Regarding Particle-Flow and Jet Formation from Displaced Vertices

[Samuel Alipour-fard]

Dear Delphes team,

I have been performing analyses with long lived dark particles which decay to dijets at displaced vertices, and have been running into a bit of an issue with jet reconstruction.

In particular, it seems that tracks originating at different displaced vertices often hit the HCAL in similar locations. As a result, the standard jet algorithms identify them as being in the same jet, whereas with the track information it's very obvious by eye that they should not be. Unfortunately, this significantly disrupts my ability to accurately simulate background rates for my search.

I would have expected Delphes particle flow algorithms to be perfectly suited to solve this issue. However, when I implement particle flow jet reconstruction it doesn't seem to behave much differently than the other algorithms with respect to catching tracks from different vertices in the same jets. Explicitly, I have implemented particle flow reconstruction by following the instructions on ticket #903, and have changed the input of my FastJetFinder to "set InputArray EFlowMerger/eflow". I have also played around with both GenJet and Jet objects to no effect.

Is there some subtlety to using the particle flow jet reconstruction that I may be missing? I noticed that in Section 3.2.1 of the Delphes 3 manual, it seems like the ParticleFlow jets might be different objects than the Jet and GenJet objects I've been playing around with, but I haven't noticed any other branches in ROOT that would contain them.
On the other side, if you would expect that ParticleFlow wouldn't be well-suited to this task, I would be curious to know if there were any other methods within Delphes that could be useful for accurately reconstructing jets coming from displaced vertices.

Thanks very much for your time.

Best,
Sam