---
title: Double-Helicity Flip Structure Function
url: https://www.emergentmind.com/topics/double-helicity-flip-structure-function
type: topic
---

# Double-Helicity Flip Structure Function

Searching arXiv for recent and foundational papers on double-helicity-flip structure functions and gluon transversity.
arXiv search query: double-helicity-flip structure function gluon transversity generalized parton distributions
The **double-helicity-flip structure function** is a leading-twist observable associated with the transfer of two units of helicity in hard scattering, and it appears in distinct but related forms in forward and off-forward QCD descriptions. In forward deep inelastic scattering from polarized spin-one or higher targets, it is conventionally denoted $\Delta(x,Q^2)$ and is identified with the double-helicity-flip helicity amplitude component $A_{+-,-+}=A_{-+,+-}$ of the hadronic tensor; it exists only for targets of spin $J\ge 1$ and does not mix with quark distributions at leading twist [1606.04505]. In the generalized parton distribution framework for off-forward processes, the corresponding physics is carried by gluon helicity-flip, or gluon transversity, GPDs $H_T^g$, $E_T^g$, $\tilde H_T^g$, and $\tilde E_T^g$, which encode $\Delta\lambda_g=2$ transitions and are nonzero for a spin-$1/2$ nucleon only away from the forward limit, where orbital angular momentum in the $t$ channel compensates the two-unit gluon helicity transfer [1407.6119]. The topic therefore links inclusive DIS on spin-one targets, exclusive reactions such as deeply virtual Compton scattering, and nonperturbative modeling via crossed-channel partial waves, dual parametrization, and lattice QCD.

## 1. Definition and helicity content

In the forward DIS formulation for a polarized spin-one target, the hadronic tensor is written as  
\[
W_{\mu\nu}(p,q,E',E) = \frac{1}{4\pi} \int d^4x\, e^{iq\cdot x} \langle p',E' | [j_\mu(x),j_\nu(0)]|p,E\rangle,
\]
with polarization vectors $E$ and $E'$ satisfying $p^{(\prime)}\cdot E^{(\prime)}=0$ and $E^{(\prime)2}=-1$. Its polarization dependence factors as  
\[
W_{\mu\nu}(p,q,E,E')=E'^{*\alpha} E^\beta W_{\mu\nu,\alpha\beta}(p,q),
\]
and the helicity decomposition
\[
W_{\mu\nu,\alpha\beta}(p,q)=\sum_{hH,h'H'}P(hH,h'H')_{\mu\nu,\alpha\beta}A_{hH,h'H'}(p,q)
\]
isolates the double-helicity-flip component through
\[
A_{+-,-+}=A_{-+,+-}\equiv \Delta(x,Q^2).
\]
The corresponding contribution to the hadronic tensor is
\[
W_{\mu\nu,\alpha\beta}^{\Delta=2}=\left[P(+-,-+)_{\mu\nu,\alpha\beta}+P(-+,+-)_{\mu\nu,\alpha\beta}\right]\Delta(x,Q^2).
\]
This channel vanishes if $E'=E$ or if the target spin is averaged, so it requires specific polarization combinations [1606.04505].

The defining physical statement is that two units of helicity are transferred between the probe and the target. For a spin-$1/2$ target this is impossible at twist-2 in the forward limit, which is why the forward structure function $\Delta(x,Q^2)$ exists only for targets of spin $J\ge 1$ [1606.04505]. By contrast, in off-forward kinematics the same two-unit helicity-transfer mechanism is realized for the nucleon through gluon helicity-flip GPDs, because the $t$ channel can carry orbital angular momentum [1407.6119].

A useful distinction follows. In forward DIS on spin-one systems, the double-helicity-flip structure function is an inclusive tensor observable. In exclusive off-forward processes, the same helicity content is encoded in specific helicity amplitudes or Compton form factors built from gluon transversity GPDs. This suggests that “double-helicity-flip structure function” denotes a family of observables unified by the $\Delta\lambda=2$ selection rule rather than a single universal scalar object.

