Accessing the Gluon Momentum Fraction of Nucleons through the Gradient Flow

This paper presents a lattice QCD calculation of the nucleon's gluon momentum fraction using the gradient flow method with nonperturbative renormalization, yielding a final result of 0.482(35) at 2 GeV in the MS-scheme.

The Gist

The Big Picture: Who is Driving the Bus?

Imagine a proton (the building block of atomic nuclei) as a busy, high-speed bus zooming down a highway. Inside this bus, there are two types of passengers: quarks (the "stars" of the show) and gluons (the "glue" holding everything together).

For decades, physicists have known how much of the bus's total speed (momentum) is carried by the quarks. But the gluons have been a mystery. We know they are there, and we know they are heavy with energy, but we haven't been able to get a precise count of how much of the bus's speed they are actually carrying.

This paper is like a team of detectives using a super-powerful microscope (called Lattice QCD) to finally take a census of the gluons and answer the question: "How much of the proton's momentum belongs to the glue?"

The Challenge: The Gluon is a Ghost

Why is this so hard?

1. The Signal is Faint: In the quantum world, quarks are loud and clear, but gluons are like ghosts. They are hard to see because the "noise" in the calculation is often louder than the "signal." It's like trying to hear a whisper in a rock concert.
2. The Math is Messy: When you try to calculate the momentum of these gluons on a computer, the numbers tend to blow up to infinity unless you use very specific, tricky math tricks to clean them up.

The Detective's Toolkit: Three New Tricks

To solve this mystery, the authors (the HadStruc Collaboration) used three clever techniques to improve their "microscope":

1. Distillation: The "High-Definition Filter"

Imagine trying to take a photo of a fast-moving car in the rain. The picture is blurry. Distillation is like putting a high-tech filter on your camera lens. It doesn't just take a picture of every single raindrop (which creates noise); instead, it focuses only on the essential shape of the car.

• In the paper: They use a mathematical filter to smooth out the "jagged" edges of the quantum data, making the gluon signal much clearer and easier to measure.

2. The Variational Method: Tuning the Radio

Imagine you are trying to listen to a specific radio station, but there are many other stations playing at the same time, creating static.

• In the paper: The proton isn't just one state; it's a mix of the "ground state" (the calm, normal proton) and "excited states" (protons vibrating with extra energy). The Variational Method is like a sophisticated radio tuner. By using a large "basis" of different mathematical antennas, they can tune out the static (the excited states) and isolate the pure signal of the ground-state proton.

3. Gradient Flow: The "Heat Press"

This is the star of the show. Imagine you have a crumpled piece of paper with a drawing on it. The drawing is hard to read because of the wrinkles. If you put it under a heat press, the wrinkles smooth out, and the drawing becomes clear.

• In the paper: The "Gradient Flow" is a mathematical process that "smooths" the quantum fields over time. As they apply this "heat," the messy, infinite quantum fluctuations (the wrinkles) disappear, leaving behind a clean, finite value that represents the gluon's momentum. It's a way to clean the data without losing the physics.

The Result: The Final Count

After running millions of simulations on supercomputers (using data from the Jefferson Lab and other facilities), they applied their "heat press" and "filters."

They found that at a specific energy scale (2 GeV), the gluons carry 48.2% of the proton's momentum.

• The Number: $0.482 \pm 0.035$.
• What it means: Almost half of the proton's "oomph" comes from the glue holding it together, not just the quarks. This result matches well with what we see in real-world experiments (like smashing particles together at the Large Hadron Collider), which gives us confidence that our theoretical understanding is correct.

Why Does This Matter?

Think of the proton as a car. We used to think the engine (quarks) did all the work. Now we know the transmission and the chassis (gluons) are doing nearly half the work too.

Understanding exactly how the momentum is shared helps us:

1. Decode the Universe: It tells us how the strong nuclear force, which holds the universe together, actually works.
2. Predict Future Experiments: As we build bigger particle colliders, we need precise maps of the proton's interior to predict what will happen when we smash them together.

Summary

This paper is a triumph of "cleaning up the data." By using a new method called Gradient Flow (smoothing the quantum fields) combined with advanced statistical tricks, the team successfully measured the "ghostly" momentum of gluons. They proved that the glue inside the proton is responsible for nearly half of its total speed, bringing our theoretical calculations in line with reality.

Technical Summary

1. Problem Statement

The distribution of momentum among quarks and gluons inside a nucleon is a fundamental open problem in nuclear physics. While Parton Distribution Functions (PDFs) describe this experimentally, determining the gluon momentum fraction ($\langle x \rangle_g$) from first principles using Lattice Quantum Chromodynamics (QCD) has historically been challenging.

• Signal-to-Noise Ratio: Gluon operators suffer from poor signal-to-noise ratios compared to quark operators, making statistical extraction difficult.
• Renormalization: Previous lattice calculations relied on either perturbative one-loop renormalization or non-perturbative RI/MOM schemes, which introduce specific systematic uncertainties.
• Discrepancy: There is a notable spread in central values and significant uncertainties in existing lattice determinations of $\langle x \rangle_g$ compared to global phenomenological fits (e.g., from DIS and Drell-Yan data).

