---
title: 'Valence Pion GPDs: Insights from Theory & Lattice'
url: https://www.emergentmind.com/topics/valence-pion-generalized-parton-distributions-gpds
type: topic
---

# Valence Pion GPDs: Insights from Theory & Lattice

Searching arXiv for recent and foundational papers on valence pion GPDs to ground the article in the literature.
Valence pion generalized parton distributions describe the leading-twist quark content of the pion in off-forward kinematics, correlating longitudinal momentum fraction with momentum transfer and, at zero skewness, with transverse spatial structure. For a spin-0 target there is a single chiral-even quark GPD, usually denoted \(H_\pi^q(x,\xi,t)\), with the forward limit reducing to the pion parton distribution function and the lowest Mellin moment reproducing the electromagnetic form factor. In the valence sector, the literature combines symmetry-preserving continuum constructions based on Dyson–Schwinger and Bethe–Salpeter equations, covariant chiral quark models, light-front overlap representations, and direct lattice-QCD extractions at zero skewness, yielding a technically coherent but method-dependent picture of pion structure [1501.03699] [2310.10579].

## 1. Operator definition and kinematics

For a spinless target, the quark GPD is defined through the light-front bilocal vector correlator. In one standard convention,
$$
H(x,\xi,t) = \frac{1}{2} \int \frac{\textrm{d}z^-}{2\pi} \, e^{i x P^+ z^-}
\left\langle P+\frac{\Delta}{2} \left| \bar{q}\left(-\frac{z}{2}\right)\gamma^+q\left(\frac{z}{2}\right) \right |P-\frac{\Delta}{2}\right\rangle_{z^+=0,\,z_\perp=0},
$$
with average momentum \(P=(P'+P)/2\), momentum transfer \(\Delta=P'-P\), skewness
$$
\xi=-\frac{\Delta^+}{2P^+},
$$
and invariant momentum transfer \(t=\Delta^2\) [1501.03699]. The support is \(x\in[-1,1]\). The regions \(|x|>\xi\) and \(|x|<\xi\) are the DGLAP and ERBL domains, respectively, and time-reversal invariance enforces evenness in \(\xi\), \(H(x,\xi,t)=H(x,-\xi,t)\) [2411.18130].

At zero skewness, \(H_\pi^q(x,0,t)\) admits the standard impact-parameter interpretation. The corresponding transverse density is
$$
q(x,\boldsymbol{b}_\perp)=
\int \frac{d^2 \boldsymbol{\Delta}_\perp}{(2\pi)^2}\,
e^{-i \boldsymbol{\Delta}_\perp\cdot \boldsymbol{b}_\perp}\,
H_\pi^q\!\left(x,0,-\boldsymbol{\Delta}_\perp^2\right),
$$
so that \(\xi=0\) is the kinematic limit in which longitudinal momentum and transverse position can be discussed simultaneously [2310.10579].

## 2. Exact constraints and limiting relations

Three structural constraints organize the theory of valence pion GPDs. The first is the forward limit,
$$
H^\pi(x,0,0)=
\begin{cases}
q_v^\pi(x), & x>0,\\
-\bar q^\pi(-x), & x<0,
\end{cases}
$$
which identifies the positive-\(x\) region with quarks and the negative-\(x\) region with antiquarks [1501.03699]. The second is the electromagnetic sum rule,
$$
\int_{-1}^{1} dx\, H^\pi(x,\xi,t)=F_\pi(t),
$$
which is \(\xi\)-independent by polynomiality [1507.05824]. The third is polynomiality of Mellin moments,
$$
\mathcal{M}_n(\xi,t)=\int dx\,x^n H(x,\xi,t),
$$
which, for the spin-0 pion, yields even polynomials in \(\xi\) of degree at most \(n+1\) [1501.03699].

The soft-pion limit is a further nontrivial constraint. In the form used in Dyson–Schwinger studies,
$$
H(x,\xi=1,t=0)=\frac{1}{2}\,\phi\!\left(\frac{1+x}{2}\right),
$$
so maximal skewness connects the pion GPD to the pion distribution amplitude [1507.05824]. A central point in the continuum literature is that this relation is not recovered by a simplified algebraic triangle model unless the axial-vector Ward–Takahashi identity is enforced; rainbow–ladder-consistent numerical DSE/BSE solutions restore the theorem [1501.03699]. In the nonlocal chiral quark model, the strict chiral-limit statement at \(\xi=1,t=0\) is that the isoscalar pion GPD vanishes pointwise, which yields a vanishing second Mellin moment in the soft-pion-theorem context, consistent with Polyakov–Weiss results [2411.18130].

