Hyper-Kamiokande: Scaling the Water Cherenkov Concept

Super-Kamiokande has been operating in the Kamioka mine in central Japan since 1996. It is arguably the most productive single detector in the history of neutrino physics: first evidence of atmospheric neutrino oscillations (1998), precision solar neutrino measurements, supernova burst sensitivity, proton decay searches, and three decades of continuously refined systematics. Its fiducial mass of 22.5 kilotons of ultrapure water has held world-leading status for water Cherenkov detectors since its commissioning.

Hyper-Kamiokande is what comes after. Construction at the Tochibora mine, 1 kilometre underground and 8 kilometres from the original Super-K site, began in 2020. The new detector is an order of magnitude larger than its predecessor — 258 kilotons of fiducial water, 71 metres tall, 68 metres in diameter. It will operate alongside an upgraded beam from J-PARC and a suite of near detectors, targeting the same physics frontier as DUNE but through fundamentally different detector technology and baseline.

This article walks through the design choices, the physics goals, and the specific complementarity with DUNE that motivates running two massive experiments on the same problem at the same time.

Scaling water Cherenkov

Water Cherenkov detection works on a simple principle: charged particles travelling through water faster than light in water produce a conical shockwave of photons (Cherenkov radiation). Arrays of photomultipliers surrounding the water volume detect the arrival times and positions of these photons, reconstructing the particle trajectory and energy.

The technique has several virtues. Water is cheap, chemically stable, and radioactively pure to extremely low levels with modest purification. Photomultipliers are mature technology. Energy reconstruction is calorimetric and well-understood. The sensitivity spans from a few MeV (solar and supernova neutrinos) up to tens of GeV (atmospheric and accelerator neutrinos).

The main trade-offs: Cherenkov reconstruction produces a ring of light rather than a full track, so spatial resolution is coarser than LArTPCs (DUNE’s technology). Event types are distinguished primarily by ring topology — sharp-edged rings for muons, fuzzier rings for electrons, multiple rings for neutral currents with photon conversion. Particle identification above GeV energies becomes less reliable. These limitations are exactly what the LArTPC approach addresses, and why both technologies are being deployed in parallel.

Hyper-Kamiokande addresses the water Cherenkov’s statistical limits by scaling volume. A 258-kiloton fiducial mass is 11.5 times the Super-K mass, producing 11.5 times the event rate for any given beam or atmospheric neutrino source. Combined with a more intense J-PARC beam, the statistical precision on ${\delta }_{CP}$ reaches the 5σ discovery level within 10 years of operation.

The detector

Physical parameters:

• Cylindrical tank: 68 m diameter × 71 m height
• Total mass: 260 kt (ultrapure water)
• Fiducial mass: 258 kt (inner region, excluding edge cuts)
• Inner detector PMTs: 20,000 × 50-cm Box-and-Line PMTs for scintillation readout, covering 20% of inner surface
• Outer detector veto: 3,200 × 20-cm PMTs in a 60-kt outer annular region
• Depth: 1,750 metres water-equivalent overburden at the Tochibora site
• Photomultiplier innovation: The Box-and-Line design offers roughly 2× the photon-counting efficiency of the original Super-K HPK R3600 tubes, improving energy resolution in the MeV-to-GeV range

The detector uses the same proven Super-K liquid-filtration and purification systems, scaled up, so contamination levels can match or exceed Super-K’s radiopurity. The mine site was selected in part because its geology permits the required cavern size without excessive rock-support interventions.

The beam

The J-PARC proton accelerator at Tokai on the Pacific coast feeds a proton beam onto a graphite target, producing pions that decay into muon neutrinos. The beam is aimed 2.5° off the line to the Kamioka mine — this “off-axis” geometry narrows the energy spectrum, giving a peaked beam near 0.6 GeV rather than a broad spectrum.

For Hyper-Kamiokande operation, J-PARC’s beam power is being upgraded from 500 kW (T2K-era) to 1.3 MW by the mid-2020s. The T2K near-detector complex is being upgraded in parallel (ND280 upgrade, IWCD — Intermediate Water Cherenkov Detector). The combination of more beam, more target mass, and better near-detector constraints gives Hyper-K substantially improved sensitivity to both CP violation and oscillation-parameter precision.

Physics goals

CP violation. The primary physics driver. Hyper-K’s statistical power dominates T2K by more than an order of magnitude, reaching 5σ sensitivity on CP violation for about 60% of possible ${\delta }_{CP}$ values within 10 years of running. For values near the current T2K-preferred region of ${\delta }_{CP}\sim -\pi /2$, discovery could come as early as 2030.

