Slides for presentation:


New Atlantis displays of low pt muon and high pT muon. Show clear bending in xy in ID. The track stops in the ID, but one can see calorimeter and muon spectrometer hits in the lower part of the detector as well. More difficult to see bending in the Muon spectrometer (right below).

xy view
high pT muon


Used all the data08.cosmag*IDCosmic.recon.ESD* files available. Found total of 3.294.471 events. For the MC I used valid2.108867.CosSimIDVolSolOnTorOn.recon.ESD.s540_d167_r676 with 7.397.860 events.

Plotted the TRT high to low treshold ratio vs. E/p for muon bremsstrahlung candidates (MBC) and electron ionisation candidates (EIC) defined as below. The plot shows clearly that more of the EIC is at higher values for TRT ratio and E/p.

TRT high/low treshold hits vs. E/p
blue lines showing the standard ATLAS elec cuts

Table below show the percentage for the two groups in the two regions.

Object TRT<=0.08 && E/p<=0.8 TRT>0.08 && 0.8< E/p >2.5
EIC 0.614286 0.385714
MBC 0.977242 0.0227577


First electron seen by me in ATLAS. In 740236 event I found 70 electrons after using the following cuts (inspired by the egamma group cosmic analysis).

  • Energy in cluster > 3GeV
  • Match in phi between track and cluster: -0.02 < phi < 0.02
  • At least two ID tracks reconstructed
  • At lest 25 TRT hits

E/p for my electron candidates

If I keep all the cuts above, but only require one ID track, this will give me a sample with muon bremsstrahlung candidates since the photon created will not give any track. When I want to have an electron I must have two tracks since the muon and the electron created from the ionisation will have one track each. Below is the TRT high/low treshold ratio for these two groups (Muon bremsstrahlung candidates and Electron ionisation candidates):

Produced data from 95 data08 samples. Want to find electrons by using the following variables:

Track match in Phi between cluster and track
Number of TRT hits
Ratio of cell energies in 3x7 vs. 7x7 cell blocks in the EM calorimeter
Number of TRT hits


Try to pick out object after the following criteria:

  • events with high-level trigger (L2) track candiidate reconstructed online in the barrel inner detector (IDCosmic)
  • filtering for EM cluster candidates with transverse energy above 3GeV and with a loose track match in phi
  • they are required to satisfy the medium electron cuts (see here).
  • If two ID barrel tracks are reconstructed: electron ionisation candidate
  • If one ID barrel track is reconstructed: muon bremsstrahlung candidate


x vs. y plot with and without EC hits. See that the plot without any EC hits is cleaner (as expected).

without EC
with EC


Animated gif of the muons going through the Inner Detector can be found here.


To see how the muons are going through the detector I have plotted 666 tracks going through some part of the ID volume. This shows really that the cosmics are coming in all possible angels (from above). The picture is a bit messy for 666 tracks so I show also a picture with only 63 tracks. What is even more fun is to sit and watch when the tracks appear. Almost like being at point 1 taking data. smile

666 tracks
63 tracks

Got the p from the Electron track particle. Calculated the E/p for electrons. Most of the electron have E/p close to zero, the distribution does not have any peak around 1 as expected. This indicate that most of the electron picked up by the egamma algorithm (put in the ElectronCollection) are crap. However I use the ATLAS cuts for electrons for the TRT high to low-treshold ratio [>0.08] and E/p [0.8<E/p<2.5] to define good electrons (see plot to the right below). Out of 241 electrons, only 13 of them seems to pass these cuts. This is expected since the number of electrons expected in ATLAS from cosmic rays is extremely small. Plan is to pick out one of the nice electrons and look at it in ATLANTIS.

E/p for electrons
Ratio TRT high to low treshold vs E/p for electrons


The (*elecItr)->detailValue(egammaParameters::EoverP is obsolete and removed from 15.3.0, should be replaced by hand-made calculation of E(cluster)/p(track). For release 15.3.1 it returns -999. Need to get the momenta from the electron track.


Atlantis display

low pT muon (15 and 9 GeV)
high pT muon (45 and 32 GeV)

Phi, CosTheta and Eta of muon tracks

cos Theta (ID)
cos Theta (SPEC)
eta (ID)
eta (SPEC)
phi (ID)
phi (SPEC)

DeltaR between Muon Spectrometer tracks and ID tracks

SPEC (y>0) - ID
SPEC (y<0) - ID
SPEC (y>0) -SPECL (y<0)

pT between

If a track start at y>0: compare last pT (position 1) with first surface in ID (position 2).
If a track starts at y<0: compare last pT in ID (position 3) with first surface in SPEC (position 4)
Would expect tracks that belong to each other (same muon) would make a peak at some energy (compare with Magnar). While tracks from different muons would be smeared out.
SPEC (y>0) - ID first
SPEC (y<0) - ID last
ID first - ID last
SPEC (y>0) last - SPEC (y<0) first

Track information

number of SPEC trks vs. ID trks
number of different tracks


Atlantis display


pT, eta, phi


--++ Muons vs electrons

elec E/p
muon E/p
TRT high/low treshold hits vs E/p for electrons
TRT high/low treshold hits vs E/p for muons

-- EirikGramstad - 2009-09-11

Edit | Attach | Watch | Print version | History: r19 < r18 < r17 < r16 < r15 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r19 - 2010-10-11 - EirikGramstad
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback