Precise measurement of the $tar{t}$ production cross-section and lepton differential distributions in $eμ$ dilepton events

Measurements of the inclusive and differential top-quark pair ($t\bar{t}$) cross-sections in proton–proton collisions at $\sqrt{s} = 13$ TeV, using 140 fb$^{-1}$ of data collected by the ATLAS experiment at the Large Hadron Collider, are presented. Events with an opposite-charge $eμ$ pair and $b$-tagged jets are selected, and the inclusive cross-section measurement is used to determine the top-quark pole mass via the dependence of the predicted cross-section on $m_t^\mathrm{pole}$. Complementary measurements of $eμb\bar{b}$ production, treating both $t\bar{t}$ and $tW$ events as signal, are also provided.

💡 Research Summary

The ATLAS Collaboration presents a high‑precision measurement of the top‑quark pair (tt̄) production cross‑section in proton‑proton collisions at a centre‑of‑mass energy of 13 TeV, using the full Run 2 dataset corresponding to an integrated luminosity of 140 fb⁻¹. The analysis focuses on the dileptonic eμ channel, selecting events with an opposite‑sign electron–muon pair and at least two b‑tagged jets. A double‑b‑tag method is employed: events with exactly one and exactly two b‑tagged jets are counted (N₁ and N₂), and two simultaneous equations relating these counts to the inclusive cross‑section σ_tt̄, the lepton selection efficiency (ε_{eμ}), the per‑jet b‑tag efficiency (ε_b), and a tagging‑correlation factor (C_b) are solved. Background contributions are estimated using a mixture of simulation and data‑driven control samples, including same‑sign eμ events to constrain fake‑lepton backgrounds.

The inclusive cross‑section is measured to be
σ_tt̄ = 829.3 ± 1.3 (stat) ± 8.0 (syst) ± 7.3 (lumi) ± 1.9 (beam) pb,
corresponding to a total relative uncertainty of about 1.2 %. The fiducial cross‑section, defined for leptons with p_T > 20 GeV and |η| < 2.5, is σ_fid_tt̄ = 14.04 ± 0.02 (stat) ± 0.10 (syst) ± 0.12 (lumi) ± 0.03 (beam) pb. The measurement assumes a top‑quark mass of 172.5 GeV; the dependence of the result on the mass is (1/σ) dσ/dm_t = −0.29 %/GeV.

By comparing the measured σ_tt̄ with next‑to‑next‑to‑leading‑order (NNLO) QCD predictions computed with the Top++ program using the NNPDF3.1_no‑top PDF set (which does not include top‑quark data), a Bayesian likelihood fit yields a pole mass of
m_t^{pole} = 172.8 +1.5 / −1.7 GeV.
This extraction provides an independent validation of the top‑quark mass determination, complementary to direct reconstruction methods.

In addition to the inclusive result, the analysis delivers ten single‑differential and three double‑differential cross‑sections as functions of lepton and dilepton kinematic variables for both the pure tt̄ → eμ process and the combined eμ b b̄ final state (including tW contributions). Representative distributions of the leading lepton transverse momentum (p_T) and pseudorapidity (η) are shown, with total uncertainties as low as 0.3 % for normalized spectra in certain regions.

The measured differential spectra are compared with a variety of state‑of‑the‑art Monte‑Carlo generators: Powheg+PYTHIA8, Powheg+Herwig7, aMC@NLO+PYTHIA8, Powheg MiNNLO+PYTHIA8, and Powheg bb4l+PYTHIA8. Generators that incorporate NNLO matrix elements matched to parton showers (Powheg MiNNLO) provide the best overall description of the data, especially in the high‑p_T tail where older Powheg hvq predictions tend to underestimate the cross‑section. The comparisons also include PDF variations (PDF4LHC15, CT14, MMHT14, PDF4LHC21) and show that the data are sensitive enough to discriminate among them, offering valuable input for future PDF fits.

Systematic uncertainties are dominated by b‑tag efficiency and its correlation, lepton identification and trigger efficiencies, jet energy scale and resolution, background modelling, and theoretical uncertainties on the signal modelling (PDF, α_s, QCD scale). The double‑b‑tag technique, together with the large dataset, reduces the statistical component to a negligible level, allowing systematic effects to be the limiting factor.

Overall, this work achieves a sub‑percent precision on the tt̄ inclusive cross‑section, extracts the top‑pole mass with a competitive uncertainty, and provides a suite of high‑precision differential measurements that challenge current theoretical predictions. The results constitute a benchmark for future LHC analyses, aid in the tuning of Monte‑Carlo generators, and contribute to the global effort of testing the Standard Model at the highest accessible energy scales.