2. Methodology

The authors employed a novel combination of advanced lattice techniques to address the signal-to-noise problem and improve renormalization.

A. Lattice Setup

• Ensemble: Calculations were performed on a single Wilson-clover ensemble with $N_f = 2+1$ dynamical flavors.
• Parameters: Pion mass $M_\pi = 358$ MeV (unphysical, heavier than physical), lattice spacing $a \approx 0.094$ fm, and volume $32^3 \times 64$.
• Statistics: 1,121 gauge configurations generated using a rational hybrid Monte Carlo (rHMC) algorithm.

B. Signal Improvement Techniques

To mitigate the poor signal-to-noise ratio of gluon operators, the authors integrated three key methods:

1. Distillation: A gauge-covariant smearing technique that factorizes correlation functions into "elementals" (interpolators) and "perambulators" (propagators). This allows for the reuse of expensive Dirac inversions across different operators, significantly reducing computational cost and storage requirements.
2. Variational Method (GEVP): A basis of six interpolating operators (including local nucleon operators and hybrid operators with covariant derivatives) was used to construct a correlation matrix. Solving the Generalized Eigenvalue Problem (GEVP) isolates the ground state and suppresses excited-state contamination.
3. Summed GEVP (sGEVP): Applied to three-point correlators to further reduce systematic uncertainties arising from excited states by summing over insertion times.

C. Renormalization via Gradient Flow

Instead of traditional RI/MOM schemes, the authors utilized the Gradient Flow (Wilson flow) for renormalization:

• Mechanism: Gauge fields are evolved in a fictitious flow time $\tau$ via a diffusion equation. This acts as a non-perturbative, gauge-invariant ultraviolet regulator.
• Matching: The flowed lattice operators are matched to the $\overline{\text{MS}}$ scheme at a scale $\mu = 2$ GeV using perturbative matching coefficients calculated to Next-to-Leading Order (NLO) and Next-to-Next-to-Leading Order (NNLO).
• Operator Choice: The calculation focuses on the $O_B$ operator (a specific combination of the energy-momentum tensor components) which offers a better signal-to-noise ratio in the nucleon rest frame compared to the $O_A$ operator.

D. Statistical Analysis

• Bayesian Model Averaging (BMA): Used extensively in all fit procedures (two-point and three-point correlators, and flow-time extrapolation). This technique averages results over different fit models (varying data cuts, functional forms like linear/quadratic/cubic) to reduce model dependence and provide robust error estimates.

3. Key Contributions

• Novel Renormalization Scheme: This work demonstrates the first application of gradient flow renormalization specifically for the nucleon gluon momentum fraction, offering a gauge-invariant alternative to RI/MOM.
• High-Precision Signal Extraction: By combining distillation with the variational method, the authors successfully extracted the gluon signal from a large ensemble (1,121 configs) despite the inherent noise of disconnected gluon loops.
• Systematic Control: The use of sGEVP and BMA provides a rigorous framework for controlling excited-state contamination and fit-model dependence, which are major sources of systematic error in this field.

4. Results

The authors report the gluon momentum fraction at a renormalization scale of $\mu = 2$ GeV:
$$\langle x \rangle_g(\mu = 2 \text{ GeV}) = 0.482(35)$$

• Uncertainty: The quoted error (0.035) is purely statistical. Systematic uncertainties (from excited states, perturbative truncation, and flow-time dependence) were analyzed but not fully included in the final error budget due to the single ensemble used.
• Consistency: The result is consistent with recent phenomenological global fits (which typically cluster around 0.40–0.50) and aligns well with other recent lattice determinations (e.g., MSU 23, MIT 24, $\chi$QCD 21).
• Perturbative Check: Comparing NLO and NNLO matching coefficients showed a relative change of ~6.6%, suggesting that missing higher-order perturbative terms are subdominant to current statistical uncertainties.

5. Significance and Limitations

• Significance: This result validates the gradient flow as an effective and robust tool for renormalizing gluon operators in lattice QCD. It provides a competitive determination of $\langle x \rangle_g$ that helps bridge the gap between lattice calculations and experimental phenomenology.
• Limitations:
  • Single Ensemble: The calculation was performed on a single lattice spacing, volume, and pion mass. Consequently, the result contains unquantified discretization, finite-volume, and unphysical quark-mass artifacts.
  • Quark Mixing: Mixing between gluon and quark operators was neglected in the matching coefficients. While numerical evidence suggests this effect is small, a full analysis is deferred to future work.
  • Systematic Budget: A complete error budget including continuum and chiral extrapolations is not yet available.

Conclusion:
The paper represents a significant step forward in lattice QCD calculations of nucleon structure. By successfully applying gradient flow renormalization and advanced variance reduction techniques, the HadStruc Collaboration has produced a precise determination of the gluon momentum fraction, demonstrating that gradient flow is a viable and powerful method for tackling the challenging problem of gluon structure in hadrons. Future work will focus on performing the calculation on multiple ensembles to control systematic uncertainties and include quark-gluon mixing effects.