The possibility of volcanic resurfacing on icy moons was first voiced by Lewis (1971, 1972) and subsequently addressed by Consalmagno and Lewis (1978), but it was not until after the Voyager flybys of Jupiter and Saturn that we saw spectacular evidence for past (and possibly present) tectonic and volcanic activity on moons such as Europa, Ganymede, and Enceladus. Our views of the Uranian moons and of Triton have yielded even more surprising images of cryovolcanic activity (Stevenson, 1982: Stevenson & Lunine, 1986: Melosh & Janes, 1989: Murchie, 1990: Kargel & Strom, 1990: Schenk, 1991: Schenk & Jackson, 1993).

Some of the most important work to emerge as a result of these observationscentred around improving our knowledge of the phase diagram for the ammonia-water system in the ranges of temperatures and pressures which are of relevance to the interiors of icy satellites. The phase diagram reproduced at the top of the previous page (from Kargel, 1992) represents almost ninety years of concerted effort, perhaps some of the most important being made by Rollet and Vuillard (1956), who were the first to identify the dihydrate phase near 32wt% NH₃. Early work of planetary relevance was undertaken by McKinnon and Meadows (1984) who studied the rheological properties of ammonia-water liquids, and Johnson et al (1983, 1984, 1985) who looked into the melting behaviour of the monohydrate and hemihydrate phases. Johnson and Nicol (1987) subsequently constructed the phase diagram for compositions 0-33 wt% NH₃ up to pressures of 5GPa. Croft et al (1988) expanded on this research, determining the equation of state of ammonia-water liquids and deriving densities for liquid and solid phases from 0-100 wt% NH₃ at temperatures from 170-300K and pressures up to 10kbar.

Jeffrey Kargel has since studied the phase relations and physical properties of ammonia-water liquids and a wide range of other substances of cryovolcanic interest, with a particular view to understanding the petrology of cryomagmas (Kargel, 1987, 1988, 1990, 1991, 1992: Kargel et al, 1991). A full review of cryovolcanic processes is inappropriate to the context of this work; for one of the best studies on the subject refer to Kargel (1990).

Kargel applied phase relations in the quinary system NH₃, H₂O, CO₂, CH₂O, and CH₃OH to explain the observed difference in viscosity between the Jovian & Saturnian satellites, and the Uranian satellites. Ammonia reacts with carbon dioxide in the presence of water to form a carbonate or carbamate (depending on the mole fraction of ammonia), and possibly urea. Ammonia also reacts with methanal to form a white organic salt, hexamethylenetetramine (HMT, which exists as a hexahydrate in the presence of water). Water causes methanal to polymerise to form polyoxymethylene (POM), a familiar plastic in the chemical industry. The phase diagram for the system water, methanol, ammonia consists of a number of stoichiometric hydrates and ammoniates.

Kargel found that carbon dioxide and methanal tend to sequester ammonia from the cryomagma leaving a water rich solution of methanol. He identified a plane of critical saturation (analogous to the critical plane of silica saturation in the basalt tetrahedron) on either side of which were produced liquids of radically different rheological properties (see below). The ammonia undersaturated side of the plane yields eutectic melts which are typically water-methanol solutions containing abundant salts, such as ammonium chloride (sal ammoniac) or ammonium sulphate (mascagnite). These are collectively referred to as brines and have very low viscosities, close to the viscosity of water. The ammonia rich side of the critical plane produces ammonia-water-methanol melts that are several orders of magnitude more viscous than brine cryolavas. Brine melts are widely thought to be responsible for the thin, sheet-like extrusions on Ganymede and Enceladus, whereas the ammonia rich magmas are more likely to be responsible for the thick, ropy flows on Miranda and Ariel. The figure here (from Kargel et al, 1991) shows the viscosity of cryomagmas and silicate lavas as a function of the surface gravity of the body upon which the lava is emplaced. The diagram illustrates that an ammonia-water eutectic melt on Titan would have roughly the same viscosity as a terrestrial basaltic andesite.

The illustration here is a representation of the pseudo-quaternary system water-ammonia-carbon dioxide-methanal. (The system is pseudo-quaternary because a fifth component, methanol, is considered.) The blue alkemade triangle between water, HMT hydrate, and the ammonium amino-carbonate NH₄NH₂CO₃, is the plane of ammonia saturation. On the ammonia undersaturated side of the plane, all of the free ammonia is taken up to form ammonium salts and the melt is predicted to be extremely methanol rich (~70% methanol, 30% water, melting near 170K). On the ammonia saturated side of the plane there is sufficient free ammonia for melting to occur, broadly, in the ternary sub-system H₂O-NH₃-CH₃OH, yielding a 47% water, 23% ammonia, 30% methanol fluid at a eutectic temperature of ~153K. This solution differs in viscosity from understaurated melts by four orders of magnitude (see above.).

