Neutrino Atlas
EN
  1. Learn
  2. How neutrino detectors work

Chapter 2: Sources and detection · Lesson 2.3

How neutrino detectors work

Confirmed

A neutrino detector never sees the neutrino itself, only the charged particles, atoms or heat it leaves behind in the rare event of an interaction. Because those events are so rare, every detector is a compromise between huge target mass, sensitive readout and extreme protection against everything else.

Cherenkov light from a charged particle in waterA charged particle faster than light in water emits a cone of light at about 41 degrees, which appears as a ring on the photomultiplier wall.θ ≈ 41°charged particle (β ≈ 1)Water, n = 1.33Photomultiplier wallRing on the wallcos θ = 1 / (n β)
How a water Cherenkov detector sees a neutrino. The neutrino itself is invisible; it produces a charged particle (here a muon or electron) that moves faster than light travels in water (n = 1.33). The particle emits a cone of light at cos θ = 1/(nβ), about 41° for β ≈ 1. Where the cone meets the wall, the photomultipliers record a ring (right). Its sharpness tells muons from electrons, its timing gives the direction.
Inverse beta decay and the delayed coincidenceAn electron antineutrino converts a proton into a positron and a neutron. The positron gives a prompt flash, the captured neutron a delayed one.ν̄e from a reactorpe⁺Prompt: positron annihilates (2 × 511 keV)nDelayed: neutron captureν̄e + p → e⁺ + nflash 1flash 2Δt ≈ 30 µs (Gd) to 200 µs (H)time
Inverse beta decay, ν̄e + p → e⁺ + n, the reaction behind Cowan and Reines’ discovery and today’s reactor experiments (threshold 1.8 MeV). The positron annihilates at once (prompt signal); the neutron wanders and is captured microseconds later, emitting gamma rays (delayed signal) – about 8 MeV on gadolinium, 2.2 MeV on hydrogen. Demanding both signals in a short time window suppresses background by orders of magnitude. Schematic.
Four ways to detect neutrinosFour schematic detectors: a water Cherenkov tank with a light cone, a liquid-scintillator sphere with isotropic light, a liquid-argon time projection chamber with drifting electrons, and strings of light sensors in ice.Water Cherenkovcone of light → ring; direction, flavourLiquid scintillatorbright, isotropic light; low thresholdsLiquid-argon TPCionisation drifts to wires; 3D imagesIce / water telescopesensor strings in km³; high energieselectric field → drift
Water Cherenkov detectors such as Super-Kamiokande see the cone of light from fast charged particles as a ring on the wall. Liquid scintillators such as Borexino, KamLAND and JUNO glow much more brightly and can see lower energies, but the light carries no direction. Liquid-argon time projection chambers such as ICARUS, MicroBooNE and DUNE record the ionisation tracks in three dimensions like a digital bubble chamber. Neutrino telescopes such as IceCube and KM3NeT instrument a cubic kilometre of natural ice or sea water. Schematic.

Cherenkov calculator →

Beginner

Catching something that almost never stops

Trillions of neutrinos from the Sun pass through your body every second, and almost none of them leave a trace. A neutrino feels only the weak interaction, a force with such a short range that a neutrino could cross a light-year of lead and still have a good chance of coming out the other side. To catch a few, physicists build very large detectors and wait. The more material a neutrino has to cross, the more likely it is that one of them, by pure chance, hits an atom.

When that happens, the neutrino passes some of its energy to a particle in the detector, usually an electron or a piece of an atomic nucleus. The detector cannot see the neutrino, but it can see this struck particle. Every detector design is a different trick for making that tiny event visible.

Counting atoms

The first successful solar neutrino detector, at the Homestake mine in South Dakota, used a tank of 615 tonnes of a cleaning fluid rich in chlorine. Occasionally a neutrino turned a chlorine atom into an argon atom. Every few weeks Raymond Davis Jr. flushed the tank and counted the argon atoms, typically only a dozen or so1. Detectors of this kind are called radiochemical. They count how many neutrinos arrived, but not when or from which direction.

