⚛️ high-energy experiments

Search for a top-philic heavy resonance in association with top quarks in $pp$ collisions at $\sqrt{s} =$ 13 TeV and 13.6 TeV with the ATLAS detector

Using 13 TeV and 13.6 TeV proton-proton collision data collected by the ATLAS detector, a search for a top-philic vector boson ( $Z'$ ) produced in association with top quarks and decaying into a top-quark pair found no significant excess over the Standard Model background, excluding $Z'$ masses below 1150 GeV at 95% confidence level for a benchmark coupling scenario.

Original authors: ATLAS Collaboration

Published 2026-09-22

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: ATLAS Collaboration

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The universe, as we understand it today, is built from a small set of fundamental particles that interact through specific forces. Among these, the top quark stands out as the heaviest known particle, a massive building block that plays a unique role in the stability of the physical laws governing our world. Because of its immense weight, the top quark is thought to be a sensitive window into new physics, potentially revealing hidden forces or particles that the current standard model of physics cannot explain. Physicists have long suspected that there might be heavier, undiscovered particles that prefer to interact with the top quark above all others, acting as a bridge to a deeper layer of reality. Finding such a particle would be a monumental step, offering clues to why the universe has the mass and structure it does.

To hunt for these elusive particles, researchers at the Large Hadron Collider in Switzerland smash protons together at nearly the speed of light, recreating the intense energy conditions that existed just moments after the Big Bang. In a recent analysis, the ATLAS collaboration, a team of thousands of scientists, examined data from these high-energy collisions to search for a specific type of heavy particle known as a top-philic resonance. This hypothetical particle would be produced alongside top quarks and then decay back into a pair of top quarks. The team looked for this signature in a vast dataset collected over several years, covering two different energy levels: 13 trillion electron volts and 13.6 trillion electron volts. To catch these rare events, they focused on a very specific and clean signal: collisions that produced either two leptons (electrons or muons) with the same electric charge, or at least three leptons. This specific combination of particles is extremely rare in standard background processes, making it an ideal place to spot a new particle if it exists.

The researchers employed a sophisticated computer program based on artificial intelligence to sift through the millions of collision events. This program was trained to distinguish between the messy, common background events and the clean, rare signals that would indicate the presence of the new heavy particle. It analyzed the energy and direction of the particles produced, looking for patterns that matched the predicted behavior of the top-philic resonance. The team defined specific regions of interest where they expected to find the signal, carefully accounting for the known background processes that could mimic the result. They also used control regions, where they knew only background events should occur, to ensure their models of the standard physics were accurate and reliable.

After analyzing the data, the team found no evidence of the new particle. The number of events they observed matched the predictions of the standard model perfectly, with no unexpected excesses that would suggest a new discovery. This null result is, in itself, a significant scientific achievement. By not finding the particle, the researchers were able to set strict limits on its possible existence. They determined that if this top-philic particle does exist, it must be heavier than 1150 gigaelectron volts for a specific theoretical scenario, and heavier than 1750 gigaelectron volts for another. These limits are much higher than what was previously known, effectively ruling out the existence of such a particle in the lower mass ranges that were accessible to earlier experiments.

The study also explored different ways the new particle might interact with the top quark, testing various theoretical possibilities regarding its internal structure. In every scenario tested, the data remained consistent with the standard model, and no signal emerged. The addition of new data from the most recent run of the collider significantly improved the sensitivity of the search, allowing the team to probe deeper into the mass range than ever before. While the hunt for this specific particle continues at higher energy levels, this work has successfully closed the door on a wide range of possibilities, guiding future theories and experiments toward the remaining, unexplored territories of the subatomic world.

Technical Summary: Search for a Top-Philic Heavy Resonance in Association with Top Quarks

Problem and Motivation
The top quark, being the heaviest elementary particle in the Standard Model (SM) with a mass near the electroweak symmetry-breaking scale, is a primary probe for physics beyond the Standard Model (BSM). Many BSM theories, such as composite Higgs models, predict new particles that couple preferentially to the top sector ("top-philic") while having suppressed couplings to light quarks and gluons. These models often introduce a new colour-singlet vector boson, denoted $Z'$ , which can be produced in association with top quarks ( $t\bar{t}Z'$ or $tZ'$ ) and decay into a top-antitop pair ( $Z' \to t\bar{t}$ ). Such processes result in final states with multiple top quarks (three or four), which are rare in the SM and thus offer a sensitive channel for discovery. While previous searches have constrained such resonances, this analysis aims to improve sensitivity by utilizing the full Run 2 dataset combined with early Run 3 data, targeting final states with high jet and $b$ -jet multiplicities.

