---
title: 'Transition GPDs: Nucleon & Resonance Dynamics'
url: https://www.emergentmind.com/topics/transition-generalized-parton-distributions-gpds
type: topic
---

# Transition GPDs: Nucleon & Resonance Dynamics

Transition generalized parton distributions (transition GPDs) are the non-forward extensions of the GPD framework in which the initial and final hadronic states are different. In their most common baryonic usage they describe \(N\to N^\ast\) and \(N\to \Delta\) transitions through gauge-invariant light-front quark and gluon operators; in broader non-diagonal settings they also encompass flavor-changing nucleon transitions such as \(p\leftrightarrow n\) and hadron-to-two-hadron transitions such as \(\pi\to\pi\pi\). They enter collinear-factorized amplitudes of hard exclusive reactions, interpolate between transition form factors and higher local matrix elements of the QCD energy–momentum tensor (EMT), and provide a framework for tomographic and mechanical characterization of resonances and other non-diagonal hadronic transitions. They are distinct from transition distribution amplitudes (TDAs), which are based on three-quark operators and govern backward kinematics rather than the forward generalized Bjorken regime [2405.15386][2401.13207][2410.11388][1109.0179].

## 1. Scope, terminology, and relation to adjacent frameworks

The term “transition GPDs” is used in more than one sense. In the baryon-resonance program it denotes bilocal quark or gluon correlators between the ground-state nucleon and an excited baryon, for example \(N\to \Delta(1232)\), \(N\to P_{11}(1440)\), \(N\to D_{13}(1520)\), or \(N\to S_{11}(1535)\). In hard exclusive pion–nucleon scattering it can also denote flavor off-diagonal nucleon transitions such as \(p\to n\), where the hadronic states differ but remain in the nucleon multiplet. A further extension is hadron-to-two-hadron transition GPDs, such as \(\pi\to\pi\pi\), which unify hadron-to-resonance transitions like \(\pi\to \rho\), \(\pi\to f_0\), and \(\pi\to f_2\) within a single non-diagonal object depending on the invariant mass and decay angles of the two-hadron system [2405.15386][2401.13207][2410.11388].

| Class | Operator content | Typical regime |
|---|---|---|
| \(N\to N^\ast\), \(N\to\Delta\) transition GPDs | Gauge-invariant light-front bilocal quark/gluon operators | Forward hard exclusive electroproduction |
| \(p\leftrightarrow n\) transition GPDs | Chiral-even bilocal quark operators | Hard exclusive pion–nucleon scattering |
| Hadron-to-two-hadron transition GPDs | Bilocal light-cone quark operators between one- and two-hadron states | Non-diagonal DVCS |
| TDAs | Light-cone three-quark operators | Backward meson electroproduction |

A recurrent source of confusion is the relation between transition GPDs and TDAs. The 2011 review on GPDs and TDAs emphasizes that TDAs are baryon-to-meson matrix elements of tri-local three-quark operators and appear in backward kinematics, whereas bilocal transition GPDs enter forward-type factorization formulae. The distinction is structural rather than terminological: the two objects obey different operator definitions, different symmetry constraints, and different factorization theorems [1109.0179].

## 2. Operator definitions and kinematic variables

For baryon resonances, the twist-2 light-ray quark operators are defined on the light front with a lightlike vector \(n\) and a Wilson line ensuring gauge invariance. The chiral-even operators are
\[
\mathcal{O}^{(+)}_\Gamma(\tau)=\bar q\!\left(-\frac{\tau n}{2}\right)\Gamma\,q\!\left(\frac{\tau n}{2}\right),
\qquad
\Gamma=\gamma^+,\ \gamma^+\gamma^5,
\]
and the chiral-odd operator is
\[
\mathcal{O}^{(\mathrm{odd})}_T(\tau)=\bar q\!\left(-\frac{\tau n}{2}\right) i\sigma^{+i} q\!\left(\frac{\tau n}{2}\right),
\qquad i=1,2.
\]
With \(P=(p+p')/2\), \(\Delta=p'-p\), \(\xi=-\Delta^+/(2P^+)\), and \(t=\Delta^2<0\), the vector transition GPDs are defined by
\[
\int\frac{dz^-}{4\pi}e^{ixP^+z^-}
\langle N^\ast(p')|\bar q(-\tfrac{z^-}{2})\gamma^+ q(\tfrac{z^-}{2})|N(p)\rangle
=
\frac{1}{2P^+}\,
\bar U_{N^\ast}(p')
\left[
H^{N\to N^\ast}\gamma^+
+
E^{N\to N^\ast}\frac{i\sigma^{+\alpha}\Delta_\alpha}{2M_N}
\right]
U_N(p),
\]
with analogous definitions for \((\tilde H,\tilde E)\) and for the chiral-odd set \((H_T,E_T,\tilde H_T,\tilde E_T)\). Transition GPDs have support \(x\in[-1,1]\) and the same DGLAP/ERBL partition as ordinary nucleon GPDs: \(|x|>\xi\) in the DGLAP region and \(|x|<\xi\) in the ERBL region, with continuity at \(x=\pm \xi\). Their scale dependence follows the usual QCD evolution equations for twist-2 light-ray operators [2405.15386].

