Stop Reconstruction with Tagged Tops

TL;DR

Problem: Stop squark searches at the LHC face severe combinatorics and overwhelming tt backgrounds in semi-leptonic channels. Approach: The paper first assesses a standard semi-leptonic analysis and then develops a hadronic fat-jet strategy using a Cambridge/Aachen-based top tagger (HEPTopTagger) to identify two boosted hadronic tops and reconstruct their momenta. Findings: Semi-leptonic channels are unlikely to yield a discovery, while the two-top hadronic approach yields S/B around unity and discovery-level significance with 10 fb^-1, and provides a handle to measure the stop mass via m_T2 endpoints. Significance/Impact: The work demonstrates that boosted-top tagging can unlock a viable path to discovering and characterizing top partners at the LHC and points to broad utility of such taggers in multi-top, high-m multiplicity final states.

Abstract

At the LHC combinatorics make it unlikely that we will be able to observe stop pair production with a decay to a semi-leptonic top pair and missing energy for generic supersymmetric mass spectra. Using a Standard-Model top tagger on fully hadronic top decays we can not only extract the stop signal but also measure the top momentum. To illustrate the promise of tagging tops with moderate boost we include a detailed discussion of our HEPTopTagger algorithm.

Figures (6)

• Figure 1: Normalized $m_{T2}$ distributions for the stop signal ($m_{\tilde{t}} = 340$ GeV) and the $t\bar{t}$ background, after reconstructing two (real of fake) hadronic top quarks. The hypothetical LSP mass we set to $m_{\tilde{\chi}^0_{1}} = 0$ GeV (left) or to the correct value of $m_{\tilde{\chi}^0_{1}} = 98$ GeV (right).
• Figure 2: Left: partonic $\Delta R_{bjj}$ vs $p_T$ distribution for a Standard Model $t\bar{t}$ sample. Right: the same correlation, but only for tagged top quarks and based on the reconstructed kinematic properties.
• Figure 3: Distribution of all events in the $\arctan m_{13}/m_{12}$ vs $m_{23}/m_{123}$ plane. We show $t\bar{t}$ (left). $W$+jets (center) and pure QCD jets (right) samples. More densely populated regions of the phase space appear in red.
• Figure 4: Number of tagged tops in the $t\bar{t}$, QCD jets and $W$+jets channels for a varying assumed top mass in the tagging algorithm. The actual top mass in the sample is 172.3 GeV. Shown is the number of tagged first (left) and second (right) tops.
• Figure 5: Left: number of tops, tagged tops, mis-tagged tops from QCD jets and for $W$+jets for $1~{\rm fb^{-1}}$. Center: fraction of hadronic tops whose main parton-level decay products are within a C/A distance of $\Delta R_{bjj}=1.5, 1.2, 0.9$. Right: tagging efficiencies, normalized to the top line of the central panel.