Mass ordering. Through atmospheric neutrino measurements, which the detector records continuously alongside beam neutrinos. Hyper-K’s atmospheric measurement reaches 3-4σ mass-ordering sensitivity within its first few years — less definitive than DUNE’s stronger matter-effect-driven beam measurement, but independent and earlier.

Precision oscillation parameters. ${\theta }_{23}$, including its octant determination, to better than 1% precision. Measurement of the atmospheric mass splitting $|\Delta m_{31}^2|$ to ~0.1% precision.

Supernova neutrinos. A galactic core-collapse supernova at 10 kpc would produce roughly 75,000 events in Hyper-K, versus ~7,500 in Super-K and ~3,000 in DUNE. The statistics would allow precision measurement of the neutrino spectrum versus time, revealing the dynamics of core bounce, shock stalling, and eventual shock revival.

Proton decay. Hyper-K’s mass and depth extend the proton decay sensitivity beyond existing limits. For the channel $p\to e^{+}{\pi }^0$, the sensitivity reaches $10^{35}$ years — pushing well into the range predicted by minimal SU(5) grand unification.

Solar neutrino upturn. Hyper-K’s improved photomultipliers and mass should resolve the low-energy upturn of the solar neutrino spectrum below 5 MeV — a key test of the MSW matter-oscillation framework that Super-K could not resolve at the statistical precision needed.

Complementarity with DUNE

The two experiments address the same physics with complementary strengths. A simplified comparison:

Dimension	Hyper-Kamiokande	DUNE
Target mass (fiducial)	258 kt	40 kt (4 × 10 kt modules)
Technology	Water Cherenkov	Liquid argon TPC
Baseline	295 km	1300 km
Beam	Narrow-band, 0.6 GeV peak	Broad-band, 0.5-5 GeV
Matter effects	Modest (< 10%)	Strong (30%+)
Mass ordering sensitivity	Weaker, from atmospheric	Stronger, from beam
CP violation sensitivity	Strong (high statistics)	Strong (matter-CP separation)
Supernova ν_e channel	Limited	Excellent (⁴⁰Ar target)
Supernova ν̄_e channel	Excellent	Moderate
Solar neutrinos	Sensitive to ⁸B and upturn	Requires upgrade
Proton decay e⁺π⁰	Best channel	Excellent in K⁺ modes
First data	2027	2029

A combined analysis of HK + DUNE data in the 2030s will reach sensitivities neither experiment achieves alone. If the two experiments disagree on oscillation parameters at some level, that tension itself will be a signal of beyond-Standard-Model physics.

Timeline

• 2010-2015: Proposal and design phase
• 2018: Approval by Japanese government
• 2020: Construction begins at Tochibora
• 2023: Main cavern excavation completed
• 2024-2026: Tank construction and PMT installation
• 2025: T2K-II and J-PARC beam upgrade complete
• 2027: First physics data expected
• 2030-2035: CP-violation discovery era (combined with DUNE)
• 2035+: Mature precision era

Japan’s investment in Hyper-Kamiokande — approximately 64 billion yen ($600M USD) for the detector alone — represents the largest single-experiment investment in Japanese physics since Super-Kamiokande itself. The international collaboration includes over 400 researchers from 21 countries, with significant contributions from the UK, Canada, Korea, and continental Europe.

Why build both?

The natural question is whether investing in two parallel experiments is duplicative. The community’s answer is that the combination is more than the sum of its parts, for three reasons.

Cross-calibration. The two experiments measure the same physics with different technologies and systematics. Agreement between them is a strong consistency check; disagreement would signal either beyond-Standard-Model physics or a previously unknown systematic in one (or both) of the measurement chains.

Complementary sensitivities. DUNE’s matter effects distinguish ordering cleanly; Hyper-K’s statistics measure ${\delta }_{CP}$ precisely. Combined, they break the degeneracies that hamper either experiment alone.

Redundancy against catastrophic uncertainty. If an unknown cross-section or flux uncertainty skews DUNE’s results, Hyper-K provides an independent check, and vice versa. In high-precision physics, having two world-leading experiments on the same problem is not extravagance but insurance.

The 2030s will see the lepton-sector CP violation question resolved at 5σ level through these two programmes. Whether the answer is the CP phase near $-\pi /2$ favoured by current data, or something unexpected, the measurement will anchor the neutrino Standard Model for the remainder of the 21st century.