.2 µ 0 0 0 1 2 3 4 5 1 10 100 00 150 200 250 300 350 Reconstructed Neutrino Energy(GeV) E (GeV) ν • NOvA is sensitive to many different nu+A interaction channels. • Cross sections in NOvA’s energy range suffer from high uncertainties in neutrinos and no measurements below 3 GeV for antineutrinos. • Nice overlap with currently running experiments, as well as future experiments in the US. L. Cremonesi “Cross-sections at NOvA ND” 4

The NOvA Near Detector Wavelength- shifting fibres routed to a single cell on an Avalanche Photodiode Preliminary NO NO A ND Data A ND Data ν ν Beam ~1 hour 
 20 of data! • Tracking calorimeter hits / 50ns 15 • 77% hydrocarbon by mass, 16% chlorine, 6% TiO 2 10 3 • Muon catcher (steel + NOvA cells) at downstream 10 end to range out ~2GeV muons. 10 μ s 
 5 • O(10) ns single hit timing resolution. NuMI pulse 0 215 220 225 230 Hit time ( s) L. Cremonesi “Cross-sections at NOvA ND” 5 µ

Simulation Beamline+Flux: G4NuMI nu interactions & 
 FSI modelling: GENIE Detector response: 
 GEANT4 Readout electronics & DAQ: Custom simulation routines L. Cremonesi “Cross-sections at NOvA ND” 6

ν μ CC inclusive • σ (E) and flux-averaged double differential cross section in muon kinematics variables • σ (E) measurements are kinematically restricted this phase space due to limited statistics and low efficiency NOvA Simulation 2.5 Reco Muon Kinetic Energy (GeV) 4 10 Events/(8.09E+20 POT) 2 3 10 1.5 2 10 1 10 0.5 1 0.5 0.6 0.7 0.8 0.9 1 Reco Cos θ µ L. Cremonesi “Cross-sections at NOvA ND” 7

ν μ CC inclusive: Reco + Selection Muon Catcher Top View Beam Side Muon View Catcher • Hits associated in time and space are used to form a candidate interaction. • Vertices, tracks and showers are reconstructed from these hits. L. Cremonesi “Cross-sections at NOvA ND” 8

ν μ CC inclusive: Reco + Selection Muon Catcher Top View Beam Side Muon View Catcher • Solid box is Fiducial Volume • Containment uses nearest projected distance to an edge (dashed box is rough approximation). • Events with hadronic activity in or near the muon catcher are excluded L. Cremonesi “Cross-sections at NOvA ND” 9

ν μ CC inclusive: PID • Use a kNN to separate signal and background tracks based on 4 NOvA Preliminary variables: Simulated selected events 80 Simulated background Data • track length Shape-only 1- syst. range σ 20 ND area norm., 3.72 × 10 POT 60 • dE/dx along track Events • scattering along track 40 3 10 • fraction of track planes w/ single 
 20 particle dE/dx 0 0 5 10 15 NOvA Preliminary Length of primary track (m) 6 7 Simulated selected events 10 10 × Simulated background Simulated Selected Events Data 0.6 Simulated Background Shape-only 1- syst. range σ Data 6 10 20 Full 1- σ syst. range ND area norm., 3.72 10 POT × 20 ND POT norm., 3.72 10 POT × Events 0.4 Events 5 10 0.2 4 10 3 10 0.0 − 3 − 2 − 1 0 1 dE/dx Log-likelihood 0 0.2 0.4 0.6 0.8 1 Muon ID L. Cremonesi “Cross-sections at NOvA ND” 10