The number of independent distributions depends on the quantum numbers of the final state. For \(N\to \Delta(1232)\), \(J^P=3/2^+\), 16 twist-2 quark GPDs are required: four vector, four axial, and eight tensor. For \(N\to P_{11}(1440)\), \(J^P=1/2^+\), the counting matches the nucleon case: two vector and two axial GPDs. For \(N\to D_{13}(1520)\), \(J^P=3/2^-\), the chiral-even sector contains \(4+4\) GPDs, while \(N\to S_{11}(1535)\), \(J^P=1/2^-\), contains \(2+2\) chiral-even GPDs [2405.15386].

In hard exclusive pion–nucleon scattering with charged pion beams, the diffracted nucleon necessarily changes flavor, so the relevant transition GPDs are \(p\to n\) or \(n\to p\). At leading twist only chiral-even quark GPDs contribute, because the pion is a pseudoscalar and chiral symmetry holds at leading power. The flavor structure reduces to the isovector combination
\[
\mathcal{F}^{[u\bar d]}_{pn}
=
\mathcal{F}^{[d\bar u]}_{np}
=
\mathcal{F}_p^u-\mathcal{F}_p^d,
\]
with \(\mathcal{F}\in\{F,\tilde F,H,E,\tilde H,\tilde E\}\) [2401.13207].

## 3. Mellin moments, polynomiality, and physical interpretation

Lorentz covariance implies polynomiality of Mellin moments. For a generic baryonic transition,
\[
\int_{-1}^{1}dx\,x^{n-1}\,H^{N\to N^\ast}(x,\xi,t)
=
\sum_{k=0}^{n}A^{N\to N^\ast}_{n,k}(t)\,\xi^k,
\]
with analogous relations for \(E,\tilde H,\tilde E\) and the chiral-odd sector. The \(n=1\) moments reduce to local vector and axial currents and therefore to transition electromagnetic and axial form factors. For \(N\to\Delta\), the vector moments relate to the Jones–Scadron multipoles \(G_M^\ast(t)\), \(G_E^\ast(t)\), and \(G_C^\ast(t)\), while the axial moments relate to the Adler form factors \(C_5^A\), \(C_6^A\), \(C_3^A\), and \(C_4^A\). For \(N\to P_{11}\), the vector first moments give the Dirac and Pauli transition form factors, and the axial moments are constrained by PCAC and pion-pole dominance [2405.15386].

The \(n=2\) moments involve the quark part of the QCD EMT and generate gravitational transition form factors \(A(t)\), \(B(t)\), and \(D(t)\), or alternatively \(M_2(t)\) and \(d_1(t)\). In the transition context these quantities encode mass/energy distributions, angular momentum transfer, and mechanical properties such as pressure and shear. For \(N\to\Delta\), the second moment of the dominant vector GPD obeys
\[
\int_{-1}^{1}dx\,x\,G_1(x,\xi,0)
=
\frac{8}{3}\,J^{u-d}_{p\to \Delta^+}.
\]
At zero skewness, the transverse Fourier transform
\[
q^{N\to N^\ast}(x,\mathbf b_T)
=
\int \frac{d^2\Delta_T}{(2\pi)^2}
e^{-i\mathbf b_T\cdot\Delta_T}
H^{N\to N^\ast}(x,\xi=0,-\Delta_T^2)
\]
defines an impact-parameter density that is diagonal in \(x\) and \(\mathbf b_T\) on the light front while connecting different hadronic states through operator quantum-number transfer. The mechanical interpretation uses the usual \(D\)-term relations for pressure \(p(r)\) and shear \(s(r)\), with the stability condition \(\int_0^\infty dr\,r^2p(r)=0\) [2405.15386].

