Beam Halo in CMS



Machine-induced secondary particles are produced in showers which are initiated by collisions of the beam with residual gas inside the LHC vacuum chamber or by interactions of the beam halo with limiting apertures. In addition to these scattering processes, charged particles are deflected by the magnetic field of the beam optics. This so-called beam halo particles are one of the main sources of beam background at the LHC. In CMS we consider only secondary particles that are generated outside the CMS cavern. For example proton-beam gas collisions that directly occur in the active detector volume, so called beam gas events, are not included in our simulation.

CMS Notes related to this topic

Talks related to this topic

Forums and hn's

Even if Beam Halo particles are not asynchronous particles, the Beam Halo discussion is ongoing on the:

Beam Halo People

  • Marco Cardaci Group
  • Dominique Mangeol Group
  • Michael Tytgat Group ???
  • MS, vecbos+j ratio team Group ???
  • ???

BeamHaloGenerator in CMSSW

BeamHaloGenerator short manual

You can find an extensive manual on CMS IN-2005/004. Here a quick primer on the use of the parameters appearing in GeneratorInterface/BeamHaloGenerator/test/test.cfg :

  • GENMOD = 1 give the possibility to produce beam halo particles from a parametrization of the file produced with MARS(?) ( /afs/ ), a machine simulator.
  • GENMOD = 2 give the possibility to produce beam halo particles direcly from the MARS(?) file.
  • GENMOD = 3 only available for 1_8_X onward, give the possibility to produce beam halo particles from two MARS files ( /afs/ and /afs/ ), one for the proton-beam gas interaction along the beam and the other for the tails effects.
  • LHC_B1 = 1/0 and LHC_B2 = 1/0 give the possibility to switch on/off (1/0) the LHC beams. Notice: The new MARS files contain beam 2 description only. So, if you wish to use the GENMOD = 3, please switch on only beam 2.
  • IW_MUO = 1/0 and IW_HAD = 1/0 give the possibility to switch on/off (1/0) the use of halo hadrons and halo muons.
  • shift_bx = n give the possibility to switch in time of n bounch spacing the halo particles in order to get in-time and out-of-time beam halo contributions. The use of this parameter is deprecated.
  • BXNS = 25. bounch spacing. Default is 25. ns.

From 1_8_1 onward you will find the files:

  • GeneratorInterface/BeamHaloGenerator/test/BeamHalo_mixing_example_180.cfg
  • GeneratorInterface/BeamHaloGenerator/test/BeamHaloMixing_NoSignal180.cfi

If you want to mix beam halo with pile up or comsics, you will need to modify SimGeneral/MixingModule/data/mixLowLumPU_4sources.cfi and GeneratorInterface/BeamHaloGenerator/test/BeamHaloMixing_NoSignal180.cfi which will have to point to the appropriate source files and then finally run cmsRun BeamHalo_mixing_example_180.cfg

In 1_6_11 you will find:

  • GeneratorInterface/BeamHaloGenerator/test/BeamHalo_mixing_example.cfg
  • GeneratorInterface/BeamHaloGenerator/test/BeamHaloMixing_Signal.cfi
  • GeneratorInterface/BeamHaloGenerator/test/BeamHaloMixing_NoSignal.cfi
  • GeneratorInterface/BeamHaloGenerator/test/mix_4sources.cfi

If you want to mix beam halo with pile up or comsics, you will need to modify mix_4sources.cfi and BeamHaloMixing_Signal.cfi which will have to point to the appropriate source files and then finally run cmsRun BeamHalo_mixing_example.cfg

BeamHaloGenerator in LXR

The BeamHaloGenerator and the MixingModule has been modified by Ursula Berthon, Claude Charlot and Dominique Mangeol in 1_6_11 onward, 1_8_1 onward and 2_0_0 onward, in order to give the possibility to mix Pile Up, Cosmics and Beam Halo from the two beams, have a look:

The BeamHaloGenerator has been modified by Nadia Erin Adam and Dominique Mangeol in 1_8_X and 2_X_Z in order to give the possibility to use beam-gas and tails files from MARS Machine simulation, have a look for instance at 1_8_4:

How to run in 5 seconds BeamHaloGenerator

  • For CMSSW_1_8_4 from your afs area copy and paste the following commands:
scramv1 project CMSSW CMSSW_1_8_4
cd CMSSW_1_8_4/src
eval `scramv1 runtime -csh`
cvs co -r CMSSW_1_8_4 GeneratorInterface/BeamHaloGenerator
cd GeneratorInterface/BeamHaloGenerator
scramv1 b
cd test
rm bgas.123
ln -s /afs/ bgas.123
rm tails.1
ln -s /afs/ tails.1
rm input.txt
ln -s /afs/ input.txt

At this point you have to choose the cfg file to run either: cmsRun test.cfg or cmsRun test_mars_beamgas.cfg or cmsRun test_mars_beamtails.cfg or cmsRun BeamHalo_mixing_example_180.cfg

  • For CMSSW_1_6_11 from your afs area copy and paste the following commands:
scramv1 project CMSSW CMSSW_1_6_11
cd CMSSW_1_6_11/src
eval `scramv1 runtime -csh`
cvs co -r CMSSW_1_6_11 GeneratorInterface/BeamHaloGenerator
cd GeneratorInterface/BeamHaloGenerator
scramv1 b
cd test
rm input.txt
ln -s /afs/ input.txt

in this case you can run cmsRun test.cfg or cmsRun BeamHalo_mixing_example.cfg.

Beam Halo Samples

Privately produced samples

- Beam Halo particles samples, each of 10,000 events, made with GENMOD = 1, LHC_B1 = 1, LHC_B2 = 0, EG_MIN = 10., EG_MAX = 7000., for muons, hadrons and different HF wires thresholds and made with stardard .cfg's for gene-simu-digitization and gene-simu-digi-reconstruction in CMSSW_1_4_3, CMSSW_1_4_5, CMSSW_1_5_0_pre4, CMSSW_1_5_1 are in /castor/ The use of these samples produced only for purpose of testing is somehow deprecated.

- Beam Halo samples produced in CMSS_1_6_7 are in /castor/ among them, 500K-evts-beam-halo files at GEN-SIM level:


with obvious meaning of the name covention. Also 10K-evts files with different random-number-seeds in the same castor directory:


where X = 1 or 2, Y = Muons or Hadrons and Z = from 1 to 50.

- Beam Halo samples produced in CMSS_1_8_4 with beam halo particles from LHC_B2 are in /castor/ :


Officially produced samples

Beam Halo from DBS : Only the in-time samples should be taken into account for use. Those off-time have not been actually shifted.

Beam Halo Production Plans




TODO Implement the possibility to have multiple sources in CRAB. eek!

TODO Write plans on how to collect beam halo from real data Edit topic

Minimal scenario production

  • GEN-SIM in 1_6_7 beam halo (muons/hadrons)*(beam1/beam2) 500 K evts samples in /castor/ DONE smile TOTAL 2M evts
  • DIGI-RECO with CMSSW_1_6_11 the GEN-SIM samples with time shift from bx=-3; -2; -1; +1; +2;+3;+4;+5 (50K evts each) TOTAL 400 K evts
  • DIGI-RECO with CMSSW_1_6_11 the GEN-SIM samples for the whole (-3;5) bx window for 1 beam (100k) and for 2 beams (100k) for hadrons and for muons. TOTAL 200 K evts

To be mixed with GEN-SIM beam1+beam2 for hadrons and muons:

  • QCD 0<Pt_hat<600GeV (12 bins) 12*50K(30K?) events (the upper Pt_hat limit could certainly still be decreased)
  • Wenu+Jets 0<Pt_hat<600GeV (12 bins) 12*50K events (true MET; I don't think there is 50k events in each bins, it is more like 30k)
  • Wmunu+Jets 0<Pt_hat<600GeV (12 bins) 12*50K events (same comment)
  • Wtaunu+Jets (inclusive samples) 100K
  • Znunu+Jets 0<Pt_hat<600GeV (12 bins) 12*50K events
  • Zee+Jets 0<Pt_hat<600GeV (12 bins) 12*50K events (extra clean samples, if something is coming on top of it we should see it)
  • Zmumu+Jets 0<Pt_hat<600GeV (12 bins) 12*50K events (same comments)
  • ttbar+0jets (Alpgen) 200K
  • Mono-jet M_D = 5TeV, n = 2 100K (the sample now exist at GEN level)DONE smile

USE: With these samples, we could do some basic MHT/MET studies and some first approach of MET and Jet cleaning. We would not be able to understand the influence on level-1 trigger rate (because of the too low stat for QCD)

Extended scenario production

  • For the bins between 0<Pt_hat<300GeV about 300K/bin (9 bins) for the other bins 150K/bin (5 bins) (More for the QCD samples, so we can estimate the QCD trigger rate. On the side of QCD that would require more statistics and Pt_hat bins up to 1TeV. This is based on some rate calculation performed with Spring07 QCD data which had relatively small statistics)
  • Wtau nu full stat
  • ttbar -> increase the stat to 1M
  • LM1 -> 100K (or full stat)
  • LM5 (? I am not sure of the number, the one which has low MET and low Jet Et) full stat
  • One could add also a few QCD samples with both Halo and Pile-Up to have an idea where we are going.

USE: These samples could be useful for checking effects of MET cleaning; rate efficiency comparison etc...

Other issues related to production

  • We should consider getting the MARS simulation program with the CMS cavern setup, so we can make samples ourselves.
  • What we need to do for the DataMixingModule in order to be able to overlay BH from real data to MC signals is the same thing which has been done for the MixingModule (not really complicated)
  • Notice for the 10 TeV run: The most urgent, is rather to get something for the start-up. (moving from 7 to 5 or the inverse should be rather easy) The particle spectrum is hardcoded into the generator...
  • We don't have something for the start up , we should ask MIB people to have something which could be used to get a new spectrum. With the current beam halo, we assume to run (properly) at 5TeV, that would mean to slightly shift lower value the energy spectrum which is used in the generator to emulate the spectrum of the halo particles, but that would not be for the start-up...

Beam Halo Cleaning Plans


Quoting Dijet Group Wiki

  • Step 7: Data Cleanup

Must remove events with high pT jets from comic rays, beam-halo, noise. The distribution of MET/SumEt and MHT/SumHT needs to be studied in the real data to find an efficient cut point to remove beam-halo/cosmics/catasetrophic noise. Distributions of jet Num Towers, EMF, Num Tracks, CHF, eta, phi, Pt, dijet mass, dijet delta-phi, and others in the Barrel need to be studied before and after the cleanup cut. This is part of a program of Jet ID in JetMET. Studies need to be done in Monte Carlo, Global Run data, and early collision data. The following cuts have been identified so far with >99% efficiency in Monte Carlo for QCD:

MET / SumET < 0.3
EM Fraction of Jet > 0.1
Number of CaloTowers in Jet >= 3
Number of Tracks in Jet >= 1


  • Need to implement in the software and to test it. Michael Tytgat Group ???

-- MarcoCardaci - 24 Apr 2008

Edit | Attach | Watch | Print version | History: r9 < r8 < r7 < r6 < r5 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r9 - 2008-04-29 - MarcoCardaci
    • 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