|
|
|
|
|
 |
|
|
 |
|
|
The dominant background to diphotons with large MET comes from QCD events
where a jet or vertex mismeasurement leads to "fake MET".
The MET resolution should be very similar for events with photons and with jets faking photons,
since a jet that fakes a photon fragments into one or a few neutral pions.
This class is called "Sideband event".
The comparison in MET shape between Signal and Sideband events
are used to estimate QCD background.
This is done by normalizing two MET distributions in the region of low MET (MET < 20 GeV).
The expected number of QCD background is 0.29 +0.50-0.25(stat) +-0.24(syst) with MET > 45 GeV.
The uncertainty in the background at large MET is estimated to have a 60% statistical uncertainty
and is taken from the variation in the prediction of the selection requirements on the di-EM sample
(or sideband events), and 70% from the fit variation (partially from fitting choices
and partially from statistical uncertainties of the sample).
The fraction of events from true gammagamma sources is estimated
using the shower shape and conversion probability,
where background from a neutral pion decay to gammagamma is subtracted.
|
|
|
The other dominant background comes from events with "true MET".
If a real electron is produced but its track is not reconstructed,
then it will pass all photon cuts and produce a background to photons.
The track may be lost due to tracking inefficiency, or more likely, to a hard bremsstrahlung
where the track would be cut away.
The usual sources of electrons are W, Zgamma, b and conversions.
There may be several in events (WW, etc.) or they may be produced with a photon, such as Wgamma.
The photon may be real or fake. We are mostly concerned with events with true MET since it is difficult to fake
- this means the largest background from electrons comes from Wgamma.
All sources are estimated at once using the data by taking a sample of identified e-gamma events
and multiplying by the rate at which good electrons are falsely identified as a photon,
The fake rate is estimated (from Z0->ee events) to be roughly 1% for selection criteria used here.
The systematic uncertainty in the e-gamma background is estimated to be 32%,
dominated by differences between the sample of transverse energy spectrum and the fake rate
which is averaged over all transverse energy.
0.10+-0.05(stat)+-0.03(syst) events are predicted with MET > 45 GeV.
|
|
|
Non-collision backgrounds from cosmic rays or beam-related problems can contribute
one or more fake photons and/or MET. The dominant background is from sources which produce
both photons which are fake. Using the data this background is estimated
using a sample of events with no primary collision
and two electromagnetic (EM) clusters, multiplied by the rate at which EM clusters from cosmic rays
pass the photon identification cuts. For MET > 45 GeV 0.21+-0.05+-0.14 events are predicted.
|
|
|
The total expected number of background events is 0.60 +0.50-0.26(stat) +-0.28(syst) with MET > 45 GeV.
Non-collision background is excluded from MET spectrum below.
|
|
|
