Skip to main content

Case Study: Super-Kamiokande and the Discovery of Neutrino Oscillations


In the 1970s theoretical physicists were hot on the trail of what they believed would be a Grand Unified Theory. They hoped that this theory would join three of the four forces of nature together into one theoretical construct. It would provide a unified explanation for the electromagnetic, strong, and weak interactions. This theory, named SU(5) after the mathematical group that it contained, predicted that the proton, until then thought to be an absolutely stable particle, would decay into other particles. The observation of proton decay would therefore be evidence for grand unification, and several experimental groups were formed to attempt this search. Unfortunately, this search would not be an easy one, as the predicted lifetime of the proton (or bound neutron) was 1028–30 years!

Luckily, proton decay is a statistical process. If one were to watch a single proton and wait for it to decay, one would expect to wait a very long time, but if a large number of protons were watched simultaneously, say, 1030 protons, one of these protons would be expected to decay in a year. Thus, the search for nucleon decay requires massive detectors. A search with a sensitivity of 1033 years requires a detector with approximately 1033 nucleons. Since there are 6×1029 nucleons per ton of material, this implies detectors of kiloton scale. Over the past two decades, there have been two types of nucleon decay detectors: water Cherenkov detectors such as IMB, Kamiokande, and Super-Kamiokande; and fine grain sampling detectors such as NUSEX, Frejus, and Soudan. The largest and most sensitive of these detectors is the Super-Kamiokande water Cherenkov detector, which was completed and began taking data in 1996.

Water Cherenkov Detectors

A water Cherenkov detector consists of a large volume of purified water viewed by photomultiplier tubes (very sensitive photocells that convert light into an electrical signal). These massive detectors employ water both as an inexpensive target material and the detection medium. When a charged particle passes through a transparent material with a speed faster than the speed of light in that material, it emits light (Cherenkov light). The light is given off at a constant angle to the particle track. The angle is determined by the particle speed and the index of refraction of the medium. For a charged particle traveling in water at a speed close to that of light in vacuum, this angle is about 42°. In the case of a track that begins and ends in the detector, a ring of Cherenkov light is formed on the detector wall. By measuring the time of arrival of each photon at a photomultiplier tube, and by knowing the angle that the photons made with the particle track, it is possible to fully reconstruct the particle track, including its direction and its beginning (vertex) and ending points.

The Super-Kamiokande detector (see Figure 1) is a 50,000 metric ton (55,000 ton), water Cherenkov detector located at a depth of 1,000 m (3,300 ft) of rock in the Mozumi zinc mine in central Japan. It consists of a cylindrical stainless steel tank, 39 m (128 ft) in diameter and 41 m (135 ft) high, filled with purified water. The detector is optically segmented into an inner volume, 34 m (112 ft) in diameter and 36 m (118 ft) in height, and an outer (anticoincidence) region, 2.5 m (8 ft) thick, on the top, bottom, and sides of the inner volume. The inner detector is viewed by 11,146 photomultiplier tubes, 50 cm (20 in.) in diameter, uniformly distributed on the inner boundary. The total mass of water inside the surface of the inner detector defined by the photomultiplier tubes is 32,000 metric tons (35,000 tons). The fiducial mass, defined to be 2 m (6.5 ft) inside the photomultiplier tube plane, is 22,500 metric tons (25,000 tons).

The outer annulus of the detector is called an anticoincidence region as it is used to tag entering charged particles as well as to attenuate low-energy


gamma rays and neutrons, which cause undesirable background events in the sensitive volume. The outer annulus also complements the measurement of particle energy in the inner detector by measuring the energy loss due to exiting particles. This outer detector region is viewed by 1,885 photomultiplier tubes, 20 cm (8 in.) in diameter, with additional light collection afforded by attached plastic plates. The walls of the anticoincidence region are reflective to enhance light collection. The photomultiplier tubes are mounted facing outward on the same superstructure as the 50 cm (20 in.) photomultiplier tubes of the inner volume. Also, an optical barrier is mounted on the same structure to separate the inner and outer regions.

