Neutrinos were first detected in 1956 by Fred Reines of the University of California at Irvine and the late George Cowan. They showed that a nucleus undergoing beta decay emits a neutrino with the electron, a discovery that was recognized with the 1995 Nobel Prize for Physics. Ever since the discovery, physicists have wanted to know whether neutrinos have mass.

Direct attempts to determine the mass of the neutrino have been based on the conservation of energy and momentum in, for example, the beta decay of tritium. But results have only provided upper limits on the mass. The new evidence instead depends on a fascinating quantum mechanical process known as neutrino oscillations.

To explain this we first need to know that neutrinos are elementary particles from the family of leptons. All leptons have a spin of 1/2, but some (like the electron) are charged, while neutrinos have no charge. Three types of neutrinos exist, and are distinguished by the way in which they interact. The neutrino emitted in beta decay is now called the electron neutrino, and the others are the muon and tau neutrinos.

The muon neutrino is emitted along with the muon – a charged lepton like the electron but 200 times heavier – in the decay of the pion. The muon neutrino was first detected by Leon Lederman, Jack Steinberger and Mel Schwartz at the Brookhaven National Laboratory in 1962, a discovery for which they shared the Nobel prize in 1988. The tau neutrino has never been detected directly but is known to be emitted in the decay of the tau, the heaviest lepton.

A neutrino oscillation means that one type of neutrino gradually transforms into another as it moves. Indications of such oscillations have been reported before, but the new work provides the first convincing proof of their existence.

The collaboration of 120 US and Japanese scientists measured neutrinos produced in the atmosphere by cosmic rays. SuperKamiokande, a Cerenkov detector containing 50 000 tonnes of ultrapure water and located a kilometre below ground in the Kamioka mine near Takayama, can detect electron and muon neutrinos but not tau neutrinos.

Neutrinos can enter the detector from either above or below – neutrinos from above travel a relatively short distance through the atmosphere, whereas neutrinos from below travel much further through the Earth. Since the cosmic-ray flux is known to be the same from both directions (except for a small geomagnetic effect), and since neutrinos interact so weakly that they penetrate the Earth, it was expected that the same number of neutrinos would enter the detector from both directions. But the SuperKamiokande team found that the number of muon neutrinos entering the detector from below was half the number coming from above. The electron neutrinos, however, were unaffected.

The only explanation for this finding is that muon neutrinos had oscillated into tau neutrinos, which cannot be detected by SuperKamiokande. The distance required for one oscillation must be between 100 and 10 000 km, which means that neutrinos travelling through only the atmosphere would not experience significant oscillations, while neutrinos also travelling through the Earth have a large probability for oscillation.

If neutrinos do have mass, it might be expected that each type of neutrino should have a different mass and a different mass eigenstate. However, nearly all theories of neutrino mass predict that the three types of neutrinos are well defined mixtures of several mass eigenstates. As time passes or the neutrino moves, the relative phases of the different components in this mixture change. Consequently, a state that is originally, say, muon-type, gradually transforms into another type of neutrino. Neutrino oscillations therefore result from the mass of the neutrino.

This quantum mechanical argument is similar to that for an electron orientated at an angle to a magnetic field. In this case the electron is in a mixture of spin-up and spin-down eigenstates, and it precesses around the magnetic field. For the neutrinos, the length of oscillation (analogous to the precession time) depends on the mass difference between the eigenstates. The SuperKamiokande result suggests that both the muon and tau neutrinos are made up of approximately equal mixtures of two mass eigenstates. The mass difference between these eigenstates is such that the squares of the masses differ by 10^-2-10^-3 (eV)².

The discovery of neutrino mass is significant for several reasons. For example, the pioneering work of Raymond Davis of the University of Pennsylvania since the 1960s has prompted many experiments to measure neutrinos produced in nuclear reactions in the Sun. These experiments have consistently yielded a neutrino flux less than half that calculated from solar models. It has been thought for some time that the only simple explanation for this anomaly was in terms of neutrino oscillations that convert electron-type neutrinos to another type.

To relate this to the atmospheric result, there must be a third mass eigenstate, very close to one of the other two, that is a mixture of the electron neutrino with the muon and tau neutrinos. The mass difference needed to explain the solar neutrinos is much smaller than that needed to account for the atmospheric result. These three mass eigenstates can explain the “disappearance” of both solar and atmospheric neutrinos, but they are not directly related.

The SuperKamiokande result could also be important for big bang theory, which predicts that the universe contains a large background of neutrinos. If the neutrino mass was 1 eV, this would suggest that neutrinos account for more mass in the universe than all of the protons and neutrons put together. Neutrinos would therefore represent a significant amount of “dark matter”, mass that we cannot see but is predicted to exist by cosmological models. The simplest interpretation of the solar and atmospheric results is that the heaviest neutrino has a mass of 0.1 eV. However, neutrino oscillations depend only on the differences in mass, so it is possible that all three masses are 1 eV or greater but that the mass differences are much smaller. Thus it is not clear how this new evidence relates to the dark matter problem.

But the biggest prize may be the impact the result could have on the development of grand unified theories, which attempt to explain the weak, electromagnetic and strong interactions in terms of a single interaction. A key component of these theories is the symmetry between quarks and leptons. Indeed, there are fundamental similarities between these particles. For a start, both quarks and leptons have a spin of 1/2 and experience the same weak interactions. Furthermore, both types of particle have three families or generations. The first generation of quarks comprises the up and down quarks, the constituents of protons and neutrons. The second generation consists of the strange and charm quarks and is 10-100 times heavier, while the third and heaviest generation comprises the bottom quark and the recently discovered top quark. Similarly the three families of leptons are the electron, muon and tau, and their associated neutrinos.

In certain grand unified theories this underlying symmetry becomes exact at some high-energy scale. These models often assume that neutrinos acquire mass in the same way as other particles do. A very interesting possibility – postulated in 1977 by Murray Gell-Mann, Pierre Ramond and Dick Slansky – is that the neutrino masses are inversely proportional to the high-energy scale at which the quark-lepton symmetry is broken. If this so-called “see-saw” mechanism is true, this measurement of the neutrino mass may be the first window into new physics at energies far beyond those accessible with particle accelerators.