⚛️ high-energy theory

Gauge-Invariant Off-Shell Mass

This paper defines a gauge-invariant and renormalized off-shell mass function by extending the pinch technique to arbitrarily long fermion lines, thereby canceling gauge-dependent contributions locally via Ward-Takahashi identities to yield a process-independent self-energy that matches the physical mass on-shell and provides an infrared-finite scalar mass for off-shell comparisons.

Original authors: Kang-Sin Choi, Hyeseon Im

Published 2026-07-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Kang-Sin Choi, Hyeseon Im

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to weigh a ghost.

In the world of quantum physics, particles like electrons aren't solid little billiard balls; they are more like fuzzy clouds of activity. When physicists try to measure the "mass" of one of these clouds, they usually run into a weird problem: the answer changes depending on how you look at it. It's as if the weight of a cloud depended on whether you were looking at it from the left, the right, or through a pair of blue-tinted glasses. In the old way of doing things, if the particle wasn't sitting still (what physicists call "on-shell"), its mass seemed to be a trick of the light, shifting and changing based on the mathematical "gauge" (the set of rules) the scientist chose to use.

This paper by Kang-Sin Choi and Hyeseon Im says: "Stop the trickery. Let's find the real weight, no matter how the particle is moving."

The Problem: The Shifting Shadow

Think of a particle moving through space as a surfer riding a wave. In the old mathematical models, the "mass" of the surfer seemed to change depending on the angle of the sun (the gauge). If you calculated the mass while the surfer was mid-air (off-shell), the number you got was different if you used a different angle of sunlight. This made it impossible to say, "This is the mass of the surfer," because the answer kept sliding around.

The authors argue that this sliding isn't a feature of nature; it's a flaw in the math. They want a definition of mass that is gauge-invariant. That means the mass should be the same number whether you look from the left, the right, or through blue glasses. It should be a solid, unchanging property of the particle, even when it's zooming around at high speeds or sitting in a virtual state inside a larger interaction.

The Solution: The "Pinch" Technique

To fix this, the authors use a clever trick called the Pinch Technique. Imagine you have a long, wiggly rope (the path of the particle) with knots tied in it (the interactions). In the old math, the "gauge-dependent" mess was like a sticky goo smeared all over the rope, making it impossible to tell where one knot ended and another began.

The authors realized that if you look closely at the knots, you can "pinch" the rope. When you pinch a specific section, the sticky goo cancels itself out perfectly, leaving behind a clean, pure section of the rope. They call this segment-locality.

Here is the magic: They showed that you don't need to look at the whole rope or wait for the particle to stop moving to see the clean mass. You can just look at a tiny segment between two knots. By using a mathematical rule called the Ward-Takahashi identity (think of it as a strict accounting rule that says "what goes in must come out"), they proved that the messy, shifting parts cancel out right there, locally, without needing to check the rest of the universe.

The Result: A Mass That Moves

By using this "pinch" method, they defined a new off-shell mass function, $m(q)$ .

What it is: It's a function that tells you the mass of the particle at any speed or energy level ( $q$ ), not just when it's sitting still.
Why it's special: It is gauge-invariant. No matter which mathematical "glasses" you wear, you get the exact same mass function.
The connection to reality: When the particle is finally sitting still (on-shell), this new function gives you the exact physical mass we already know and measure. But when the particle is zooming around (off-shell), it gives you a smooth, consistent curve of mass that doesn't wobble or change based on your math tools.

They also found a way to make this mass function infrared-finite. In the messy world of quantum physics, calculations often blow up with infinite numbers when dealing with very low-energy (soft) photons. The authors showed that by separating the mass into a "scalar" part, they get a version that stays finite and clean, making it directly comparable to other powerful methods like lattice QCD (which uses supercomputers to simulate the universe on a grid).

What They Ruled Out

The authors are very clear about what this is not.

They are not saying the mass is a single, fixed number for the whole universe. The mass does change with momentum, but it changes in a predictable, gauge-invariant way.
They are not relying on the particle being "on-shell" (sitting still) to define the mass. In fact, they explicitly reject the idea that mass only exists as a single number at the pole. They argue that the mass is a function, $m(q)$ , that exists everywhere, not just at one point.
They are not suggesting that the "virtual" particles inside an interaction are ill-defined. On the contrary, they prove that the internal lines of a Feynman diagram (the virtual particles) have a well-defined, gauge-invariant mass, just like real particles do.

