Measurement of charged-current muon neutrino-argon interactions without pions in the final state using the MicroBooNE detector

Using 1.3×10²¹ protons on target, the MicroBooNE collaboration reports new flux-integrated differential cross-section measurements for charged-current muon neutrino-argon interactions without final-state pions, finding good agreement with event generators in single-differential distributions but more limited agreement in double-differential ones.

The Gist

Imagine you are trying to understand how a specific type of invisible bullet (a neutrino) behaves when it hits a very dense, frozen block of argon (a noble gas).

This paper is a report from the MicroBooNE experiment, a giant, high-tech camera buried underground at Fermilab. Their goal was to take a "snapshot" of what happens when these invisible bullets hit the argon, but with a very specific rule: we only care about the shots where no pions (a type of particle) are created in the explosion.

Here is the story of how they did it, explained simply:

1. The Setup: A Giant 3D Camera

Think of the MicroBooNE detector as a massive, 85-ton tank filled with liquid argon that is colder than outer space. Inside this tank, there are thousands of tiny wires acting like the pixels on a camera.

• The Trigger: A beam of muon neutrinos (the "bullets") is fired from the surface down into the tank.
• The Interaction: Occasionally, a neutrino hits an argon atom. When it does, it creates a spark of light and leaves a trail of electric charge on the wires.
• The Result: The computer reconstructs this into a 3D image, showing exactly where the particles went.

2. The Challenge: Finding the "Ghost"

Neutrinos are notoriously shy. They rarely hit anything. And when they do, the resulting explosion is messy.

• The Goal: The scientists wanted to study a specific type of collision called "CC0π". This stands for "Charged Current, Zero Pions."
• The Analogy: Imagine throwing a ball at a wall of bricks. Sometimes, the ball bounces off cleanly (this is what they want). Sometimes, the ball hits a brick, and the brick shatters into tiny pieces (pions).
• The Problem: In the past, it was hard to tell the difference between a "clean bounce" and a "messy shatter" if the pieces were too small to see. Previous experiments only looked at cases where they could see a proton (a piece of the argon nucleus) flying out. This new study says, "Let's look at all the clean bounces, even if no proton is visible."

3. The Detective Work: Sorting the Noise

The tank is full of "noise." Cosmic rays (particles from space) constantly rain down on the detector, looking like neutrino hits.

• The Filter: The team used a sophisticated AI (a "Boosted Decision Tree") to act like a super-smart detective. It looks at the shape of the tracks left by particles.
  • Muons leave long, straight tracks (like a laser beam).
  • Pions often break apart or look different.
  • Protons are heavy and stop quickly.
• The Selection: They filtered out everything that looked like a pion or a cosmic ray. They kept only the events where a muon was created, and no pions were found.

4. The Measurement: Unfolding the Truth

Here is the tricky part: The camera doesn't see the "true" collision perfectly. It sees a blurry version because of how the particles move through the liquid argon and how the wires read them.

• The Analogy: Imagine trying to guess the speed of a car by looking at its blurry tire tracks in the mud. You know the mud smears the tracks, so you have to do some math to "un-smear" the picture and figure out how fast the car actually was.
• The Math: The scientists used a technique called unfolding (specifically Wiener-SVD) to reverse the blurring. They took the blurry data they collected and mathematically reconstructed what the collision must have looked like in reality.

5. The Results: Checking the Recipe Books

Now that they had the "true" picture of these collisions, they compared it to the "recipe books" (computer models) that physicists use to predict how neutrinos behave.

• The Models: They tested several different computer programs (GENIE, GiBUU, NEUT, NuWro). These programs are like different chefs trying to guess the recipe for a neutrino collision.
• The Verdict:
  • Good News: Most of the chefs got the "single ingredient" measurements right (e.g., how fast the muon was going).
  • The Twist: When looking at the relationship between the speed and the angle of the muon (the "double-differential" measurement), only a few chefs got it right.
  • The Winner: The GiBUU 2025 and NEUT models did the best job of predicting the complex 3D dance of the particles. The older models tended to underestimate how often these clean collisions happen, especially when the muon was moving forward.

