Analyzing Monojet Dark Matter

Purpose of this page

In order to have information about the steps of the analysis centralized, we could write down any significant information, making it up to dated.

How to produce ntuples

How to produce distributions from LHE

Will come soon.

How to do full simulation

Will come soon.

Producing Ntuples from MiniAOD

These are the instruction on how to create Ntuples from MiniAOD via CRAB3.

Show Hide

cmsrel CMSSW_7_4_1
cd CMSSW_7_4_1/src
cmsenv

git cms-merge-topic 9003
rm -rf RecoEgamma/ElectronIdentification/data
git clone https://github.com/cms-data/RecoEgamma-ElectronIdentification.git RecoEgamma/ElectronIdentification/data

rm -rf RecoEgamma/PhotonIdentification/data
git clone https://github.com/cms-data/RecoEgamma-PhotonIdentification.git RecoEgamma/PhotonIdentification/data

cleanMuonJet=0;      // (1) for Muon control regtion    // (0) for Muon signal regtion
cleanElectronJet=0; // (1) for Electron control regtion // (0) for Electron signal regtion
cleanPhotonJet=0;   // (1) for Photon control regtion  // (0) for Photon signal regtion

Samples for Run2 (data, background and signal)

List of samples from Data Run2

The following table contains the list of Data samples (Run2015B-PromptReco-v1), as well as the correspondent number of events and luminosity.

These are the instructions to compute luminosity:

Show Hide

crab report -d <path_to_your_crab_directory/>

The following table shows the paths to the Data Ntuples in T2_SPRACE which includes events after the good vertex selection:

• offlineSlimmedPrimaryVertices
• !isFake && ndof > 4 && abs(z) < 24 && position.Rho < 2

The following table contains the list of Data samples (Run2015B-17Jul2015-v1) INTENDED to be produced in Ntuple format, as well as the correspondent number of events and luminosity. According to MiniAOD Analysis Documentation, this datasets correctly include E_T^missing filters. It is important to use the recommendations provided by MET group concerning Spring15 MC and 50 ns about:

• Met corrections for Run-II; and
• MET Filters for Run II.

List of MC samples for background (Run2 Spring15)

The following table contains the list of MC background samples (Run2 Spring15) with correspondent number of events and cross sections (from different sources). As mentioned in the Exo Diboson twiki page, "If only the lepton decay of the W is included in the sample, the cross section is multiplied by BR(W->lv) = 0.322 with l = e, mu, tau", and the references for cross sections come from:

• [1] https://twiki.cern.ch/twiki/bin/viewauth/CMS/StandardModelCrossSectionsat13TeVInclusive
• [2] https://twiki.cern.ch/twiki/bin/view/LHCPhysics/TtbarNNLO
• [3] https://twiki.cern.ch/twiki/bin/viewauth/CMS/SingleTopSigma
• [4] DAS

The following table shows the paths to the MC background Ntuples in T2_SPRACE site which includes events after the good vertex selection:

• offlineSlimmedPrimaryVertices
• !isFake && ndof > 4 && abs(z) < 24 && position.Rho < 2

List of MC samples for background (Phys14)

The following table contains the list of MC background samples (Phys14) with correspondent number of events and cross sections.

The following table shows the paths to the MC background Ntuples in T2_SPRACE site which includes events after the good vertex selection.

List of MC samples for signal (i2HDM)

The following table shows the path of MC samples for i2HDM signal in T2_SPRACE, as well as the number of generated events, the number events after the good primary vertex selection, and the cross section. There are samples considering cases of h1h1 and h1h2 DM production.

List of MC samples for signal (EFT)

The following table shows the path of MC samples for EFT signal in T2_SPRACE, as well as the number of generated events, the number events after the good primary vertex selection, and the cross section (from Monojet twiki page). These are samples with Axial-Vector and Vector cases.

Kinematic selections

This is the list of the standard Monojet selection applied in the signal region:

• 1 or 2 jets in the event
• p_T(Jet1) > 110 GeV
  • CH fraction > 0.2
  • NH fraction < 0.7
  • EM fraction < 0.7
  • CEM fraction < 0.99
• p_T(Jet2) > 30 GeV
  • NH fraction < 0.7
  • EM fraction < 0.9
• ΔΦ(Jet1, Jet2) < 2.5
• lepton veto
• photon veto
• MET > 200 GeV (PF MET with Type-1 corrections)

Kinematic distributions

Here we create distributions from Ntuples

Results from Physics14 samples

These figures were made with Physics14 samples for background while Spring15 samples are not ready. It does not have any selection other than good primary vertex.

