Over the 45 years since their discovery, neutrinos have changed from being a physics oddity into one of experimental physics’ most powerful tools. Here, pioneers John Bahcalland Raymond Davisrelate the evolution of the study of extra-terrestrial neutrinos, and provide some stimulating pointers for astronomers and physicists embarking on new observational ventures.
The possibility of observing solar neutrinos began to be discussed seriously following Holmgren and Johnston’s experimental discovery in 1958 that the cross-section for the production of beryllium-7 by the fusion of helium-3 and helium-4 was more than a thousand times as large as had been previously believed. This led to Willy Fowler and Al Cameron suggesting that boron-8 might be produced in the Sun in sufficient quantities (from beryllium-7 and protons) to produce an observable flux of high-energy neutrinos from boron-8 beta decay.
Looking inside the Sun
We begin our story in 1964, when we published back-to-back papers in Physical Review Letters, arguing that it was possible to build a 100 000 gallon detector of perchloroethylene that would measure the solar neutrino capture rate on chlorine. Our motivation was to use neutrinos to look into the interior of the Sun and thereby test directly the theory of stellar evolution and nuclear energy generation in stars. The particular development that made us realize that the experiment could be done was the demonstration by John Bahcall, in late 1963, that the principal neutrino absorption cross-section on chlorine was 20 times as large as had been previously calculated, owing to a super-allowed nuclear transition to an excited state of argon.
Proposals, then and now
If you have a good idea today, you are likely to require many committees, many years and many people to get the project from concept to observation. The situation was very different in 1964.
As Ray Davis was a member of the Brookhaven chemistry department, we presented our case to Dick Dodson, who was
chairman of the Brookhaven chemistry department, and to laboratory director Maurice Goldhaber. Dodson was excited about the possibility of supporting a fundamental new direction within the chemistry department. Goldhaber, on the other hand, was sceptical about all astrophysical calculations, but was intrigued by the nuclear physics of the neutrino analogue transition. Following only a few weeks of consideration, the project received the required backing from Brookhaven, and Dodson and Davis visited the Atomic Energy Commission (AEC) to inform the people in the chemistry division of the plans to begin a solar neutrino experiment. The way was paved by Charlie Lauritsen and Fowler, who had strong scientific and personal connections with the AEC as a result of their wartime work. The project received a warm welcome at the AEC.
A small team, comprising Davis, Don Harmer (on leave from Georgia Tech) and John Galvin (a technician who worked part-time on the experiment), designed and built the experiment. Kenneth Hoffman, a young engineer, provided expert advice on technical questions. The money came from Brookhaven’s chemistry budget. Neither of us remember a formal proposal ever being written to a funding agency. The total capital expenditure to excavate the cavity in the Homestake Gold Mine in South Dakota, build the tank and purchase the liquid was $0.6 million (in 1965).
Solar neutrino experiments
During 1964-1967, Fred Reines and his group worked on three solar neutrino experiments in which recoil electrons produced by neutrino interactions would be detected by observing the associated light in an organic scintillator. Two of the experiments, which exploited the elastic scattering of neutrinos by electrons, were actually performed and led to a higher than predicted upper limit on the boron-8 solar neutrino flux. The third, which was planned to detect neutrinos absorbed by lithium-7, was abandoned after the initial chlorine results showed that the solar neutrino flux was low.
These experiments introduced the technology of organic scintillators into the arena of solar neutrino research, a technique that will only finally be used in 2001 when the BOREXINO detector begins to detect low-energy solar neutrinos. Also during this period, Bahcall investigated the properties of neutrino-electron scattering and showed that the forward peaking from boron-8 neutrinos is large – a feature that was incorporated 25 years later in the Kamiokande (and later SuperKamiokande) water Cherenkov detectors.
The first results from the chlorine experiment were published in Physical Review Lettersin 1968, again in a back-to-back comparison between measurements and standard predictions. The initial results have been remarkably robust; the conflict between chlorine measurements and standard solar model predictions has lasted over three decades.
The main improvement has been in the slow reduction of the uncertainties in both the experiment and the theory. The efficiency of the Homestake chlorine experiment was tested by recovering carrier solutions, by producing argon-37 in the tank with neutron sources and by recovering chlorine-36 inserted in a tank of perchloroethylene. The solar model was verified by comparison with precise helioseismological measurements.
