Determination of b-jet energy scale using Z→bb signal
Authors
Julien Donini, Tommaso Dorigo, Mia Tosi
University of Padova and INFN
Kenichi Hatakeyama
Rockefeller University
Christopher Neu, Daniel Whiteson
University of Pennsilvania
Shawn Kwang, Melvin Shochet
University of Chicago
Tomonobu Tomura
University of Tsukuba
Contact the Authors

  • Abstract

    • We present a measurement of the energy scale of b-jets using the Z → bbbar signal based on 584 pb-1 of CDF Run II data. We extract a signal of 5674 ± 448(stat.) Z → bbbar decays in low-radiation central dijet events. We measure, using an unbinned likelihood technique, a b-jet energy scale factor of

    k = 0.974 ± 0.011 (stat)+0.017-0.014(syst) = 0.974 +0.020-0.018(stat+syst.).

    This measurement is relevant for b-jets in an energy range close to that of jets from Z decays.

    Using the reconstructed Z candidates we also measure the Z boson cross section times branching ratio in the bbbar final state:
    σZxBR(Z->bb) = 1578 ± 123(stat)+624-391(syst) pb = 1578 +636-410(stat+syst) pb.

    (Theory (NLO): σZxBR(Z->bb) = 1129 ± 22 pb.)

  • Figures

    • Description gif eps
    Kinematical distributions (data)
     Jet Et distributions for the two leading jets in experimental data before and after event selection (two leading jets with Et > 22 GeV and |η|<1.0).     		  eps1 eps2

    Kinematical distributions (data and MC)
     Additional kinematical characteristics of double tagged data. Left: Et of third jet for data (red histogram) and Z→bb MC (blue histogram). Right: azimuthal difference (ΔΦ) between the leading two jets for data and MC.     		  eps1 eps2

    Optimization of kinematical cuts
     Expected uncertainty on the jet energy scale factor returned by pseudo-experiments as a function of the cut on third jet Et (left plot) or the ΔΦ azimuthal angle between the two leading used to select the signal region (right plot).     		  eps1 eps2

    Signal PDF
     Dijet mass probability density function for Z→bb MC signal for 20 different values of scale factor, ranging from 0.9 to 1.09.     		  eps1 eps2

    Background template
     Data-driven background dijet mass distribution (red dots) fit with a parameterization (black line).     		eps

    Test of fitting procedure (1)
     Pull mean (left plot) and pull sigma (right plot) on fitted b-jet energy scale factor as a function of the input scale factor.     		  eps1 eps2

    Test of fitting procedure (2)
     Pull mean (left plot) and pull sigma (right plot) on number of events of fitted signal as a function of the input scale factor.     		  eps1 eps2

    Sideband fits
     Result of fit to the dijet mass sideband regions (mjj<60 GeV || mjj>120 GeV) of selected data with different background models. Fit Χ2 is shown as a function of the kinematical variables (ΔΦ between leading two jets and Et of third jet) that define each different background models.     		  eps1 eps2

    b-jet energy scale result
     Result of a constrained unbinned likelihood fit performed to double tagged dijets data (blue points). The signal is constrained to the number of expected events (gaussian constraint: 4630 ± 727 Gaussian). The data-driven background shape and Monte-Carlo signal p.d.f are shown respectively in green and red. The fit returns 5674 ± 448 events of signal and a b-JES of 0.974 ± 0.011 (errors are statistical only). The inset on the upper right shows the data minus background distribution (blue points) and the signal shape (in red) normalized to the fitted number of events of signal.     		eps

    Cross check (1)
     Result of constrained likelihood fit performed on a smaller 50-200 GeV dijet mass window (left plot). This fit yields 5912 ± 448 event of signal and a b-jet energy scale factor of 0.978 ± 0.011 (errors are statistical only).     		eps

    Cross check (2)
     Result of likelihood fit performed with an iterative procedure to correct the background from signal contamination, without prior assumption on the Z->bb cross section. This fit yields 6467 ± 504 event of signal and a b-jet energy scale factor of 0.976 ± 0.010 (errors are statistical only). This measured number of signal events is used in the σZxBR(Z->bb) calculation.     		eps

  • Tables

    • Description gif eps
    Statistics of analyzed data eps

    Signal acceptance
     Breakdown of all sources of uncertainty on the expexted number of signal events.     		eps

    MC acceptance
     Fraction of expected Z→bb, Z→cc and W→cs events in signal region and background regions, for double tagged events (++) and taggable events (00).     		eps

    Systematic uncertainties on b-jet energy scale
     Systematic uncertainies on fitted b-jet energy scale and number of events of signal.     		eps

    Other systematics
     Other sources of systematic uncertainies.     		eps

    Systematics on σZxBR(Z->bb)
     Summary of all sources of systematic uncertainties on the σZxBR(Z->bb) measurement.     		eps

    Maintained by Julien Donini
    Created: Wed Mar 21 20:38:52 CET 2007