These constraints also delimit recurrent misconceptions. A model may reproduce a plausible forward PDF or form factor while still violating the soft-pion theorem or breaking valence symmetry. Conversely, continuity at the crossover \(x=\pm\xi\), correct support, and \(\xi\)-polynomiality are not optional technical details but defining consistency conditions.

## 3. Dyson–Schwinger and Bethe–Salpeter constructions

A major continuum route begins by computing Mellin moments from the triangle diagram in an impulse approximation and then reconstructing the full GPD through double distributions. In the algebraic DSE/BSE-inspired model, the moments are generated from
$$
2( P \cdot n )^{m+1} \, \mathcal{M}_m(\xi,t)
=
\mathrm{Tr}\!\int \frac{\mathrm{d}^4k}{(2\pi)^4}
(k \cdot n)^m \,
i\Gamma_\pi\,
S\!\left( k - \frac{\Delta}{2} \right)
\, i\gamma \cdot n \,
S\!\left( k+\frac{\Delta}{2} \right)
\, i\bar{\Gamma}_\pi\,
S( k - P ),
$$
with a dressed-quark propagator
$$
S(p)=\big[-i\gamma\cdot p+M\big]\Delta_M(p^2),\qquad
\Delta_M(s)=\frac{1}{s+M^2},
$$
and a pion Bethe–Salpeter vertex
$$
\Gamma_\pi(k,p)=
i\gamma_5\,\frac{M}{f_\pi}\,M^{2\nu}
\int_{-1}^{1}dz\,\rho_\nu(z)\,
\big[\Delta_M(k_{+z}^2)\big]^\nu,
\qquad
\rho_\nu(z)=R_\nu(1-z^2)^\nu,
$$
with \(\nu=1\) chosen to recover the asymptotic pion distribution amplitude [1501.03699].

The reconstructed GPD is then written in double-distribution form,
$$
H(x,\xi,t)=
\int_{|\alpha|+|\beta|\le1} d\beta\,d\alpha\,
\delta(x-\beta-\alpha\xi)\,
\left[F(\beta,\alpha,t)+\xi\,G(\beta,\alpha,t)\right].
$$
This implementation guarantees support, polynomiality, continuity at \(x=\xi\), and the form-factor sum rule by construction, and no separate D-term is introduced in that particular model [1501.03699]. The resulting \(F_\pi(t)\) and forward valence PDF were reported to agree very well with available data.

A related DSE treatment emphasized that the impulse approximation is incomplete in the valence sector. In that analysis, the triangle approximation alone breaks the \(x\leftrightarrow 1-x\) symmetry of the forward valence PDF, and additional terms corresponding to derivative insertions at the pion vertex are required to restore the symmetry analytically [1507.05824]. An earlier rainbow–ladder study sharpened this point further by arguing that the impulse approximation is generally invalid owing to omission of contributions from the gluons which bind dressed-quarks into the pion, and by proposing an improved \(\xi=0\) construction that yields a positive-definite impact-parameter distribution and a qualitatively sound picture of the pion’s dressed-quark structure at a hadronic scale [1411.6634].

Within that \(\xi=0\) construction, the valence GPD is written as
$$
H_\pi^{\rm v}(x,0,-\mathpzc{t})
=
q_\pi^{\rm v}(x)\,
\mathcal{N}(\mathpzc{t})\,
\frac{F_\pi(\mathpzc{t})}
{\left[1+\frac{\mathpzc{t}}{4M^2}(1-x)^2\right]^2},
$$
which makes explicit the nontrivial \(x\)–\(t\) correlation that is absent in a purely factorized \(q(x)F_\pi(t)\) ansatz [1503.08645]. The same line of work also relates the \(\xi=\pm1\) limit to the pion distribution amplitude through the soft-pion theorem.

## 4. Chiral and light-front quark-model realizations

The Nambu–Jona-Lasinio model provides a covariant Bethe–Salpeter realization of pion GPDs with an explicit DD\(+\)D-term structure. In the chiral limit and at \(t=0\), the model yields the simple piecewise expression
$$
H^{\pi^+\,\text{tot}}_u(x,\xi,0)
=
\frac{1}{2}\,\theta(\xi-x)\theta(\xi+x)
+\theta(x-\xi)\theta(1-x),
$$
and the isoscalar GPD obeys \(H^{I=0}(x,1,0)=0\) [1010.2974]. The model also makes polynomiality explicit through tabulated generalized form factors and supplies a low matching scale, \(Q_0^{\rm LO}\simeq0.29~\mathrm{GeV}\) and \(Q_0^{\rm NLO}\simeq0.43~\mathrm{GeV}\), from which LO and NLO evolution can be performed [1010.2974].