For more than 20 years the best estimates for the observational result and for the theoretical prediction have remained essentially constant. The discrepancy between the standard solar model prediction and the chlorine observation became widely known as “the solar neutrino problem”.
Very few people worked on solar neutrinos during 1968-1988. The chlorine experiment was the only solar neutrino experiment to provide data in these two decades. It is not easy for us to explain why this was the case; we certainly tried hard to interest others in doing different experiments and we gave many joint presentations. Each of us had one principal collaborator during this long period – Bruce Cleveland (experimental) and Roger Ulrich (solar models).
A large effort to develop a chlorine experiment in the Soviet Union was led by George Zatsepin, but it was delayed by the difficulties of creating a suitable underground site for the detector. Eventually the effort was converted into a successful gallium detector, SAGE, led by Vladimir Gavrin and Tom Bowles, which gave its first results in 1990.
Only one year after the first (1968) chlorine results were published, Vladimir Gribov and Bruno Pontecorvo proposed that the explanation of the solar neutrino problem was that neutrinos oscillated between the state in which they were created and a state that was more difficult to detect. This explanation, which is the consensus view today, was widely disbelieved by nearly all of the particle physicists whom we talked to in those days.
In the form in which solar neutrino oscillations were originally proposed by Gribov and Pontecorvo, the process required that the mixing angles between neutrino states should be much larger than the quark mixing angles, something that most theoretical physicists believed, at that time, was unlikely. Ironically, a flood of particle theory papers explained, more or less “naturally”, the large neutrino mixing angle that was decisively demonstrated 30 years later in the SuperKamiokande atmospheric neutrino experiment.
One of the most crucial events for early solar neutrino research occurred in 1968 while we were relaxing after a swim at the CalTech pool. Gordon Garmire (now a principal scientist with the Chandra X-ray satellite) came up to Davis, introduced himself and said that he had heard about the chlorine experiment. He suggested that it might be possible to reduce significantly the background by using pulse risetime discrimination, a technique used for proportional counters in space experiments. The desired fast-rising pulses from argon-37 Auger electrons are different from the slower-rising pulses from a background gamma or cosmic ray.
Davis went back to Brookhaven and asked the local electronic experts if it would be possible to implement this technique for the very small counters that he used. The initial answer was that the available amplifiers were not fast enough to be used for this purpose with the small solar neutrino counters. However, within about a year, three first-class Brookhaven electronic engineers, Veljko Radeca, Bob Chase and Lee Rogers, were able to build electronics fast enough to be used to measure the risetime in Davis’s counters.
This “swimming-pool” improvement was crucial for the success of the chlorine experiment and the subsequent radiochemical gallium solar neutrino experiments – SAGE, GALLEX and GNO. Measurements of the risetime as well as the pulse energy greatly reduce the background for radiochemical experiments. The backgrounds can be as low as one event in three months.
In 1978, after a decade of disagreement between the Homestake neutrino experiment and standard solar model predictions, it was clear that the subject had reached an impasse and a new experiment was required. The chlorine experiment is, according to standard solar model predictions, sensitive primarily to neutrinos from a rare fusion reaction that involves boron-8 neutrinos. These are produced in only 2 of every 104 terminations of the basic proton-proton fusion chain. In early 1978 there was a conference of interested scientists at Brookhaven to discuss what to do next. The consensus was that we needed an experiment that was sensitive to the low-energy neutrinos from the fundamental proton-proton reaction.
The only remotely practical possibility appeared to be another radiochemical experiment, this time with gallium-71 (instead of chlorine-37) as the target. However, a gallium experiment (originally proposed by Russian theorist V A Kuzmin in 1965) was expensive – we needed about three times the world’s annual production of gallium to do a useful experiment.
In an effort to generate enthusiasm for a gallium experiment, we wrote another Physical Review Letters paper, this time with a number of interested experimental colleagues. We argued that a gallium detector was feasible and that a gallium measurement, which would be sensitive to the fundamental proton-proton neutrinos, would distinguish between broad classes of explanations for the discrepancy between prediction and observation in the chlorine-37 experiment. Over the next five or six years, the idea was reviewed a number of times in the US, always very favourably. A blue-ribbon panel headed by Glenn Seaborg enthusiastically endorsed both the experimental proposal and the theoretical justification.
