ÐÓ°ÉÔ­´´

In attempting to understand how the Sun shines, physicists and astronomers
have been unexpectedly confronted with a mystery: the case of the missing
neutrinos. These exotic particles, produced in the Sun, have been observed
in an experiment that has been running for a quarter of a century at the
bottom of a gold mine. But fewer neutrinos have been detected than standard
theories predict. There are two possible reasons for this. Astronomers could
be wrong about the way the Sun works. Or physicists may need to rethink
their theories of how weakly interacting particles behave. The discrepancy
has generated a flood of theoretical proposals and several new experiments
to test these ideas. The latest, which many hoped would provide the definitive
answer, has added yet another twist to the tale.

Physicists and astronomers believe that the Sun shines because of the
conversion of hydrogen nuclei (protons) to helium nuclei (alpha particles)
in the solar core(see Figure 1a). In the process, positive electrons
(positrons) and neutrinos (n) are also produced along with about 25 million
electronvolts (MeV) of thermal energy for every four protons burned; one
electronvolt is the energy an electron acquires by passing through a potential
of one volt. About 600 tonnes of hydrogen are converted to helium every
second in the Sun’s central regions, providing the energy that we know as
sunlight and making life on Earth possible.

The standard model of how the Sun works is constructed with the aid
of a computer from the most accurate data concerning solar nuclear reactions,
the most precise available physical description of the solar interior, and
the general equations that are thought to govern the evolution of stars.
Since we know more about the Sun than about any other star and since the
Sun is believed to be in an easy-to-calculate, stable middle-aged state,
a precise test of the standard solar model provides a unique and critical
test of how stars produce energy and evolve.

Neutrinos are elementary particles that travel at essentially the speed
of light. As well as being produced in the Sun they are produced on Earth
in nuclear reactions that involve natural radioactivity, in nuclear fission
reactors, and in high-energy physics accelerators. Unlike the traditional
messengers of astronomy, light particles called photons, neutrinos interact
only weakly with matter. This weak interaction enables them to escape directly
from the centre of the Sun and to provide astronomers with otherwise inaccessible
information about the centre of our nearest star.

The Sun acts as a natural particle accelerator, producing a beam of
neutrinos that can be used to probe the so-called electro-weak interactions
that act between sub-atomic particles. The theory of elec-troweak interactions
offers a single, extraordinarily precise description of all the known electric,
magnetic and weak interactions among elementary particles. There are three
types of neutrino: electron-type, muon-type and tau-type. According to the
standard model of electroweak interactions, neutrinos have no mass, travel
at essentially the speed of light, and never change from one type to another.
Solar neutrino experiments test the standard formulation of electroweak
theory on energy scales or timescales that cannot be achieved with traditional
laboratory experiments. Neutrinos from the Sun offer a glimpse of what
is happening deep inside a star and so allow us to test in detail theories
of how the stars shine as a result of nuclear reactions, and how they evolve.

Deep physics

Because neutrinos interact only weakly with matter, large detectors,
typically made up of hundreds or thousands of tonnes of material, are needed
to capture them. These detectors must be placed far underground in sheltered
places like mines and specially built deep laboratories. Otherwise, the
rare occasions when astronomical neutrinos trigger something observable
in the detectors would be confused with the more numerous background interactions
caused by high-energy, strongly interacting particles known as cosmic rays
that come to us from various parts of the Galaxy.

For two decades, beginning in 1967, the only operating solar neutrino
experiment was carried out by the chemist Raymond Davis, now at the University
of Pennsylvania. Davis uses an underground tank in the Homestake gold mine
in South Dakota, containing more than 600 tonnes of a dry-cleaning fluid
called perchloroethylene. A quarter of the chlorine atoms that occur naturally
in this fluid are the isotope chlorine-37, rather than the more common chlorine-35.
A chlorine-37 atom conveniently and efficiently captures neutrinos to become
a radioactive argon-37 atom. Davis detected only about one-quarter of the
solar neutrinos predicted by theoretical calculations. This disparity between
theory and observation is known as the solar neutrino problem.