Results on Proton Decay

Background for nucleon decay arises from interactions between the target material and muons and neutrinos produced by cosmic-ray interactions in the upper atmosphere. By locating the detectors underground, experimenters can reduce cosmic-ray muons to a manageable level, but since neutrinos penetrate large thicknesses of material easily, some neutrino background is unavoidable. The vast majority of atmospheric neutrino interactions bear little resemblance to nucleon decay, but a small fraction are indistinguishable (based on topology and kinematic parameters) from the signal. At present, Super-Kamiokande has found no candidates that can be unambiguously ascribed to baryonnumber-violating nucleon decay, but as described below, the study of the neutrino backgrounds led to a very important discovery.

Grand unified theories continue to predict a broad range of possible proton lifetimes. There is evidence that the fundamental approach to unification


is sound, and nucleon decay is one of the few accessible regimes where grand unified theories can be directly confronted with experimental data. Further progress in the detection of this unique process may be crucial to the future development of particle physics; this strong possibility indicates that the search for evidence of nucleon decay should be pursued with renewed vigor.

Atmospheric Neutrinos: From Background to Signal

The search for nucleon decay naturally yields a sample of atmospheric neutrino interactions fully contained in Super-Kamiokande's inner detector volume. Although these atmospheric neutrinos present a background to some physics searches (e.g., proton decay and extraterrestrial neutrino astronomy), they are a rich source of physics in their own right. They provide, for example, a unique opportunity for the study of the neutrinos themselves. In fact, the story of Super-Kamiokande is a classic example of that old bit of wisdom in experimental science which states: One person's background is another person's signal.

Atmospheric neutrinos are generated in cosmicray interactions in the atmosphere: a cosmic ray (either proton or nucleus) interacts, producing a cascade of secondaries; the secondaries (typically pions and kaons) may either interact again or decay. This cascade continues until the primary particle's energy is dissipated in the atmosphere, primarily by ionization. In such a cascade, the number of positively charged secondaries will be nearly equal to the number of negatively charged secondaries. Thus, if one assumes that all low-energy pions decay and that their muon secondaries also decay, simple counting of the daughters allows the following conclusion: there will be twice as many muon neutrinos as electron neutrinos. These estimates are borne out by sophisticated Monte Carlo calculations of the atmospheric flux. The atmospheric flux covers a broad range of neutrino energies from 10s of MeVs to 10s of TeVs and beyond.

Super-Kamiokande is the largest existing underground detector of atmospheric neutrinos. During its first 1,489 days of operation, it collected about 12,000 events in which a neutrino interacted in the water and the resulting charged particle was fully captured and identified in the detector. This is by far the world's largest sample of atmospheric neutrino interactions and constitutes a valuable resource—the interpretation of these events has given new insight into the question of neutrino oscillations.

In addition, the detector records a vast number of entering (meaning some light is seen in the anti-coincidence volume of the detector) particles, about three each second. Of these, a small fraction, about one every other day, enters from below the horizon. These are of special interest since they are mainly muons produced in muon neutrino interactions in the rock surrounding the detector. Such neutrinos travel large distances, up to 12,000 km between production and interaction. As shall be seen in the next section, the distance neutrinos have traveled since production is of central importance.

Neutrino Oscillations

There are three known types of neutrinos: the electron neutrino, the muon neutrino, and the tau neutrino. Although neutrinos are one of the fundamental constituents of matter, until recently, whether or not they had any mass was unknown. If neutrinos do have mass, then it is possible for one type to spontaneously change into another and then back again as it travels through space. This process is called neutrino oscillations. The parameters describing this process involve the difference in the square of the individual neutrino masses (δm2) and the probability of the oscillation occurring (known as the mixing probability). The mixing probability is written as the square of the sine of twice the mixing angle, sin22 θ . Full mixing would be sin22 θ = 1.0. The mathematics of this process predicts that the oscillation length (or wavelength of the oscillation) should be proportional to the energy of the neutrino divided by δm2. This means that a very small δm2 produces a very long oscillation length.

