Precision measurement of neutrino oscillation parameters with 10 years of data from the NOvA experiment

Using a dataset with double the previous exposure and improved analysis techniques, the NOvA experiment reports its most precise single-experiment constraints on atmospheric neutrino mass splitting and mixing parameters, showing a mild preference for the normal mass ordering that strengthens to 87% probability when combined with Daya Bay data.

The Gist

Imagine the universe is filled with invisible, ghostly particles called neutrinos. These particles are the ultimate escape artists: they have almost no mass, no electric charge, and they can pass through entire planets (like Earth) without bumping into a single atom. They are so elusive that trillions of them are passing through your body right now, and you wouldn't feel a thing.

However, these ghosts have a secret superpower: they can change their identity.

This paper from the NOvA experiment is like a 10-year-long detective story about these shape-shifting ghosts. Here is the story, broken down into simple terms.

The Setup: A High-Speed Ghost Train

Imagine a giant particle accelerator at Fermilab in Illinois. It's like a massive cannon that shoots a beam of "ghost train cars" (neutrinos) toward a detector in Minnesota, 500 miles away.

• The Near Detector (The Train Station): This is located right next to the cannon. It counts the ghosts as they leave the station to see what they look like when they start.
• The Far Detector (The Destination): This is a massive, 14,000-ton box filled with plastic pipes and liquid, sitting in Minnesota. It waits to catch the ghosts after their long journey.

The Mystery: The Shape-Shifting Game

In the world of neutrinos, there are three "flavors" (identities): Muon, Electron, and Tau.

• The beam starts as almost 100% Muon ghosts.
• As they travel 500 miles, some of them magically transform into Electron ghosts.
• Some Muon ghosts disappear entirely (they turn into Tau ghosts, which are hard to see).

The NOvA team spent 10 years watching this happen. They collected data from 2014 to 2024, doubling the amount of data they had before. It's like watching a magic show for a decade to finally figure out the trick.

The Big Discoveries

1. The "Mass Ordering" Puzzle (Who is the Heaviest?)

Neutrinos have three different mass states (let's call them Light, Medium, and Heavy). But scientists didn't know the order.

• Normal Ordering: Light < Medium < Heavy.
• Inverted Ordering: Heavy < Medium < Light.

The Result: The NOvA data is like a scale that is tipping slightly. It says, "We are 87% sure the order is Normal (Light < Medium < Heavy)." It's not a slam-dunk victory yet, but it's the strongest hint we've had so far. Think of it like a jury that is leaning heavily toward "Guilty," but wants just a little more evidence before the final verdict.

2. The "Mixing" Angle (How much do they swap?)

The paper found that the Muon ghosts are swapping identities almost perfectly. It's like a dance where the partners switch places so often that they spend about half their time as one person and half as the other. This "maximal mixing" is a very specific, beautiful symmetry in nature that the data confirms.

3. The "CP Violation" (Why is there more matter than antimatter?)

This is the biggest mystery of the universe: Why does the universe exist? If the Big Bang created equal amounts of matter and antimatter, they should have annihilated each other, leaving nothing but light. But we are here, so something must have tipped the balance.

Neutrinos might hold the key. If they behave differently than "anti-neutrinos" (their evil twins), it could explain the imbalance.

• The Result: The NOvA data is still a bit fuzzy here. It's like trying to hear a whisper in a noisy room. The data suggests there might be a difference, but it's not loud enough to shout "We found it!" yet. They need more data to be sure.

How They Did It (The Detective Work)

To solve this, the scientists didn't just count ghosts; they had to build a perfect simulation of the universe.

• The "Ghost Map": They used supercomputers to predict exactly how the ghosts should behave if their theories were right.
• The "Noise Filter": The detectors are so sensitive that cosmic rays (particles from space) hit them constantly, creating "static" on the line. The team had to use advanced AI and deep learning to filter out the cosmic noise and find the real ghost signals.
• The "Double Check": They compared their results with other experiments (like Daya Bay in China) to make sure they weren't making a mistake. When they combined their data with Daya Bay's, the confidence in the "Normal Ordering" jumped from 70% to 87%.

