Constraining Neutrino Interaction Uncertainties for Neutrino Oscillation Measurements at the T2K Experiment

This paper details the methodology and results of constraining neutrino flux and interaction uncertainties using the T2K ND280 near detector to improve systematic control in oscillation measurements, while also validating the robustness of the approach and demonstrating the potential benefits of the upgraded detector for future precision.

The Gist

Imagine the T2K experiment as a massive, high-stakes game of "Where's Waldo?" but instead of finding a person in a crowd, scientists are trying to find specific patterns in how invisible particles called neutrinos change their identity as they travel.

Here is a simple breakdown of what this paper does, using everyday analogies.

1. The Big Picture: The Long Journey

Neutrinos are ghostly particles that rarely interact with anything. In the T2K experiment, a beam of these particles is shot from a facility in Tokai, Japan, all the way to a giant detector called Super-Kamiokande, 295 kilometers (about 183 miles) away.

As they travel, these neutrinos "oscillate," meaning they switch flavors (like a chameleon changing colors). Scientists want to measure exactly how often this happens to understand the fundamental laws of the universe.

2. The Problem: The "Blurry Camera"

To measure this change, scientists need to know two things:

1. What was sent? (The starting number and type of neutrinos).
2. What arrived? (The number and type that made it to the far detector).

The problem is that the "camera" used to see the neutrinos isn't perfect. When a neutrino hits an atom in the detector, it creates a messy explosion of other particles. To figure out how much energy the original neutrino had, scientists have to guess based on the debris.

The Analogy: Imagine trying to guess the speed of a car that crashed into a wall by looking at the scattered pieces of the bumper. If your theory about how bumpers break is slightly wrong, your guess about the car's speed will be wrong, too.

In the past, the biggest source of error in T2K wasn't the number of neutrinos; it was the uncertainty in how they crash into atoms (the "crash theory").

3. The Solution: The "Control Room" (ND280)

To fix this, T2K has a "Control Room" detector called ND280, located just 280 meters from the source. This detector sees the neutrinos before they have a chance to change colors.

This paper is all about upgrading the software and rules used to interpret what happens in this Control Room. The scientists are essentially saying: "Let's look at the crash debris right here, refine our crash theory, and use that to make a much better prediction for what happens 295 kilometers away."

4. What Did They Actually Do? (The Upgrades)

The paper details three major upgrades to their "crash theory" software:

• Better Sorting (New Event Selections):
  Previously, they grouped all the crash debris together. Now, they are using a more detailed sorting system. They are specifically tagging events that have protons (heavy particles) or photons (light particles) in the debris.

  • Analogy: Instead of just counting "car parts," they are now separating "headlights" from "tires" and "engines." This helps them understand exactly how the crash happened.

• A New "Crash Manual" (Interaction Models):
  They updated the theoretical models that predict how neutrinos interact with atomic nuclei. They added new "knobs" and "dials" to the software.

  • Analogy: Imagine the old manual said, "If a car hits a wall, it breaks like this." The new manual says, "Actually, it depends on the car's weight, the wall's material, and the angle. Here are 50 different ways it could break, and we will adjust the manual based on what we actually see."

• Refining the Beam Map (Flux Prediction):
  They improved their map of the neutrino beam itself, using new data from a separate experiment (NA61/SHINE) to better predict how many neutrinos are in the beam and what their energies are.

5. The Results: Does the New Theory Work?

The scientists took their new, complex software and ran it against the actual data collected at the Control Room (ND280).

• The Fit: They adjusted their "knobs" until the software's prediction matched the real data.
• The Outcome: The new model fits the data very well. The "p-value" (a score of how well the theory matches reality) is high (57.5%), meaning the theory is a good description of what's happening.
• The Surprise: When they looked at the "knobs" they turned, they found that the universe behaves slightly differently than their original "best guess" manual suggested. For example, they had to tweak how neutrinos interact with protons inside the nucleus to make the math work.

6. The "Stress Test" (Robustness)

To make sure they didn't just get lucky, they ran a series of "what-if" scenarios. They asked: "What if our theory is totally wrong in a specific way? Would our method still catch the neutrinos correctly?"

They simulated data using completely different, alternative theories of how neutrinos crash. They found that even if the real world worked like one of these alternative theories, their new method would still be able to constrain the errors and give a reliable result for the main experiment.

7. The Bottom Line

This paper doesn't discover a new particle or solve the mystery of the universe's origin. Instead, it does the unglamorous but vital work of calibrating the ruler.

By refining how they measure the neutrino "crashes" at the near detector, they have significantly reduced the "fuzziness" in their measurements. This means that when they look at the data from the far detector (Super-Kamiokande) to measure neutrino oscillations, they can be much more confident that their results are real and not just a mistake in their math.

In short: They built a better map and a sharper lens for the Control Room, ensuring that the long-distance measurements of the neutrinos are as precise as humanly possible.