Prompted by this puzzle, Japanese researchers, together with physicists
from the University of Pennsylvania developed a second experiment in 1987
in an underground laboratory in the Japanese Alps. This experiment, currently
led by Yoji Totsuka of the University of Tokyo, detects the scattering of
incoming solar neutrinos by electrons in 680 tonnes of ultra-pure water.
The scattered electrons are observed with the aid of large photosensitive
detectors that collect the characteristic light emitted by fast-moving electrons.
The direction in which this light is emitted shows that the neutrinos come
from the Sun.

The water experiment, called Kam-iokande II, has confirmed the discrepancy
between theory and observation. The energy threshold for detecting neutrinos
in the Japanese experiment – that is, the lower limit of neutrino energies
it can detect – is much higher than for the chlorine experiment and the
discrepancy with theory is less. The degree of disagreement with standard
calculations appears to depend upon the energy of the solar neutrinos.

Most of the neutrinos that the chlorine and water experiments were designed
to detect come from a rare reaction in which beryllium-7, an unstable isotope,
captures a proton to form radioactive boron-8. The rate of this reaction
is very slow. In order to get close enough to fuse, the proton and the beryllium
nuclei must overcome a large energy barrier due to the repulsion of their
electric charges. This reaction is calculated to occur only once in every
5000 times that four hydrogen nuclei are burned as in Figure 1a. Despite
this, the predicted number of neutrinos detected in the chlorine and the
water experiments is dominated by the rare reaction because it produces
high-energy neutrinos to which the detectors are much more sensitive.

Physicists and astronomers believe that most solar neutrinos are produced
in the fundamental initial reaction in the solar energy-generating process
of Figure 1a, the so-called proton-proton (p-p) reaction (Figure 1b). In
this reaction, a proton (p) decays to a neutron in the vicinity of another
proton, forming a heavy hydrogen nucleus called deuterium (hydrogen-2),
and emitting a positron and a neutrino. The neutrinos that this reaction
produces have energies of less than 0.4 MeV, too low to be detected by the
chlorine and the water experiments. Astrophysicists can calculate to within
2 per cent the number of p-p neutrinos that should be produced per unit
of time according to the standard solar model.

Sun on the spot

In a crucial test of both the standard solar model and the physics
of how neutrinos behave, the low-energy p-p neutrinos are currently being
observed in two experiments with large gallium detectors in underground
observatories: one in Russia, where 60 tonnes of gallium metal are used,
the other in Italy where the detector contains 30 tonnes of gallium in the
form of a gallium chloride solution.

The isotope gallium-37 is converted into germanium-71 when gallium absorbs
a neutrino. The germanium isotope is radioactive and can be extracted chemically
and counted in a manner similar to the way argon-37 is extracted and counted
in the chlorine experiment. American scientists are participating in the
Russian experiment, which is called SAGE (for Soviet American Gallium Experiment)
and is being carried out in the Baksan Laboratory which was specially excavated
underneath the Andyrchi mountain massif in the North Caucasus region. The
European experiment, called Gallex, involves scientists from Germany, France,
Italy, Israel, and the US and is taking place in the Gran Sasso Underground
Laboratory in Italy. In June 1990, at a scientific meeting at CERN in Geneva,
Tom Bowles of Los Alamos National Laboratory, who leads the American team,
and Vladimir Gavrin of the University of Moscow, who leads the Russians,
created a sensation by announcing that the preliminary results of the SAGE
experiment indicated that most of the basic p-p neutrinos (see Figure 1b)
were missing. If these results were correct, the consequences would be
revolutionary: new physics would be needed to explain why so few p-p neutrinos
were being detected.

