Reinterpretation of searches with LLPs

Interested people: Andre Lessa, Sezen Sekmen, Jose Zurita, Sabine Kraml, Andrea Coccaro, Zhen Liu

Guidelines for reinterpretation material for LLP searches were discussed based on the following two analyses:

• ATLAS-SUSY-2018-42: dE/dx - A significant amount of reinterpretation material was provided, however it is not sufficient to fully reproduce the analysis. Additional information required for recasting and validation:
  • Information for all search regions, not only for discovery.
  • Separate efficiencies for C1N1 and C1C1 models for the EWKino case.
  • Exclusion curve using only de discovery regions.
• CMS-PAS-SUS-21-006: disappearing track - Newly published. Reint material in the making. Information to be provided and further studies:
  • Full analysis implementation in ADL. Input format Delphes.
  • Disappearing track efficiencies vs. truth level LLP decay length, perhaps pT, eta. Efficiencies provided as tables within the ADL file, and are emulated via hit-and-miss.
  • dE/dx values provided vs. p and mass, or perhaps p/m with resolution. Also provided as tables within the ADL file.
  • Cutflows for selected benchmarks.
  • Efficiency maps for all 49 bins in all SMSs as text tables / ROOT trees. Separate efficiencies for C1N1 and C1C1 models for the EWKino case.
  • Optional study:
    • Check disappearing track tagging / dE/dx behavior in LLPs other than charginos.
    • Verify tagging efficiency as a function of the visible LLP daughter pT (pion/lepton).
    • Study complementarity between disappearing track searches and searches for displaced leptons.

LLP efficiencies mainly applied on GEN objects and their daughters. Usually done through tracing daughters in hepmc or Delphes GEN information. For the Delphes case, including the whole GEN collection leads to extremely large files, and tracing daughters can be tricky (due to radiation and showering).

Propose having a module that filters LLPs and their daughters from the Delphes GEN branch and saves them in the Delphes output. A simple implementation by Andre Lessa can be found here (HSCPFilter module). Plan to make it available for public use, and further, discuss with the Delphes team the possibility to make it a recognized part of Delphes releases. Disappearing track ADL will be built based on this output.

Any models to be tested ?