Flashes of blue light

Most modern detectors look for light. If a neutrino knocks an electron to very high speed in water, the electron travels faster than light does in water. It then emits a faint blue cone of light, much as a jet faster than sound creates a sonic boom. This glow is called Cherenkov radiation. The Japanese detector Super-Kamiokande is a tank holding 50,000 tonnes of ultrapure water, lined with more than 11,000 light sensors2. The ring of light on the walls reveals the direction and energy of the particle. IceCube uses the same effect in a cubic kilometre of Antarctic ice5.

Other detectors fill their tanks with a special oil, a liquid scintillator, that glows whenever a charged particle passes through it. It produces far more light than water, so it can detect lower-energy neutrinos, though it shows the direction less clearly.

Electric eyes

The light from a neutrino event is extremely dim, sometimes only a few dozen particles of light. It is recorded by photomultiplier tubes, glass bulbs that turn a single light particle into a measurable electrical pulse. Large detectors use thousands of them, each the size of a beach ball in the largest designs.

Other tricks

  • Heavy water, in which ordinary hydrogen is replaced by deuterium, let the SNO detector in Canada count all three kinds of neutrino at once3.
  • Tanks of liquid argon act like a giant digital camera: the particle tracks leave trails of free electrons that are pulled to fine wire grids and read out as a 3D image6.
  • Ultra-cold crystals, called bolometers, measure the tiny rise in temperature caused by a single nuclear event7.

Why underground

The biggest problem is not catching neutrinos but ignoring everything else. Cosmic rays rain down on the Earth's surface all the time, and natural radioactivity in rock, air and even in the detector materials produces signals that can look like neutrinos. Physicists therefore build detectors deep underground, in mines or under mountains, where a kilometre or more of rock filters out cosmic rays. They also surround the detector with water or other shielding and select extremely clean materials. A neutrino detector is as much an exercise in silence as in sensitivity.

Quiz

Test your knowledge

4 questions on this level · Print as worksheet

  1. 1What did Raymond Davis Jr. count in the Homestake tank?
  2. 2What produces Cherenkov radiation in a water detector?
  3. 3What is the advantage of a liquid scintillator over water?
  4. 4Why are neutrino detectors built deep underground?

Intermediate

Three ways to register an interaction

A neutrino interacts through charged-current reactions, which create a charged lepton, or neutral-current reactions, which only transfer energy and momentum. The cross-section is tiny: around 10⁻⁴⁴ cm² for MeV neutrinos on electrons, rising roughly linearly with energy to about 10⁻³⁸ cm² per nucleon at 1 GeV. Detectors therefore need large masses, from tonnes to gigatonnes. They register the secondary particles through three basic signals: light (Cherenkov or scintillation), ionisation charge, or heat and phonons. Radiochemical detectors are a fourth, indirect category.

Radiochemical detectors

Bruno Pontecorvo proposed in 1946 to detect neutrinos via ³⁷Cl + νₑ → ³⁷Ar + e⁻, threshold 0.814 MeV. Raymond Davis Jr. built this at Homestake with 615 tonnes of perchloroethylene, 1480 m underground, and extracted the radioactive ³⁷Ar (half-life 35 days) by helium purging1. The gallium experiments SAGE and GALLEX/GNO used ⁷¹Ga + νₑ → ⁷¹Ge + e⁻, threshold 0.233 MeV, low enough for the dominant pp neutrinos from the Sun. Radiochemical detectors integrate over weeks and give no energy, time or direction for single events.

Water and ice Cherenkov detectors

A charged particle faster than light in the medium, v > c/n, emits Cherenkov light on a cone with

cos θ = 1 / (n β)θ: Cherenkov angle; n: refractive index (1.33 for water); β = v/c. For relativistic particles in water θ ≈ 42°.

Photosensors on the walls record the ring pattern. Its timing gives the vertex, its orientation the direction, its total light the energy, and its sharpness separates muons (sharp rings) from electrons (fuzzy rings caused by showers). Kamiokande (1983) and IMB pioneered the method; Super-Kamiokande (1996) holds 50 kilotonnes of water viewed by 11,146 photomultipliers of 50 cm diameter, 1000 m below Mount Ikeno2. Since 2020 gadolinium dissolved in the water captures neutrons and tags inverse beta decay. IceCube instruments a cubic kilometre of Antarctic ice with 5160 optical modules on 86 strings at depths of 1450 to 2450 m5.

