DUNE: The Deep Underground Neutrino Experiment

The Deep Underground Neutrino Experiment — DUNE — is the largest long-baseline neutrino experiment ever built, and the centrepiece of the next generation of oscillation physics. When its four 10-kiloton liquid argon detector modules are fully installed a mile underground in South Dakota, and when Fermilab’s upgraded proton accelerator sends a high-intensity neutrino beam 1300 kilometres through the Earth to meet them, DUNE will spend the 2030s attacking the three central open questions in neutrino physics: the mass ordering, CP violation in the lepton sector, and the ability to read a galactic supernova in unprecedented detail.

Construction began in the mid-2010s. The first detector module reached its target at Sanford Underground Research Facility in 2024 and is being populated with instrumentation. The beam starts around 2029. First physics results are expected by the early 2030s, with full-sensitivity measurements running through the decade. The experiment is the largest and most expensive investment in basic research in US particle physics in a generation.

This post walks through what DUNE actually is, how its design choices map onto the physics it hopes to measure, and why the experiment matters even if — or especially if — its results are exactly what the Standard Model predicts.

Two sites, one experiment

DUNE is a two-site experiment. The neutrino beam is produced at Fermilab in Illinois, where a 120 GeV proton beam from the Main Injector strikes a graphite target, producing a cascade of pions and kaons. A focusing horn selects positive or negative mesons — switchable to run neutrino or antineutrino mode — which then decay in a 194-metre tunnel to produce a collimated beam of muon neutrinos (or antineutrinos) at energies between 0.5 and 5 GeV, peaking near 2.5 GeV. The beam is aimed 5.8 degrees below horizontal, passing through the Earth’s crust and mantle to emerge at the far detector.

The far detector is the Sanford Underground Research Facility (SURF) in Lead, South Dakota, a former gold mine now converted to a physics laboratory. The far detector complex sits at a depth of 1478 metres (the “4850 Level”), where rock overburden reduces the cosmic-ray muon flux by about six orders of magnitude compared to the surface. Four detector modules, each containing 10 kilotons of liquid argon in its fiducial volume, will be installed in excavated caverns the size of major airport terminals.

The baseline — the distance from beam source to far detector — is 1300 km. This is the result of a careful optimisation: it is long enough that matter effects in the Earth meaningfully modulate the oscillation probability (required for mass-ordering determination), but short enough that the beam intensity at the far detector remains observable with practical detector mass.

A near detector at Fermilab, 574 metres from the beam target, characterises the unoscillated beam and provides the spectral reference needed to extract oscillation parameters precisely.

Why liquid argon

Liquid argon time-projection chambers (LArTPCs) are DUNE’s signature technology. In an LArTPC, charged particles traversing the argon ionise the medium, producing free electrons that drift under an applied electric field toward a collection plane of fine wire grids. By recording the arrival time and position of each ionised electron, the detector reconstructs every charged-particle trajectory in three dimensions with millimetre resolution. Scintillation light emitted by the argon simultaneously triggers the event timing and provides calorimetric information.

Compared to water Cherenkov detectors (Super-Kamiokande, Hyper-Kamiokande) and scintillator detectors (JUNO, Borexino), LArTPCs offer:

• Spatial resolution of ~1 mm — sufficient to distinguish electrons from photons by the early ionisation profile
• Full 3D reconstruction of all charged tracks
• Excellent energy resolution (~3–5% at GeV energies) through calorimetric readout
• Particle identification via energy loss and topology
• Sensitivity to low-energy signals (keV-scale) that water Cherenkov cannot see

The downside is complexity: LArTPC construction requires ultra-high-purity argon (sub-ppb contamination), large cryogenic infrastructure, and sensitive electronics that must operate at 87 K. The ProtoDUNE demonstrator at CERN (2018–2024) established the technology at the kiloton scale; the DUNE far detectors scale it by a factor of 40.

The headline physics

Mass ordering. The sign of $\Delta m_{31}^2$ — whether $m_3$ is the heaviest mass eigenstate or the lightest — is not yet known. DUNE determines it by comparing ${\nu }_{\mu }\to {\nu }_e$ appearance rates in neutrino and antineutrino modes. Over the 1300-km baseline, matter effects in the Earth enhance the neutrino oscillation (and suppress the antineutrino) for normal ordering, and vice versa for inverted ordering. The asymmetry reaches tens of per cent at the spectral peak, well above the few-per-cent statistical sensitivity DUNE will achieve in its first few years. The mass-ordering determination is expected at 5σ significance within 3–4 years of beam operation.