Two months ago, many of the same scientists assembled in Granada, Spain,
to hear from Till Kirsten of the Max Planck Institute for Nuclear Physics
in Heidelberg the first results from the Gallex collaboration (see New ÐÓ°ÉÔ­´´,
Science, 11 July). For several months prior to this meeting, conflicting
rumours pulsed through the scientific community. Both astronomers and physicists
were looking to the results of the Gallex experiments, in combination with
those already announced from the SAGE experiment, to provide unequivocal
answers to basic questions. Is our conventional understanding of how the
Sun shines incorrect? Or do we need new physics beyond the textbook theory
of electroweak interactions?

When the Gallex results were announced on 1 June, the long-awaited answer
was a resounding ‘maybe’. The results identify neither the astronomers nor
the physicists as clear culprits in the solar neutrino problem. The measured
rate of 83+-21 solar neutrino units (an SNU is a convenient unit for measuring
the rate at which solar neutrinos are detected) is significantly different
from the rate of 132+-7 SNU calculated with the standard solar model. But
it is not far enough off, considering the large experimental uncertainties,
to demand new physics rather than new astronomy. The Gallex experiment has
not solved the solar neutrino problem, but it has made a major advance by
observing the fundamental, low-energy p-p neutrinos for the first time.

At the same meeting, the SAGE researchers presented other interesting
results. These indicate that more neutrinos are being detected in their
experiment than was originally suggested. It seems likely that the SAGE
and Gallex experiments will eventually yield similar answers once they have
been running for a few more years and uncertainties with the equipment and
statistics have been reduced.

Missing neutrinos

The core of the solar neutrino problem remains the low counting rate
observed in Davis’s chlorine experiment. His result is especially difficult
to explain when combined with the more moderate deficit of high-energy neutrinos
detected by the Japanese Kamiokande II pure water experiment. Both these
experiments are sensitive primarily to the same rare neutrinos from radioactive
boron. Hans Bethe of Cornell University and I have argued that the only
way to reconcile the published results of these two experiments is to infer
that the physicists’ standard electroweak model does not correctly predict
the behaviour of the electron-type neutrino.

So far, four solar neutrino experiments have measured the rates at
which neutrinos of different energies arrive at Earth and all four have
found rates outside the range predicted on the basis of the combined standard
solar and standard electroweak models. But some scientists have expressed
reservations about whether we really need new physics.

When will we know the final answer? Not until 1994, or perhaps even
1996. By 1994, the two gallium experiments should yield results of higher
statistical accuracy, which may yet point us towards a solution to the solar
neutrino problem. If not, by 1996 two powerful new experiments will be operating
with large counting rates and high statistical accuracy. One of these, called
the Super-Kamiokande experiment, is being constructed in Japan. It is a
much improved, much larger version of the Kamiokande II pure water experiment.
The second new experiment, the Sudbury Neutrino Observatory (SNO), is being
built in an INCO nickel mine near Sudbury in Ontario. The SNO experiment
will capture neutrinos with a thousand tonnes of precious heavy water (D2O),
in which deuterium replaces ordinary hydrogen.

Both of these new experiments will detect neutrinos using photosensitive
detectors, similar to those developed for the Kamiokande II experiment,
which collect light emitted by the fast-moving electrons produced by the
neutrino interactions. Most importantly, they will also use electronic means
to measure the energies of individual neutrinos and so determine the energy
spectrum of neutrinos that reach us – that is, the relative numbers of neutrinos
with different energies.

This energy spectrum is a crucial test: errors in astrophysical theory
can only change the total numbers of neutrinos from a given nuclear reaction;
they do not affect the neutrino energy spectrum. On the other hand, some
attractive theoretical explanations contradict the standard electroweak
model and imply that the neutrino energy spectrum can be changed. The new
measurements of the neutrino energy spectrum should tell us whether it is
correct to infer deviations from the standard model of particle physics
on the basis of the chlorine and Kamiokande II experiments. It also appears
likely that an Italian-American collaboration called Borexino will be able
to make a crucial diagnostic test by observing neutrinos of a specific and
relatively low energy produced by beryllium nuclei in the Sun.

