ASTROBIOLOGY
Volume 3, Number 4, 2003
© Mary Ann Liebert, Inc.
Europa
Sulfate Content of Europa’s Ocean and Shell:
Evolutionary Considerations and Some Geological
and Astrobiological Implications
WILLIAM B. MCKINNON1 and MICHAEL E. ZOLENSKY2
ABSTRACT
Recent models for the origin of Jupiter indicate that the Galilean satellites were mostly derived from largely unprocessed solar nebula solids and planetesimals. In the jovian subnebula the solids that built Europa were first heated and then cooled, but the major effect was
most likely partial or total devolatilization, and less likely to have been wholesale thermochemical reprocessing of rock 1 metal compositions (e.g., oxidation of Fe and hydration of
silicates). Ocean formation and substantial alteration of interior rock by accreted water and
ice would occur during and after accretion, but none of the formation models predicts or implies accretion of sulfates. Europa’s primordial ocean was most likely sulfidic. After accretion
and later radiogenic and tidal heating, the primordial ocean would have interacted hydrothermally with subjacent rock. It has been hypothesized that sulfides could be converted
to sulfates if sufficient hydrogen was lost to space, but pressure effects and the impermeability of serpentinite imply that extraction of sulfate from thoroughly altered Europa-rock
would have been inefficient (if indeed Mg sulfates formed at all). Permissive physical limits
on the extent of alteration limit the sulfate concentration of Europa’s evolved ocean to 10%
by weight MgSO4 or equivalent. Later oxidation of the deep interior of Europa may have also
occurred because of water released by the breakdown of hydrated silicates, ultimately yielding S magma and/or SO2 gas. Geological and astrobiological implications are considered. Key
Words: Europa—Io—Extraterrestrial oceans—Sulfate—Accretion—Meteorites—Hydrothermal systems. Astrobiology 3, 879–897.
INTRODUCTION
A
for the formation of Europa’s water 1 ice layer argues for a hypersaline sulfate composition (Kargel, 1991; Kargel
et al., 2000; Spaun and Head, 2001). The model is
PROMINENT MODEL
based on the plausible compositional affinity between the bulk of Europa and the most volatilerich meteorite class, CI carbonaceous chondrites
(e.g., Fanale et al., 1977; Mueller and McKinnon,
1988). These meteorites are extensively, and apparently isochemically, aqueously altered (e.g.,
1 Department
of Earth and Planetary Sciences and McDonnell Center for the Space Sciences, Washington University, St. Louis, Missouri.
2 NASA Johnson Spacecraft Center, Houston, Texas.
879
880
Frederiksson and Kerridge, 1988; Zolensky and
McSween, 1988, and references therein), with
abundant Mg and Ca sulfates (epsomite and gypsum) as well as some Na and other minor sulfates. The central tenet in Kargel (1991) and
Kargel et al. (2000) is that a Europa that accreted
from carbonaceous chondritic materials essentially matches Europa’s observed bulk density,
and regardless of whether the sulfates formed in
“satellitesimal” precursors or by aqueous alteration within Europa, early stages of heating
would have caused epsomite to melt incongruently to form a buoyant hypersaline brine. The
brine would have erupted to the surface to form
an aqueous sulfate crust/ocean (by weight, ,30%
MgSO 4, ~5% Na2SO4 , plus minor additional sulfates). Later additions of sulfate-poor waters from
the breakdown of, for example, gypsum (which
is much less soluble than epsomite) and hydrated
silicates such as serpentine would then have led
to a variety of compositional and structural evolutionary paths for Europa’s icy outer layer. The
emphasis in the model, however, remains on hypersaline compositions (Kargel et al., 2000).
More generally, sulfates formed in planetesimal or satellitesimal precursors could undergo
bulk melting during accretion or, if formed by
aqueous alteration within Europa, could erupt directly to the surface without necessarily going
through stages of precipitation and remelting. In
the former case the crust formed would have the
composition of the accreting sulfates [by weight,
,45% MgSO 4, ,15% Na2SO 4, if CI-chondrite like
(Kargel, 1991)], whereas the sulfate salinity in the
latter case (formation within Europa) would not
be so constrained and would depend on, for example, the water/rock ratio during alteration.
Nonetheless, sulfate dominance would be expected because of the high solar abundance of
sulfur and the solubility of many sulfates, and the
dilution on Europa of more “traditional” alkali
halides such as NaCl (Fanale et al., 1977; Kargel,
1991; Kargel et al., 2000; Zolotov and Shock,
2001b). Frozen sulfate brine is also closely compatible with Galileo Near Infrared Mapping Spectrometer (NIMS) surface spectral analysis, especially for spectra obtained from the so-called
non-ice end-member in areas of relatively recent
geological activity or impact exhumation (McCord et al., 1998, 1999, 2001, 2002). This spectral
interpretation, however, is not without its critics
(Johnson et al., 2004).
The above scenario is logically internally consistent, but needs to be reconsidered (McKinnon,
MCKINNON AND ZOLENSKY
2002b). We argue here that the composition of
the Europan ice water 1 ice layer may not be
very sulfate-rich and, in principle, may contain
no sulfates at all (although it should contain sulfur). The basis for our arguments comes from advances in understanding meteorite genesis and
alteration, as well as new models of satellite formation.
SULFATE IN
CARBONACEOUS METEORITES
Carbonaceous chondrites are oxidized and
hydrated compared with ordinary chondrites.
The most volatile-rich among them, the CI chondrites, contain major amounts of magnetite
(Fe3 O4 ), carbonates [e.g., (Ca,Mg)CO3 ], gypsum
(CaSO 4 ? 2H 2 O), and epsomite (MgSO 4 ? 7H 2O).
Of these minerals, sulfates, and magnesium sulfates in particular, require particularly oxidizing
conditions to form (e.g., Lewis, 1982). Such conditions have not been shown to arise naturally in
theoretical models of meteorite aqueous alteration, whether they involve equilibrium calculations (Zolensky et al., 1989; Rosenberg et al., 2001)
or reaction path models in which the temperature
increases monotonically (Schulte, 1997).
Moreover, the magnesium sulfate veins in CI
chondrites (e.g., Orgueil) are likely the result of
terrestrial exposure (Gounelle and Zolensky,
2001). Sulfate veins in Orgueil and other CI chondrites [and one CM (Lee, 1993)] have been taken
as prima facie evidence of a late-stage aqueous alteration event [from Sr isotopic data, possibly
within the last 10 Myr for Orgueil (MacDougall
et al., 1984)]. Gounelle and Zolensky (2001) found
that these sulfate veins were not identified or reported in the literature when the meteorites in
question originally fell (1864 for Orgueil), and it
was not until the early 1960s that sulfate veins began to be noted and studied in different CI chondrites. Pieces of Orgueil exposed to humid ambient air (or even sealed) for more than 130 years
are now covered with an efflorescence of magnesium sulfate (clearly not present at the time of
the meteorite’s fall). This should not be surprising: A characteristic occurrence of epsomite on
Earth is as florets formed on moist, sulfide-bearing cave walls (e.g., Hurlbut, 1971).
It is not certain whether the magnesium sulfates in Orgueil and other carbonaceous chondrites formed by oxidation (corrosion) of meteoritic sulfides, which is certainly kinetically
SULFATE CONTENT OF EUROPA’S OCEAN
plausible for this soluble sulfate, or whether the
sulfate veining and efflorescence are simply a remobilization due to humidity of extraterrestrial
sulfates originally present. Airieau et al. (2001)
measured oxygen isotopes in CI sulfates and
found them consistent with a terrestrial origin via
sulfide corrosion (i.e., the D17 O is terrestrial-like).
Lewis and Prinn (1984, p. 143) claim that up to
one-half of the water reported in CI chondrites
may be terrestrial in origin. In the latter case, isotopic exchange with terrestrial water may have
occurred, so the amount of additional, terrestrial
water cannot be accurately determined. Sulfate
oxygen, however, does not readily exchange with
solvent water, so it is thought that sulfate oxygen
isotopic compositions preserve the signature of
the oxidation process (Airieau et al., 2001). It is
worth noting that one of us (M.E.Z.) has only observed sulfates in chondritic interplanetary dust
particles that have been exposed to air for years.
The remarkable carbonaceous chondrite, Tagish Lake, which fell in northwest Canada in 2000,
881
is one of the most chemically and mineralogically
primitive meteorites known (Brown et al., 2000).
Its organic fraction is different from that in other
carbonaceous chondrites (e.g., a much reduced
variety of amino acids), and based on spectral
similarities the meteorite has been hypothesized
to be our first sample of a P-type or D-type asteroid (Hiroi et al., 2001). P- and D-type asteroids
predominate in the outer asteroid belt and Trojan groups (e.g., Gradie et al., 1989), and are thus
arguably even better compositional analogues for
the volatile-rich Galilean satellites than C-type
carbonaceous asteroids (and meteorites derived
from them), which predominate in the main belt.
Significantly, sulfides but no sulfates have been
reported for samples of this meteorite (Brown et
al., 2000; Gounelle et al., 2001; Keller and Flynn,
2001; Zolensky et al., 2002) (Fig. 1).
These three lines of evidence—(1) theoretical
difficulties in accounting for the highly oxidized
nature of sulfate-bearing carbonaceous chondrites; (2) a likely terrestrial origin for magnesium
FIG. 1. Backscattered electron image of the primitive Tagish Lake carbonaceous chondrite (modified from Zolensky
et al., 2002). Shown are phyllosilicate aggregates and a pyrrhotite (iron-sulfide)-rich rim (light gray). No sulfates are present in Tagish Lake, and if Europa’s bulk composition is analogous, its ocean is likely sulfate deficient.
