---
title: Gluonic Mixing Angle in Meson Spectroscopy
url: https://www.emergentmind.com/topics/gluonic-mixing-angle
type: topic
---

# Gluonic Mixing Angle in Meson Spectroscopy

Searching arXiv for recent and foundational papers on gluonic mixing angles in pseudoscalar and scalar sectors.
The **gluonic mixing angle** is a parameter used to quantify the admixture of gluonic degrees of freedom—most commonly a pure glue or gluonium state \(|G\rangle\)—into hadronic mass eigenstates that are otherwise described in terms of quark-flavor basis states. In the pseudoscalar sector, it is typically introduced in analyses of \(\eta\), \(\eta'\), and pseudoscalar glueball mixing, where it measures the extent to which the physical \(\eta'\) or \(\eta\) departs from a purely quark-antiquark configuration. In the scalar sector, closely related constructions quantify glueball–quarkonium mixing through either explicit angles or coupling-based surrogates. Across phenomenology, lattice QCD, and QCD sum rules, the term therefore refers not to a single universal convention, but to a family of parameters that encode quark–gluon configuration mixing in a chosen basis and model [1301.6987], [2205.12541].

## 1. Formal definitions in quark-flavor and glueball bases

In the \(\eta\)–\(\eta'\) system, two quark bases are commonly used. In the octet–singlet basis, the physical states are written as
\[
\begin{pmatrix} \eta \\ \eta' \end{pmatrix}
=
\begin{pmatrix}
\cos\theta_P & -\sin\theta_P\\
\sin\theta_P & \cos\theta_P
\end{pmatrix}
\begin{pmatrix}
\eta_8\\
\eta_0
\end{pmatrix},
\]
with
\[
|\eta_{0} \rangle = \frac{1}{\sqrt{3}}\, |u\bar{u} + d\bar{d} + s\bar{s} \rangle,\qquad
|\eta_{8} \rangle = \frac{1}{\sqrt{3}}\, |u\bar{u} + d\bar{d} - 2s\bar{s} \rangle.
\]
In the nonstrange–strange basis one defines
\[
|\eta_{q}\rangle = \frac{|u\bar{u} + d\bar{d}\rangle}{\sqrt{2}},\qquad
|\eta_{s}\rangle = |s\bar{s}\rangle,
\]
and
\[
\begin{pmatrix} \eta \\ \eta' \end{pmatrix}
=
\begin{pmatrix}
\cos\phi_{P} & -\sin\phi_{P}\\
\sin\phi_{P} & \cos\phi_{P}
\end{pmatrix}
\begin{pmatrix}
\eta_{q}\\
\eta_{s}
\end{pmatrix}.
\]
The angles are related by
\[
\theta_{P} = \phi_{P} - \arctan{\sqrt{2}}.
\]
In this formulation, \(\phi_P\) is the quark-flavor mixing angle, while \(\theta_P\) is the corresponding octet–singlet angle [1301.6987].

The **gluonic mixing angle** appears when a gluonium basis state is added. A widely used pseudoscalar parameterization writes
\[
\eta' =
\cos\psi_{G}\,\sin\psi_{P}\,|\eta_q\rangle
+
\cos\psi_{G}\,\cos\psi_{P}\,|\eta_s\rangle
+
\sin\psi_{G}\,|G\rangle,
\]
\[
\eta =
\cos\psi_{P}\,|\eta_q\rangle
-
\sin\psi_{P}\,|\eta_s\rangle.
\]
Here \(\psi_P\) plays the role of the quark-flavor mixing angle, while \(\psi_G\) is the **gluonic mixing angle** that measures mixing between the quark-flavor sector and the pure gluonium state in the \(\eta'\). In this convention the gluonium content is
\[
\sin^2\psi_G,
\]
which is interpreted directly as the probability of finding the gluonic component in \(|\eta'\rangle\) [1301.6987].

A more general phenomenological framework uses three Euler-type angles. One writes
\[
|\eta\rangle = X_\eta |N\rangle + Y_\eta |S\rangle + Z_\eta |G\rangle,
\qquad
|\eta'\rangle = X_{\eta'} |N\rangle + Y_{\eta'} |S\rangle + Z_{\eta'} |G\rangle,
\]
with
\[
|N\rangle \equiv \frac{1}{\sqrt{2}}(u\bar u + d\bar d),\qquad |S\rangle \equiv s\bar s,\qquad |G\rangle.
\]
The gluonic angles \(\Phi_{G1}\) and \(\Phi_{G2}\) control the glue admixtures in \(\eta\) and \(\eta'\), respectively, so that approximately
\[
Z_\eta^2 = \sin^2\Phi_{G1},\qquad Z_{\eta'}^2 \approx \sin^2\Phi_{G2}
\]
when \(\Phi_{G1}\) is small [1008.2829].

