First look at Wgg

P. BellUniversité de Genève

Informal Wg/Zg meeting30th June 2011

W + 2 photon production at LHC- Sensitive to Anomalous Quartic Coupling of form WWgg

- Generic (non-ATLAS) study in arXiv:0907.5299:- Compared/cross-checked LO predictions of Baur and

Eboli/Lietti MC generators

- Assumed:- √s = 14 TeV- 25GeV min. charged lepton pT, - 15GeV min. photon pT- 25Gev min. MET- |eta|<2.5 for charged leptons and photons- DR = 0.4 between the three objects

- Detector response (smearing and efficiency) modelled by “PGS” package

accepted signal cross section 8.6fb (electron and muon channels combined)

now is the time to start this analysis!

NB: More recent MC studies (arxiv.org/abs/1103.4613, arxiv.org/abs/1001.2688) show large NLO corrections (k-factor > 3, non-uniform)

First look at data (electron channel)

- Started to slim and look for events in EGAMMA and later SMENU D3PDs (to period H)- Cutflow not refined: essentially took the most certain of the Wg selection and asked for two photons:

- Egamma GRL + e20medium trigger- At least one tight electron with cl_pT > 25GeV inside |eta| < 2.47, crack region

removed; no further medium electrons; el oq flag applied- At least two tight photons with cl_pT > 15GeV inside |eta| < 2.47, crack region

removed; ph oq flag applied- MET_RefFinal > 15GeV - deltaR between the three EM objects > 0.4

- Results appearing at https://twiki.cern.ch/twiki/bin/view/AtlasProtected/WGammaGamma

Found 12 candidate events in B-G2 – might be interesting to examine each

- Cross-check of results would be good

MC reviewExisting datasets:- Based on MadGraph (SM couplings) e.g. from

MC10.118615.Pythia_MadGraph_Wminushadgammagamma.pyMC10.118616.Pythia_MadGraph_Wminuslepgammagamma.pyMC10.118617.Pythia_MadGraph_Wplushadgammagamma.pyMC10.118618.Pythia_MadGraph_Wpluslepgammagamma.py

Existing LO generators:- “W2Pho” (http://projects.hepforge.org/w2pho/) based on Lietti MC, cross-checked with Baur MC, has AQGC vertex and Les Houches output, could be used to generate official samples- ALPGEN can do W + 1 or 2 photons

At NLO:-arxiv.org/abs/1103.4613 and-arxiv.org/abs/1001.2688 => Both refer to private MC generators, have sent a request to obtain them. SM only.

Summary

- Expect ~9 Wgg events in 1fb-1 according to LO predictions + simple detector simulation

- Started to look for candidate events in electron channel (found 12 in period B-G2)- need to refine and cross-check cut flow- need to start on muon channel

- Which generator to use?– For eventually measuring a cross-section and ultimately setting QGC limits

- Need to understand NLO effect - can we make do with existing LO generators?- Ideally want NLO predictions with AQGC terms Should demonstrate to the theorists we are serious

- Not forgetting backgrounds – misidentified jets passing the photon selection expected to dominate – hope to use a “fake rate” method currently being developed for Z’ analysis.

- Rough plans- could make an “observation” and first total cross-section measurement soon- better to use differential distributions (Mgg) for probing AQGCs

Backup

NomenclatureTwo lowest dimension effective Lagrangian terms that give rise to purely quartic couplings involving at least two photons are:

Both terms generate Anomalous QGCs of the form WWgg and ZZgg

The coupling parameters b0 and bc (units GeV-2) are distinguished for the W and Z vertices and studied separately

NB: in LEP (and other) analyses the same couplings are expressed a0/L2 and ac/L2

All 4 parameters are zero in the SM

The anomalous couplings scale with the square of the photon energies, so a substantial improvement in the sensitivity can be expected at the LHC over the results from LEP.

Form factorsAny deviation of the parameters aW,Z

0,c/L2 from zero value will eventually violate unitarity so the cross section rise has to be regulated by a form factor(Not necessary in the LEP analyses)

Usual convention:

(i.e., for values of Mgg above LFF the AQGCs get suppressed)

If we work in a unitarity safe region, meaning up to some limit of Mgg, no need to apply FF and so can compare limits on bare couplings