July exercise activities

We will be able to speed up our real heavy ion data analysis, if we prepare and practice analysis procedure before having heavy ion runs. July exercise is aimed to do real analysis with MC data as real data. With this exercise, researchers would know what is missing, what is needed, how to do and efficiently organize jobs in their PInG (Physics Interest Group). For di-lepton PInG, I take the jobs that 1) cross check results between official analyzer for July exercise and improved previous official analyzer and 2) have distributions of muons with July exercise data.

Used code and samples

• Used analyzer : DimuonAnalyzer (on the cvs : /UserCode/CmsHi/Z0Ana/DimuonAnalyzer)
  • This analyzer has verbose log (using std::cout) to see the daughter of Z0. To avoid this, open the test/dimuonanalyzer_cfg.py and put False at doZ0check.
• Used Hydjet MC sample : Hydjet MinBias + Z0 signal embedded (on the CERN castor : /castor/cern.ch/user/d/dmoon/cms370/Hydjet_MinBias_2.76TeV_Z0_Emb_Reco/)

HYDJET MC comparison

• Aim : Comparison of MC distributions with real July exercise data. This is a cross check to the group analyzer.
  • Cross check various distributions of single muons and dimuons
  • Using skimmed data files (RECO → skim0).
    • Skim0 configuration for di-lepton group : "skim0" selects events which have at least 1 muon. And it saves only those events into another RECO_skim0 file.
    • After skimming, size of data files are quite reduced. It makes analysis more simple and save T2 data storage.
• Related topic : MC comparison
• List of presentations
  1. Single and dimuon distributions for HardEnriched data / MinBias data / Hydjet MC comparison
• List of root files (produced by DimuonAnalyzer)
  • Below root files were made with older version of DimuonAnalyzer. The lastest version of DimuonAnalyzer produces more NTuple than below root files.
  1. Z0 embedded July exercise data
  2. Minbias July exercise data
  3. Z0 embedded hydjet MC
• In this study, HardEnriched and MinBias data are made by AMPT. They have sharp peak at Z0 mass region, but Hydjet MC isn't. Below histogram have double peaks. This gives 2nd topic to be investigated.

Mass shift in the HYDJET dimuon mass distribution

• Motivation : Why does peak's are appear at 60~80 GeV/c² when dimuon's p_t > 12 GeV/c? It seems Z0's mass peak is little bit shifted.
• Related topic : Mass Shift
• List of presentations
  1. 22 July dilepton meeting : GEN unlike dimuon mass plot distinguished by muon's mother in the Z0 embedding sample
  2. 28 July dilepton meeting : Why sometimes so energetic gamma come out from the muon? A muon lose its momentum a lot when it decays into very energetic gamma and muon. So dimuon's mass falls into 60~80 GeV/c².
  3. Daughter of Z0 and the lastest result
• Conclusions (updated at 1 September)
  • This is not the problem of muon reconstraction.
  • Dimuon mass peak at 60~80 GeV/c² come from Z0.
  • pgun only sample don't have this problem. So it seems it is a bug in the generator.

Gen muons in the Hydjet MC

• Notice : For example, 94+muon means dimuon's mothers are pdgID 94 and muon.
• Mother pdgID of single muons (descending order)
  • Hydjet MB + Z0 embedding : Z0, pid94, muon, B, D, charm or bottom baryons(Λ^-_c, Λ⁰_b, anti Σ⁰_b), J/psi, Υ, η
  • Hydjet MinBias : B, D, bottom baryons(Λ⁰_b, anti Λ⁰_b), J/psi, φ
• Daughter pdgID of Z0 : This picture shows two different decays of 94+muon. In the shape, upper value means pdgID and lower value means its p_t

Dimuon mass plot with GEN muon

• Black : total dimuon pairs
• Blue : dimuon pairs with its mother particles.
  • Numbers in the histogram (located in the right middle) is the entries of Blue dotted histogram.
• SLIDE : plots from various mothers
  • In the Hydjet MB sample, there are very few dimuons. Also, for dimuons, B and D are the only mother of dimuon.