Other workers have contributed to our understanding of the behaviour of ammonia-water liquids at high pressures (Nicol et al, 1989: Boone & Nicol, 1991: Hogenboom et al, 1994) and their rheological properties (Goldsby & Kohlstedt, 1994).

Kargel (1990) indicated that eutectic melts should be slightly less dense than the solid phase and thus be able to rise buoyantly through an ammonia hydrate-water ice crust*. As these liquids are erupted onto the surface, perhaps to depths of several kilometres, the harder it would be for ammonia-rich residual liquids to reach the surface, and are therefore more likely to stall at depth, crystallising as plutons.

To extrapolate from here and consider the possibility of ice volcanism on Titan is an act of purest speculation; there is no evidence to support the contention that volcanism has been, or is currently, active on the surface. However, given the observation of past ice volcanism on so many outer solar system satellites, in particular the small Saturnian satellites such as Enceladus where activity appears to be very recent if not contemporary, it seems altogether unreasonable to assume that ice volcanism has not occurred on Titan in the past at the very least. Given that a large supply of ammonia-water liquid may be present beneath the crust its also seems feasible that volcanism or plutonism could continue up to the present.

Above: Cryoclastic volcanism on Titan? Not according to Ralph Lorenz. Image from Moore & Hardy, 1972.

This subject has received surprisingly scant consideration. Lorenz (1996) investigated and constrained some of the likely parameters of volcanism on Titan, determining (based on radiogenic heat flow considerations alone) a maximum resurfacing rate of 2x10E^-3m/yr. If we are to take the view that the fluids present today have strongly differentiated (Kargel, 1990) then all of this material will solidify at depth and evidence for it may be subtle and difficult to detect with the rather limited coverage that Cassini will be able to give us. Lorenz (1996) assumed that melt products would be capable of reaching the surface and found that, whilethe morphology of volcanic features would be rather small (of the order of a few hundred metres), Cassini would be able to resolve those features.

Past.

An ocean of ammonia-water several hundred kilometres deep at the surface of Titan has the potential to interact at two boundaries; top and bottom. At the bottom fluids can circulate through the upper layers of the outer rocky core in a manner analogous to the hydrothermal systems along the Earth's mid-ocean ridges. At the top, infalling meteoritic material sediments through the upper reaches of the ocean, reacting to form new minerals.

Engel et al (1994) studied the reactions of exogenic silicate material with the ammonia-water ocean. At the postulated surface temperatures of >300K, silicates give up their sodium and potassium ions to the solution, being replaced in the crystal lattice by the ammonium ion. The result of these interactions is the production of a suite of ammonium silicate minerals (see appendices I & III), that could conceivably become trapped in the ice Ih crust as it solidifies. A great deal of evidence already point to a surface composition similar to that of Callisto*, and ammonium bearing illites have been tentatively identified in spectra of Callisto (Calvin & Clark, 1993). As reactions proceed to equilibrium the ocean solution becomes richer and richer in dissolved potassium and sodium. Volcanism throughout Titan's history should then bring up cryofluids bearing these ions in solution (see Kargel, 1990 & 1992, for phase relations of alkali metals in ammonia-water liquids). The important point, in terms of providing a test for volcanic resurfacing, is that radiogenic potassium (⁴⁰K) will decay to ⁴⁰Ar and this gas will be released into the atmosphere by meteorite impacts. It would seem, at first glance, that the atmospheric concentration of ⁴⁰Ar should be directly related to the amount of liquid which has been erupted from the interior. This depends on the amount of argon that was originally accreted into Titan. If Titan's nitrogen were accreted as N₂ clathrate (instead of as ammonia), and assuming a solar ratio of N₂/Ar of 11.7 (Owen, 1982a) then the mole fraction of ⁴⁰Ar in Titan's atmosphere would entirely overwhelm the small amount outgassed through volcanic activity. Alternatively, if we are to accept the kinetic inhibition models and believe that Titan's atmospheric nitrogen is the result of the photolysis of ammonia, then the mole fraction of accreted ⁴⁰Ar will be so small that the supply due to volcanic activity will not only dominate but be readily detectable by the Huygens descent probe.

