Excluding MeV-scale QCD axions by $K_L \to π^0π^0 a$ at KTeV

This paper reexamines the viability of MeV-scale QCD axions by deriving new constraints from kaon decay measurements, particularly the KTeV $K_L \to \pi^0 \pi^0 e^+ e^-$ data, and concludes that the previously suggested parameter window is effectively excluded even after accounting for theoretical uncertainties.

The Gist

Imagine the universe is a giant, complex machine, and for decades, physicists have been trying to fix a specific glitch in its engine called the "Strong CP Problem." To fix it, they proposed a tiny, invisible part called the QCD Axion. Think of this axion as a "ghost particle" that is so light and elusive that it has been hiding in plain sight, evading detection.

For a long time, scientists thought these ghosts had to be incredibly heavy (in particle physics terms) or incredibly light. But a few years ago, a clever group of theorists suggested a new hiding spot: a "MeV-scale" axion. This is a ghost with a mass of about 10 million electron-volts (MeV)—heavy enough to be interesting, but light enough to slip through the cracks of previous experiments. They proposed a specific set of rules (like a secret handshake) that would allow this axion to exist without being caught by our current detectors.

The Detective Work: The KTeV Experiment
In this paper, the authors act as detectives revisiting an old crime scene. They looked at data from an experiment called KTeV, which took place at Fermilab. The KTeV experiment was designed to watch a specific type of particle, the Kaon (specifically the $K_L$), decay into other particles.

Think of the Kaon as a unstable soap bubble. Usually, it pops into two smaller bubbles (pions) and sometimes a pair of electrons and positrons ($e^+e^-$). The KTeV team had previously said, "We only saw this happen 6.6 times out of a billion." They set a strict limit: if it happened more often, they would have seen it.

The New Theory: The "Ghost" in the Machine
The authors asked: "What if the Kaon didn't just pop into electrons directly? What if it first created our 'ghost' axion, which then immediately turned into an electron-positron pair?"

If this happened, the final result would look exactly the same to the detector: two pions and an electron pair. However, the path the particles took would be slightly different, like a car taking a scenic detour instead of a highway. The KTeV team had assumed the car took the highway (Standard Model physics). The authors realized that if the car took the scenic detour (the axion), the KTeV detector might have missed the subtle differences in how the data was analyzed.

The Simulation: A Digital Twin
To test this, the authors built a digital twin of the KTeV experiment. They used powerful computers to simulate millions of Kaon decays, but this time, they included the "scenic detour" (the axion). They programmed the simulation to mimic the real detector's eyes, ears, and blind spots, including how it measures energy and position.

They found that the KTeV detector is actually very good at spotting this specific type of "scenic detour." Even if the axion is there, the detector would have seen it.

The Verdict: The Ghost is Gone
The results were devastating for the "MeV-scale axion" theory.

1. The KTeV Limit: When they applied the KTeV data to their new simulation, they found that the "MeV-scale axion" would have produced way more events than KTeV allowed. It's like saying, "If this ghost existed, we would have seen 1,000 of them, but we only saw 6."
2. The Safety Net: The authors also checked other experiments (like NA62 and E949) and the magnetic properties of electrons. While some of these left tiny loopholes open, the KTeV result slammed the door shut on almost all of them.
3. The "Fine-Tuning" Loophole: The only way the axion could still hide is if the universe is incredibly "fine-tuned." Imagine a scale where two heavy weights (mathematical corrections) are placed on opposite sides. For the axion to survive, these weights must cancel each other out perfectly to three decimal places, reducing the axion's production by a factor of 1,000. The authors argue this is highly unlikely and unnatural.

The Conclusion
In simple terms, this paper says: "The MeV-scale QCD axion is likely dead."

The specific version of the axion that was supposed to be light and sneaky has been exposed by the KTeV experiment. Unless the laws of physics are rigged with an incredibly precise, unnatural cancellation (a "fine-tuning" miracle), this particle does not exist in the mass range scientists were hoping for.

