JUNO Reactor Neutrino Oscillation Experiment

• JUNO is a large-scale reactor antineutrino oscillation experiment that uses a massive liquid scintillator detector to capture fine oscillation spectra.
• It achieves sub-percent precision on oscillation parameters and over 3σ sensitivity to neutrino mass ordering through advanced PMT and calibration technologies.
• The experiment’s versatile design supports studies of solar, supernova, atmospheric neutrinos, and searches for exotic processes and non-standard interactions.

The Jiangmen Underground Neutrino Observatory (JUNO) is a large-scale, medium-baseline reactor antineutrino oscillation experiment located in Guangdong Province, China. Its central scientific objectives are the determination of the neutrino mass ordering (also called the neutrino mass hierarchy) via vacuum-dominated oscillatory interference and sub-percent precision measurements of the solar and atmospheric neutrino oscillation parameters. JUNO's strategy relies on detecting electron antineutrinos from two major reactor complexes at a distance of approximately 52.5 km, leveraging a massive liquid scintillator (LS) detector with unprecedented energy resolution to resolve the fine structure of the oscillation spectrum. The experiment additionally functions as a versatile observatory for solar, supernova, atmospheric, and geo-neutrino studies, as well as searches for rare and exotic processes.

1. Experimental Configuration, Detector, and Site

JUNO is sited in an underground laboratory, beneath approximately 650–700 m of rock overburden (≈ 1800 m.w.e.), which suppresses cosmic-ray muon backgrounds to manageable levels. The central detector comprises a 20 kton volume of linear alkylbenzene–based LS enclosed within a 35.4 m acrylic sphere. This volume is instrumented with 17 612 20-inch photomultiplier tubes (PMTs) and 25 600 3-inch PMTs, yielding total photocathode coverage exceeding 75% and a light yield of ≳1200 p.e./MeV (Cerrone, 25 Mar 2024, Petitjean et al., 2021). The PMT system employs both conventional dynode and novel microchannel-plate (MCP) technologies, with quantum efficiencies reaching ~35%. The energy resolution target is σ_E/E ≤ 3%/√E(MeV), achieved through high light yield, PMT density, and optimized LS optical properties (Zhan, 2015, He, 2014, Li, 2014).

Surrounding the central sphere, a 43.5 m-diameter ultra-pure water pool equipped with 2400 PMTs provides an active muon veto via the water-Cherenkov effect, further assisted by a plastic-scintillator top tracker. The system achieves cosmic muon rejection efficiency >99% and supports <0.5% residual cosmogenic isotope background rates (He, 2014, Li, 2014, Petitjean et al., 2021).

The experiment relies on antineutrinos emitted from the Yangjiang (6 × 2.9 GW_th) and Taishan (2 × 4.6 GW_th) nuclear reactor complexes, located at baselines of 52–53 km from the detector. In aggregate, these sources provide 26.6–36 GW_th (project phase-dependent), resulting in ~16,000 IBD events per year (Stock, 12 May 2024, Abusleme et al., 18 Nov 2025, Li, 2014). The baseline configuration maximizes sensitivity to both the “solar” Δm²{21} and “atmospheric” Δm²{31,32} frequency modes by placing the detector at the first solar oscillation maximum (Δ_{21} ≈ π/2) (Cerrone, 25 Mar 2024).

2. Oscillation Formalism and Mass Ordering Sensitivity

The survival probability for reactor electron antineutrinos in the three-flavor framework is given in vacuum as:

$$P_{ee}(E) = 1 - \cos^4\theta_{13}\sin^2 2\theta_{12}\sin^2\Delta_{21} - \sin^2 2\theta_{13}[\cos^2\theta_{12}\sin^2\Delta_{31} + \sin^2\theta_{12}\sin^2\Delta_{32}]$$

where $$\Delta_{ij} \equiv 1.267\,\Delta m^2_{ij}[\mathrm{eV}^2]\,L[\mathrm{km}]/E[\mathrm{MeV}]$$ (Cerrone, 25 Mar 2024, Abusleme et al., 18 Nov 2025, He, 2013, Li, 2014). At JUNO's baseline and the reactor energy range (1.8–8 MeV), both the slow “solar” (Δm²{21} ~ 7.5×10⁻⁵ eV²) and fast “atmospheric” (|Δm²{31,32}| ~ 2.5×10⁻³ eV²) oscillation modes are manifest and give rise to an interference pattern whose fine structure encodes the neutrino mass ordering.

