⚛️ high-energy experiments

Measurement of Neutrino Oscillation Parameters at JUNO and Indication of the Neutrino Mass Ordering

Based on 207.2 live days of reactor neutrino data, the JUNO experiment achieved world-leading 1% precision in measuring atmospheric mass splitting and improved solar oscillation parameters, providing a $2.1\sigma$ indication favoring the normal neutrino mass ordering.

Original authors: The JUNO Collaboration

Published 2026-09-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: The JUNO Collaboration

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Neutrinos are the most abundant massive particles in the universe, yet they are also the most elusive. Trillions of them pass through your body every second, ghost-like and nearly massless, rarely interacting with the matter around them. For decades, scientists believed these particles were weightless, but a series of discoveries revealed a profound truth: neutrinos have mass, and they change their identity as they travel. This phenomenon, known as oscillation, occurs because the three types of neutrinos—electron, muon, and tau—are not distinct, solid objects but rather mixtures of three different mass states. As a neutrino moves through space, these mass states interfere with one another, causing the particle to shift back and forth between its three flavors. While physicists have mapped out much of this behavior, a fundamental question has remained unanswered: which of the three mass states is the heaviest? The answer determines the "mass ordering" of the neutrino family, a detail that could reshape our understanding of the universe's evolution and the fundamental laws of physics.

To solve this puzzle, a massive experiment called JUNO, located deep underground in China, has begun to listen to the faint whispers of neutrinos. The facility sits beneath a hill in Guangdong, shielded by hundreds of meters of rock to block out cosmic radiation. At its heart lies a giant sphere filled with 20,000 tons of liquid scintillator, a special oil that flashes with light when a neutrino interacts with it. This sphere is surrounded by thousands of ultra-sensitive cameras, or photomultiplier tubes, designed to catch even the faintest glimmer of light. The experiment focuses on antineutrinos, the antimatter counterparts of neutrinos, which are constantly being produced by two nearby nuclear power plants. As these particles travel roughly 52.5 kilometers to the detector, they oscillate in a complex pattern that depends on their mass differences. By measuring the energy of the arriving particles with extreme precision, scientists can decode the interference pattern and determine the ordering of the masses.

In a recent analysis, the JUNO collaboration reported a major step forward in this quest, using data collected over 207 days. The team identified more than 8,000 candidate events where an antineutrino had struck a proton in the detector, creating a positron and a neutron. This specific interaction, known as inverse beta decay, provides a clear signature that allows researchers to distinguish the signal from background noise. The analysis was rigorous, relying on a "blinding" strategy where the scientists calibrated their equipment and tested their methods without looking at the final results to ensure no unconscious bias influenced the outcome. They also incorporated data from a smaller, high-resolution detector called TAO, located right next to one of the reactors, to get a precise picture of the neutrino beam before it began its journey. This combination allowed the team to measure the energy of the neutrinos with a precision of about one percent, a level of accuracy that had never been achieved before for this type of measurement.

The results of this study provide the most precise measurements to date for two key parameters that govern how neutrinos mix and change. The team determined the values for the solar oscillation parameters with unprecedented accuracy, narrowing down the uncertainty to just a few percent. More importantly, they achieved a world-leading precision of approximately one percent for the atmospheric mass splitting, a value that describes the difference between the heaviest and the lightest neutrino masses. While the data alone showed multiple possible solutions due to the complexity of the signal, the researchers combined their findings with existing measurements from other experiments to isolate the correct physical answer. This synthesis of data allowed them to pinpoint the mass splitting values for both possible scenarios: normal ordering, where the third mass state is the heaviest, and inverted ordering, where it is the lightest.

When the researchers compared these new, highly precise measurements against data from long-distance accelerator experiments, a clear preference emerged. The combined evidence favored the normal ordering scenario over the inverted one. The statistical significance of this preference reached a level of 2.1 standard deviations, meaning there is a roughly two percent chance that the result is a fluke. While this does not yet constitute a definitive discovery, which typically requires a much higher level of certainty, it represents the first time a JUNO-based analysis has pointed so strongly toward the normal ordering. The team also performed a Bayesian analysis, a different statistical approach that yielded a factor of over ten in favor of the normal ordering. These findings mark a critical milestone for the experiment, demonstrating that JUNO is capable of the extreme precision required to solve the mass ordering mystery. As the detector continues to collect data over the coming years, the statistical power will grow, promising to turn this indication into a definitive measurement that will finally reveal the hidden hierarchy of the neutrino family.

Technical Summary: Measurement of Neutrino Oscillation Parameters at JUNO and Indication of the Neutrino Mass Ordering

Problem and Context
The discovery of neutrino oscillations confirmed that neutrinos possess non-zero masses and that flavor eigenstates are mixtures of mass eigenstates. Within the standard three-flavor framework, the remaining unknowns include the octant of the mixing angle $\theta_{23}$ , the CP-violating phase $\delta_{CP}$ , and the neutrino mass ordering (NMO). The NMO determines whether the mass spectrum follows a normal ordering (NO, $\Delta m^2_{31} > 0$ ) or an inverted ordering (IO, $\Delta m^2_{31} < 0$ ). While long-baseline accelerator and atmospheric neutrino experiments probe the NMO through matter-induced effects, the Jiangmen Underground Neutrino Observatory (JUNO) aims to resolve this through reactor antineutrino ( $\bar{\nu}_e$ ) disappearance. This approach exploits the energy-dependent imprint of vacuum-dominated oscillations, offering sensitivity independent of $\delta_{CP}$ and $\theta_{23}$ .