In hadron-to-two-hadron transitions the polynomiality structure is richer because the GPD depends, in addition to \((x,\xi,t)\), on variables describing the final two-hadron system. For the unpolarized isoscalar \(\pi\to\pi\pi\) transition GPD,
\[
\int_{-1}^1 dx\, x^N H^{\pi\to\pi\pi}_S(x,\xi,t;W_{\pi\pi}^2,\alpha,t')
=
\sum_{i,j\ge 0}^{i+j\le N}
(-2\xi)^i[(1-\xi)\alpha]^{N-i-j}\,
\mathcal A_{N+1,i,j}(t;W_{\pi\pi}^2,t'),
\]
so the Mellin moments are polynomial in \(\xi\) and \(\alpha\) of order \(N\). Here time reversal does not constrain transition GPDs, and both even and odd powers of \(\xi\) can appear in Mellin moments [2410.11388].

## 4. The \(N\to\Delta\) channel: complete leading-twist definition and forward-limit transition densities

The \(N\to\Delta\) channel has become the principal testing ground for structural questions in transition GPD theory. A 2024 reanalysis of the leading-twist chiral-even sector showed that the vector and axial \(N\to\Delta\) correlators each require four independent covariant structures, not three. In the vector case, the complete basis contains an additional structure proportional to \(n^\alpha\), and likewise in the axial case. The corresponding GPDs, usually denoted \(h_X\) and \(C_X\), are specific to the nonlocal operator and have vanishing first Mellin moments:
\[
\int_{-1}^{1}dx\,h_X(x,\xi,t)=0,
\qquad
\int_{-1}^{1}dx\,C_X(x,\xi,t)=0.
\]
Their inclusion is required to ensure completeness and polynomiality of higher moments. Earlier parametrizations either omitted these \(n^\alpha\)-proportional terms or introduced a fourth term that becomes linearly dependent after contraction with the light-cone vector. Light-front multipole analysis and crossed-channel \(SO(3)\) partial-wave analysis independently confirm the necessity of the full four-structure basis in both vector and axial sectors [2501.00185].

This correction also clarifies a misconception imported from diagonal GPDs. In non-diagonal \(N\to\Delta\) matrix elements, parity and hermiticity still constrain the decomposition, but time reversal does not reduce the vector sector to the diagonal-nucleon counting. Current conservation reduces only the local vector current basis; it does not remove the extra twist-2 nonlocal structures. The complete leading-twist chiral-even definition is therefore not merely a change of convention but a structural requirement for phenomenology and model building [2501.00185].

Transitions between unequal-mass baryons admit a special light-front forward limit. Setting
\[
\xi=0,\qquad \Delta^+=0,\qquad \boldsymbol{\Delta}_T=0
\]
leaves a nonzero light-front energy transfer
\[
\Delta^-=\frac{m'^2-m^2}{2P^+}\neq 0,
\]
while the invariant momentum transfer remains \(t=0\). In this limit one can define forward-limit transition parton densities,
\[
f_I(x)\equiv H_I(x,\xi=0,t=0).
\]
For \(N\to\Delta\) the additional \(X\)-type structure produces a new tensor-polarized transition parton density,
\[
f_Q(x)\equiv H_X(x,\xi=0,t=0),
\]
proportional to the \(1/2\to 3/2\) spin-transition tensor. Current conservation implies the zero-sum rule
\[
\int_{-1}^{1}dx\,f_Q(x)=0,
\]
and the second moment is related to a local form factor \(F_4\),
\[
\int_{-1}^{1}dx\,x\,f_Q(x)=2F_4(t).
\]
A chiral quark–soliton model estimate based on large-\(N_c\) gives \(f_Q(x)\) as an even function of \(x\) that changes sign at \(x\sim 0.1\) for \(x>0\); the paper states that its magnitude is at least an order of magnitude smaller than typical nucleon PDFs with \(\int dx\,f(x)=\mathcal O(1)\), while also emphasizing that the \(x\to 0\) region is not reliably predicted in that approximation [2507.18402].