## 2. Pseudoscalar \(\eta\)–\(\eta'\) mixing and gluonium fraction

The most explicit phenomenological use of a gluonic mixing angle in the supplied literature occurs in charge-exchange production of \(\eta\) and \(\eta'\) with \(\pi^-\) and \(K^-\) beams. The underlying idea is that the two beams filter different flavor components: \(\pi^-\) beam preferentially probes the \(|\eta_q\rangle\) component, whereas \(K^-\) beam preferentially probes the \(|\eta_s\rangle\) component. In the quark-only picture, the cross-section ratios at \(t'=0\) satisfy
\[
R_{\pi} = \frac{\sigma(\eta')}{\sigma(\eta)} = \tan^{2}\phi_{P},
\qquad
R_{K} = \frac{\sigma(\eta')}{\sigma(\eta)} = \cot^{2}\phi_{P}.
\]
If a gluonic component is included and assumed not to contribute directly to production, these relations become
\[
R_{\pi} = \cos^{2}\psi_{G}\,\tan^{2}\phi_{P},
\qquad
R_{K} = \cos^{2}\psi_{G}\,\cot^{2}\phi_{P},
\]
so that
\[
R_{\pi}R_{K} = \cos^{4}\psi_{G}.
\]
This directly ties the measurable cross sections to the gluonic mixing angle \(\psi_G\) [1301.6987].

Using 32.5 GeV/c \(\pi^-\) and \(K^-\) beams in the GAMS-\(4\pi\) setup, the measured ratios at \(t'=0\) were
\[
R_{\pi}(\eta'/\eta)= 0.56 \pm 0.04,
\qquad
R_{K}(\eta'/\eta)= 1.30 \pm 0.16.
\]
From the \(\pi^-\) beam alone, the quark-flavor mixing angle was extracted as
\[
\phi_{P} = (36.8 \pm 1.0)^\circ,
\]
corresponding to
\[
\theta_{P} = -17.9 \pm 1.0^\circ
\]
in the octet–singlet basis. However, the \(K^-\)-beam result would imply
\[
\phi_{P} = (41.3 \pm 1.8)^\circ
\]
if one naively ignored gluonium, leading to an inconsistency at about \(2.5\sigma\). The product
\[
R_{\pi}R_{K} = 0.72 \pm 0.09
\]
also deviates from the quark-only expectation \(R_\pi R_K = 1\). Within the adopted two-angle scheme, this yields
\[
\sin^{2}\psi_{G}= 0.15 \pm 0.06.
\]
The same paper cites the KLOE value
\[
\sin^{2}\psi_{G} = 0.12 \pm 0.04
\]
from radiative meson decays, and finds consistency within uncertainties [1301.6987].

This construction gives the gluonic mixing angle a sharply operational meaning: it is the parameter that suppresses the quark components of \(\eta'\) by \(\cos\psi_G\) and assigns a gluonic probability \(\sin^2\psi_G\). The interpretation nonetheless depends on specific dynamical assumptions: planar-diagram dominance, channel selectivity for \(|\eta_q\rangle\) and \(|\eta_s\rangle\), and negligible direct production of the gluonic basis state in charge-exchange reactions. The paper explicitly leaves open the possibility that the discrepancy could reflect limitations of the planar-diagram approximation rather than a literal gluonium component, although its stated conclusion favors the latter [1301.6987].

## 3. Extraction from quarkonium decays and phenomenological fits

A different implementation appears in analyses of \(J/\psi\) and \(\psi(2S)\to VP\) decays, where \(V\) and \(P\) denote vector and pseudoscalar mesons. In that framework, the physical \(\eta\) and \(\eta'\) are expanded in the \(|N\rangle\), \(|S\rangle\), and \(|G\rangle\) basis with coefficients \(X_i\), \(Y_i\), and \(Z_i\). The gluonic fractions are then defined as
\[
Z_\eta^2 = |\langle G|\eta\rangle|^2,\qquad Z_{\eta'}^2 = |\langle G|\eta'\rangle|^2.
\]
The angles \(\Phi_{G1}\) and \(\Phi_{G2}\) govern the admixtures in \(\eta\) and \(\eta'\), respectively, while the quark-mixing angle \(\theta_P\) determines the \(N\)–\(S\) rotation [1008.2829].