* The Galileo space-craft has recently been used to attempt to characterize the non-water-ice components of the surface of Callisto (and the other Galilean satellites). The Near Infra-red Mapping Spectrometer (NIMS) found only very weak water ice signatures on Callisto, suggesting that water ice represents between 0-10% of the surface minerals. Five relatively strong bands have been attributed to CH, CO₂, CN, SO₂, and an SH-radical containing material. (McCord et al, 1997)

No study of hydrothermal interactions at the base of the ocean has yet been made. However Shock and McKinnon (1993) considered the interaction of a liquid mantle and rock core on Triton, and this may prove useful in assessing the types of chemical reaction that might have occurred at Titan's core-mantle boundary. It should be noted that Shock and McKinnon took the composition of Triton's mantle to be similar to that given in Mumma et al (1993) for the composition of comets. Triton almost certainly formed in a different chemical environment to Titan, plausibly directly from the PSN. The rock:ice ratio is higher and the mantle is likely to be far richer in CO, CO₂, and N₂. Nevertheless there are grounds for drawing limited comparisons. A certain amount of carbon monoxide and methanal may have been accreted into Titan anyway, and more is likely to have been supplied by cometary impacts. Taking into account these oxygen bearing compounds, then a range of simple organic compounds can be produced by hydrothermal reprocessing; including simple carboxylic acids, alcohols, amines, and amino acids. These are of specific interest because of the effects that they have on the melting properties and rheologies of cryomagmas. Methanal sequesters ammonia, and methanol lowers the melting point and is concentrated in the residual liquid (Kargel, 1990). The detection of these and other compounds may place constraints on the level of cryovolcanic activity and the degree of hydrothermal interaction at the core-mantle boundary. This is not something that Huygens will be capable of achieving. Such a detailed study requires the investigation of modal mineralogies and petrographic textures of individual cryolava samples. Our ability to acquire such samples is certainly in excess of fifty years in the future.

Contemporary volcanic activity ought to be rather easy to detect. The Cassini Visible and Infra-red Mapping Spectrometer (VIMS) will map parts of the surface through the atmospheric window at 5.1µm. An ammonia-water cryomagma at a peritectic temperature of 176K will emit twenty times more radiation at this wavelength than the surrounding landscape radiating at ~94K, contrasting quite strongly, particularly in images of the night hemisphere of Titan (Baines et al, 1992).

Every solid body, regardless of whether it has an atmosphere or not, shows evidence of surficial alteration by impact cratering: We would not expect Titan to differ in this respect. There are, however, some erosional forces that are unique to bodies with atmospheres and with flowing liquids at the surface. Titan is blessed with an atmosphere, and quite possibly with surface liquids as well. This would imply that erosion and deposition due to aeolian, pluvial, fluvial, and marine forces will be expressed in the surface morphology.

As has already been discussed the crust, or 'bedrock' of Titan is expected to be the familiar low pressure phase ice Ih, perhaps with an admixture of ammonia hydrates and smaller proportions of more complex compounds. The degree to which weathering and erosion are possible depends on the mechanical strength and the solubility of this bedrock in hydrocarbon fluids.

The latter problem was investigated by Rebai et al (1983). Their group reported a surprisingly large value for the solubility of water in cryogenic liquids. This was subsequently shown to be erroneous and a revised figure was produced (Rest et al, 1990). Sagan and Dermott (1982) first considered the dissolution of a water ice crust via the mechanism of tidal erosion. They argued that tidal currents in contained ocean basins would be sufficient to remove any land masses, producing the global ocean their dynamic constraints demanded. Lunine and Stevenson (1985a) applied the same model using the solubility values of Rebai et al (1983), and found that the dissolution of kilometre scale water ice features was possible over the age of the solar system. Following the publication of revised solubility data, Lunine (1993) calculated that chemical erosion should not be a significant process. However the solubility of ammonia in cryogenic hydrocarbons is six or seven orders of magnitude greater than that of water. In fact the solubility of ammonia in these liquids is of the same rough order of magnitude as calcite in water (Krauskopf & Bird, 1985, quoted in Lorenz & Lunine, 1996). Though the solubility of hydrates of ammonia remains to be determined it is likely to be (significantly?) larger than that of water.