A complementary nonlocal chiral quark model analyzes valence pion GPDs across DGLAP and ERBL regions with two nonlocal topologies: a type-(a) contribution entering both regions and a type-(b) contribution entering only ERBL. The resulting valence GPD is continuous at \(x=\pm\xi\), although its first derivative exhibits kinks, and the model reports a total valence momentum fraction of \(\approx0.93\) at the model scale, with the deviation from unity attributed to nonlocal-current gauge-invariance issues [2411.18130]. The same framework performs one-loop evolution with \texttt{APFEL++} to \(\mu^2=4~\mathrm{GeV}^2\) and \(100~\mathrm{GeV}^2\), finding that produced sea quarks and gluons are strongly suppressed as \(\xi\) becomes nonzero and are largely confined within the ERBL region.

Light-front constituent approaches usually restrict the analysis to \(\xi=0\). In a light-cone quark model with Brodsky–Huang–Lepage wave functions and Melosh–Wigner rotation, the pion GPD is represented by a diagonal overlap,
$$
H_\pi^q(x,0,-\boldsymbol{\Delta}_\perp^{\,2})
=
\int \frac{d^2\boldsymbol{k}_\perp}{16\pi^3}\;
\psi^{*}\!\left(x,\boldsymbol{k}_\perp + \frac{1-x}{2}\boldsymbol{\Delta}_\perp\right)
\psi\!\left(x,\boldsymbol{k}_\perp - \frac{1-x}{2}\boldsymbol{\Delta}_\perp\right)\;
S(x,\boldsymbol{k}_\perp,\boldsymbol{\Delta}_\perp),
$$
with starting scale \(\mu_0^2=0.246~\mathrm{GeV}^2\) [2107.02402]. In that model, \(H(x,0,t)\) peaks at \(x\approx0.5\) at \(-t=0\), while increasing \(-t\) decreases the magnitude and shifts the peak toward higher \(x\).

Basis light-front quantization with holographic confinement, longitudinal confinement, and color-singlet NJL interactions gives another valence-only realization. For the pion, the overlap formulas at \(\zeta=0\) generate both \(H(x,0,t)\) and the chiral-odd \(E_T(x,0,t)\), and Mellin moments define generalized form factors
$$
A^q_{n0}(t)=\int_0^1 dx\,x^{n-1}H^q(x,0,t).
$$
At \(\mu^2=4~\mathrm{GeV}^2\), the model reports \(A^{\pi,q}_{20}(0)=0.244\pm0.018\), \(B^{\pi,q}_{T10}(0)=0.229\pm0.004\), and an average transverse shift \(\langle b_y^\perp\rangle_1=0.162\pm0.003~\mathrm{fm}\), in reasonable agreement with lattice-QCD benchmarks quoted in the same work [2110.05048].

## 5. Lattice-QCD determinations at zero skewness

Direct \(x\)-dependent lattice determinations of the valence pion GPD use Large-Momentum Effective Theory. An exploratory first study computed the valence quasi-GPD at \(\xi=0\) on a \(2+1+1\)-flavor HISQ ensemble with \(a\approx0.12~\mathrm{fm}\), \(L\approx3~\mathrm{fm}\), \(m_\pi\approx310~\mathrm{MeV}\), and boosts up to \(P_z\approx1.74~\mathrm{GeV}\) [1904.12376]. It reproduced the \(t=0\) PDF limit and found that \(\int dx\,H_v^\pi(x,0,t)\) agrees with the pion form factor from the \(z=0\) matrix element within \(1\sigma\), but the errors were too large to discriminate among different \(x\)–\(t\) parametrizations.

The first physical-pion-mass \(x\)-dependent calculation at \(\xi=0\) used clover valence fermions on \(2+1+1\) HISQ ensembles with \(a\approx0.09~\mathrm{fm}\), \(L\approx5.5~\mathrm{fm}\), and boost momentum \(1.73~\mathrm{GeV}\), together with hybrid renormalization and NNLO LaMET matching to \(\overline{\mathrm{MS}}\) at \(\mu=2~\mathrm{GeV}\) [2310.10579]. The calculation covered \(Q^2=-t\in\{0,0.19,0.39,0.77,0.97\}~\mathrm{GeV}^2\), reproduced \(F_\pi(Q^2)\), predicted higher moments \(A_{30}\) and \(A_{40}\), and provided the first lattice-QCD pion tomography \(q^\pi(x,b)\). A central result was that the transverse profile narrows as \(x\) increases, indicating stronger transverse localization of large-\(x\) valence partons.

