$J/\psi$ Photoproduction from Threshold to HERA: Leading-Twist Convolution, Small-$x$ Pathology, and Eikonal Unitarization

This paper resolves the discrepancy between leading-twist predictions and experimental data for $J/\psi$ photoproduction by demonstrating that modern small-$x$-singular PDFs distort threshold reconstructions but can be successfully reconciled with both near-threshold and HERA measurements through a minimal eikonal unitarization framework anchored by OPE sum rules.

The Gist

The Big Picture: A Bridge Between Two Worlds

Imagine the universe of particles is divided into two neighborhoods:

1. The "Hard" Neighborhood: Where particles are tiny, fast, and follow strict, predictable rules (like billiard balls).
2. The "Soft" Neighborhood: Where particles are fuzzy, messy, and follow complex, chaotic rules (like a crowded dance floor).

The J/ψ particle (a heavy "charmonium" particle) is like a tiny, super-dense marble. Because it is so small and heavy, it can peek into the "Soft" neighborhood (the inside of a proton) without getting lost in the crowd. Scientists want to use this marble to take a "X-ray" of the proton to see how the gluons (the glue holding the proton together) are arranged.

The paper investigates how well our current maps (mathematical models) predict what happens when we shoot this marble at a proton, from very slow speeds (near the threshold) to incredibly fast speeds (like at the HERA accelerator).

---

The Problem: Two Broken Maps

The author, Arkadiy Syamtomov, tried to use the best modern maps of the proton's internal glue (called PDFs or Parton Distribution Functions) to predict the results. He found that the maps had two major "glitches" depending on how you used them.

Glitch #1: The "Zoom-In" Distortion (The Moment-Based Approach)

Imagine you are trying to guess the shape of a mountain range by looking at a few specific points (moments) on a map.

• The Old Map (1999): The mountains were gentle hills. The points lined up perfectly, and the prediction was smooth.
• The New Map (Modern Data): The new maps show that near the edge of the world (very low energy, or "small-x"), the mountains are actually jagged, vertical cliffs.
• The Result: When the scientists tried to use the "points" method with these new jagged cliffs, the math got confused. It tried to fit a smooth curve through the jagged cliffs and ended up predicting a mountain that was 17 to 20 times steeper than reality.
• The Analogy: It's like trying to draw a smooth road through a canyon using only a few GPS dots. If the dots are on the edge of a cliff, your computer might draw a road that goes straight up into the sky instead of following the valley. This method failed near the starting line (threshold).

Glitch #2: The "Speeding Bullet" (The Direct Convolution Approach)

So, the scientists tried a different method: Direct Convolution. Instead of guessing from points, they just calculated the path directly, like simulating a car driving down the road.

• The Good News: This method worked perfectly near the starting line. It matched the new, slow-speed data from the GlueX experiment beautifully. It didn't get confused by the jagged cliffs.
• The Bad News: When they sped the car up to highway speeds (HERA energies), the simulation went crazy. It predicted the car would be 7 to 12 times faster than what the actual speedometers (experiments) showed.
• The Analogy: Imagine a car that accelerates perfectly from a stop sign but then ignores the speed limit and the friction of the road, zooming off into space. The math said, "More glue means more speed!" but reality said, "No, there's a limit."

---

The Solution: The "Traffic Cop" (Eikonal Unitarization)

Why did the fast car zoom too fast? Because the math assumed the proton was an empty highway where the car could just keep accelerating. But in reality, as the car gets faster, the "glue" (gluons) in the proton gets so dense that it acts like a traffic jam.

The author introduced a concept called Eikonal Unitarization.

• The Metaphor: Imagine the proton is a crowded dance floor. At slow speeds, the J/ψ marble can slip through easily. But at high speeds, the marble is so energetic that it starts bumping into the dancers (gluons) so hard that the dancers push back.
• The Fix: The author added a "Traffic Cop" to the math. This cop says, "Okay, you can go fast, but you can't go faster than the density of the crowd allows."
• The Result: When this "Traffic Cop" (saturation effect) was applied, the prediction for the high-speed run dropped down and matched the real-world data perfectly. The "overshoot" was fixed.

---

The Hidden Hero: The "Real Part"

One of the most interesting findings is about the Real Part of the interaction.

