Electron Reconstruction

Results of C.Charlot, C.Rovelli and Y.Sirois presented at Physics Meeting, 22 June 05 and at Physics Meeting, December 05

• The electron tracks are followed right through to the ECAL, providing additional possibilities for ECAL/tracker matching
• A measurement of the track momentum at the ECAL end of the track, together with the momentum measurement at the vertex, provides a measure of the amount of energy radiated in the tracker material
• The momentum error measurement may be more realistic than that obtained from the KF tracking

• By using the mode of the GSF probalility density function (PDF) rather than the weighted mean, a better measurement of the electron momentum at the vertex is obtained.

	Transverse momentum at vertex for pT = 10 GeV/c electrons. Distributions obtained considering only the mode of the PDF (blue) and using the weighted mean of all the Gaussians which enter the PDF (red) are shown.
GIF EPS

• Trajectory seeding: default track reconstruction strategy, with two hits in the pixel detector. Regional seeding was used to reduce combinatorics: only seeds in a (large) region matching an ECAL supercluster were considered. The seeds corresponding to all candidate superclusters were collected, followed by a single cleaning and single smoothing step.
• The supercluster was obtained using the default ECAL clustering and superclustering, but the seed cluster E_T threshold was lowered to 1 GeV.
• Trajectory building: the Bethe-Heitler parametrization for the energy loss was used. An `infinite` chi2 cut is applied to collect hits up to the outer radius of the tracker.
• Trajectory fitting: the Gaussian Sum Filter was used both in the forward and in the backward fit.
• Comparisons are shown below of the results obtained with this configuration and the results obtained with KF tracks. The KF track results are given for both the default track reconstruction, and the HLT electron cluster-seeded KF tracks where the seeding is tuned for background rejection in HLT, and the tracking is tuned for measurement of the momentum at the vertex.

	Distribution of the number of reconstructed hits per track for pT = 10 GeV/c electrons. Distributions are shown for electron GSF tracks (blue), for default KF tracks (red) and for HLT electron tracks (black).
GIF EPS

		Delta eta between the reconstructed and the generated track for pT = 10 GeV/c (left) and  for pT = 30 GeV/c (right) electrons. Distributions are shown for GSF tracks (blue), default KF tracks (red) and the HLT electron tracks (black).
GIF EPS	GIF EPS

		Delta phi between the reconstructed and the generated track for pT = 10 GeV/c (left) and  for pT = 30 GeV/c (right) electrons. Distributions are shown for GSF tracks (blue), default KF tracks (red) and the HLT electron tracks (black).
GIF EPS	GIF EPS

		Reconstructed momentum at vertex as a fraction of the generated momentum for pT = 10 GeV/c (left) and  for pT = 30 GeV/c (right) electrons. Distributions are shown for GSF tracks (blue), default KF tracks (red) and the HLT electron tracks (black).
GIF EPS	GIF EPS

	Difference between the track momentum at vertex and at the outermost hit versus the true momentum carried away by bremsstrahlung photons for pT = 10 GeV/c electrons. The weighted mean of the gaussians of the PDF is used.
GIF EPS