## 2. Operator basis and partonic interpretation

At leading twist in the forward spin-one case, the double-helicity-flip channel arises from a tower of purely gluonic operators in the light-cone OPE,
\[
\frac{1}{2}\mathcal{T}_{\mu\nu}^{\Delta=2}(q) = \sum_{n=2,4,\ldots} \frac{2^n q^{\mu_1}\ldots q^{\mu_n}}{(Q^2)^n}C_n(Q^2,\mu^2)\mathcal{O}_{\mu\nu\mu_1\ldots\mu_n}(\mu^2),
\]
with
\[
\mathcal{O}_{\mu\nu\mu_1 \ldots \mu_{n}}(\mu^2)= S\left[G_{\mu\mu_1} \overleftrightarrow{D}_{\mu_3}\ldots \overleftrightarrow{D}_{\mu_{n}}G_{\nu \mu_2}\right],
\]
where $S$ denotes symmetrization and trace subtraction. The Wilson coefficients are
\[
C_n(Q^2,\mu^2)= -\frac{\alpha_s(Q^2)}{2\pi}\textrm{Tr} \mathcal{Q}^2 \frac{2}{n+2}.
\]
Moments of the structure function are
\[
M_n(Q^2)=\int_0^1 dx\, x^{n-1}\Delta(x,Q^2)=C_n(Q^2,Q^2)\frac{A_n(Q^2)}{2},
\]
with reduced matrix elements $A_n$ parameterizing the forward matrix element of the gluonic operators [1606.04505].

The parton-model relation is given by
\[
\Delta(x,Q^2) = -\frac{\alpha_s(Q^2)}{2\pi}\textrm{Tr}\mathcal{Q}^2\,x^2 \int_x^1 \frac{dy}{y^3}\,\delta G(y,Q^2),
\]
where
\[
\delta G(x,\mu^2) = g_{\hat{x}}(x,\mu^2)-g_{\hat{y}}(x,\mu^2).
\]
Here $\delta G$ measures the difference between linearly polarized gluons along transverse $\hat x$ and $\hat y$ directions in a transversely polarized target, making explicit that the double-helicity-flip structure function probes gluonic quadrupole structure rather than helicity-nonflip spin content [1606.04505].

In the off-forward nucleon case, the relevant twist-2 gluon helicity-flip operator is the symmetric traceless tensor projected as $\mathbb{S}\, G^{+i}G^{j+}$, and its matrix element is parameterized by four gluon transversity GPDs:
\[
\frac{1}{P^+} \int \frac{d\lambda}{2\pi} e^{i x P^+ \lambda}
\langle p'| \mathbb{S}\, G^{+i}(-\lambda n/2)\, G^{j+}(\lambda n/2) |p\rangle
= \mathbb{S}\; \frac{1}{2 P^+} \frac{P^+ \Delta^j - \Delta^+ P^j}{2 m P^+}\,
\bar U(p')\Big[
H_T^g i \sigma^{+i}
+ \tilde H_T^g \frac{P^+ \Delta^i - \Delta^+ P^i}{m^2}
+ E_T^g \frac{\gamma^+ \Delta^i - \Delta^+ \gamma^i}{2 m}
+ \tilde E_T^g \frac{\gamma^+ P^i - P^+ \gamma^i}{m}
\Big] U(p).
\]
These distributions are real-valued, $H_T^g$, $E_T^g$, and $\tilde H_T^g$ are even in $\xi$, $\tilde E_T^g$ is odd in $\xi$, and all are even in $x$ by charge conjugation [1407.6119].

## 3. Selection rules, forward limits, and target-spin dependence

The most important selection rule is the forward-limit restriction. For a spin-$1/2$ target, the forward gluon helicity-flip distribution vanishes: a nucleon cannot absorb two units of gluon helicity at zero momentum transfer without orbital angular momentum [1407.6119]. This is why there is no forward twist-2 gluon transversity distribution in the nucleon. Off forward, however, the $t$-channel exchange can carry orbital angular momentum and compensate $\Delta\lambda_g=2$, allowing nonzero matrix elements and making exclusive observables sensitive to gluon transversity.

For spin-one targets the situation is different. The forward double-helicity-flip structure function $\Delta(x,Q^2)$ exists precisely because the target can absorb two units of helicity at leading twist. The lattice-QCD study of the $\phi$ meson emphasized this point explicitly: the observable exists only for targets of spin $J\ge 1$ and does not mix with quark distributions at leading twist [1606.04505].

