Measurement of $\pi^0$ Production in $\bar{\nu}_{\mu}$ Charged-Current Interactions in the NOvA Near Detector

The NOvA collaboration presents the most precise measurement to date of muon antineutrino-induced charged-current neutral pion production on a hydrocarbon target, finding that while results agree with GENIE predictions, other models underestimate the cross section in the $\Delta$(1232) resonance region.

The Gist

The Big Picture: Catching Ghosts in a Fish Tank

Imagine the universe is filled with neutrinos. These are tiny, ghost-like particles that zip through everything—planets, stars, and even your body—without ever saying "hello." They rarely stop to talk to anything.

The NOvA experiment is like a massive, high-tech fish tank (the "Near Detector") built specifically to catch these ghosts when they do decide to bump into something. This paper is about a specific type of collision: when an antineutrino (a ghost with a negative charge) smashes into a carbon atom and creates a neutral pion (a short-lived particle that instantly explodes into two flashes of light).

Why do we care? Because understanding exactly how these collisions happen helps scientists figure out the secrets of the universe, like why there is more matter than antimatter. But to do that, we need to know the "rules of the road" for these particles. This paper is a new, very precise rulebook.

---

1. The Setup: A Beam of Ghosts

The scientists at Fermilab create a beam of antineutrinos by shooting protons (like tiny bullets) into a block of graphite. This creates a shower of particles that decay into antineutrinos. They focus this beam using giant magnetic "horns" (like a funnel for water) and shoot it toward their detector.

• The Analogy: Imagine firing a stream of invisible darts at a wall made of carbon bricks. Most darts pass right through, but occasionally, one hits a brick and causes a tiny, spectacular explosion of light.

2. The Detective Work: Sorting the Noise

The detector is filled with liquid plastic (scintillator) that glows when a particle passes through. When a collision happens, it leaves a trail of light. The problem? The detector sees everything. It sees the main collision, but also background noise, other types of particles, and particles that bounced around inside the atom before escaping.

• The Analogy: Imagine you are at a loud concert trying to hear one specific singer. You have to filter out the bass, the drums, the crowd cheering, and the guy next to you talking on his phone.
• The Filter: The scientists used a super-smart computer program (an AI called a "Convolutional Neural Network") to act as a bouncer. It looked at the "shape" of the light trails.
  • Did it look like a muon (a heavy cousin of the electron)?
  • Did it look like two photons (light particles) coming from a pion explosion?
  • If the answer was yes, the event was kept. If not, it was tossed out.

3. The Measurement: Counting the Pieces

Once they isolated the "good" collisions, they measured the speed and direction of the pieces flying out (the muon and the pion). They wanted to see if their predictions matched reality.

• The Prediction: They used a computer model called GENIE to guess what would happen. It's like a weather forecast for particle physics.
• The Reality: They compared the computer's forecast to the actual data from the detector.

4. The Surprise: The "Delta" Glitch

Here is the main discovery of the paper:

The computer model (GENIE) was mostly right, but it made a specific mistake. It underestimated how often a specific type of collision happened involving a particle called the Delta resonance (think of it as a "super-excited" version of a proton).

• The Analogy: Imagine you are baking cookies. Your recipe (the model) says you should get 10 cookies per batch. You bake 100 batches and count 120 cookies. You realize your recipe is slightly off when the oven is set to a specific temperature (the "Delta region").
• The Result: The data showed that in the "Delta region" (where the energy is just right to create this excited state), the actual number of pions produced was about 20% higher than the model predicted.

Other computer models (called NuWro and NEUT) were even worse; they predicted even fewer pions than the data showed.

5. Why This Matters

You might ask, "So what if the model is off by 20%?"

• The Analogy: If you are trying to navigate a ship across the ocean using a map, and the map says the water is 20% shallower than it really is, you might run into rocks.
• The Stakes: In neutrino physics, scientists are trying to measure tiny differences in how neutrinos change their "flavor" (oscillate) over long distances. To do this, they need to know exactly how neutrinos interact with matter. If the "interaction rules" (cross-sections) are wrong, the whole calculation for the oscillation experiment could be wrong.

