Transcript Photon candidates

H -> gg introduction
Satoru Uozumi
Jan-9th 2010 KNU/TAMU meeting
• Introduction for students
- Photon signal
- Higgs generation and decay
• Photon ID at CDF
• ….
Intro – photon signal at CDF
•Silicon Vertex Detector
(precise position measurement)
•Tracking Chamber
(momentum measurement)
•1.4 Tesla Solenoidal magnet
• Calorimeter (Electromagnetic/Hadron)
(energy measurement)
• Muon Detector (yellow & blue)
Photon signal has
• no associated track
(because of no electric charge)
• small and dense shower signal
in electromagnetic calorimeter,
similar with electron
Intro - The CDF Calorimeter
e/g
Shower
signal
CES
signal
Photon
Photon
Not photon
Track
– Hadronic part
– Electromagnetic part (cell size ~ 20cm) for
photon/electron energy measurement.
Shower-max detector (CES) for shower position
measurement.
In principle, it’s the same also in CMS
Identification of Photon Signal
Photon candidates: isolated electromagnetic showers in the calorimeter,
with no charged tracks pointing at them
Sungwon Lee
2003 CTEQ Summer School
4
However, photon-ID is more difficult in CMS…
Simplest two-photon event at CDF
(photon event is not always like such simple !)
• Clusters only in
EM calorimeter
• No associated track
to the clusters
Simulated Higgs -> 2 photon event at CMS
In CMS, there are more background tracks
and clusters due to more QCD background
and pile-up events (=multiple collisions at
one bunch crossing).
CDF Photon ID for Higgs/exotic search
• “standard CDF photon ID”
for Randall-Sundrum Graviton
and H->gg search (CDF 8858 etc.)
• Criteria is similar with electron
selection, but without track.
• Main background : collimated
photons from 0 decay in jet
EM Calorimeter
g
0
(in jet)
g
g
Preshower Detector
Shower Maximum Detector
Neural Network approach for
CDF Photon ID
• CDF 5791 … Electron/Photon ID with Neural Net
JetNet package, S/B improved 1.9 -> 2.6 with Run I data
• CDF 10282 … multi-variable photon ID using TVMA package
Photon signal with e+e- conversion in CDF
Photon signal is accessible also with electron data
through conversion.
However the conversion events are not used for
exotic particles search with photons in CDF.
Photon Identification – learn from CDF
•
Usually jet contains one or more 0 mesons which decay to photons
–
we are really interested in direct photons (from the hard scattering)
–
but what we usually have to settle for is isolated photons (a
reasonable approximation)
Isolation: require less than e.g. 2 GeV within e.g. R = 0.4 cone
•
This rejects most of the jet background, but leaves those cases where a
single 0 or  meson carries most of the jet’s energy
•
This happens perhaps 10–3 of the time, but since the jet cross section is
103 times larger than the isolated photon cross section, we are still left
with a signal to background of order 1:1
There are a number of different technique to distinguish photons from 0
backgrounds. (see below)
1. Conversion Probability: g’s to convert in a preshower detector
2. Shower Profile: 2 g’s from 0 will produce EM showers with broader
lateral and smaller longitudinal profiles
3. Reconstruction: requires good EM/angular resolution (fixed target)
Additional issue for photon ID at LHC – pile-up of multiple collisions
Sungwon Lee
2003 CTEQ Summer School
Photon with e+e- conversion in CMS ?
• Due to tracker material:
- photons have >50% probability to convert into e+e- pair
- May give access to the photon signal ?
