ÐÓ°ÉÔ­´´

Astronomers are mightily puzzled. They know that stars and the dense
clouds where stars are born are rich sources of chemical building blocks
– molecules, atoms and ions. But what lurks in the colder regions of near-empty
space between stars – regions vast enough for organic molecules to wander
for decades, perhaps centuries, without meeting and reacting with another
one? Astronomers have come up with various explanations, ranging from dust
grains and naked DNA or viruses to cellulose, chlorophyll and more recently,
buckyballs. Now a trail of spectroscopic clues stretching back over seventy
years is leading them closer to the true identity of the materials that
make up this diffuse interstellar medium.

The thrill of the chase for scientists, many of whom met in May in Boulder,
Colorado, for the first international conference on the topic, comes from
the possibility that these mystery molecules might be novel chemicals. What
makes this puzzle all the more compelling is that these molecules may provide
the molecular seeds for life, not only on Earth, but throughout the Universe.

To probe the chemical composition of stars and interstellar material,
astronomers use telescopes fitted with spectroscopic detectors. These separate
starlight into its spectral components and record the intensity of the light
at each wavelength. This information is plotted as a series of peaks, or
‘bands’, of intensity against wavelength. The bands indicate the wavelengths
at which atoms and molecules in space have either emitted or absorbed electromagnetic
radiation.

Atomic fingerprinting

Emission spectra are clearest and strongest as they are recorded directly.
Absorption spectra are typically produced when radiation from a star, for
example, passes through cooler matter which absorbs certain wavelengths
of radiation. These wavelengths show up as dark lines or bands superimposed
on the emission spectrum of the star. Every atom and molecule absorbs and
emits radiation at specific wavelengths, depending on its chemical composition.
So the bands, whether absorption or emission, represent a unique ‘fingerprint’
of the atom or molecule. To identify atoms and molecules in space, researchers
compare the height, width and position of these bands, recorded across the
electromagnetic spectrum, with the fingerprints of known molecules.

As early as 1919, astronomers noticed a puzzling series of unusually
broad absorption bands at visible and near-infrared wavelengths, superimposed
on the emission spectra of certain stars. The best studied is HD 183143
in the constellation of Sagitta in the northern hemisphere. Its starlight
should show a smooth distribution of intensity across the spectrum, but
from some perspectives there are dips in the signal – a loss of light at
various wavelengths. These dips point to some mystery matter in the diffuse
clouds of interstellar space, lying between the star and the observer. The
pattern of this spectrum, called the diffuse interstellar bands (DIBs),
provides the best clue to the chemical make-up of the diffuse interstellar
medium. But despite decades of intense effort since these first observations,
not a single DIB has been matched with a molecule. The lack of any absorption
spectra for chemicals under interstellar conditions that could verify the
identity of these molecules forced most astronomers to give up and move
on.

George Herbig of the University of Hawaii in Honolulu is one who has
stuck it out for over thirty years. ‘I’ve been at this more years than I
care to think about,’ he says. ‘There have been so many good possibilities,
but they’ve all slipped away.’ If anything, the DIBs problem has grown even
more daunting. Thanks to the development of electronic detectors with greater
resolution and accuracy, the past decade has seen a great improvement in
astronomers’ ability to distinguish absorption bands originating in the
diffuse interstellar medium from those that represent the background ‘noise’
of molecules in stars and the Earth’s atmosphere. But this advance has also
resulted in an explosion in the number of mystery bands detected.

Fifteen years ago there were 35 DIBs. Today there are 150 well-known
ones, at wavelengths ranging from the blue part of the visible spectrum
to the near infrared. And they are still being discovered; another 40 were
announced at the Boulder conference.

In comparison, the dense clouds where most molecules in space are thought
to originate seem well understood. Such clouds include the ‘molecular’ clouds
in the spiral arms of galaxies where stars are born. During its life, each
star ejects huge quantities of elements. In the dense clouds, ions and simple
molecules often meet and react on the surface of dust grains. At the edges
of the clouds, atoms and molecules are often struck by ultraviolet radiation,
which knocks out some of their electrons to produce ions. These can combine
with molecular hydrogen or with other simple molecules to form more complex
chemicals. So far, at least 80 kinds of polyatomic molecule have been found
in dense clouds, ranging from simple diatomic species, such as hydrochloric
acid, to 11-carbon chains and biochemically important chemicals such as
amino acids, ammonia and formaldehyde.

Patrick Thaddeus, a radio astronomer at Harvard University’s Center
for Astrophysics in Cambridge, Massachusetts, has helped to identify as
many as a quarter of them. ‘Discovering the architecture of these materials,
which are hammered out by synthetic processes in space and sculpted by starlight,
is a very important physical and chemical problem,’ he says. It is one that
could have a direct bearing on the question of how complex the starting
materials for life on Earth were. Did they build up to complex structures
on Earth? Or are complex organic molecules wandering around in space, prepackaged
and ready to rain down and start life on hospitable planets? ‘There may
have been, in the very beginning, much more chemical complexity than was
ever believed to exist,’ says Thaddeus.