What's Next?
This doesn't mean axions don't exist at all. It just means they aren't the "MeV-scale" ones. Scientists will now have to look for axions that are either much heavier (like the original theories suggested) or much lighter, or perhaps they need to invent a completely new theory to fix the Strong CP problem. The "easy" hiding spot has been found and cleared out.

Technical Summary

1. Problem Statement

The QCD axion is a leading candidate for solving the Strong CP problem. While standard axion models require a high decay constant ($f_a \gtrsim 10^8$ GeV), a 2017 proposal by Alves and Weiner suggested a viable "MeV axion" scenario where $f_a \sim 1$ GeV, corresponding to an axion mass $m_a \sim 10$ MeV.

This specific model evades many standard constraints by imposing three conditions:

1. Flavor Restriction: The axion couples only to the first generation of fermions (up, down, electrons).
2. Charge Ratio: The Peccei-Quinn (PQ) charges of the up and down quarks are set to $Q_u = 2$ and $Q_d = 1$. This suppresses the axion-pion mixing ($\theta_{a\pi}$), evading constraints from $\pi^+ \to e^+\nu_e a$.
3. Electron Charge: A non-zero electron PQ charge ($Q_e$) is introduced to allow the axion to decay visibly ($a \to e^+e^-$) with a short lifetime, relaxing constraints from invisible decay searches (e.g., $K^+ \to \pi^+ X_{inv}$).

Despite these conditions, the viability of the MeV axion in the mass range $2m_e \leq m_a \leq 30$ MeV remained an open question, particularly regarding constraints from rare kaon decays. Previous analyses had not thoroughly re-examined the limits from the $K_L \to \pi^0\pi^0 e^+e^-$ decay channel measured by the KTeV experiment.

2. Methodology

The authors perform a comprehensive re-examination of the MeV axion model using a combination of effective field theory (EFT) calculations and a dedicated Monte Carlo (MC) simulation of the KTeV detector.

A. Theoretical Framework

• Effective Lagrangian: The authors utilize the chiral Lagrangian to derive the axion's interactions with mesons. They focus on the $\Delta S = 1$ effective interactions, specifically the octet enhancement mechanism.
• Mixing Angles: They calculate the mixing angles between the axion and neutral mesons ($\pi^0, \eta_{ud}, \eta_s$). A key feature of the model is the suppression of $\theta_{a\pi}$ (making the axion "pion-phobic") while maintaining significant mixing with $\eta_{ud}$ and $\eta_s$.
• Decay Channels:
  • $K^+ \to \pi^+ a$: Analyzed under the assumption that the octet enhancement is dominated by a higher-order operator ($g'_8 \gg g_8$), which suppresses this decay rate compared to the leading-order prediction, allowing the model to survive $K^+$ decay bounds.
  • $K_L \to \pi^0\pi^0 a$: This is the primary channel of interest. The decay proceeds via the mixing of the axion with off-shell neutral mesons ($\pi^0, \eta_{ud}, \eta_s$) in the $K_L \to \pi^0\pi^0 \phi^*$ transition. The branching ratio is proportional to the square of the effective mixing combination: $(\theta_{a\pi} - \theta_{a\eta_{ud}} - \sqrt{2}\theta_{a\eta_s})^2$.

B. Experimental Reinterpretation (KTeV)

The KTeV experiment set an upper limit on the Standard Model (SM) process $K_L \to \pi^0\pi^0 e^+e^-$ at $6.6 \times 10^{-9}$ (90% C.L.). The authors reinterpret this limit for the axion signal $K_L \to \pi^0\pi^0 a$ (with $a \to e^+e^-$).