Why Does This Matter?

You might ask, "Who cares about argon collisions?"

• The Big Picture: Future experiments (like DUNE) will use massive tanks of liquid argon to study why the universe is made of matter instead of antimatter. To do this, they need to understand neutrino interactions perfectly.
• The Bridge: This paper provides a "gold standard" measurement. It helps physicists tune their computer models so that when they look at neutrinos in the future, they can be sure their models are correct. It's like calibrating a telescope before you try to map the stars.

In a nutshell: MicroBooNE took a huge dataset of neutrino hits, used AI to filter out the junk, mathematically cleaned up the blurry images, and found that while our current computer models are getting better, they still need a little more tuning to perfectly predict how neutrinos dance with argon atoms.

Technical Summary

1. Problem and Motivation

Neutrino oscillation experiments (such as DUNE and Hyper-Kamiokande) aim to determine the neutrino mass ordering and measure CP violation. Achieving the required precision demands an unprecedented understanding of neutrino-nucleus interaction models.

• The Challenge: A major source of systematic uncertainty in oscillation analyses is the modeling of "quasielastic-like" (CCQE-like) interactions. In many detectors (like Water Cherenkov), the signal is defined by a single ring (no pions), but the detector cannot distinguish between true CCQE events and events where a pion was produced but absorbed or not detected (Final State Interactions - FSI).
• The Gap: Previous MicroBooNE measurements focused on $\nu_\mu$ CC interactions with at least one reconstructed proton ($1\ell1p$). However, a significant portion of CCQE-like events involves no final-state protons (e.g., due to FSI absorption or multinucleon interactions). There was a lack of high-statistics data on the $\nu_\mu$ CC0$\pi$ channel (no pions) that includes events with zero protons ($0\ell0p$) to fully map the kinematics of these interactions on Argon.
• Goal: To provide a flux-integrated, differential cross-section measurement for $\nu_\mu$ CC0$\pi$ interactions on Argon, relaxing the proton requirement to include all topologies, thereby offering a benchmark for neutrino event generators and facilitating comparisons with Cherenkov detectors.

2. Methodology

Experimental Setup

• Detector: MicroBooNE, a Liquid Argon Time Projection Chamber (LArTPC) located at Fermilab's Booster Neutrino Beam (BNB).
• Dataset: Full exposure of $1.3 \times 10^{21}$ protons on target (POT) collected between 2015 and 2020. This is approximately double the dataset used in previous MicroBooNE CC0$\pi$ measurements.
• Beam: Predominantly $\nu_\mu$ (93.7%) with an average energy of $\sim 0.8$ GeV.

Event Selection and Reconstruction

The analysis employs a topological selection strategy using the Pandora multi-algorithm framework and an XGBoost (Gradient Boosted Decision Tree) for particle identification (PID).

• Signal Definition: $\nu_\mu$ CC interactions on $^{40}$Ar with:
  • Final-state muon momentum: $0.1 < p_\mu < 2.0$ GeV/c.
  • No neutral pions ($\pi^0$).
  • No charged pions with $p_\pi > 70$ MeV/c.
  • Crucial Update: Unlike previous works, no final-state protons are required.
• Background Rejection:
  • Cosmic-ray muons are rejected using topological likelihoods and calorimetric Bragg peak analysis.
  • Neutral Current (NC) and $\nu_e$ CC backgrounds are suppressed using PID BDTs that distinguish muons from protons/pions based on $dE/dx$ profiles and track topology.
  • Sideband control regions (NC, CCN$\pi$, and Out-of-Fiducial-Volume) were used to validate background modeling.
• Efficiency: The overall selection efficiency is 13%, with a sample purity of 71%. The efficiency drops significantly for high-momentum muons due to the requirement that muons be fully contained within the detector.

Cross-Section Extraction

• Unfolding: The measured reconstructed distributions were unfolded to the true particle level using the Wiener-SVD (Singular Value Decomposition) technique. This method regularizes the unfolding to preserve statistical power while correcting for detector resolution and bin migration.
• Observables: The results are presented as:
  1. Single-differential cross sections in muon momentum ($p_\mu$).
  2. Single-differential cross sections in muon scattering angle ($\cos \theta_\mu$).
  3. Double-differential cross sections in the joint space of ($p_\mu$, $\cos \theta_\mu$).

3. Key Contributions

1. First Inclusive CC0$\pi$ Measurement on Argon: This is the first high-statistics measurement of the CC0$\pi$ channel on Argon that includes events with no final-state protons. This expands the kinematic reach and provides a more complete picture of the "CCQE-like" signal.
2. Extended Kinematic Range: The muon momentum range was extended up to 2 GeV/c (previously limited to lower momenta in similar analyses).
3. Comprehensive Generator Comparison: The data was compared against a suite of modern neutrino event generators (GENIE v3.2.0, GiBUU 2025, NEUT 5.4.0.1, and NuWro 21.09.2) using a consistent framework (NUISANCE).
4. Validation of Unfolding: The paper rigorously validates the unfolding procedure using "fake data" tests with alternative models (NuWro) and extreme parameter variations, demonstrating robustness against model bias.

4. Results

Data vs. Simulation

• Event Rates: The observed event rates in 1D and 2D distributions show good agreement with the central value MicroBooNE-tuned GENIE model.
• Backgrounds: The dominant backgrounds are $\nu_\mu$ CC with pions ($\sim 5.5\%$) where pions are misidentified, and Cosmic Ray activity ($\sim 7.8\%$). Sideband validations confirmed these backgrounds are well-modeled.

Generator Performance

The comparison of unfolded data with various generators revealed:

• 1D Distributions: All tested generators (GENIE, GiBUU, NEUT, NuWro) provide a reasonable description of the single-differential distributions ($p_\mu$ and $\cos \theta_\mu$ individually).
• 2D Distributions (Correlated): The double-differential measurement provides a stricter test.
  • Best Performers: GiBUU 2025 and NEUT showed the best agreement with the correlated 2D data (high $p$-values).
  • Underperformers: NuWro and the standard GENIE G18-10a tune showed significant discrepancies ($p < 5\%$) in the 2D space, particularly failing to describe the angular distribution as a function of momentum.
• Normalization: Most generators tend to underpredict the normalization, especially for forward angles and intermediate muon momenta ($0.5 - 0.7$ GeV/c).

Uncertainties

The total uncertainty in the extracted cross sections is approximately 10–20%, dominated by:

• Neutrino interaction modeling (GENIE tune parameters).
• Neutrino flux uncertainties.
• Detector response (charge collection, electric field).
• Statistical uncertainties (data and MC).

5. Significance

• Model Tuning: The results provide critical constraints for tuning neutrino interaction models, specifically regarding multinucleon effects (2p2h) and Final State Interactions (FSI) that lead to pion-less final states. The failure of some generators in the 2D space highlights the need for improved modeling of the correlation between muon momentum and angle.
• Oscillation Physics: By characterizing the CC0$\pi$ channel on Argon (the target for DUNE), this measurement reduces systematic uncertainties for future long-baseline oscillation experiments. It helps distinguish between true CCQE events and background events that mimic them.
• Cross-Experiment Comparison: The topological definition of the signal (no pions) allows for direct comparison with Water Cherenkov experiments (like T2K and Hyper-K), bridging the gap between different detector technologies and target nuclei.
• Methodological Benchmark: The use of Wiener-SVD unfolding and rigorous sideband validation sets a standard for future LArTPC cross-section measurements.

In conclusion, this paper delivers a high-precision benchmark dataset for $\nu_\mu$-Argon interactions, revealing that while current generators capture 1D trends, they struggle with the full 2D kinematic correlations, pointing to specific areas where nuclear physics modeling requires refinement.