The sign of Δm²{31} (normal ordering, NO, versus inverted ordering, IO) results in an energy-dependent shift of the high-frequency oscillatory phase relative to the low-frequency envelope; this “beat” effect produces subtle spectrum distortions at the percent level around 2–6 MeV, resolvable only with high statistics and σ_E/E ≤ 3%/√E. The mass ordering sensitivity is conventionally quantified via Δχ² ≡ χ²_min(wrong ordering) – χ²_min(true ordering). JUNO, after ~6 years, is projected to achieve a Δχ² ≈ 9–11, corresponding to ≳ 3σ significance, from the reactor spectrum alone; inclusion of an external 1% prior on |Δm²{μμ}| (from accelerator experiments) can raise this to Δχ² > 16 (Cerrone, 25 Mar 2024, Collaboration et al., 28 May 2024, Stock, 12 May 2024, Li, 2014, He, 2014, Zhan, 2015).

3. Precision Oscillation Parameter Measurement

JUNO is designed for sub-percent determination of the key oscillation parameters governing the lepton flavor mixing matrix (PMNS):

Parameter	Projected Precision (6 yr)	Reference
$\sin^2\theta_{12}$	0.4–0.5%	(Stock, 12 May 2024, Cerrone, 25 Mar 2024, Collaboration et al., 2022)
$\Delta m^2_{21}$	0.3–0.6%	(Stock, 12 May 2024, Cerrone, 25 Mar 2024, Collaboration et al., 2022)
$|\Delta m^2_{31}|$	0.2–0.3%	(Cerrone, 25 Mar 2024, Collaboration et al., 2022, Li, 2014)

Recent first data (59.1 days, August–November 2025) have already improved the world-precision on $\sin^2\theta_{12}$ and $\Delta m^2_{21}$ by a factor of 1.6, with values $\sin^2 \theta_{12} = 0.3092\,\pm\,0.0087$ and $$Δm^2_{21} = (7.50\,\pm\,0.12)\times10^{-5}\,\mathrm{eV}^2$$, confirming the design and analysis methodology (Abusleme et al., 18 Nov 2025). Full six-year exposure will anchor global fits and enable tests of PMNS unitarity at the ∼0.5% level (Collaboration et al., 2022, Cerrone, 25 Mar 2024).

4. Detector Technologies, Calibration, and Systematics Control

JUNO's performance is predicated on advanced detector technologies and rigorous control of systematics:

• Photosensors: High-quantum efficiency 20″ MCP-PMTs and conventional dynode PMTs, with overall ≥75% optical coverage, provide timing ≤3 ns and charge resolution ~30–35% (1 p.e.); the 3″ PMT array affords dynamic range and redundancy (Petitjean et al., 2021, Zhan, 2015).
• Liquid Scintillator: Purification via vacuum/molecular distillation and alumina columns achieves attenuation lengths >20–30 m at 430 nm. The optimized LS cocktail (2.5 g/L PPO, bis–MSB) ensures light yield ≳10,000 photons/MeV (Li, 2014, He, 2014).
• Calibration: Extensive calibration—involving radial/axial deployment of radioactive sources (γ, β, neutron), laser and UV-LED fiber systems, and abundance of in situ physics handles (e.g., spallation neutrons, Bi–Po cascades, α decays)—controls absolute energy scale and nonlinearity to sub-percent (<1%) levels across the fiducial volume (Stock, 12 May 2024, Petitjean et al., 2021, Collaboration et al., 2022).
• Systematic Mitigation: Reactor flux normalization and shape uncertainties are addressed using the dedicated TAO near detector (see below); energy-scale nonlinearity is self-calibrated using multiple oscillation peaks and calibration anchors (Forero et al., 2017, Capozzi et al., 2020). Muon-induced $\nu$ backgrounds (e.g., $^9$Li, $^8$He), fast neutrons, geoneutrinos, and accidental coincidences are suppressed by optimized selection and the muon veto (Stock, 12 May 2024, Li, 2014).