Liquid scintillator

Organic scintillators emit around 10⁴ photons per MeV deposited, roughly fifty times more than Cherenkov light in water. This enables thresholds well below 1 MeV and good energy resolution, but the light is isotropic, so direction information is largely lost. KamLAND (1000 t) detected reactor antineutrinos via inverse beta decay, using the delayed coincidence between the positron flash and the neutron capture about 200 µs later. Borexino (about 280 t) at Gran Sasso measured individual solar neutrino components by elastic scattering on electrons4; JUNO in China uses 20 kilotonnes.

Heavy water

SNO placed 1000 tonnes of D₂O in an acrylic sphere 2 km underground near Sudbury, viewed by 9456 photomultipliers3. Deuterium allowed three reactions: charged current (only νₑ), neutral current ν + d → p + n + ν (all flavours equally), and elastic scattering. Comparing them proved solar neutrino flavour change in 2002.

Liquid-argon TPCs, germanium and bolometers

A liquid-argon time projection chamber, proposed by Carlo Rubbia in 19778, drifts ionisation electrons in an electric field to wire planes, producing millimetre-scale 3D images of every track. ICARUS, MicroBooNE and the future DUNE far detector, four modules of about 17 kilotonnes each, use this method6. High-purity germanium detectors and cryogenic bolometers achieve excellent energy resolution and are the main tools for neutrinoless double beta decay searches; CUORE operates 988 tellurium dioxide crystals near 10 mK7.

Backgrounds and the move underground

At the surface roughly one cosmic muon crosses each square centimetre per minute. Laboratories such as Gran Sasso, Kamioka and SNOLAB lie under 1 to 2 km of rock and reduce this flux by factors of about a million or more. The remaining backgrounds come from uranium, thorium and potassium in rock and materials, radon gas, and isotopes activated by cosmic rays. Experiments fight them with water shields, muon vetoes, radiopure materials and a fiducial volume that excludes the outer layers of the target.

Quiz

Test your knowledge

4 questions on this level · Print as worksheet

  1. 1What is the typical cross-section for MeV neutrinos on electrons?
  2. 2Why could the gallium experiments see solar pp neutrinos while chlorine could not?
  3. 3How does a water Cherenkov detector tell muons from electrons?
  4. 4Which reactions on deuterium let SNO compare neutrino flavours?

Expert

Design drivers

Every neutrino detector balances four quantities: target mass M, which sets the event rate R = Φ σ N_target ε; energy threshold and resolution; event topology information (direction, particle identification, vertex); and the background rate in the signal region. Cross-sections span many orders of magnitude, from σ(νₑe) ≈ 10⁻⁴⁴ cm² at 10 MeV through σCC/E ≈ 0.7 × 10⁻³⁸ cm²/GeV for νμ on isoscalar nucleons in the deep-inelastic regime, to the coherent enhancement of CEvNS, which scales roughly with the square of the neutron number but deposits only keV-scale nuclear recoils. No technology is optimal across this range.

Radiochemical counting

Chlorine and gallium experiments measure an integrated capture rate in SNU (10⁻³⁶ captures per target atom per second). Homestake exposed 615 t of C₂Cl₄ (2.2 × 10³⁰ ³⁷Cl atoms) and recovered on the order of ten ³⁷Ar atoms per extraction, counted by their 2.82 keV Auger electrons in miniature proportional counters1. Extraction efficiency was verified with carrier gas and, for gallium, with ⁵¹Cr and ³⁷Ar sources; those calibrations later produced the gallium anomaly.

