¿ìè¶ÌÊÓƵ

Some kind of extraordinary activity is going on at the centre of one
in ten of all the galaxies in the Universe. Very often, the centre of one
of these galaxies produces so much light that it outshines the rest of the
galaxy. These ‘active’ galaxies can be a hundred times brighter than a normal
galaxy and, because of this, they are visible over huge distances.

Astronomers have long been fascinated by active galaxies, but no telescope
is powerful enough to allow us to see what lies at their centre, the area
known as the active galactic nucleus or AGN. The most popular candidate
is a black hole, but the evidence remains inconclusive and some astronomers
want to reopen the debate. The keystone of their argument is a theory that
first surfaced 25 years ago, and is now gaining support.

A group of active galaxies called quasars set the ball rolling several
decades ago. Quasars (Quasi-Stellar objects) are the brightest, and in some
cases the most distant, objects in the universe. When they were discovered
in 1963, astronomers put forward theories to try to explain them. An initial
frontrunner was the starburst theory, which posits massive bouts of star
formation at the centres of these galaxies as an explanation for their brightness.
Then, in 1969, Donald Lynden-Bell of the Uni-versity of Cambridge proposed
that quasars could derive their power from material falling into black holes
at their cores. These black holes may be equivalent in mass to a billion
(109) Suns.

According to this theory, the amount of energy that can be released
by matter falling into a black hole is so enormous that it easily matches
the energy radiated by even the brightest AGN. ‘It is very difficult to
come up with an energy generation mechanism, other than a black hole, which
does not require extreme parameters to make it work,’ explains David Clements
of the University of Oxford. ‘For example, black holes process matter into
energy much more efficiently than normal thermonuclear fusion in stars.’
So the black hole swiftly became the standard explanation of how active
galaxies generate energy.

The starting point for any theory of active galaxies has to be to make
some sense of their remarkable variety. Detailed and subtle classifications
now exist, derived from profiles of the precise frequencies at which the
nuclear region emits energy. From these emission line profiles astronomers
can deduce what gases are present in the AGN and the conditions in which
they exist. Gas atoms radiate energy as their electrons move from a higher
energy state to a lower one. Which of these atomic transitions is favoured
(and hence the frequency of the energy emitted) depends on the range of
density of the gas, its temperature and its state of ionisation. The width
of the emission lines (the observed range of frequencies from a particular
transition) depends on the speed at which the emitting clouds move to and
fro relative to the observer . Broad emission lines are produced by fast-moving
clouds travelling in various directions relative to the line of sight. Slower-moving
clouds produce narrower lines. AGNs fall into four categories known as Seyfert
galaxies, radio galaxies, quasars and blazars.

The first active galaxies were catalogued in 1943 by the American astronomer
Carl Seyfert, following a survey he had made of spiral galaxies in which
he noticed that some spirals and barred spirals have extremely bright nuclei.
Very often, these nuclei were too small to be picked out or ‘resolved’ in
photographs; yet, compared with the rest of the galaxy, they shone brilliantly.
Examining the spectra of these regions, Seyfert discovered strong emission
lines. A Seyfert galaxy can be classified as ‘type 1’ or ‘type 2’, depending
on the width of its hydrogen emission lines. In Seyfert type 1 galaxies,
the hydrogen lines are very broad, indicating that the clouds producing
them are travelling as fast as 10 000 kilometres per second. In Seyfert
type 2 galaxies, the hydrogen lines are much narrower, indicating velocities
of around 1000 kilometres per second.

A year later, amateur astronomer Grote Reber detected a strong radio
source in the constellation of Cygnus in the Milky Way, which turned out
to be the first of many radio galaxies to be discovered. These objects
are almost always elliptical galaxies, possessing two huge ‘lobes’, one
on each side of the visible part of the galaxy, which radiate electromagnetic
energy. Some have jets of particles and radiation spurting out of them which
appear to have been emitted from the core of the galaxy, and astronomers
think that whatever causes the radio emission derives from the galaxy’s
nucleus.