## 5. Factorization theorems and process classes

In generalized Bjorken kinematics, transition GPDs enter hard exclusive amplitudes in direct analogy with ordinary DVCS and DVMP. For transition DVCS,
\[
T_{\mathrm{DVCS}}^{N\to N^\ast}(Q^2,\xi,t)
\propto
\sum_q e_q^2\int_{-1}^{1}dx\,
C^{(q)}(x,\xi,Q^2)\,
H^{N\to N^\ast}(x,\xi,t)
+\text{(E, }\tilde H,\tilde E\text{ contributions)}.
\]
In transition DVMP, the helicity amplitudes are convolutions of hard subprocess amplitudes with transition GPDs. For \(\gamma^\ast p\to \pi^- \Delta^{++}\), the longitudinal twist-2 amplitudes depend on the axial transition GPDs \(\tilde G_i\), the transverse twist-3 amplitudes depend on tensor GPDs \(G_{Ti}\), and the longitudinal channel contains a prominent pion-pole contribution. Large-\(N_c\) relations connect several \(N\to\Delta\) transition GPD combinations to isovector nucleon GPDs and transversity GPDs [2405.15386].

A complementary extraction strategy uses hard exclusive pion–nucleon scattering. In the “two-stage” single diffractive hard exclusive process paradigm, the amplitudes for
\[
N+\pi\to N'+\gamma\gamma
\qquad\text{and}\qquad
N+\pi\to N'+\ell^+\ell^-
\]
factorize into the pion distribution amplitude, transition GPDs, and perturbatively calculable short-distance coefficients. The diphoton channel is complementary to the Drell–Yan-type dilepton channel because its hard kernels contain a “special integral” with a dynamically generated pole \(x_p(\xi,z,\theta)\), which scans the DGLAP region \(|x|\ge \xi\) and enhances sensitivity to the \(x\)-dependence of the transition GPDs. By contrast, the exclusive dilepton channel reduces at leading order to moments such as
\[
\tilde F_0^\pm(\xi,t)=P\int_{-1}^1dx\,\frac{\tilde F^\pm(x,\xi,t)}{x-\xi},
\]
together with crossover-line values \(\tilde F^\pm(\xi,\xi,t)\), so its angular dependence is universal and does not by itself resolve shadow-GPD ambiguities. At leading order the timelike photon is purely longitudinal, and only the polarized transition GPD contributes in the dilepton channel [2401.13207].

Hadron-to-two-hadron transition GPDs furnish a further non-diagonal generalization. For the spinless process \(\gamma^\ast \pi\to \gamma\pi\pi\), the hadronic tensor is parameterized by transition Compton form factors
\[
\mathcal H^{\pi\to\pi\pi}(\xi,t;W_{\pi\pi}^2,\Omega_\pi^\ast)
=
\int_{-1}^1 dx\, C^+(x,\xi)\,H^{\pi\to\pi\pi}(x,\xi,t;W_{\pi\pi}^2,\Omega_\pi^\ast),
\]
\[
\tilde{\mathcal H}^{\pi\to\pi\pi}(\xi,t;W_{\pi\pi}^2,\Omega_\pi^\ast)
=
\int_{-1}^1 dx\, C^-(x,\xi)\,\tilde H^{\pi\to\pi\pi}(x,\xi,t;W_{\pi\pi}^2,\Omega_\pi^\ast),
\]
with
\[
C^\pm(x,\xi)=\frac{1}{x-\xi+i\epsilon}\pm \frac{1}{x+\xi-i\epsilon}.
\]
The framework establishes soft-pion theorems at the \(2\pi\) threshold, partial-wave expansions in the \(\pi\pi\) decay angles, and a dispersive Muskhelishvili–Omnès representation that ties each partial wave to \(\pi\pi\)-scattering phases. In the narrow-width limit it reduces to the usual hadron-to-resonance transition GPD picture, for example \(\pi\to \rho\) [2410.11388].

The contrast with TDAs remains sharp. TDAs are defined through matrix elements of light-cone three-quark operators,
\[
\widehat O^{\alpha\beta\gamma}_{\rho\tau\chi}(z_1,z_2,z_3)
=
\varepsilon_{c_1c_2c_3}
\Psi_\rho^{c_1,\alpha}(z_1)\,
\Psi_\tau^{c_2,\beta}(z_2)\,
\Psi_\chi^{c_3,\gamma}(z_3),
\]
and govern backward processes such as backward meson electroproduction and backward DVCS, where the hard amplitude factorizes schematically as \(T_H\otimes\)TDA\(\otimes\)DA rather than coefficient function \(\otimes\) transition GPD [1109.0179].