The spin-one GPD case provides a complementary perspective. For the deuteron, the chiral-odd quark helicity-flip correlator involves nine leading-twist transversity GPDs $H_i^{qT}(x,\xi,t)$, and helicity transitions with $\Delta\lambda=\pm 2$ are nonzero only if the target can supply two units of angular momentum. In a light-cone convolution model this occurs through the deuteron D-wave component and nonzero momentum transfer; with a pure S-wave, $\Delta\lambda=\pm 2$ amplitudes vanish [1902.03681]. This is consistent with the broader rule that double-helicity-flip signals require either intrinsic target spin $J\ge 1$ in the forward limit or off-forward orbital angular momentum in the crossed channel.

Discrete symmetries strongly constrain these observables. In the helicity-flip GPD sector, hermiticity and time-reversal invariance imply real-valued GPDs; support is $x\in[-1,1]$; Mellin moments satisfy polynomiality; and the $\xi$-parity assignments distinguish the even and odd distributions [1407.6119]. A plausible implication is that any phenomenological parametrization not organized around these symmetry constraints will fail to reproduce the correct analytic structure.

## 4. Crossed-channel SO(3) analysis and partial-wave structure

A central development for the off-forward double-helicity-flip sector is the crossed-channel ${\rm SO}(3)$ partial-wave analysis of helicity-flip GPDs. The strategy is to analytically continue Mellin moments to the cross channel, construct $N\bar N$ GDAs, and expand matrix elements with definite operator helicity $J_3$ in Wigner $d$-functions $d^J_{J_3,|\lambda'-\lambda|}(\theta)$ [1407.6119]. This identifies the linear combinations of GPDs that are diagonal in the $t$-channel partial-wave basis and are therefore the natural objects for modeling.

For quark helicity flip, the explicit Wigner functions used are
\[
d^J_{\pm1,0}(\theta) = (\pm1)\frac{1}{\sqrt{J(J+1)}} \sin\theta\, P'_J(\cos\theta),
\]
and
\[
d^J_{\pm1,1}(\theta) = \frac{1}{J(J+1)} (1 \pm \cos \theta)\Big[ P'_J(\cos \theta) + \cos \theta\, P''_J(\cos \theta) \mp P''_J(\cos \theta) \Big].
\]
Crossing back to DVCS kinematics gives
\[
\cos \theta \to \frac{1}{\xi \beta} + O(1/Q^2), \qquad \beta=\sqrt{1-\frac{4m^2}{t}}.
\]
In practice, hadron masses are neglected for helicity counting, but retained in Dirac spinors to count tensor structures [1407.6119].

The main result for quark transversity GPDs is the identification of combinations suited for expansion in derivatives of Legendre polynomials. The combinations
\[
\tau\, \tilde H_T^q(x, \xi, t) - \frac{1}{2} E_T^q(x, \xi, t),
\qquad
- H_T^q(x, \xi, t) + \tau\, \tilde H_T^q(x, \xi, t) - \frac{1}{2} E_T^q(x, \xi, t)
\]
are expanded in $P'_J(1/\xi)$, while
\[
H_T^q(x, \xi, t) + \tau\, \tilde H_T^q(x, \xi, t) \pm \tau\, \tilde E_T^q(x, \xi, t)
\]
are expanded in
\[
P'_J(1/\xi) + \frac{1 \mp \xi}{\xi}\, P''_J(1/\xi),
\]
with $\tau \equiv \Delta^2/(4m^2)$ [1407.6119].

The same pattern applies in the gluon sector, but with operator helicity $J_3=0,\pm2$, so the relevant partial waves involve $d^J_{\pm2,0}(\theta)$, $d^J_{\pm2,1}(\theta)$, and $d^J_{0,0}(\theta)$ [1407.6119]. This is the direct off-forward counterpart of the double-helicity-flip structure function: the $\Delta\lambda_g=2$ content is represented by those gluon transversity combinations that project onto the $J_3=\pm2$ partial waves.

The selection rules for $J^{PC}$ of the contributing $t$-channel exchanges coincide with those of Ji and Lebed, and both natural and “unnatural” parity towers appear, because the tensor operator does not have definite parity [1407.6119]. This crossed-channel structure is not merely formal: it determines which combinations are stable under phenomenological fitting and which resonance exchanges can contribute.

