Preliminary Results for CCQE Scattering with the MINOS Near Detector

of 45/45
reliminary Results for CCQE Scattering with the MINOS Near Detector Mark Dorman, UCL On behalf of the MINOS Collaboration UINT 09, CC/NC QE Scattering, 19 th May 2009 1
  • date post

    05-Feb-2016
  • Category

    Documents

  • view

    44
  • download

    0

Embed Size (px)

description

Preliminary Results for CCQE Scattering with the MINOS Near Detector. Mark Dorman, UCL On behalf of the MINOS Collaboration. 1. NUINT 09, CC/NC QE Scattering, 19 th May 2009. Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook. - PowerPoint PPT Presentation

Transcript of Preliminary Results for CCQE Scattering with the MINOS Near Detector

  • Preliminary Results for CCQEScattering with the MINOSNear DetectorMark Dorman, UCL

    On behalf of the MINOS Collaboration

    NUINT 09, CC/NC QE Scattering, 19th May 20091

  • 2Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Overview Introduction to the NuMI beam and MINOS Near Detector

    MINOS MC simulation and -CC QE scattering theory

    -CC event selection and sample characteristics

    -CC QE event selection and sample characteristics

    Axial-vector mass fit technique and parameters

    Fit results and outlook for the analysisNUINT 09, CC/NC QE Scattering, 19th May 2009

  • 3 The NuMI BeamNUINT 09, CC/NC QE Scattering, 19th May 2009 The distance between the target and first horn can be changed to give a variable beam energy. In the low energy configuration the beam comprises:Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 4 The Near Detector (ND)NUINT 09, CC/NC QE Scattering, 19th May 2009Steel PlanePMT DarkBoxMulti-AnodePMTWLS FibresStrips on adjacent planesare mounted orthogonal to allow for 3D event reconstruction.Segmentation: 5.94cm longitudinal 4.1cm transverseScint.: 1cm thick, 4.1cm wide beam 1km from target

    0.98 kton

    282 steel planes

    B = 1.2 T High rates so ND is instrumented with deadtime-less QIE electronics.Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 5 Event GenerationNUINT 09, CC/NC QE Scattering, 19th May 2009 NEUGEN3 event generator

    QE: BBA05 vector form factors, dipole FA with MAQE = 0.99 GeV

    Resonances: Rein-Seghal model

    DIS: Bodek-Yang modified LO model, tuned to e and n data in resonance/DIS overlap region

    Coherent production

    Nuclear model: relativistic Fermi Gas (Pauli-blocking of QE)

    FSIs: nucleons and pions

    Hadronization: KNO-scaling -> PYTHIA/JETSET as W increasesIntroduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 6 The Axial-Vector MassNUINT 09, CC/NC QE Scattering, 19th May 2009 The uncertainty in the QE cross section is dominated by the uncertainty in the axial-vector form factor. FA is written using a dipole form as:Dipole FormKnown from neutron beta-decay experiments.

    The quasi-elastic axial-vector mass. Can measure MAQE by looking at the Q2 distribution for a QE-like event sample. Both the shape and rate of the distributions are affected by changes to MAQE:1 GeV neutrinoson free nucleonIntroduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 7 Selecting CC EventsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We first remove the majority of NC events by requiring a reconstructed track and then further enrich the sample using a nearest neighbour technique (kNN).

    The kNN combines variables that differentiate between muon tracks and the pion or proton tracks that can be reconstructed in NC events.NDGood agreement between data and MC.NDCC Selection EfficiencyNC Contamination

  • 8 Energy Spectra and Flux TuningNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Different beam configurations sample different regions in parent hadron xf and pT.

    We fit the data and tune our FLUKA hadron production model.

    The fits also include nuisance parameters for beam optics effects, NC cross section and ND energy scales. This flux-tuning procedure has been very successful and all of the MC distributions shown in my talk will use the tuned hadron production model.

  • 9 CC Sample Kinematic CoverageNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook These sub-samples correspond roughly to kinematic regions where the event generator is using a different piece of the interaction model.The MINOS ND sees interactions over a large region of x and Q2.

  • 10 Safe and Low Q2 DIS SamplesNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Default MC with flux tuning (before any other fitting or selection) models the DIS samples reasonably well. Expect systematics from flux, cross section model and intra-nuclear re-scattering to be significant though.

  • 11 Transition and QE/ SamplesNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Again, expect systematics from flux, cross section model and final state interactions to be significant but the flux-tuned default MC shows some large disagreements with the data for QE and resonance interactions.

  • 12 Transition and QE/ SamplesNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookSuggests a mis-modeling of nucleareffects for resonance interactions. Thereis no Pauli-blocking of resonance eventsin the default MC.Suggests a mis-modeling of boththe rate and shape of QE interactionsat higher Q2 as well as at low Q2.

  • 13 Pauli-Blocking Resonance InteractionsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We have not yet included a better nuclear model for resonances but have applied Pauli-blocking (and with increased Fermi momenta) as an effective low Q2 suppression. This substitute does improve data/MC agreement.