As previously mentioned, atmospheric neutrinos are produced in the interactions of primary cosmic rays in the atmosphere. They penetrate the Earth from all directions, arriving almost isotropically at the detector. On their way to the detector, they travel distances from about 10 km (when coming directly from above) to about 12,000 km (when they come from below the detector). Such large oscillation distances are not available with other, humanmade neutrino beams in this energy range and for these flavors of neutrinos. The large distances together with the low mean energy of the neutrinos provides an opportunity to study oscillations with δm2 as low as a few times 10-5 eV2.

The flux is composed of roughly equal proportions of electron and muon neutrinos and antineutrinos. Since it is believed that lepton flavor is conserved in near-instantaneous particle interactions, the flavor of an incoming neutrino can be determined by identifying the lepton type produced in the detector. In Super-Kamiokande, this is accomplished by examining in detail the nature of the particle tracks.

Electron tracks are separated from muon tracks by taking advantage of a difference in patterns of illuminated phototubes caused by the difference in the mechanisms of energy deposition by these two classes of particles. An electron track is a showering track that result from the cascade shower of electrons and positrons formed as the electron passes through the water. This showering track produces a fuzzy ring in the detector. A muon, on the other hand, does not shower and produces a sharply defined Cherenkov ring pattern.

The energy spectrum of the incoming atmospheric neutrinos is mostly below the threshold needed for the production of a tau, so any oscillation to tau neutrinos manifests itself predominantly as a loss of muon neutrinos.

The most accurately predicted feature of the atmospheric neutrinos (±5%) and historically the first to be studied is the ratio of the muon neutrino to the electron neutrino flux. If oscillations of any kind take place, this ratio could change. The IMB and Kamiokande detectors did find a significant deficit of muon neutrinos while recording approximately the predicted number of electron neutrinos. Taken at face value, this suggested neutrino oscillations, but, of course, such an interpretation relied on the accuracy of the Monte Carlo simulations of neutrino production in the atmosphere. Super-Kamiokande was in an ideal position to decide this issue. The enormous size of Super-Kamiokande yielded a very large number of atmospheric neutrinos to study. If neutrino oscillations were the reason for the muon neutrino deficit, it would be possible to see a difference in the number of muon neutrinos coming from above the detector, where the path length of the neutrino since production was small, from those coming from beneath the detector, where the path was through the entire Earth. This comparison essentially eliminates reliance on atmospheric simulations since it compares the downward-traveling events with the upward-traveling ones. Conclusive evidence of the oscillation was observed and announced in 1998. A significant, angular dependent deficit of muon-type neutrinos was observed. As the number of electron-type neutrinos showed no such angular dependence, it was clear at this point that the muon-type neutrinos were turning into either tau neutrinos or some previously undiscovered (sterile) form of neutrino. Subsequent studies of Super-Kamiokande's data confirmed that the muon neutrinos were predominantly converting into tau neutrinos, whereas the existence of a sterile neutrino has become increasingly unlikely.

As mentioned earlier, another source of information for oscillation studies is provided by the neutrinos that interact in the rock surrounding the detector. These interactions mainly produce muons that enter the detector from below. The observed rate of these upward-traveling muons is compared with computer simulations based on a variety of models for atmospheric neutrino fluxes, neutrino interactions, and the transport of interaction products through the rock. Once again, the observations are significantly below the predictions and are consistent with the contained neutrino interaction sample interpretation, namely, muon neutrinos turning into tau neutrinos. Since this analysis is dependent on atmospheric simulations, the sensitivity to mixing angle is limited mainly by the 20 percent systematic uncertainty in the flux. In order to overcome this limitation, the ratio of the rate of up-going tracks that stop in the detector and the rate of those that traverse the detector and exit were studied. This ratio is flux-independent. Since the median energy of stopping muons is much smaller than that of the exiting ones, the ratio is sensitive to oscillations, and the same oscillation parameters as with the other two analyses are obtained, thus confirming the oscillation interpretation and hence the existence of massive neutrinos with an entirely different event sample.