## 5. Representation in dual parametrization, Mellin moments, and resonance exchange

The dual parametrization and Mellin–Barnes approaches provide a theoretically consistent realization of the partial-wave picture. In this framework, conformal moments diagonalize LO evolution, and the further expansion in $t$-channel SO(3) partial waves factorizes the $\xi$ dependence through $P_J(\cos\theta)$ and their derivatives [1407.6119]. The generic form is
\[
F(x, \xi, t) = \sum_J (2J+1)\,F_J(t)\,\mathcal{P}_J(1/\xi),
\]
where $\mathcal{P}_J$ is either $P'_J(1/\xi)$ or an appropriate combination of $P'_J(1/\xi)$ and $P''_J(1/\xi)$, depending on operator and hadron helicities [1407.6119].

Polynomiality is ensured by construction because Mellin moments in $x$ are polynomials in $\xi$ of bounded degree. In the helicity-flip sector this is tied to the generalized form factor decomposition of local operators and the $\xi$-parity constraints on $H_T$, $E_T$, $\tilde H_T$, and $\tilde E_T$ [1407.6119]. This suggests that the partial-wave basis is not only a convenient expansion scheme but an implementation of the fundamental Lorentz-covariant polynomiality requirement.

A complementary covariant representation expresses the matrix element as an infinite sum over spin-$J$ resonance exchanges,
\[
\langle N(p')|\, \hat O \, |N(p)\rangle
\sim \sum_{R_J} \sum_{\rm pol.} \frac{1}{\Delta^2 - M_{R_J}^2}
\big[ \langle N(p') R_J(\Delta) | N(p) \rangle \big] \otimes \big[ \langle 0| \hat O | R_J(\Delta) \rangle \big].
\]
This reproduces the same SO(3)-suitable combinations and supports concrete resonance models, such as $f_2(1270)$ exchange for gluon helicity flip and an analogous $b_1$ exchange for quark helicity flip [1407.6119].

The 2025 string-based Mellin–Barnes study is relevant methodologically because it constructs analytic, Regge-based conformal moments satisfying polynomiality, crossing symmetry, and support by construction, and evolves them to $\mu=2$ GeV at NLO [2508.00817]. However, that work explicitly does **not** include chiral-odd quark transversity $H_T$, $E_T$, $\widetilde H_T$, $\widetilde E_T$ nor the gluon double-helicity-flip transversity GPDs for the spin-$1/2$ nucleon [2508.00817]. It is therefore directly pertinent to GPD model-building technology but not to the double-helicity-flip gluon transversity channel itself.

## 6. Relation to observables in DIS, DVCS, and exclusive reactions

In inclusive DIS on spin-one targets, the double-helicity-flip structure function is the coefficient of the $\Delta=2$ piece of the hadronic tensor and is accessed through polarization configurations that isolate the helicity amplitude $A_{+-,-+}$ [1606.04505]. Because this observable vanishes for spin-averaged or identical-polarization configurations, its extraction requires tensor-polarized targets and a dedicated helicity analysis.

In hard exclusive processes on the nucleon, gluon helicity-flip GPDs contribute at leading twist to the $O(\alpha_s)$ DVCS amplitude and to various hard exclusive reactions [1407.6119]. The Compton form factors are given schematically by convolutions such as
\[
\mathcal{F}_T^{g}(\xi, t, Q^2) \sim \int_{-1}^{1} dx\, C_T^{g}(x, \xi, Q^2)\,\{H_T^g, E_T^g, \tilde H_T^g, \tilde E_T^g\}(x, \xi, t),
\]
and the double-helicity-flip sensitivity is governed by the helicity amplitudes that project operator helicity $J_3=\pm2$ on the gluon side [1407.6119].

The practical exclusive observable is an azimuthal harmonic. The coefficient of the $\cos(3\phi)$ modulation in DVCS interference terms serves as a proxy for gluon double-helicity-flip effects, and the $(3\phi)$ modulation of the interference term in the unpolarized beam–longitudinally polarized target asymmetry measured by HERMES was identified as a possible indication of sizable gluon transversity [1407.6119]. Sensitivity depends on $x$, $\xi$, $t$, and $Q^2$, with larger $|t|$ enhancing orbital angular momentum and therefore the off-forward gluon helicity-flip signal [1407.6119].