Technical Summary

Technical Summary: Constraining Neutrino Interaction Uncertainties for Neutrino Oscillation Measurements at the T2K Experiment

Problem Statement
Neutrino oscillation experiments, such as T2K (Tokai-to-Kamioka), rely on precise measurements of neutrino flux and interaction cross-sections to extract oscillation parameters (mixing angles, mass splittings, and the CP-violating phase $\delta_{CP}$). A primary limitation in current and future long-baseline experiments is the systematic uncertainty arising from neutrino-nucleus interaction models. While the near detector (ND280) constrains the product of flux and cross-section, differences in detector acceptance, target materials (carbon vs. oxygen), and neutrino energy spectra between the near and far detectors prevent perfect cancellation of systematic errors. Consequently, uncertainties in the theoretical modeling of neutrino interactions—particularly regarding nuclear effects like spectral functions, final state interactions (FSI), and multi-nucleon processes—have become a dominant source of error, potentially limiting the sensitivity of future experiments like Hyper-Kamiokande and DUNE.

Methodology
This paper details a comprehensive near-detector analysis using data from the ND280 off-axis detector, corresponding to an exposure of $1.97 \times 10^{21}$ protons on target (POT) in neutrino mode and $1.63 \times 10^{21}$ POT in antineutrino mode. The analysis employs a simultaneous fit of a parameterized neutrino flux, interaction model, and detector response to binned data across multiple event selections.

Key methodological updates include:

1. Refined Event Selections: The analysis introduces new event categories based on proton and photon tagging. The $\nu_\mu$ CC0$\pi$ sample is split into zero-proton and proton-tagged subsets, and photon-tagged samples separate resonant $\pi^0$ production from multi-pion events. This increases the number of samples from three to five per Fine-Grained Detector (FGD) in neutrino mode.
2. Updated Flux Model: The flux prediction incorporates 2010 NA61/SHINE hadron-scattering data from a replica of the T2K target, including charged kaon and proton yields. This reduces flux uncertainties by nearly 50% in the 2–7 GeV range. Improvements were also made to the modeling of horn-cooling water interactions.
3. Expanded Interaction Model Parameterization: The baseline model uses the NEUT v5.4.0 event generator. The uncertainty parameterization was significantly expanded to include theory-driven variations:
   • CCQE: Uncertainties on the Spectral Function (SF) model, including Short-Range Correlations (SRC) normalization, shell occupancy, shell momentum shapes, removal energy shifts, Pauli blocking, and optical potential corrections.
   • 2p2h: Normalization and shape uncertainties split by initial state pair types (np, nn) and neutrino/antineutrino modes.
   • Single Pion Production (SPP): Updated tuning of resonant parameters ($M_A^{RES}$, $C_5^A$) using bubble chamber data, plus new uncertainties for resonance decay kinematics and $\pi^0$ normalization.
   • FSI: New parameters for nucleon FSI ("nucleon fate") and expanded pion FSI uncertainties.
4. Fitting Framework: The analysis utilizes two independent frameworks: a hybrid-frequentist approach (using Minuit2) and a Bayesian approach (using Markov Chain Monte Carlo). Both minimize an extended binned negative log-likelihood that includes statistical terms and systematic penalty terms.

Key Results

• Fit Quality: The fit yields a total p-value of 57.5% ($\chi^2 = 5628.82$), indicating the model adequately describes the data, though this is lower than the previous analysis (74%), likely due to the increased sensitivity of the new selections.
• Parameter Shifts: Several interaction parameters were pulled significantly from their priors:
  • The nucleon axial mass ($M_A^{QE}$) was pulled higher, reshaping the CCQE cross-section.
  • Shell normalization parameters for carbon and oxygen shifted, altering the spectral function shape.
  • Pauli blocking and optical potential parameters were pulled to suppress low-energy transfer events.
  • The Bodek-Yang correction for Deep Inelastic Scattering (DIS) was effectively turned off in the fit.
• Uncertainty Reduction: Post-fit uncertainties on event rates were significantly reduced. For example, the total uncertainty on CCQE events dropped from 13.2% (prefit) to 2.0% (postfit), and for 2p2h events from 50.5% to 12.2%.
• Benchmarking: The postfit model was compared against external cross-section measurements:
  • Agreement improved for $\nu_\mu$ CC0$\pi$ differential cross-sections on Carbon and Oxygen targets ($\chi^2$/dof improved from 118.7/58 to 59.9/58).
  • Agreement improved for CC1$\pi^+$ measurements in muon kinematics.
  • However, the postfit model showed worse agreement with CC0$\pi$ measurements involving hadron kinematics (specifically transverse momentum imbalance $\delta p_T$), suggesting the model's predictive power for proton kinematics remains limited despite the fit being driven by lepton kinematics.

Significance and Claims
The paper claims that the updated methodology provides a robust control over systematic uncertainties for the T2K oscillation analysis reported in Ref. [17]. By employing a more refined parameterization of interaction uncertainties and leveraging new event selections (proton and photon tagging), the analysis successfully constrains the flux and cross-section models.

The authors emphasize that while the fit improves agreement with muon kinematics, the residual discrepancies in hadron kinematics (proton momentum and transverse imbalance) highlight the need for further improvements in nuclear modeling. The study demonstrates that the current uncertainty model is sufficient to cover variations for current statistics-limited analyses but identifies specific areas (nuclear ground state models and FSI) where future developments are necessary for the next generation of experiments. The paper concludes that the upgraded ND280, with improved acceptance and lower hadron thresholds, is expected to further enhance these constraints.