Methodology
The analysis utilizes proton-proton collision data collected by the ATLAS detector at the Large Hadron Collider (LHC) at centre-of-mass energies of $\sqrt{s} = 13$ TeV (2015–2018, 140 fb $^{-1}$ ) and $\sqrt{s} = 13.6$ TeV (2022–2023, 56 fb $^{-1}$ ).

Event Selection: The search targets events containing either exactly two same-sign leptons (electrons or muons) or at least three leptons. These signatures are chosen to suppress the dominant SM backgrounds. Selected events must contain at least six jets, with at least two being $b$ -tagged, and a scalar sum of transverse momenta ( $H_T$ ) greater than 500 GeV.
Signal Model: The signal is interpreted within a simplified model featuring a $Z'$ boson coupled to top quarks with a coupling strength $c_t$ and a chirality angle $\theta$ . The interaction Lagrangian allows for both left- and right-handed couplings. The dominant production modes considered are $t\bar{t}Z'$ and $tZ'$ .
Background Estimation: The dominant backgrounds are $t\bar{t}t\bar{t}$ production and $t\bar{t}$ production in association with a $W$ , $Z$ , or Higgs boson. Non-prompt (fake) leptons and charge misidentification ( $Q_{misID}$ ) backgrounds are estimated using data-driven methods and template fits in dedicated control regions (CRs). Specifically, CRs are defined to constrain the normalisation of $t\bar{t}W$ and various non-prompt lepton components (heavy-flavour decays, photon conversions, and low-mass virtual photons).
Discrimination Strategy: A transformer-based neural network, implemented in the Salt framework, is employed to discriminate signal from background. The classifier is parameterised as a function of the hypothesised $Z'$ mass. It processes both event-level features (e.g., $H_T$ , jet multiplicity) and object-level features (e.g., $p_T$ , $\eta$ , $b$ -tagging scores). The output is a discriminant variable, $D_{signal}$ , used in the final statistical interpretation.
Statistical Analysis: A profile-likelihood fit is performed simultaneously across all signal regions (SRs) and control regions. Systematic uncertainties, including experimental (jet energy scale, $b$ -tagging) and modelling (generator choice, scale variations) sources, are incorporated as nuisance parameters.

Key Contributions

Integration of Run 3 Data: This is the first ATLAS search for top-philic resonances to combine Run 2 and Run 3 data, increasing the integrated luminosity to 196 fb $^{-1}$ and extending the energy reach to 13.6 TeV.
Advanced Machine Learning: The analysis introduces a mass-parameterised transformer-based classifier for signal-background discrimination, moving beyond traditional cut-based or simpler multivariate approaches.
Comprehensive Background Constrain: The use of multiple dedicated control regions allows for a simultaneous fit that constrains the normalisation of the dominant $t\bar{t}W$ and non-prompt backgrounds directly from data, reducing reliance on pure Monte Carlo predictions.
Chirality Sensitivity: The analysis explores the full range of the chirality angle $\theta$ , investigating the interplay between $t\bar{t}Z'$ and $tZ'$ production modes, which have different dependencies on $\theta$ .

Results
No significant excess of events over the expected Standard Model background is observed. The results are interpreted as 95% confidence level (CL) upper limits on the production cross-section times branching ratio.

Exclusion Limits: For a benchmark scenario with coupling strength $c_t = 1$ and chirality angle $\theta = 0$ (purely left-handed coupling), $Z'$ boson masses below 1150 GeV are excluded. The expected exclusion limit for this scenario is 1250 GeV.
Coupling Strength: For a stronger coupling of $c_t = 3$ , the observed exclusion extends to masses below 1750 GeV (for $\theta = \pi/2$ ) and 2000 GeV (for $\theta = 0$ ).
Improvement: The addition of Run 3 data increases the sensitivity by approximately 30% at a mass of 1.5 TeV. The observed cross-section limits represent an improvement of roughly a factor of 15 over previous ATLAS searches in the single-lepton channel.
Parameter Space: Two-dimensional contours in the $(c_t^2, \theta)$ plane are presented for $Z'$ masses of 1000 GeV and 2000 GeV, demonstrating the sensitivity to the coupling structure.

Significance
The paper claims that this search represents a significant improvement in sensitivity to top-philic $Z'$ production compared to previous results using only Run 2 data. This enhancement is attributed to the inclusion of Run 3 data, improved object performance and selection criteria, and the application of the transformer-based neural network. The results provide stringent constraints on simplified models of top-philic resonances, particularly in the multi-top final state, ruling out a wide range of masses and coupling configurations. The absence of a signal reinforces the validity of the Standard Model up to the TeV scale for these specific top-philic scenarios while setting the stage for future high-luminosity searches.

More like this