Whatever these molecules are, astronomers expect them to be different
from other molecules in space because conditions in the diffuse interstellar
medium are particularly severe. Blazing starlight, unimpeded by dust, would
continually bombard any molecules with ultraviolet photons. This would be
enough to tear simple molecules into atoms or other fragments. Bigger molecules
would have their electrons knocked out and be converted into ions.

The diffuse interstellar medium is likely to contain molecular ions
or free radicals – highly reactive groups of atoms having unpaired electrons
in their outer shells. ‘The shocking thing is that molecules survive at
all in the diffuse medium – it’s too harsh an environment,’ says Louis
Allamandola, an astrochemist at the NASA-Ames Research Center in Moffet
Field, California. But scientists are convinced not only that they do survive,
but that they are homing in on their quarry.

A decade ago, half of all interested astronomers would have argued that
DIBs are caused by solid grains of carbon. Today most believe that the culprits
are probably carbon-based molecules in the gas phase. Several pieces of
circumstantial evidence brought about this shift in thinking.

Magic molecule?

One was the discovery of carbonaceous molecules in the dense clouds.
If they are synthesised in interstellar space, perhaps the DIB molecules
originate there too, even if they later change their form while in the diffuse
interstellar medium. Another was that, even though the quality and quantity
of DIBs data have increased, the peaks have remained consistently broad
and without any substructure. This implies that the DIBs are caused by big
molecules, perhaps complex carbon-based molecules, rather than by dust grains
which would have produced DIBs of varying size. One thing that is certain
is that the DIBs are not due to one magic molecule, as many researchers
once believed. A single kind of molecule simply could not produce 127 absorption
bands.

In 1987, astronomers found that the intensity of the DIBs changes as
their perspective, or ‘line-of-sight’, shifts. Most intriguingly, the size
of certain bands grows or shrinks in unison. This led astronomer Jacek Krelowski
of the Nicholas Copernicus University in Torun , Poland, and others to propose
that these groups of DIBs represent families of related molecules. Krelowski
suggests that each chemical family requires the same set of physical conditions
to survive in the diffuse interstellar medium, so their concentrations vary
together as their local environment changes. But what sort of chemical families
were they?

Important clues have now emerged from experiments in the laboratory.
Chemists have made large, complex carbonaceous molecules and measured their
absorption spectra under the harsh conditions that approximate to the interstellar
medium – extremely cold, only a few degrees above absolute zero, bright
ultraviolet light, and widely spaced molecules. Late last year, John Maier
and his coworkers at the University of Basel in Switzerland found evidence
that the mystery molecules are families of carbon radicals, chains which
are rich in triple bonds, or ‘unsaturated’. While the idea is not new –
it is plausible given the abundance of hydrogen and carbon in the Universe
– it has remained untested because of the difficulty of studying such highly
reactive radicals under ‘interstellar’ conditions.

Working with ions at just 5 kelvin, Maier’s group made unsaturated,
negatively charged carbon chains 2 to 12 carbon atoms long and mixed all
the even numbered ones (which just happened to be the ones they could select)
with neon, an inert chemical, to prevent the chemical species reacting.
Next, the researchers shone white light through a solid mixture of carbon
chains trapped in frozen neon and measured the absorption spectra. They
found strong matches with 15 DIBs in the visible and near-infrared wavelengths.

Although Maier admits he cannot tell specifically which carbon chain
is responsible for which band, he says the overall pattern and the high
number of coincidences between the observed bands and the bands measured
for the carbon chains in the laboratory is significant. ‘It’s not just a
one band-one molecule match,’ he says. ‘We seem to have picked out a subset
of related molecules that are out there.’ Since publishing these results,
the Basel group has made additional matches. Their evidence suggests that
a significant fraction of molecules in the diffuse interstellar medium may
be unsaturated carbon chains consisting of 6 to 16 carbon atoms: ‘We may
have identified up to 20 bands, or perhaps 15 per cent of all DIBs,’ says
Maier. ‘That would be quite a lot and a good start.’ Harold Kroto, professor
of chemistry at the University of Sussex, says this breakthrough is causing
great excitement in the research community. ‘The experiment looks right,
feels right and makes chemical sense,’ he says.

Searching the ultraviolet

In line with previous astronomical observations, Maier found no absorption
lines in the ultraviolet part of the spectrum. But recent studies, results
of which were presented at the Boulder conference, show that they do exist.
Jason Cardelli and his team from the University of Wisconsin at Madison
have definitive evidence, in the shape of data collected by the Hubble Space
Telescope, for a DIB in the ultraviolet (UV). While searching for clues
about the abundance of heavy elements in the interstellar medium, they came
across an unusually broad and asymmetric band at 1369 angstroms that they
could not match to any known element or molecule.

Cardelli believes there are two reasons why past observations failed
to iden-tify DIBs in the UV. First, instrumentation before Hubble did not
provide the required spectral resolution. Secondly, researchers may have
expected DIBs in the UV to look similar to those in the visible. But as
Cardelli points out, they appear much weaker and narrower.

If the carbon chain theory does turn out to be correct one, where could
such molecules come from? Thaddeus says such unsaturated carbon chains could
have developed from the simpler carbonaceous molecules known to exist in
the dense clouds. He proposes that the chains grow and dissociate under
the impact of starlight by a mechanism similar to that operating for carbon
clusters in an arc – the method chemists routinely use to make buckyballs.

This technique involves passing an electric current between two graphite
rods in a plasma reactor containing helium. The vaporised carbon (soot)
contains fullerenes. The carbon atoms tend first to build up into longer
and longer chains before forming rings, bow ties and eventually football-shaped
fullerenes.

Some researchers, however, argue that some DIBs may be caused by carbonaceous
molecules of a more elaborate ring-based design. The strongest candidates
are the polyaromatic hydrocarbons, or PAHs, and are commonly found on Earth
in the blackened tips of barbecued steaks, car exhaust and cigarette smoke.
The chemicals have also been found in meteorites, and infrared emission
spectroscopy suggests they may inhabit some molecular clouds, such as the
one known as the Orion Nebula in the glowing region of Orion’s sword tip.
This is part of a dense cloud that is completely shielded by dust from UV
radiation and where molecules are abundant. ‘If the PAHs are so stable and
in so many objects, there’s no reason to think they wouldn’t be found in
the interstellar medium,’ Allamandola of the NASA-Ames Research Center asserts.

It is 10 years since PAHs were first proposed as the source of DIBs,
but only recently have researchers obtained spectroscopic evidence to support
this theory. Last year a group led by Allamandola reported that they found
close matches between DIBs and the absorption spectra of two PAHs, naphthalene
and pyrene. In a technique similar to that of the Basel group, Allamandola’s
team trapped pyrene and naphthalene, in frozen neon and argon matrices at
temperatures of a few kelvin. The researchers then mimicked starlight by
shining UV light on the materials to create ions. The matrix holds the ions
close enough together so that their spectra could be measured, but far enough
apart so that they do not react. The absorption spectra indicated that positively
charged pyrene molecules are the cause of the DIB at 4430 angstroms, in
the visible blue region of the spectrum (see New ÐÓ°ÉÔ­´´, Science, 13
March 1993).

Despite these advances there can be no final verdict on what produces
DIBs, says Herbig, until researchers find out exactly how much their neon
or argon matrices are shifting the position of the absorption peaks. Some
argue that the matrix shift is negligible enough to give significant matches
between experiment and observations of DIBs. But most researchers do not
consider the work to be definitive proof. Moreover, says Herbig, a candidate
molecule must have several absorption peaks that match the positions and
relative sizes of observed DIBs to rule out the effect of chance matches.

Absorbing light

An improved experimental technique being developed by the Basel group
and a team at the University of Colorado and the Boulder office of the National
Oceanic and Atmospheric Administration may provide the clincher. This method
dispenses with the matrix, allowing scientists to directly measure the absorption
spectra of molecular ions in the gas phase at interstellar temperatures.
But with this method large obstacles remain. The biggest is making enough
molecules to take a measurement. ‘Because ions are such fragile species
and so reactive, it’s hard to maintain enough molecules to do an absorption
experiment on them,’ explains Veronica Bierbaum, a chemist at the University
of Colorado. Her group is having some success at generating positively charged
PAH ions, she says, but it still needs concentrations 10 000 times higher
before it can measure the absorption spectra using the best instruments
available.

Meanwhile, other candidates are beginning to appear. In May two scientists
working in the Netherlands, B. Foing at the European Space Agency and P.
Ehrenfreund at the Leiden Observatory, announced they had discovered evidence
that positively charged C-60 molecules could account for some DIBs. The
researchers who discovered fullerenes in the mid-1980s were trying to crack
the DIBs puzzle. Other fullerene molecules also remain candidates because
they are likely to form in space, especially in hot carbon stars, and are
stable enough to withstand UV bombardment in the diffuse interstellar medium.
There is the possibility that they have not been detected there yet: ‘If
C-60 comes out of a star, it will have all kinds of things stuck to it because
it’s a perfect reagent. And we don’t know how that would modify its spectrum,’
explains Kroto.

In the end, solving the DIBs problem will probably require several approaches
and the answer will be complex, says Adolf Witt, an astrophysicist at the
University of Toledo in Ohio. ‘I don’t picture the diffuse interstellar
bands as being the result of something that you can write a molecular formula
for, but instead as clusters of primarily amorphous hydrocarbons that don’t
reveal their presence in easily read lines,’ he says. The diffuse interstellar
medium could contain a stew of all the proposed chemicals: unsaturated carbon
chains, PAHs, fullerenes, chlorin and other porphyrins, as well as amorphous
hydrocarbon clusters. ‘It’s hard to say what these molecules will turn out
to be,’ says Herbig. ‘They might be something so exotic that even the chemists
don’t know them. I imagine the answer will come along and we’ll all say:
why didn’t we think of that earlier?’

Karen Schmidt is a writer based in North Carolina.