Efficient Identification of Boosted Semileptonic Top Quarks at the LHC

TL;DR

This paper develops and tests new methods to efficiently identify highly boosted semileptonic top quarks at the LHC, improving signal discrimination without relying on b-tagging or missing energy.

Contribution

It introduces novel isolation variables and simple cuts that enhance the detection of boosted semileptonic top quarks, aiding new physics searches at the LHC.

Findings

Isolation variable can reject light QCD jets with embedded muons at 10^3 level.

Simple pT-scaled  R cut effectively discriminates W-strahlung background.

Baseline cuts enable potential discovery of KK gluons up to 4.5 TeV.

Abstract

Top quarks produced in multi-TeV processes will have large Lorentz boosts, and their decay products will be highly collimated. In semileptonic decay modes, this often leads to the merging of the b-jet and the hard lepton according to standard event reconstructions, which can complicate new physics searches. Here we explore ways of efficiently recovering this signal in the muon channel at the LHC. We perform a particle-level study of events with muons produced inside of boosted tops, as well as in generic QCD jets and from W-strahlung off of hard quarks. We characterize the discriminating power of cuts previously explored in the literature, as well two new ones. We find a particularly powerful isolation variable which can potentially reject light QCD jets with hard embedded muons at the 10^3 level while retaining 80~90% of the tops. This can also be fruitfully combined with other cuts for O(1) greater discrimination. For W-strahlung, a simple pT-scaled maximum \Delta R cut performs comparably to a highly idealized top-mass reconstruction, rejecting an O(1) fraction of the background with percent-scale loss of signal. Using these results, we suggest a set of well-motivated baseline cuts for any physics analysis involving semileptonic top quarks at TeV-scale momenta, using neither b-tagging nor missing energy as discriminators. We demonstrate the utility of our cuts in searching for resonances in the top-antitop invariant mass spectrum. For example, our results suggest that 100 fb^{-1} of data from a 14 TeV LHC could be used to discover a warped KK gluon up to 4.5 TeV or higher.