Cherenkov detectors

The Frank–Tamm yield in water is about 340 photons per cm between 300 and 600 nm for β ≈ 1, roughly 1 % of a scintillator's light. Super-Kamiokande's 11,146 inner photomultipliers give about 40 % photocathode coverage and roughly 6 photoelectrons per MeV, sufficient for a solar threshold of a few MeV, angular resolution for electron direction of order 25° at 10 MeV, and e/μ separation by ring sharpness at better than 99 %2. Gadolinium sulphate, added in 2020 at 0.01 % and later raised to about 0.03 % Gd, gives neutron captures with an 8 MeV gamma cascade. In ice, IceCube's 5160 digital optical modules sample photons scattered over tens of metres; reconstruction relies on a depth-dependent ice model and distinguishes tracks from cascades5.

Scintillators and the delayed coincidence

Linear alkylbenzene or pseudocumene with PPO as a fluor yields ~10⁴ photons/MeV, reduced at high ionisation density by quenching (Birks' law). Borexino reached about 500 photoelectrons per MeV and radiopurity of the order of 10⁻¹⁹ g/g in ²³⁸U and ²³²Th, which made spectroscopy of ⁷Be, pep, pp and CNO neutrinos possible4. Residual ¹⁴C (¹⁴C/¹²C ~ 10⁻¹⁸) sets a hard floor near 0.2 MeV. Inverse beta decay, ν̄ₑ + p → e⁺ + n with threshold 1.806 MeV, provides a nearly background-free delayed coincidence: a prompt positron signal followed by a 2.2 MeV capture gamma on hydrogen (τ ≈ 200 µs) or an ~8 MeV cascade on gadolinium (τ ≈ 30 µs). Pulse-shape discrimination separates α from β/γ events.

Heavy water

SNO's 1000 t of D₂O offered νₑ + d → p + p + e⁻ (CC), νx + d → p + n + νx (NC, threshold 2.22 MeV) and νx + e⁻ → νx + e⁻ (ES)3. Neutron detection evolved over the three phases: capture on deuterium, on ³⁵Cl after adding 2 t of NaCl, and finally in ³He proportional counters, each with different systematic uncertainties.

Liquid-argon TPCs

Liquid argon (87 K, 1.39 g/cm³) gives ~4 × 10⁴ ionisation electrons and a similar number of 128 nm scintillation photons per MeV before recombination. At 500 V/cm electrons drift at about 1.6 mm/µs, so metre-scale drift lengths require oxygen-equivalent impurities below 0.1 ppb. Wire or pixel readout yields bubble-chamber-quality images with dE/dx for electron–photon separation, crucial for νₑ appearance. Atmospheric argon contains ³⁹Ar at about 1 Bq/kg, tolerable for GeV physics but limiting for low-energy searches. DUNE's far detector at SURF will comprise four modules of roughly 17 kt each, 1.5 km underground and 1300 km from Fermilab6.

Germanium and bolometers

For ⁷⁶Ge (Qββ = 2039 keV) enriched germanium detectors reach FWHM resolutions near 0.1 %, and GERDA demonstrated background indices around 10⁻³ counts/(keV·kg·yr). A bolometer measures ΔT = E/C(T); because the Debye heat capacity scales as T³, temperatures of ~10 mK make single MeV events visible. CUORE operates 988 TeO₂ crystals (742 kg) at about 10 mK and set T½(¹³⁰Te, 0νββ) > 3.5 × 10²⁵ yr with over 2 t·yr of exposure (2025)7. CUPID will add scintillating crystals for α rejection.

Photosensors

Photomultipliers offer gains of ~10⁷, single-photon timing of ~1–3 ns and quantum efficiencies of 25–30 %. Large-area tubes dominate cost-driven designs; JUNO combines about 17,600 tubes of 50 cm with 25,600 tubes of 7.6 cm. Silicon photomultipliers and wavelength shifters are used in liquid argon and xenon.

