¿ìè¶ÌÊÓƵ

On the night of 24 February 1987, an assistant at the Las Campanas observatory
in Chile glanced up and saw a new star. A Canadian astronomer at the observatory
stared in disbelief at the star’s image on a photograph. An amateur astronomer
in New Zealand gazed in excitement through his telescope eyepiece. Oscar
Duhalde, Ian Shelton and Albert Jones had discovered the brightest supernova
in almost 400 years.

The last supernova visible to the naked eye was seen in 1604, long before
astronomers knew enough to call it a supernova, and several years before
they first turned the telescope to the sky. The story of supernovae really
begins in the 1930s, when Fritz Zwicky – a Swiss astronomer working in the
US – realised that some temporary stars in other galaxies were too luminous
to be ordinary novae, and coined the term supernova. An ordinary nova is
a merely a puff of gas blown off the surface of a compact star. Zwicky realised
that a supernova was a star that literally blew itself apart.

In a remarkably prescient paper with his colleague Walter Baade in 1933,
Zwicky suggested that ‘supernovae represent the transitions from ordinary
stars into neutron stars, which in their final stages consist of extremely
closely packed neutrons’. This was only three years after the discovery
of the neutron in the laboratory, and astronomers took 34 years to discover
the first neutron stars, in the shape of the radio-emitting ‘pulsars’. Baade
and Zwicky also concluded that supernovae produce extremely fast subatomic
particles called cosmic rays.

Following Zwicky’s lead, astronomers have discovered hundreds of supernovae
in distant galaxies. But until recently this research formed rather a backwater,
and few astronomers wanted to follow up the discoveries. Even when the most
powerful telescopes were turned on the supernovae, they could reveal little
about an explosion that is taking place millions of light years away.

That’s why the new star of five years ago, unromantically christened
Supernova 1987A, rocked astronomers around the world (‘Supernova: the cosmic
bonfire’, ¿ìè¶ÌÊÓƵ, 5 November 1987). It lay in the nearest galaxy
to our own, the Large Magellanic Cloud, at a distance of only about 170
000 light years. Not only was the supernova visible to the naked eye, but
its radiations could be picked up by telescopes operating at all wavelengths
from gamma ray to radio waves, and those that detect neutrinos. Astronomers
had even photographed the star before it erupted.

Supernova 1987A has focused the minds of all astronomers on these most
spectacular events . It has been a particularly exciting few years for astronomers
already working on supernovae, such as theorist Stan Woosley, of the Lick
Observatory in California. ‘It was like Christmas all the time, with new
results coming in every week,’ Woosley recalls. ‘I was on an adrenaline
high for a year after the discovery.’ With hindsight, we can now pick out
the scientific highlights.

The single most important event occurred without anyone knowing about
it. Twenty hours before the discovery, a burst of ghostly subatomic particles
called neutrinos swept past the Earth: several million million neutrinos
passed through your body. Neutrinos are produced by nuclear reactions such
as some of those in stars. A few of the neutrinos – 19 in total – were detected
in two large tanks of water deep underground in the US and Japan. These
tanks were particle detectors designed to look for the decay of protons
which some fundamental theories of physics predict.

The neutrinos undoubtedly came from Supernova 1987A. The energy and
length of the burst showed that they were generated in a region that reached
a temperature of 50 000 million K for just a few seconds. These are exactly
the conditions at the centre of a star when it collapses to become a neutron
star. Calculations also show that the visible light emitted represented
only a tiny part of the exploding star’s energy budget. More than 99.9 per
cent of the energy was carried away by neutrinos – enough to power all the
stars in our Galaxy for several years.

This was only the second time that astronomers had detected neutrinos
from a star. The first had been rather embarrassing. Researchers had found
that the Sun produced only one-third of the amount of neutrinos that the
theorists have predicted. And the Sun, after all, is the star we supposedly
know most about. The number of neutrinos from the exotic environment of
a supernova was so close to what was expected, however, that theorists have
taken heart that the basic theories of stars are not far wrong. The neutrinos
have also been a boon to scientists in fields of research ranging from relativity
theory to cosmology .