Production of SM Higgs
Gluon fusion
… dominant at LHC
Vector boson fusion
… 2nd at LHC
Two jets associated
with Higgs
qqbar -> WH, ZH
ttH, bbH processes
SM Higgs decay
LEPEWWG Summer2003
• If the Higgs mass is in the SM preffereded region,
many decay channels will be expected
-> measurement of Higgs coupling can be measured to various particles
• 2 gamma decay is one of the important channel in this region
Possibility of Enhancement
of non-SM Hgg decays
H->gg Branching Fraction
no couplings to fermions
(Fermiophobic Higgs)
no couplings to
top,bottom quarks
no couplings to
down-type fermions
Standard Model
Higgs Mass, GeV
S.Mrenna, J.Wells, Phys. Rev. D63, 015006 (2001)
in general we should be prepared for any H->gg
branching fraction ( up to 1.0 ) due to new physics
H -> gg
• First analysis … inclusive search
• Easy - Just reconstruct masse from two photon candidates
• Sharp mass peak can be observed
(thanks to good ECAL resolution)
• Background is large, though (S/B < a few%),
Background can be estimated from mass sideband.
Still important to understand components of backgrounds.
CMS 14 TeV
1 fb-1
(signal scaled x10)
Reducible background
• g+jet
• jet+jet
58 % jet gluon
4.7 % jet quark
37.3 % mixture
Partonshadrons (« jets »)
Irreducible background
• Born(aQED2)
• Bremsstrahlung
(correction to Born)
• Box gggg (as2aQED2)
Calculation by Pythia (for ATLAS)
Mgg (GeV)
Vector Boson Fusion analysis
• Inclusive H -> gg analysis is simple, but BG is large
• If we focus on VBF process, S/B can be significantly improved
- In addition to gg signal, require two forward jets
- Also require no 3rd jet in central region
• Study at ATLAS (LAPP-EXP-2008-09)
Prospect on photon-ID
• CDF photon-ID - “like electron without track”
• Difference at CMS :
- finer ECAL granularity
- Pile-up events
- material budget
• At first, need to know what has been done for
the CMS H->gg analysis, then find out what we
can do
Backups
Run 1  Run 2
• The TeVatron is a broad-band quark and gluon collider
Number
of
Events
Huge statistics
for precision physics
at low mass scales
Formerly rare processes
become high statistics
processes
Increased reach
for discovery physics
at highest masses
Run 1
Run 2
Extend the third orthogonal axis:
the breadth of our capabilities
Sungwon Lee
2003 CTEQ Summer School
subprocess s
21
Identification of Photon Signals
Photon candidates: isolated electromagnetic showers in the calorimeter,
with no charged tracks pointing at them
Central Calorimeter
• Signals
• CDF/DØ uses two techniques for
determination of photon signal;
1. EM Shower width
2. Conversion Probability
• CDF measures the transverse
profile at start of shower
(preshower detector) and
at shower maximum
• DØ measured longitudinal
shower development at start
of shower
g
• Backgrounds
g
0

g
Preshower Detector
Shower Maximum Detector
CES has better separation,
CPR better at high Et (Et>35)
Sungwon Lee
g-g
g-Jet
Jet-jet
CES
24±6%
28±8%
48±7%
CPR
29±23% 40±28%
30±23%
2003 CTEQ Summer School
22
Photon Purity Estimators
• CDF
• DØ
Each ET bin fitted
as sum of:
(a) = photons
(b) = bgd w/o tracks
(c) = bgd w/ tracks
E1: E in the 1st
calo section
For every photon, using the conversion
and profile info., CDF find the fraction of
candidates with this info. (extracted
signals statistically)
Sungwon Lee
DØ model longitudinal energy
depositions of photon’s and jets and
perform a statistical comparison to
data using the discriminant variable to
determine the photon purity.
2003 CTEQ Summer School
23
DiBoson Production
Test Gauge Boson Self Interactions
SM Higgs searches
Resonance searches: Look for excess in kinematical
distributions: ET(g) , 3body mass, lepton PT
Complementarity with LEP experiments:
Probing at higher √s
W-g final state
EM Calorimeter
For Wg/Zg Photon Id is crucial:
Main backgrounds:
0→gg,
jets faking photon
Fake Rates:
0.2% @JetET=10 GeV
0.05% JetET>25GeV
g
0
ET(g)>7 GeV
g
g
Pre-Shower
Detector
R(g,l)>0.7
|g|<1.1
Cal & Trk Iso
Shower
Maximun
Detector