Fitting PDG2010 \(J/\psi\to VP\) branching fractions with gluonium allowed in both \(\eta\) and \(\eta'\), the preferred case with \(s_V=0\) and \(x=0.82\) gave
\[
\Phi_{G1} = 0.04 \pm 0.05,\qquad
\Phi_{G2} = 0.53 \pm 0.24.
\]
These correspond to
\[
Z_\eta^2 = 0.002 \pm 0.002,\qquad
Z_{\eta'}^2 = 0.30 \pm 0.24.
\]
The same analysis found, without gluonium,
\[
\theta_P = (-14.04 \pm 2.37)^\circ,
\]
equivalent to a quark-flavor angle
\[
\phi \equiv \theta_P + 54.7^\circ \approx 40.7^\circ.
\]
When gluonium is included, the fitted quark angle shifts to
\[
\theta_P = (-10.21 \pm 4.48)^\circ,
\]
albeit with larger uncertainty [1008.2829].

These results support a now-standard phenomenological pattern: negligible gluonium in \(\eta\), possible but not statistically decisive gluonium in \(\eta'\). The central value \(\Phi_{G2}\approx 0.53\) rad corresponds to roughly \(30^\circ\), but the quoted uncertainty is large enough that zero gluonic content is not strongly excluded. The analysis itself emphasizes that “the large uncertainty prevents us from definitely saying that gluonium content is present or not” [1008.2829]. This suggests that the gluonic mixing angle in phenomenological decay fits is often better interpreted as a model-dependent nuisance parameter constrained by flavor-sensitive channels and OZI-suppressed amplitudes rather than as a sharply determined observable.

## 4. Lattice QCD and direct \(\eta\)–glueball mixing

A more direct realization of the concept is provided by lattice QCD studies of isoscalar pseudoscalar–glueball mixing. In \(N_f=2\) QCD at \(m_\pi\approx 350\) MeV, a two-state mixing framework was constructed in the basis of a quark-antiquark isoscalar pseudoscalar \(|\eta\rangle\) and a pseudoscalar glueball \(|G\rangle\). In the dominant two-state subspace,
\[
H^{(2)} =
\begin{pmatrix}
m_\eta & x_1\\
x_1 & m_G
\end{pmatrix},
\]
and the mixing angle is defined through the usual diagonalization. In the small-mixing limit,
\[
|\theta| \approx \frac{|x_1|}{m_G - m_\eta}.
\]
The physical states are then
\[
|\eta_{\rm phys}\rangle \simeq \cos\theta\,|\eta\rangle + \sin\theta\,|G\rangle,
\qquad
|G_{\rm phys}\rangle \simeq -\sin\theta\,|\eta\rangle + \cos\theta\,|G\rangle.
\]
In this approach the gluonic mixing angle is literally the rotation angle between a quarkonium-like pseudoscalar and a pseudoscalar glueball [2205.12541].

The correlators were computed using the distillation method on an anisotropic \(16^3\times 128\) lattice, with 6991 configurations. The pseudoscalar meson operator was built from smeared quark bilinears, while the glueball operator was constructed from Wilson loops projected onto the \(A_1^{-+}\) irrep. Fitting diagonal and off-diagonal correlators yielded
\[
m_\eta=714(6)(16)\ \text{MeV},
\qquad
m_G \simeq 2.48(5)\ \text{GeV},
\]
and a mixing energy
\[
|x|=107(15)(2)\ \text{MeV}.
\]
From this the mixing angle was determined as
\[
|\theta|=3.46(46)^\circ.
\]
A second operator choice yielded a consistent but less precise value,
\[
|\theta| = 2.5(1.2)^\circ.
\]
The quoted conclusion is that the \(\eta\)–\(G\) mixing is tiny [2205.12541].

The physical interpretation is particularly clear in this framework. A mixing angle of about \(3.5^\circ\) implies
\[
\sin^2\theta \approx 0.0036,
\]
so the gluonic component in the \(\eta\) is of order \(0.4\%\), and the quark component in the glueball-like state is likewise at the sub-percent level. The same work argues, via the Witten–Veneziano relation, that the \(N_f=2\) \(\eta\) mass maps to an \(\mathrm{SU(3)}\) singlet counterpart near \(981\) MeV, suggesting that the large singlet pseudoscalar mass is generated predominantly by the \(U_A(1)\) anomaly rather than by large direct mixing with a pseudoscalar glueball [2205.12541]. A plausible implication is that “gluonic content” in the pseudoscalar singlet sector should be carefully distinguished from anomaly-induced topological mass generation: the former is state mixing, the latter is an effective mass mechanism.

