wiki:MA5LongLivedParticle

MadAnalysis 5 LLP (Long Lived Particle)

LLP with official Delphes 3.4.1

No pile-up simulation

The Delphes package considers only the final state (the mother particles are only used in tagger modules). Therefore Delphes considers only the decay product of the LLP. The daughter particles (neutral or charged) are correctly propagated to the calorimeters and they are treated like prompt particles (same efficiency, same resolution).

Taking the example of a LL neutralino decaying into leptons, the corresponding leptons are reconstructed like prompt leptons by Delphes. The simulation is wrong.

With pile-up simulation

What is the effects of the pile-up subtraction procedure to a LLP particle? To be studied into details.

LLP with special Delphes package for MA5

Delphes-LLP package is a special improvement of the Delphes package. Associated to MadAnalysis 5 v1.6, it allows the user to perform phenomenological investigations with long-lived particles or to recast existed analysis devoted to this kind of signature.

Current status: the package handles neutral LLPs which decay into leptons in the first layers of the CMS tracker (such as neutralino in the framework of RPV-SUSY). Realistic efficiency is applied to the tracks and D0 parameter is accessible in MadAnalysis 5.

Package[update 2017/11/06]: ma5track_v1.6.tgz

Changes in the Delphes dataformat:

for leptons:

  • D0: impact parameter in the tranverse plane wrt the closest approach point
  • DZ: impact parameter in the longitudinal plane wrt the closest approach point
  • Xd, Yd, Zd: coordinates of the closest approach point
  • Tp, Xp, Yp, Zp: coordinates of the vertex where the lepton is coming from (MC Truth)

New modules in Delphes:

  • MA5EfficiencyD0: applying efficiency to tracks according to the D0 and DZ parameters

Efficiency functions is coming from 8-TeV tracking performance of CMS, describing in the following wiki (only D0 variable is used): https://twiki.cern.ch/twiki/bin/view/CMSPublic/DisplacedSusyParametrisationStudyForUser

set EfficiencyFormula { (d0<=20) * (-5.06107e-7 * d0**6 + 0.0000272756 * d0**5 - 0.00049321 * d0**4 + 0.00287189 * d0**3 + 0.00522007 * d0**2 -0.0917957 * d0 +  0.924921) +
                        (d0>20) * (0.00) }

Changes in the MA5 dataformat:

D0:
event.mc()->muons()[i].d0();
event.mc()->electrons()[i].d0();

DZ:
event.mc()->muons()[i].dz();
event.mc()->electrons()[i].dz();

Closest approach point:
event.mc()->muons()[i].closestPoint().X()
event.mc()->muons()[i].closestPoint().Y()
event.mc()->muons()[i].closestPoint().Z()

Displaced vertex:
event.mc()->muons()[i].vertexProd().X()
event.mc()->muons()[i].vertexProd().Y()
event.mc()->muons()[i].vertexProd().Z()

Recast experimental analyses

Ongoing development

  • Next development: the package will handle charged LLPs which decay into leptons in the first layers of the CMS tracker (such as neutralino in the framework of RPV-SUSY). To reach this goal, the magnetic field of the detector must be applied to the charged LLP before decaying in order to curve the trajectory. This effect will be described by the MA5LLParticlePropagator module.
Last modified 3 years ago Last modified on 11/07/17 11:13:12