The first quasar was detected when a radio source known as 3C₂73
was identified with an optical object in 1963. Despite its star-like appearance,
3C₂73 is not a star. From its spectrum, it had to be an AGN.
The red shift of its emission lines indicated that it was a huge distance
from the Earth, so the amount of energy it was emitting had to be vast.
In many ways the emission lines made it look like an exceptionally powerful
Seyfert galaxy.

The fourth class of active galaxies contains blazars, very energetic
and rapid varying emitters of radiation. The first blazar was discovered
in 1929, but at the time astronomers thought it was a variable star. It
was not until the 1970s that Joseph Miller of Lick Observatory in California
showed it to be the superbright nucleus of an elliptical galaxy.

One of the most interesting theories to be presented about AGNs suggests
that most active galaxies are the same type of object. According to this
theory, the differences observed from Earth are simply a result of astronomers
having to view them at different angles. This idea has led to various attempts
to reconcile these observed differences, ranging from theories which seek
to link a couple of types of AGN, to more ambitious ones which explain all
active galaxies as a single type of object.

A leading proponent of this approach is Robert Antonucci of the University
of California at Santa Barbara. He has suggested recently that some of the
evidence of orientation effects is so robust that astronomers are beyond
the stage of testing the theory, and need now only refine it. But not all
astronomers are convinced, and even Antonucci admits that some of the refinements
are fairly major ones. However, all supporters of such unified schemes agree
that black holes are a plausible explanation.

As material begins to fall into a black hole, rotation of the galaxy
surrounding it makes the material swirl like bath water around a plug hole.
This creates an ‘accretion disc’ of material which heats up as it swirls
around, rubbing against itself and radiating the energy this generates into
space.

This theory also accounts for differences in the observed emission spectra
of different active galaxies. Surrounding the accretion disc is a thick,
dusty, doughnut-shaped ring called the torus, which blocks the observer’s
view from certain angles. Because the accretion disc and torus lie in the
equatorial plane of the black hole, very little radiation from the central,
broad line region can escape that way. Instead it is channelled along the
rotation axis of the black hole. This produces two jets which beam powerful
radiation and subatomic particles into space. Viewing the AGN along the
jets will provide a very different view from looking at it along the accretion
disc.

Hydrogen clouds in the neighbourhood of the torus produce the emission
lines detectable from Earth. The speed with which the clouds orbit, and
hence the width of the emission lines they produce, depends on how close
they are to the black hole. The closer they are, the faster they orbit and
the broader the lines they emit. Between the accretion disc and the torus
are hydrogen clouds which produce the broad emission lines; above and below
the torus are the hydrogen clouds which produce the narrower emission lines.

In the unified scheme, every type of active galaxy can be explained
by the angle at which the object is being viewed. The view down the jet
gives a blazar. Moving away from the jet down to about 45 degrees allows
the observer to look down into the broad line region of the galaxy. Depending
upon the power of the central region, the observer will see a QSO or a
Seyfert type 1. Still farther off the axis of the jet, the broad line region
begins to be obscured by the dust torus. The narrow line region is still
clearly visible, however, and so the observer will see a Seyfert type 2.
If the active galaxy is a strong emitter at radio wavelengths, it will appear
as a radio galaxy from this orientation.

James Hough of the University of Hertfordshire, who researches different
types of active galaxy, favours a more modest version of this scheme. ‘There
are different levels of unification,’ he says. ‘Grand unified theories try
to encompass every type of AGN, both radio loud and quiet. There are also
smaller-scale unified theories which try to unify Seyfert 1s and 2s, for
instance. Really, those kind of schemes are the ones which most people feel
happier about.’ He thinks radio loud AGNs could be entirely different from
those that are radio quiet. ‘Radio loud AGNs seem to have elliptical galaxies
as hosts and radio quiet ones predominantly seem to be in spiral galaxies.
So it may be that you do actually end up with two distinct categories. The
grand unified schemes come back to the crucial issue: do the radio quiet
AGNs have a radio loud component that you just don’t see? My guess is that
they don’t and it just does not exist.’