882
sulfate veins in CI carbonaceous chondrites; and
(3) lack of any sulfates in the freshly fallen Tagish Lake meteorite—considerably weaken the
theoretical underpinning for a sulfate-dominated
Europan ocean. On the other hand, the evidence
against an extraterrestrial origin for magnesium
sulfate in Orgueil and other CI chondrites (or at
least veins thereof) does not easily extend to calcium sulfate, which has a much lower solubility
and presumably slower reaction kinetics. If gypsum formed by aqueous alteration on asteroid
parent bodies, then only some additional oxidation would have been necessary to stabilize epsomite. Before discussing how such sulfates could
have formed, we first review formation models
for Europa.
PREDICTED COMPOSITION OF EUROPA
Predicted compositions for the regular satellites of Jupiter and Saturn go back to the pioneering works of Lewis (1971), Pollack and
Reynolds (1974), and Prinn and Fegley (1981).
The predictions are based on the concepts of thermochemical equilibrium in a condensing gas of
solar composition within a planetary subnebula
whose pressure and temperature decline with
distance from the planet. The most detailed chemical models for the jovian subnebula were offered
by Prinn and Fegley (1981), who predicted a hydrated and oxidized rocky component once temperatures had fallen sufficiently. In particular, no
free Fe-Ni metal was predicted; rather, iron
would appear as FeS and Fe3O4, and with Mg in
hydrated silicates such as serpentine. In strict
chemical equilibrium, carbon would remain
largely uncondensed in the gas phase as CH4
(Prinn and Fegley, 1989; Fegley, 1998, 1999b) and
only begin to appear as minor ammonium bicarbonate and/or carbamate at low temperatures
(well below the threshold for water-ice condensation) (Prinn and Fegley, 1981).
Such chemical models implicitly assume that a
suitable accretion physics exists because of the
fact that the satellites formed. It has, however,
long been recognized that the standard model of
a satellite-forming subnebula, essentially a miniature version of the solar nebula, faces great dynamical difficulties (Lunine and Stevenson, 1982;
Stevenson et al., 1986). Specifically, cooling and
condensation times of subnebular materials
would have been much longer than accretion
times, so early “satellitesimals” could have been
MCKINNON AND ZOLENSKY
lost to gas drag; moreover, tidal interactions between larger satellites and the subnebula would
have been strong enough that these satellites
should have migrated inward and been lost before the subnebula dispersed (presumably when
the solar nebula itself dissipated).
Such considerations form the background for
the recent study of Canup and Ward (2002).
Based on recent numerical studies of Jupiter’s formation, in which accretion of material into the
jovian subnebula continues at a reduced rate even
after Jupiter grows large enough to open a gap in
the solar nebula (Bryden et al., 1999; Lubow et al.,
1999), Canup and Ward (2002) showed that the
jovian satellites could have accreted over an extended period of time ($106 years) in a “gasstarved” subnebular disk. Because their subnebula is not as massive in comparison with the
satellites throughout most of its lifetime, satellites
as large as Europa and Ganymede could ultimately accrete and survive. While temperatures
in the inner jovian subnebula might have been
hot enough to vaporize water ice (or prevent its
condensation) (Canup and Ward, 2002), the densities and pressures in their models are low
enough to kinetically inhibit thermochemical
equilibrium among the non-ice phases during the
brief subnebular lifetime of any particular solid
particle (brief when compared with the total accretion time scale).
The implication is that Europa and the other
icy Galileans accreted from solar nebular condensates that underwent little further chemical
processing in the subnebula. Conceivably, the accreting materials could have been pristine solar
nebular condensates at 5 AU, but given that
Jupiter’s formation occurred several million years
after the formation of the solar nebula (in the
dominant core-accretion model), it is plausible
that the materials underwent some thermal and
aqueous processing on (solar-orbiting) planetesimal parent bodies. One could also ask whether
the solar nebula near Jupiter at this late stage contained sufficient solids of small enough sizes to
be entrained in the accretion flow across Jupiter’s
tidal gap. While this requires study, it seems
likely that such a population of small particles
would have been created continually through collisions among larger planetesimals, especially if
their velocities were stirred up (increased) by the
now massive Jupiter or density waves generated
in the nebula (cf. Consolmagno et al., 2003).
It should be noted that the surfaces of the dark
P and D asteroids of the outer asteroid belt and
883
SULFATE CONTENT OF EUROPA’S OCEAN
Trojan clouds are interpreted spectrally as consisting of largely anhydrous minerals and water
ice [i.e., surfaces that have not been significantly
aqueously altered aqueous (Jones et al., 1990;
Lebofsky et al., 1990)]. Presumably, these asteroids formed sufficiently late or far enough from
the Sun that heating from the decay of short-lived
radionuclides, such as 26 Al, or electromagnetic induction was ineffective (as opposed to the heavily altered C and related type asteroids of the
main belt). [Heating due to the decay of longlived radionuclides is only important in this regard for the largest asteroids (see, e.g., McKinnon, 2002a).] The spectral data obtained to date
support this hypothesis (e.g., Rivkin et al., 2002),
though some amount of hydrated minerals may
have escaped detection on very dark surfaces
(Cruikshank et al., 2001). Given that collisional
evolution would have exposed the aqueously altered interiors of P- and D-type asteroids if such
interiors formed, the implication of the formation
model of Canup and Ward (2002) is that much of
the material that accreted to build the Galilean
satellites was a mixture of essentially primordial
solar nebular solids: largely anhydrous silicates
(Prinn and Fegley, 1981), Ni-Fe alloy, iron sulfide
(e.g., Lauretta et al., 1996), water ice, and carbonaceous matter (Prinn and Fegley, 1989; Fegley, 1999a) (and CO2 as clathrate or ice if the
solar nebular temperature were low enough). [Indeed, the model of Hersant et al. (2001) predicts
that temperatures in the solar nebula near Jupiter
ultimately drop below 50 K, but this model neglects the heating due to irradiation of the flared
surface of the solar nebula by the Sun (see
D’Alessio et al., 1999).]
The “traditional” accretion disk origin for regular satellites has also been recently reexamined
by Mosqueira and Estrada (2003a,b). They argue
that once the subnebula cools sufficiently, satellites form so rapidly (,103–4 years for Europa)
that gas drag loss is no longer an issue; furthermore, they argue that once large satellites form,
they may open tidal gaps in the subnebula and,
in the Io-to-Ganymede region at least, deplete the
gas in the inner nebula. This depletion stabilizes
these large satellites against so-called type II migration over the lifetime of the jovian subnebula
(see Ward, 1997). In the model of Mosqueira and
Estrada (2003a), the solids that built the satellites
came in as planetesimals during the hydrodynamic collapse of proto-Jupiter’s gas envelope,
breaking up and dissolving in the envelope depending on their strength and volatility. It is an
open question, given the short time scales for
hydrodynamic collapse and satellite formation
(,105 years), as to whether thermochemical equilibrium would have been achieved as proposed
in the jovian subnebula model of Prinn and Fegley (1981, 1989). A study of thermochemistry during formation of Saturn’s accretion disk found
that thermochemical equilibrium (e.g., CO R
CH4) was not achieved (Mousis et al., 2002). If this
result is applicable to Jupiter and to gas-grain reactions in general, then even if Europa accreted
in an initially denser (i.e., not gas-starved) subnebula (Mosqueira and Estrada, 2003a,b), its initial composition would have been dominated by
local solar nebula planetesimals. Thermochemical equilibrium must be considered on a case-bycase basis, however, and one can imagine that
over the jovian subnebular lifetime the pressure
and temperature conditions in Europa’s accretion
zone might have been favorable to magnetite formation but not serpentine formation, for example (see discussion in Fegley, 1999b).
In summary, an examination of recent work on
the origin of Jupiter and its satellites strongly indicates that, irrespective of model type or details,
the materials that built the Galilean satellites were
predominantly derived from largely unprocessed
solar nebula solids and planetesimals. These solids
and planetesimals were drawn mostly from the
solar nebula region near Jupiter. In the jovian subnebula the solids were first heated and then
cooled, but the major effect was most likely partial or total devolatilization and loss of water and
volatile organics, and less likely to have been
wholesale thermochemical reprocessing and
reequilibration of rock 1 metal compositions (e.g.,
oxidation of Fe and hydration of silicates). None
of the formation models predicts or implies that
sulfur would be delivered to Europa in the form
of sulfates. Despite the long-standing history, and
utility, of assuming that the accreting rock 1 metal
component of the Galilean satellites may have resembled sulfate-bearing carbonaceous chondrite
(e.g., Fanale et al., 1977; Mueller and McKinnon,
1988; Kargel, 1991; Kargel et al., 2000), we conclude
that this assumption is no longer warranted.
COULD SULFATES HAVE
FORMED ON EUROPA?
If sulfates were not originally incorporated into
Europa, the question is then whether they could
have formed on Europa itself. To address this
884
MCKINNON AND ZOLENSKY
question, we first reexamine models for sulfate
formation on carbonaceous chondrite parent bodies. As mentioned above, published theoretical
models of aqueous alteration of carbonaceous
meteorite compositions have yet to predict sulfates as a reaction product. Late-stage precipitation of sulfates in CI meteorites is therefore typically explained as the result of an unspecified
late-stage oxidation event (Zolensky et al., 1989,
1993), if the sulfates are indeed extraterrestrial in
origin (Gounelle and Zolensky, 2001).