Why Should You Care?

You might ask, "Why spend billions of dollars and 10 years to study invisible ghosts?"

1. Understanding the Universe: If we figure out the mass of neutrinos, we might finally understand how the universe formed and why we exist.
2. New Physics: The Standard Model (our current rulebook for physics) is incomplete. Neutrinos are the cracks in the wall where new, exciting physics might be hiding.
3. Technology: The tools built to catch these ghosts (massive detectors, AI, super-fast computers) often lead to breakthroughs in medicine, imaging, and computing.

The Bottom Line

The NOvA experiment has taken a giant step forward. They have measured the "atmospheric mass splitting" (a fancy way of saying the difference in weight between the heaviest and lightest neutrino) with 1.5% precision. That is incredibly precise!

They haven't solved the whole mystery yet, but they have handed the next generation of scientists a very strong clue: The universe likely follows the "Normal" mass order, and the ghosts are dancing in a very specific, symmetric way.

The search continues, but with this new data, we are closer than ever to understanding the ghostly rules that govern our reality.

Technical Summary

Here is a detailed technical summary of the paper "Precision measurement of neutrino oscillation parameters with 10 years of data from the NOvA experiment."

1. Problem Statement

Neutrino oscillation physics has established that neutrinos have mass and mix, but several fundamental parameters remain poorly constrained or unknown within the three-flavor paradigm. Key open questions addressed by this work include:

• Neutrino Mass Ordering (MO): Determining whether the third mass eigenstate ($\nu_3$) is heavier (Normal Ordering, NO) or lighter (Inverted Ordering, IO) than the other two.
• The $\theta_{23}$ Octant: Determining if the mixing angle $\theta_{23}$ is less than or greater than $\pi/4$ (maximal mixing).
• CP Violation: Measuring the CP-violating phase $\delta_{CP}$ to understand the matter-antimatter asymmetry of the universe.
• Precision: Achieving sub-percent precision on the atmospheric mass splitting ($|\Delta m^2_{32}|$) to test the consistency of the three-flavor model and resolve degeneracies with reactor experiments.

2. Methodology

Experimental Setup:

• Beam: The experiment utilizes the NuMI (Neutrinos at the Main Injector) beam from Fermilab, producing a high-purity $\nu_\mu$ beam (94% pure in neutrino mode, 93% $\bar{\nu}_\mu$ in antineutrino mode) with a wrong-sign contamination of ~5-6%.
• Detectors: Two functionally identical liquid scintillator detectors are used:
  • Near Detector (ND): Located 1 km from the target to measure the unoscillated beam.
  • Far Detector (FD): Located 810 km away, positioned 14.6 mrad off-axis to select a narrow energy band peaking at 1.8 GeV.
• Dataset: The analysis uses data collected from February 2014 to March 2024, representing a 96% increase in neutrino-mode exposure compared to previous NOvA results (totaling $26.61 \times 10^{20}$protons-on-target in neutrino mode). Antineutrino exposure remains unchanged at$12.50 \times 10^{20}$ POT.

Analysis Techniques & Improvements:

• Event Reconstruction: Uses a combination of classical algorithms and deep learning (Convolutional Neural Networks) for particle identification (PID).
• Energy Estimation:
  • $\nu_\mu$: Sum of muon track length and hadronic calorimetric energy.
  • $\nu_e$: Quadratic function of hadronic and electromagnetic shower energies.
• New Selections:
  • Low-energy Sample: Introduced to cover the 0.5–1.5 GeV range, previously underexplored but sensitive to MO.
  • Peripheral Selection: Includes high-confidence events with vertices outside the fiducial volume to increase efficiency.
• Simulation & Systematics:
  • Flux/Interaction: Uses Geant4 for transport and a custom Genie 3.0.6 configuration (tuned to NOvA ND data) for interactions, including a Valencia MEC model for multi-nucleon effects.
  • New Uncertainties: Added uncertainties for pion production in the resonant-to-DIS transition region, proton/pion inelastic scattering (via Geant4Reweight), and neutron detection (using the Menate model to correct Geant4 underestimation).
• Statistical Framework:
  • Frequentist: Poisson likelihood with Profiled Feldman-Cousins intervals.
  • Bayesian: Markov Chain Monte Carlo (MCMC) to compute posterior probability densities.
  • External Constraints: Results are presented with and without constraints from the Daya Bay reactor experiment on $\sin^2 2\theta_{13}$ (1D) and the joint $\sin^2 2\theta_{13}$–$\Delta m^2_{32}$ (2D) constraints.

