Search for light long-lived particles in $pp$ collisions at $\sqrt{s}=13$ TeV using displaced vertices in the ATLAS inner detector

Abstract

A search for long-lived particles (LLPs) using 140 fb$^{-1}$ of $pp$ collision data with $\sqrt{s} = 13$ TeV recorded by the ATLAS experiment at the LHC is presented. The search targets LLPs with masses between $5$ and $55$ GeV that decay hadronically in the ATLAS inner detector. Benchmark models with LLP pair production from exotic decays of the Higgs boson and models featuring long-lived axion-like particles (ALPs) are considered. No significant excess above the expected background is observed. Upper limits are placed on the branching ratio of the Higgs boson to pairs of LLPs, the cross-section for ALPs produced in association with a vector boson, and, for the first time, on the branching ratio of the top quark to an ALP and a $u/c$ quark.

Figures (9)

• Figure 1: Distributions of $\mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1}$ for the observed data (black points) and the estimated background (filled histogram) with its uncertainty after the background-only fit to data in the six SRs described in the text. The signal expectation for the Higgs portal model with $m_{s}=55$$\text{Ge V}$ and $c\tau_{s}=100$ mm is shown in the solid line, scaled to $\textrm{BR}(H\rightarrow ss \rightarrow 4b)=1\%$. The signal expectation for ALP production in association with a $Z$ boson scaled to $\sigma(qq\rightarrow Za)=5\times 10^{-3}$ pb, and for $t \rightarrow aq$ scaled to $\mathrm{BR}(t \rightarrow aq) = 0.1\%$ are shown in the dashed and dotted lines, respectively, for $m_{a}=55$$\text{Ge V}$ and $c\tau_{a}=100$ mm. The observed data in the 1- and 2-lepton (VBF) regions corresponds to an integrated luminosity of 140 (37.5) fb$^{-1}$. The ratio between the data and estimated background is shown in the bottom panel.
• Figure 2: The 95% confidence level limits on the (upper) Higgs boson branching ratio $H \rightarrow ss \rightarrow 4q$, (middle) $q\overline{q}\rightarrow Va$ cross-section where $V=W$ or $Z$, and (lower) $t \rightarrow aq$ branching ratio shown as a function of the mean proper decay length $c\tau$ of the long-lived particle. The observed limits are shown with a solid line. The expected limits and corresponding $\pm1\sigma$ uncertainty bands for $m_{s/a} = 55$ GeV are shown with dashed lines and shaded bands, respectively. In the upper plot, the limits shown are on the Higgs boson branching ratio $H \rightarrow ss \rightarrow 4b$ for $m_s=16,40,55$$\text{Ge V}$, and $H \rightarrow ss \rightarrow 4c$ for $m_s=5$$\text{Ge V}$. In the upper plot, the observed limits for the Higgs Portal model from the previous ATLAS search EXOT-2018-57 are shown with the dot-dashed lines.
• Figure 3: Example Feynman diagrams for the three benchmark models considered in the analysis. Figure (a) shows an example diagram for the Higgs portal model, in which the Higgs boson is produced via Vector Boson Fusion, and the long-lived $s$ particles decay to pairs of $b$-quarks. Figure (b) shows an example of the $Va$ ALP production mode, in which the ALP $a$ is produced in association with a $Z$ boson, with $a \rightarrow gg$ and $Z \rightarrow \ell^+ \ell^-$. Figure (c) shows an example diagram of ALP production via the exotic top-quark decay, with $t \rightarrow ac$ and $a \rightarrow c\overline{c}$.
• Figure 4: Distributions of $\mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1}$ for the observed data (black points) and the background prediction (teal filled histogram) with its uncertainty in the three $n_{\mathrm{DV}} = 1$ validation regions with $0.7 < \mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1} < 0.9$. The ratio between the data and predicted background is shown in the bottom panel. The background estimates are computed using events in the three CRs with $\mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1} < 0.7$.
• Figure 5: Distributions of $\mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1}$ for the observed data (black points) and the background prediction (blue) with its uncertainty in the photon validation region (VR) for events with $n_{\mathrm{DV}} = 1$ and $n_{\mathrm{DV}} \geq 2$. The ratio between the data and predicted background is shown in the bottom panel. The background estimate is computed using events in the photon VR with $\mathrm{BDT}_{j_0} \times \mathrm{BDT}_{j_1} < 0.7$.