The reason for the non-appearance of sulfates
in theoretical models is illustrated in Fig. 2, which
is an example calculation from the thesis work of
Schulte (1997). Figure 2 illustrates the redox evolution of a closed system composed of the major
elements in CM chondrites and water with dissolved C and N. The calculation starts at low temperature and low hydrogen fugacity, fH2 (,partial pressure). Upon reaction the fH2 increases
due to oxidation reactions such as
3Fe 1 4H2 O R Fe3 O4 1 4H 2
(1)
and the fluid becomes alkaline due to the release
of OH2 (cf. Zolensky et al., 1989; Rosenberg et al.,
2001). With increasing temperature the fH2 increases further, and ultimately becomes more reducing than the fayalite-magnetite-quartz (FMQ)
buffer, meaning that magnetite, such as formed
in Reaction 1, would no longer be stable (at least
in the system Si-Fe-O-H). The aqueous oxidation
of, for example, iron to magnetite, Fe 1 FeS to
FIG. 2. fH2 (in bars) as a function of increasing temperature during aqueous alteration of CM chondrite
composition, according to the theoretical model of
Schulte (1997). PPM is the pyrrhotite-pentlandite-magnetite buffer. Modified from Schulte (1997).
tochilinite, and olivine to serpentine and brucite
all release H2 , and it has been increasingly recognized that the resulting fluids are very reducing (e.g., Schulte, 1997; Rosenberg et al., 2001). On
Earth, the aqueous alteration of peridotite can
lead to sufficient hydrogen accumulation to stabilize and form native iron (see Frost, 1985;
O’Hanley, 1996; McCollom and Seewald, 2001),
which is otherwise rarely found on the Earth’s
surface. Clearly, such very reducing conditions
are no way to form sulfate.
In Fig. 2, fH2 values can exceed 1 bar, which
leads to a plausible mechanism for sulfate formation on small asteroids: hydrogen escape (Zolotov
and Shock, 2001c). The lithostatic pressure distribution within a spherical asteroid of uniform density r is given as a function of radius, r, by:
1
2p
r2
P(r) 5 } r2GR2 1 2 }2
3
R
2 < 60 kPa 3
r
R
r
} 2 1 } 2 11 2 } 2
1}
2,000 kg/m
10 km
R
2
3
2
2
2
(2)
where R is the surface radius, and G is the gravitational constant. Hydrogen partial pressures
greater than lithostatic are not sustainable, so hydrogen must degas or vent from smaller asteroids
that were heated to the ice melting temperature
or beyond [such as by 26Al decay (see Wilson et
al., 1999; Young et al., 1999)], and indeed from the
near surface of any asteroid undergoing aqueous
alteration. From the perspective of Fig. 2, if the
fH2 is physically constrained to be low as the asteroid heats up, the asteroid evolves to become
more oxidizing with respect to FMQ or other mineral buffers. Rather than being closed, the chemical system is now open (to space), or physically
“buffered” to low fH2. Sulfate formation is, in
principle, possible, and it is a reasonable inference that hydrogen loss is a major cause of latestage oxidation events on meteorite parent bodies. Whether such events actually led to the
formation of magnesium sulfate in the low-T-altered CI chondrites is, however, questionable
(Gounelle and Zolensky, 2001).
Can the concept of oxidation by means of hydrogen loss be applied directly to a large icy satellite like Europa, as argued by Zolotov and Shock
(2001a,b, 2003)? The answer depends on a critical
examination of the conditions within an accreting
and evolving Europa and further comparison with
conditions likely obtained in carbonaceous mete-
885
SULFATE CONTENT OF EUROPA’S OCEAN
orite parent bodies. Zolotov and Shock (2001a,
2003) showed that aqueous sulfate stability is
achieved when the fH2 is low enough [typically,
between the FMQ and HM (hematite-magnetite)
buffers] at elevated temperatures and alkaline
conditions. As noted above, alkaline conditions are
a typical result of aqueous alteration of chondritic
materials. In contrast, even though the loss of exsolving H2 from aqueous fluids on small asteroids
is guaranteed, the pressures within Europa are
comparatively great. The central pressure within a
primordial Europa of uniform composition would
have been, based on Eq. 2, on the order of 3 GPa.
Values of fH2 ,l bar or somewhat greater should
have been easily sustainable against exsolution in
Europa’s deeper interior, which would not have allowed for sulfate stability (e.g., Fig. 2). Major loss
of H2 from Europa’s ocean would have been physically possible, but drawdown of dissolved hydrogen to low enough fH2 for sulfate stability in the
rocky interior would have been restricted to depths
in diffusive or hydrothermal communication with
the ocean (we quantify this below). In summary,
sulfate formation on Europa is indeed theoretically
possible via the mechanism of Zolotov and Shock
(2001a,c, 2003). However, the proportion of accreted sulfide (and sulfur in organics) that could be
so converted requires careful evaluation.
the incoming planetesimals, h the fraction of the
impact energy retained (Schubert et al., 1981;
–
McKinnon et al., 1997), and Cp(T) the mean heat
capacity averaged over the interval T 2 Te 5 DT:
1
–
Cp (T) 5 }
DT
5E
T
Te
CpdT 1 DHpc
6
(4)
where DHpc represents the sum of all enthalpy
changes due to phase transitions between T and
Te . The first term in the parentheses in Eq. 3 is the
gravitational potential energy of the incoming
material, per unit mass, while the second term is
the specific kinetic energy at infinity. The latter
can be neglected if the jovian subnebula was dynamically cold, which would pertain to accretion
by sweep-up of smaller satellitesimals (as opposed to oligarchic growth). The heat capacity
term in Eq. 4 is averaged because the heat capacities of natural materials such as ice/water or
rock are temperature-dependent and to account
for the latent heat of melting and vaporization of
the water ice fraction.
Figure 3 shows the increase in accretion temperature with increasing radius as a function of
h. Given Europa’s distance from Jupiter, Te is
taken to be 250 K, and Europa’s density is taken
to be 3 3 103 kg m23 , representing a mixture of
THE PRIMORDIAL OCEAN
As Europa accreted, the heat released would
have melted accreted ice and led to the formation
of a primordial ocean. How much of the ice
melted and how much was able to react with
metal, sulfide, silicates, and organic matter depended on how fast Europa accreted. More than
sufficient accretional energy was available. The
potential energy, per unit mass, of a uniform, accreting, spherical satellite is 3GM/5R, where M is
the mass of the satellite. For Europa, this is 1,200
kJ kg2 1, which can be compared with the latent
heat of water-ice melting (335 kJ kg21 ) or vaporization (2,260 kJ kg2 1 at 1 bar) and the specific
heat of rock (,1 kJ kg21 K2 1 at STP).
The temperature distribution T(r) in a symmetrically accreting uniform sphere can be written as:
1
4p
h
,y.2
T(r) 5 }
} rGr2 1 }
–
2
Cp (T) 3
21T
e
(3)
where r is now the instantaneous radius, ,y. the
mean encounter velocity, Te the temperature of
FIG. 3. Accretional temperature profile for Europa as
a function of radius, r, for various impact heat retention
fractions, h. Melting and vaporization of the ice fraction
(10% by weight) are incorporated into a temperaturedependent heat capacity. Temperatures are buffered by
ice phase changes for h #0.15, whereas steam-H2 atmospheres are raised by for greater values of h. Surface temperatures are limited to lie below those established by a
convective adiabat (see Lunine and Stevenson, 1982).
886
rock 1 metal (90%) and ice (10% by mass), with
an appropriate (but simplified) heat capacity of
[1.4 1 0.2 3 (T/250 K)] kJ kg2 1 K21 [based on
Eisenberg and Kauzmann (1969) and Robie et al.
(1978), incorporating water-ice phase changes,
and valid up to ,800 K.] [Technically, 250 K lies
above the water-ice condensation temperature at
Europa’s position in the subnebula (Prinn and
Fegley, 1981; Canup and Ward, 2002), but we assume that Europa accreted materials from a range
of semimajor axes. The precise value of Te does
not affect our arguments.] The factor h is empirical, but can be derived from more detailed calculations of the deposition and redistribution of
impact energy, plus radiative and possibly convective cooling (Lunine and Stevenson, 1982).
Based on earlier models of lunar and terrestrial
planet accretion, Schubert et al. (1981) took h
,0.1–0.4 to be plausible, while not ruling out
higher values. An explicit numerical calculation
of the temperature profile of an accreting
Ganymede by Coradini et al. (1982) implied h
,0.4. The curves in Fig. 3 do not illustrate the
moderating effect of radiative and convective
heat loss on surface temperatures at any given
time (i.e., the temperature maximum is always in
the outer portion of the accumulating body, not
at the very surface).
If Europa accreted very rapidly, as in the model
of Mosqueira and Estrada (2003a), then a larger
value of h, $0.4, is preferred. In this case almost
all of the ice should melt promptly (only the deep
interior remains cool). Considering that ,15–30%
of Europa by volume would have been ice, based
on the gravitationally derived limits to the ocean
1 shell depth today (Anderson et al., 1998), expulsion of melted and vaporized primordial ice
would have been very efficient. There would
have been limited time for chemical interaction.
The degree of chemical interaction would depend
on the fragment size distribution and permeability of the accreted materials. If the permeability
were high [samples of Tagish Lake are quite
porous, and when immersed in water disintegrate to primordial mush (Zolensky et al., 2002)]
and the grain size very small, as implied by the
altered matrix material in CI and CM chondrites
(Buseck and Hua, 1993), then at least serpentinization would have proceeded to completion
even if accreted minerals were only in contact
with hot water for 102 years [based on the experiments of Wegner and Ernst (1983)].
Internal temperatures would have been
buffered to some degree by the latent heat of
MCKINNON AND ZOLENSKY
melting and vaporization of the water ice fraction, but at the end of accretion the expelled water—the primordial ocean—would have been hot
and enveloped by a steam-H2 atmosphere (the
jovian subnebula would still be present). An upper limit to the surface temperature assumes a
convective adiabat connected the surface to the
subnebula (Lunine and Stevenson, 1982). For the
simplified case of a “dry” steam adiabat, the temperature difference DT is GM/R Cpg, where Cgp is
the steam heat capacity. Assuming the water
value (4.2 kJ kg21 K21) as a good approximation
for steam at 0.1 GPa pressures (Burnham et al.,
1969), DT #500 K by the end of accretion (Fig. 3).
