Toponium: Bound State of Top Quarks

• Toponium is a bound state of a top quark and an anti-top quark, characterized by nearly degenerate pseudoscalar and vector states and an ultra-small Bohr radius (~8×10⁻¹⁸ m).
• NRQCD and various potential models predict its binding energies and fine splittings with corrections on the order of 0.1–0.2 GeV, enabling precision extraction of QCD and electroweak parameters.
• Experimental evidence from threshold production at the LHC, using advanced kinematic and machine-learning analyses, provides robust signatures and potential probes for BSM effects.

Toponium is the bound state of a top quark ($t$) and an anti-top quark ($\bar t$), forming the most massive and smallest quarkonium system in the Standard Model. Unlike lighter quarkonia, toponium experiences comparable formation and decay times due to the top quark's large weak width and extremely short lifetime, probing the highest energy and shortest distance scales attainable in QCD. Theoretical, phenomenological, and experimental advances over the past several years have culminated in precision modeling, evidence for quasi-bound states following threshold production at the LHC, and dedicated proposals for extracting key QCD and electroweak parameters from toponium observables.

1. Quantum Numbers, Potential Models, and Spectral Properties

Toponium admits the same $n^{2S+1}L_J$ classification as other quarkonia, with primary S-wave ground states: the pseudoscalar ($\eta_t: J^{PC}=0^{-+}$) and the vector ($\psi_t: 1^{--}$) (Bai et al., 17 Jun 2025, Wang et al., 2024, Najjar et al., 13 Nov 2025, Zhang et al., 18 Mar 2026). Due to the top quark’s mass ($m_t\simeq 172.4$–$173.3$ GeV), the reduced mass $\mu = m_t/2$ is much larger than $\Lambda_{\rm QCD}$, and relativistic, spin, and nonperturbative corrections are parametrically suppressed ($\lesssim O(100~\mathrm{MeV})$) (Wang et al., 2024). The typical binding energies are

$\bar t$0

with $\bar t$1, $\bar t$2, and lead to ground-state masses $\bar t$3–$\bar t$4 GeV, nearly degenerate for spin singlet and triplet states (Jiang et al., 2024, Bai et al., 17 Jun 2025, Najjar et al., 13 Nov 2025, Zhang et al., 18 Mar 2026). Empirical and theoretical models (Cornell, screened, logarithmic, Dyson–Schwinger) all consistently find that fine and hyperfine splittings are at most $\bar t$5–$\bar t$6 GeV (Wang et al., 2024, Jiang et al., 2024, Zhang et al., 18 Mar 2026).

The Bohr radius is extremely small: $\bar t$7 (Fu et al., 2024), confirming that toponium is the most compact known QCD bound state and that perturbative treatments are valid throughout, with linear confining corrections subdominant.

2. Decay Channels, Branching Ratios, and Lifetime

Toponium decay is dominated by the weak decay of its constituent tops, with total resonance widths set by $\bar t$8–$\bar t$9 GeV (Fu et al., 2024, Wang et al., 2024, Bai et al., 17 Jun 2025). Annihilation decays $n^{2S+1}L_J$0, $n^{2S+1}L_J$1 are strongly suppressed by $n^{2S+1}L_J$2 and the smallness of $n^{2S+1}L_J$3 or $n^{2S+1}L_J$4. Detailed width calculations yield (Fu et al., 17 Apr 2025, Jiang et al., 2024, Wang et al., 2024):

• $n^{2S+1}L_J$5–$n^{2S+1}L_J$6 MeV
• $n^{2S+1}L_J$7–$n^{2S+1}L_J$8 keV
• $n^{2S+1}L_J$9–$\eta_t: J^{PC}=0^{-+}$0 keV

Branching fractions are thus:

• $\eta_t: J^{PC}=0^{-+}$1
• $\eta_t: J^{PC}=0^{-+}$2
• $\eta_t: J^{PC}=0^{-+}$3

Reflecting the dominance of constituent decay, direct annihilation searches are challenging at hadron colliders, but final-state kinematic features and resonance structures at threshold provide discriminating power (Fu et al., 17 Apr 2025, Jiang et al., 2024, Wang et al., 2024, Francener et al., 5 Feb 2025).