The Takeaway

This paper is like a quality control report for the universe's instruction manual.

1. We built a better detector: We caught way more events than ever before (6 times more than the previous record).
2. We found a bug: The current "instruction manual" (the GENIE model) underestimates how often certain explosions happen.
3. We fixed the map: By providing these precise numbers, the scientists are helping other researchers update their models so that future experiments (like the ones trying to solve the mystery of the universe's existence) can navigate correctly.

In short: The ghosts are hitting the wall more often than we thought, specifically in one specific way, and now we have the data to fix our maps.

Technical Summary

1. Problem Statement

Precise knowledge of neutrino and antineutrino cross-sections on nuclear targets in the few-GeV energy range is critical for the success of future neutrino oscillation experiments. A specific challenge arises from neutral pion ($\pi^0$) production. In oscillation experiments searching for electron neutrino ($\nu_e$) appearance, the decay of a $\pi^0$ into two photons ($\gamma\gamma$) can mimic the electromagnetic shower of an electron, creating a significant background.

While neutral-current (NC) $\pi^0$ production is the dominant background, it is difficult to constrain directly because the outgoing lepton is not detected. Therefore, precise measurements of charged-current (CC) $\bar{\nu}_\mu$ induced $\pi^0$ production are essential to constrain the kinematic models used to predict NC backgrounds. Prior to this work, there was only one previous measurement of $\bar{\nu}_\mu$ CC $\pi^0$ production (by MINERvA at 3.6 GeV), leaving a gap in data at the lower average energy (~2 GeV) relevant to current long-baseline experiments like NOvA.

2. Methodology

Experimental Setup

• Experiment: The NOvA experiment uses the NuMI (Neutrinos at the Main Injector) beam at Fermilab.
• Detector: The measurement utilizes the NOvA Near Detector (ND), a 300-ton liquid scintillator tracking calorimeter located 1 km from the target. The target material is primarily hydrocarbon (CH).
• Beam: An antineutrino-enhanced beam (polarity reversed horns) with a peak energy of ~2 GeV. The exposure corresponds to $11.38 \times 10^{20}$ Protons-On-Target (POT). The beam purity is ~92% $\bar{\nu}_\mu$, with small wrong-sign ($\nu_\mu$) and electron neutrino components.

Event Selection and Reconstruction

• Signal Definition: The signal is defined as $\bar{\nu}_\mu + CH \to \mu^+ + \pi^0 + X$.
• Kinematic Cuts: To ensure signal purity and efficiency, the analysis restricts the muon momentum ($0.5 \le p_\mu < 2.5$ GeV/c) and angle ($\theta_\mu < 60^\circ$).
• Reconstruction:
  • Muon: Identified via a Boosted Decision Tree (MuonID) and track length.
  • $\pi^0$: Reconstructed from two electromagnetic-like (EM-like) prongs (photons) using a Convolutional Neural Network (CNN) to distinguish EM showers from hadrons. The $\pi^0$ momentum is the vector sum of the two photon momenta.
  • Variables: Differential cross-sections are measured as functions of $\pi^0$ momentum/angle, muon momentum/angle, squared four-momentum transfer ($Q^2$), and the hadronic invariant mass proxy ($W^{EXP}$).

Statistical Analysis

• Template Fit: To separate signal from backgrounds, a simultaneous fit is performed using six templates:
  1. Signal-like (CCN$\pi^0$): Contains true $\mu^+$ and $\pi^0$ (includes $\bar{\nu}_\mu$ signal and $\nu_\mu$ induced $\pi^0$).
  2. Sidebands: Defined by low EM-scores, low invariant mass ($m_{\gamma\gamma}$), or high NCID scores to constrain backgrounds like CC0$\pi^0$ (no $\pi^0$), NC interactions, and other misreconstructed events.
• Unfolding: The D'Agostini iterative unfolding algorithm (2 iterations) is applied to correct for detector resolution effects (smearing and bin migration) to extract the true differential cross-sections.