In truth, DT is smaller still, as the temperature profile actually follows a gentler, but more complex,
“wet” adiabat (Lunine and Stevenson, 1982). [The
surface temperature would have almost certainly
been less than the critical temperature (Seward
and Franck, 1981), so an ocean would have existed
(as opposed to a pure steam-H2 envelope).]
The ocean that would have formed in this
rapid-accretion scenario should have contained
sulfur dissolved from soluble organics (if volatile
organics survived heating in the subnebula) plus
aqueous sulfide. Given that the solubilities of
native sulfur, HS2 , and H2 S are relatively low
(Vaughan and Craig, 1978), such a primordial
ocean could not be described as sulfur-rich. Although the high temperatures at the end of accretion thermodynamically favor the conversion
of sulfide to sulfate (Zolotov and Shock, 2003), the
high fH2 values at this time due to the presence
of subnebular H2 would have prevented it. Until
the solar nebula and jovian subnebula were banished [possibly preceded by gap opening at Europa, or across the inner subnebula (Mosqueira
and Estrada, 2003b)], Europa could not have lost
its hydrogen and undergone irreversible oxidation. We also note from Fig. 3 that at no time
do accretional temperatures exceed the Fe-FeS
eutectic (minimum melting) value, so prompt
metallic core formation and sequestration of S in
such a core did not occur.
In comparison, the protracted formation of Europa in the model of Canup and Ward (2002) allows for a much cooler Europa (h ,, 0.1), as long
as the satellitesimals are small enough that efficient conductive communication is maintained
with the growing, radiating satellite surface over
the accretion time scale (which they favor to explain an incompletely differentiated Callisto). If
the bulk of Europa accreted cold, then ice melting would have been driven by the release of ra-
887
SULFATE CONTENT OF EUROPA’S OCEAN
diogenic heat (at the rate of ,1 kJ kg2 1 Myr 21 for
U, Th, and 40 K; see Fig. 19 in Greeley et al., 2004)
and tidal heat (of unknown power) if the Laplace
resonance were primordial (Peale and Lee, 2002).
This heating would have been largely from the
inside out, and buoyancy of the released water
would have driven a large-scale hydrological
flow to the surface. How much would have made
it out depends on the extent of aqueous interaction along the way. It has long been recognized
that Europa, with a density of 3,015 kg m23 , could
be almost entirely composed of oxidized and hydrated rock (e.g., Kargel, 1991; Kargel et al., 2000).
That is, rehydrating the rock 1 metal interior of
Europa [from the models of Anderson et al.
(1998)] would essentially consume the ocean as it
exists today (or nearly so). Thus in this slow-coolaccretion-and-later-heating scenario Europa’s
primordial ocean could have been thin (compared with today’s). In terms of sulfur, it would
have been dominantly sulfidic due to the massive
quantities of H2 released during the contemporaneous alteration of the interior.
In summary, Europa’s primordial ocean was
most likely sulfidic, whether formed promptly
during accretion in the presence of the jovian subnebula or bled out over the next few 107 years as
the interior warmed and (if necessary) ice melted
and altered accreted minerals. In the former case,
a modest to substantial portion of Europa’s primordial water was likely trapped in the outer
layers in hydrated minerals (e.g., serpentine),
whereas in the latter case most of the primordial
water was likely trapped throughout the interior
(Fig. 4).
COMPOSITIONAL EVOLUTION
OF THE OCEAN
Following accretion, Europa’s primordial
ocean would have interacted hydrothermally
with the subjacent rock “core.” The hydrothermal
circulation would have been driven by the release
of radiogenic and tidal heat and by the heat
released by the serpentinization of unaltered
olivine and pyroxene (e.g., Kelley et al., 2001). The
depth of this circulation, which is necessary to deliver hydrogen-degassed oceanic water to the interior if sulfate is to form, is not known. Fractureassisted hydrothermal circulation at terrestrial
mid-ocean ridges extends to at least 10 km (300
MPa) (see Solomon et al., 2003). If we assume a
similar pressure scale for early Europan hy-
FIG. 4. Internal structures for Europa following (a) rapid accretion in a dense, minimum-mass subnebula
(Mosqueira and Estrada, 2003a,b) or (b) slow accretion in a thin, gas-starved subnebula (Canup and Ward, 2002).
In (a) ice is melted and expelled to the surface to form an ocean, with a minimum of time (,103 years) to aqueously
alter nebular or subnebular condensates; an aqueously unaltered interior is the end-member case. In (b) delayed melting of accreted ice leads to efficient alteration of the interior, though enough water should be left over to form a relatively thin ocean (and ice shell, not illustrated). The relatively high-temperature extraction in (a) implies a greater
concentration of sulfur species than in (b), but for both cases the primordial ocean is reduced and sulfidic. Rock-plusmetal interiors are modeled with the solar composition mineralogies in Mueller and McKinnon (1988).
888
drothermal systems, 300 MPa is reached at ,100
km total depth for a “thin” (,45-km-deep) ocean
overlying an oxidized and hydrated but undifferentiated solar composition interior; for a primordial anhydrous and reduced interior (an
end-member assumption for the fast-accretion
scenario; in reality, a fast-accreted Europa should
end up partially if not substantially altered), the
ocean is substantially thicker (,145 km deep),
and the same pressure is reached ,15 km below
the ocean floor [based on structural models similar to those in McKinnon and Desai (2003)]
(Fig. 4). [These models are calculated in an updated version of the ICYMOON structural code
(Mueller and McKinnon, 1988). Improvements include the ability to handle internal oceans, liquid
or solid metallic cores, variable oxidation states
(Fe partitioning), and partially differentiated
mantles.] The hydrous and anhydrous interior
models imply alteration zones equivalent to
,10% to 3.5% of Europa’s rock 1 metal (+ organics) by volume, respectively, and provide potential limits on the total amount of sulfur that
could end up in the ocean as quenched hydrothermal sulfate (Zolotov and Shock, 2003).
The total conversion of 3.5% of Europa’s sulfur
to sulfate would be a substantial input to a thick
primordial ocean [i.e., a thickness similar to today’s (see Greeley et al., 2004)]. The CI abundance
of S, on a water-free basis, is 6.6% by weight,
which corresponds to an input of ,1 3 1020 kg of
S, or ,8.5% by weight MgSO 4 if in solution as
this particular sulfate. The total conversion of
10% of Europa’s sulfur, in the case of a thin primordial ocean, provides more sulfur than can be
physically accepted in solution by the thin ocean.
We adopt the maximum possible concentration
in solution of 37% by weight MgSO 4 (at 70°C)
from Kargel (1991) as the maximum possible for
the hydrothermally interacting ocean, because either heating or cooling of this hot brine would
lead to precipitation of MgSO4 minerals and stifling of hydrothermal transport. Although this
sulfate concentration is quite high, at this stage
the bulk of Europa’s water would be locked up
at depth in hydrated silicates (in contrast to the
end-member thick ocean case, where the interior
is dry; Fig. 4). Later dehydration reactions as Europa heated up should ultimately have released
this trapped water, though, and (if no additional
sulfur is added) expanded and diluted the ocean
to ,10% by weight MgSO4.
These concentrations might be unrealistically
MCKINNON AND ZOLENSKY
high. The ability of hydrothermal systems to penetrate thoroughly altered “Europa rock” at depth
(which, in analogy with carbonaceous chondrites,
would be dominated by “matrix” serpentine,
saponite, magnetite, etc.) is an important limiting
factor. We expect that porosity would have been
largely eliminated by precipitation, mineral bulking, and pressure-induced deformation during
initial postaccretion alteration (e.g., serpentinization). The intrinsic permeability of serpentine
itself is quite poor [10212 –102 11 darcys at low
pressure and 34°C, where 1 darcy < 102 12 m2
(MacDonald and Fyfe, 1985)], and this permeability refers only to water; diffusion of NaCl and,
presumably, complexes such as SO 422 are two orders of magnitude or more slower. The critical
temperature gradient necessary for hydrothermal
circulation, from Turcotte and Schubert (2002,
Chapter 9), can be scaled to Europan conditions:
1
dT
100km
} $ 3 3 104 }
dz
b
10
m
} 2 K km
2 1}
k
2
220
2
21
(5)
where b is the thickness of the hydrothermal
layer, and k is the permeability [10220 m2 is appropriate for serpentine at 300°C (MacDonald
and Fyfe, 1985)]. It is clear from Eq. 5 that hydrothermal circulation in a massive, nonporous, serpentinitic Europan mantle is impossible. Aqueous interaction between the ocean and subjacent
rock, once serpentinization and other hydration
and oxidation reactions (e.g., magnetitization) are
close to complete, requires closely spaced penetrating fractures (which may not be possible in
hot, ductile serpentinite under pressure) and diffusion of fluid from the fractures. [Note that this
should not be a problem for meteorite parent bodies (asteroids), with their low overburden pressures and multiple avenues for fracture generation
(i.e., impacts, gas exsolution, and the volume
changes due to serpentinization itself).] Sulfate diffusion in these circumstances on Europa will be
even more limited. Based on NaCl diffusion measurements (MacDonald and Fyfe, 1985), SO422
might diffuse ,1 km in 100 Ma (at 300°C), but this
distance is likely to be a severe overestimate.
In summary, hydrothermal extraction of sulfate from thoroughly altered Europa rock may be
extremely inefficient. This is most relevant to the
slow-accretion scenario, wherein the extent of hydration and oxidation of the interior is greatest in
the postaccretionary epoch. Even for the fast-ac-
889
SULFATE CONTENT OF EUROPA’S OCEAN
cretion scenario, in which the degree of alteration
in the outer rock layers could be initially less
after accretion, hydrothermally driven serpentinization of primordial, near-surface Europarock should be geologically rapid, and may occur as fast as water can diffuse to reaction sites
(Wegner and Ernst, 1983; MacDonald and Fyfe,
1985). It follows that sulfate formation, if it is to
occur, occurs later, and is thus subject to the same
crack and porosity closure and diffusive transport restrictions.
