Sweeping the pion chimney for axion-like particles with KOTO

This paper demonstrates that recasting data from the J-PARC KOTO experiment's search for $K_L \to 6\gamma$ decays can establish novel limits on prompt axion-like particles in the challenging "pion chimney" mass range and extend sensitivity to a broader mass spectrum by incorporating displaced decays.

The Gist

Imagine the universe is filled with a vast, invisible ocean of particles. Among the most famous "fish" in this ocean are particles called pions (specifically the neutral pion, $\pi^0$). For decades, physicists have been trying to catch a new, ghostly type of fish called an Axion-Like Particle (ALP). These ALPs are so elusive that they usually slip right through our nets.

However, there is a specific, tricky spot in the ocean where the ALPs are almost the exact same size and weight as the pions. The authors of this paper call this spot the "Pion Chimney."

The Problem: The Chimney is a Blind Spot

Usually, scientists look for ALPs by seeing if they decay (break apart) into light particles (photons) far away from where they were created. This "delay" helps them tell the ALP apart from the common pion.

But in the "Pion Chimney," the ALP is so similar to the pion that it decays immediately, right next to where it was born. It's like trying to spot a specific twin in a crowd of identical twins who are standing right next to each other. Because they look so much alike and happen at the same time, standard experiments can't tell them apart. This has left a gap in our knowledge where we simply don't know if these ALPs exist or not.

The Solution: The KOTO Experiment as a Detective

The authors propose a clever new way to catch these "chimney" ALPs using data from the KOTO experiment in Japan.

Think of the KOTO experiment as a high-speed camera taking pictures of Kaons (another type of particle) as they fly through a detector and break apart.

• The Standard Event: Usually, a Kaon breaks into three pions ($\pi^0$). Each pion instantly turns into two flashes of light (photons). So, the camera sees six flashes of light ($3 \times 2 = 6$).
• The New Search: The authors ask: "What if one of those pions was actually a sneaky ALP?" If a Kaon breaks into two pions and one ALP ($2\pi^0 + a$), and the ALP also turns into two flashes of light, the camera still sees six flashes of light.

To the camera, the two events look identical. But the authors realized that the math behind the scenes is different.

The Trick: The "Weighted Average" Illusion

Here is the creative analogy: Imagine you are trying to guess the weight of a mystery object by looking at how it bounces off a wall.

• If the object is a standard pion, it bounces in a very predictable way, and when you calculate its "reconstructed mass" (what the computer thinks it weighs), it lands perfectly on the known weight of a Kaon.
• If the object is a chimney ALP, it is slightly heavier or lighter than a pion. When the computer tries to do the math assuming it's a pion, the numbers get confused. The "reconstructed mass" of the Kaon shifts slightly to the left or right.

The authors showed that if these ALPs exist, they wouldn't just add a little noise to the data. Instead, they would create new, distinct peaks (hills) in the graph of the Kaon's mass, sitting right next to the main hill. It's like hearing a second, slightly higher-pitched note played alongside a main note; you can hear the difference even if you can't see the instrument.

What They Did

1. Simulated the Scene: They built a computer model of the KOTO detector to see exactly how it "sees" these six flashes of light.
2. Checked the Data: They looked at real data from KOTO (collected from 200 trillion protons hitting a target) to see the "hill" of the standard Kaon mass.
3. The Search: They scanned the data for those extra, shifted hills that would appear if ALPs were hiding in the Pion Chimney.

The Results

• No Ghosts Found (Yet): They didn't find any new hills in the data. This means ALPs in this specific mass range are rarer than we thought, or they don't exist at all.
• New Limits: Because they didn't find them, they can now draw a new "fence" around the Pion Chimney. They can say with confidence: "If these ALPs exist, they must be weaker than this specific limit." This is the first time anyone has been able to set such strict rules for this specific, hard-to-probe mass range.
• Future Potential: They also showed that if we look at the data differently (allowing for ALPs that travel a tiny bit before decaying), we could potentially find ALPs that are even lighter than the pion.

The Bottom Line

This paper is like a detective saying, "We couldn't find the thief in the crowded room, but by analyzing exactly how the shadows fell on the wall, we now know exactly where the thief couldn't have been hiding." They have successfully swept the "Pion Chimney" clean, ruling out a whole class of potential new particles that were previously invisible to science.

Technical Summary

Technical Summary: Sweeping the Pion Chimney for Axion-Like Particles with KOTO

Problem Statement
Probing Axion-Like Particles (ALPs) in the mass range immediately adjacent to the neutral pion ($m_{\pi^0}$), a region termed the "pion chimney" (defined here as $120 \le m_a \le 150$ MeV), presents significant experimental challenges. This difficulty arises from two primary factors:

1. Resonant Mixing: At the effective theory level, ALP-pion mixing is resonantly enhanced when $m_a \approx m_{\pi^0}$. Consequently, ALPs that would typically be long-lived become prompt decays. This severely limits the sensitivity of searches relying on displaced vertices to separate signal from background, as well as searches for invisible decays.
2. Background Rejection: Precision diphoton measurements typically apply cuts on the diphoton invariant mass ($m_{\gamma\gamma}$) to avoid overwhelming $\pi^0 \to \gamma\gamma$ backgrounds. These cuts inherently exclude the mass window near $m_{\pi^0}$, rendering standard recasts of data (e.g., from KTeV) ineffective in this specific regime.