For spin-one hadrons, exclusive channels provide a different decomposition of the same angular-momentum content. In the deuteron convolution model, quark helicity-flip amplitudes
\[
A_{\lambda'+;\lambda-}(x,\xi,t)=\frac{1}{2}\left[T^{q\,1}_{\lambda'\lambda}(x,\xi,t)+iT^{q\,2}_{\lambda'\lambda}(x,\xi,t)\right]
\]
show that $\Delta\lambda=\pm2$ transitions are exactly zero in a pure S-wave and grow with $|t|$ when D-wave components are present [1902.03681]. The inclusive forward double-helicity-flip structure function, however, is generated at leading twist by **gluon** transversity rather than by quark chiral-odd transversity [1902.03681]. This is an important point of potential confusion: quark transversity GPDs and the forward double-helicity-flip DIS structure function probe related angular-momentum physics but belong to different operator channels.

A useful comparison is provided by the photon. For a real spin-one photon target, the off-forward double-helicity-flip channel is encoded in a GPD $E_1(x,\Delta^\perp)$ that exhibits the expected quadrupole factor $(\Delta^1)^2-(\Delta^2)^2$ and vanishes in the forward limit [1301.6895]. This reinforces the general statement that $\Delta\lambda=2$ amplitudes require either nonzero momentum transfer or target spin at least one, depending on the channel under consideration.

## 7. Lattice-QCD results, bounds, and phenomenological status

The first lattice-QCD investigation of the leading moment of the twist-two double-helicity-flip gluonic structure function was carried out for the $\phi$ meson, using the $n=2$ local operator
\[
\mathcal{O}_{\mu\nu\mu_1 \mu_{2}} = S\left[G_{\mu\mu_1}G_{\nu \mu_2}\right].
\]
A key theoretical advantage is that this operator “transforms irreducibly as $(2,2)$ under the Lorentz group and does not mix with quark-bilinear operators of the same dimension under renormalization,” although it mixes into higher-twist four-quark operators and lattice hypercubic symmetry induces additional mixing patterns that must be controlled by irrep selection [1606.04505].

The exploratory calculation used a Lüscher–Weisz gauge action; a Clover-improved Wilson fermion action with one level of stout link smearing; an isotropic $24^3 \times 64$ lattice with $a=0.1167(16)$ fm; $N_f=2+1$ sea quarks; $m_\pi=450(5)$ MeV; strange mass tuned to $m_\phi=1040(3)$ MeV; $N_{\textrm{cfg}}=1042$ configurations; 96 source locations per configuration; and 100032 measurements total [1606.04505]. The extracted reduced matrix element for the leading moment was
\[
A_2 \sim 0.23(2)(5),
\]
where the first uncertainty is statistical and the second is an estimated $20\%$ renormalization uncertainty [1606.04505].

The same study examined the gluonic analogue of the Soffer bound. At the distribution level,
\[
|\delta G(x)|\le \frac{1}{2}\left(f_1(x)+\frac{1}{2}f_{1LL}(x) + g_1(x)\right),
\]
and for the first Mellin moments,
\[
|A_2|\le \frac{1}{24} (5 B_{2,1} - 6 B_{2,2}).
\]
Using the extracted $A_2$, $B_{2,1}$, and $B_{2,2}$, the ratio
\[
|A_2| \big/ \left[\frac{1}{24}(5 B_{2,1} - 6 B_{2,2})\right]
\]
was found to be consistent with $0.8$–$1.0$ within uncertainties, meaning the bound is saturated at the level of $80\%$–$100\%$ [1606.04505].

These results are exploratory and the limitations are explicit: a single lattice spacing and volume, heavier-than-physical pions, neglected renormalization factors, neglected operator mixing for some channels, and omitted annihilation diagrams [1606.04505]. Nevertheless, the robust signal demonstrates feasibility for more complex studies in nucleons and light nuclei, where off-forward or spin-one double-helicity-flip observables become accessible.

From the phenomenological side, promising channels include tensor-polarized DIS on spin-one targets, DVCS via azimuthal-harmonic analysis, exclusive vector meson production, and two-meson production [1407.6119; 1606.04505]. For the nucleon, gluon transversity is the leading-twist gateway to double-helicity-flip effects because quark chiral-odd transversity GPDs do not contribute to DVCS or single-meson leptoproduction at leading twist due to chiral selection rules [1407.6119]. For nuclei with $J\ge 1$, the forward structure function $\Delta(x,Q^2)$ remains a particularly clean probe of “exotic glue,” since neither nucleons nor pions can transfer two units of helicity to the nuclear target [1606.04505].

Source: https://www.emergentmind.com/topics/double-helicity-flip-structure-function