We further stress that while sulfate formation
by this means is geochemically plausible (albeit
geophysically difficult), there remain additional
reasons for caution, as previously discussed in
part:
1. Sulfates in CI chondrites, particularly aqueously mobile sulfates, may be a product of
terrestrial alteration (Gounelle and Zolensky,
2001).
2. Sulfates are not present in the Tagish Lake meteorite (Zolensky et al., 2002).
3. In addition, sulfates are not present in the thermally metamorphosed CI chondrites Yamato
82162 and 86029 (Zolensky et al., 1993; Tonui
et al., 2003).
4. Also, sulfates are not produced in experiments
where material of CI chondritic chemistry is
reacted with water at temperatures between
300°C and 800°C, even when buffered at HM
(Scott et al., 2002).
Admittedly, the latter experiments were performed at 1.5 GPa, a pressure reached only at
great depth in Europa (,500 km), but the lack of
sulfates in naturally and experimentally heated
CI chondrite is not encouraging.
We note here that the “hot plate” experiments
of Fanale et al. (2001), in which samples of the CM
chondrite Murchison were leached and boiled,
did yield both Mg and Ca sulfate in solution. Performed in open air [an oxygen fugacity ( f O2 ) of
0.21 bar], this result is not unexpected. While a
perfectly fine demonstration of thermochemistry
in action, these experiments are of no direct relevance to any circumstances likely to have ever existed on Europa.
High temperature evolution and dehydration
As indicated above, it is also important, with
regard to Europa’s ocean chemistry, to consider
the longer-term heating and dehydrating of the
interior. Radiogenic heating alone will bring Europa’s interior to the Fe-FeS eutectic (as an example relevant to metallic core formation) by ,1 Gyr
(see Sec. 15.4 in Greeley et al., 2004). Well before
this, serpentine will break down to talc 1 olivine
1 water, at ,500–600°C for Europa interior pressures (Kitahara et al., 1966; Greenwood, 1976),
with talc and tremolite dewatering occurring at
somewhat higher temperatures, depending on
pressure (Greenwood, 1976; Jenkins et al., 1991).
For the rapid-accretion scenario, in which the
interior may be only partially altered, dewatering
may be less significant (but see below). For the
slow-accretion scenario, most of Europa’s ocean
originates at this later time, emplaced as hot aqueous fluids erupt from the interior. The fluids
would be reactive, and one of the first effects
upon release would be to lower the local fH2
(Frost, 1985), which may provide the oxidative
drive for reactions such as pyrite production:
6FeS 1 4H2 O R Fe3O4 1 3FeS 2 1 4H 2
(6)
This reaction is important below ,1,000 K at low
pressure (Lewis, 1982), and is one of the pathways
Lewis (1982) offers to explain Io’s surface S and
SO 2, in that pyrite melts incongruently at 1,016 K
(at low pressure) to form pyrrhotite and a sulfurrich liquid. This sulfur should be able to readily
escape to the surface, along with SO2 vapor in
chemical equilibrium with it (Lewis, 1982). Even
the release of small amounts of hydrated water,
accompanied by pyrite formation, could be important for Io, and thus by extension, to a rapidly
accreted Europa.
Anhydrite (CaSO4) formation is also possible
in Europa’s dehydrating interior, as well as
MgSO 4 formation if the fO2 and/or temperature
become great enough. It is not known whether
the H2 released by Reaction 6 would permit
MgSO 4 formation [modeling or experiments
along the lines of Scott et al. (2002) would be valuable in this regard]. In any event, Reaction 6 is favored at dewatering temperatures, and once any
unreacted water escapes further alteration of
pyrite to sulfate will not be possible.
We note that water escaping the heated interior of Europa should ultimately move rapidly
to the surface in fluid-filled cracks, and would
not have the opportunity to chemically interact
with upper, cooler layers beyond the vicinity of
the crack walls. Thus the aqueous fluids that
890
MCKINNON AND ZOLENSKY
would erupt at this time (potentially most of the
ocean) should also be predominantly sulfidic
(and hydrogen-charged). We also note that
pyrite formation does not interfere with metallic core formation [cores are inferred from
Galileo gravity studies for both Io and Europa
(Anderson et al., 1998, 2001; Greeley et al., 2004)].
The relevant effect, once the pyrite melts, is an
overall reduction of the amount of pyrrhotite/
troilite and an increase in the amount of magnetite. FeS is not eliminated, and the importance
of eutectic melting is maintained (see Greeley
et al., 2004).
As the interior of Europa continues to heat, the
assemblage pyrrhotite–pyrite–magnetite should
pass into the stability field of anhydrite (Lewis,
1982). The formation of anhydrite from calcium
“clinopyroxene,” CaSiO 3, would proceed by the
reactions:
CaSiO3 1 FeS +2O2 R CaSO4 1 FeSiO 3 (7)
and
3CaSiO 3 1 Fe3O4 1 3S
1 4O 2 R 3CaSO 4 1 3FeSiO 3
(8)
where the clinopyroxenes are represented as Mgfree end-members for simplicity (Luhr, 1990). Sulfur is available for Reaction 8 once pyrite melts,
but both reactions require additional oxygen. If
some H2O remains available for release from particularly temperature-stable hydrated silicates
(Ca-tremolite being the obvious candidate), then
anhydrite should form (Luhr, 1990). In this case
anhydrite becomes a stable and relatively infusible Europan mantle mineral, and would be a
reservoir of S that could be supplied to later-forming silicate melts, which in turn would be a source
of volcanic SO 2 vapor.
The implication of the high temperature evolution of Europa’s interior for the ocean models
is that venting of SO2 will occur, but unless anhydrite is an important mantle mineral, it is the
eruption of elemental sulfur that will provide the
dominant flux of sulfur to the surface (i.e., Reaction 6 vs. Reactions 7 and 8). This is potentially
important for the slowly accreted Europa scenario, depending on how far Reaction 6 proceeded, but may not be as important for the
quickly accreted Europa scenario, if little hydrated material is placed at depth. The total
amount of SO 2 vapor released to the ocean in either case is likely modest in comparison with the
permissive upper limits on dissolved sulfate derived above for hydrothermal alteration [on the
order of 0.5% by weight when crudely scaled
from Io, accounting for Europa’s H2O fraction
and presumably much lower level of silicate volcanic activity (cf. Kargel et al., 2000)]. Although
speculative at this point, in that we have no evidence for it, a massive subcrust of sulfur on Europa (beneath the ocean) is a serious possibility
(cf. Kargel et al., 2000). In our opinion, it is a more
likely barrier to present-day ocean–rock hydrothermal interactions than sulfates precipitated
from a saturated ocean, based on our upper limit
sulfate concentrations.
DISCUSSION
We have attempted to describe the possible accretionary and evolutionary paths Europa may
have taken, based on recent modeling of formation of the Galilean satellites, and focused on implications for ocean chemistry and sulfur chemistry in particular. We argue that, following initial
formation of a sulfidic ocean, inputs of oxidized
sulfur (sulfate), due to its much greater solubility, would have tipped the balance in favor of an
oxidized ocean.
Previous work also concluded that an oxidized,
sulfate-bearing ocean is likely, but with different
levels of salinity (Table 1). Kargel (1991) and
Kargel et al. (2000), using CI carbonaceous chondrite as a compositional analogue, favor extremely sulfate-rich chemistries as exemplified
by the low temperature peritectic in the H2 OMgSO 4-Na2SO4 system (~20% by weight MgSO 4).
Zolotov and Shock (2001b) based detailed ocean
chemistry models on extraction of volatiles from
a relatively unaltered (and metal-bearing) carbonaceous chondrite type (CV), with a level of
partial extraction chosen to match that appropriate for the terrestrial upper crust and ocean (with
respect to the bulk silicate Earth). Their best estimate (K1a ) model contains 0.1 molal SO 422 , equivalent to 1.2% by weight MgSO4 . Our permissive
upper limit is intermediate: #10% by weight
MgSO 4 or equivalent. [Unfortunately, the conductivity limits implied by the Galileo magnetometer discovery of an induced field at Europa
(Zimmer et al., 2000) do not provide useful constraints on sulfate concentration. For a broad
range of ocean depths consistent with the gravity data, the minimum conductivities necessary
891
SULFATE CONTENT OF EUROPA’S OCEAN
TABLE 1.
Reference
Kargel et al. (2000)
Zolotov and Shock (2001b)
Johnson et al. (2004)
This work
This work
ESTIMATED SULFATE CONCENTRATION
IN
EUROPA’S OCEAN
Amount
Source
,20 wt% MgSO4
,1.2 wt%
,0.002 wt%
(,),10 wt%
,0.5 wt%?
Peritectic sulfate melt
Partial extraction from BSEa
Iogenic contribution
Hydrothermal oxidation
SO2 gas from volcanismb
a BSE,
bulk silicate Earth (analogy).
volcanic activity is unknown but certainly less than that of Io. Sulfur volcanism (.1 kilometer thick globally averaged?) is also plausible.
b Europan
to account for the induced field are ,0.08–0.14 S
m21 , which are 1/30 to 1/20 that of terrestrial
seawater, respectively (see Greeley et al., 2004).
The implied NaCl salinity scales accordingly, and
can be met by even the partial extraction model
of Zolotov and Shock (2001b). In this case the
minimum sulfate concentration needed is 0%. If
the conductivity is due to sulfate alone, ,1% by
weight is the minimum implied.]
