Ghost Modes and Gauge Puzzles: New Calculation Pins Down How Quark Distributions Evolve

Deep inside every proton and nucleus lies a three-dimensional landscape of quarks and gluons, and one of the most powerful tools for mapping it is the transverse momentum dependent distribution, or TMD. Unlike ordinary parton distribution functions, which only track how much longitudinal momentum a quark carries, TMDs also record the quark’s motion sideways, offering a genuinely three-dimensional portrait of the hadron. A new theoretical study published in The European Physical Journal C by Tolga Altinoluk and Guillaume Beuf of the National Centre for Nuclear Research in Warsaw, together with Jamal Jalilian-Marian of Baruch College and the CUNY Graduate Center, has now carried out a meticulous one-loop calculation of the quark TMD in a notoriously tricky gauge, and in doing so has settled a subtle question about where the physics of evolution actually resides.

The stakes are higher than they might sound. TMDs are not static objects: when quantum corrections are included, they acquire a dependence on the energy scale at which the proton is probed, and that dependence is governed by the celebrated Collins-Soper-Sterman, or CSS, evolution equations. These equations resum large logarithms that would otherwise wreck perturbative predictions for processes such as semi-inclusive deep inelastic scattering, where an electron knocks a single quark out of a proton and the small sideways kick of the struck quark carries essential information. Verifying that the CSS equations emerge correctly from a first-principles calculation in an unusual gauge is a stringent consistency test of quantum chromodynamics, the theory of the strong force.

The Warsaw and New York team chose to work in the light-cone gauge, a framework in which the gauge field component along a chosen light-like direction is set to zero. This gauge is the natural habitat of the Color Glass Condensate, the effective theory describing the dense gluon fields that dominate hadrons and nuclei at high scattering energies. Because the same gauge is routinely used in saturation physics, performing TMD calculations in it builds a bridge between two communities that have historically developed parallel toolkits: TMD factorization on one side and gluon saturation on the other. But the light-cone gauge comes with a catch. Even after imposing the gauge condition, a residual gauge freedom remains, and the gluon propagator acquires an extra denominator that is singular and demands a prescription to be handled consistently.

The authors attacked the problem using the background field formalism, in which the non-perturbative content of the target is separated from the perturbative quantum fluctuations. This method, long a workhorse of small-x saturation physics, allowed them to compute the one-loop corrections to the quark TMD while keeping careful track of every divergence. Ultraviolet divergences were regulated with dimensional regularization, while the rapidity divergences, which plague TMDs specifically and cannot be tamed by dimensional regularization, were handled with the so-called pure rapidity regulator, a modern device that isolates rapidity singularities without contaminating the ultraviolet or collinear structure of the calculation.

The central result is reassuring: no matter which prescription the authors used to tame the extra singularity of the light-cone gauge propagator, the one-loop renormalization of the quark TMD reproduced the standard CSS evolution equations. With the Mandelstam-Leibbrandt prescription, widely regarded as the most mathematically consistent choice because it permits Wick rotation and preserves power counting, the calculation went through cleanly, with no unphysical infrared divergences appearing anywhere. The ultraviolet and rapidity renormalization factors combined to yield exactly the expected scale dependence, including the double logarithmic term that drives the resummation of Sudakov-type logarithms.

Yet the journey to that result revealed something genuinely surprising about where the double logarithm comes from. In the Mandelstam-Leibbrandt setup, the crucial double-log contribution did not arise from the familiar ladder diagrams in which a gluon is exchanged between the quark lines. Instead, it emerged from diagrams in which a gluon propagator terminates on the transverse segment of the gauge link located at future light-cone infinity, and specifically from a ghost-like zero-mode contribution. The Mandelstam-Leibbrandt prescription, when written out, effectively introduces a spurious pole at zero longitudinal momentum, a remnant of the residual gauge freedom, and it is this ghost-like object, lurking in the propagator, that carries the double logarithm responsible for CSS resummation in this gauge.

The picture changes dramatically when the authors switched to a different family of prescriptions, all defined in terms of a single light-cone vector. This family includes the Cauchy Principal Value prescription as well as consistent versions of the retarded and advanced prescriptions, and it interpolates smoothly between them through a single parameter. With any member of this family, the diagrams attached to the transverse gauge link at infinity vanish entirely, and the double-logarithmic contribution to the CSS evolution is instead produced by the quark-to-quark ladder diagram. Remarkably, the final answer at the level of each individual diagram turned out to be independent of the free parameter of the family, confirming that the choice of asymptotic boundary condition for the gluon field, a choice experimentalists and theorists often make implicitly, is irrelevant for higher-order corrections to parton distributions.

There was, however, an unwelcome surprise hiding in the single-vector prescriptions. Although the CSS evolution equations emerged intact, the calculation with these prescriptions produced leftover infrared divergences that survived all the way to the final next-to-leading-order correction to the quark TMD, even though the virtuality of the background fields should have acted as a physical infrared regulator. These divergences are not the familiar soft or collinear singularities of QCD but artifacts of the prescriptions themselves. The authors note that because the offending term factorizes from the TMD with a trivial dependence on the transverse separation, it does not spoil the standard matching between TMDs and collinear distributions at this order, but they caution that further study is needed to determine whether these artifacts could cause trouble at higher orders or in matching coefficients.

The calculation also highlights a striking gauge-dependence in the internal anatomy of the result. In the target light-cone gauge used here, the light-like segments of the gauge link in the TMD operator definition collapse to trivial identity matrices, and only the transverse Wilson line at infinity does real work. In the complementary projectile light-cone gauge, the situation is reversed: the transverse link becomes inert and rapidity divergences flow instead through the light-like segments. That the same CSS equations emerge from both arrangements is a powerful demonstration of gauge invariance in action, but it also shows how differently the same physics can be distributed across diagrams depending on the gauge and prescription choices.

The authors see several natural extensions of this work. Applying the same background-field framework to the gluon TMD in the target light-cone gauge is an immediate next step, as is repeating the quark calculation in the projectile gauge, which is heavily used in saturation phenomenology. Such studies should further clarify the relationship between the Color Glass Condensate and TMD factorization, and may point the way toward a unified formalism containing both. For now, the message of this paper is one of robustness: the CSS evolution of the quark TMD is a genuine, prescription-independent property of QCD, even if the ghostly zero-modes of the light-cone gauge turn out to be its unexpected carriers in one corner of the theory’s gauge space.

Subject of Research: One-loop renormalization of the quark transverse momentum dependent distribution in light-cone gauge and the origin of CSS evolution

Article Title: One-loop renormalization of quark TMD in the light-cone gauge: CSS evolution

Article References: Altinoluk, T., Beuf, G., & Jalilian-Marian, J. (2026). One-loop renormalization of quark TMD in the light-cone gauge: CSS evolution. The European Physical Journal C, 86(9), Article 1077. https://doi.org/10.1140/epjc/s10052-026-16305-8

Image Credits: AI Generated

DOI: 10.1140/epjc/s10052-026-16305-8

Keywords: quark TMD, light-cone gauge, CSS evolution, Mandelstam-Leibbrandt prescription, background field method, rapidity divergence, QCD factorization, gluon saturation, Color Glass Condensate, Wilson lines, renormalization, particle physics