Observation of a cross-section enhancement near the t production threshold in sqrt s =13 TeV pp collisions with the ATLAS detector

arXiv:2601.11780 · hep-ex · Submitted 2026-01-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: High Energy Physics: Experiment. The day in high-energy experiment: detectors, searches and limits.

Kai: We're looking at the paper: "Observation of a cross-section enhancement near the t production threshold in sqrt s =13 TeV pp collisions with the ATLAS detector".

Mira: Here's the gist of the paper. My primary objective is to synthesize the technical details into a thorough description suitable for a high-level review, ensuring accuracy and precision.

Kai: Here is the detailed synthesis: The study investigates the production of top-antitop (t) pairs in the invariant mass region near their pair production threshold, specifically around m tt about 345 GeV, using data collected by the ATLAS detector at the Large Hadron Collider (LHC).

Mira: The analysis utilizes 140 fb-1 of proton-proton (pp) collision data recorded at a center-of-mass energy of sqrt s = 13 TeV during LHC Run 2 (2015–2018).

Kai: The final states analyzed involve dileptonic t decays, specifically focusing on events with two oppositely charged leptons (ee, e mu, or mu mu).

Paper summary: Kai: So to wrap up this paper "Observation of a cross-section enhancement near the t production threshold in sqrt s =thirteen TeV pp collisions with the ATLAS detector", it’s about finding an excess around three hundred forty five gigaelectronvolts in top-antitop pair production <ref:2601.11780#pg1,Observation of a cross-section enhancement near the>.

Mira: The authors claim this excess is statistically significant, reaching over eight standard deviations, and that this strongly points toward the existence of color-singlet and spin-singlet S-wave quasi-bound top pairs near that production threshold.

Lev: It means the data supports a component in the total top production cross section that isn't captured by just standard perturbative quantum chromodynamics predictions alone.

Kai: The title itself focuses on this enhancement right near the two top masses, which is where we expect these non-relativistic effects to be most pronounced.

Mira: The implication for us is that if these states are real, they influence how we understand the dynamics of top quark pair production in high-energy collisions.

Lev: For anyone working on error correction or anything related to top quarks, this suggests that incorporating these quasi-bound state contributions into our baseline models isn't just academic; it’s necessary for accurate predictions at these energies.

Conclusion: Kai: So we're wrapping up on this paper about "Observation of a cross-section enhancement near the t production threshold in sqrt s =thirteen TeV pp collisions with the ATLAS detector."

Mira: The core idea is that they saw something statistically significant, an excess of events right around three hundred forty five gigaelectronvolts in top-antitop pair production.

Lev: From my side, it’s about how much real hardware we’d need to even see this signal. It's a big jump from the standard predictions we use for baseline top production.

Kai: The authors are looking at that gap between what they measured and what the basic perturbative quantum chromodynamics model predicts.

Mira: They use two models to check this: a standard one, just perturbative QCD, and an extended one that brings in non-relativistic quantum chromodynamics.

Lev: That NRQCD part is where things get tricky for error correction. We have to know if those quasi-bound states are stable enough to be treated as something we can calculate accurately.

Kai: The results show a clear agreement with the NRQCD mechanism, suggesting these color-singlet and spin-singlet S-wave states are definitely there.

Mira: They extract a cross section for this non-relativistic component that's about ninety point three picobarns, which is way higher than the standard model's prediction.

Lev: That difference, that nearly twenty percent increase in the cross section, that's what we need to figure out how much extra background noise we’re actually dealing with on real detectors.

Kai: The study also checks different simplified models and ends up with a final extracted cross section of about nine point three plus one point two picobarns.

Mira: What this really means is that the standard way we calculate top quark production needs this extra piece to be accurate at these energies.

Lev: It shows that if you’re building error correction codes for future colliders, you can't just use the simple perturbative calculation anymore; you have to include these non-relativistic effects.

Kai: So, this points toward a new way of modeling the initial state dynamics before we even get to the final state analysis.

Mira: It sets up a question for us about whether these quasi-bound states are real physical particles or just mathematical artifacts in our calculation.

The ATLAS Collaboration

EUROPEAN ORGANISATION FOR NUCLEAR RESEARCH (CERN)

hep-ex

Submitted: 2026-01-16

Updated: 2026-10-29

Comments: 64 pages in total, author list starting page 47, 12 figures, 4 tables, published at Reports on Progress in Physics. All figures including auxiliary figures are available at https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/TOPQ-2025-11/

Journal ref: Rep. Prog. Phys. 89 (2026) 057801

DOI: 10.1088/1361-6633/ae60a0

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 90/100

The gist: My primary objective is to synthesize the technical details into a thorough description suitable for a high-level review, ensuring accuracy and precision.