How Sure Are They?

The authors are extremely confident, but they are careful to distinguish between what they have proven and what they have demonstrated.

They have proven the cancellation mechanism mathematically. They showed that the gauge-dependent parts cancel out exactly using the Ward-Takahashi identity. This isn't a guess; it's an algebraic certainty.
They have demonstrated this explicitly for the simplest case (Compton scattering with one loop) and argued that it holds for any number of loops and any number of external particles through a logical induction (a step-by-step proof that if it works for one step, it works for the next).
They note that for non-Abelian theories (like the strong force in QCD), the logic holds, but the full proof for any number of particles is still an open structural problem they are leaving for future work. However, for the specific case of the quark self-energy, they show it works perfectly.

The Big Picture

Think of the mass of a particle not as a single weight on a scale, but as a running score in a video game. In the old days, the score would glitch and change numbers depending on which camera angle you picked. This paper fixes the camera. Now, no matter how you zoom in, pan out, or change the angle, the score (the mass) updates smoothly and consistently.

The authors have successfully defined a "running mass" that is real, consistent, and gauge-invariant. They have turned the fuzzy, shifting concept of a virtual particle's mass into a solid, well-defined function that physicists can use to understand how particles behave when they are not sitting still. It's a step toward seeing the "ghosts" of the quantum world as clearly as the real things.

Technical Summary: Gauge-Invariant Off-Shell Mass

Problem Statement
In quantum field theory (QFT), the concept of mass is fundamentally distinct from classical mechanics. While the bare mass $m_B$ is a Lagrangian parameter, the physical mass is determined by the pole of the dressed propagator, $S(q) = i/(\not{q} - m_B - \Sigma(q))$ . The pole mass is gauge-independent, as established by the Nielsen identity. However, defining an off-shell mass function $m(q)$ for arbitrary virtuality $\not{q} \neq m$ has remained problematic. The conventional self-energy $\Sigma(q)$ is gauge-dependent off-shell because the gauge-parameter dependence of the boson propagator (specifically the longitudinal term proportional to $(1-\xi)$ in $R_\xi$ gauges) is distributed across the global topology of Feynman diagrams. Consequently, the off-shell mass function has lacked a definition that is simultaneously gauge-invariant, process-independent, and scheme-independent. Existing definitions, such as the pole mass (a single number) or the $\overline{\text{MS}}$ running mass (scheme-dependent), fail to provide a gauge-invariant function of momentum that describes the fermion propagation at arbitrary virtuality.

Methodology
The authors generalize the Pinch Technique (PT) to define a gauge-invariant, renormalized off-shell mass function. The core methodology relies on a segment-local application of the Ward-Takahashi Identity (WTI) rather than the standard PT approach which often relies on on-shell projections and the Dirac equation for external legs.

Segment-Local WTI: The authors demonstrate that the cancellation of gauge-dependent terms is local to the internal fermion line segment. When the WTI acts at a vertex adjacent to an internal fermion line, it generates a difference of propagators, $S(q) - S(q-\ell)$ . This "pinching" mechanism telescopes within the segment bounded by external photon vertices.
Generalized Self-Energy ( $\hat{\Sigma}$ ): By extending the PT to an arbitrarily long fermion line with $n+1$ external photon vertices, the authors decompose the amplitude into "propagator-like" and "vertex-like" contributions. The gauge-dependent pieces (proportional to $1-\xi$ ) from self-energy, vertex, and box topologies are reorganized. The propagator-like pinch contributions are absorbed into a generalized self-energy $\hat{\Sigma}(q)$ .
Boundary Term Cancellation: The WTI firing at the endpoints of the fermion line generates boundary terms (shifted external leg propagators). These are shown to cancel against external leg self-energies and crossed (u-channel) amplitudes. Crucially, this cancellation does not affect the internal self-energy $\hat{\Sigma}(q)$ , which depends only on the internal momentum $q$ .
Renormalization: The authors perform on-shell renormalization on $\hat{\Sigma}(q)$ to define the mass function $m(q) = m + \hat{\Sigma}_{\text{ren}}(q)$ . This subtraction removes infrared divergences associated with wave-function renormalization and ensures the residue at the pole is unity.
Non-Abelian Extension: The framework is extended to QCD. For the quark line, the result is a color-factor ( $C_F$ ) multiple of the QED result. For gluon segments, the WTI is replaced by the pinch decomposition of the three-gluon vertex and the inclusion of ghost loops (Slavnov-Taylor identities), maintaining the segment-local cancellation structure.
All-Orders Induction: The construction is extended to all orders in perturbation theory via induction, utilizing the exact WTI for dressed propagators and vertices.

