Direct Measurement of the W Boson Width
 
Valeria Bartsch, Daniel Beecher, Ilija Bizjak, Mark Lancaster,
Sarah Malik, Emily Nurse, Troy Vine, David Waters
University College London


&GammaW = 2032 ± 73 MeV/c2
(the world's most precise single direct measurement)

This is now published as Phys. Rev. Lett. 100 071801 (2008). The PRL is here

 
Transverse Mass Fits

Contents
  • Abstract
  • Measurement Technique
  • Signal Modelling
  • Energy & Momentum Scales
  • Hadronic Recoil
  • Backgrounds
  • Width Fits
  • Uncertainty Tables
  • New World Averages



    •  
    Download plots in eps format by clicking on the plot.
    To view the full gif format, right-click and select "View Image."


     
     
     Abstract

    We present the first direct measurement of the W boson width in approximately 350 pb-1 of proton-antiproton data collected by the CDF II detector. Fitting to the high transverse mass tail in both electron and muon channels, we find &GammaW = 2032 ± 73 (stat.+syst.) MeV/c2, making this the world's single most precise direct measurement.

     
     
     Measurement Technique

    We model the transverse mass distribution over the range 50-200 GeV. We normalise to the data in the region 50-90 GeV and fit for the width in the high transverse mass region 90-200 GeV. The tail region, separated from the bulk of the resolution effects that determine the shape of the transverse mass distribution close to MW, is sensitive to the width of the Breit-Wigner lineshape, as indicated in the following figure.


     
     
     Signal Modelling

    We model W production and decay at leading order, modified to take into account the effect of higher order QCD and QED corrections. Parton distribution uncertainties affect the W rapidity distribution, which propagates into an uncertainty on the fitted width (see figure below). We measure the boson pT distribution in Z events and compare to a NLO resummed calculation (identical to RESBOS). We fit for the non-perturbative form factor g2 and a linear distortion parameter B (see figures below). B is consistent with zero and is set to zero. Uncertainties on g2 and B are propagated into an uncertainty on the fitted width.


     
     
     Simulation

    Material scale factor. We tune the amount of material by fitting to the high (E/P) tail in wide bins :

     
     
     Momentum Scale and Resolution

    The track momentum scale and resolution are found from the M(μμ) distribution in Z→μμ data. The delta curvature distribution is taken from W→μν events in the full GEANT Monte Carlo. Delta curvature for the fast simulation is obtained by sampling the distribution and multiplying by a constant factor: Fres. The value of Fres is found by tuning to M(μμ) in the Z→μμ data. The momentum scale is found by tuning to M(μμ) in the Z→μμ data. The non-gaussian tail fraction in the delta curvature distribution is varied and fitted for using the E/p distribution in W→eν data.

     
     
     Calorimeter Scale and Resolution

    The Central Electromagnetic Calorimeter (CEM) scale and resolution are found from the M(ee) distribution in Z→ee data and the E/p distribution in W→eν data.
     
     
     Recoil Calibration


    The recoil is obtained by fitting to Z data and minimum-bias data.

    Z recoil fits :
     
     
     Backgrounds

    Muon channel:

    Summary of background composition :
    • Background transverse mass distributions : (eps) (gif)
    The Decay In-Flight (DIF) background is found by fitting the chi2 distribution. DIF events are selected by reversing the d0 cut and Zmumu events are used to get the expected signal distribution.
    • Chi2 fit (tracks without silicon hits) (eps) (gif)
    • Chi2 fit (tracks with silicon hits) (eps) (gif)
    The multi-jet background is found by fitting the met distriution. Multi-jet events are selected by making a reverse track isolation cut.

    Electron channel:

    Summary of background composition :
    • Background transverse mass distributions : (eps) (gif)
    The multi-jet background is found by fitting the met distriution. Multi-jet events are selected by reversing electron ID cuts.
     
     
     Width Fits

    The muon and electron transverse mass distributions are fitted for the width.
    • muon fit result : &GammaW = 1948 ± 67 (stat.) MeV/c2 : (eps) (gif)
    • electron fit result : &GammaW = 2118 ± 60 (stat.) MeV/c2 : (eps) (gif)
    Combining the two taking fully into account all correlations yields &GammaW = 2032 ± 73 (stat.+syst.) MeV/c2. The combination probability is 20%.

    As a cross check the muon and electron transverse momentum distributions and the missing et distributions are fitted for the width:
  • muon et fit: (eps) (gif)
  • electron et fit: (eps) (gif)
  • muon met fit: (eps) (gif)
  • electron met fit: (eps) (gif)
    •  
     
     Uncertainty Tables

     
     
     New World Averages

    This measurement is compared to other direct width measurements here
    The CDF run-2 measurement is the most precise measurement from a single experiment.
    It improves the uncertainy on the world average from 60 to 47 MeV.