Some of the most convincing evidence that black holes are the source
of the activity of active galaxies takes the form of phenomena that require
large gravitational fields and compact regions of energy generation. For
instance, the gravitational field of a black hole is required to cause matter
to be ejected from the AGN at speeds close to that of light. This bulk motion
of material can be observed by astronomers. Rapid variation in brightness
of the AGN is another property conventionally explained by black holes.
Telescopes are not yet powerful enough to pick out the AGN in active galaxies:
they simply see the whole thing as a blur. For that blur to brighten on
the timescale of, for example, a day means that the whole AGN region must
vary in brightness in that time. Since information can only travel at the
speed of light, the region emitting the radiation must be less than one
light day across – about the same size as our Solar System. This fits in
perfectly with the proportions astronomers calculate from theory for an
accretion disc around a black hole. However, both bulk motion and rapid
variability are confined to the minority of AGNs.

But not everyone accepts the black hole theory. Roberto Terlevich of
the Royal Greenwich Observatory is among the sceptics. He points out that
most active galaxies are of the radio quiet variety and believes they can
be explained by conventional astrophysics, without recourse to exotic ideas
about black holes. One of Terlevich’s collaborators, Brian Boyle of the
University of Cambridge, says their motivation is a simple one: ‘We know
young stars exist and we know massive stars exist. We do not know that
black holes exist. So why not try to explain the AGN phenomena with things
that we know exist rather than more esoteric things such as black holes?’
Terlevich, Boyle and others have developed a theory in which the energy
to power an AGN comes from bursts of star formation and supernova explosions
in young galaxies. The ideas are based on surveys of quasars, QSOs and galaxies
done in the past few years. These surveys have provided what astronomers
call the luminosity function for these objects, which tells astronomers
how the brightness is distributed among individual quasars, QSOs and galaxies.

Terlevich and Boyle put forward their version of the starburst hypothesis
because of some observations that no one had taken much notice of before:
‘The very high red shift quasars are observed to have a large amount of
metals,’ explains Boyle. ‘Metals’ is the term astronomers use for the elements
heavier than helium, which make up 2 per cent of the Universe. Nearly all
these elements are synthesised in the hearts of massive stars and then thrown
out into the interstellar medium by supernova explosions. Boyle continues:
‘That means that a significant amount of star formation has gone on. So
given that the rate of star formation must be very high, we asked ourselves
how bright would these young stars appear? When you work it out, the luminosity
of young stars alone is comparable to the luminosity you get from an active
galactic nucleus itself.’

In other words, energy comes from the stars themselves, and there is
no need for black holes. According to the details of their theory, the activity
and evolution of an AGN can be split into four phases. During the first
3 million years, the central region of the galaxy becomes a very bright
emission nebula as massive stars form and begin to shine thanks to the nuclear
fusion taking place at their cores. Some of the stars formed at the heart
of the galaxy are extremely massive – perhaps up to 100 times the mass of
the Sun.

A meeting of masses

Then the most massive stars of 40 to 60 solar masses or more begin to
develop into a class of object known as Wolf-Rayet stars. These are very
hot, massive stars which shine with between 100 000 and 1 million times
the brightness of the Sun. The prodigious amount of radiation released by
the star lifts its outer layers and expels them into space. These gently
expanding shells of hydrogen radiate energy because they are being excited
by the radiation from the star. The radiation they emit is confined to narrow
spectral lines because the shells are only expanding at relatively low velocities.
Thus the galaxy begins to take on the appearance of a Seyfert type 2. This
phase usually lasts for about 1 million years.

Before reaching the Wolf-Rayet stage, the stars undergo a highly unstable
phase which is believed to occur just before the onset of helium fusion
in their cores. The star Eta Carinae in our own Galaxy is currently at this
stage of development. At this point in a star’s life, large amounts of dust
are produced, which begin to obscure the central regions of the galaxy.

