User:Amble/1

The 17-keV neutrino was indicated by the results of beta decay endpoint experiments during the 1980s and 1990s. These results were never consistently replicated, and it is now known that, while neutrinos do have mass, the heaviest neutrino mass can be no more than 0.3 eV. To do: ''The 0.3 eV limit must be for the three known generations only. Can still be a heavy sterile neutrino. What are current limits on mixing angle? Check out arXiv:1204.5379v1 [hep-ph] 18 Apr 2012. Having trouble finding limits for keV-scale sterile neutrino that isn't dark matter.''

The apparent detection initially came from experiments performed by John J. Simpson of the University of Guelph using implanted tritium sources. These were apparently replicated by the same group during the 1980s using a variety of nuclei. The results remained controversial, however, because they were not confirmed by outside groups, and in particular because experiments using a different type of detector (magnetic spectrometers) did not see the signal. Around 1991 there was a resurgence in interest as three groups outside the Guelph group announced detection of a signal consistent with a 17 keV neutrino with ~1% mixing probability. These promising signs were followed by definitive experiments that established within a year or two that no 17 keV neutrino existed.

There was not found to be any particular error on the part of the experimenters, and in fact no single experimental artifact could explain the 17 keV neutrino signal in all of the detector types and nuclear species that were used. Instead, a variety of unrelated effects happened to produce a signal that could mimic the same mixing with a 17 keV neutrino. The experimental techniques and detector studies that arose from the effort to elucidate the nature of the 17 keV neutrino signal led to advances in semiconductor detector technology that have been applied to other areas of particle physics and astrophysics. In addition, the 17-keV neutrino continues to be of great interest for the history and philosophy of science and for the discussion of how to interpret unexpected or apparently conflicting experimental data.

Background
The neutrino was originally predicted to exist by Wolfgang Pauli in order to explain how beta decay could respect conservation of energy. The very light or massless, weakly interacting neutrino could carry away a part of the energy of each decay, with the electron carrying away the rest. This partition of energy explains why the electrons have a continuous spectrum of energy rather than a sharp line.

According to the classic Standard Model of particle physics, the three generations of neutrinos are all massless particles and travel at the speed of light. It has long been suspected that the neutrinos may have a small mass, either through the Dirac mechanism or the Majorana mechanism. This would have many interesting consequences. If the neutrinos have a mass, however small, it is possible for neutrino oscillations to occur. This allows a neutrino to transform from one "flavor" to another as it travels. Neutrino oscillations can explain the deficit of solar neutrinos detected on Earth. The current strong evidence that neutrinos have a mass comes primarily from neutrino oscillations, originally detected by SNO, Kamiokande, and others.

A massive neutrino can have significant consequences for astrophysics and cosmology. Neutrinos can compose part of the non-baryonic dark matter that seeds the formation of structure in the universe and accounts for most of the mass of galaxies and galaxy clusters. If the majority of the dark matter is composed of very light neutrinos, they will behave as hot dark matter that tends to escape easily from gravitational wells during structure formation. This "free-streaming" will smooth out or erase inhomogeneity, tending to prevent the formation of structure. Massive neutrinos can also have consequences for the mechanism of supernova explosions.

During the 1980s and early 1990s it was expected that neutrinos may have a mass in the range 1-10 eV. This would be large enough to serve as a viable dark matter candidate consistent with structure formation; but also small enough to avoid problems with supernova mechanisms. A neutrino with a much larger mass > 1 keV would be problematic for theoretical as well as observational reasons, since there was no straightforward way to accommodate such a large neutrino mass within the Standard Model. By the late 1980s accelerator experiments had limited the electron neutrino mass to no more than 10 eV and the mixing angle with the muon neutrino to no more than 0.3%. This implies that, within the Standard Model, a very massive neutrino mixing significantly in beta decays could only be the tau neutruno.

An additional possibility is that there may be one or more additional neutrino flavors beyond the three types included in the Standard Model. These "sterile neutrinos" would not participate in the weak nuclear force, but might be produced by neutrino oscillation.

Experimental approach
A nonzero mass will modify the shape of the spectrum of electrons from a beta-decaying nucleus. The modification is seen very near the endpoint. The beta spectrum will not reach the total energy (Q value) of the decay, but ends slightly below the Q-value by an amount corresponding to the neutrino mass. Because this gap can be very small, the neutrino mass may be more easily detected as a change in slope or "kink" in the spectrum.

Tritium is often chosen as the beta-emitting source because of its very low Q value of 18.6 keV. For this reason it continues to be used by modern experiments such as KATRIN that continue to search for modifications to the endpoint of the beta decay spectrum in order to measure the neutrino mass. If the electron neutrino had a mass of 17 keV, it would be impossible for a tritium beta decay to produce an electron with more than 1.6 keV of kinetic energy. This would be easily detectable and is therefore easily ruled out by the detection of any tritium decays with more than 1.6 keV in the electron. The scenario of interest during the 1980s and early 1990s is a slightly more complicated one: although the electron neutrino may have an undetectably small mass, a second neutrino species could have a larger mass and a substantial mixing angle. This second, heavy neutrino could be a muon neutrino or tau neutrino, or alternatively an additional sterile neutrino. In a small fraction of the beta decays the initial electron antineutrino would change flavor, showing up as a heavy antineutrino and leaving the electron with a small kinetic energy. The tritium decay experiments could therefore measure the hypothetical mass and mixing angle by detecting an excess of beta radiation with very low energies around 1 keV.

