Mixed Longitudinal-Transverse Structure Functions

• Mixed longitudinal-transverse structure functions are observables that distinguish coupled longitudinal and transverse fluctuations in systems like deep-inelastic scattering and turbulence.
• They are extracted via Rosenbluth separation in DIS and modeled using multifractal frameworks in turbulence, enabling precise determination of nucleon structure functions and analysis of intermittency.
• Nuclear modifications and small-x analytic techniques further refine our understanding of L–T mixing effects, guiding future experimental probes and theoretical improvements.

Mixed longitudinal-transverse structure functions characterize the statistics and dynamics of coupled longitudinal and transverse fluctuations in systems ranging from deep-inelastic electron–nucleon scattering to fully developed fluid turbulence. These observables arise in the decomposition of cross sections or correlators into longitudinal and transverse components and their mixed moments, allowing precise separation or joint analysis of underlying physical processes. They play a central role in extracting fundamental quantities such as the nucleon structure functions, probing nuclear modifications, and modeling intermittency in turbulence.

1. Formalism of Longitudinal–Transverse Structure Functions

In inclusive electron–nucleon scattering, the differential cross section in the laboratory frame, under the one-photon-exchange approximation, is decomposed in terms of unpolarized structure functions $F_1(x,Q^2)$ (transverse) and $F_2(x,Q^2)$ (combined longitudinal and transverse) (Christy et al., 2011):

$$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$

where $L_{\mu\nu}$ is the leptonic tensor and $W^{\mu\nu}$ the hadronic tensor. Lorentz and gauge invariance dictate the decomposition:

$$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$

The absorption cross sections for transverse ($\sigma_T$) and longitudinal ($\sigma_L$) virtual photons are related to the structure functions via:

$$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$

$$F_2(x,Q^2) = \frac{K\,\nu}{4\pi^2\alpha(1+\nu^2/Q^2)}\,\left[\sigma_T+\sigma_L\right]$$

The longitudinal structure function, $F_2(x,Q^2)$0, isolates the longitudinal response:

$F_2(x,Q^2)$1

where $F_2(x,Q^2)$2.

2. Rosenbluth Separation and Experimental Extraction

Rosenbluth (L–T) separation is the principal method for isolating $F_2(x,Q^2)$3 and $F_2(x,Q^2)$4 experimentally. The reduced cross section at fixed $F_2(x,Q^2)$5 and variable virtual-photon polarization parameter $F_2(x,Q^2)$6 is expressed as:

$F_2(x,Q^2)$7

By measuring $F_2(x,Q^2)$8 at multiple $F_2(x,Q^2)$9 values (e.g., different electron angles), a linear fit allows extraction of $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$0 (intercept at $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$1) and $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$2 (slope). This enables direct determination of $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$3 and $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$4 without reliance on models for $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$5. Rosenbluth separations have yielded high-precision longitudinal and transverse structure functions in previously unexplored $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$6 and $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$7 regimes (Christy et al., 2011).

Beam energies deployed at Jefferson Lab experiments span 1–6 GeV (Hall C) and up to 11 GeV post-12 GeV upgrade, with data taken at multiple angles to cover $$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$8–$$\frac{d^2\sigma}{d\Omega\,dE'} = \frac{\alpha^2}{Q^4}\,\frac{E'}{E}\,L_{\mu\nu}W^{\mu\nu}$$9. Point-to-point systematic uncertainties on cross sections are typically $L_{\mu\nu}$0, with normalization uncertainties $L_{\mu\nu}$1, and statistical errors per bin $L_{\mu\nu}$2–$L_{\mu\nu}$3. Radiative corrections (Mo–Tsai-type) are applied, with uncertainties $L_{\mu\nu}$4 (Christy et al., 2011).

3. Mixed Structure Functions in Turbulence: Multifractional Framework

In fully developed turbulence, mixed structure functions parameterize statistical correlations between longitudinal and transverse velocity increments at scale $L_{\mu\nu}$5:

$L_{\mu\nu}$6

The scaling behavior in the inertial range ($L_{\mu\nu}$7) is governed by exponents $L_{\mu\nu}$8:

$L_{\mu\nu}$9

A unified multifractal formalism introduces a joint Hölder-singularity spectrum $W^{\mu\nu}$0 for the local scaling exponents $W^{\mu\nu}$1 of longitudinal and transverse increments, yielding via Laplace’s method:

$W^{\mu\nu}$2

(Buaria, 18 Jan 2026)

Explicit models for $W^{\mu\nu}$3 (e.g., bivariate log-normal or log-Poisson forms) enable closed-form expressions for the exponent surfaces $W^{\mu\nu}$4, which generalize classical single-variable intermittency laws.