To our great frustration and disappointment, the gallium experiment was never funded in the US, although many of the experimental ideas that gave rise to the Russian experiment (SAGE) and the German-French-Italian-Israeli-US experiment (GALLEX) largely originated at Brookhaven. Physicists strongly supported the experiment and said that the money should come out of an astronomy budget; astronomers said it was great physics and should be supported by the physicists. The US Department of Energy (DOE) could not get the nuclear physics and the particle physics sections to agree on who had the financial responsibility. In a desperate effort to break the deadlock, Bahcall was even the principal investigator of a largely Brookhaven proposal to the US National Science Foundation (which did not support proposals from DOE laboratories). A pilot experiment was performed with 1.3 tons of gallium by an international collaboration (Brookhaven, Pennsylvania, MPI Heidelberg, IAS Princeton and the Weizmann Institute), which developed the extraction scheme and the counters eventually used in the GALLEX full-scale experiment.
In strong contrast with what happened in the US, Moissey Markov, head of the Nuclear Physics Division of the Russian Academy of Sciences, helped to establish a neutrino laboratory within the Institute for Nuclear Research, participated in the founding of the Baksan neutrino observatory, and was instrumental in securing 60 tons of gallium free for Russian scientists for the duration of a solar neutrino experiment.
The SAGE Russian-US gallium experiment went ahead under the leadership of Gavrin, Zatsepin (Institute for Nuclear Research, Russia) and Bowles (Los Alamos), while the mostly European experiment (GALLEX) was led by Till Kirsten (Max Planck Institute, Germany). Both had a strong but not primary US participation.
The two gallium experiments were performed during the 1990s and gave very similar results, providing the first experimental indication of the presence of proton-proton neutrinos. Both experiments were tested by measuring the neutrino rate from an intense laboratory radioactive source.
There were two dramatic developments in the solar neutrino saga, one theoretical and one experimental, before the gallium experiments produced observational results. In 1985 two Russian physicists proposed an imaginative solution to the solar neutrino problem that built on the earlier work of Gribov and Pontecorvo and, more directly, the insightful investigation by Lincoln Wolfenstein (Carnegie Mellon).
Alexei Smirnov and Stanislav Mikheyev showed that, if neutrinos have masses in a relatively wide range, then a resonance phenomenon in matter (now universally known as the MSW effect) could efficiently convert many of the electron-type neutrinos created in the interior of the Sun to more difficult to detect muon and tau neutrinos. The MSW effect can work for
small or large neutrino mixing angles. Because of the elegance of the theory and the possibility of explaining the experimental results with small mixing angles (analogous to what happens in the quark sector), physicists immediately began to be more sympathetic to particle physics solutions to the solar neutrino problem. More importantly, they became enthusiasts for new solar neutrino experiments.
The next big breakthrough also came from an unanticipated direction. The Kamiokande water Cherenkov detector was developed to study proton decay in a mine in the Japanese Alps and set an important lower limit on the proton lifetime. In the late 1980s the detector was converted by its Japanese founders, Masatoshi Koshiba and Yoji Totsuka, together with some US colleagues, Gene Beier and Al Mann of the University of Pennsylvania, to be sensitive to the lower energy events expected from solar neutrinos.
With incredible foresight, these experimentalists completed their revisions to make the detector sensitive to solar neutrinos in late 1986, just in time to observe the neutrinos from Supernova 1987a emitted 170 000 years earlier. (Supernova and solar neutrinos have similar energies – about 10 MeV – much less than the energies relevant for proton decay.) In 1996 a much larger water Cerenkov detector (with 50 000 tons of pure water) began operating in Japan under the leadership of Yoji Totsuka, Kenzo Nakamura, Yoichiro Suzuki (from Japan), and Jim Stone and Hank Sobel (from the US).
So far, five experiments have detected solar neutrinos in approximately the numbers (within a factor of two or three) and in the energy range (less than 15 MeV) predicted by the standard solar model. This is a remarkable achievement for solar theory, because the boron-8 neutrinos that are observed primarily in three of these experiments (chlorine, Kamiokande and its successor SuperKamiokande) depend on approximately the 25th power of the central temperature. The same set of nuclear fusion reactions that are hypothesized to produce the solar luminosity also give rise to solar neutrinos. Therefore, these experiments establish empirically that the Sun shines by nuclear fusion reactions among light elements in essentially the way described by solar models.
