Water Cherenkov detectors

Neutrinos are very weakly interacting, so in order to acquire a good neutrino data sample, you need both an intense neutrino source, to maximise the number of neutrinos, and a very large detector, to maximise the number of chances for each neutrino to interact


Using water as the detection medium is a way to provide a very large target mass at reasonable cost—all the world's largest neutrino detectors are water Cherenkov experiments.

How it works

Nothing can travel faster than the speed of light in a vacuum. However, when light travels through a transparent medium such as water, its speed is slowed down by the refractive index of the medium: the refractive index of water at room temperature is 1.33, so light in water is travelling at about 3/4 of the speed of light in a vacuum.

Charged particles, however, are not slowed down by the refractive index, so a high-energy charged particle—say, a 500 MeV muon, which is travelling at 0.977c—will be travelling faster than light in water. 

The effect of this is similar to an aeroplane travelling faster than the speed of sound in air: you get an optical "boom" analogous to the aircraft's sonic boom. In effect, the particle outruns its own electric field, as shown in the diagram.

In a time Δt, the particle travels a distance βcΔt and a photon travels a distance cΔt/n, where n is the refractive index of the medium.

This means that the angle between the particle trajectory (the red line) and the trajectory of the coherent wavefront (the cyan arrows) is given by cos θ = 1/βn ≈ 1/n when the particle is relativistic enough that β ≈ 1 (as is usually the case). For water at room temperature, this angle is 41º.

In a water Cherenkov detector, the Cherenkov radiation is detected, usually by photomultiplier tubes (PMTs), and the cone of emission reconstructed. The axis of the cone gives the direction of the particle, and the light yield gives the particle energy. 

Only charged particles with β > 1/n can be detected: this gives a threshold total energy of about 0.8 MeV for electrons, 160 MeV for muons and 1.4 GeV for protons and neutrons (somewhat higher in practice since the amount of Cherenkov radiation emitted exactly at threshold is negligible).

Neutrinos are detected in water Cherenkovs when they interact by W exchange, converting into the equivalent charged lepton (muon or electron for νμ or νe respectively), or when they elastically scatter off electrons (when the recoil electron can be detected). 

Identifying ντ is more difficult, because of the short lifetime of the tau, but has been achieved by Super-K using the fact that tau decays often yield fast pions in the detector.

Advantages and Disadvantages

A big advantage of using this technique is that very large target masses can be instrumented at comparatively modest cost. Super-Kamiokande is a cylindrical water tank 40 m in diameter by 40 m high, containing 50 kilotons of water (of which 22.5 kt is fiducial mass).

The IceCube neutrino telescope instruments a cubic kilometre of the Antarctic ice cap, and the KM3NeT project plans similar-sized deployments in the Mediterranean Sea.

Water Cherenkov neutrino telescopes have a distinguished history: work done using Super-Kamiokande gained Nobel Prizes for Masatoshi Koshiba in 2002 and for Takaaki Kajita in 2015 (the other half of the 2015 prize was also won by a water Cherenkov: the Sudbury Neutrino Observatory in Canada).

Another advantage is that muons and electrons can be separated quite efficiently using ring morphology. A muon will typically produce a clean, sharp-edged ring, whereas an electron will scatter more and will produce a much fuzzier ring. 

This enables a detector like Super-K to tag the flavour of the incoming neutrino, which is critical for νμ → νe oscillations where it is necessary to detect and identify the νe.

The principal disadvantage of the technique is its comparatively high threshold. In particular, protons and neutrons are only detected if their momenta exceed about 1.1 GeV/c, which means that—for example—in T2K, any nucleons ejected from the nucleus with which the neutrino interacted will not be detected, since the T2K neutrino beam energy peaks at about 600 MeV. 

This can make it difficult to classify events observed in Super-K accurately. In addition, the very large size of these detectors makes surrounding them with a magnet impractical, so the charge of the produced lepton is not determined.

Gadolinium loading

The charge identification issue can be addressed, at least in the simple quasi-elastic scattering process where no additional particles are produced, by identifying the final-state nucleon as either a proton (implying the reaction νμ+ n → μ + p, or the equivalent for other flavours) or a neutron (implying ν̅μ+ p → μ+ + n). 

In T2K, the produced proton cannot be detected directly as it is below Cherenkov threshold. However, the produced neutron can in principle be detected by observing the delayed γ radiation emitted when it is captured by a nucleus. 

This technique was used in the original Reines and Cowan discovery experiment, which used cadmium chloride as its neutron target (cadmium has a very large neutron capture cross section).

Cadmium is extremely toxic and thus unpleasant to work with, but gadolinium has a similarly huge neutron capture cross section and gadolinium sulphate is water soluble. Therefore, dissolving Gd2(SO4)3 in the water of a water Cherenkov detector could provide a way to distinguish between neutrino and antineutrino reactions.

The EGADS experiment, a "scale model" of Super-Kamiokande installed in the same mine, has proved that a water Cherenkov experiment can run successfully with 0.2% Gd2(SO4)3 loading (and that it is possible to design a water purification system that will not remove the gadolinium!). 

The next step is ANNIE, which will demonstrate the use of Gd loading in a neutrino beam and measure the neutron yield from neutrino interactions. 

On the basis of the EGADS test, the Super-Kamiokande Collaboration decided in June 2015 to go ahead with Gd2(SO4)3 in Super-K. Our experience with ANNIE should prove very useful in interpreting the data from this new phase of Super-K.