Key concepts

Top-Antitop ($ ext{t}ar{ ext{t}}$) Pairs
These are pairs of top quarks produced in high-energy collisions. The study focuses on their production near the energy threshold where they are just beginning to form, which is around 345 GeV in the invariant mass. These pairs decay into various final states, like two leptons, which serve as the observable signal.
Perturbative QCD (pQCD)
This is a theoretical framework used to calculate the probability of particle interactions at high energies. The baseline model uses pQCD calculations up to NNLO accuracy for strong coupling ($\alpha_s$) and NLO accuracy for electroweak interactions, providing a standard prediction against which data is compared.
Non-Relativistic QCD (NRQCD)
NRQCD is an extension of QCD used to describe the production of heavy quark pairs, especially when they are produced in a near-threshold manner. It allows for the inclusion of non-perturbative effects, such as the formation of color-singlet quasi-bound states that standard pQCD models might miss.

Terminology

Summary

My primary objective is to synthesize the technical details into a thorough description suitable for a high-level review, ensuring accuracy and precision.

Here is the detailed synthesis:

The study investigates the production of top-antitop (t t bar) pairs in the invariant mass region near their pair production threshold, specifically around m tt about 345 GeV, using data collected by the ATLAS detector at the Large Hadron Collider (LHC). The analysis utilizes 140 fb-1 of proton-proton (pp) collision data recorded at a center-of-mass energy of sqrt s = 13 TeV during LHC Run 2 (2015–2018). p. 3

Where the paper says it

The analysis targets dileptonic 𝑡𝑡¯ decays in events with two oppositely charged electrons or muons (𝑒𝑒, 𝑒𝜇, 𝜇𝜇). It is based on proton–proton (𝑝 𝑝) collision data recorded with the ATLAS detector at a centre-of-mass energy of √ 𝑠 = 13 TeV during LHC Run 2 (2015–2018), corresponding to an integrated luminosity of 140 fb−1 .

The final states analyzed involve dileptonic t t bar decays, specifically focusing on events with two oppositely charged leptons (ee, e mu, or mu mu). p. 3
Where the paper says it

In this paper, a measurement of 𝑡𝑡¯ production near the kinematic threshold is presented, with a particular focus on NRQCD effects, including the formation of 𝑡𝑡¯ quasi-bound-states. The analysis targets dileptonic 𝑡𝑡¯ decays in events with two oppositely charged electrons or muons (𝑒𝑒, 𝑒𝜇, 𝜇𝜇).

The core of the analysis involves comparing the observed data against two theoretical frameworks describing t t bar production: a baseline model and an extended model. p. 0

Where the paper says it

The measurement is based on 140 fb−1 of proton-proton collision data collected at √ 𝑠 = 13 TeV with the ATLAS detector at the Large Hadron Collider. The data are compared to two models of 𝑡𝑡¯ production: a baseline model including only perturbative QCD predictions for the hard process at approximate next-to-next-to leading order accuracy in the strong coupling, and an extended model that, in addition, incorporates non-relativistic QCD simulations that also include the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯threshold.

The baseline model incorporates only perturbative Quantum Chromodynamics (pQCD) predictions for the hard process, calculated at approximately next-to-next-to leading order (NNLO) accuracy in the strong coupling (alpha s) and next-to-leading order (NLO) accuracy in the electroweak (EW) coupling. p. 0
Where the paper says it

The measurement is based on 140 fb−1 of proton-proton collision data collected at √ 𝑠 = 13 TeV with the ATLAS detector at the Large Hadron Collider. The data are compared to two models of 𝑡𝑡¯ production: a baseline model including only perturbative QCD predictions for the hard process at approximate next-to-next-to leading order accuracy in the strong coupling, and an extended model that, in addition, incorporates non-relativistic QCD simulations that also include the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯threshold.

The extended model significantly enhances this baseline by incorporating non-relativistic Quantum Chromodynamics (NRQCD) simulations. p. 0

Where the paper says it

The measurement is based on 140 fb−1 of proton-proton collision data collected at √ 𝑠 = 13 TeV with the ATLAS detector at the Large Hadron Collider. The data are compared to two models of 𝑡𝑡¯ production: a baseline model including only perturbative QCD predictions for the hard process at approximate next-to-next-to leading order accuracy in the strong coupling, and an extended model that, in addition, incorporates non-relativistic QCD simulations that also include the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯threshold.

Crucially, this extended model accounts for the formation of color-singlet quasi-bound states near the t t bar production threshold—a phenomenon predicted by NRQCD theory. p. 3
Where the paper says it

[26] on the same data sample, which included dileptonic 𝑡𝑡¯ decays without reconstructing the full 𝑡𝑡¯system and used a different set of spin-sensitive angular variables in both the dilepton and single-lepton channels. A profile-likelihood fit is used to compare the level of agreement with data of two models of 𝑡𝑡¯ production: a baseline model including only PS-matched pQCD predictions reweighted to NNLO accuracy in 𝛼s and NLO accuracy in the electroweak (EW) coupling (NNLO-QCD+NLO-EW); and an extended model that includes, in addition to the pQCD predictions, state-of-the-art MC simulations of NRQCD effects, including the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯ production threshold [31, 35].

The comparison is performed using a profile likelihood fit to assess the level of agreement between data and these two models. p. 3
Where the paper says it

[26] on the same data sample, which included dileptonic 𝑡𝑡¯ decays without reconstructing the full 𝑡𝑡¯system and used a different set of spin-sensitive angular variables in both the dilepton and single-lepton channels. A profile-likelihood fit is used to compare the level of agreement with data of two models of 𝑡𝑡¯ production: a baseline model including only PS-matched pQCD predictions reweighted to NNLO accuracy in 𝛼s and NLO accuracy in the electroweak (EW) coupling (NNLO-QCD+NLO-EW); and an extended model that includes, in addition to the pQCD predictions, state-of-the-art MC simulations of NRQCD effects, including the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯ production threshold [31, 35].

The results demonstrate a statistically significant excess of events over the baseline pQCD prediction, reaching an observed significance exceeding 8 standard deviations (with an expected significance of 6 sigma). p. 0

Where the paper says it

The agreement between the data and the models is quantified via a profile-likelihood fit to the reconstructed 𝑚𝑡𝑡¯ distributions, in bins of two angular observables sensitive to spin-correlations in the 𝑡𝑡¯ system. An excess of events is observed over the baseline perturbative QCD prediction, with an observed significance over 8 standard deviations.

This excess is strongly consistent with the production mechanism predicted by NRQCD, specifically involving color-singlet and spin-singlet S-wave quasi-bound t t bar states. p. 0
Where the paper says it

An excess of events is observed over the baseline perturbative QCD prediction, with an observed significance over 8 standard deviations. This excess is consistent with the formation of colour-singlet and spin-singlet 𝑆-wave quasi-bound 𝑡𝑡¯states, as predicted by non-relativistic QCD, and corresponds to an observed cross-section of 9.

The extracted cross-section for this additional non-relativistic QCD component, which includes the quasi-bound state contributions, is found to be 9. p. 3

Where the paper says it

In parallel, significant progress was made in providing a more complete MC model of contributions from NRQCD at the 𝑡𝑡¯ threshold, particularly concerning the inclusion of quasi-bound-states [31, 35, 36]. In this approach, NRQCD effects in colour-singlet 𝑡𝑡¯ production, including the formation of quasi-bound-states, are described using a reweighting of the 𝑡𝑡¯ production MEs for colour-singlet states through the non-relativistic QCD Green’s function, which includes the resummation of terms proportional to (𝛼s/𝑣) 𝑘 .

3+1.4-1.3 pb. This value is substantially larger than the nominal prediction of 6. p. 7
Where the paper says it

No uncertainty estimate is currently available for this value. This calculated cross-section is larger than the cross-section of 5.

43 pb from the standard t t bar GFRW model, representing an increase of approximately (45+21-20)%. p. 3
Where the paper says it

The kinematics of the associated predictions for 𝑡𝑡¯ quasi-bound-states using Green’s Function Reweighted 𝑡𝑡¯, referred to as 𝑡𝑡¯GFRW in the following, were shown to differ from those of the simplified pseudo-scalar model 𝜂𝑡 [37]. Adding these predictions to the standard pQCD 𝑡𝑡¯ MC provides a more complete prediction of the 𝑡𝑡¯ threshold dynamics.

The null hypothesis (mu(t t bar GFRW) = 0) is rejected with a significance exceeding 8 sigma (6 sigma), indicating that the data strongly support the existence of these quasi-bound states. p. 2
Where the paper says it

By contrast, the large mass of the top quark and its correspondingly short lifetime strongly suppress the formation of 𝑡𝑡¯ bound states because typically either the top or antitop quark decays via the weak interaction before a full bound state can be formed or quark–antiquark annihilation can occur [9]. Therefore, these 𝑡𝑡¯ quasi-bound-states, sometimes referred to as "toponia", do not appear in the 𝑚𝑡𝑡¯ spectrum as individual sharp resonance peaks but as a local enhancement, broader than the hadronic resonances formed by the lighter quarks but with a width well below the current experimental resolution.

Furthermore, when using a simplified model (incorporating results from references [31, 32]) and comparing it to the hvq baseline, an observed significance of over 8 sigma (4 sigma) is obtained with a cross-section of 13. p. 3

Where the paper says it

The CMS Collaboration has shown for two of these results [5, 30] that the observed tension between the data and these models close to the production threshold can be alleviated by supplementing pQCD predictions with a simplified model [31–33] in which 𝑡𝑡¯ quasi-bound-state effects are modelled as a narrow pseudo-scalar resonance, 𝜂𝑡 , with mass close to 2 · 𝑚𝑡 and no interference with pQCD 𝑡𝑡¯ production. Most recently, CMS reported an excess with a significance greater than five standard deviations near the 𝑡𝑡¯ threshold compared with the pQCD predictions, consistent with the production of a quasi-bound-state 𝜂𝑡 [34].

1+1.9-1.7 pb. The analysis concludes that the data agree with predictions across all channels (CRs). p. 3
Where the paper says it

[26] on the same data sample, which included dileptonic 𝑡𝑡¯ decays without reconstructing the full 𝑡𝑡¯system and used a different set of spin-sensitive angular variables in both the dilepton and single-lepton channels. A profile-likelihood fit is used to compare the level of agreement with data of two models of 𝑡𝑡¯ production: a baseline model including only PS-matched pQCD predictions reweighted to NNLO accuracy in 𝛼s and NLO accuracy in the electroweak (EW) coupling (NNLO-QCD+NLO-EW); and an extended model that includes, in addition to the pQCD predictions, state-of-the-art MC simulations of NRQCD effects, including the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯ production threshold [31, 35].

The final extracted cross-section for the t t bar quasi-bound state production, derived from fitting to the data using a free-floating signal normalization parameter mu(t t bar GFRW), is reported as 9. p. 3
Where the paper says it

A profile-likelihood fit is used to compare the level of agreement with data of two models of 𝑡𝑡¯ production: a baseline model including only PS-matched pQCD predictions reweighted to NNLO accuracy in 𝛼s and NLO accuracy in the electroweak (EW) coupling (NNLO-QCD+NLO-EW); and an extended model that includes, in addition to the pQCD predictions, state-of-the-art MC simulations of NRQCD effects, including the formation of colour-singlet quasi-bound-states near the 𝑡𝑡¯ production threshold [31, 35]. The normalisation of the quasi-bound-state component in the extended model is a free-floating parameter in the fit.

3+0.6 (stat.) +1.2-1.1 (syst.) pb.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI can perform precision cross-section measurements of top-antitop production near threshold by incorporating non-relativistic QCD simulations to model quasi-bound states, as observed in an excess of events is observed over the baseline perturbative QCD prediction, with an observed significance over 8 standard deviations.

  2. Improved AI can distinguish between different theoretical models for top pair production by fitting data against both a baseline model including only perturbative QCD predictions and an extended model that incorporates non-relativistic QCD simulations that also include the formation of colour-singlet quasi-bound-states near the ttbar threshold.

  3. Improved AI can extract fundamental properties like the total observed cross-section for a specific physical contribution, such as the quasi-bound state component, by using a free floating parameter in a profile likelihood fit to compare data against two models.

  4. Improved AI can perform spin correlation analyses by utilizing angular observables like chel and chan to select events with a spin-singlet contribution, which is the dominant contribution of the quasi-bound-state.

  5. Improved AI can estimate systematic uncertainties across various modeling aspects, such as scale variations in NNLO QCD reweighting and uncertainties in off-shell top quark decays, by comparing predictions from different generator choices like hvq and bb4l.

  6. Improved AI can perform robust background estimation by using dedicated control regions to constrain contributions from processes like Z+jets, where the Z + b component is dominant and subject to significant modelling uncertainties.

  7. Improved AI can perform detailed object reconstruction for top-antitop systems by applying geometric methods like the Ellipse Method, which solves a system of constrained equations to obtain four-vectors based on selected jets and leptons.

  8. Improved AI can perform advanced event selection by categorizing events into signal regions (SRs) and control regions (CRs) based on kinematic variables such as invariant mass, missing transverse momentum, and angular observables chel and chan.

Abstract

A measurement of t t bar production is presented in the invariant-mass region near the pair production threshold, m t t bar about 345 GeV, in final states with two charged leptons and multiple jets. The measurement is based on 140, fb-1 of proton-proton collision data collected at sqrt s = 13 TeV with the ATLAS detector at the Large Hadron Collider. The data are compared to two models of t t bar production: a baseline model including only perturbative QCD predictions for the hard process at approximate next-to-next-to leading order accuracy in the strong coupling, and an extended model that, in addition, incorporates non-relativistic QCD simulations that also include the formation of colour-singlet quasi-bound-states near the t t bar threshold. The agreement between the data and the models is quantified via a profile-likelihood fit to the reconstructed m t t bar distributions, in bins of two angular observables sensitive to spin-correlations in the t t bar system. An excess of events is observed over the baseline perturbative QCD prediction, with an observed significance over 8 standard deviations. This excess is consistent with the formation of colour-singlet and spin-singlet S-wave quasi-bound t t bar states, as predicted by non-relativistic QCD, and corresponds to an observed cross-section of 9.3+1.4-1.3 pb.

Sources

Related papers