CP violation. The Dirac phase ${\delta }_{CP}$ of the PMNS matrix controls whether leptons and antileptons oscillate at the same rate. Current T2K and NOvA data favour ${\delta }_{CP}$ near $-\pi /2$, but at only 2–3σ combined significance. DUNE’s longer baseline and broader spectrum give it sensitivity to the second oscillation maximum, where the CP-violating effect is enhanced relative to the matter effect. The experiment is designed to reach 5σ discovery of CP violation (if $|\sin {\delta }_{CP}|$ is close to 1) within 10 years of running, and to measure ${\delta }_{CP}$ to better than 10° precision.

Precision measurement of other PMNS parameters. DUNE will measure ${\theta }_{23}$ — including its octant (above or below $\pi /4$) — to better than 1% precision. The octant question is sensitive to flavor-symmetry model predictions. It will also improve measurements of $|\Delta m_{31}^2|$ by an order of magnitude over current global-fit uncertainties.

Beyond the beam: supernova, proton decay, and exotica

DUNE is not just a beam-physics experiment. The 40-kiloton far-detector mass, underground location, and sensitivity to low-energy signals make it simultaneously one of the most capable observatories for several non-beam physics programmes.

Galactic supernova detection. A core-collapse supernova anywhere in our galaxy would produce thousands of neutrino events in DUNE, with unique sensitivity to the ${\nu }_e$ channel through ${\nu }_e+{}^{40}\text{Ar}\to e^{-}+{}^{40}{\text{K}}^{\ast }$. This reaction is forbidden by kinematics in water Cherenkov detectors — the ${}^{16}\text{O}$ nuclear matrix element is suppressed. DUNE would measure the neutronization burst (the first 10 ms of core collapse) at unprecedented statistical precision, directly probing the supernova mechanism and testing proto-neutron-star formation dynamics.

Proton decay. Grand-unified theories predict proton decay at half-lives of $10^{33}$ to $10^{36}$ years. The primary signature — a muon or positron plus a kaon or pion — is exactly what LArTPCs can resolve. DUNE’s sensitivity in the $p\to K^{+}\bar{\nu }$ channel is expected to reach $6\times 10^{34}$ years, better than any existing experiment.

Atmospheric neutrinos. The 40-kiloton volume records hundreds of atmospheric neutrino events per year, providing independent mass-ordering constraints and sensitivity to tau-neutrino appearance through matter effects.

Beyond-Standard-Model searches. Sterile neutrinos at the near detector, non-standard interactions, exotic neutrino couplings, dark matter searches — the list of BSM physics sensitivity DUNE will have is long.

Timeline and status

• 2013: Proposal submitted as a successor to LBNE
• 2015–2017: CD-1 and CD-3 approvals; proto-DUNE construction at CERN
• 2017: Long Baseline Neutrino Facility (LBNF) excavation begins at SURF
• 2019: ProtoDUNE results demonstrate LArTPC performance at kiloton scale
• 2022–2024: Cavern excavation at SURF completed; first detector module components delivered
• 2024: First detector module (horizontal-drift design) instrumentation begins
• 2028: Second module (vertical-drift design) installation complete
• 2029: First beam expected; initial physics runs
• 2030–2035: Mass ordering determination, initial CP measurements
• 2035+: Full four-module configuration, CP precision era

Delays are possible: the scale of the underground construction and cryogenics logistics is unprecedented, and the PIP-II accelerator upgrade at Fermilab (required for the beam) is an additional dependency. But the programme has institutional support across US, European, and South American collaborators, and construction is proceeding on a realistic schedule.

Why this matters

By the time DUNE reaches full sensitivity, the Standard Model of particle physics will be 60 years old. The mass ordering, CP violation, and precise PMNS parameters are the last unmeasured quantities in the Standard Model’s minimal three-flavor extension. DUNE is, in this sense, a completion experiment: it will finish measuring the four parameters needed to fully describe neutrino mixing in the Standard Model.

If those measurements turn up no surprises, the Standard Model’s lepton sector will be as quantitatively specified as its quark sector. That is not a negative result — it is the kind of closure that comes only once a generation and is the precondition for whatever comes next.

If, instead, DUNE finds tensions — the wrong mass ordering, an unexpected ${\delta }_{CP}$, inconsistencies with JUNO or Hyper-Kamiokande — then beyond-Standard-Model physics has an experimental anchor. Either outcome is worth the 20-year investment.

DUNE is unique in one practical sense: it is the only currently-funded experiment capable of simultaneously addressing all three open long-baseline questions with single-experiment consistency. Its results will define the 2030s for neutrino physics.