ν μ CC inc: efficiency and background NOvA Simulation NOvA Simulation 0.3 0.3 0.25 0.8 -like 0.2 0.2 µ µ Efficiency ν ν -like / 0.6 / Purity µ 0.15 ν Anti- µ ν 0.4 Non- 0.1 0.1 0.05 0.2 0 0 1 2 3 4 Reconstructed Neutrino Energy (GeV) 0 1 2 3 4 True Neutrino Energy (GeV) • Selection efficiency is dominated by containment cut. • Backgrounds are small near the 2 GeV peak, larger in the tails of the spectrum. • Uncertainties are at the level of a few %. L. Cremonesi “Cross-sections at NOvA ND” 11

ν μ CC inc: Summary of Uncertainties NOvA Simulation 2.5 0.1 Reco Muon Kinetic Energy (GeV) 0.08 Statistical Uncertainty 2 0.06 1.5 0.04 1 0.02 0.5 0 0.5 0.6 0.7 0.8 0.9 1 Reco Cos θ µ • Statistical uncertainties are typically <2%. • Systematics are still being assessed, but we expect for the differential measurement ~10% highly correlated (normalisation) flux uncertainties, and all the other systematics combined to be 5-8% • σ (E) measurement systematics will be similar, although systematics from energy scale uncertainties will be larger on the rising and falling edges of the spectrum. L. Cremonesi “Cross-sections at NOvA ND” 12

ν e CC inclusive: Overview σ (E) and flux-averaged single • differential cross-section as a function of the electron kinematics for energies between 1 and 3 GeV. • Challenging because (by design) there are ~1% of ν e. • We have shown preliminary results on this channel in the past. That analysis is now superseded with a different event identification developed in the oscillation analysis. L. Cremonesi “Cross-sections at NOvA ND” 13

ν e CC inclusive: CVN • NOvA uses a Convolutional Neural Network (CNN) where a series of image filters are applied to hit map images to extract features associated with an interaction • Not limited to features chosen a priori • CNNs extract features of varying complexity and learn correlations • 30% effective increase in exposure • First CNN implementation on a HEP result • Inputs are image representations of our events where “RGB” calibrated hit information • Does not require previous reconstruction: no reconstruction inefficiencies • Inspired by animal visual cortex • Kernels, Filters or Convolutional Layers extract features of varying levels of complexity L. Cremonesi “Cross-sections at NOvA ND” 14

ν e CC inclusive: CVN • A convolutional visual network (CVN) is then trained on these filters. • 30% effective increase in exposure in the Far Detector for the oscillation analysis L. Cremonesi “Cross-sections at NOvA ND” 15

ν e CC incl: PID Fraction (%) Interaction 51 . 2 ν e CC ¯ 4 . 4 ν e CC 21 . 1 ν µ CC 23 . 0 NC 0 . 18 Other • Currently using a cut (CVN > 0.85) that optimises the FoM of S/ √ ( S+B) • Backgrounds are significant, and we are investigating potential driven data constraints. L. Cremonesi “Cross-sections at NOvA ND” 16

ν e CC incl: Efficiency and Purity • Xsec, FSI and calibration systematics included in error bands. • Uncertainties on efficiency and backgrounds is between 5-10%. • Data-driven constraints on the efficiency and backgrounds are being explored. L. Cremonesi “Cross-sections at NOvA ND” 17

ν μ CC π 0 • Signal: ν μ -CC events with at least one primary π 0 in the final state. • π 0 production vital for ν e appearance searches • Flux-averaged differential cross sections in final state kinematics L. Cremonesi “Cross-sections at NOvA ND” 18

ν μ CC π 0 : PID Use non-muon shower variables to form a π 0 identifier: • Bragg peak identifier. • Energy per hit. • Photon gap from vertex. • Number of missing planes. Fit signal and background MC to data in each kinematic bin. L. Cremonesi “Cross-sections at NOvA ND” 19

ν μ CC π 0 • Signal is dominantly RES (38.3%) and DIS (61.3%). • Uncertainty (~15%) is systematic dominant. • Plan to report flux-averaged differential cross section in final state muon and pion kinematics. • Α t final stage of internal review. Results soon! L. Cremonesi “Cross-sections at NOvA ND” 20