Supernova 1987A did not fulfil all the predictions, however, and the
discrepancies have given astronomers a chance to refine their theories of
how stars shine. Since Zwicky’s time, astronomers have learnt that there
are two kinds of supernova, imaginatively called Type I and Type II. The
spectrum of Supernova 1987A showed lines from hydrogen, which designated
it as Type II. According to the orthodoxy, a Type II supernova is the explosion
of a massive star that has swollen to become a red supergiant, hundreds
of times larger than the Sun.

For the first time, astronomers had records of a supernova before it
exploded. Yes, the star was a heavyweight, about 20 times heavier than the
Sun. But, no, it was not a red supergiant when it exploded: it was blue,
indicating a much hotter star – and only 50 times the size of the Sun. After
five years, astronomers have worked out why. The star was different from
old stars in the Milky Way because it was born in a galaxy that has a much
smaller proportion of heavy elements. Both galaxies are made mainly of hydrogen
and helium, but the Large Magellanic Cloud has only one-quarter the amount
of elements heavier than helium.

In old stars in our Galaxy, the heavier elements are good at absorbing
the radiation coming up from within the star. This makes the star swell
to supergiant size, with the outer region billowing up and down in large
convective currents like a pan coming to the boil. If a star has less of
these heavy elements, radiation can travel relatively unhindered into space
and the star is therefore smaller and hotter – a blue giant.

According to Woosley, an old star in the Large Magellanic Cloud can
exist in either of these states, and can change from one to the other. The
star that exploded as Supernova 1987A was born about 20 million years ago,
and swelled into a red supergiant late in life. Sometime between 10 000
and 50 000 years before the explosion, it shrank to become a blue giant.

Last year, the Hubble Space Telescope found more evidence for this change
from red supergiant to blue giant. Lying around the supernova is a faint
ring of gas, tipped up to our view so that looks it like an oval. It lies
one-third of a light year from the supernova, so close that no telescope
on the ground can show it clearly. But the Hubble telescope – even with
its misshapen main mirror – can reveal the ring clearly.

There is only one reasonable way to explain the ring. When the star
was a red giant, a steady wind of gases blew off from its surface. The wind
blew fastest from the poles and more slowly from the equator, so the gases
formed an hour-glass shape around the star. When the star changed to become
a blue giant, the nature of the wind changed. It became more tenuous, but
faster. As this gale swept outwards, it ran into the waist of the hour-glass
and compressed the denser gas there into the glowing ring that we now observe.

Just a few weeks after the explosion, the International Ultraviolet
Explorer satellite launched in 1978 to study the spectrum of ultraviolet
sources had detected spectral lines that clearly come, not from the fast-expanding
gases of the supernova, but from stationary gas nearby that had been lit
up by the explosion. With hindsight, this must have been the gas in the
ring. Nino Panagia of the European Space Agency has used this spectral measurement
to work out the distance to the supernova. The ultraviolet light brightened
over a period of several months, showing that the ring has a radius of 0.68
light years. The Hubble telescope found that it appears 1.66 arcseconds
across (where an arcsecond is roughly 1/2000 the apparent diameter of the
Moon). To appear this size, the ring must lie 169 000 light years from us.

The supernova has thus provided a direct and accurate distance to the
Large Magellanic Cloud. Previous estimates had relied on comparing stars
in the galaxy with those in the Milky Way, and gave conflicting estimates
around 160 000 light years. Astronomers then found the distance to farther
galaxies by comparing their stars to those in the Large Magellanic Cloud,
so the new measurement suggests that all other galaxies are slightly more
distant than previously had been thought.