Methodology
This analysis is based on 207.2 live days of data collected by JUNO between August 30, 2025, and May 18, 2026, representing a significant extension of the initial 59.1-day dataset. The detector, located beneath Dashi Hill in Guangdong, China, utilizes a 20-kiloton liquid scintillator central detector instrumented with approximately 43,000 photomultiplier tubes.

Event Selection and Calibration: Inverse beta decay (IBD) candidates ( $\bar{\nu}_e + p \to e^+ + n$ ) were selected using a refined set of criteria, including an expanded fiducial volume ( $R < 17.2$ m) and improved muon vetoing. A comprehensive calibration program using deployed sources (e.g., $^{68}$ Ge, $^{137}$ Cs, Am-C) and natural sources ( $^{214}$ Po) established the energy scale and non-linearity models. The energy resolution at the detector center was measured to be approximately 3.5% ( $\sigma/E$ ) for 68Ge $\gamma$ -rays.
Background Suppression: Dominant backgrounds, particularly cosmogenic $^9$ Li/ $^8$ He, were suppressed via spallation neutron and muon-track vetoes. Accidental backgrounds were reduced by a factor of $\sim$ 40 using a likelihood-ratio cut based on spatio-temporal distributions. The geoneutrino flux was estimated by leaving the rate unconstrained in the fit while constraining the $^{232}$ Th-to- $^{238}$ U ratio.
Analysis Framework: A joint frequentist binned- $\chi^2$ fit was performed on the JUNO data, 90,280 IBD candidates from the Taishan Antineutrino Observatory (TAO), and Daya Bay near-detector data. The fit utilized the Combined Neyman–Pearson (CNP) prescription for statistical variances. To resolve the multiple degenerate local minima in $\Delta m^2_{31}$ caused by the limited dataset, the analysis incorporated the Daya Bay measurement of $|\Delta m^2_{31}|$ as a constraint.
Mass Ordering Determination: To test the mass ordering, constraints on $\Delta m^2_{31}$ from long-baseline (LBL) accelerator experiments (NuFit) were added as a pull term. Both frequentist and Bayesian analyses were conducted to evaluate the preference for NO over IO.

Key Contributions

First JUNO Measurement of $\Delta m^2_{31}$ : The paper reports the first measurement of the atmospheric mass splitting using JUNO data, achieving a world-leading precision of approximately 1%.
Improved Solar Parameters: The analysis provides the world's most precise measurements of the solar oscillation parameters, $\sin^2 \theta_{12}$ and $\Delta m^2_{21}$ .
Integration of TAO Data: This is the first oscillation analysis to incorporate reactor antineutrino spectrum data from the TAO satellite detector, which serves as a near-detector reference for the unoscillated spectrum.
Geoneutrino Flux Measurement: The dataset constitutes the world's largest geoneutrino candidate sample to date, providing the first measurement of the geoneutrino flux at the JUNO site.

Results

Oscillation Parameters:
- Atmospheric Splitting: Under the Normal Ordering (NO) hypothesis, $\Delta m^2_{31} = +2.509^{+0.027}_{-0.025} \times 10^{-3} \text{ eV}^2$ . Under the Inverted Ordering (IO) hypothesis, $\Delta m^2_{31} = -2.482^{+0.025}_{-0.026} \times 10^{-3} \text{ eV}^2$ .
- Solar Parameters: $\sin^2 \theta_{12} = 0.3036 \pm 0.0064$ (2.1% relative precision) and $\Delta m^2_{21} = (7.388 \pm 0.078) \times 10^{-5} \text{ eV}^2$ (1.1% relative precision).
Mass Ordering Indication: The combined analysis of JUNO data with Daya Bay and LBL constraints favors Normal Ordering over Inverted Ordering.
- Frequentist: The preference for NO yields a $\Delta \chi^2 = 4.86$ , corresponding to a one-sided significance of $2.1\sigma$ (or $2.3\sigma$ for the two-sided convention) for rejecting IO. This significance holds across all possible values of $\delta_{CP}$ .
- Bayesian: A Markov Chain Monte Carlo analysis yields a Bayes factor of 10.18 in favor of NO.
Backgrounds: The best-fit geoneutrino rate is $2.12 \pm 0.32$ counts per day, corresponding to an efficiency-corrected flux of $(73 \pm 11)$ TNU.

Significance
The paper establishes JUNO as a precision neutrino experiment, marking an important milestone toward a high-significance determination of the neutrino mass ordering. The precision measurement of $\theta_{12}$ provides a stringent test of lepton-flavor mixing patterns and may help reveal underlying flavor symmetries. Furthermore, the precise determination of the two neutrino mass-squared differences offers tight constraints on new-physics scenarios beyond the standard three-flavor oscillation paradigm. While the current dataset results in multiple local minima for $\Delta m^2_{31}$ , the integration of external constraints allows JUNO to identify the physical minimum and provide the first indication of the mass ordering within a JUNO-based analysis. The results also demonstrate the feasibility of using reactor antineutrino data to probe the mass ordering independently of matter effects and CP violation.

More like this