While they wait for these experiments physicists and astrophysicists
are guided by an aesthetic sense. There is a beautiful theory, developed
between 1978 and 1986, that resolves the solar neutrino problem by contradicting
the standard assumption that neutrinos have no mass. Named the MSW effect
after the three scientists who devised it – Stanislav P. Mikheyev and Alexei
Yu. Smirnov of Moscow University and Lincoln Wolfenstein of the Carnegie
Institute of Technology in Pittsburgh – it describes how neutrinos can lose
their sense of identity when interacting with the large number of electrons
in the Sun. Solar fusion reactions produce electron-type neutrinos which
can be detected relatively easily by existing experiments. According to
the MSW theory, as they interact with the many electrons in the Sun these
electron-type neutrinos are converted into another type of neutrino, either
a ‘muon-type’ or a ‘tau-type’, that is much more difficult to detect. If
the MSW theory is correct, the long-sought solar neutrinos are not really
missing – merely hard to observe.

Testing aesthetics

By comparing the rate of two different reactions called ‘charge current’
and ‘neutral current’ reactions in their heavy water experiment, the SNO
researchers will be able to test directly the MSW prediction. The charge
current reaction records only electron-type neutrinos; the neutral current
reaction registers neutrinos of all types. If the standard electroweak theory
is correct, and neutrinos do not change their type, the flux of electron-type
neutrinos measured with the charge current reaction and with the neutral
current reaction will be the same because standard theory predicts that
the Sun produces only electron-type neutrinos.

One formulation of the MSW theory that agrees with the Gallex, the chlorine
and the Kamiokande II experimental results implies that the electron-type
neutrino is physically slightly mixed with a neutrino of a different type
that has a tiny mass of about 0.003 eV. This mass is more than a thousand
times smaller than existing terrestrial experiments can measure. Solar
neutrino experiments are sensitive to such tiny values of the mass because
the neutrino beam in the centre of the Sun encounters enormous amounts of
solar matter in travelling from the centre of the Sun to the underground
detectors on Earth.

With our level of understanding of the Sun, calculations of neutrino
emission from the solar interior can be done with relatively high precision.
For the past quarter of a century, many astronomers have therefore taken
the position: ‘We understand the Sun, so if there is a problem, it must
be in the physics of the neutrino.’ Until relatively recently, many physicists
adopted a similar attitude towards astronomy, suggesting that the solar
neutrino problem proves that astrophysicists don’t know what they are talking
about. Solar neutrino research to date hinges on the question of who is
right.

The first quarter-century of solar neutrino astronomy produced a scientific
mystery, the missing solar neutrinos. In the next few years experimenters
expect to solve this mystery, or at least to identify the principal culprit
and to point us towards either a more complete theory of stellar energy
generation or a better theory of neutrino propagation. If we are lucky,
they might do both.

In any event, astronomers and astrophysicists are likely to be pleased
by the present situation. When the first detailed calculation of solar neutrino
fluxes was carried out in 1962, a few outspoken physicists doubted whether
astrophysicists could calculate the results of solar evolution with sufficient
accuracy to make a solar neutrino experiment meaningful. Now the situation
is different: physicists are debating whether the estimated 15 per cent
theoretical uncertainty (approximately one standard deviation) in the higher-energy,
boron-8 solar neutrino flux is sufficiently conservative. Since this rare
neutrino reaction is estimated to occur only once in every 5000 times the
Sun burns four hydrogen nuclei to form a helium nucleus – that is, the
energy production is right to one part in 5000 – any astronomical uncertainties
are small and the basic correctness of the solar calculations is no longer
being questioned. Solar neutrino experiments have confirmed with impressive
accuracy the fundamental idea that stars shine by nuclear fusion reactions
among light elements. Astronomers and astrophysicists have plenty to be
pleased about.

John N. Bahcall is a professor of Natural Sciences at the Institute
for Advanced Study in Princeton, New Jersey. Further reading: Neutrino Astrophysics
by John N. Bahcall, Cambridge University Press, 1989.