Lorenz and Lunine (1996) explored the role of erosion on Titan's surface in depth, in particular the influence of liquids on the landscape. Applying new solubility figures they found that pure water ice is unlikely to be chemically eroded, but that hundred-metre-scale patches of glacial ammonia or ammonia hydrates may be eroded in as little as a few tens of thousands of years. This has the interesting corollary that a karst landscape could develop involving subterranean caverns and passages, and maybe even analogues of stalactites and stalagmites!

Lorenz (1994) highlighted a fascinating implication for the effect of bedrock solubility in Titan's surface fluid on crater morphology. Lorenz demonstrated that craters should fill with liquid ethane to produce crater lakes, and that standard variations on crater shape (such as central peaks, multiple rings, and domed centres from viscous relaxation of the substrate could generate 'ring' and 'bullseye' lakes. Furthermore, an assessment of the magnitude and direction of tidal currents indicates that these rings and bullseyes could become smeared out to form 'horseshoes' oriented towards the anti-Saturn point (0°N, 180°E). In the event that neither chemical or mechanical erosion is equal to the task of smearing out these lakes we may observe some unusual drainage patterns as lakes break their banks along one margin and then drain back as tides rise and fall.

Lorenz and Lunine (1996) also applied their solubility study to the solid hydrocarbons thought to coat the surface (see also Raulin, 1987). Short chain hydrocarbons are more soluble than longer ones and rapidly saturate in the surface fluid, precipitating out as 'benthic' sediments. While other solids, such as nitriles and tholins, may not saturate in the oceans it may be possible for them to be leached by the much larger solvent volumes that circulate through the regolith (methane recycling due to internal heat flow and leaching by rainfall). In fact the solvent volume circulating through the soil may be as much as 10⁵ times greater than the purported ocean volume. The evaporation of methane at the surface could leave an evaporite deposit; evaporation at depth could leave a distinct chemical horizon in the soil.

The effectiveness of physical weathering is dependent on the mechanical strength of the agent under attack. On the Earth we are well aware that chalk is eroded very easily, sandstones less so, and granites least of all. The strength of the material depends on how well it is bound together and how hard its components are. In a given rock the softest constituents are destroyed quickest, and the detrital grains are usually the hardest. Quartz, with a hardness of 7, is the commonest terrestrial detrital grain.

Ice which we encounter every day is not much harder than gypsum and really not terribly strong. As temperatures fall the properties of ice change drastically. The rigidity of ice at 100K is 4x10E¹⁰ dyne cm^-2, about ¹/₃ the value for granite at the Earth's surface (15x10E¹⁰ dyne cm^-2), and its hardness at 197K is 6. At the temperatures we find at the surface of Titan water ice is a relatively strong material, petrologically speaking, and at least as hard as quartz.

One of the most basic forms of physical weathering on earth is the freeze-thaw cycle, where the formation of ice crystals (with an associated 9% volume increase) in cracks can break rocks apart. A related phenomenon is the growth of other crystals, such as halite, in rock crevasses. However the atmosphere on Titan is so massive that it has a considerably longer radiative time constant than our own atmosphere (Flasar et al, 1981, calculated a value of about 140 years), so diurnal temperature variations are likely to be negligible. Added to this is the observation of a surprising uniformity in temperature with latitude (roughly a 4K difference from equator to pole). The possibility of diurnal weathering thus becomes vanishingly small (Lorenz & Lunine, 1996).

Here is a revised and updated Eosean cycle for the lower half of the troposphere illustrating the contributions from volcanism and the cycling of fluids through the regolith.

Rainfall is one of the primary weathering agents both on Earth and, in the distant past, on Mars. However, Lorenz (1993a, 1993b) indicated that, while rainfall could be quite heavy, little of it may actually reach the ground, thus reducing its importance in the weathering process. Lorenz and Lunine (1996) made a quantitative study of the erosive power of rainfall on Titan. They calculated that erosion rates due to rainfall are probably somewhere between one hundred and ten thousand times lower than on Earth, giving an upper limit of 1 metre every million years. This rate is comparable with the rates observed in terrestrial periglacial environments and the lower limit of the rates estimated for the martian highlands (Craddock & Maxwell, 1993).

Whether or not methane which is able to survive the descent to the surface is able to run off to form streams and rivers depends on the methane humidity of the atmosphere and the rate of delivery to the surface. Given our current knowledge of both of these it seems unlikely that rivers can occur on Titan. Should Cassini observe widespread drainage networks this would argue for a considerably 'wetter' climate. Lorenz and Lunine draw an analogy with the caliche topped mesas of southern Nevada: Light rainfall on Titan may permit the cementation of the uppermost layer of the regolith (similar to calcrete). Where stronger rainfall were able to penetrate this hardpan layer, preferential erosion of the underlying soil might produce isolated mesas.

