Search for events with one displaced vertex from long-lived neutral particles decaying into hadronic jets in the ATLAS muon spectrometer in $pp$ collisions at $\sqrt{s}=13$ TeV

TL;DR

The paper reports a search for events with a single displaced vertex from long-lived neutral particles decaying to hadronic jets in ATLAS, using 140 fb$^{-1}$ of 13 TeV pp data. It leverages a dedicated MS displaced-vertex reconstruction algorithm and two analysis channels (muon RoI cluster trigger and Z-associated lepton triggers) to maximize sensitivity across a wide LLP lifetime range. A data-driven ABCD method with neural-network–defined control planes estimates background, and signal efficiencies are extrapolated over lifetimes using MC with lifetime validation. No excess is observed, and stringent upper limits are set on LLP production cross sections times branching fractions for scalar portal, baryogenesis, ALP, and dark photon models, including the strongest ATLAS constraints for certain ALP lifetimes and branching fractions $B\ge 1\%$ for the Higgs portal case over $c\tau$ ranges from a few cm to tens of meters.

Abstract

A search for events with one displaced vertex from long-lived particles using data collected by the ATLAS detector at the Large Hadron Collider is presented, using 140 fb$^{-1}$ of proton-proton collision data at $\sqrt{s} = 13$ TeV recorded in 2015-2018. The search employs techniques for reconstructing vertices of long-lived particles decaying into hadronic jets in the muon spectrometer displaced between 3 m and 14 m from the primary interaction vertex. The observed number of events is consistent with the expected background and limits for several benchmark signals are determined. A scalar-portal model and a Higgs-boson-portal baryogenesis model are considered. A dedicated analysis channel is employed to target Z-boson associated long-lived particle production, including an axion-like particle and a dark photon model. For the Higgs boson model, branching fractions above 1% are excluded at 95% confidence level for long-lived particle proper decay lengths ranging from 5 cm to 40 m. For the photo-phobic axion-like particle model considered, this search produces the strongest limits to date for proper decay lengths greater than $\mathcal{O}(10)$ cm.

Figures (18)

• Figure 1: Diagrams of the (a) scalar portal and (b) baryogenesis models. The LLPs are represented by double lines and labeled $s/\chi$, and the final-state SM fermions are labeled as $f$Strassler:2006ri.
• Figure 2: Diagram of the ALP production through an intermediate, off-shell $\PZ$ boson. The long-lived ALP is represented by a dashed line and labeled $a$, while the final-state SM charged leptons are labeled as $\ell$, and the SM gluons as $g$.
• Figure 3: Diagrams of the (a) dark photon and (b) -associated scalar models. The long-lived dark photon is represented by $Z_d$, the long-lived scalars are labeled $s$, the final-state SM charged leptons are labeled as $\ell$, and the final-state SM fermions are labeled as $f$.
• Figure 4: Efficiency for the Muon RoI Cluster trigger against the LLP decay position for some scalar portal samples in the (a) MS barrel and (b) MS endcaps. These distributions represent events that meet the data quality requirements and feature a reconstructed primary vertex. These efficiency profiles are derived exclusively from MC simulation and do not undergo any corrections for mis-modeling, as outlined in Section \ref{['sec:RoICluTrig']}. The vertical lines in the plots denote relevant detector boundaries. "HCal end" signifies the outer boundary of the hadronic calorimeter, while "RPC 1/2" indicates the first/second stations of RPC chambers. "TGC 1" represents the first stations of TGC chambers, and "L/S" denotes the Large or Small sectors.
• Figure 5: Cumulative displaced vertex efficiency in the barrel where the vertex is required to be isolated from (a) jets and (b) high- tracks, depicted as a function of the selected $\Delta R$, when (c) the sum of low- tracks in a $\Delta R = 0.2$ cone around the vertex direction must be less than a specified cut in , and (d) when the sum of low- tracks in a $\Delta R$ cone around the vertex is required to be less than $5~\text{Ge V}\xspace$, as a function of the $\Delta R$ value. The vertical lines indicate the isolation values utilized in the RoI-triggered channel. Discrepancies between multijet and data distributions are attributed to non-collision background, absent in MC simulations. The $\Delta R$ selection on the jets (a) and high- tracks (b) was tightened to 0.8 to suppress residual non-collision background contributions in the ABCD plane as described in Section \ref{['sec:Bkg1MSVx']}.