Methodology
The authors propose recasting data from the J-PARC KOTO experiment, specifically the $K_L \to 3\pi^0 \to 6\gamma$ dataset collected during the first run ($2 \times 10^{14}$ protons on target), to search for the process $K_L \to 2\pi^0 a \to 6\gamma$.

• Simulation and Reconstruction: The authors developed a custom weighted Monte Carlo (MC) generator to emulate the KOTO detector simulation. This includes the beamline geometry, the CsI electromagnetic calorimeter (CsIC) response, and the specific event selection criteria (six reconstructed photon hits with $E \ge 50$ MeV, isolated by 150 mm).
  • The simulation approximates the KOTO reconstruction procedure, which assumes the $K_L$ decay vertex lies on the $z$-axis and uses a perturbative method to reconstruct the vertex and invariant mass ($m_{KL}^{rec}$) based on photon pairings.
  • Crucially, the simulation applies the pion mass constraint ($m_{\pi^0}^2 = 2E_i E_j (1 - \cos \theta_{ij})$) to all photon pairs, including those originating from the ALP decay.
• Signal Morphology: The core insight is that if an ALP with mass $m_a \neq m_{\pi^0}$ is reconstructed assuming it is a pion, the kinematic mismatch induces a shift in the reconstructed $K_L$ mass. The authors derive that the shift in the reconstructed mass peak, $\delta m_{KL}^{rec}$, is approximately proportional to the mass difference: $\delta m_{KL}^{rec} \simeq -1.2 (m_a - m_{\pi^0})$. This results in single or multipeak structures shifted relative to the standard $m_{KL}$ peak.
• Statistical Analysis:
  • Background Modeling: Since a first-principles calculation of the $m_{KL}^{rec}$ spectrum is not feasible, the authors employ a phenomenological approach. They fit the Standard Model (SM) background (the $3\pi^0$ peak plus continuum tails) using a Voigt-like profile modified by rational polynomial prefactors. The Akaike Information Criterion (AIC) selects a specific model ($f_{4,3}$) with 9 parameters.
  • Hypothesis Testing: Rather than subtracting the background fit and analyzing residuals, the authors perform a combined signal-plus-background fit. They introduce a parameter $\rho = \text{Br}[K_L \to 2\pi^0 a] / \text{Br}[K_L \to 3\pi^0]$ and calculate the change in $\chi^2$ ($\delta\chi^2$) relative to the SM-only hypothesis. Limits are set at 90% Confidence Level (CL) corresponding to $\delta\chi^2 < 1.64$.
• Beyond the Chimney: The paper also explores a modified reconstruction strategy for ALPs outside the chimney mass range (specifically $m_a < m_{\pi^0}$). This involves relaxing the mass constraint for one photon pair and searching for a "bump" in the reconstructed diphoton invariant mass ($m_a^{rec}$). This approach also accounts for displaced decays of longer-lived ALPs by solving for a decay displacement parameter $\lambda$ that satisfies the $K_L$ mass constraint.

Key Contributions and Results

1. Novel Limits in the Pion Chimney: Using the $2 \times 10^{14}$ POT dataset, the authors derive novel limits on $\text{Br}[K_L \to 2\pi^0 a]$ for ALP masses between 120 and 150 MeV. The extracted limits are of the order of a few $\times 10^{-4}$ to a few $\times 10^{-3}$.
2. Coupling Constraints: These branching ratio limits are translated into constraints on ALP-SM Wilson coefficients ($c_{gg}$ and $c_{dR}$) for three benchmark scenarios. The analysis demonstrates that KOTO calibration data can probe regions of the parameter space that are typically inaccessible to other searches due to the "pion chimney" difficulty.
3. Displaced Decay Sensitivity: For ALPs outside the chimney, the modified reconstruction technique ($m_a^{rec}$) yields branching ratio limits comparable to those derived from the $m_{KL}^{rec}$ spectrum within the chimney. The study notes that while longer lifetimes ($c\tau = 1-10$ cm) offer some separation from prompt backgrounds, acceptance losses due to decays occurring outside the fiducial volume or beyond the calorimeter can weaken these limits.
4. Validation: The MC simulation faithfully reproduces the observed $K_L \to 3\pi^0 \to 6\gamma$ signal peak and shape, validating the reconstruction approximations used.

Significance and Claims
The paper claims that this approach provides a method to "sweep" the pion chimney, a mass range notoriously difficult to probe. By leveraging the precise kinematic reconstruction of the $K_L$ mass in existing calibration data, the authors demonstrate that KOTO can set competitive, novel limits on prompt ALPs without requiring new data collection or dedicated triggers.

The authors maintain a modest tone regarding their claims:

• They emphasize that the limits are prospective. The results rely on a phenomenological description of the background and a simplified detector simulation.
• They acknowledge that model dependence from the choice of lineshape and the treatment of bin-bin correlations (which are currently neglected) could affect the precision of the limits.
• They state that a "fully-fledged analysis within an experimental framework" is required to validate the precise confidence levels and to incorporate correlations and detailed background modeling (such as accidental overlays from halo neutrons).
• They note that while the technique is powerful for neutral kaon decays, it generally remains less constraining than limits from charged kaon decays ($K^+ \to \pi^+ a$) in most models, though it offers unique sensitivity in the specific chimney region where charged kaon searches may be less effective due to different kinematic or background constraints.

In conclusion, the work establishes a proof-of-concept that existing KOTO datasets can be recast to probe the pion chimney, potentially covering the entire mass range with future, larger datasets ($10^{21}$ POT), provided systematic uncertainties are rigorously controlled.