Not surprisingly, Supernova 1987A has tied down the theory of supernova
explosions more convincingly. Since the pioneering work by Zwicky and Baade,
astronomers had developed two theories of how the collapse of a star’s core
could blow its outer layers into space. According to one theory, the outer
gases begin to fall inwards, but when they hit the solid surface of the
newly formed neutron star a reverse shock wave rips outwards through the
star. On the other hand, the outer gases might be flung out by the flood
of neutrinos from the core, like steam lifting the lid of a kettle. Supernova
1987A has confirmed that both mechanisms are important. A reverse shock
began to move out from the star’s centre, but it slowed down and was about
to stop when the neutrinos set it going again.

After its initial flash of light in February 1987, Supernova 1987A faded
a little and then began to brighten again. By mid-May, it was 250 million
times more brilliant than the Sun. The only obvious source for this luminosity
was energy from the decay of radioactive elements created by nuclear reactions
in the star. Woosley had already suggested that the supernova produced some
radioactive nickel-56. This decays quickly to cobalt-56, which in turn decays
into stable iron-56 over a period of several months.

Confirmation was not long in coming. First, the supernova began to fade
at just the rate that cobalt-56 decays. Then, towards the end of 1987, satellites
and instruments borne on high-altitude balloons picked up gamma rays with
the precise energy emitted by cobalt-56. And over the next couple of years,
astronomers at the Anglo-Australian telescope saw the optical spectral line
from cobalt actually fading in the supernova’s spectrum as this element
turned into iron in front of their eyes. (‘Portrait of an exploding star,’
¿ìè¶ÌÊÓƵ, 24 February 1990). The mass of cobalt can be calculated independently
from these three measurements, and the figures agree well. The supernova
created a mass of nickel – which decayed to cobalt – equal to 7 per cent
of the Sun’s mass.

This discovery has shed light on the second, completely different, kind
of supernova. A Type I supernova is probably the explosion of a small white
dwarf star, pushed over its natural limit of stability when extra matter
falls on it from a companion star. Type I supernovae are extremely bright,
and astronomers had long suspected that radioactivity makes them shine.
Now, Supernova 1987A has shown that a star explosion can easily produce
nickel-56, and that its energy can be turned into light. There is now no
doubt that Type I supernovae are lit up by the same radioactive decay: this
kind of explosion converts 20 times as much matter into radioactive nickel,
so in consequence a Type I supernova is correspondingly brighter.

For the first couple of years, Supernova 1987A faded at just the decay
rate of cobalt-56 – allowing for the fact that some of the energy escaped
directly as gamma rays and X-rays. Now, the cobalt-56 has virtually all
disappeared. The main source of energy should be radioactive cobalt-57,
which was produced in smaller quantities but has a longer half-life. The
light from the supernova is indeed fading at a slower rate now, similar
to the decay curve of the heavier cobalt isotope. So far, however, the newly
launched satellite, the Compton Gamma Ray Observatory, has not detected
the characteristic radiation from cobalt-57. There could be another source
of energy: the neutron star born in the explosion.

Astronomers have been waiting patiently since 1988 for the underlying
neutron star to show itself. A fast-rotating neutron star can produce regular
pulses of radiation (a pulsar), but no-one has detected such a signal from
Supernova 1987A. There was a false alarm in 1989, when a team of American
astronomers claimed they had found rapid pulses from the supernova, but
they turned out to be interference from the TV camera on the telescope.

A neutron star should show up in other ways. Astronomers had expected
some of the gas from the explosion to fall back on to the neutron star’s
surface, and this would provide a source of energy that should make them
shine as brightly as 30 000 Suns. But the supernova’s light has already
dropped below that level. Theorists have suggested two explanations. The
more conservative view is that the neutron star has a strong magnetic field
which is preventing the gas from falling in. On the other hand, so much
gas may have fallen back at an earlier stage that the neutron star was pushed
over its natural weight limit, and collapsed to become a black hole.

At the moment, the supernova is now only a few hundred times brighter
than the Sun – around one per cent the brightness of the star before it
exploded. Astronomers are having difficulty following the star as it fades
further, because it lies near to two other stars which are now brighter
than the supernova itself.