Systematic Uncertainties

Uncertainties are evaluated using a "multi-universe" method where parameters for flux, interaction models (GENIE), and detector response are varied. Key sources include:

• Flux: ~8% uncertainty on the $\bar{\nu}_\mu$ flux.
• Interaction Modeling: Uncertainties in resonance (Delta) and Deep Inelastic Scattering (DIS) models, and Final State Interactions (FSI).
• Detector: Energy calibration, light yield, and aging effects.
• Total Systematic: ~14.9%, dominating over statistical uncertainty (5.5%).

3. Key Contributions

• High-Statistics Dataset: This measurement represents a six-fold increase in statistics compared to the previous MINERvA measurement on a similar target, providing the most precise data on antineutrino-induced $\pi^0$ production to date.
• Kinematic Differential Cross-Sections: The paper provides the first high-precision differential measurements of $\bar{\nu}_\mu$ CC $\pi^0$ production in the 1–5 GeV range for:
  • $\pi^0$ momentum and angle.
  • Muon momentum and angle.
  • $Q^2$ and $W^{EXP}$.
• Model Validation: It offers a rigorous test of neutrino interaction generators (GENIE, NuWro, NEUT) specifically in the antineutrino regime, which is less constrained than the neutrino regime.

4. Results

Comparison with GENIE (v3.0.6)

• Overall Agreement: The data generally agree well with the NOvA-tuned GENIE model (G18 10j 00 000 configuration).
• Delta Resonance Deficit: A systematic underestimation of the data is observed in the $\Delta(1232)$ resonance region (low $W^{EXP}$ < 1.4 GeV and low $\pi^0$ momentum < 0.4 GeV/c).
  • In the $\pi^0$ momentum peak bin (~0.2 GeV/c), GENIE underestimates the cross-section by approximately 20%.
  • The $W^{EXP}$ distribution shows the data are systematically above the GENIE prediction in the resonance-dominated region, suggesting the model underestimates $\Delta^0$ production.
• FSI Impact: Including Final State Interactions (FSI) in the simulation significantly improves the agreement with the $\pi^0$ momentum distribution, particularly at low momenta, by accounting for charge exchange and absorption.

Comparison with Other Generators

• NuWro and NEUT: Both models significantly underestimate the cross-section across most of the kinematic range (by 10–30% at the peak).
• $\chi^2/NDF$: GENIE provides the best overall description of the data (average $\chi^2/NDF \approx 0.77$ for $p_{\pi^0}$), whereas NuWro and NEUT show larger discrepancies (e.g., $\chi^2/NDF \approx 2.04$ and $3.06$ respectively for $p_{\pi^0}$).

Kinematic Features

• The cross-section peaks at a $\pi^0$ momentum of 0.2 GeV/c and an angle of 27$^\circ$.
• The muon momentum distribution shows data lying slightly above the model prediction for $p_\mu > 1$ GeV/c, indicating a preference for a "harder" muon spectrum than predicted.

5. Significance

• Oscillation Physics: These results provide critical constraints on the modeling of $\pi^0$ production, which is the primary background for $\nu_e$ appearance searches. Improved models will reduce systematic uncertainties in oscillation parameter measurements (e.g., $\theta_{23}$ and $\delta_{CP}$).
• Model Tuning: The observed deficit in the $\Delta(1232)$ resonance region highlights the need for further tuning of interaction models, particularly regarding resonance axial masses and FSI parameters in the antineutrino sector.
• Data Availability: The paper releases the full covariance matrices and cross-section data, enabling the broader community to refine neutrino interaction generators and improve the precision of future neutrino experiments.

In conclusion, this paper establishes a new benchmark for antineutrino interaction physics, revealing specific shortcomings in current simulation models regarding resonance production and providing the necessary data to refine them for the next generation of neutrino oscillation experiments.