Update 05.18.2018 -- Run 9 embedding: Jet Reconstruction Efficiency (R = 0.3, Charged)

We need to know the jet reconstruction efficiency to correct the unfolded measured spectrum, so this is a quick study of the reconstruction efficiency for R = 0.3 charged jets.  There are essentially two dimensions here to play around with: which embedding configuration to use (FF vs. RFF), and which jets to select (recoil, off-axis, all, etc.).  In all situations, the QA cuts applied to the particle- and detector-level jets are the same:

  • areaJet > 0.2
  • |etaJet| < 0.7
  • pTjet > 0.2 GeV/c

The 'pTjet' above refers to the raw reconstructed jet pT, no corrections have been made for the UE or JES.  The efficiencies below are presented as a function of particle-level 'pTcorr = pTjet - rho*areaJet'.  To select different populations of jets, I varied the "recoil" condition.  There were three cases:

  • |dFjet - pi| < pi/4: this is our typical condition.  It selects jets roughly back-to-back with the trigger.
  • |dFjet - pi| < pi/2: this selects all jets in the hemisphere opposite the trigger.
  • |dFjet - pi| < pi: this selects everything.

Then, I checked the FF and RFF configurations for each case separately.  Here are the RFF efficiencies:

Here are the FF efficiencies:

I then combined the RFF and FF configurations: we saw previously that the response between the two configurations is consistent (see my posts on 05/13/2018 and 05/14/2018).

I also compared the RFF, FF, and RFF+FF efficiencies to each other for each population of jets:

As expected, the RFF and FF efficiencies are consistent with each other.  I was also curious about how our trigger condition (charged hadrons with 'pTtrg = 9 - 20 GeV/c' and '|etaTrg| < 0.9' after all of our usual track QA cuts) impacted the efficiency, so I made a "min bias" population of jets where I let the trigger 'pT' be anything (0 - 100 GeV/c) and the trigger eta be anything (< 2) but still required that the reconstructed vertex satisy '|Vz| < 55' and '|Vr| < 2'.  I also varied the jet population in a similar manner as above, but -- of course -- here it doesn't mean anything since the trigger can be anything.

The "min bias" FF efficiency looks weird, so I may have made a mistake when I was producing the "min bias" FF sample...

Either way, I've attached a zip file with all the ROOT files, and plots of all of the individual efficiencies; the detector-level spectra plotted in those are not smeared.  The plotted "detector-level" 'pTcorr' distribution is the corresponding particle-level 'pTcorr' for each detector-level jet that was successfully matched.