• Monte Carlo Simulation: They developed a dedicated MC simulation to reproduce the KTeV detector geometry, beam characteristics, and resolution effects.
  • Geometry: Simulated the $K_L$ beam line, drift chambers (DC), and CsI calorimeter.
  • Kinematics: Generated $K_L \to \pi^0\pi^0 a$ events using the matrix elements derived from the chiral Lagrangian.
  • Reconstruction: Implemented the KTeV event reconstruction algorithm, including vertex fitting using $\chi^2_{track}$ (drift chambers) and $\chi^2_{cal}$ (calorimeter), and applied the same background rejection cuts used in the original analysis (e.g., invariant mass cuts, photon conversion vetoes).
• Efficiency Calculation: They determined the signal efficiency ($\epsilon_a$) for various axion masses and lifetimes. They validated their simulation by reproducing the SM efficiency for $K_L \to \pi^0\pi^0 e^+e^-$ within 25% of the reported KTeV value.
• Recasting the Bound: The new bound on the axion branching ratio is derived by scaling the KTeV limit:
  $$\text{Br}(K_L \to \pi^0\pi^0 a) < \text{Br}(K_L \to \pi^0\pi^0 e^+e^-)_{\text{KTeV}} \times \frac{\epsilon_{\text{SM}}}{\epsilon_a} \times \frac{1}{\text{Br}(a \to e^+e^-)}$$

3. Key Contributions

1. First Dedicated KTeV Analysis for MeV Axions: While the KTeV limit was known, this is the first study to perform a full detector simulation and recast the limit specifically for the MeV axion kinematics, accounting for the specific matrix elements and detector acceptance.
2. Handling Theoretical Uncertainties: The authors address the large $O(1)$ uncertainties in the chiral Lagrangian mixing angles (specifically $\theta_{a\eta_{ud}}$ and $\theta_{a\eta_s}$) by treating them as free parameters in a 2D plane analysis, rather than relying solely on leading-order predictions.
3. Comprehensive Constraint Synthesis: They combine the new KTeV bound with existing constraints from:
   • $K^+ \to \pi^+ e^+e^-$ (BNL/E949)
   • $K^+ \to \pi^+ \gamma\gamma$ (E949)
   • $K^+ \to \pi^+ X_{inv}$ (NA62)
   • Electron anomalous magnetic moment ($(g-2)_e$)
   • Beam-dump and collider experiments.

4. Results

• Exclusion of the Viable Window: The analysis demonstrates that the KTeV bound on $K_L \to \pi^0\pi^0 a$ is extremely stringent. Even when accounting for the theoretical uncertainties in the mixing angles, the predicted branching ratio for the MeV axion (based on the viable parameter space) exceeds the KTeV upper limit by orders of magnitude.
• Fine-Tuning Loophole: The only region that survives the KTeV bound requires a "fine cancellation" among the mixing angles ($\theta_{a\pi} \approx \theta_{a\eta_{ud}} + \sqrt{2}\theta_{a\eta_s}$) to suppress the production rate by three orders of magnitude relative to the leading-order prediction.
  • This cancellation is highly unnatural and falls outside the expected range of theoretical uncertainties (parameterized by the chiral expansion parameter $\Delta_\chi \approx 0.73$).
  • Even if this cancellation occurs, the remaining parameter space is further constrained by the electron $(g-2)_e$ measurement and $K^+ \to \pi^+ \gamma\gamma$ limits.
• Benchmark Points: The authors tested three benchmark points (A, B, C) with varying masses and electron charges. In all cases, the KTeV bound, combined with other limits, excludes the parameter space unless the fine-tuning condition is met.

5. Significance

• Ruling Out the MeV Axion: The paper concludes that the specific MeV-scale QCD axion model proposed by Alves and Weiner (with $f_a \sim 1$ GeV) is excluded by current experimental data, primarily due to the $K_L \to \pi^0\pi^0 e^+e^-$ search at KTeV.
• Implications for Axion Physics: This result reinforces the conventional wisdom that the QCD axion decay constant must be large ($f_a \gtrsim 10^8$ GeV), pushing the axion mass to the $\mu$eV scale (or requiring heavy axion models with additional mass contributions).
• Methodological Impact: The study provides a robust framework for recasting rare kaon decay limits for light new particles. The efficiency maps and simulation tools developed can be applied to other light scalar or pseudoscalar searches in the MeV–100 MeV mass range.

In summary, the paper definitively closes the "MeV axion" loophole by demonstrating that the predicted rate of $K_L \to \pi^0\pi^0 a$ is incompatible with the KTeV null result, unless one invokes highly unnatural fine-tuning of chiral Lagrangian parameters.