3. Key Contributions

• Dataset Expansion: Doubled the neutrino-mode exposure compared to the previous major NOvA publication, significantly reducing statistical uncertainties.
• Refined Systematics: Introduced new systematic uncertainty categories (pion production transition, neutron modeling) and improved detector simulation (Menate model for neutrons, Geant4Reweight for hadrons).
• Novel Event Selections: Implementation of the "Low-energy" sample to probe the 0.5–1.5 GeV region, enhancing sensitivity to the mass ordering.
• Dual Statistical Approach: Provides a comprehensive comparison between Frequentist and Bayesian analyses, ensuring robustness of the results.

4. Key Results

Oscillation Parameters (Conditional on Mass Ordering, with 1D Daya Bay constraint):

• Atmospheric Mass Splitting ($|\Delta m^2_{32}|$):
  • Normal Ordering (NO): $2.431^{+0.036}_{-0.034} \times 10^{-3} \text{ eV}^2$
  • Inverted Ordering (IO): $-2.479^{+0.036}_{-0.036} \times 10^{-3} \text{ eV}^2$
  • Precision: Achieves a fractional uncertainty of 1.5%, the most precise single-experiment measurement to date.
• Mixing Angle ($\sin^2 \theta_{23}$):
  • Preferred value: $0.55^{+0.02}_{-0.06}$.
  • The data favors regions close to maximal mixing ($\sin^2 \theta_{23} = 0.5$) in both orderings, with no significant preference for a specific octant.
• CP Phase ($\delta_{CP}$):
  • Best fit (NO): $0.87^{+0.30}_{-0.90} \pi$.
  • Highest Posterior Density (HPD) for NO: $0.93^{+0.21}_{-0.89} \pi$.
  • The data remains compatible with both CP conservation and CP violation.

Mass Ordering Preference:

• Without External Constraints: NOvA data shows a mild preference for Normal Ordering with a Bayes factor of 2.4 (70% posterior probability).
• With 1D Daya Bay Constraint: Preference strengthens to a Bayes factor of 3.3 (77%).
• With 2D Daya Bay Constraint: Preference significantly strengthens to a Bayes factor of 6.6 (87% posterior probability).
• Frequentist Significance: The Inverted Ordering is disfavored at 1.4$\sigma$ (p-value 0.16) with the 1D constraint, rising to 1.6$\sigma$ (p-value 0.11) with the 2D constraint.

Goodness of Fit:

• Posterior-predictive p-values (PPP) are approximately 0.48 across all samples, indicating excellent agreement between the data and the three-flavor oscillation model.

5. Significance

• Precision Benchmark: This work sets the current standard for the precision of $|\Delta m^2_{32}|$ from a single long-baseline experiment (1.5% uncertainty).
• Complementarity: The high precision of NOvA's $|\Delta m^2_{32}|$ measurement allows for powerful synergy with reactor experiments (Daya Bay). The 2D constraint from Daya Bay breaks degeneracies, pushing the preference for Normal Ordering to 87%.
• Future Outlook: While the preference for Normal Ordering is growing, the results are not yet definitive (requiring $>5\sigma$). The paper highlights that future data from current and upcoming experiments (DUNE, Hyper-Kamiokande) combined with NOvA's precision will be crucial for resolving the mass ordering and discovering CP violation.
• Validation of Three-Flavor Paradigm: The consistency of NOvA results with other experiments (T2K, Super-K, IceCube, MINOS) and reactor data reinforces the validity of the standard three-flavor oscillation framework.