3. Production Mechanisms and Collider Phenomenology

The dominant production mode at hadron colliders is gluon fusion, $\eta_t: J^{PC}=0^{-+}$4, calculable at leading order via S-wave NRQCD projection: $\eta_t: J^{PC}=0^{-+}$5 Including PDF convolution and threshold resummation yields total cross sections at $\eta_t: J^{PC}=0^{-+}$6 TeV: $\eta_t: J^{PC}=0^{-+}$7 (Jiang et al., 2024, Bai et al., 17 Jun 2025, Fu et al., 17 Apr 2025, Desai et al., 27 Jan 2026), consistent with LHC Run 2 measurements. Central exclusive production via $\eta_t: J^{PC}=0^{-+}$8 or central gluon fusion is subdominant ($\eta_t: J^{PC}=0^{-+}$9(attobarn)) and is only accessible at the FCC with $\psi_t: 1^{--}$0 events/ab$\psi_t: 1^{--}$1 (Francener et al., 5 Feb 2025).

At $\psi_t: 1^{--}$2 machines, vector ($\psi_t: 1^{--}$3) resonance formation in threshold scans enables ultra-precise extraction of $\psi_t: 1^{--}$4 and tuning of $\psi_t: 1^{--}$5 for maximal toponium yield (Fu et al., 2024, Bai et al., 17 Jun 2025). $\psi_t: 1^{--}$6 leptonic and hadronic cross sections at $\psi_t: 1^{--}$7 are at the $\psi_t: 1^{--}$8–$\psi_t: 1^{--}$9 fb level (Fu et al., 17 Apr 2025). The scalar channel is suppressed by the smallness of direct $m_t\simeq 172.4$0 production at $m_t\simeq 172.4$1 machines.

4. Experimental Evidence and Signal Characterization

Recent ATLAS and CMS results exhibit a statistically significant excess in $m_t\simeq 172.4$2 production near threshold ($m_t\simeq 172.4$3 GeV) with local significances exceeding $m_t\simeq 172.4$4 (Najjar et al., 13 Nov 2025, Bai et al., 17 Jun 2025, Fuks et al., 3 Nov 2025). This excess is interpreted as evidence for a spin-0, color-singlet resonance ($m_t\simeq 172.4$5) (Bai et al., 17 Jun 2025, Fuks et al., 3 Nov 2025, Aguilar-Saavedra, 2024). Comprehensive multivariate analyses exploiting spin density matrices, angular correlations, and quantum information observables yield further discrimination between toponium and the continuum (Antozzi et al., 26 Feb 2026, Desai et al., 27 Jan 2026).

LHC data are analyzed through both template fits of kinematical variables—such as reconstructed $m_t\simeq 172.4$6 mass, rapidity separation ($m_t\simeq 172.4$7), and $m_t\simeq 172.4$8 between leptons—and boosted decision trees incorporating spin and entanglement observables. The inclusion of Coulombic Green's function reweighting in event generators further improves agreement with experimental data and tightens systematic uncertainties (Fuks et al., 3 Nov 2025, Desai et al., 27 Jan 2026, Antozzi et al., 26 Feb 2026).

The LHC single-lepton and dilepton channels are both viable for observation, with the single-lepton analysis recently shown to give $m_t\simeq 172.4$9 significance in full Run 2 data for shape observables such as the reconstructed transverse momentum $173.3$0 and minimal lepton–jet angular separation (Fuks et al., 3 Sep 2025).

5. Theoretical and Computational Frameworks

Toponium formation is treated in NRQCD (nonrelativistic QCD), potential models (Coulomb, Cornell, screened, logarithmic), QCD sum rules, and covariant Dyson–Schwinger equations (rainbow–ladder truncation) (Jiang et al., 2024, Zhang et al., 18 Mar 2026, Muraki et al., 2024, Najjar et al., 13 Nov 2025). All approaches yield consistent predictions for masses, wavefunctions, and binding energies. In NRQCD, the Coulombic regime is justified by the ultra-short Bohr radius and the negligible effect of linear confinement (Fu et al., 2024).