Key Contributions and Results

Definition of $\hat{\Sigma}(q)$ : The paper constructs a generalized self-energy $\hat{\Sigma}(q)$ that is strictly gauge-invariant ( $\partial \hat{\Sigma}/\partial \xi = 0$ ) and process-independent. It is shown that $\hat{\Sigma}(q)$ is exactly equal to the conventional self-energy evaluated in the Feynman gauge ( $\xi=1$ ):
$\hat{\Sigma}(q) = \Sigma_\xi(q) \big|_{\xi=1}$
This equality holds for arbitrary virtuality and is derived without invoking the Dirac equation for external states.
The Off-Shell Mass Function: The mass function is defined as $m(q) = m + \hat{\Sigma}_{\text{ren}}(q)$ $m (q) = m + \hat{Σ}_{ren} (q)$ . This function possesses four key properties:
1. Gauge-invariant: Independent of the gauge parameter $\xi$ for all $q$ .
2. Process-independent: Determined solely by the internal segment locality.
3. Scheme-independent: When expressed in terms of on-shell physical parameters (pole mass), it yields a unique result.
4. Physical limit: At the pole ( $\not{q}=m$ ), $m(m) = m$ , recovering the physical pole mass.
Infrared-Finite Scalar Mass: The authors identify an infrared-finite scalar mass function $\hat{M}(q^2)$ derived from the same gauge-invariant self-energy. Unlike the Dirac-valued $m(q)$ , which carries infrared logarithms in its residue, $\hat{M}(q^2)$ is multiplicatively renormalization-group invariant and directly comparable to mass functions computed in Schwinger-Dyson equations and Lattice QCD (typically in Landau gauge).
Explicit Calculation: The mechanism is explicitly demonstrated using the minimal off-shell Compton amplitude (s-channel) with external legs off-shell. The calculation shows that the gauge-dependent contributions from self-energy, vertex, and box diagrams cancel exactly within the internal segment, leaving only the Feynman-gauge result.
Non-Abelian Verification: The paper verifies that for the quark self-energy in QCD at one loop, the gauge-invariant result is simply $C_F$ times the QED result in the Feynman gauge. The non-abelian contributions (three-gluon vertices and ghosts) cancel into other sectors (modified three-gluon vertex and u-channel) and do not alter the segment-local self-energy.

Significance and Claims
The paper claims to resolve the structural difficulty of defining a gauge-invariant off-shell mass. By shifting the focus from global diagram topology to the segment-local action of the WTI, the authors establish that the gauge dependence of the self-energy is an artifact of the gauge-fixing procedure that can be systematically removed without relying on on-shell conditions for external particles.

The significance lies in:

Unification: It unifies the concept of mass from a single on-shell number to a continuous, gauge-invariant function of momentum.
Comparability: It provides a theoretical object ( $\hat{M}(q^2)$ ) that is directly comparable with non-perturbative determinations (Schwinger-Dyson and Lattice) which are currently gauge-dependent.
Conceptual Shift: It asserts that the distinction between "real" (on-shell) and "virtual" (off-shell) fermions is kinematic rather than dynamical. Both are described by the same gauge-invariant dressed propagator $\hat{S}_{\text{ren}}(q) = i/(\not{q} - m(q))$ , differing only in the momentum argument.
Robustness: The construction holds to all orders in perturbation theory and extends to non-Abelian gauge theories, relying on the exact WTI and pinch decomposition rather than specific truncations or on-shell projections.

The authors emphasize that this is a definition of the mass function within the theory, not a proposal for a new experimental observable, though it clarifies the theoretical status of virtual particles and their mass properties.