• In physics, interactions have a "Real" part and an "Imaginary" part (don't worry, "Imaginary" just means a specific mathematical component, not "fake").
• Near the starting line (Threshold): The "Imaginary" part (the actual collision) is almost zero because the car is barely moving. However, the "Real" part is huge.
• The Analogy: It's like pushing a heavy boulder. Before the boulder actually rolls (collision), you are already exerting a massive amount of force just trying to get it to budge. The paper found that near the start, the "force" (Real part) is doing almost all the work, anchored by a specific constant value derived from the math.

---

Summary for the General Audience

1. The Goal: Scientists want to understand how heavy particles interact with protons to learn about the "glue" holding matter together.
2. The Conflict: Modern maps of the proton's glue are very accurate but have a weird "spike" at low energies.
3. The Failure:
   • Method A (using points) got confused by the spike and predicted a result that was way too steep.
   • Method B (direct calculation) worked at low speeds but predicted results that were way too high at high speeds.
4. The Fix: The high-speed failure happened because the math ignored the "crowd control" of the proton. When the scientists added a "traffic jam" rule (unitarization) to the math, the predictions finally matched reality at all speeds.
5. The Takeaway: The universe has a speed limit for these interactions. Even if the "glue" gets denser, the interaction can't grow infinitely; it eventually saturates, like a sponge that can't hold any more water.

In short: The paper fixed a broken prediction by realizing that at high speeds, the proton isn't an empty highway, but a crowded room that pushes back.

Technical Summary

1. Problem Statement

The paper addresses the theoretical description of near-threshold $J/\psi$ photoproduction on nucleons ($\gamma N \to J/\psi N$) across a wide energy range, from the production threshold ($W \approx 4.0$ GeV) to high energies at HERA ($W \approx 300$ GeV).

The core problem is a qualitative tension arising when modern Next-to-Next-to-Leading Order (NNLO) gluon Parton Distribution Functions (PDFs) are applied to the Operator Product Expansion (OPE) framework:

1. Threshold Pathology: Using modern PDFs in a moment-based reconstruction yields unphysical threshold exponents ($a \approx 17\text{--}20$), whereas experimental data and previous analyses suggest $a \approx 1\text{--}2$.
2. High-Energy Overshoot: While a direct convolution approach fixes the threshold behavior, it significantly overshoots HERA experimental data by a factor of 7–12 at high energies ($W \gtrsim 90$ GeV).

The study aims to resolve these discrepancies by analyzing the interplay between modern gluon distributions, target-mass corrections (TMC), dispersion relations, and unitarization effects.

2. Methodology

The author employs a multi-faceted theoretical framework combining perturbative QCD, dispersion theory, and phenomenological unitarization:

• Framework: The analysis is based on the Vector Meson Dominance (VMD) model, relating the photoproduction amplitude to the forward $J/\psi$-nucleon scattering amplitude.
• OPE and Sum Rules: The forward amplitude is analyzed using the Operator Product Expansion (OPE) for short-distance interactions. The imaginary part of the amplitude is linked to the gluon distribution moments.
  • Target-Mass Corrections (TMC): Crucially, the study includes TMCs, which modify the Mellin moments via an $x$-dependent weight function $T_n(x)$ without introducing new higher-twist operators.
  • PDFs: Four modern NNLO gluon PDF sets are used: ABMP16, MSHT20, CT18, and NNPDF4.0.
• Two Reconstruction Approaches:
  1. Moment-Based Reconstruction: Fits the cross-section ansatz $\sigma \propto (s-s_{th})^a / s^{a+1}$ to the $n=4$ and $n=6$ Mellin moments.
  2. Direct Convolution: Computes the cross-section directly via a partonic convolution integral (Eq. 17), which naturally restricts the integration to the kinematically relevant $x$-range near threshold.
• Dispersion Relations: The real part of the amplitude is reconstructed using a once-subtracted dispersion relation. The subtraction constant $M_{\psi N}(0)$ is fixed by the OPE sum rules.
• Unitarization: To address the high-energy overshoot, a minimal eikonal unitarization is applied to the imaginary part of the amplitude, introducing an energy-dependent saturation scale.