4. Nuclear Modifications and L–T Mixing

In nuclear deep-inelastic scattering, the assumption that the longitudinal-transverse structure function ratio $W^{\mu\nu}$5 is unchanged in nuclei is theoretically violated. Nucleons possess finite transverse momentum $W^{\mu\nu}$6 relative to the photon direction, giving rise to mixing probability proportional to $W^{\mu\nu}$7. The impulse-approximation convolution formalism leads to matrix convolution of nuclear and nucleon structure functions—off-diagonal light-cone distributions $W^{\mu\nu}$8 explicitly couple $W^{\mu\nu}$9 and $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$0 in the nuclear case (Kumano, 23 Jun 2025):

$$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$1

Numerical estimates for the deuteron indicate $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$2–$$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$3 relative nuclear modification in $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$4 at medium and large $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$5 and low to moderate $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$6. The modification is suppressed but not vanishing at high $$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$7.

Nuclear Modification Table

$$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$8	$$M\,W^{\mu\nu} = \left(-g^{\mu\nu}+\frac{q^\mu q^\nu}{q^2}\right)\,F_1 + \text{[longitudinal term]} \frac{F_2}{p\cdot q}$$9 at $\sigma_T$0 GeV$\sigma_T$1	$\sigma_T$2 GeV$\sigma_T$3	$\sigma_T$4 GeV$\sigma_T$5
0.2	1.01	1.005	1.002
0.5	1.04	1.02	1.01
0.8	1.05	1.025	1.015

Mixing corrections vanish in the Bjorken limit ($\sigma_T$6), but persist at finite $\sigma_T$7—especially relevant for contemporary and planned DIS facilities. These effects are critical for precise extraction of neutron structure functions from nuclear target data and for flavor decompositions (Kumano, 23 Jun 2025).

5. Analytical Methods and Small-$\sigma_T$8 Regime

The relation between $\sigma_T$9 and $\sigma_L$0 is central to the extraction of mixed longitudinal–transverse structure functions, particularly at small $\sigma_L$1. Froissart-bounded parametrizations (e.g., the BDH form) for $\sigma_L$2 and QCD evolution via DGLAP equations enable analytic solution for $\sigma_L$3, both in leading and next-to-leading order (Kaptari et al., 2019):

$\sigma_L$4

with $\sigma_L$5, $\sigma_L$6 and $\sigma_L$7 analytic in fit parameters.

At $\sigma_L$8, $\sigma_L$9, satisfying the Froissart bound $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$0 as $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$1. NLO corrections further improve agreement with HERA measurements.

At ultra-high energies (e.g., cosmic neutrino scattering), these analytic structure functions can be directly inserted into cross-section integrals, with additional TMCs, heavy-quark threshold matching, and saturation corrections as required.

6. Physical Interpretation and Applications

Mixed longitudinal-transverse structure functions embody the coupled dynamics of stretching and rotational motions (turbulence) or photon absorption modes (DIS). In turbulence, transverse intermittency is governed by mixed exponents $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$2 rather than by purely transverse increments; the local viscous cutoff is dictated by the longitudinal Reynolds number, leading to inheritances of intermittency properties across gradient directions (Buaria, 18 Jan 2026).

In nuclear DIS, physical origin of L–T mixing is the finite transverse momentum of nucleons inside the nucleus, which affects both precision flavor decompositions and interpretations of gluon dynamics especially at small $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$3. Future high-precision experimental programs at Jefferson Lab (large $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$4) and Electron-Ion Colliders (small $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$5) are crucial for validating and quantifying these effects (Kumano, 23 Jun 2025, Christy et al., 2011).

7. Future Directions and Open Problems

• Jefferson Lab’s upgraded capabilities will extend Rosenbluth separation to $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$6 GeV$$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$7 and $$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$8, refining constraints on large-$$F_1(x,Q^2) = \frac{K\,M}{4\pi^2\alpha}\,\sigma_T(x,Q^2)$$9 PDFs and higher-twist corrections (Christy et al., 2011).
• New techniques such as spectator-tagging and mirror-nuclei experiments will minimize nuclear-model ambiguities in neutron structure function measurements (Christy et al., 2011).
• Quantitative study of L–T mixing in nuclei at small $$F_2(x,Q^2) = \frac{K\,\nu}{4\pi^2\alpha(1+\nu^2/Q^2)}\,\left[\sigma_T+\sigma_L\right]$$0 awaits new high-luminosity electron-ion colliders where gluon dynamics and nuclear shadowing can be systematically probed (Kumano, 23 Jun 2025).
• In turbulence, the joint multifractal model provides a predictive framework for small-scale intermittency, validated up to $$F_2(x,Q^2) = \frac{K\,\nu}{4\pi^2\alpha(1+\nu^2/Q^2)}\,\left[\sigma_T+\sigma_L\right]$$1 in DNS, but exploration of more complex multifractal spectra remains an evolving frontier (Buaria, 18 Jan 2026).
• On the theoretical side, improved TMC implementations in collinear factorization and higher-twist modeling are required to match the unprecedented experimental precision.

A plausible implication is that a unified treatment of mixed longitudinal-transverse structure functions, integrating QCD, nuclear, and turbulence dynamics, is necessary for a fully accurate description of multi-modal fluctuations and their interplay at all scales.