  • 14 Selecting CC QE EventsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We select a QE-enriched sample using a simple cut on the hadronic shower energy and only consider events with muons that stop in the ND:Ehad < 250 MeV for QE sampleEfficiency: 53%, Purity: 61%

  • 15 Data/MC Comparison for QE SampleNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 16 Data/MC Comparison for QE SampleNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookTotal # of events selected:Data: 344,736MC: 292,501Data wants more low Q2 suppressionand a flatter spectrum at higher Q2.

  • 17 Fitting for an Effective MAQENUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Use MINUIT to minimize:# observed events in bin i# expected events in bin iFit parametersShift in systematicparameter j from nominal1 errorfor systematicparameter jMC normalization Will show results from 2 shape-only fits to Q2QE (>0.3 and >0.0 GeV2).

    The most important systematic effects are considered directly in the fits whilst other uncertainties are studied via their impact on the fit results.

    Based on observations of the transition sample we apply Pauli-blocking to the resonance interactions. We also apply a correction for the Coulomb interaction between the nucleus and outgoing lepton.

  • 18 Fit Parameter: MAQE-ScaleNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Treated as a free parameter in the fits. The effect on the QE sample is asymmetric due to the non-QE background. Note that these figures are absolutely normalized although the fits only consider the Q2 shape.Nominal = 0.99 GeV

  • 19 Fit Parameter: Muon Energy ScaleNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We take a 1 error of 2% for use in the penalty term (this is the published MINOS uncertainty for stopping muons). This parameter changes Q2 via the muon-only kinematic reconstruction.Nominal = 1.00

  • 20 Fit Parameter: MARES-ScaleNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We take a 1 error of 15% for use in the penalty term (again, this is the published MINOS uncertainty). This parameter can change both the shape and rate of the resonance background in the QE sample.Nominal = 1.12 GeV

  • 21 Fit Parameter: Low Q2 SuppressionNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook We fit a scaling of the Fermi momentum used to Pauli-block QE events. This is an effective parameter to account for mis-modeled nuclear effects and is only used when we fit for MAQE below Q2=0.3 GeV2. Nominal = 1.00

  • 22 More on Effective Low Q2 SuppressionNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Now looking at true Q2 for true CC QE events. We take an assumed 1 error of 30% on this parameter which corresponds to removing about 50% of the lowest Q2 events. There are limitations for this parameter

  • 23 Examples of Nuclear Model SpreadNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookComparing FG to SF there is a Q2 shape changepersisting out to 0.6 GeV2. Comparing RFGM toRDWIA there is only a Q2 shape change out to 0.2 GeV2.

  • 24 Results for >0.3 GeV2 FitNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Reduced 2: 1.545 -> 0.836

    MC Scale Factor: 1.311 -> 1.103MINOS Preliminary

  • 25 >0.3 GeV2 Fit Contour and CorrelationsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Contour drawn from MINUIT and minimized at every point with respect to the other fit parameter; the MARES-scale. There is a relatively strong anti-correlation between these two fit parameters.MINOS Preliminary

  • 26 Systematic Errors and >0.3 GeV2 ResultNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookIncludes effects suchas mis-calibration andmis-reconstruction.Changes to intra-nuclearre-scattering parameterssuch as pion elastic orin-elastic scattering, chargeexchange, secondary pionproduction, the formation time and nucleon absorption. Investigated via changesto the value of the Fermimomentum used to applyPauli-blocking.Effective MAQE = 1.26 +0.12-0.10 (fit) +0.08-0.12 (syst) GeVMINOS PreliminaryMINOS Preliminary

  • 27 Results for >0.0 GeV2 FitNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Reduced 2: 216.5 -> 1.023

    MC Scale Factor: 1.180 -> 1.140MINOS Preliminary

  • 28 >0.0 GeV2 Fit ContoursNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Contours are drawn from MINUIT and minimized at every point with respect to the other fit parameters; the MARES-scale and either the QE Fermi momentum scale or stopping muon energy scale.

  • 29 Systematic Errors and >0.0 GeV2 ResultNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookChange in both data andMC -> sanity check.Re-weight the MC for changes to the normalization of lowmultiplicity channels.Vary flux tuning withinerrors from beam fits.Effective MAQE = 1.19 +0.09-0.10 (fit) +0.12-0.14 (syst) GeVFit errors are reduced because of theextra statistics from the low Q2 region.Additional systematic error is increased due to inclusion oflow Q2 region.MINOS PreliminaryMINOS Preliminary

  • 30 SummaryNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook From a fit above the region where we might expect serious nuclear effect mis-modeling we find a best fit effective MAQE = 1.26 +0.12-0.10+0.08-0.12 GeV best represents our data.

