Structure-dependent radiative corrections to $e^+ e^- \to \pi^+ \pi^- \gamma$ in the GVMD approach

This paper computes next-to-leading order radiative corrections to the $e^+ e^- \to \pi^+ \pi^- \gamma$ process by incorporating the non-perturbative pion structure via the generalized vector meson dominance model, comparing these results with the naive scalar QED approach to quantify model uncertainties for radiative return experiments at flavor factories.

The Gist

Imagine you are trying to measure the size of a very specific, invisible object (a pion) by smashing two high-speed particles together and watching what flies out. This is what physicists do at "flavor factories," which are giant particle accelerators.

To get a precise measurement, they need to understand every tiny detail of how the particles interact. However, there's a problem: the pion isn't a simple, solid marble. It's more like a fuzzy, squishy cloud made of smaller parts.

For a long time, scientists calculated these interactions using a "naive" method. They treated the pion as if it were a perfect, point-like dot (like a billiard ball). This is easy to calculate, but it's not entirely accurate because it ignores the pion's "fuzziness."

The Problem: The "Fuzzy" Cloud

In this paper, the authors (a team of Italian physicists) say, "Hey, we need to stop pretending the pion is a dot. We need to account for its actual, complex structure."

They focus on a specific event: an electron and a positron (matter and antimatter) collide, create two pions, and shoot out a bright flash of light (a photon). This is called $e^+e^- \to \pi^+\pi^-\gamma$.

The "flash of light" is crucial. It acts like a brake, slowing down the collision so the scientists can study the pions at lower energies. But, the light doesn't just come from the collision; it can also be emitted by the pions themselves as they fly apart. This is called Final State Radiation (FSR).

The Old Way vs. The New Way

• The Old Way (Scalar QED): Imagine calculating the path of a car by assuming it's a single point. You ignore the fact that the car has a suspension, a trunk, and a driver. This works okay for a straight road, but if the car hits a bump (a complex interaction), your prediction might be off.
• The New Way (GVMD): The authors use a method called Generalised Vector Meson Dominance (GVMD). Think of this as realizing the car is actually a team of dancers. The "pion" is the whole group, and the "vector mesons" (like the $\rho$ and $\omega$ particles) are the individual dancers holding hands. When the light (photon) hits the pion, it's not hitting a dot; it's interacting with this whole dancing troupe.

What Did They Do?

The team performed a massive, complex calculation (a "one-loop" calculation, which is like checking every possible detour a particle could take) to see how much the "fuzziness" of the pion changes the results.

They compared their new, "fuzzy" model against the old "dot" model for four different experimental setups (like different camera angles and speeds):

1. KLOE: Looking at low-energy collisions.
2. BESIII: Medium energy.
3. B-factories: High energy.

The Results: Why It Matters

Here is the punchline, translated into everyday terms:

• For the "Total Amount" (Invariant Mass): If you just count how many pions are produced, the difference between the "dot" model and the "fuzzy" model is tiny—about 0.1% to 0.3%. It's like the difference between weighing a bag of apples on a kitchen scale versus a high-precision lab scale. It's small, but noticeable.
• For the "Direction" (Angles and Asymmetry): This is where it gets interesting. If you look at where the pions are flying (the angles), the difference jumps to 1% to 2%.
  • Analogy: Imagine throwing a ball. If you treat the ball as a dot, you predict it goes straight. If you realize the ball is a spinning, squishy cloud, you realize the wind (the interaction) pushes it slightly left or right. The "fuzzy" model predicts a different direction than the "dot" model.

Why Should You Care?

This might sound like abstract physics, but it has a real-world impact on one of the biggest mysteries in science today: The Muon $g-2$ Puzzle.

Scientists are trying to measure the magnetic properties of a particle called the muon. Theoretical predictions and experimental measurements don't quite match up. To fix this, they need to calculate the "background noise" (the vacuum polarization) with extreme precision.

The pion is a major part of that background noise. If the models used to calculate the pion's behavior are slightly wrong (because they treat it as a dot instead of a fuzzy cloud), the final answer for the muon mystery will be wrong.

The Bottom Line

This paper is like a software update for the tools physicists use to simulate particle collisions.

• Before: The software assumed particles were simple dots.
• Now: The software knows particles are complex, fuzzy clouds.

By making this update, the authors have reduced the "uncertainty" in their calculations. This helps experimentalists at places like the KLOE and BESIII experiments know exactly how much their "fuzzy" pion models might be skewing their data. It's a small step in the math, but a giant leap toward solving the mystery of why the universe's fundamental particles behave the way they do.

Technical Summary

1. Problem Statement

The paper addresses the need for high-precision theoretical predictions for the process $e^+e^- \to \pi^+\pi^-\gamma$ (radiative return), which is crucial for measuring the electromagnetic pion form factor $F_\pi(q^2)$. This measurement is essential for the data-driven calculation of the leading-order hadronic vacuum polarization (HVP) contribution to the muon anomalous magnetic moment ($a_\mu$), a key component in resolving the "new muon $g-2$ puzzle."

The Core Issue:
Current standard tools (e.g., the event generator Phokhara and the BabaYaga@NLO generator) utilize the Factorised Scalar QED (F$\times$sQED) approach. In this model:

• The pion is treated as a point-like scalar particle in the calculation of Next-to-Leading Order (NLO) virtual corrections.
• The non-perturbative structure of the pion is only introduced via an overall form factor multiplication.
• Limitation: This neglects the internal structure of the pion in the loop integrals for Final-State Radiation (FSR) and Initial-Final State Interference (IFI). Recent studies suggest this approximation introduces systematic uncertainties that are significant for sub-percent precision goals.