5. TAO Satellite Detector and the Role of External Inputs

The Taishan Antineutrino Observatory (TAO) is an integral near detector positioned ∼30–44 m from a Taishan reactor core. With a 2.8 t Gd-doped LS target, 95% SiPM optical coverage, and energy resolution ≈1.5%/√E, TAO registers ≳2000 IBD events/day, providing a direct measurement of the reactor antineutrino spectrum at sub-percent statistical and systematic precision (Steiger, 2022, Capozzi et al., 2020). This reference spectrum allows JUNO to correct for fine-structure "sawtooth" features and residual shape systematics in the unoscillated spectrum, boosting sensitivity to mass ordering and reducing flux-induced uncertainties in parameter measurements (Forero et al., 2017, Steiger, 2022).

Joint analyses of JUNO and TAO further mitigate the impact of micro-structure of reactor spectra (arising from ∼10³ β-decay branches), suppressing their influence on the oscillation fit by reducing the residual model uncertainty in each 50 keV bin to <0.2% (Capozzi et al., 2020, Forero et al., 2017).

6. Sensitivity to New Physics and Extended Program

JUNO's spectral and topological event reconstruction capabilities yield not only precision standard-oscillation results but also sensitivity to diverse new-physics effects:

• Quantum Decoherence and Wavepacket Effects: JUNO's baseline and resolution allow sensitivity to quantum decoherence or wavepacket separation signaling loss of oscillation amplitude, testing σ_x (wavepacket spatial width) down to ~10⁻¹² m and differentiating decoherence from standard oscillations even for σ_x ≳ current experimental bounds (Marzec et al., 2022, Gouvêa et al., 2020, Collaboration et al., 2021).
• Exotic Damping: JUNO can distinguish exponential damping signatures (from, e.g., invisible ν₃ decay, non-standard interactions, absorption, or wavepacket effects) by their differential impact on oscillation harmonics, probing parameters well beyond existing limits (Collaboration et al., 2021).
• Non-Standard Interactions (NSI): Even in the presence of scalar NSI (e.g., a nonzero η{ee} in the propagation Hamiltonian), JUNO's unique sensitivity ensures robust extraction of Δm^2{21} and θ_{12} to sub-percent precision; NSI-induced degeneracies with standard parameters are surmountable only with global analyses combining reactor, solar, and accelerator data (Gupta et al., 2023).
• Multipurpose Physics: The low backgrounds and large mass enable secondary science including detection of supernova burst neutrinos (~5000 events for a galactic SN), diffuse supernova background, solar neutrinos (⁷Be, ⁸B, pep, CNO chain), atmospheric and geo-neutrinos, and rare decays and exotic (e.g., sterile or dark-sector) searches (Stock, 12 May 2024, Li, 2016).

7. Chronology, Status, and Outlook

JUNO was approved in 2013, with excavation and civil engineering initiated shortly thereafter (Li, 2014, He, 2014). Detector and electronics integration progressed through the 2020s, culminating in full assembly, filling, and commissioning by late 2024 (Stock, 12 May 2024). The first data run began in August 2025, with initial results achieving world-leading precision on solar oscillation parameters after just 59.1 live days (Abusleme et al., 18 Nov 2025). The six-year physics run is projected to achieve the primary goals—3σ or better mass ordering sensitivity and sub-percent mixing parameter measurements—by 2031, with expanded multipurpose programs ongoing (Stock, 12 May 2024, Cerrone, 25 Mar 2024).

The combination of massive LS, 3%/√E resolution, strict control of reactor and detector systematics, and the presence of a near detector (TAO) positions JUNO as a flagship experiment for medium-baseline oscillometry and precision lepton sector studies, complementing and informing next-generation accelerator and atmospheric neutrino projects worldwide (Kudenko, 2017, Cerrone, 25 Mar 2024, Li, 2014).