ÐÓ°ÉÔ­´´

Geology began as the study of the Earth. As human endeavour has reached
out into space, the earth sciences have expanded to include other planetary
bodies. First to attract geologists’ attention were our Moon and the other
terrestrial planets – Mercury, Venus and Mars. These are all similar to
the Earth in composition, with a dense core of metallic material overlain
by a mantle of silicate rock. But these may not be the only places to look
for the processes that shape the surface of a planet. The bodies that orbit
the outer planets much farther from the Sun are just as interesting and
there are more of them. Although most are made of ice on top of a rocky
core, these worlds carry evidence of processes remarkably like those on
the terrestrial planets.

Voyager 1 provided our first detailed views of the outer planets and
their moons in 1979, when it passed Jupiter’s four large satellites. Before
the images arrived on Earth, most people expected these moons to be dead
places, looking like our own Moon with its innumerable craters. Only the
largest rocky planets were supposed to be geologically active, with volcanoes,
faults and maybe even plate tectonics.

Researchers did not believe that there were any processes that could
renew the surface of a relatively small, solid planetary body. They reasoned
that without continual renewal, the surfaces of such bodies should have
become densely covered with craters as a result of impacts by meteorites
and comets over the 4.5 billion years that they have existed.

Although Voyager’s pictures soon proved them wrong, there was logic
behind the researchers’ thinking. Heat is the source of the energy that
is expressed as geological activity at the surface of our planet. Volcanoes
can be expected only if a planet or a moon is hot enough to produce molten
rock at shallow depths. For such a body to be tectonically active, with
widespread faulting and other deformation at the surface, it must not only
be hot inside but its temperature must increase with depth quickly enough
to force the interior to transport heat outwards by convection, as well
as by conduction.

Conduction is the flow of heat through a material, whereas in convection
the hot material itself flows. In Earth and Venus, below their rigid outer
shell or lithosphere, the rest of the mantle is solid but weak enough to
flow, carrying heat, at speeds of around a centimetre per year. The flow
enables the plates into which the Earth’s lithosphere is broken to move
around and is also a direct cause of volcanic activity in places, such as
Hawaii, that lie directly above rising plumes of hot mantle material. Earth
has seven major plates, giving it a distinctive style of plate tectonics.
Venus appears to have many smaller plates, making it rather different from
the Earth in detail but broadly similar in that it has young, and probably
still active, volcanoes and fault belts.

Earth and Venus are so hot inside because the rocks from which they
are built contain radioactive isotopes, which release heat when their nuclei
decay. Many radioactive isotopes exist naturally on Earth, and four in particular
are important heat sources today: potassium-40, uranium-235, uranium-238
and thorium-232. These isotopes have very long half-lives (the time it takes
for half the nuclei of a given isotope to decay, which is used as a measure
of its persistence), of around a billion years. Although they have been
decaying since the planet formed, enough radioactive material has survived
to keep the large terrestrial planets hot through radiogenic heating.

The smaller terrestrial bodies – Mars, Mercury and the Moon – have been
geologically dead, or at least comatose, for a very long time. This is not
because their mantles contain weight-for-weight less of the important radioactive
isotopes than the larger planets, but simply because they have greater surface
areas in proportion to their volumes. The heat generated by radioactive
decay within them escapes by conduction without setting off convection,
although these bodies were evidently more active in the past when these
isotopes, and some short-lived ones, now vanished, were more abundant.

But what of the satellites of the outer planets? They are less dense
than Earth, and cannot have the same composition. The densities of most
of them indicate that they are composed of a mixture of roughly 60 per cent
ice and 40 per cent rock. The presence of ice was first demonstrated by
spectroscopic analysis with telescopes nearly 40 years ago. Modern infrared
measurements estimate surface temperatures ranging from about -170 °C
on Jupiter’s moons to -235 °C around Neptune. Under such conditions,
ice at the surface of one of these moons is far too strong to flow as glacier
ice does on Earth; instead it behaves like rock. This gives these satellites
outer layers that are comparable to the Earth’s lithosphere.

Ice does not contain any of the radioactive isotopes that provide the
Earth’s main source of heat and the present rate at which radioactive decay
could produce heat in the rocky parts of icy satellites is probably too
low to soften and mobilise ice, let alone rock, except in the deep interiors
of the largest of them. In the distant past, when radioactive isotopes were
more abundant in their rocky components, radiogenic heating could have been
enough to allow the denser rocky material to sink through the ice to form
a distinct core. But shallow convection of ice driven by radiogenic heating,
powerful enough to lead to icy volcanism and surface deformation, cannot
be contemplated for any icy satellite during the most recent 90 per cent
of its life.

So imagine the surprise when Voyager 1 sent back pictures from Io, the
innermost satellite of Jupiter, showing volcanoes in the process of erupting.
And this was not all. The next satellite out, Europa, was revealed as having
a young icy surface, cracked like an eggshell and with only half a dozen
recognisable impact craters. Its neighbour, Ganymede, was seen to be crossed
by belts of curious grooved terrain, clearly the result of some sort of
tectonic activity, although they must be old because even the youngest belts
of grooves are quite heavily cratered. Only the outermost of Jupiter’s large
satellites, Callisto, has the sort of drab, cratered surface that almost
everyone expected.

The volcanoes on Io are a spectacular mixture of vents that form plumes
spraying sulphur or sulphur dioxide a hundred kilometres or more into space
and volcanic craters filled with hot molten material. Recent flows of material
can be traced for considerable distances away from Io’s volcanoes, and there
has been much debate over whether these ‘lavas’ are made of silicate rock,
like lava on Earth, or sulphur. It is probable that molten sulphur and molten
rock have both played important roles in Io’s recent volcanic history. Io
is a rocky exception to the icy rule among the satellites of the outer planets,
being slightly denser and more massive than the Moon. As on the Moon, Mars
and Mercury, its radiogenic heat production should have declined to a level
too low for volcanism today.

Perhaps the only people not surprised by Io’s current activity were
Stanton Peale at the University of California in Santa Barbara and his colleagues
at the NASA Ames Research Center, also in California. Blending scientific
prowess and impeccable timing in equal measure, they published a paper in
the journal Science a few days before Voyager 1 reached Jupiter, in which
they predicted that Io would be very hot inside, perhaps even molten, as
a result of tidal heating. This happens by the same mechanism that drives
tides in the oceans on Earth, the gravitational attraction between planet
and moon.

Tidal heating of Io’s interior comes about because it is deformed by
the gravitational pull of the planet; this may supply enough energy to power
geological activity at the surface. The gravitational pull of Jupiter raises
a tidal bulge about 1 kilometre high in the centre of the side nearest to
the planet. If Io were rotating more rapidly than it orbits the planet,
this tidal bulge would have to move backwards around Io to stay opposite
the planet. This cannot easily happen, and resistance to movement translates
into a drag on the rotation of the moon.

Such tidal resistance has slowed down the rotation of almost all satellites,
Io and our own Moon included, until they rotate at the same rate as they
orbit the planet. This phenomenon, known as ‘captured rotation’, results
in the satellite keeping the same face permanently towards its planet, and
is the reason why we only ever see one side of the Moon from Earth.

While a satellite is being brought into captured rotation, the stresses
that tend to keep the tidal bulge in position opposite the planet, would
be huge, and the resulting dissipation of energy within the interior of
the satellite would probably be enough to start internal convection and
widespread melting. But all this comes to an end very early on, almost certainly
by the time the satellite is 10 million years old. Once a moon settles into
captured rotation, there appears to be no reason for further heating and
little chance of further geological activity.

When a planet has several satellites, gravitational interactions often
make the orbits of neighbouring satellites simple multiples of one another,
a phenomenon called orbital resonance. Around Jupiter, for example, for
every orbit of Ganymede, Europa completes two orbits, and Io four. This
means that Io overtakes Europa at the same points in its orbit each time.
The regular gravitational tug from Europa pulls Io’s orbit out of shape,
and Io’s orbital speed varies accordingly, while its axial spin continues
at a constant rate. The result is that, seen from Jupiter, Io’s rotation
alternates between being slightly ahead of and slightly behind its average
position.

In consequence, the height and location of the tidal bulge on Io varies
as it fights to stay exactly opposite Jupiter. It is this flexing that generates
the heat that powers Io’s volcanoes. Of course, nothing is for free, and
orbital energy – controlling the satellite’s speed and distance from the
planet – is lost at the same rate as the interior of Io gains heat.

The effect is not confined to Io. Europa’s very young icy surface is
a result of tidal heating through orbital resonance with Ganymede. Ganymede
itself, though dead now, was kept active by tidal heating for considerably
longer than Jupiter’s outermost large satellite, Callisto. Orbital resonance
between Saturn’s satellites is responsible for young (and possibly current)
activity on Enceladus.

Present or past tidal interactions can also account for episodes of
enhanced heat production within the seven or so other icy satellites that
have evidently had long and complex geological histories. The most remarkable
of these lie among the satellites of Uranus and Neptune. Here there are
clear signs of volcanism, with vents and lava flows that look for all the
world like volcanic features on the Earth and Moon, except that they must
be made of ice.

Although the sources of heat differ on the terrestrial and icy bodies,
the results are remarkably alike. The similarity between the contorted surface
of the icy moons and rock formations on Earth arises from the properties
of their unusual varieties of ice. The composition of the ice on different
bodies arises from where they condensed within the spinning cloud of gas
and dust that formed the Solar System. Among the satellites of Jupiter the
ice is likely to be frozen water contaminated by dissolved salts, notably
sulphates of magnesium and sodium, that would have been leached out from
the rocky parts of these bodies. At Saturn or beyond, the progressively
lower temperatures while the satellites were forming would have allowed
volatile substances such as ammonia, methane, methanol and nitrogen to condense
as well as water. These molecules would become trapped within crystals of
water-ice, and in some cases would even have formed distinct ices. Experimental
work on these exotic mixtures is indicating that they give ice a complex
set of melting and crystallisation properties which mimic those seen in
the mixtures of silicate minerals that make up rocks .

In particular, when ice mixtures melt they do so in a way known as ‘partial
melting’. For example, ice formed from water (H₂O) mixed with
ammonia (NH₃) consists of an intergrowth of two distinct types
of crystal: pure water-ice and ammonia hydrate (NH₃. H₂O).
When this mixture of crystals melts, at a pressure low enough to be comparable
to conditions near the surface of the icy moons, all the ammonia hydrate
crystals and some of the water-ice crystals melt first. This produces a
fluid known as a ‘partial melt’, consisting of approximately 33 per cent
ammonia mixed with 67 per cent water. It leaves behind a solid residue of
water-ice crystals. Experiments show that this melting happens at -97 °C,
so melting is comparatively easy. Because the melt contains a higher proportion
of ammonia than the ice from which it formed, when it freezes again the
ice formed is also richer in ammonia than the original ice.

Experimental data are still scarce, but it seems likely that the presence
of more than one contaminant in the original ice, which is the likely case,
should result in even more complicated melting behaviour, especially considering
the different compositions of melt that would form at higher pressures.
In this way, a volcanically active icy satellite could develop a series
of ice types, just as many igneous rock types form by partial melting of
silicate rocks within the Earth.

The melts, whether completely liquid or mixtures of liquid and crystals,
would flow across the surface of an icy world in ways that match the flows
of volcanic rocks on the terrestrial planets. For example, the ammonia-water
melt described above is considerably more viscous than water. Under the
low surface gravity of an icy satellite it would flow like basalt, the least
viscous type of lava common on the Earth. Basalt tends to produce thin lava
flows that can travel large distances relatively quickly, producing landscapes
dominated by vast tracts of flat or gently sloping lava flows. A melt containing
about 80 per cent crystals would form a ‘mush’ of liquid and crystals and
it would flow more like rhyolite, the most viscous type of lava common on
Earth. This gives an entirely different terrain, dominated by short, steep-sided
lava flows. And if, as seems likely, there was methanol in the ammonia-water
melt, it would be just as viscous even if no crystals had begun to form.
Even dissolved salts increase the viscosity of the first-formed melt way
above that of pure water, although they do not lower the melting point by
more than a few degrees.

Another parallel with terrestrial processes is what happens as a melt
produced inside a moon approaches the surface. If the confining pressure
underground decreases in such a way that bubbles of vapour form as the melt
rises, the scene is set for an explosive eruption. As the bubbles expand,
suddenly and violently, they can fragment the melt, showering the surrounding
region with shards of frost. On Earth, pyroclastic eruptions such as the
June 1991 eruption of Mount Pinatubo in the Philippines happen in this way
when the pressure of the expanding bubbles overcomes the cohesive strength
of the molten rock.

It is incidental whether the source of heat for planetary activity is
radioactive decay or tidal processes, as both can lead to internal convection
and surface activity. On the icy bodies, the melting and crystallisation
relationships shown by the impure ices of the outer Solar System repeat
all the important characteristics of melting and crystallisation found in
silicate rocks. To reach an understanding of how planets function, geologists
should look beyond the Earth, and even the terrestrial planets. There are
more than a dozen worlds beyond Mars that have developed variations on the
same theme.

* * *

OF MELTS AND MIXTURES . . .

Most rock in the terrestrial planets consists of mixtures of silicate
minerals. Such rocks begin to melt at considerably lower temperatures than
would any of its constituent minerals in isolation. The partial melting
that rock undergoes is easiest to see in a rock made of two minerals, such
as anorthite, a calcium aluminium silicate, and diopside, a calcium magnesium
silicate.

A ‘phase diagram’ illustrates how mixtures of this type melt and crystallise.
A solid mixture of 70 per cent anorthite and 30 per cent diopside would
consist of these two types of crystals only, and in these proportions, until
the temperature rose to 1275 °C. At this point all the diopside and
some of the anorthite crystals would melt, yielding a partial melt with
a composition shown by the position of the lowest point between the two
curves – in this case 40 per cent anorthite and 60 per cent diopside. The
melt would be richer in the chemicals that make up diopside than was the
original rock.

Anorthite crystals could continue to melt only if the temperature were
to rise further, and the mixture would melt completely at about 1480 °C.
The melt, if it were all pooled together, would then have the same composition
as the original solid. However, if the molten mixture separated from the
solid as the melting progressed, the melt would never regain the original
composition of the solid.

Mixtures of ices of the sort to be expected on the icy moons of Saturn
and beyond behave in the same way. The second phase diagram shows the melting
of ice made of intergrown crystals of water-ice and ammonia hydrate. A mixture
of 80 per cent water-ice and 20 per cent ammonia hydrate crystals would
produce a partial melt whose composition again matches the lowest point
between the two curves – equivalent to 33 per cent water-ice and 67 per
cent ammonia hydrate, at a temperature of -97 °C. This leaves a residue
of water-ice crystals which could continue to melt only if the temperature
rose further.

Phase relationships depend on pressure as well as temperature, and in
fact there is no need to heat either ice or silicate rock to begin partial
melting. The drop in pressure as material moved upwards by solid-state convection
in a terrestrial planet or icy satellite would be enough to do the trick.