We reiterate, however, that the use of the CI
carbonaceous chondrite analogue is not warranted. The extraterrestrial formation of sulfates
within them, particularly soluble and thus mobile
magnesium sulfates, is questionable (Gounelle
and Zolensky, 2001). Neither is it likely that Europa accreted from rocks of similar composition
to CI chondrites, based on recent formation models (Canup and Ward, 2002; Mosqueira and
Estrada, 2003a,b). Rather, we argue that Europa
accreted from a mixture of water ice (,10% by
weight) and a mostly anhydrous precursor to
CI-chondrite-like rock (,90% by weight) (cf.
Browning and Bourcier, 1998). CI-chondrite-like
in this case means having solar abundances of
non-ice elements (unlike CV chondrites, which
are volatile depleted), but may differ in the
amount and type of carbonaceous material—Tagish Lake is a prime example (Zolensky et al., 2002).
The use of terrestrial proxies to estimate the
fractionation of elements such as S to Europa’s
surface (Zolotov and Shock, 2001b) is understandable, but a major aim of our paper has been
to consider elemental fractionation under more
plausible conditions based on our understanding
of Europa’s formation and evolution. In terms of
composition, accretion, and subsequent thermal
and chemical history, we find that Europa shares
little in common with the Earth.
There are two major pathways in which sulfate
could be supplied to Europa’s ocean. One, proposed by Zolotov and Shock (2001a,c, 2003), postulates a loss of alteration-derived hydrogen to
space, so that the fH2 (partial pressure) falls into
the sulfate stability field for alkaline hydrothermal fluids. However, as we have noted, a number of difficulties arise when this pathway is considered in terms of Europa’s size, gravity, and
petrological history, as opposed to the CI parent
bodies for which the model was originally developed. If CI magnesium sulfate is indeed terrestrial, this model also loses most of its empirical motivation as well, as neither theoretical
models nor experiments have yet demonstrated
sulfate formation under conditions appropriate to
Europa or carbonaceous asteroids.
The second pathway is late oxidation of the
deep interior of Europa, via water released by the
breakdown of hydrated silicates (serpentine, talc,
tremolite) as temperatures increased because of
radiogenic and tidal heating (but prior to actual
melting and metallic core formation). Although a
hypothesis at this stage, this pathway is similar
in concept to the proposal of Lewis (1982) for deriving Io’s surface sulfur from such an oxidized
interior. Lewis (1982) conceived the immediate
source of the sulfur magma on Io to be a material similar in mineralogy and oxidation state to
CM chondrites (with or without C or H2 O). He
also allowed for a somewhat less oxidized interior as long as the source region contained pyrite,
which yields a sulfur-rich melt upon heating. Io
has an oxidized, sulfur- and SO2 -bearing surface.
This fact and especially the observation that volcanic gases from Io’s upper mantle indicate an
fO2 similar to the low end of the range found in
the Earth’s upper mantle [,2–3 log units below
FMQ (Zolotov and Fegley, 2000; cf. Blundy et al.,
1991)], indicate that Io’s interior is indeed oxidized with respect to metal-bearing chondritic
compositions. On this basis, the second pathway
for supplying oxidized sulfur (though not necessarily sulfate) to Europa’s ocean is on sounder observational footing. In this case, most of the sulfur reaching the surface should have been in the
892
form of sulfur magma, but presumably enough
gaseous SO 2 would have been supplied to oxidize the ocean.
[That Io’s mantle fO2 is now below FMQ is consistent with the evolutionary scenario of Lewis
(1982). Following extraction of S to the surface
and Fe3O4 to the core, and in the absence of anhydrite, the only source of labile oxygen would
be FeO in silicates and minor amounts of Fe2 O3
in garnet and spinel (magnetite being an endmember) (Blundy et al., 1991). Continuous loss of
SO 2 to the Io torus, due to tidal heating and magnetospheric sputtering, is more than sufficient to
deplete the amount of Fe31 in spinel, if in similar abundance to that in the Earth’s upper mantle, and thus to drive the fO2 of Io’s mantle from
FMQ to its present value over geologic time.]
Europa accreted more water than Io. If this water was largely trapped at depth in hydrated silicates, which is most likely if Europa accreted
slowly in a gas-starved subnebula, then the potential existed for a more complete conversion of
FeS to pyrite and magnetite (if not anhydrite), and
thus, an even thicker surface layer of S than seen
on Io today. This sulfur would have formed a subcrust at the base of the ocean. Although it may
seem an exotic possibility, it should be remembered that S is cosmically abundant. It is mostly
terrestrial prejudice that prevents us from seeing
S as a potentially important crust-forming material (as the Earth, and its mantle in particular, is
sulfur depleted).
We have so far said little of the role of carbon
in Europa. Although largely beyond the scope of
this paper, we point out that carbon may have
played an important role, especially if substantial
amounts of involatile organic matter or graphite
were accreted. Carbon, if present in sufficient
amounts, may have buffered the fO2 of the interior during its thermal evolution, although Lewis
(1982) argued for Io that C would have been degassed at high temperatures (.1,000 K) as CO2 .
At lower temperatures or higher pressures (depending on the oxidation state), C could have
been retained within Europa and carbonate could
have formed and acted as a sink—and thus competed with S—for Ca, Mg, Fe, etc. (e.g., Blundy et
al., 1991; Kargel et al., 2000; Scott et al., 2002).
Whether buffering to lower fO2 or promoting carbonate at the expense of sulfate, the overall effect
of carbon would have been to reinforce the upper limits on the amount of possible SO 422 in the
ocean, as derived above.
MCKINNON AND ZOLENSKY
Eventually, Europa’s rocky interior would
have dehydrated and partially melted to produce
an Fe-rich core and a silicate mantle. If tidal heating is sufficient, partial melting in the mantle and
present-day volcanism are possible (see Greeley
et al., 2004). Unless completely sealed off by a
massive sulfur layer, hydrothermal interaction
between the ocean and volcanic rock is likely. If
terrestrial seafloor hydrothermal interactions are
any guide, the oxidation state of circulating fluids would be buffered by the volcanic host rock
to reducing values (fO2 at or below FMQ) for temperatures below 500°C, but with greater degrees
of oxidation possible at higher T (McCollom and
Shock, 1998).
If hydrothermal fluids were to circulate mainly
in fresh igneous units typical of the terrestrial
seafloor, sulfate would not survive in Europa’s
ocean. The equilibration of hydrothermal fluids
with such rocks at lower temperature would convert sulfate to relatively insoluble sulfide [and
terrestrial seafloor black smokers are, of course,
venting dissolved sulfides (e.g., Von Damm,
1990)]. On Europa, at least one of three things
must be happening to preserve oceanic sulfate:
(1) Individual hydrothermal systems must persist
at least until the rocks themselves are sufficiently
oxidized, possibly because of hydrogen loss
(Zolotov and Shock, 2003); (2) hydrothermal circulation must occur in relatively oxidized [e.g.,
anhydrite-bearing (Luhr, 1990)] igneous rock to
begin with (i.e., in igneous rocks that differ in this
important regard from their terrestrial counterparts); or, simply, (3) hydrothermal circulation
does not occur. Even assuming hydrothermal
equilibration in the stability field of sulfate, the
situation late in Europa’s history would be one of
conversion of preexisting oceanic sulfide or dissolved sulfur and SO 2, or of maintenance of
oceanic sulfate, and not one of extraction to the
ocean of new, “juvenile” sulfate. Thus our 10%
by weight MgSO 4 or equivalent upper limit on
the sulfate concentration of Europa’s evolved
ocean is not changed.
GEOLOGICAL AND
ASTROBIOLOGICAL IMPLICATIONS
The implications of the evolutionary models
and constraints we have developed for Europa’s
ocean and shell composition are profound. There
is no longer any compelling reason to invoke
SULFATE CONTENT OF EUROPA’S OCEAN
an initial hypersaline sulfate ocean on Europa.
Rather, the initial oceanic conditions were likely
to be reducing (from initial aqueous alteration) to
very reducing (from subnebular H2). Given that
the solubilities of reduced and elemental sulfur
species are low compared with those of sulfates,
the early ocean would not have been sulfur rich.
SO 2 might have entered the ocean later (following Europa’s dehydration stage), or SO422 may
have formed hydrothermally to a modest degree
from existing, more reduced sulfur compounds.
Given these constraints, the total sulfate content
of the ocean would thus almost certainly be less
than the eutectic compositions proposed by
Kargel et al. (2000). We derived a generous upper
limit of 10% by weight MgSO4 and, while it is difficult to be more precise, argued that the total sulfate content was likely far less.
In principle the ocean may have remained sulfidic from its inception, but the example of Io is
a strong argument that there are processes that
can deliver oxidized sulfur compounds (S and
SO 2 ) to the outer rocky layers of a Galilean satellite core. This is especially true for a slowly accreted Europa, appropriate to formation in a
gas-starved disk (Canup and Ward, 2002), in
which most of Europa’s water would have initially been trapped at depth, predominantly as
hydrated silicates. Later hydrothermal conversion of an oxidized, sulfate-bearing ocean to a
reduced, sulfidic one is also possible if Europa’s
rocky interior became sufficiently volcanically
active and the volcanic rocks were sufficiently
reducing. The opposite is also true, in that a sulfidic ocean could be converted to a sulfate-bearing one via hydrothermal interaction with sufficiently oxidized volcanic rocks. The petrological
and hydrological evolutions of Europa’s interior
require detailed thermochemical and physical
modeling before more definitive conclusions can
be drawn.
In the case of less-than-eutectic concentrations,
partial freezing of the ocean will tend to create a
floating shell of pure ice. Geological processes, of
course, allow for rapid freezing of ocean water
and incorporation of existing sulfates and other
salts into the shell as well in certain circumstances. For example, in shell separation and
wedge-shaped band formation (Schenk and McKinnon, 1989; Sullivan et al., 1998; Prockter et al.,
2002), it is plausible that some oceanic water rises
in vertical dikes and freezes against cold ice walls.
During chaos formation, oceanic waters could
893
reach the surface or near surface, rapidly freeze,
and become incorporated in the shell (e.g., Greenberg et al., 1999; O’Brien et al., 2002). Thus, the
shell could, locally or regionally, have nearly the
same composition as the ocean. As long as the
composition is more ice-rich than the eutectic,
however, any thermal processes that lead to melting should create a relatively dense brine that
would tend to drain down toward the ocean (e.g.,
Head and Pappalardo, 1999). Thus, with further
geological processing of this sort, a sulfate-bearing shell would naturally tend to evolve towards
a sulfate-poor or sulfate-free composition.
Frozen hydrated sulfate can explain Europa’s
infrared signatures (e.g., McCord et al., 2002). It is
also conceivable that the sulfate observed at the
surface is produced radiolytically (Johnson et al.,
2004) from frozen ocean water containing dominantly reduced sulfur species, but whether aqueous sulfur concentrations would be sufficient in
this case is doubtful. Johnson et al. (2004), in fact,
argue for a predominantly Iogenic sulfur source.
They estimate a maximum S flux, at Europa’s antapex of orbital motion, equivalent to a 1 cm thickness of sulfur per 107 years. Generously assuming
this implantation is global and that the geological
turnover time for the surface is also 107 years, this
implies that ~3 3 104 kg of radiolytically produced
H2SO4 per m2 has been supplied to the ocean over
geologic time. The maximum concentration of radiolytic sulfate that could build up in the Europan
ocean would be small, ,0.002% by weight.
From an astrobiological perspective, the concentration of sulfate salts in the oceanic layer is
likely to be low to moderate at best [and alkali
halides concentrations should be low compared
with terrestrial values simply due to dilution
(Kargel et al., 2000; Zolotov and Shock, 2001b)].
Massive sulfate hydrate beds at the seafloor, which
might otherwise have formed during ocean cooling (Kargel et al., 2000), should also be absent, and
the ocean should not be cut off by them from chemical interaction with the rocky interior. On the
other hand, we argue here that massive (multikilometer-thick) sulfur beds are a distinct possibility, and they would be an astrobiological disaster from the point of view of hydrothermally
hosted life (Shock et al., 2000). At the very least,
geological and astrobiological models should not
simply assume that Europa’s ocean is at or near
sulfate saturation. Rather, serious consideration
should be given to a range of lower sulfate concentrations (1–10% by weight) as well as to an
894
MCKINNON AND ZOLENSKY
SO 2-rich ocean floored by sulfur, and possibly,
SO 2 -clathrate. Further work on accretion models
and possible thermochemistries in the jovian subnebula, the possible hydrothermal origin of sulfates, and the origin of Io’s oxidized, sulfur-rich
surface would also be of great value.
ACKNOWLEDGMENTS
We thank Misha Zolotov and Everett Shock for
extensive and ongoing discussions, Mitch Schulte
for permission to use his figure in advance of
publication, and A. Ruzicka for a useful review.
This work was supported by grants from the
NASA Exobiology and Planetary Geology &
Geophysics Research Programs (W.B.M.) and
Planetary Materials & Geochemistry (M.E.Z.).
ABBREVIATIONS
fH2, hydrogen fugacity; FMQ, fayalitemagnetite-quartz; f O2 , oxygen fugacity; HM,
hematite-magnetite.
REFERENCES
Airieau, S.A., Farquhar, J., Jackson, T.L., Leshin, L.A.,
Thiemens, M.H., and Bao, H. (2001) Oxygen isotope
systematics of CI and CM chondrite sulfate: implications for evolution and mobility of water in planetesimals [abstract 1744]. In Lunar and Planetary Science Conference XXII [book on CD-ROM], Lunar and Planetary
Institute, Houston.
Anderson, J.D., Schubert, G. Jacobson, R.A., Lau, E.L.,
Moore, W.B., and Sjogren, W.L. (1998) Europa’s differentiated internal structure: inferences from four Galileo
encounters. Science 281, 2019–2022.
Anderson, J.D., Jacobson, R.A., Lau, E.L., Moore, W.B.,
and Schubert, G. (2001) Io’s gravity field and interior
structure. J. Geophys. Res. 106, 32963–32969.
Blundy, J.D., Brodholt, J.P., and Wood, B.J. (1991) Carbonfluid equilibria and the oxidation state of the upper
mantle. Nature 349, 321–324.
Brown, P.G., Hildebrand, A.R., Zolensky, M.E., Grady,
M., Clayton, R.N., Mayeda, T.K., Tagliaferri, E., Spalding, R., MacRae, N.D., Hoffman, E.L., Mittlefehldt,
D.W., Wacker, J.F., Bird, J.A., Campbell, M.D., Carpenter, R., Gingerich, H., Glatiotis, M., Greiner, E., Mazur,
M.J., McCausland, P.J., Plotkin, H., and Rubak Mazur,
T. (2000) The fall, recovery, orbit, and composition of
the Tagish Lake meteorite: a new type of carbonaceous
chondrite. Science 290, 320–325.
Browning, L. and Bourcier, W. (1998) Constraints on the
anhydrous precursor mineralogy of fine-grained materials in carbonaceous chondrites. Meteor. Planet. Sci. 33,
1213–1220.
Burnham, C.W., Holloway, J.R., and Davis, N.F. (1969)
Thermodynamic properties of water to 1000°C and
10,000 bars. Geol. Soc. Am. Special Paper 132, 1–96.
Bryden, G., Chen, X., Lin, D.N.C., Nelson, R.P., and Papaloizou, J.C.B. (1999) Tidally induced gap formation
in protostellar disks: gap clearing and suppression of
protoplanetary growth. Astrophys. J. 514, 344–367.
Buseck, P.R. and Hua, X. (1993) Matrices of carbonaceous
chondrite meteorites. Annu. Rev. Earth Planet. Sci. 21,
255–305.
Canup, R.M. and Ward, W.R. (2002) Formation of the
Galilean satellites: conditions of accretion. Astron. J. 124,
3404–3423.
Consolmagno, G.J., Weidenschilling, S.J., and Britt, D.T.
(2003) Forming well-compacted meteorites by shock
events in the solar nebula. Meteor. Planet. Sci. 38(Suppl.),
A128.
Coradini, A., Federico, C., and Lanciano, P. (1982)
Ganymede and Callisto: accumulation heat content. In
Comparative Study of the Planets, edited by A. Coradini
and M. Fulchignoni, Reidel, Dordrecht, The Netherlands, pp. 61–70.
Cruikshank, D.P., Dalle Ore, C.M., Roush, T.L., Geballe,
T.R., and Owen, T.C. (2001) Constraints on the composition of Trojan asteroid 624 Hektor. Icarus 153, 348–360.
D’Alessio, P., Calvet, N., Hartmann, L., Lizano, S., and
Canto, J. (1999) Accretion disks and young objects. II.
Tests of well-mixed models with ISM dust. Astrophys.
J. 527, 893–909.
Eisenberg, D. and Kauzmann, W. (1969) The Structure and
Properties of Water, Oxford University Press, New York.
Fanale, F.P., Johnson, T.V., and Matson, D.L. (1977) Io’s
surface and the histories of the Galilean satellites. In
Planetary Satellites, edited by J.A. Burns, University of
Arizona Press, Tucson, pp. 379–405.
Fanale, F.P., Li, Y.-H., De Carlo, E., Farley, C., Sharma,
S.K., Horton, K., and Granahan, J.C. (2001) An experimental estimate of Europa’s “ocean” composition independent of Galileo orbital remote sensing. J. Geophys.
Res. 106, 14595–14600.
Fegley, B. (1998) Iron grain catalyzed methane formation
in the jovian protoplanetary subnebulae and the origin
of methane on Titan. Bull. Am. Astron. Soc. 30, 1092.
Fegley, B. (1999a) Chemical and physical processing of
presolar materials in the solar nebula and the implications for preservation of presolar materials in comets.
Space Sci. Rev. 90, 239–252.
Fegley, B. (1999b) Kinetics of gas-grain reactions in the solar nebula. Space Sci. Rev. 92, 177–200.
Fredericksson, K. and Kerridge, J.F. (1988) Carbonates
and sulfates in CI chondrites: formation by aqueous activity on the parent body. Meteoritics 23, 35–44.
Frost, B.R. (1985) On the stability of sulfides, oxides, and
native metals in serpentinite. J. Petrol. 26, 31–63.
Gounelle, M. and Zolensky, M.E. (2001) A terrestrial origin for sulfate veins in CI1 chondrites. Meteor. Planet.
Sci. 36, 1321–1329.
SULFATE CONTENT OF EUROPA’S OCEAN
Gounelle, M., Zolensky, M.E., Tonui, E., and Mikouchi, T.
(2001) Mineralogy of Tagish Lake, a unique type 2 carbonaceous chondrite [abstract 1616]. In Lunar and Planetary Science Conference XXXII [book on CD-ROM], Lunar and Planetary Institute, Houston.
Gradie, J.C., Chapman, C.R., and Tedesco, E.F. (1989) Distribution of taxonomic classes and the compositional
structure of the asteroid belt. In Asteroids II, edited by
R.P. Binzel, T. Gehrels, and M.S. Matthews, University
of Arizona Press, Tucson, pp. 316–335.
Greeley, R., Chyba, C., Head, J.W., McCord, T., McKinnon, W.B., Pappalardo, R.T., and Figueredo, P.H. (2004)
Geology of Europa. In Jupiter—The Planet, Satellites, and
Magnetosphere, edited by F. Bagenal, T.E. Dowling, and
W.B. McKinnon, Cambridge University Press, Cambridge, UK, in press.
