杏吧原创

PLANETS huddle in cosy families around the light and warmth of their parent
stars. At least that鈥檚 true of all the planets we know about鈥攖he nine in
our own Solar System, and eighty or so that we have spotted orbiting nearby
stars. But in fact family life might be the exception. Most planets could be
orphans, wandering in the cold and dark between the stars.

True, there are still no definitive sightings of these lonely planets, but
over the past three years various clues from theories, simulations and finally
observations of one distant star cluster all seem to point to their
existence.

The most astonishing evidence appeared this June, in a paper by Kailash Sahu
of the Space Telescope Science Institute in Baltimore (Nature, vol 411,
p 1022). Sahu and his team pointed their telescope at the Galactic bulge, the
central region of our Galaxy. Not that they were interested in the bulge
itself鈥攖he real attraction was a cluster of stars called M22, which lies
between the bulge and the Earth.

The astronomers were looking for any evidence of hidden matter in the cluster
betraying itself by 鈥済ravitational lensing鈥�, in which the light of a more
distant star is temporarily magnified by the gravity of an intervening object.
Other such searches have always turned up dim, low-mass stars a few tenths the
mass of the Sun.

The big surprise was that whereas they found only one gravitational lensing
event by a star, there were six from much lighter objects. In all six there was
no gravitational lensing effect from a heavier companion, so these light objects
must have been far from any star. What could they be but free-floating
planets?

Six may not sound like a lot, but it hints at a huge number of planets in
total. Because a planet has a much weaker gravitational field than a star, it
has to stray far closer to your line of sight to produce a detectable lensing
effect. As the planets produced six times as many events as stars, there must be
far more than six times as many planets. If they have around the mass of
Jupiter, that means roughly 60 free-floating planets per star. If they are more
like Earth, then we are talking about 600 free-floating planets per star.

Where could all these lone planets have come from? The conventional picture
is that planets are born in the dense clouds of gas and dust that surround young
stars. If that is the only way they can form, then the free planets in M22 must
somehow have been ejected from their nursery.

The prime suspects are other stars. The gravity of an interloper star could
rudely turf some planets out into space if it comes close enough. Calculations
show that the crucial distance is about five times the distance of the outermost
planet鈥攕o any star coming within five times the distance to Pluto, or
about 30 billion kilometres, could eject planets from our Solar System. Even
that distant an approach would be very rare in the Sun鈥檚 neighbourhood, but in a
globular cluster such as M22, close stellar encounters are frequent because the
stars are so crowded.

Hurley and his colleague Michael Shara have simulated this process on a
specially designed computer. 鈥淕RAPE鈥� machines鈥攕hort for GRAvity PipE
line鈥攚ere developed for such astrophysical simulations by Jun Makino and
his group at the University of Tokyo. Purpose-built for a single task, they turn
in world-beating performance for relatively low cost.

The 100,000 stars of a typical globular cluster are still a bit much to
handle, so Hurley and Shara simulated an open cluster, a bit like a globular
cluster but with a more manageable number of stars鈥攖heir simulations used
10,000. Each had a mass between 0.1 and 100 times that of the Sun, and 1 in 10
stars was given a single Jupiter-mass planet. They then ran their simulation for
the equivalent of 4 billion years, the age of the Solar System.

What they found was that 30 per cent of the stars left in the cluster were
stripped of their planets. 鈥淚n globular clusters, the fraction could easily be
higher than 50 per cent,鈥� says Hurley.

But as we don鈥檛 live in a globular cluster, the big question is whether
free-floating planets are common outside those crowded stellar environments, in
the Milky Way at large. The answer could well be yes, says Hurley, since most
stars are born in large batches from giant clouds of gas and dust. In these
stellar nurseries, encounters between stars could easily free young planets,
creating a significant population of planets roving the Galaxy.

Meanwhile, a planetary scientist at the California Institute of Technology in
Pasadena has another reason to believe free-floating planets could be out there.
David Stevenson thinks that young planets might also toss out their
siblings.

According to Stevenson, some simulations of planet formation produce not one
or two Earth-mass planets in each solar system, but more like 10. Most get
swallowed by their parent star or ejected into interstellar space by close
encounters with embryonic giant planets. 鈥淚f 1 in 10 stars has a planetary
system and 10 planets are lost from each, then you get as many free planets as
stars,鈥� says Stevenson (Nature, vol 400, p 32).

The observations are equivocal. In 1999, Patrick Roche of Oxford University
and Philip Lucas of the University of Hertfordshire claimed to have seen
free-floating planets in the Orion Nebula鈥攁 relatively nearby star-forming
region. But earlier this year, a team led by Mark McCaughrean of the
Astrophysical Institute Potsdam, Germany, showed that the free-floating specks
were probably too heavy to be planets鈥攎ore likely they were very small
stars called brown dwarfs.

In 2000, researchers saw more suspicious specks in another young star
cluster, Sigma Orionis, but again no one is sure whether these are planets or
brown dwarf stars.

But even if the observational evidence is limited, the simulations seem
clear-cut: if either marauding stars or cuckoo siblings are at work, then the
whole of our Galaxy must be filled with dark planets, wandering space for
eternity. It鈥檚 very unlikely that they would ever get recaptured by other stars.
Encounters would be rare, and in any case wandering planets would almost
certainly fly straight through any solar system without meeting any
resistance.

But these wanderers might still make an impression. If a planet passed
through the cloud of dormant comets surrounding our Solar System, its gravity
could send some of them falling inwards. Some scientists blame comets for mass
extinctions of life on Earth. We may owe some of these episodes to the petulance
of orphaned planets.

And the M22 result could have even more profound implications, forcing
planetary scientists to rethink some of their most cherished ideas. Remember
that the number of lensing events implied tens or hundreds of free planets for
every star. And if roughly 50 per cent of planets are stripped from stars, you
have to double those figures to get the number of planets each star started out
with鈥攑erhaps a thousand planets per star. 鈥淚t鈥檚 an extraordinary result,鈥�
says Hurley.

Stevenson鈥檚 alternative source of free planets could go some way to
explaining the shortfall, but certainly not all the way. The numbers are just
too huge. Could it be that our Solar System is unusual in having only nine
planets, and 100 or 1000 is the norm?

According to Hurley, Japanese astronomers have run simulations that make
these numbers more plausible. Shigeru Ida of the Tokyo Institute of Technology
and Eiichiro Kokubo of the National Astronomical Observatory have shown that in
a condensing planetary system where the density of rocky material is low, many
low-mass bodies like the Earth can form, perhaps as many as 50 to 100. But these
simulations assume that heavy atoms are reasonably plentiful鈥攖hey鈥檙e
thought necessary to form planetary cores. M22, though, has a low abundance of
heavy atoms鈥攋ust a 50th that of the Sun.

鈥淭he M22 observations are a real puzzle,鈥� admits Steinn Sigurdsson of
Pennsylvania State University. 鈥淭he amount of mass in planets is extraordinarily
large. I鈥檇 almost have to say that, to account for so much mass, planets would
have to form as free-floaters, without host stars.鈥�

Stevenson agrees. 鈥淓jection might not be necessary,鈥� he says. 鈥淭he
planet-mass bodies might have formed directly, exactly the same way most stars
form, by the collapse and fragmentation of a cloud of gas and dust, in which
some of the fragments are very small.鈥�

Few have considered this scenario. But earlier this year Alan Boss of the
Carnegie Institution in Washington DC published the results of a new model of a
collapsing gas cloud. Crucially, he included the effects of magnetic fields. The
field can act like a spring, stopping the cloud from collapsing fully and
forming a single massive object at its core, so that it rebounds and breaks up
into several smaller bodies instead. Boss wouldn鈥檛 call these objects
planets鈥攑referring to reserve that word for things that form around
stars鈥攂ut his simulation could explain the strange findings from M22.

On the other hand, M22 may be a dud. Because it was so extraordinary and
unexpected, we should take the result with a pinch of salt until someone else
confirms it.

The situation is complicated by another globular cluster called 47 Tucanae.
Observations of nearby stars have turned up several giant planets orbiting
extremely close to their parent stars. But 47 Tucanae seems to have none of
these close-in Jupiters. A large team co-led by Sigurdsson used the Hubble Space
Telescope to look for evidence of such planets passing in front of their parent
stars. They expected to see about 20 close-in Jupiters with orbital periods of
less than five days. 鈥淲e found none,鈥� says Sigurdsson. So did the stars of 47
Tucanae form no planets at all? And does that call into question the M22
result?

Because 47 Tucanae has a higher abundance of heavy atoms than M22, it ought
to have a better chance of containing planets. But it also has an exceptionally
high density of stars. 鈥淭hey are so close together that it鈥檚 possible that the
intense heat from young, massive stars disrupts protoplanetary discs so no
planets form in the first place,鈥� says Hurley. Then again, maybe the lack of
close-in giant planets can be explained if stars in 47 Tucanae were stripped of
all their planets. That would strengthen the case for orphan worlds.

If there really are so many planets drifting between the stars, it seems a
terrible waste. After all, in the cold and dark of interstellar space, such
planets must surely be lifeless.

Not necessarily. These worlds will produce their own heat from the decay of
radioactive elements such as thorium-232 deep in the planetary core, just like
on Earth. Though their surfaces may be frigid, deep within them there could be
enough warmth and energy to sustain primitive life.

And Stevenson believes that his particular brand of interstellar planet might
be even more hospitable. Ejected early on from their planetary systems, they
will be swathed in a thick mantle of molecular hydrogen, the stuff of the
protoplanetary nebula. The blanket of hydrogen would press down on the surface,
creating a pressure 1000 times that on Earth.

This will act as a potent greenhouse gas. According to his calculations, it
could trap radioactive heat so well that it would keep the surface at room
temperature for many billions of years. 鈥淲ith the greenhouse effect operating,
it is quite possible for an interstellar planet to stay warm for at least 10
billion years鈥攖wice as long as the Earth has so far existed,鈥� says
Stevenson. Beneath such an atmosphere there might easily be oceans of liquid
water.

The lack of starlight as an energy source from life could be a problem.
According to Stevenson, the energy available from radioactive rocks is only a
10,000th that available from the Sun on Earth. Nevertheless, he sees no
fundamental reason why primitive micro-organisms could not harness it somehow.
鈥淚t could be that most of the life in our Galaxy is on dark interstellar planets
and not on sunny worlds like our own,鈥� he says. 鈥淚nterstellar planets could be
the most common abodes for life in the Galaxy.鈥