(larger version) (eps version)
This page summarizes the PR results for the X(3872) -> J/Psi PiPi with J/Psi -> ee at CDF which were approved by the B group on 9th June 2005.
The reconstruction of the X(3872) -> J/Psi pi+ pi- with
J/Psi -> e+e- is very challenging experimentally in
the complex hadronic environment at the Tevatron
collider.
The CDF experiment deploys a dedicated trigger for
this J/Psi decay channel. However, a highly efficient
method to select electrons at high purity is needed to
separate the signal from the more than 10 times higher
background from mainly pions.
Electrons are identified using
the sophisticated NeuroBayes(R).
Multiple neural networks are deployed in this analysis.
One network identifies electrons by analysing
the change in curvature of the reconstructed track
as the particle traverses the matter in the detector.
The next network combines this information with
measurements of the energy deposited in the electromagnetic
calorimeters, specific energy-loss (dE/dx) in the central
drift chamber and information from the time-of-flight
detector.
This allows to efficiently select electrons from the
complex hadronic environment at high purity, removing (almost) all
non-electron particles from the original selection.
Electrons and positrons identified this way are combined to
form a J/Psi -> ee candiate. The resulting invariant mass spectrum
is shown in the figure below:
(larger version) (eps version)
The main remaining source of background
are conversion electrons from the process photon -> e+e-.
A dedicated neural network has been trained to identify electrons
and positrons which originate from this conversion process.
Rejecting J/Psi candidates where either the electron or the positron
originate from the conversions, this source of background can be
removed efficiently and the invariant J/Psi mass distribution is
observed at a very low background level.
(larger version) (eps version)
J/Psi particles decaying to electron-positron pairs lose a
substantial amount of energy from the
process J/Psi -> (e+ e- gamma) and Bremsstrahlung as
the light electrons/positrons traverse the detector.
This results in the strong radiative tail observed in the
above figures.
A further dedicated network determines
whether the electron or the positron forming the J/Psi has lost
more energy due to the above processes.
To correct for the energy loss, a photon is added to the (e+e-)
vertex. The energy of the photon is determined such that the
then fitted J/Psi mass agrees with the nominal J/Psi mass
(given by the Particle Data Group). The direction of the photon
is determined such that it is parallel to the
electron or positron which has lost more energy.
The J/Psi -> ee identified by the procedure outlined above are then combined with two oppsitely charged pions to form am X(3872) candidate. The selection cuts to suppress the background follow the steps taken in the observation of the X(3872) for the case where the J/Psi decays into two muons (See Phys. Rev. Lett. 93, 072001 (2004) for details).
In detail, the selection criteria are:
The NeuroBayes neural network package:
X(3872) - Theory:
X(3872) - Experimental results:
X(3872) - conference contributions:
