New measurement of $K^+\to\pi^+\nu\bar\nu$ branching ratio at the NA62 experiment

The NA62 experiment at CERN presents a new measurement of the $K^+\to\pi^+\nu\bar\nu$ branching ratio using combined 2016–2024 data, yielding a result of $(9.6^{+1.9}_{-1.8})\times10^{-11}$ that is compatible with the Standard Model prediction with better than 20% precision.

The Gist

The Big Picture: Hunting a "Ghost" Particle

Imagine you are trying to find a specific, incredibly rare coin hidden in a massive, chaotic pile of trash. This coin is the K+ →π+ν¯ν decay.

In the world of particle physics, this is known as a "golden mode." It's a process where a Kaon (a type of unstable particle) turns into a Pion (another particle) and two invisible "ghosts" (neutrinos).

• Why is it special? According to our current rulebook of physics (the Standard Model), this should happen very rarely—about 1 time in every 10 billion Kaons.
• Why do we care? Because if we find it happening more or less often than the rulebook predicts, it's a smoking gun for "New Physics." It could mean there are hidden particles or forces we don't know about yet, potentially revealing secrets about the universe at scales as huge as 100,000 times the size of a proton.

The Detective Agency: The NA62 Experiment

The NA62 experiment at CERN (in Switzerland) is like a high-tech detective agency built specifically to catch this ghost.

• The Setup: They shoot a beam of protons at a target to create a stream of Kaons. These Kaons zoom through a vacuum tube (the "decay volume").
• The Challenge: Most Kaons decay in common, boring ways. The experiment has to filter out billions of "boring" events to find the few "ghost" events. It's like trying to hear a whisper in a rock concert.
• The Gear: They use a massive array of detectors (like giant cameras and speed traps) to track every particle. If a particle doesn't match the "ghost" signature perfectly, it gets tossed out.

The New Update: 2023–2024 Data

This paper reports on new data collected recently. Here is what they did differently and what they found:

1. Turning Down the Volume

For a while, the particle beam was so intense (too many particles crashing into each other) that the detectors got overwhelmed, like a microphone blowing out from too much volume.

• The Fix: In 2023, they turned the beam intensity down to 75%.
• The Result: Surprisingly, this made the experiment better. It's like turning down the music at a party so you can actually hear the conversation. This change reduced "background noise" (false alarms) while keeping the signal clear.

2. Smarter Software (The AI Upgrade)

The team upgraded their software with new "brainy" algorithms (using Transformers and Neural Networks).

• The Analogy: Imagine trying to find a specific person in a crowd. The old software was like a security guard checking IDs. The new software is like a super-smart AI that recognizes a person's gait, height, and clothing style instantly, even if they are partially hidden.
• The Result: They caught more of the real Kaons and filtered out the fake ones much more effectively.

3. The Double-Dip

By combining the new 2023–2024 data with all the data collected since 2016, they effectively doubled their sample size. They now have a much larger "search party" looking for the ghost.

The Verdict: Did They Find It?

Yes! And the results are beautiful.

• The Count: They found 84 candidate events (the ghost particles) across all their data.
• The Prediction: The Standard Model predicted they should find about 23 events in this specific new batch, and the math worked out perfectly when they added the old data.
• The Measurement: They calculated the "branching ratio" (the probability of this happening) to be 9.6 in 100 billion.
  • Previous measurements were a bit shaky (like a blurry photo).
  • This new measurement is sharp and precise (like a high-definition photo).

Why This Matters

The result is 9.6 ± 1.9.

• The "± 1.9" is the margin of error.
• The Standard Model predicted roughly 8.4 to 8.6.
• The new measurement overlaps perfectly with the prediction.

The Takeaway:
Think of the Standard Model as a weather forecast that says "It will rain tomorrow." The NA62 experiment went outside, measured the rain, and said, "You were right! It rained exactly as predicted."

This is a huge victory for the Standard Model. It means our current understanding of the universe is incredibly robust. However, in science, "boring" is also good news because it tells us we need to look even harder and build even better detectors to find the tiny cracks in the theory where the real new physics might be hiding.

In short: The NA62 team caught the ghost, counted it, and confirmed that the universe is behaving exactly as the rulebook says it should—so far! They will keep collecting data until 2026 to see if the pattern holds up or if a surprise appears.

Technical Summary

1. Problem and Motivation

The decay $K^+ \to \pi^+ \nu\bar{\nu}$ is a Flavor Changing Neutral Current (FCNC) process that is highly suppressed in the Standard Model (SM). It proceeds via electroweak box and penguin diagrams, heavily suppressed by the Glashow-Iliopoulos-Maiani (GIM) mechanism and the small Cabibbo-Kobayashi-Maskawa (CKM) matrix element governing $t \to d$ transitions.

• Theoretical Significance: The SM predicts a branching ratio of approximately $(8.4 \text{--} 8.6) \times 10^{-11}$ with a precision better than 8%. The theoretical uncertainty is dominated by CKM parameters ($V_{cb}$ and $\gamma$), while intrinsic theoretical uncertainty is only $\sim 3\%$.
• New Physics Sensitivity: This "golden mode" is exceptionally sensitive to physics Beyond the Standard Model (BSM), probing mass scales up to $\mathcal{O}(100 \text{ TeV})$. Deviations from the SM prediction could indicate new particles or interactions.
• Previous Status: Previous measurements by BNL (E787/E949) and early NA62 data (2016–2022) provided evidence and a 5$\sigma$ observation, but with large uncertainties ($\sim 25\%$). The goal was to reduce this uncertainty to the 20% level to rigorously test the SM.

2. Methodology and Experimental Setup

The analysis utilizes data collected by the NA62 experiment at the CERN SPS between 2023 and 2024, combined with previous datasets (2016–2022).

Experimental Configuration

• Beam: A 75 GeV/c secondary beam with a 6% $K^+$ component is generated by directing a 400 GeV proton beam onto a beryllium target.
• Detector Upgrades (2021–2024):
  • Intensity Optimization: In 2023, beam intensity was reduced to 75% of the maximum design intensity to optimize detector performance and reduce saturation effects.
  • KTAG Upgrade: The radiator gas in the differential Cherenkov counter (KTAG) was changed from $N_2$ to $H_2$, reducing the material budget from 3.9% $X_0$ to 0.7% $X_0$.
  • Tracking Algorithms: A transformer-based 4D beam-tracking algorithm was developed for the GTK (beam spectrometer), reducing the probability of missing a true $K^+$ from 6% to 4%.
  • PID Improvements: A Convolutional Neural Network (CNN) based calorimetric Particle Identification (PID) was implemented using LKr and MUV1/2 data to better suppress $\mu^+ \to \pi^+$ misidentification, particularly when photon clusters overlap with muon tracks.
• Veto Systems: Comprehensive veto systems (LAV, IRC, SAC, LKr, MUV) reject background from $\pi^0$ decays, upstream interactions, and beam halo. A specific offline veto using the first LAV station within a 2 ns window was introduced to suppress upstream background events.

Analysis Strategy

• Normalization Channel: The analysis uses $K^+ \to \pi^+ \pi^0$ as the normalization channel to cancel systematic uncertainties related to beam flux and detection efficiency.
• Signal Selection:
  • Events are selected based on the squared missing mass variable: $m^2_{miss} = (P_K - P_\pi)^2$.
  • The analysis is binned into six categories based on the $\pi^+$ momentum (15–45 GeV/c).
  • Signal Regions (SR) are defined to suppress dominant background modes ($K^+ \to \pi^+ \pi^0$, $K^+ \to \mu^+ \nu$, etc.) by a factor of $\mathcal{O}(10^4)$.
• Background Estimation:
  • In-FV Backgrounds: Evaluated using control regions and kinematic tail fractions derived from data or Monte Carlo (MC) simulations.
  • Upstream Backgrounds: Evaluated using an "Upstream Reference Sample" (URS) where $K/\pi$ matching is inverted. A template fit is used to extrapolate these backgrounds to the signal region, corrected for matching probabilities.

3. Key Contributions

1. Dataset Expansion: The 2023–2024 dataset doubled the effective signal sample size compared to previous analyses while proportionally reducing background levels.
2. Algorithmic Innovations: The implementation of transformer-based tracking and CNN-based PID significantly improved reconstruction efficiency and background rejection, particularly for challenging overlap scenarios.
3. Operational Optimization: The strategic reduction of beam intensity and the gas change in KTAG allowed the experiment to operate at peak sensitivity, improving the efficiency of additional selections ($\epsilon_{RV}$) by 14%.
4. Comprehensive Combination: The paper presents the first combined result of the full 2016–2024 dataset, integrating the new high-statistics run with previous observations.

4. Results

• 2023–2024 Dataset:
  • Observed events in the Signal Region (SR): 33.
  • Expected background: $11.9^{+2.9}_{-2.3}$.
  • Measured Branching Ratio:
    $$B(K^+ \to \pi^+ \nu\bar{\nu}) = (7.2^{+2.3}_{-2.1}) \times 10^{-11}$$
• Combined 2016–2024 Dataset:
  • Total observed events: 84.
  • Total expected background: $30^{+4}_{-3}$.
  • Final Measured Branching Ratio:
    $$B(K^+ \to \pi^+ \nu\bar{\nu}) = (9.6^{+1.9}_{-1.8}) \times 10^{-11}$$
    (Where the uncertainty is $+1.8/-1.6$ (stat) and $+0.8/-0.6$ (syst))

5. Significance and Conclusion

• Precision: The combined result achieves a total precision of 20%, a significant improvement over previous measurements.
• Standard Model Compatibility: The result is fully compatible with the SM prediction (which ranges roughly from $7.9 \times 10^{-11}$ to $8.6 \times 10^{-11}$ depending on the CKM fit used).
• Statistical Significance: The observation excludes the background-only hypothesis with a significance exceeding 6$\sigma$.
• Implications: The measurement confirms the SM prediction for this rare decay with high precision, placing stringent constraints on BSM models that predict deviations in FCNC processes. The result is consistent with previous BNL experiments and the KOTO experiment's limits on the neutral mode ($K_L \to \pi^0 \nu\bar{\nu}$).
• Future Outlook: With data-taking continuing until 2026, the total dataset is expected to increase by another 50%, potentially pushing the precision below 10% and further probing the frontiers of flavor physics.