Show Hide

Results from Run2 Spring15 samples

See Chang-Seong's presentation in the Monojet meeting on 10/Aug/2015.

Computing limits

As preliminary results, only statistical uncertainties are taken into acccount.

Results from Run2 Spring15 samples

From preliminary results, the following figures show the expected limits using Spring15 MC samples for background. To compute these limits, we use the expected number of events got from page 19 of Chang-Seong's presentation in the Monojet meeting on 10/Aug/2015.

Limits considering EFT Dark Matter model

• Figure on the letf (right) shows theoretical cross-sections for Axial Vector (Vector) Dark Matter. It is considered Lumi = 40.03 pb^-1. It takes into account the standard Monojet selection above, which leads to EFT signal samples with efficiencies from 11 to 14%. EFT signals present very low cross sections, as is showed in the table above. Here is the data card for Axial-Vector with DM mass = 1000 GeV using Lumi = 40.03 pb^-1:

Show Hide

imax    1       number of channels
jmax    6       number of backgrounds
kmax    1       number of nuisance parameters
-----------------------------------------------------------------------------------
bin                     results
observation             940
-----------------------------------------------------------------------------------
bin                     results         results         results         results         results         results         results
process                 signal          ZJetsToNuNu     WJetsToLNu      Top             QCD             DYJetsToLL      Diboson
process                 0               1               2               3               4               5               6
rate                    5.95086         436.9           337.1           11              69.7            4.5             11.5
-----------------------------------------------------------------------------------
backgUncert     lnN     -               1.047837        1.0545832       1.301818        1.119799        1.471111        1.294783

• Now we show the same plots, but including observed limits. Since the signal is not optimized because we are using the standard monojet selections, the number of observed events (940) is higher than the total number of expected events (871), leading to observed limits (30 - 40 pb) higher than the expected median limits (20 - 28 pb).

• Figure on the letf (right) shows theoretical cross-sections for Axial Vector (Vector) Dark Matter. It is considered Lumi = 5 fb^-1. It takes into account the standard Monojet selection above. EFT signal samples have from 11 to 14% of efficiency. As expected, the increasing in luminosity did not bring any significant improvement. Here is the data card for Axial-Vector with DM mass = 1000 GeV using Lumi = 5 fb^-1:

Show Hide

imax    1       number of channels
jmax    6       number of backgrounds
kmax    1       number of nuisance parameters
-----------------------------------------------------------------------------------
bin                     results
observation             0
-----------------------------------------------------------------------------------
bin                     results         results         results         results         results         results         results
process                 signal          ZJetsToNuNu     WJetsToLNu      Top             QCD             DYJetsToLL      Diboson
process                 0               1               2               3               4               5               6
rate                    743.3           54571.6         42105.9         1373.97         8705.97         562.078         1436.42
-----------------------------------------------------------------------------------
backgUncert     lnN     -               1.047837        1.0545832       1.301818        1.119799        1.471111        1.294783

• Figure on the letf (right) shows theoretical cross-sections for Axial Vector (Vector) Dark Matter. It is considered Lumi = 5 fb^-1. It takes into account the standard Monojet selection above + E_T^missing > 500 GeV. EFT signal samples have from 1.5 to 3% of efficiency. The high E_T^missing cut improves the expected limits by 1 order of magnitude. The limits are between 0.3 and 0.6 pb, being compatible with the 0.4 pb found in the Search for New Physics in Monojet Final State (AN-15-072).

Limits considering i2HDM model

• These figures show limits considering the i2HDM model (with different λ₃₄₅ for each mass point). It is considered:
  • left figure: Lumi = 40.03 pb^-1 and standard Monojet selection, which leads to i2HDM signal samples with efficiencies from 17 to 21% and expected limits around 15 pb;
  • right figure: Lumi = 5 fb^-1 and standard Monojet selection + E_T^missing > 500 GeV, which leads to i2HDM signal samples with efficiencies from 0.5 to 1.1% and expected limits around 1 pb.

Recent studies about limit calculations is in ThetaLimits.

-- assantos - 2015-07-24

Comments