NRQCD-based modeling incorporates binding and threshold effects via energy-dependent Green’s functions, implemented in Monte Carlo event generators through reweighting, as in MadGraph5_aMC@NLO and custom UFO models (Fu et al., 17 Apr 2025, Fuks et al., 3 Sep 2025). Recursive Jigsaw Reconstruction (RJR), RestFrames techniques, and boosted decision trees are employed to optimize reconstruction and classification, offering $173.3$1–$173.3$2 improvements in sensitivity compared to established reconstruction methods (Desai et al., 27 Jan 2026, Desai et al., 26 Jan 2026).

QCD sum-rule studies—including nonperturbative effects up to dimension-eight—match experimental results and confirm genuine binding ($173.3$3) for both pseudoscalar and vector channels (Najjar et al., 13 Nov 2025). Bethe–Salpeter analyses further confirm small hyperfine splittings and provide estimates for decay constants ($173.3$4–$173.3$5 GeV) (Zhang et al., 18 Mar 2026).

6. Extraction of Fundamental Parameters and Implications for QCD and BSM Physics

Observation and precision spectroscopy of toponium at hadron and lepton colliders provide a unique probe for:

• Determining $173.3$6 via threshold scans or lineshape fits to $173.3$7 MeV accuracy, an order-of-magnitude improvement over current methods (Fu et al., 2024).
• Extracting the QCD coupling $173.3$8 in the multi-hundred GeV regime via Bohr radius and spectral features (Fu et al., 17 Apr 2025).
• Testing NRQCD and the validity of perturbative QCD in the ultraviolet.
• Indirect constraints on BSM physics, including light (pseudo)scalars coupled to the top sector: shifts in the toponium binding energy or resonance mass are sensitive to additional Yukawa interactions or singlet-Higgs mixing (Matsuoka, 2024, Bai et al., 17 Jun 2025).
• The stabilization of the electroweak vacuum and the extraction of the top–Higgs Yukawa via rare decay modes (e.g., $173.3$9, $\mu = m_t/2$0) (Matsuoka, 2024, Bai et al., 17 Jun 2025).

Observed cross sections and branching ratios broadly agree with NRQCD predictions, but some analyses find an excess (by a factor $\mu = m_t/2$1) over pure Coulomb theory, potentially reflecting higher-order corrections or BSM effects (Fuks et al., 3 Nov 2025).

7. Experimental Outlook and Challenges

Toponium can be robustly isolated at the LHC in dilepton and single-lepton $\mu = m_t/2$2 final states through a combination of precision kinematic, angular, and entanglement observables (Antozzi et al., 26 Feb 2026, Desai et al., 27 Jan 2026, Desai et al., 26 Jan 2026, Fuks et al., 3 Sep 2025). Planned HL-LHC and future lepton colliders (CEPC, FCC-ee) will allow detailed measurements of toponium lineshapes, branching fractions, and the extraction of key SM parameters at unprecedented precision (Bai et al., 17 Jun 2025, Fu et al., 2024).

Limitations include the need for precise theoretical modeling of background spin/color observables, control of systematic uncertainties at the few-percent level, and the practical suppression of rare annihilation decays. P-wave ($\mu = m_t/2$3) states are predicted to be unobservable due to suppressed production; observation of exclusive toponium ($\mu = m_t/2$4) in ultraperipheral $\mu = m_t/2$5, $\mu = m_t/2$6, $\mu = m_t/2$7 collisions is unrealistic at the LHC, and only marginally viable at the FCC (Francener et al., 5 Feb 2025).

The field continues to advance rapidly with the integration of higher-order corrections, quantum information observables, and machine-learning discrimination strategies, cementing toponium as a unique window into both top-quark and QCD fundamental physics at the highest energies currently accessible.