Search for Anomalous Production of Di-photon+MET Events in 2 fb-1 of Data.

Authors:  Sasha Pronko (Fermilab); Ray Culbertson (Fermilab); Max Goncharov (Texas A&M); Shin Shan Yu (Fermilab).

A short analysis description can be found here.

Results for di-photon signal region: two tight photons (~25% of true di-photon events).

Non-collision backgrounds: cosmic ray and beam halo muons.
EWK backgrounds: mostly W+pho (e/tau->fake pho)  or W+jet (e/tau/jet->fake pho) events.
Fake MET: regular di-photon events and QCD pho-jet or jet-jet events where one or both photons are faked by a jet.


 MetSig>3.0 MetSig>4.0 MetSig>5.0
Observed
 120
 52
 34
EWK with real MET
 53.6 ± 8.9 47.3 ± 8.0 41.6 ± 7.0
Fake Met
 52.1 ± 11.5 15.4 ± 3.8 6.2 ± 2.7
Non-collision
 0.90 ± 0.32 0.85 ± 0.30 0.80 ± 0.27
Total predicted
background
 106.6 ± 14.5 63.6 ± 8.9 48.6 ± 7.5


The MET-significance distribution in di-photon signal sample (eps,gif).



The observed MET distribution in all events from di-photon signal sample (eps,gif).



The observed MET distribution in di-photon signal sample after metsig>5.0 cut (eps,gif).


Results for di-photon control region: two loose photons (~5% of true di-photon events), at least one photon must fail tight cuts.
This control sample has the same types of backgrounds as our signal sample.
It was used to test and tune the MET resolution model as well as other background estimation techniques.


 MetSig>3.0 MetSig>4.0 MetSig>5.0
Observed
 126
 63
 41
EWK with real MET
 68.5 ± 11.4 55.1 ± 9.9 43.6 ± 8.3
Fake Met
 61.3 ± 16.5 15.6 ± 5.1 4.2 ± 3.5
Non-collision
 1.29 ± 0.47 1.20 ± 0.43 0.85 ± 0.38
Total predicted
background
 131.1 ± 20.1 71.9 ± 11.1 48.7 ± 9.0


The MET-significance distribution in di-photon control sample (eps,gif).



The observed MET distribution in all events from di-photon control sample (eps,gif)

 

The observed MET distribution in di-photon control sample after metsig>5.0 cut (eps,gif).
Results for Z->ee control region: two central electrons, 86 GeV/c2<MZ<97.5 GeV/c2.
This control sample was used to test and tune the MET resolution model.


 MetSig>3.0 MetSig>5.0
Z(tau)+pho
 0
 0
Z(e)+pho
 2.72 ± 0.37(stat) 1.43 ± 0.27(stat)
Z->tau+tau
 0.23 ± 0.13(stat) 0.15 ± 0.11(stat)
ttbar
 1.45 ± 0.08(stat) 1.40 ± 0.08(stat)
WW
 2.97 ± 0.07(stat) 2.57 ± 0.06(stat)
WZ
 5.44 ± 0.04(stat) 4.97 ± 0.04(stat)
ZZ
 6.53 ± 0.06(stat) 5.94 ± 0.06(stat)
Fake MET
 46.1 ± 2.1 ± 10.0 4.0 ± 0.6 ± 2.8
Total background
 65.4 ± 2.1(stat) 20.5 ± 0.7(stat)
Observed
 61
 16


The MET-significance distribution in Z->ee control sample (eps,gif).



 The observed MET distribution in all events from Z->ee control sample (eps,gif).


The observed MET distribution in Z->ee control sample after metsig>5.0 cut (eps,gif).
Created by Sasha Pronko (pronko@fnal.gov)