As with rain, glaciers have played an important role in reshaping the landscapes of both Earth and Mars (Kargel & Strom, 1992, 1996). Methane is the most likely substance to form glaciers on the surface of Titan; in its pure form it freezes at just 90.7K. However nitrogen forms a solid solution with methane (Connolley et al, 1980), considerably lowering its melting point. (See Eluszkiewicz & Stevenson, 1990b, for a study of the rheology of nitrogen/methane solids.) Under the current climatological conditions it is very unlikely indeed that a glacial nitrogen/methane solid solution can exist on Titan. Possible climate changes that could cause the atmosphere to freeze onto the surface and so yield a more favourable environment have already been discussed. Lorenz and Lunine (1996) calculated that such a freeze out should produce either a global layer of glacial nitrogen/methane 100 metres thick or, if preferentially deposited at the poles, 'ice' sheets up to 1km thick. They further went on to calculate that the presence of U-shaped valleys of depth 500-1000 metres would indicate a period of atmospheric freeze out of duration 0.25-100 million years.

Since it seems that two of the most important erosive processes bear so little on Titan's surface morphology, we must turn to wind to try and find a good candidate for a strong weathering mechanism. Greeley and Iverson (1985) were the first to take a quantitative approach by calculating probable saltation threshold speeds. This work was built upon by Allison (1992) and Grier and Lunine (1993), taking into account variations in particle cohesion and drag. Lorenz et al (1994) produced updated determinations of friction speeds, saltation lengths, and ripple lengths, and assessed the availability of sediment in terms of the formation of dunes which might be detected by the Cassini instruments. Lorenz et al found that Titan's current atmosphere behaves rather like that of Venus, where threshold wind speeds considerably higher than predicted friction speeds make saltation unlikely. Added to which, the supply of sediment might be rather limited. Windblown sediments abrade rocks to produce more particles. As observed by Greeley and Iverson (1985) the low friction speeds, short saltation lengths, and near vertical impact angles render the winds of Titan rather ineffective at abrasion. Thus the only materials available in abundance that can act as transportable sedimentary grains are those produced by meteorite impact and volcanic processes. The probable absence of cryoclastic activity noted by Lorenz (1996) means that volcanic activity can also be effectively ruled out as a sedimentary source. Given this paucity of sediment supply, Lorenz et al found that even fairly small sub-aerial dunes would require long periods of time (30 million years for a 10 metre high feature) to accumulate. Furthermore, given the superior ability of fluids to move seafloor sediments, and so produce ripples and dunes, the detection of these landforms would appear to suggest a sub-marine origin. Models of thinner and thicker atmospheres do not greatly increase the probability of the formation of sub-aerial wind blown deposits.

Lorenz and Lunine (1995) briefly considered the activity of the winds on Titan, noting the new Global Circulation Model (GCM) of Hourdin et al (1995a) that predicted very low surface wind speeds (~1m s^-1) and matched the high altitude superrotation seen during the occultation of the star 28 Sagitarii*. This very low wind speed means low sediment transport rates. Lorenz and Lunine quote a transport rate fully six orders of magnitude smaller even than on Mars.

* 28 Sagitarii. The occultation of the K-giant star SAO187255 (28 Sagitarii) on July 3rd 1989 allowed astronomers to infer a number of key facts about Titan's atmosphere. Firstly, that it displays an axial symmetry that indicates Titan's spin vector to be aligned with that of Saturn (Sicardy, 1991). Secondly, that the opacity of the northern limb was about three times greater than the southern limb, presumably a seasonal effect (Sicardy, 1991: Yelle et al, 1992). And thirdly, that the upper atmosphere is rotating in about 26 hours, implying wind speeds of ~ 100 metres per second (Hourdin et al, 1991: Hubbard & Sicardy, 1992). Subsequent models (Del Genio et al, 1991: Hubbard et al, 1991: Hourdin et al, 1994, 1995b) have since been able to reproduce this high level superrotation and suggest very weak surface winds. Close study of the ethane emission lines in Titan's atmosphere by Kostiuk et al (1997) have largely confirmed these findings, observing prograde wind speeds of ~70 metres per second.