3. Key Contributions and Findings

A. The "Small-x Pathology" in Moment-Based Reconstruction

The paper identifies a specific mechanism causing the failure of the moment-based approach with modern PDFs:

• Mechanism: Modern PDFs exhibit a small-$x$ singularity ($g(x) \sim x^{\alpha_g-1}$ with $\alpha_g < 0$). This singularity disproportionately enhances lower-order moments ($A_4$) compared to higher-order moments ($A_6$) because the weight $x^{n-2}$ suppresses the small-$x$ region less for lower $n$.
• Result: The ratio $\tilde{A}_6 / \tilde{A}_4$ (TMC-corrected) is reduced by a factor of $\approx 1.7$ compared to older scaling parametrizations.
• Consequence: To satisfy the sum-rule constraints with this compressed ratio, the fitted threshold exponent $a$ is driven to unphysical values of $a \approx 17\text{--}20$. This results in a cross-section that rises too steeply near threshold and falls off too rapidly ($\sim W^{-4}$) at high energy, failing to match HERA data.

B. Success of the Direct Convolution at Threshold

• The direct convolution approach avoids the moment-based reconstruction artifact.
• By restricting the integration domain to $x > \epsilon_0/\lambda$, the method naturally excludes the problematic small-$x$ region near threshold (where $x \to 1$).
• Result: This approach yields smooth, monotonically rising cross-sections that successfully describe near-threshold data (GlueX, Cornell, Gittelman) for all four modern PDF sets without requiring unphysical exponents.

C. The High-Energy Overshoot and Unitarization

• The Tension: While the direct convolution fixes the threshold, it overshoots HERA data by a factor of 7–12 at $W \gtrsim 90$ GeV.
• Diagnosis: This is an intrinsic feature of leading-twist convolution with any PDF possessing a small-$x$ singularity. The growth of the gluon density feeds directly into the cross-section, violating unitarity bounds at high energies.
• Resolution (Eikonal Unitarization): The author proposes a minimal eikonal correction:
  $$\text{Im } M_{\text{unit}} = M_{\text{sat}}(W) \left[ 1 - \exp\left( -\frac{\text{Im } M_{\text{conv}}}{M_{\text{sat}}(W)} \right) \right]$$
  where $M_{\text{sat}}(W)$ is an energy-dependent saturation amplitude fitted to HERA data.
• Outcome: This screening factor suppresses the leading-twist result at high $W$ (reducing the cross-section by $\sim 7$ at 100 GeV) while leaving the threshold region unchanged (since $\text{Im } M \to 0$ there). This reconciles the theory with the full $W$-range data ($\chi^2/N_{\text{data}} \approx 3/21$).

D. Role of the Real Part

• Near threshold, the real part of the amplitude dominates the cross-section.
• The real part is anchored by the OPE subtraction constant $M_{\psi N}(0) \approx 36\text{--}39 \text{ GeV}^{-2}$ and amplified by the dispersive integral.
• At threshold ($W \approx 4.1$ GeV), $\text{Re } M \approx 150$ while $\text{Im } M \approx 0$, confirming that the observable is driven almost entirely by the dispersive (real) component.

4. Significance

• Theoretical Consistency: The paper demonstrates that the "small-x pathology" is an artifact of the reconstruction ansatz (moment-based fitting) rather than a failure of the underlying OPE or the PDFs themselves.
• Phenomenological Validation: It validates the direct convolution method as the robust tool for describing threshold $J/\psi$ photoproduction with modern NNLO PDFs.
• Unitarity Constraints: It provides strong evidence that leading-twist QCD alone is insufficient to describe $J/\psi$ photoproduction at HERA energies. The data necessitates unitarization or gluon saturation effects, which suppress the cross-section growth predicted by linear DGLAP evolution.
• Proton Structure: The work reinforces the connection between threshold $J/\psi$ photoproduction and the gravitational form factors of the nucleon (via the subtraction constant), offering a probe for the gluonic contribution to the nucleon mass.

5. Conclusion

The study concludes that while modern NNLO PDFs successfully describe the near-threshold behavior of $J/\psi$ photoproduction via direct convolution, they inherently violate unitarity at high energies due to small-$x$ singularities. A minimal eikonal unitarization successfully reconciles the theory with experimental data across the entire energy spectrum, confirming that saturation effects are essential for high-energy heavy quarkonium production.