Backgrounds

  • Cosmic muons: about 1 cm⁻² min⁻¹ at sea level; reduced by roughly 10⁶ at Gran Sasso (3800 m w.e.) and by more than 10⁷ at SNOLAB (6010 m w.e.); see underground depth.
  • Cosmogenic activation: ¹¹C in scintillator, ⁹Li and ⁸He that mimic inverse beta decay, ⁶⁸Ge and ⁶⁰Co in germanium and copper exposed at the surface.
  • Radioactivity: ²³⁸U and ²³²Th chains, ⁴⁰K, ²²²Rn and its daughters ²¹⁰Pb, ²¹⁰Bi, ²¹⁰Po; ⁸⁵Kr and ³⁹Ar in noble liquids.
  • Neutrons from (α,n) reactions and fission in rock, shielded by water, polyethylene or self-shielding of the target.
  • Irreducible neutrino backgrounds: atmospheric neutrinos for proton decay, solar neutrinos for dark-matter searches (the neutrino fog).

Quiz

Test your knowledge

4 questions on this level · Print as worksheet

  1. 1How were the ³⁷Ar atoms extracted at Homestake counted?
  2. 2About how many photoelectrons per MeV does Super-Kamiokande collect?
  3. 3Why do bolometers operate near 10 mK?
  4. 4What sets a hard energy floor near 0.2 MeV in organic scintillators like Borexino?

Open questions

Next-generation detectors push in two directions at once: ever larger targets for oscillation and astrophysics (Hyper-Kamiokande, DUNE, JUNO, IceCube-Gen2) and ever lower backgrounds and thresholds for double beta decay and coherent scattering. Hybrid Cherenkov–scintillation readout and isotope-loaded scintillators are active areas of development.

Terms to explore

People

On the timeline

  1. 1946Pontecorvo proposes the chlorine-argon detector
  2. 1967Germanium detector searches for 0νββ
  3. 1977Rubbia proposes the liquid-argon TPC
  4. 1983Large water Cherenkov detectors begin operation
  5. 1985Chen proposes a heavy-water solar detector
  6. 1996Super-Kamiokande begins operation
  7. 2010IceCube is completed at the South Pole
  8. 2020Gadolinium is added to Super-Kamiokande

References

  1. R. Davis Jr., D. S. Harmer, K. C. Hoffman, Search for neutrinos from the Sun, Phys. Rev. Lett. 20, 1205 (1968), doi:10.1103/PhysRevLett.20.1205
  2. S. Fukuda et al. (Super-Kamiokande Collaboration), The Super-Kamiokande detector, Nucl. Instrum. Meth. A 501, 418 (2003), doi:10.1016/S0168-9002(03)00425-X
  3. J. Boger et al. (SNO Collaboration), The Sudbury Neutrino Observatory, Nucl. Instrum. Meth. A 449, 172 (2000), doi:10.1016/S0168-9002(99)01469-2
  4. G. Alimonti et al. (Borexino Collaboration), The Borexino detector at the Laboratori Nazionali del Gran Sasso, Nucl. Instrum. Meth. A 600, 568 (2009), doi:10.1016/j.nima.2008.11.076
  5. M. G. Aartsen et al. (IceCube Collaboration), The IceCube Neutrino Observatory: instrumentation and online systems, JINST 12, P03012 (2017), doi:10.1088/1748-0221/12/03/P03012
  6. B. Abi et al. (DUNE Collaboration), Deep Underground Neutrino Experiment (DUNE) Far Detector Technical Design Report, Volume I, JINST 15, T08008 (2020), doi:10.1088/1748-0221/15/08/T08008
  7. CUORE Collaboration, Science (2025), doi:10.1126/science.adp6474; earlier: Nature 604, 53 (2022), doi:10.1126/science.adp6474, doi:10.1038/s41586-022-04497-4
  8. C. Rubbia, The liquid-argon time projection chamber: a new concept for neutrino detectors, CERN-EP-INT-77-8 (1977), cds.cern.ch/record/117852

Last updated: · Editorial standard

Cite this page

Text

Neutrino Atlas (2026). “How neutrino detectors work”. Retrieved 22 November 2026, from https://neutrino-atlas.com/learn/how-neutrino-detectors-work/

BibTeX

@misc{neutrinoatlas2026howneutrinodetectorswork,
  title        = {How neutrino detectors work},
  author       = {{Neutrino Atlas}},
  year         = {2026},
  url          = {https://neutrino-atlas.com/learn/how-neutrino-detectors-work/},
  note         = {Accessed 2026-11-22},
  language     = {english}
}