In the third phase, the galaxy maintains its appearance as a Seyfert
type 2. The star cluster producing the activity is now about 4 to 8 million
years old and the first supernovae are exploding. These supernovae are
the death throes of the truly massive stars in the cluster; they will all
have had masses more than 40 times that of the Sun. These explosions will
blow bubbles in the central regions of the galaxy and cause modest amounts
of radio emission at frequencies astronomers associate with radio quiet
AGNs.

Final explosions

The fourth and final phase in the evolution of the active galaxy is
the most dramatic. It begins when the cluster is about 8 million years old,
and continues for the next 50 million years at least. During this phase,
stars with between 8 and 25 solar masses explode as supernovae. There are
many more stars in this mass range than there are ones of over 40 solar
masses and the energy output of the central region increases dramatically.
Depending on the precise increase in energy output, the galaxy either begins
to display the characteristics of a Seyfert type 1 galaxy, or of a quasar.
As the supernova remnants expand through the dense central regions of the
galaxy they emit radiation in the ultraviolet and X-ray regions of the
spectrum.

Ken Pounds of the University of Leicester, is, however, an adherent
of the black hole theory. Using the Exosat satellite he has recently made
extensive studies of active galaxies in which he observed some which changed
their X-ray brightness over short periods of time. ‘We got really detailed
observations for about half a dozen Seyfert galaxies. Without exception,
they showed variability on timescales of tens of minutes to a few hours.
The output would often double in those times. It is very difficult to see
a mechanism by which this energy can be produced other than by accretion
on to a massive black hole. I think that is perhaps the best evidence against
the alternative supernovae model.’

Pounds sees no problem in the fact that not all AGNs show such rapid
variability, which is more evident in the lower-luminosity sources and less
pronounced in the more luminous AGNs. This is consistent with the black
hole hypothesis, he claims, because a more luminous source would be expected
to contain a bigger black hole and accretion disc which light would take
longer to travel across.

Supporters of the starburst model argue that long-term variability can
be explained by supernova flashes, the cooling of the remnants and instabilities
in the expanding supernova remnants. They are developing similar explanations
for short-term variability and flickering. One idea Terlevich has just put
forward is that short-term X-ray variability can be explained by rocks smashing
through the material ejected by the supernovae remnant. These violent occurrences
disturb the gas that is emitting the X-rays so much that flickering is produced,
he claims.

Boyle concedes that the starburst model does not yet have all the answers:
‘The radio loud properties of quasars are more difficult to explain.’ But
he points out that only one AGN in a hundred is a radio loud quasar: ‘We
consider it to be no great problem. If our model can explain 99 per cent
of the objects and we haven’t yet worked hard enough to explain the other
1 per cent, then we do not feel that it is too great a failing of the model
as such.’ He admits that the model cannot explain radio jets. He suggests
that what differentiates a radio loud quasar from a radio quiet one may
in fact be a black hole, since a black hole’s accretion disc provides a
mechanism for focusing the jets.

A frequent criticism of the starburst theory is that some QSOs radiate
more energy than starbursts can provide, but Boyle is adamant that this
is not so. A paper he and Terlevich published this year describes how calculations
developed by N. Arimoto and Y. Yoshi of the Tokyo Astronomical Observatory
can be used to show that elliptical galaxies can reach luminosities comparable
to even the brightest QSOs.

Evidence for early star formation in elliptical galaxies comes from
the old age of the stars in nearby ellipticals. Their star formation processes
should be reaching a peak at about the same time that QSOs become visible.
So could QSOs simply be the cores of elliptical galaxies in the process
of formation? Terlevich and Boyle think so. Working backwards from the luminosity
function of nearby ellipticals to calculate the rates of star formation,
they found that they could precisely reproduce the QSO luminosity.