Similar searches were performed with other beta-emitting species such as carbon-14 and sulfur-35. These have a larger Q-value, >100 keV, so that mixing to a 17-keV heavy neutrino would cause an excess of electrons at 17 keV below the endpoint.

Modern experiments such as KATRIN typically search for a deficit of electrons just below the endpoint, in order to detect the small mass of the electron antineutrino itself.

This family of experimental approaches does not require the assumption that the neutrino is a Majorana fermion. This makes it sensitive to scenarios in which some other approaches, such as neutrinoless double beta decay, would be unable to detect even a substantial neutrino mass.

Magnetic spectrometers
The beta decay spectrum had traditionally been measured using magnetic electron spectrometers. In this type of detector, the electron is deflected by a magnetic field. The path of a slow-moving, lower-energy electron will be more strongly deflected than that of a fast-moving, higher-energy electron. The position at which the electron hits the detector therefore gives a measure of its energy.

By the early 1980s magnetic spectrometers had achieved energy resolutions around 50 eV. They had been used to place limits on the mass of the electron neutrino by measuring the position and shape of the beta decay endpoint.

After the initial reports of a 17 keV heavy neutrino, similar searches were also performed using magnetic spectrometers. The results were consistently negative. Experimenters were able to measure the same decays using both solid-state detectors and magnetic spectrometers, but only the solid-state detectors showed signs of a low-energy excess consistent with a 17-keV neutrino. It was suspected that magnetic spectrometers could have nonlinearities in their energy response at low energy, causing them to miss an excess.

Solid-state detectors
The group at Guelph worked with solid-state semiconductor detectors with implanted sources. These were an advance over previous generations of detectors used with external sources because the full energy of each decay would be contained within the crystal, and surface effects can be avoided. The energy carried by an electron from a beta decay is measured through the amount of ionization produced in the crystal.

Early work with Si/Li detectors implanted with tritium was published in 1981. In these experiments the goal was to check limits on the mass of the electron neutrino by studying the position and shape of the endpoint of the beta decay spectrum. The achievable energy resolution was around 300 eV. Although the energy resolution was not as good as that of magnetic spectrometers, the implanted solid state detector offered an attractive alternative because of its ability to capture the full energy of the decay and because the different underlying technology (with a different set of possible systematics) could serve as a useful check on the limits set by spectrometers. These experiments found no evidence for a nonzero neutrino mass, giving limits somewhat less sensitive than those from magnetic spectrometers.

A second application of the same detector type was a heavy-neutrino search. This tests models where one of the other neutrino species has a large mass and a nonzero mixing angle with the electron neutrino. The signature of this mixing would be an excess of betas with energy lower than the total Q-value by an amount corresponding to the neutrino mass. Relative to the light-neutrino endpoint experiment, this heavy-neutrino application required excellent accounting for event rates and detection efficiency, but had a less stringent requirement for energy resolution.

Apparent confirmations
In late 1990 multiple outside groups appeared to confirm the existence of the 17 keV neutrino. The 14th Europhysics Conference on Nuclear Physics in Bratislava featured several independent, apparently positive results. The group at Lawrence Berkeley Lab reported data from a 14C source embedded in a germanium detector, and also preliminary results using a different experimental technique capturing the gamma rays from internal bremsstrahlung in electron capture (IBEC). The Zagreb group at Ruđer Bošković Institute presented its own IBEC data using 71Ge and 55Fe (?) as the source.

Hime and Jelley at Oxford University put out a preprint in early 1991 also providing evidence for the 17 keV neutrino. They used a Si(Li) detector and a 35S source with the electrons collimated into the detector. The results supported a 17 keV neutrino at 8σ.

Timeline

 * 1981: Initial experiments by Simpson on beta decay of tritium give no evidence for neutrino mass or mixing with heavy neutrino species.
 * 1985: Simpson reports evidence for a heavy neutrino with mass 17.1 keV and mixing angle of 3%.
 * 1986: Simpson reanalyzes the work of Ohi with 35S, finding evidence for mixing with a 17-keV neutrino.
 * 1986: Simpson paper "Evidence for a 17-keV neutrino in 3H and 35S spectra" at '86 Massive Neutrinos in Astrophysics and Particle Physics. Morionde.  Edited by O. Fackler and J. Trân Thanh Vân.  Cân't find online.
 * 1989: Simpson and Hime, 35 S sources placed next to windowless Si(Li) detector
 * 1989: Hime and Simpson, tritium implanted into germanium detector instead of Si(Li)
 * 1990: October 22-26, 14th Europhysics Conference on Nuclear Physics in Bratislava. New results reported by Berkeley and Zagreb groups.
 * 1991: Hime and Jelley, 35S source collimated into Si(Li) detector at Oxford. Supports 17 keV neutrino at 8σ.
 * 1991: Berkeley group publishes results from Ge detector with 14C dissolved in the crystal. Positive indication for 17 keV neutrino.
 * 1991: Zlimen in Croatia studying internal bremsstrahlung of 71Ge
 * 1991: Schwarzschild favorable article
 * 1991: July 25-August 1: Joint Lepton-Photon Symposium and Europhysics Conference on High Energy Physics in Geneva. Talks by Simpson and by Morrison.
 * 1993: Schwarzschild unfavorable article