Indeed, it is now fair to say that the supernova is dead. We are now
seeing a dispersing cloud of gas at what was the star’s heart. But that
does not mean that the show is over. On the contrary, a whole new field
of astronomy and science is opening up. The scene of the action is now shifting
to the super-nova’s surroundings, as the gases from the explosion crash
into the clouds of gas lying around the supernova. In the middle of 1990,
astronomers in Australia detected radio waves for the first time since the
explosion. This rebirth of activity is coming from a region between the
supernova and the ring detected by the Hubble Space Telescope. For the first
time, astronomers are witnessing the birth of a ‘supernova remnant’.

Radio astronomers have already catalogued some 150 supernova remnants
in our Galaxy: these expanding gas clouds are undoubtedly the remains of
stars that exploded hundreds or even thousands of years ago. The most famous
is the Crab Nebula, the remnant of a star seen to explode in 1054. But the
Crab Nebula is not very typical: it appears bright, even in a small telescope,
because a rapidly spinning neutron star in the centre is constantly supplying
it with energy. Most supernova remnants are faint when seen with an optical
telescope, but very powerful in their output of radio waves and X-rays.

The best-studied supernova remnants lie near us in the Milky Way, for
example, the remains of a star observed to explode in 1572 by the great
Danish astronomer Tycho Brahe. About a century later, another supernova
produced the remnant known as Cassiopeia A, an exceptionally strong radio
source. This is particularly interesting because the gas from the explosion
is currently running into small clouds that were ejected long before the
star exploded.

Using calculations of blasts from nuclear explosions in the Earth’s
atmosphere, astronomers have developed a basic theory for the motion of
a supernova shock wave as it blasts through the thin gas in space. But there
are still gaping holes in applying the theory to supernovae. For a start,
the Earth’s atmosphere is quite uniform, but the interstellar medium is
very clumpy and near a Type II supernova it also contains bits of gas shed
by the star before it exploded. The theory is not good at explaining what
happens when the shock runs into these lumps of gas, especially when we
do not know where the interstellar clouds were to begin with. In addition,
a cosmic explosion unfolds in slow-motion: in a supernova remnant that is
already several hundred years old, we have to watch for a century or more
to see any major changes.

The remnant of Supernova 1987A solves both these problems. Even within
a few years, we expect it to change rapidly: between now and 1997, the remnant
will double in size and grow to enclose eight times the volume of space.
We also know the location and nature of the surrounding gas clouds before
the shock hits them. Closest in is the ring found by Hubble. Ground-based
telescopes have shown swirls of gas and dust stretching several light years
from the supernova, while thin sheets of dust several hundred light years
away have revealed their existence by reflecting the supernova’s light in
spectacular haloes around the supernova’s position.

‘We expect to learn a lot about shock interactions,’ anticipates Roger
Chevalier, of the University of Virginia, ‘and so test our theories of the
X-ray and optical emission from supernova remnants. We don’t have a well-developed
theory for their radio emission yet, and we should learn from Supernova
1987A just what factors affect the radio output.’

Chevalier believes that the supernova’s shock is currently generating
radio waves because it has caught up with a slightly denser region in the
outflowing gases from its blue giant stage. He expects real fireworks when
the shock hits the ring, about the year 2000. Optical astronomers will see
the ring brighten dramatically as the shock wave ploughs through it, while
the radio emissions will reach their peak luminosity. Chevalier also predicts
a powerful surge of X-rays. ‘It could become a stronger X-ray source than
Cassiopeia A, because the shock is hitting the denser gas after only a few
years, rather than 300 years,’ he says.

Astronomers who study cosmic rays are also looking forward to the development
of the new supernova remnant. In general, they still believe Zwicky’s original
prediction that supernovae give rise to cosmic rays, but current theories
suggest that the individual particles are sped up to high energies in the
shock front of the supernova remnant. But they do not agree on the details.
The supernova remnant may be most efficient in generating cosmic rays when
it is young and energetic. Conversely, it may accelerate the particles best
when it is old and large, and covers a much larger volume of space.