Further support for their claims comes from Alexei Filippenko of the
University of California at Berkeley. His optical observations of a supernova
in spiral galaxy NGC 4615 in May 1987 produced some unexpected line profiles.
He thought these looked like a Seyfert type 1 nucleus or a QSO, and ascribes
the supernova’s unusual appearance to the dusty environment in which it
was embedded when it exploded. Such dusty environments are actually regions
where star formation is currently taking place, and so would closely resemble
the conditions in a starburst region. Filippenko went on to conclude that
had this object been observed in the nucleus of a galaxy, it would probably
have been classed as a Seyfert galaxy.

Edward Shayon and Dan Dowling of the University of Maryland used the
Hubble Space Telescope to look at active galaxy ARP220. They concluded that
it was powered by the massive starburst they discovered at its centre. More
recently, Clements and his collaborators have used ideas from the starburst
theory to explain the long wavelengths emanating from the apparently active
galaxy IRAS F10214+4725. The hydrogen emission is also in keeping with the
starburst interpretation. Nevertheless, other factors such as some emission
lines in the optical spectrum do not tie up. ‘At the moment, 10214 is a
bit of a special case,’ says Clements. ‘It was discovered on the basis of
its infrared emission, not by optical or radio emission, which is how every
other active galaxy has been discovered. I think that 10214 is essentially
a unique object.’ Clements suggests that IRAS F10214+4725 may be a dust-enshrouded
black hole surrounded by a starburst region. This leaves open the question
of which is producing the bulk of the luminosity.

A team led by Judith Perry of the Institute of Astronomy, Cambridge,
and John Dyson of the University of Manchester has been working on a hybrid
model for several years. For them, the stumbling block with the black hole
theory is the mechanism that fuels it, and hence the AGN. ‘A starburst region
is a natural way of containing the amount of matter you need to fuel an
active nucleus,’ explains Robin Williams, one of the team. ‘In 1985, Perry
and Dyson found that the stellar mass loss processes can naturally produce
shocks in which gas can cool and produce the broad line emission of many
AGNs. If a reasonable fraction of the mass which is lost from the stellar
cluster can be funnelled into the central black hole, then you have the
energy produced which powers the AGN.’ In this view, most of the AGN luminosity
does not come from the starburst region itself. Instead, this region provides
the matter which the black hole swallows; the release of gravitational energy
as this happens powers the activity.

Such a combination theory may turn out to be the most likely option.
Pounds does not discount starbursts as a possible expla-nation for some
AGNs: ‘We only have to look round our Galaxy to see the enormous impact
of supernovae, and there are increasing numbers of observations from nearby
galaxies of extended nuclear X-ray emission. There are regions within galaxies
where remarkable bursts of supernova activity are taking place. That could
certainly account for some of these active galaxies, particularly the lower
luminosity ones.’

Stuart Clark is a freelance astronomy writer.

* * *

Reading between the spectral lines

A prism separates white light into its constituent colours: red, orange,
yellow, green, blue, indigo, violet. This familiar pattern is known as a
continuous spectrum, often shortened to a continuum.

An absorption spectrum is produced if the white light travels through
some matter – a cloud of gas, say – before it is passed through a prism.
The atoms in the gas absorb certain wavelengths of light more strongly than
others, leaving dark lines in the continuous spectrum.

If the gas cloud is viewed at an angle away from the white light source,
however, the spectrum will not be a continuum at all but a sequence of coloured
lines. These emission lines will be at the same frequencies as the dark
lines in the absorption spectrum, since they are produced by the same atoms
in the gas. If the darkness of absorption lines, or the brightness of emission
lines, are traced on a graph, they rise in curves depending upon how fast
the gas is moving.

A gentle, hill-shaped curve is produced when gas is swirling around
at high velocities – the broader the line, the faster the emitting gas.
In the spectrum of an AGN, the broad emission lines are produced by swirling
clouds of gas, while the narrow lines are produced by those moving more
sedately.

If the galaxy containing the cloud is moving away from the observer,
then the spectral lines will be shifted towards the red end of the spectrum.
This ‘red shift’ is caused by space between the galaxies stretching as
the Universe expands. The broadening of spectral lines is a manifestation
of the Doppler effect.