2. Methodology

The authors compute the NLO radiative corrections by explicitly incorporating the non-perturbative structure of the pion into the one-loop calculations.

Theoretical Framework: Generalised Vector Meson Dominance (GVMD)

• Instead of the point-like approximation, the authors adopt the GVMD model to describe the pion-photon interaction.
• The pion form factor $F_\pi(q^2)$ is modeled as a sum of Breit-Wigner (BW) functions representing vector meson resonances ($\rho, \omega, \phi, \rho', \rho'', \rho'''$).
• Implementation Rules:
  • The GVMD form factor is inserted directly into the amplitudes by modifying the Scalar QED (sQED) vertices.
  • For a single photon vertex, the vertex is multiplied by $F_\pi(q^2)$.
  • For a two-photon vertex, it is multiplied by $F_\pi(q_1^2)F_\pi(q_2^2)$.
  • This preserves gauge invariance and allows the use of standard loop techniques with effective complex masses for the virtual photons.

Calculation Details:

• Topologies: The calculation covers four main topologies contributing to $e^+e^- \to \pi^+\pi^-\gamma$ at NLO:
  1. $(1\gamma^*)$: Corrections to the pion form factor (ISR/FSR).
  2. $(2\gamma^*, \text{ISR})$: Initial-state radiation with two virtual photons connecting to the pion pair.
  3. $(2\gamma^*, \text{FSR})$: Final-state radiation involving the pion Compton tensor ($\gamma^*\gamma^* \to \pi^+\pi^-\gamma$).
  4. $(2\gamma^*)$: General two-photon exchange.
• Technical Tools:
  • Amplitudes generated using FeynArts and FeynCalc.
  • Tensor integrals reduced to scalar integrals (up to 4-points) using hexagon.m.
  • Numerical evaluation of scalar loops via the Collier library.
  • Numerical derivatives for specific resonance terms computed in quadruple precision using LoopTools.
  • Cross-checked with an independent implementation in Form.
• Renormalization: Dimensional regularization and on-shell renormalization scheme. Infrared (IR) divergences are handled using a fictitious photon mass $\lambda$, ensuring cancellation consistent with F$\times$sQED.

3. Key Contributions

• First NLO GVMD Calculation for Radiative Return: This work extends previous structure-dependent corrections (previously limited to the $e^+e^- \to \pi^+\pi^-$ energy scan process) to the radiative process $e^+e^- \to \pi^+\pi^-\gamma$.
• Quantification of Model Uncertainty: By comparing GVMD results against the standard F$\times$sQED approach, the authors quantify the systematic error introduced by neglecting the pion's internal structure in loop corrections.
• Implementation in Event Generators: The calculations are implemented in the BabaYaga@NLO Monte Carlo event generator, matched with a Parton Shower (PS) to achieve NLOPS accuracy.
• Consistency Checks: The authors validated their implementation by comparing the exact NLO $2 \to 3$ calculation with the soft-photon limit of the $2 \to 2$ process (from a previous study), confirming agreement in the soft-photon regime.

4. Results

The authors analyzed four experimental scenarios: KLOE I (Large Angle), KLOE II (Small Angle), BESIII, and B-factories (Belle/BaBar).

Key Observations:

• Invariant Mass Distribution ($d\sigma/dM_{\pi\pi}$):
  • The GVMD corrections relative to F$\times$sQED are generally small, on the order of a few permille ($\sim 0.1\% - 0.5\%$) around the $\rho$-peak.
  • In KLOE II (SA), corrections are dominated by the $(1\gamma^*)$ term due to collinear logarithms.
  • In BESIII and B-factories, the $(2\gamma^*)$ contributions are suppressed due to the small value of $F_\pi(s)$ at high energies.
• Angular Distribution ($d\sigma/d\theta_+$):
  • Structure-dependent effects are significantly larger here, reaching the percent level ($\sim 1\% - 2\%$).
  • These corrections are dominated by the $(2\gamma^*, \text{ISR})$ Compton tensor topology, which interferes asymmetrically with the Leading Order ISR.
• Forward-Backward Asymmetry ($A_{FB}$):
  • In the KLOE I (LA) setup, the full GVMD correction to $A_{FB}$ is approximately 1% near the $\rho$-peak.
  • This observable is highly sensitive to structure-dependent corrections because the LO asymmetry is non-zero in the radiative process, making the NLO structure-dependent term a critical correction.

5. Significance

• Precision Physics: The study demonstrates that for future measurements of the pion form factor aiming at sub-percent precision, the standard F$\times$sQED approximation is insufficient. Structure-dependent corrections must be included to avoid systematic biases.
• Muon $g-2$ Impact: Since the HVP contribution to $a_\mu$ relies heavily on $e^+e^- \to \pi^+\pi^-$ data extracted via radiative return, reducing the theoretical uncertainty in these cross-sections is vital for resolving the tension between experimental $a_\mu$ measurements and Standard Model predictions.
• Future Tools: The work paves the way for a new version of the BabaYaga@NLO generator that includes NLO structure-dependent corrections in two independent formalisms (GVMD and a future dispersive approach), providing a robust tool for flavor factories (DA$\Phi$NE, BESIII, Belle II).

In conclusion, the paper establishes that while the impact on the total cross-section is small, the impact on angular observables and asymmetries is significant, necessitating the inclusion of non-perturbative pion structure in loop calculations for next-generation precision experiments.