Gluon GTMDs are generalized two-parton correlation functions that encode longitudinal momentum, transverse momentum, and spatial positions of gluons in nucleons.
They provide a unified framework bridging GPDs, TMDs, and Wigner distributions, enabling detailed phase-space imaging and insights into gluon orbital angular momentum.
Experimental access via exclusive processes like double quarkonium and J/ψ production offers practical means to measure spin–orbit correlations and other key QCD observables.
Generalized transverse momentum-dependent parton distributions (GTMDs) are the most general two-parton correlation functions within a hadron, encoding the joint dependence on the longitudinal momentum fraction, transverse momentum, and transverse spatial position of partons. Gluon GTMDs, defined for spin-½ hadronic targets, play a central role in mapping the multidimensional tomography of gluons in nucleons and nuclei. They bridge and generalize both generalized parton distributions (GPDs) and transverse momentum-dependent distributions (TMDs), serving as "mother distributions" for all two-parton correlation observables. Gluon GTMDs are key to understanding canonical orbital angular momentum (OAM) and spin–orbit correlations in QCD and are directly linked to gluon Wigner distributions, enabling phase-space imaging of the proton at small and moderate x.
1. Operator Definitions and Parametrization
The leading-twist, unintegrated, off-forward gluon GTMD correlator for a spin-½ hadronic target is defined as
Here, Faμν is the gluon field strength with color index a, Wab is the adjoint gauge link (Wilson line), P=(p+p′)/2, Δ=p′−p, x=k+/P+, ξ=−Δ+/(2P+), and k⊥ is the average transverse gluon momentum (Bhattacharya et al., 2018). These correlators can be projected onto unpolarized and helicity distributions, e.g. via Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.0 and Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.1 (Chakrabarti et al., 17 Sep 2025, Tan et al., 2024).
The leading-twist (twist-2) gluon GTMDs, parameterizing the correlator for a spin-½ target, are organized as follows:
In the small-P=(p+p′)/26 regime, gluon Wigner and GTMD distributions can be constructed from solutions to the impact-parameter-dependent Balitsky-Kovchegov equation, linking the GTMD to dipole P=(p+p′)/27-matrices and saturation physics. The latter enables numerical evaluation of angular harmonics, including elliptic GTMDs (P=(p+p′)/28 component) relevant for diffractive dijet and exclusive vector-meson production (Hagiwara et al., 2016).
4. Experimental Access and Phenomenology
Gluon GTMDs can be accessed via exclusive hard processes in hadronic and lepton-hadron collisions, with process-dependent selectivity for individual GTMDs:
Exclusive double quarkonium production
P=(p+p′)/29
Here, the amplitude at leading order is a convolution of two gluon GTMD correlators. By forming polarization and azimuthal angle combinations in the final state, one projects out bilinear GTMD structures; for instance, Δ=p′−p0 is isolated by combinations such as Δ=p′−p1 (Bhattacharya et al., 2018).
Exclusive heavy meson production at the EIC
In exclusive Δ=p′−p2 electroproduction,
Δ=p′−p3
twist-3 collinear factorization allows direct sensitivity to the Δ=p′−p4-moments of Δ=p′−p5 (OAM) and Δ=p′−p6 (spin–orbit), through the appearance of characteristic Δ=p′−p7 and Δ=p′−p8 azimuthal dependencies in the cross section (Bhattacharya et al., 24 Jan 2026). The Δ=p′−p9 modulation is polarization-independent and selects the canonical OAM, while x=k+/P+0 depends on target polarization and isolates the spin–orbit correlator.
Theoretical and projected EIC studies indicate that few-percent level asymmetries in such processes are feasible for the extraction of gluon GTMDs (Bhattacharya et al., 24 Jan 2026).
5. Small-ξ=−Δ+/(2P+)1 QCD and Saturation Domain
At small ξ=−Δ+/(2P+)2, GTMDs are related to the quantum phase-space imaging of gluons in the color glass condensate (CGC) framework. The gluon GTMD is linked to the Fourier transform of the dipole forward ξ=−Δ+/(2P+)3-matrix, enabling the calculation of both isotropic and elliptic (ξ=−Δ+/(2P+)4) GTMDs: ξ=−Δ+/(2P+)5
ξ=−Δ+/(2P+)6
(Hagiwara et al., 2016). Both components peak at transverse momentum ξ=−Δ+/(2P+)7, the saturation scale, with the elliptic harmonic being numerically sub-leading but experimentally accessible in diffractive observables.
6. Evolution Properties and Renormalization
Gluon GTMDs, as nonlocal light-cone correlators with both rapidity and UV divergences, require soft factor subtraction for proper field-theoretic definition: ξ=−Δ+/(2P+)8
with ξ=−Δ+/(2P+)9 the unsubtracted correlator and k⊥0 the gauge-invariant soft function (Echevarria et al., 2016). Evolution proceeds via coupled renormalization-group (μ) and Collins-Soper (ζ) equations: k⊥1
where k⊥2 is the adjoint cusp anomalous dimension, k⊥3 the non-cusp anomalous dimension, and k⊥4 the rapidity kernel. All gluon GTMDs—unpolarized, polarized, and helicity-flip—share an identical evolution kernel, as the evolution is spin-independent (Echevarria et al., 2016). Solutions can be obtained for resummed evolution up to NNLL accuracy using known anomalous dimensions.
7. Model Implementations and Wigner Distributions
Light-front spectator and gluon-triquark models, particularly those anchored in soft-wall AdS/QCD, enable analytic and numerical evaluation of gluon GTMDs for nonzero skewness. These models represent GTMDs as overlaps of light-cone wave functions, providing closed-form results for all F- and G-type GTMDs across k⊥5, and yield five-dimensional Wigner distributions for various polarization configurations (k⊥6, k⊥7, k⊥8, k⊥9, linearly polarized gluons) (Chakrabarti et al., 17 Sep 2025, Tan et al., 2024).
The resulting Wigner distributions exhibit characteristic symmetry and multipole patterns in transverse momentum and impact parameter space, with Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.00 and Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.01 also providing direct access to the canonical gluon OAM and spin–orbit correlation, respectively. For example, model results at 2 GeV give total gluon angular momentum Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.02, kinetic OAM Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.03, canonical OAM Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.04, and spin–orbit correlation Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.05 (Chakrabarti et al., 17 Sep 2025).
Wigner/GTMDs in the CGC and small-Wabij(x,k⊥,ξ,Δ⊥)=∫(2π)3P+dz−d2z⊥eixP+z−−ik⊥⋅z⊥⟨p′,λ′∣Fa+i(−2z)WabFb+j(+2z)∣p,λ⟩z+=0.06: (Hagiwara et al., 2016)