Probing TeV scale top-philic resonances with boosted top-tagging at the high luminosity LHC

We investigate the discovery potential of singly produced top-philic resonances at the high luminosity (HL) LHC in the four-top final state. Our analysis spans over the fully hadronic, semileptonic, and same-sign dilepton channels where we present concrete search strategies adequate to a boosted kinematic regime and high jet-multiplicity environments. We utilize the template overlap method with newly developed template observables for tagging boosted top quarks, a large-radius jet variable $M_J$, and customized $b$-tagging tactics for background discrimination. Our results show that the same-sign dilepton channel gives the best sensitivity among the considered channels, with an improvement of significance up to 10%–20% when combined with boosted top-tagging. Both the fully hadronic and semileptonic channels yield comparable discovery potential and contribute to further enhancements in the sensitivity by combining all channels. Finally, we show the sensitivity of a top-philic resonance at the LHC and HL-LHC by showing the $2\sigma$ exclusion limit and $5\sigma$ discovery reach, including a combination of all three channels.

Figure 1

The tree-level single production of the $V_1$ decaying into the fully hadronic (left), semileptonic (middle), and same-sign-dileptonic (right) channels. As all three channels exhibit disjoint final states, we set a different search strategy on each channel and estimate its sensitivity at the HL-LHC. The top and antitop pairs from the resonance decay are highly boosted, and they are characteristically distinguished by the other two nonboosted spectator tops.

Figure 2

$p_T$ (left) and $\eta$ (right) distributions of the four tops at the parton level for a benchmark point, $M_{V_1}=1.5\mathrm{TeV}$ and $c_t=2.0$.

Figure 3

Parton level production cross section of $pp\to t\overline{t}{V}_1\to t\overline{t}t\overline{t}$ (in fb, black-solid lines) at $\sqrt{s}=14\mathrm{TeV}$ in the $M_{V_1}-c_t$ parameter space. The red solid curve [obtained from $c_t^2/(2M_{V_1}^2)=5.0{\mathrm{TeV}}^{-2}$] represents current ATLAS bounds from 13 TeV LHC with $3.2{\mathrm{fb}}^{-1}$ of data [51]. Projected bounds for $300{\mathrm{fb}}^{-1}$ ($3000{\mathrm{fb}}^{-1}$) may be obtained via a naive rescaling and are shown in dashed (dotted) curves.

Figure 4

$O{v}_3^t$ distributions of the hardest $R=0.7$ fat jet with (top panel) $r_{\mathrm{sub}}=0.1$ and (bottom panel) $r_{\mathrm{sub}}=0.15$ for benchmark event samples: the Standard Model semileptonic $t\overline{t}$ process without additional jets and $jZ(\to \nu \overline{\nu })$.

Figure 5

(left) For a typical top-faking QCD jet, there is a $p_T$ hierarchy between the collinear jets, while (right) for a boosted top, its decay products share a symmetric $p_T$ scale with a relatively large angular separation. We can implement the difference of these features into the template overlap method by introducing a new measure $ty$ in Eq. (3.1).

Figure 6

$ty$ distributions of the hardest top-tagged $R=0.7$ fat jet with (top panel) $r_{\mathrm{sub}}=0.1$ and (bottom panel) $r_{\mathrm{sub}}=0.15$.

Figure 7

Invariant mass distributions of the (top panel) first and (middle panel) second hardest top jets after the basic cuts and the ditop selection. The 1.5 TeV $V_1$ resonance is then reconstructed using the boosted ditop system in the bottom panel.

Figure 8

Various distributions of (top panel) a number of isolated jets after basic cuts and ditop selection, (middle panel) an invariant mass of three isolated jets which minimizes ${\chi }^2$ defined in Eq. (4.4), and (bottom panel) $M_J$ scalar sum of the masses of large radius ($R=1.5$) jets.

Figure 9

The $b$-tag scores of (upper-left panel) the first and (upper-right panel) the second hardest top jets, (lower-left panel) the reconstructed spectator top, and (lower-right panel) isolated $r=0.4$ jets from the first two top jets (i.e., $\mathrm{\Delta }{R}_{\mathrm{j},{\mathrm{t}}_1,2}>1.1$).

Figure 10

The panels in the first row show $ty$ distributions of the (left) first and (right) second hardest top jets, and the panels in the second row represent the respective invariant mass distributions after basic Cuts and the ditop selection.

Figure 11

Invariant mass distributions of the (top panel) hardest top jet $t_{\mathrm{had}}$ and (middle panel) reconstructed leptonic top jet $t_{\mathrm{lep}}$ after basic cuts and the ditop selection. The 1.5 TeV $V_1$ resonance is reconstructed using the boosted ditop system in the bottom panel.

Figure 12

Distributions of (left panel) a number of jets isolated from $t_{\mathrm{had}}$ and $t_{\mathrm{lep}}$ by $\mathrm{\Delta }{R}_{\text{j},{\mathrm{t}}_{\text{had},\text{lep}}}>1.2$, and (right panel) $M_J$ scalar sum of the masses of large radius ($R=1.5$) jets. Basic cuts and the ditop selection are applied for both of the cases.

Figure 13

Required luminosities in ${\mathrm{fb}}^{-1}$ of the combined fully hadronic, semileptonic, and SSDL channels for $5\sigma$ discovery for HL LHC run at $\sqrt{s}=14\mathrm{TeV}$.

Figure 14

Required luminosities in ${\mathrm{fb}}^{-1}$ of the combined fully hadronic, semileptonic, and SSDL channels for (left) $2\sigma$ exclusion for HL LHC at $\sqrt{s}=14\mathrm{TeV}$. We also show current (red, solid line for $3.2{\mathrm{fb}}^{-1}$) and projected (dashed line for $300{\mathrm{fb}}^{-1}$ and dotted line for $3000{\mathrm{fb}}^{-1}$) bounds with the combined results.