Electron Reconstruction

Electron candidate definition, preselection and identification

Results of S. Baffioni et al., presented at Physics Meeting, 05 December 05

Electron candidate definition

The reconstruction of electrons in CMS starts by the reconstruction of superclusters in the ECAL which in turn are used to search for
pairs of compatible hits in the pixel layers. Each pair of pixel hits is then used to grow trajectories inward-outward. The definition of
electron candidates further involves a loose preselection so to be suited to any physics analysis involving primary electrons. Electron
candidates are defined as:

a reconstructed electron track initiated by the reconstruction od a supercluster in the ECAL matched with two hits in the pixel detector
an energy-momentum matching between the supercluster and the track E/p < 3
an eta geometrical matching between the track extrapolation from the innermost position and the supercluster Deta < 0.02
a phi geometrical matching between the track extrapolation from the innermost position and the supercluster Dphi < 0.1
a ratio of the energy deposited in the hadronic tower just behind the electromagnetic seed cluster over the energy of the seed cluster H/E < 0.2

Electron reconstruction and the definition of electron candidate has been studied in the p_T range 10-50 GeV and is illustrated on the physics channel
H->ZZ^*->e⁺e^-e⁺e-.

		Electron candidate efficiency for electrons from Higgs decays H->ZZ^*->e⁺e^-e⁺e^- after preselection and for m_H=150GeV/c² (a) as a function of p_T and (b) as a function of eta
GIF EPS	GIF EPS

Electron isolation and selection of primary electrons

<>The selection of primary electrons is based on a requirement on the transverse impact parameter of the reconstructed track. Lepton isolation requirement can be imposed as a simple an powerfull means of suppressing QCD background in multi-lepton physics channel. Using track isolation allows to avoid the complication due to external or internal bremsstralhung photons which might appear in the ECAL within an isolation cone.

		Distributions of the normalised transverse impact parameter for electrons from H->ZZ^->e⁺e^-e⁺e^-(m_H=150GeV/c²) and from the ZZ, ttbar and Zbbar backgrounds (left) and efficiency of track isolation requirement for the Higgs boson signal as a function of the rejection against the ttb background (right).
GIF EPS	GIF EPS

Electron identification

Electron identification finally makes use of shower shape and cluster-track matching variables. Their distributions often strongly depend on the electron class. Identification performances can therefore be improved using class dependant cuts or probability distribution functions. The study has been made using samples of QCD di-jet events to evaluate the fake background in the p_T range 5-50GeV/c.

		Distributions of hadronic energy over the electromagnetic energy (left) and of the sum of the 3x3 matrix over the sum of the 5x5 matrix centered on the hotest crystal (right) for electrons from the golden class (blue), from the showering class (brown) and from the fake background (black).
GIF EPS	GIF EPS

		Distributions of the second order moment of the eta projection of the shower transversal profile (left) and of the supercluster energy over the track momentum (right) for electrons from the golden class (blue), from the showering class (brown) and from the fake background (black).
GIF EPS	GIF EPS

		Distributions of the difference between the reconstructed supercluster eta (left) and phi (right) position w.r.t the track position. The track parameters are taken at the innermost track position and propagated to the ECAL.
GIF EPS	GIF EPS

		Distributions of the seed cluster energy over the track momentum estimated from the outermost track position (left), and of the difference between the seed cluster phi position and the track position at the ECAL taking track parameters from the outermost position (right).
GIF EPS	GIF EPS

Last update: 30 June 2005
Comments: please contact Chris Seez, Yves Siroi