## 6. Observables, first measurements, and outstanding systematics

The observable program for baryonic transition GPDs closely parallels that of ordinary GPDs, but with resonance reconstruction and decay angular analysis added. The key observables are differential cross sections and azimuthal modulations, beam-spin asymmetry \(A_{LU}\), longitudinal target asymmetries, Rosenbluth-separated \(\sigma_L\) and \(\sigma_T\), and spin-density matrix elements extracted from decay angular distributions. In transition DVCS, Bethe–Heitler interference gives direct sensitivity to the imaginary part of transition Compton form factors. For \(N\to\Delta\), the electroproduction amplitude factorizes into a resonance-production part and the \(\Delta\to\pi N\) decay, whose characteristic angular distributions and “autopolarization” patterns enable partial-wave separation. First JLab 12 GeV results include CLAS12 measurements of hard exclusive \(\pi^-\Delta^{++}\) electroproduction beam-spin asymmetries, where the extracted \(\sigma_{LT'}/\sigma_0\) has sign opposite to \(\pi^+n\) and larger magnitude; Hall C preliminary analyses of \(\pi^+\Delta^0\); and preliminary \(A_{LU}^{\sin\phi}\) moments for \(N\to\Delta^+\) DVCS in \(ep\to e'\gamma n\pi^+\), with clear resonance structures as functions of \(M_{n\pi^+}\) and \(-t\) [2405.15386].

The experimental program extends beyond JLab. The same white paper identifies CLAS12 luminosity upgrades and a possible 22 GeV upgrade, COMPASS/AMBER muon-beam and meson-beam measurements, J-PARC exclusive Drell–Yan with \(\pi\)-beam momenta \(10\)–\(20\) GeV, EIC/EicC far-forward detection of transition final states, and LHC ultraperipheral or fixed-target channels as relevant facilities. It also stresses the need for standardization of \(N\to\Delta\) parametrizations, complete decompositions for higher resonances, improved EMT-based mechanical interpretations, combined chiral/large-\(N_c\) EFT, lattice calculations, and systematic treatment of higher-twist corrections, radiative effects, resonance–nonresonance separation, and factorization-validity tests [2405.15386].

For pion-induced flavor-changing transitions, the exclusive diphoton and dilepton observables are argued to be physically measurable at the J-PARC and AMBER experiment energies. The diphoton channel is predicted to have larger overall rates than exclusive Drell–Yan, strong forward enhancement, and non-universal \(q_T\) shapes reflecting the \(x\)-dependent kernel, while the exclusive Drell–Yan channel has a universal \(\sin^2\theta\) shape at leading order and a large, \(q_T\)-independent single transverse target-spin asymmetry that constrains the normalization and sign of \(\tilde E\) rather than its full \(x\)-dependence [2401.13207].

In the spinless \(\pi\to\pi\pi\) case, the estimated \(e^-\pi^+\to e^-\gamma\pi^+\pi^0\) cross section near the \(\rho(770)\) resonance in JLab@12 GeV kinematics is at the level of a few picobarns. The decay-angle distributions separate longitudinal and transverse \(\rho\) polarization contributions: after integration over \(\varphi_\pi^\ast\), the longitudinal contribution scales as \(\cos^2\theta_\pi^\ast\), whereas the transverse contribution scales as \(\sin^2\theta_\pi^\ast\). The \(\varphi_\pi^\ast\) dependence develops a \(\cos(2\varphi_\pi^\ast)\) interference pattern at larger \(-t\) [2410.11388].

Taken together, these developments place transition GPDs at the intersection of exclusive reaction theory, hadron tomography, and resonance structure. The formalism now covers bilocal nucleon-to-resonance transitions, flavor-changing nucleon channels, and hadron-to-two-hadron non-diagonal processes, while the distinction from TDAs has become sharper. The principal unresolved issues are no longer definitional in the broad sense, but structural completeness in specific channels, model-independent extraction across multiple reactions, and the control of power corrections and resonance backgrounds needed to convert observables into quantitative transition tomography.

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