Nevertheless, all of the experiments disagree quantitatively with the combined predictions of the standard solar model and the standard theory of electroweak interactions (which implies that nothing much happens to the neutrinos after they are created). The disagreements are such that they appear to require some new physics that changes the energy spectrum of the neutrinos from different fusion sources.
Solar neutrino research today is very different from how it was three decades ago. The primary goal now is to understand the neutrino physics, which is a prerequisite for making more accurate tests of the neutrino predictions of solar models. Solar neutrino experiments today are all large international collaborations, each typically involving in the order of 100 physicists. Nearly all of the new experiments are electronic, not radiochemical, and the latest generation of experiments measure typically several thousand events per year (with reasonable energy resolution), compared with typically 25-50 per year for the radiochemical experiments (which have no energy resolution, only an energy threshold).
Solar neutrino experiments are currently being carried out in Japan (SuperKamiokande in the Japanese Alps), in Canada (SNO, which uses a kiloton of heavy water in Sudbury, Ontario), in Italy (BOREXINO, ICARUS and GNO, each sensitive to a different energy range and all operating in the Gran Sasso Underground Laboratory), in Russia (SAGE in the Caucasus region) and in the US (Homestake chlorine experiment). The SAGE, chlorine and GNO experiments are radiochemical; the others are electronic.
Since 1985 the chlorine experiment has been operated by the University of Pennsylvania under the joint leadership of Ken Lande and Davis. Lande and Paul Wildenhain have introduced major improvements to the extraction and measurement systems, making the chlorine experiment a valuable source of new precision data.
The most challenging and important frontier for solar neutrino research is to develop experiments that can measure the energies of individual low-energy neutrinos from the basic proton-proton reaction, which constitutes (we believe) more than 90% of the solar neutrino flux.
Solar neutrino research is a community activity. Hundreds of experimentalists have collaborated to carry out difficult, beautiful measurements of the elusive neutrinos. Hundreds of researchers have helped to refine the solar model predictions, measuring accurate nuclear and solar parameters and calculating input data such as opacities and equation of state.
Three people have played special roles. Hans Bethe was the architect of the theory of nuclear fusion reactions in stars, as well as our mentor and hero. Willy Fowler was a powerful and enthusiastic supporter of each new step and his keen physical insight motivated much of what was done in solar neutrino research. Bruno Pontecorvo opened everyone’s eyes with his original insights, including his early discussion of the advantages of using chlorine as a neutrino detector and his suggestion that neutrino oscillations might be important.
Over the next decade, neutrino astronomy will move beyond our cosmic neighborhood and, we hope, will detect distant sources. The most likely candidates now appear to be gamma-ray bursts. If the standard fireball picture is correct and if gamma-ray bursts produce the observed highest-energy cosmic rays, then very-high-energy (1015 eV) neutrinos should be observable with a km2 detector. Experiments that are capable of detecting neutrinos from gamma-ray bursts are being developed at the South Pole (AMANDA and ICECUBE), in the Mediterranean Sea (ANTARES, NESTOR) and even in space.
Looking back on the beginnings of solar neutrino astronomy, one lesson appears clear: if you can measure something new with reasonable accuracy, then you have a chance to discover something important. The history of astronomy shows that it is very likely that what you will discover will not be what you were looking for. It helps to be lucky.
A version of this article originally appeared as a Millennium Essay in J N Bahcall and R Davis Jr 2000 Publications of the Astronomical Society of the Pacific112 429). Copyright 2000, Astronomical Society of the Pacific, reproduced with permission of the editors.
J N Bahcall and R Davis Jr 1976 Solar Neutrinos: a scientific puzzle Science191 264.
J N Bahcall and R Davis Jr 1982 An account of the development of the solar neutrino problem Essays in Nuclear Astrophysicsed.
C A Barnes, D D Clayton and D N Schramm (Cambridge University Press) 243. (This article is also reprinted in Neutrino Astrophysicsby J N Bahcall (Cambridge University Press, 1989).)