## 5. Scalar-sector analogues and coupling-based mixing measures

Outside the pseudoscalar sector, the term “gluonic mixing angle” is often generalized to scalar glueball–quarkonium systems. Here conventions are less uniform. One important QCD sum-rule analysis of scalar gluonium and scalar quark mesons did not define a basis-state rotation angle directly. Instead, it introduced an effective mixing angle \(\phi\) in terms of current couplings:
\[
\langle 0|J_G|i\rangle \equiv f_{iG},\qquad
\langle 0|J_q|i\rangle \equiv f_{iq},
\]
and
\[
\tan^2\phi
=
\left|
\frac{f_{1G}f_{2q}}{f_{2G}f_{1q}}
\right|.
\]
Using Gaussian sum rules applied independently to gluonic, quark, and mixed correlators, the analysis extracted two scalar states with masses around \(1.0\) GeV and \(1.4\)–\(1.5\) GeV and found
\[
\phi = 54^\circ \pm 4^\circ,
\]
which it interpreted as near-maximal mixing of scalar gluonium and quark mesons [1209.2994].

The basis of that conclusion is a consistent pattern across the three correlators: the heavier state couples more strongly to the gluonic current, the lighter state more strongly to the quark current, yet both states couple significantly to both currents. The analysis attributes the large mixing to chiral-symmetry-violating, non-perturbative contributions in the non-diagonal correlator, especially the quark condensate, the mixed quark–gluon condensate, and instanton effects, which are not chirally suppressed [1209.2994].

By contrast, a later QCD Laplace sum-rule study of off-diagonal \(\bar qq\)–glueball correlators in the scalar sector introduced not an angle but a dimensionless mixing strength
\[
N \equiv \frac{|f_a f_b^\ast + f_a^\ast f_b|}{2|f_a' f_b'|},
\]
with the interpretation
\[
N \approx \cos\theta \sin\theta = \frac{1}{2}\sin 2\theta
\]
in a simple two-state picture. For non-strange scalar \(\bar qq\)–glueball mixing it found
\[
N_{\bar u u GG}\simeq 0.11,
\]
and for strange scalar \(\bar s s\)–glueball mixing
\[
N_{\bar s s GG}\simeq 0.17.
\]
These correspond to effective mixing angles of order \(6^\circ\) and \(10^\circ\), respectively, leading the authors to conclude that their results do not support strong quark/gluonic mixing for the \(0^{++}\) states [1903.06381]. This sharp contrast with the near-maximal value of [1209.2994] illustrates how strongly extracted gluonic mixing depends on the choice of correlator, sum-rule kernel, phenomenological ansatz, and mapping between couplings and state composition.

Lattice calculations in many-flavor theories supply a different perspective. In twelve-flavor QCD, gluonic \(0^{++}\) operators and flavor-singlet scalar fermionic operators were found to project onto nearly degenerate light scalar states in part of parameter space, suggesting strong gluonic–fermionic mixing in the scalar channel. However, that work did **not** define or extract a numerical mixing angle; it instead established the operator basis and GEVP framework required for a future mixed analysis [1310.0963]. This suggests that, in lattice practice, “gluonic mixing angle” in the scalar sector is best viewed as a derived quantity from overlap amplitudes in a combined variational basis rather than a primary observable.

## 6. Large-\(N_c\) expectations and conceptual limitations

Large-\(N_c\) counting provides a parametric expectation for gluonic mixing. For pure gluonic operators \(G\) and mesonic bilinears \(B\), mixed connected correlators scale as
\[
\langle G_1\cdots G_m B_1\cdots B_n\rangle_c \sim \mathcal{O}\big(N_c^{1-m-n/2}\big).
\]
For a glueball–meson two-point mixing correlator, this gives
\[
\langle G B\rangle_c \sim \mathcal{O}(1/\sqrt{N_c}),
\]
which implies that the off-diagonal mixing matrix element is \(\mathcal{O}(1/\sqrt{N_c})\). Since the diagonal masses are \(\mathcal{O}(1)\), a two-state mixing angle scales parametrically as
\[
\theta \sim \mathcal{O}(1/\sqrt{N_c}).
\]
This framework therefore suggests that gluonic mixing angles are naturally suppressed in the large-\(N_c\) limit, even though they need not be numerically tiny at \(N_c=3\) [1508.05773].

This large-\(N_c\) expectation aligns naturally with the small pseudoscalar \(\eta\)–glueball mixing angle found on the lattice [2205.12541], and it is also compatible with phenomenological results that place \(\eta'\) gluonic fractions at moderate rather than dominant levels [1301.6987], [1008.2829]. It is less obviously compatible with near-maximal scalar mixing claims [1209.2994]. This does not, by itself, invalidate such scalar results, because the large-\(N_c\) argument is parametric and the scalar sector is known to be especially sensitive to non-perturbative dynamics, broad resonances, and basis ambiguities. It does, however, sharpen a key conceptual distinction: a “gluonic mixing angle” extracted from a phenomenological fit is not automatically a basis-independent property of QCD.

Several recurrent limitations appear across the literature. First, the angle depends on the chosen basis: singlet–octet, quark-flavor, glueball-plus-quarkonium, or current-coupling basis. Second, different definitions map differently onto “gluonium content”: \(\sin^2\psi_G\), \(Z_{\eta'}^2\), \(\sin^2\theta\), and coupling-derived proxies are not interchangeable without further assumptions. Third, direct glueball mixing must be distinguished from anomaly-driven mass generation in the pseudoscalar singlet channel. Fourth, production-based extractions depend on reaction mechanisms such as planar-diagram dominance, while sum-rule extractions depend on OPE truncation, continuum modeling, and the interpretation of off-diagonal residues. These features explain why the same phrase can refer either to a tightly defined lattice rotation angle or to a model-dependent effective parameter.

## 7. Comparative picture across methods

The range of meanings attached to the gluonic mixing angle is summarized by the following representative determinations.

| Sector / method | Parameterization | Representative result |
|---|---|---|
| \(\pi^-\), \(K^-\) charge exchange | \(\sin^2\psi_G\) in \(\eta'\) | \(\sin^{2}\psi_{G}= 0.15 \pm 0.06\) [1301.6987] |
| \(J/\psi,\psi(2S)\to VP\) phenomenology | \(\Phi_{G2}\), \(Z_{\eta'}^2\) | \(\Phi_{G2}=0.53\pm0.24\), \(Z_{\eta'}^2=0.30\pm0.24\) [1008.2829] |
| \(N_f=2\) lattice \(\eta\)–glueball mixing | two-state angle \(\theta\) | \(|\theta|=3.46(46)^\circ\) [2205.12541] |
| Scalar Gaussian sum rules | coupling angle \(\phi\) | \(\phi = 54^\circ \pm 4^\circ\) [1209.2994] |
| Scalar Laplace sum rules | mixing strength \(N\), effective \(\theta\) | small mixing, \(N\simeq0.11\)–\(0.17\) [1903.06381] |

Taken together, these results support several general statements. In the pseudoscalar sector, \(\eta\) is consistently found to have negligible or tiny gluonic admixture, while \(\eta'\) may contain a nonzero gluonic component, though its precise size remains model dependent [1008.2829], [1301.6987], [2205.12541]. In the scalar sector, the existence of strong glueball–quarkonium mixing is method dependent and remains more controversial: some sum-rule analyses favor near-maximal mixing, others favor only weak admixture [1209.2994], [1903.06381]. This suggests that the phrase “gluonic mixing angle” should always be interpreted together with the basis choice, the channel, and the extraction framework.

A common misconception is to identify a nonzero gluonic mixing angle with a large “glueball component” in a naive probabilistic sense across all formulations. That identification is justified only in specific two-state parameterizations such as \(\sin^2\psi_G\) or \(\sin^2\theta\). In coupling-based or multi-angle schemes, the angle instead parameterizes overlaps, current couplings, or Euler rotations in a larger state space, and its numerical value need not map directly onto a basis-independent probability. Another common misconception is to equate the anomalous mass of the \(\eta'\) with direct glueball mixing. The lattice \(N_f=2\) study explicitly argues the opposite: the topology-induced interaction contributes most of the \(\eta\) mass, while the direct \(\eta\)–glueball mixing angle is tiny [2205.12541].

The gluonic mixing angle thus occupies a distinctive position in hadron spectroscopy. It is simultaneously a phenomenological descriptor of flavor-singlet mesons, a lattice-extractable parameter in controlled two-state systems, a sum-rule diagnostic of off-diagonal correlators, and a large-\(N_c\) constrained measure of quark–gluon configuration mixing. Its significance lies less in any single canonical numerical value than in the way it organizes the longstanding question of how much of an observed hadron is quarkonium, how much is glue, and how that distinction depends on the theoretical lens through which the state is examined.

Source: https://www.emergentmind.com/topics/gluonic-mixing-angle