A subsequent lattice study with \(a=0.04~\mathrm{fm}\) and valence pion mass \(300~\mathrm{MeV}\) analyzed both symmetric and asymmetric momentum-transfer frames, showed the equivalence of both at the level of Lorentz-invariant amplitudes, and matched quasi-GPDs with NNLO perturbative corrections plus leading renormalon and renormalization-group resummations [2407.03516]. In that work, the matched \(H(x,0,t)\) was stated to be reliable for \(x\gtrsim0.2\) at \(P_z=1.937~\mathrm{GeV}\), and its \(t\)-dependence at fixed \(x\) was well described by a monopole form,
$$
H(x,t)=\frac{H(x,0)}{1-t/M^2(x)}.
$$
The effective transverse radius extracted from this monopole decreases with \(x\), again supporting the picture that high-\(x\) quarks are more localized in the transverse plane [2407.03516].

## 6. Form factors, evolution, and phenomenological status

Electromagnetic and gravitational form factors are the most direct Mellin-moment observables associated with valence pion GPDs. In the nonlocal chiral quark model, the first Mellin moment is numerically \(\xi\)-independent and yields the pion electromagnetic form factor with charge normalization \(F_{\pi^+}(0)=1\), while the second Mellin moment gives \(A_\pi(t)\) and \(D_\pi(t)\); the model reports
$$
A_{\pi^+}(0)+D_{\pi^+}(0)\approx0.04,
$$
a charge radius \(\sqrt{\langle r^2\rangle_{F_{\pi^+}}}\approx0.544~\mathrm{fm}\), and gravitational radii \(\sqrt{\langle r^2\rangle_{A_\pi}}\approx0.436~\mathrm{fm}\) and \(\sqrt{\langle r^2\rangle_{D_\pi}}\approx0.436~\mathrm{fm}\) [2411.18130]. In that same framework, the negative D-term is interpreted as consistent with chiral expectations, while the omission of meson loops is noted as a likely source of underestimated D-term slopes.

A more explicitly phenomenological program constructs pion GPDs from forward PDFs and a controlled off-forward extension that satisfies support, positivity, forward limits, polynomiality, and continuity by construction. In the DGLAP region, the quark GPD takes the master-equation form
$$
H_\pi^q(x,\xi,t)\big|_{\rm DGLAP}
=
\sqrt{q_\pi(x_{\rm in})\,q_\pi(x_{\rm out})}\;
\Phi_\pi^q(x,\xi,t),
$$
with \(x_{\rm in}=(x+\xi)/(1+\xi)\) and \(x_{\rm out}=(x-\xi)/(1-\xi)\), and the soft-pion theorem is used to fix the odd-in-\(x\) D-term [2110.06052]. That analysis also extends to gluon GPDs and Compton form factors, finding that next-to-leading order corrections are significant even in the so-called valence region.

At zero skewness, a recent global QCD analysis of pion electromagnetic form-factor data adopts the factorized ansatz
$$
H_v^q(x,t,Q^2)=q_v(x,Q^2)\,\exp\!\big[t\,f_v^q(x)\big],
$$
with
$$
f_v^q(x)=\alpha'(1-x)^3\ln\frac{1}{x}+B(1-x)^3+A\,x(1-x)^2.
$$
Using xFitter, JAM21, and MAP23 pion PDFs at \(Q^2=4~\mathrm{GeV}^2\), the study finds that only two parameters are sufficient to describe the data and adopts the MAP23-based fit with
$$
\alpha'=1.196\pm0.024~\mathrm{GeV}^{-2},\qquad
A=2.384\pm0.148~\mathrm{GeV}^{-2},
$$
and total \(\chi^2/\mathrm{ndf}=161.144/139\) [2508.15073]. In impact-parameter space this ansatz yields
$$
q_v^\pi(x,\boldsymbol{b}_T)
=
\frac{q_v(x,Q^2)}{4\pi f_v(x)}
\exp\!\left[-\frac{\boldsymbol{b}_T^2}{4f_v(x)}\right],
\qquad
\langle b_T^2\rangle(x)=4f_v(x),
$$
making the transverse-width interpretation completely explicit.

Across approaches, several limitations recur. Valence-only constructions neglect explicit sea-quark and gluon degrees of freedom at the input scale; simplified algebraic DSE models may fail to enforce the axial-vector Ward–Takahashi identity; \(\xi=0\) light-front models do not by themselves test full polynomiality; and some continuum studies do not perform detailed positivity-bound checks or impact-parameter analyses [1501.03699]. These are not contradictions so much as delimitations of scope. Taken together, the literature supports a stable qualitative picture: the valence pion GPD has the correct forward and form-factor limits, exhibits nontrivial \(x\)–\(t\) correlations, narrows in transverse space as \(x\) increases, and provides the common framework linking pion PDFs, form factors, distribution amplitudes, and three-dimensional tomography.

Source: https://www.emergentmind.com/topics/valence-pion-generalized-parton-distributions-gpds