    From a fit above Q2=0.0 GeV2 we find that a best fit effective MAQE = 1.19 +0.09-0.10+0.12-0.14 GeV, along with an effective low Q2 suppression given by taking a Fermi momentum 1.28 times higher than nominal, best represents our data.

    The MINOS preliminary results, on Iron and at higher energy than K2K and MiniBooNE (2.5 GeV peak with most data from 1-6 GeV), show the same trends as these experiments.

    Our data wants more low Q2 suppression, consistent with a beyond the Fermi Gas nuclear model. The shape of the Q2 spectrum requires an increase in the relative number of high Q2 events, which we accomplish in our model by increasing MAQE.

  • 31 CommentsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook The extent to which our measured parameters should be considered effective is intrinsically linked to nuclear effect (mis-)modeling.

    In this analysis, like some other experiments, we have tried to separate the lowest Q2 behaviour from the middle part of the Q2 spectrum.

    Even though our means of applying the low Q2 suppression is too simple, the way it enters the error on the MAQE shape fit appears to be reasonable and possibly conservative.

    These results are preliminary because there are areas where we would like to make improvements (for example a more explicit study of Q2 resolution effects) and because there are some extensions that we would like to make to the results.

  • 32 Analysis OutlookNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook There are a number of routes for extending and improving this analysis: use of additional data (we have over 5 times the amount of low energy running as well as data from alternate beam configurations)

    potential measurement with anti-neutrino QE events using the 6% contribution of anti-neutrinos to our beam

    investigation of absolute rate effects (for example using our kinematic plane sub-samples)

    improved QE selection:We are developing an entirelycomplimentary QE selection procedure that looks for protontracks. In particular, this sample has a higher Q2 reach.

    [See poster by N. Mayer for details.]

  • 1 Backup SlidesNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 2 NuMI Beam PerformanceNUINT 09, CC/NC QE Scattering, 19th May 2009RUN I - 1.27x1020 POT

    RUN IIa 1.23x1020 POT RUN III 3.0x1020 POT + Total Protons (e20)2006 CC PublicationNu08 NCNu08 CCRUN IIb 0.71x1020 POT MINOS currently has over 7e20 POT!Other running, HE beam: 0.15x1020 POT CCQE AnalysisIntroduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 3 QE Scattering: Nucleon Form FactorsNUINT 09, CC/NC QE Scattering, 19th May 2009 Describe hadronic vertex in terms of nucleon form factors which define how much of each type of weak current contributes to a scatter:

    These form factors are functions of the probe strength, Q2.

    In practice, only 3 of the form factors are non-zero:where(V-type and A-type currents)Multiplied by leptonmass in cross section so neglected for our muons.Well measured through electron scattering experiments.Not well measured and only neutrinoscattering experiments can extract FA(Q2).Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook

  • 4 MiniBooNE ResultNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookPhys. Rev. Lett 100 (2008) 032301 On Carbon

    Extra low Q2 suppression wanted by data.

    Region below 0.2 is fit with ( = extra Pauli-blocking) .

    MAQE = 1.23 0.2

  • 5 K2K ResultNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookPhys. Rev. D 74 (2006) 052002 On Oxygen

    Extra low Q2 suppression wanted by data but no attempt to include in model or fit.

    Region below 0.2 is not fit.

    MAQE = 1.20 0.12

  • 6 Estimated Q2 Resolution and BiasNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook MC estimates of resolution and bias in QE-assumed reconstruction of Q2QE. Estimates are made with two methods and in either case the fractional bias is everywhere smaller than the fractional resolution.

  • 7 >0.3 GeV2 Fit: Absolute NormalisationNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook MC is now normalized to the POT of the data. The best fit parameters do not fully correct the rate but systematic errors, such as from the flux and intra-nuclear re-scattering, would likely cover the difference.

  • 8 >0.3 GeV2 Fit: Absolute NormalisationNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook 2D data over MC ratios. The histograms are only filled when both data and MC contain greater than 7 events.

  • 9 >0.0 GeV2 Fit: Absolute NormalisationNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook MC is now normalized to the POT of the data. The best fit parameters do not fully correct the rate but systematic errors, such as from the flux and intra-nuclear re-scattering, would likely cover the difference.

  • 10 >0.0 GeV2 Fit: Absolute NormalisationNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook 2D data over MC ratios. The histograms are only filled when both data and MC contain greater than 7 events.

  • 11 >0.0 GeV2 Fit: Correlation MatrixNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/OutlookMINOS Preliminary

  • 12 Comparison of Fit ResultsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Left figure shows the two sets of fit results along with the errors returned by the fits and points showing the nominal parameter values and input 1 errors. Right figure shows an overlay of 1 contours from the fits.

  • 13 Comparison of Systematic ShiftsNUINT 09, CC/NC QE Scattering, 19th May 2009Introduction MC / QE Theory Data/MC Comparisons Axial Mass Fits Summary/Outlook Comparison of the shifts in the best fit MAQE for the two fits along with the quadrature sum of the positive and negative shifts in each case.