The Astronomer Royal, Arnold Wolfendale, has been investigating cosmic
rays for many years, and he sees Supernova 1987A continuing to provide important
new science for many years. ‘There was of course a tremendous interest in
the early stages of the supernova, which has now declined somewhat, but
I believe that there will be a considerable amount of research that will
carry on – more or less independent of time – as we study the later processes
associated with the supernova remnant,’ he says.

At the time of the explosion, some supernova researchers wondered if
it might be just a nine-day wonder for those who control the world’s large
telescopes and astronomy satellites. They worried that once the excitement
had died down, no one would bother watching Supernova 1987A as its light
fizzled out. That fear has turned out to be unfounded: astronomers – of
one complexion or another – will be keeping tabs on Supernova 1987A for
centuries to come.

* * *

1: Supernova fever

Until 1987, few astronomers took any interest in supernovae. A small
band of professional and amateur astronomers looked regularly for exploding
stars in other galaxies, and found something like 15 a year.

They notified the Central Bureau for Astronomical Telegrams, in Cambridge,
Massachusetts, where Brian Marsden gave each supernova a designation by
year and order of discovery (1986A, 1986B . . . ) and informed the rest
of the world’s astronomers. But, by and large, the people with the large
telescopes did not bother to look at them.

All that has now changed. With supernova 1991BJ appearing in December,
last year saw the discovery of 62 supernovae, a new record. Although the
University of California at Berkeley has set up an automated telescope to
look for supernovae, most of the new discoveries have come from three astronomers
who are checking wide-angle photographs of the sky: Rob McNaught in Australia,
Jean Mueller in California and Christian Pollas in France.

‘Even more important than the rate of discovery,’ says Marsden, ‘is
the fact there is a terrific rate of follow-up.’ Someone on a large telescope
now investigates virtually every supernova that is discovered. By obtaining
spectra of supernovae, astronomers have traditionally assigned them to one
of two main classes (Type I or Type II); with new detailed spectra, astronomers
can now further subdivide each type into three or four subtypes.

This classification is not just ‘stamp-collecting’. Each subtype represents
a different kind of star that can explode. The similarities and differences
provide new insights into the mechanisms that can tear a whole star apart.
According to Stan Woosley, of Lick Observatory, ‘these high quality observations
of supernovae should allow us to solve the problem of the explosion mechanism
in this decade.

* * *

2: Physicists learn from the supernova’s neutrinos

The neutrinos from Supernova 1987A provided a rich harvest for physicists,
cosmologists, and astronomers.

For a start, they have given particle physicists information they could
not have gleaned from experiments on Earth. The fact that we detected any
at all shows that neutrinos must have a lifetime longer than 170 000 years.
They arrived in a bunch that was spread over only a few seconds – which
is roughly the time taken by the star’s core to collapse – so we can set
a limit on any effects that tend to make some neutrinos lag behind the others.
The Galaxy’s magnetic field would spread out the neutrinos if they had any
significant electric charge, and the observations, according to John Bahcall
of the Institute for Advanced Study in Princeton, mean that the neutrino’s
charge must be less than one million-million-millionth of the charge on
an electron.

If the neutrinos had a significant mass, they would also be spread out
over a period of more than a few seconds, with the most energetic neutrinos
travelling fastest and arriving first. Because the neutrinos did arrive
within a few seconds of one another, after a flight of 170 000 years, the
mass of the neutrino must be extremely small. Bahcall calculates it is less
than 16 electronvolts (30-millionths of the mass of the electron). This
is consistent with the most popular view among physicists, that the neutrino
– like the photon – has exactly zero mass and travels at precisely the speed
of light.

This measurement is important for cosmology, too. The Universe is filled
with neutrinos left over from the Big Bang, and if these have even a tiny
mass they could exert enough gravitational force to stop the Universe from
expanding forever. The new upper limit on their mass means that these neutrinos
cannot slow down the expansion of the Universe.