← Back

Research · gluonic matter

One Model for Gluonic Hadrons

A single set of rules for quarks and gluons, fixed on ordinary charmonium, then asked to predict hybrids, glueballs and gluelumps.

C. Farina and E. S. Swanson, in progress · 6 November 2026

Preview of early work. This page shows work in progress, before publication. The model, its parameters and every number and figure here may change. Please cite as C. Farina and E. S. Swanson, in progress.

1The question

Protons, pions and the other hadrons are made of quarks held together by gluons. In most hadrons the gluons only provide the glue. But QCD, the theory of quarks and gluons, also allows hadrons in which the gluons themselves carry energy and angular momentum:

Lattice QCD, which solves the theory numerically on a space–time grid, now gives the masses of all of these. Models are usually built for one kind of hadron at a time. The aim here is to describe all of them with one Hamiltonian and one set of parameters, so that each new sector is a test rather than a new fit.

2The model in brief

The starting point is QCD in Coulomb gauge. There the gluon field splits into an instantaneous Coulomb-like force and transverse gluons. The transverse gluons behave like particles with an energy \(\omega(k)\) that depends on their momentum \(k\); they are called quasigluons or constituent gluons. A hadron is then a small number of heavy quarks and quasigluons with the Hamiltonian

\[ \begin{equation} H=\sum_{\text{quarks}}\frac{\mathbf p_i^2}{2m}+\sum_{\text{gluons}}\omega(p_g)+\sum_{\text{pairs}}V_{ij}, \qquad M=\sum_{\text{quarks}} m + E, \tag{1}\label{eq:H} \end{equation} \]

where \(E\) is the lowest (or an excited) energy of \(H\) and \(M\) the hadron mass. The pair potentials \(V_{ij}\) contain a short-distance Coulomb part from one-gluon exchange, a linear confining part with string tension \(b\), and spin-dependent terms. They act between every pair: quark–quark, quark–gluon and gluon–gluon. Each is multiplied by a colour factor that depends on how the pair's colours combine, which is what lets one rule cover mesons, baryons, hybrids and glueballs.

Two forms of the gluon energy are tried, and each is used everywhere in the model:

\[ \begin{align} \text{option 1 (massive):}&\qquad \omega^2(k)=k^2+m_g^2\,e^{-k/b_g}, \tag{2}\label{eq:omega1}\\ \text{option 2 (Gribov):}&\qquad \omega^2(k)=k^2+\frac{M^4}{k^2}. \tag{3}\label{eq:omega2} \end{align} \]

Option 1 gives the gluon an effective mass at low momentum. Option 2 is the form proposed by Gribov and found in Coulomb-gauge studies of QCD; it grows without limit at small \(k\), so long-wavelength gluons cost a lot of energy.

Table 1. The four fitted parameters, from the charmonium fit for each option. Everything else is fixed in advance.
ParameterMeaningOption 1Option 2
\(\alpha_s\)strong coupling0.4100.416
\(b\)string tension (GeV²)0.1570.156
\(m_c\)charm-quark mass (GeV)1.3901.390
\(\sigma\)smearing scale of the one-gluon terms (GeV)1.2471.299

3How it is tested

The four parameters of Table 1 are fitted to 20 charmonium (\(c\bar c\)) levels from lattice QCD. After that nothing is refitted, apart from one overall constant per gluonic sector. This constant absorbs the poorly known rest energy of the gluon; it moves a whole spectrum up or down but cannot change the spacing between its levels. So the pattern of levels in each new sector is a prediction.

The equations are solved numerically, with each hadron's wave function expanded in Gaussian functions of the particle separations and the best combination found by diagonalising \(H\) (the Rayleigh–Ritz method).

In every figure below, orange squares are option 1, magenta triangles are option 2, and the lattice results are shown with their errors.

4Charmonium and Ωccc

Charmonium is where the model is fitted, so Figure 1 shows how well the fit works: 22 levels from 1S to 1F, with an average deviation of 13 MeV. The triply charmed baryon \(\Omega_{ccc}\) (Figure 2) is the first test: three charm quarks with the same interaction, now with the colour factor of a baryon.

Charmonium masses by multiplet from 1S to 1F, model against lattice
Figure 1. Charmonium levels (GeV) by multiplet, model against lattice QCD (Liu et al. 2012, blue circles). Absolute masses: no shift.
Triply charmed baryon levels, model against lattice
Figure 2. \(\Omega_{ccc}\) levels against lattice QCD (Padmanath et al. 2014), shown relative to \(\tfrac32\) times the \(\eta_c\) mass, which removes the uncertainty in the charm-quark mass.

5Hybrids

Add one gluon to charmonium and you get the charmonium hybrids. The lattice finds them in groups (labelled H1 to H4 in Figure 3), and the model reproduces this grouping and the order of the groups. The boxed channels are exotic: they cannot be made from a charm quark and antiquark alone. Figure 4 does the same for hybrid baryons, three charm quarks plus a gluon.

Charmonium hybrid masses in four multiplets H1 to H4, model against lattice
Figure 3. Charmonium hybrids (MeV), lowest level in each channel, against lattice QCD (Liu et al. 2012, green diamonds). Each option has one overall shift, given in the legend. Exotic \(J^{PC}\) are boxed.
Triply charmed hybrid baryon levels, model against lattice
Figure 4. Hybrid baryons \(cccg\), three levels per \(J^P\), against the hybrid-like lattice levels (Padmanath et al. 2014). One shift per option, set by the lowest \(\tfrac12^+\) and \(\tfrac32^+\) levels.

6Glueballs, gluelumps and hybrid potentials

The last three tests remove the quarks, or make them infinitely heavy. Glueballs (Figure 5) are two quasigluons bound together; this is the hardest test, and the deviations are largest here. Gluelumps (Figure 6) are one gluon held by a fixed colour source. The static hybrid potentials (Figure 7) are the energy of the gluon field between a fixed quark and antiquark as the distance \(r\) between them grows; in the usual picture of heavy hybrids, the quark and antiquark then move in these potentials.

Glueball masses by J^PC, model against lattice bands
Figure 5. Glueballs: ground states and first excitations (GeV) against lattice QCD (Morningstar and Peardon 1999; Meyer and Teper 2005), shown as bands. One shift per option.
Gluelump mass splittings, model against lattice
Figure 6. Gluelump splittings from the lowest (\(1^{+-}\)) gluelump: ground states (solid) and first radial excitations (dashed), against lattice QCD (Herr, Schlosser and Wagner 2024; Foster and Michael 1999). The model's gluon contains only the lowest (TE) type of excitation, so the TM levels are shown for comparison only.
Static hybrid potentials Pi_u and Sigma_u minus versus separation, model against lattice
Figure 7. Static hybrid potentials \(V(r)\) for the \(\Pi_u\) and \(\Sigma_u^-\) gluon configurations and their first excitations, against lattice QCD (Capitani et al. 2019, grey bands; Juge, Kuti and Morningstar 2003, crosses and diamonds). The lowest curve, \(\Sigma_g^+\), is the ordinary quark–antiquark potential without an excited gluon.

7Citation and lattice data

C. Farina and E. S. Swanson, in progress.

Earlier work on the Coulomb-gauge constituent picture: A. P. Szczepaniak and E. S. Swanson, "Coulomb gauge QCD, confinement, and the constituent representation", Phys. Rev. D 65, 025012 (2001), hep-ph/0107078.

Lattice QCD results compared with: