1.3 Unravelling the natures of the large satellites
Before the dawn of the space age, relatively little could be discovered about even the large satellites. Their orbits were well known, and from the subtle orbital perturbations caused by neighbouring satellites it was possible to deduce their masses. Measurements of their sizes enabled densities to be calculated to within about 20 per cent of the currently accepted values for the Galilean satellites, and with rather less certainty for the large satellites of the other giant planets. However, it was clear that, except for Io and Europa, these bodies are not dense enough to be composed largely of rock like the terrestrial planets.
During the 1950s, spectroscopic studies by Gerard Kuiper (Figure 8), the discoverer of Titan's atmosphere, showed that the surface of Europa is mostly clean bright water-ice, whereas that of Ganymede (which has a lower albedo) is water-ice darkened by a dusty contaminant. (We use the term 'water-ice' where necessary to make it clear that we mean frozen water, as opposed to any other kind of ice.) Spectroscopic studies have now revealed that ice dominates the surfaces of all the large satellites except Io, which is effectively a terrestrial planet in orbit about Jupiter. In the Jupiter system, the ice is dominantly frozen water, but with increasing distance from the Sun it becomes mixed with more volatile ices. There is indirect evidence for ammonia in the ices of Uranus's satellites, and on Neptune's large satellite Triton spectroscopic observations have detected frozen nitrogen, carbon dioxide, carbon monoxide and methane. A similar mixture to that on Triton coats Pluto's surface.
Gerard Kuiper, a Dutch-born American planetary scientist, discovered Titan's atmosphere in 1944 and subsequently used spectroscopy to identify carbon dioxide in the atmosphere of Mars and ice on the surfaces of Europa and Ganymede. He discovered Miranda (Uranus) in 1948 and Nereid (Neptune) in 1949. In 1951 he suggested that there should be a zone of primordial debris beyond the orbit of Neptune. Although the first body in this zone was not discovered until nearly twenty years after his death, it is generally known as the Kuiper belt.
All of this is consistent with our understanding of the nature of the materials from which the Solar System formed, under conditions of progressively lower temperatures at greater distances from the Sun.
The icy satellites came to be regarded as worlds made of ice mixed with rock because their densities are greater than any variety of ice. This was because the silicate minerals that form rock constitute the most abundant denser material known to exist in the Solar System. Whether these satellites are differentiated bodies with the rock forming a dense core surrounded by a less-dense icy mantle, or whether they are undifferentiated uniform mixtures of rock and ice was assumed to depend on their accretion histories. An undifferentiated structure would imply homogenous accretion (rock and ice simultaneously) combined with insufficient heating to trigger differentiation. A differentiated structure could result from heterogeneous accretion (rock first, then ice) or from homogenous accretion if the rate of energy release during the accretion process generated enough heat to melt or at least mobilise the ice.
If a body of average density ρav consists of a mixture of just two components, a dense one with density ρdense and a light one with density ρlight, the way to work out what fraction of the body's volume is made of each is as follows. Let the fraction made of the dense component be x. The fraction made of the light component must then be (1−x).
There is a simple equation relating these values:
(a) Use Equation 1 to calculate the fraction of Callisto's volume occupied by rock, given that Callisto's average density is 1.83×103 kg m−3. Assume the density of rock to be similar to that of chondritic meteorites, which is about 3.10×103 kg m−3 and the density of ice to be about 0.95×103 kg m−3.
(b) Suggest some factors that could make the value calculated in this way unreliable.
(a) There are various ways to work this out - here is ours. The value we are looking for is ×, so we need to rearrange the equation to isolate all the terms involving × on the same side. First, expand the bracket, to get:
Next, subtract ρlight from each side:
Rearranging this equation:
We can now divide both sides by (ρdense - ρlight) to get:
Now we can simply insert the density values we were given. Callisto's average density is ρav, ice density is ρlight and rock density is ρdense, so:
The fraction of Callisto's volume occupied by rock is about 0.41.
(b) One reason the value may be unreliable is that the densities used are for rock and ice at low pressure. In the interior of a large icy satellite the pressure might be high enough for self-compression to lead to significantly higher densities. Another reason is that the method assumes rock and ice only, and ignores the possibility that there could be an even denser component such as an iron-rich inner core.
Irrespective of whether the rock is dispersed or concentrated, the total rock content of these bodies is too low for radiogenic heating, by the decay of radioactive elements contained within the rock, to provide sufficient heat to mobilise their interiors and refresh their surfaces. In the 1960s, the average surface temperatures of the Galilean satellites were established to be lower than −150 °C using infrared telescopes. This is so low that the ice near the surface must have comparable mechanical properties to rock near the surface of a terrestrial planet. Such ice is far too cold to behave like glacier ice on Earth, which is capable of flowing downhill under its own weight. Thus, whatever their internal structure and their mode of origin, all the icy satellites at Jupiter and beyond (where surface temperatures are even lower) were assumed to have long been geologically dead, with the implication that they must be densely covered by impact craters that have built up during the past four billion years.
Just how wrong some of these suppositions were did not become apparent until close-up images of the satellites of the outer planets were sent back by spacecraft. Only the merest hints were provided by the blurry images returned by the first probes to visit Jupiter, Pioneers 10 and 11 in 1973 and 1974. The situation became much clearer thanks to the remarkable tours of the outer Solar System accomplished by the two probes of NASA's Voyager series, beginning with Voyager 1's encounter with Jupiter in March 1979 and ending with Voyager 2's fly-by of Neptune in August 1989 (Box 1). These revealed a startling diversity of landscapes on the icy satellites. Some are indeed heavily cratered, and look much like what most people expected (Figure 10).But others have a complex variety of terrain types, showing relatively few impact craters but many signs that faulting, flooding and other resurfacing processes have acted to disrupt or bury any ancient heavily cratered terrains that may formerly have existed (Figure 11).
Box 1: The Voyager project
In 1977, NASA launched two probes named Voyager to explore the outer Solar System (Figure 9). Voyager 1 flew through the Jupiter system in March 1979, and used Jupiter's gravity to redirect its trajectory towards Saturn, which it passed in November 1980. Voyager 2 used the same 'gravity assist' tactics to visit all four giant planets in turn, beginning with Jupiter in July 1979 and concluding with Neptune in August 1989.
Each of the Voyager probes weighed 825 kg, of which 105 kg was scientific instruments. These included cameras, spectrometers, polarimeters (to measure polarisation of reflected radiation) and magnetometers. Because it was designed to travel so far from the Sun, power was provided not by solar panels but by the heat produced by radioactive decay in a plutonium-rich thermoelectric generator.
With the exception of Titan, no satellite has an atmosphere thick enough to protect its surface from bombardment. The surface of an icy satellite scatters sunlight fairly evenly in all directions, which means that not even the youngest surface consists of a continuous sheet of smooth ice. Instead, any ice that was a continuous sheet originally has become broken (presumably by meteorite and micrometeorite impact) into a mass of granular fragments, with a wide range of particle sizes, in the same way that the lunar surface consists of a regolith of rock debris. Presumably the icy regolith is thinner (only a few particles in thickness) on the youngest icy surfaces and thickest (several metres or more) on the oldest surfaces.
Unfortunately, the resurfacing events on satellites such as those in Figure 11 are impossible to date. The lunar cratering timescale, which has been calibrated radiometrically (i.e., using dating methods based on the decay of radioactive isotopes), cannot be applied in the outer Solar System. This is because we can have no expectation that the Moon (1 AU from the Sun) suffered the same rate of impact bombardment as a satellite of Jupiter (5 AU from the Sun) or a satellite of Saturn at a range of nearly 10 AU. Indeed, when the size-frequency distributions of craters on the icy satellites are examined, it is found that the pattern of distribution of craters versus size range differs from the satellites of one giant planet to the next, and that each is different to that found on the Moon. ('Size-frequency distribution' is a term used to describe the relative numbers of objects - in this case, craters - across a range of sizes.) This is convincing proof that different populations of impactors affected each region of the Solar System, so it is likely that cratering rates also behaved differently in each region. We can imagine a general decrease over time, but there may have been localised flurries of cratering, such as would be caused by the impact of debris originating from a nearby satellite that had been broken apart by a single, random, exceptionally large impact. Thus, while we can be reasonably confident that a less densely cratered surface is younger than a more densely cratered surface when comparing between satellites of the same giant planet, we cannot make any such comparison between a less densely cratered surface on a satellite of, say, Jupiter, and a more densely cratered surface of a satellite of, say, Saturn.
On many individual satellites, the differences in crater density between the oldest terrain (which may be a surface that is four billion years old) and the youngest resurfaced terrain suggest that the latter is considerably younger. In order for their surfaces to have been regenerated from within, these satellites must have experienced internal heating.
There are several reasons why radiogenic heating can be ruled out as a significant factor in this heat generation:
The total rock content of these satellites is far too low to generate enough heat, unless the rock has some implausibly weird composition, with ten to a hundred times more radioactive elements than chondritic meteorites.
Any exotic rock composition would have to vary enormously within a single satellite system to explain the geologically complex surface of Enceladus (Figure 11b) in contrast to the passive densely cratered surface of its fellow satellite Rhea (Figure 10b). Rhea is twenty-five times more massive and ought to be producing more radiogenic heat than tiny Enceladus.
Radiogenic heating ought to decay gradually over time, which is not reflected in the resurfacing histories of the more active satellites.