Continuing Studies of the Neutrino

Super-Kamiokande's atmospheric and solar neutrino results, along with those of other atmospheric and solar neutrino experiments, have conclusively established the reality of neutrino mass and mixing. This discovery is among the most fundamental in particle physics within the past several decades, requiring the first significant modification in the long history of the Standard Model. The further study of neutrino mass splittings and mixing angles, an entirely new and unexplored sector of physics, has just begun but has already revealed surprises. The mixing angle between the second and third generations (inferred from atmospheric data) is apparently near maximal, whereas solar neutrino data also favor large mixing between the first and second generations. Clearly, these leptonic mixings bear little resemblance to those of the quarks—a new and fundamental revelation. New experiments have begun and others are in the construction and planning stages to explore this new and exciting avenue of study.

See also:Japanese High-Energy Accelerator Research Organization, KEK; Lepton; Neutrino; Neutrino, Discovery of


Aglietta, M., et al. "Experimental Study of Atmospheric Neutrino Flux in the NUSEX Experiment." Europhysics Letters8 , 611–614 (1989).

Agrawal, V., et al. "Atmospheric Neutrino Flux above 1 GeV. Physical Review D53 , 1314–1323 (1996).

Allison, W. W. M., et al. "Measurement of the Atmospheric Neutrino Flavour Composition in Soudan 2. " Physics Letters B391 , 491–500 (1997).

Casper, D., et al. "Measurement of Atmospheric Neutrino Composition with the IMB-3 Detector. " Physical Review Letters66 , 2561–2564 (1991).

Daum, K., et al. "Determination of the Atmospheric Neutrino Spectra with the Frejus Detector. " Zeitschrift für Physik C66 , 417–428 (1995).

Fukuda, Y., et al. "Evidence for Oscillation of Atmospheric Neutrinos." Physical Review Letters81 , 1562–1567(1998).

Fukuda, S., et al. "Tau Neutrinos Favored over Sterile Neutrinos in Atmospheric Muon Neutrino Oscillations. " Physical Review Letters85 , 3999–4003 (2000).

Hirata, K. S., et al. "Experimental Study of the Atmospheric Neutrino Flux." Physics Letters B205 , 416–420 (1988).

Henry W. Sobel

Cite this article
Pick a style below, and copy the text for your bibliography.

  • MLA
  • Chicago
  • APA

"Case Study: Super-Kamiokande and the Discovery of Neutrino Oscillations." Building Blocks of Matter: A Supplement to the Macmillan Encyclopedia of Physics. . 22 Sep. 2018 <>.

"Case Study: Super-Kamiokande and the Discovery of Neutrino Oscillations." Building Blocks of Matter: A Supplement to the Macmillan Encyclopedia of Physics. . (September 22, 2018).

"Case Study: Super-Kamiokande and the Discovery of Neutrino Oscillations." Building Blocks of Matter: A Supplement to the Macmillan Encyclopedia of Physics. . Retrieved September 22, 2018 from

Learn more about citation styles

Citation styles gives you the ability to cite reference entries and articles according to common styles from the Modern Language Association (MLA), The Chicago Manual of Style, and the American Psychological Association (APA).

Within the “Cite this article” tool, pick a style to see how all available information looks when formatted according to that style. Then, copy and paste the text into your bibliography or works cited list.

Because each style has its own formatting nuances that evolve over time and not all information is available for every reference entry or article, cannot guarantee each citation it generates. Therefore, it’s best to use citations as a starting point before checking the style against your school or publication’s requirements and the most-recent information available at these sites:

Modern Language Association

The Chicago Manual of Style

American Psychological Association

  • Most online reference entries and articles do not have page numbers. Therefore, that information is unavailable for most content. However, the date of retrieval is often important. Refer to each style’s convention regarding the best way to format page numbers and retrieval dates.
  • In addition to the MLA, Chicago, and APA styles, your school, university, publication, or institution may have its own requirements for citations. Therefore, be sure to refer to those guidelines when editing your bibliography or works cited list.