Greenberg, R., Hoppa, G.V., Tufts, B.R., Geissler, P., Riley, J., and Kadel, S. (1999) Chaos on Europa. Icarus 141,
263–286.
Greenwood, H.J. (1976) Metamorphism at moderate temperatures and pressures. In The Evolution of the Crystalline Rocks, edited by D.K. Bailey and R. Macdonald,
Academic Press, London, pp. 187–259.
Head, J.W. and Pappalardo, R.T. (1999) Brine mobilization during lithospheric heating on Europa: implications for formation of chaos terrain, lenticula texture,
and color variations. J. Geophys. Res. 104, 27143–27155.
Hersant, F., Gautier, D., and Huré, J.-M. (2001) A two-dimensional model for the primordial nebula constrained
by D/H measurements in the solar system: implications
for the formation of giant planets. Astrophys. J. 554,
391–407.
Hiroi, T., Zolensky, M.E., and Pieters, C.M. (2001) The
Tagish Lake meteorite: a possible sample from a D-type
asteroid. Science 293, 2234–2236.
Hurlbut, C.S. (1971) Dana’s Manual of Mineralogy, 18th
edit., John Wiley & Sons, New York.
Jenkins, D.M., Holland, T.J.B., and Clare, A.K. (1991) Experimental determination of the pressure-temperature
stability field and thermochemical properties of synthetic tremolite. Am. Mineral. 76, 458–469.
Johnson, R.E., Carlson, R.W., Cooper, J.W., Paranicas, C.,
Moore, M.H., and Wong, M. (2004) Radiation effects on
the surfaces of the Galilean satellites. In Jupiter—The
Planet, Satellites, and Magnetosphere, edited by F. Bagenal, T.E. Dowling, and W.B. McKinnon, Cambridge
University Press, Cambridge, UK, in press.
Jones, T.D., Lebofsky, L.A., Lewis, J.S., and Marley, M.S.
(1990) The composition and origin of the C, P, and D
asteroids: water as a tracer of thermal evolution in the
outer belt. Icarus 88, 172–192.
Kargel, J.S. (1991) Brine volcanism and the interior structure of asteroids and icy satellites. Icarus 94, 368–390.
Kargel, J.S., Kaye, J.Z., Head, J.W., Marion, G.M., Sassen,
R., Crowley, J.K., Ballesteros, O.P., Grant, S.A., and
Hogenboom, D.L. (2000) Europa’s crust and ocean: origin, composition, and the prospects for life. Icarus 148,
226–265.
Keller, L.P. and Flynn, G.J. (2001) Matrix mineralogy of
the Tagish Lake carbonaceous chondrite: TEM and
895
FTIR studies [abstract 1639]. In Lunar and Planetary Science Conference XXXII [book on CD-ROM], Lunar and
Planetary Institute, Houston.
Kelley, D.S., Karson, J.A., Blackman, D.K., Fruh-Green,
G.L., Butterfield, D.A., Lilley, M.D., Olson, E.J.,
Schrenk, M.O., Roe, K.K., Lebon, G.T., Rivizzigno, P.,
and the AT3–60 Shipboard Party (2001) An off-axis hydrothermal vent field near the Mid-Atlantic ridge at 30°
N. Nature 412, 145–149.
Kitahara, S., Takanouchi, S., and Kennedy, G.C. (1966)
Phase relations in the system MgO-SiO2–H2O at high
temperatures and pressures. Am. J. Sci. 264, 223–233.
Lauretta, D.S., Kremser, D.T., and Fegley, B. (1996) The
rate of iron sulfide formation in the solar nebula. Icarus
122, 288–315.
Lebofsky, L.A., Jones, T.D., Owensby, P.D., Feierberg,
M.A., and Consolmagno, G.J. (1990) The nature of low
albedo asteroids from 3-mm spectrophotometry. Icarus
83, 12–26.
Lee, M.R. (1993) The petrography, mineralogy and origins of calcium sulphate within the Cold Bokkeveld CM
carbonaceous chondrite. Meteoritics 28, 53–62.
Lewis, J.S. (1971) Satellites of the outer planets: their physical and chemical nature. Icarus 15, 174–185.
Lewis, J.S. (1982) Io: geochemistry of sulfur. Icarus 50,
105–114.
Lewis, J.S. and Prinn, R.G. (1984) Planets and Their Atmospheres: Origin and Evolution, Academic Press, Orlando, FL.
Lubow, S.H., Siebert, M., and Artymowicz, P. (1999) Disk
accretion onto high-mass planets. Astrophys. J. 526,
1001–1012.
Luhr, J.F. (1990) Experimental phase relations of waterand sulfur-saturated arc magmas and the 1982 eruptions of El Chichón Volcano. J. Petrol. 31, 1071–1114.
Lunine, J.I. and Stevenson, D.J. (1982) Formation of the
Galilean satellites in a gaseous nebula. Icarus 52, 14–39.
MacDonald, A.H. and Fyfe, W.S. (1985) Rate of serpentinization in seafloor environments. Tectonophysics 116,
123–135.
MacDougall, J.D., Laugmair, G.W., and Kerridge, J.F.
(1984) Early solar system activity: Sr isotopic evidence
from the Orgueil CI meteorite. Nature 307, 249–251.
McCollom, T.M. and Seewald, J.S. (2001) A reassessment
of the potential for reduction of dissolved CO2 to hydrocarbons during serpentinization of olivine. Geochim.
Cosmochim. Acta 65, 3769–3778.
McCollom, T.M. and Shock, E.L. (1998) Fluid-rock interactions in the lower ocean crust: thermodynamic models of hydrothermal interaction. J. Geophys. Res. 103,
547–575.
McCord, T.B., Hansen, G.B., Fanale, F.P., Carlson, R.W.,
Matson, D.L., Johnson, T.V., Smythe, W.D., Crowley,
J.K., Martin, P.D., Ocampo, A., Hibbitts, C.A., Granahan, J.C., and the NIMS Team (1998) Salts on Europa’s
surface detected by Galileo’s Near Infrared Mapping
Spectrometer. Science 280, 1242–1245.
McCord, T.B., Hansen, G.B., Matson, D.L., Johnson, T.V.,
Crowley, J.K., Fanale, F.P., Carlson, R.W., Smythe,
W.D., Martin, P.D., Hibbitts, C.A., Granahan, J.C.,
Ocampo, A., and the NIMS Team (1999) Hydrated salt
896
minerals on Europa’s surface from the Galileo Near-Infrared Mapping Spectrometer (NIMS) investigation. J.
Geophys. Res. 104, 11827–11851.
McCord, T.B., Orlando, T.M., Teeter, G., Hansen, G.B.,
Sieger, M.T., Petrik, N.G., and Keulen, L.V. (2001) Thermal and radiation stability of the hydrated salt minerals epsomite, mirabilite, and natron under Europa environmental conditions. J. Geophy. Res. 106, 3311–3319.
McCord, T.B., Teeter, G., Hansen, G.B., Sieger, M.T., and
Orlando, T.M. (2002) Brines exposed to Europa surface
conditions. J. Geophy. Res. 107 (E1), 4.1–4.6, doi:10.1029/
2000JE001453.
McKinnon, W.B. (2002a) On the initial thermal evolution
of Kuiper Belt objects. In Proceedings of Asteroids, Comets,
Meteors (ACM 2002), Publication SP-500, ESA Publications Division, Noordwijk, The Netherlands, pp. 29–38.
McKinnon, W.B. (2002b) Sulfate content of Europa’s ocean
and shell: geological and astrobiological implications.
Bull. Am. Astron. Soc. 34, 914–915.
McKinnon, W.B. and Desai, S.S. (2003) Internal structures
of Galilean satellites: what can we really tell? [abstract
2104]. In Lunar and Planetary Science Conference XXXIII
[book on CD-ROM], Lunar and Planetary Institute,
Houston.
McKinnon, W.B., Simonelli, D.P., and Schubert, G. (1997)
Composition, internal structure, and thermal evolution
of Pluto and Charon. In Pluto and Charon, edited by S.A.
Stern and D.J. Tholen, University of Arizona Press, Tucson, pp. 295–343.
Mosqueira, I. and Estrada, P.R. (2003a) Formation of the
regular satellites of giant planets in an extended
gaseous nebula. I: Subnebula model and accretion of
satellites. Icarus 163, 198–231.
Mosqueira, I. and Estrada, P.R. (2003b) Formation of the
regular satellites of giant planets in an extended
gaseous nebula. II: Satellite migration and survival.
Icarus 163, 232–255.
Mousis, O., Gautier, D., and Bockelée-Morvan, D. (2002)
An evolutionary turbulent model of Saturn’s subnebula: implications for the origin of the atmosphere of Titan. Icarus 156, 162–175.
Mueller, S. and McKinnon, W.B. (1988) Three-layered
models of Ganymede and Callisto: compositions, structures, and aspects of evolution. Icarus 76, 437–464.
O’Brien, D.P., Geissler, P., and Greenberg, R. (2002) A
melt-through model for chaos formation on Europa.
Icarus 156, 152–161.
O’Hanley, D.S. (1996) Serpentinites: Records of Tectonic and
Petrological History, Oxford University Press, New York.
Peale, S.J. and Lee, M.H. (2002) A primordial origin of the
Laplace relation among the Galilean satellites. Science
298, 593–597.
Pollack, J.B. and Reynolds, R.T. (1974) Implications of
Jupiter’s early contraction history for the composition
of the Galilean satellites. Icarus 21, 248–253.
Prinn, R.G. and Fegley, B. (1981) Kinetic inhibition of CO
and N2 reduction in circumplanetary nebulae: implications for satellite composition. Astrophys. J. 249, 308–317.
Prinn, R.G. and Fegley, B. (1989) Solar nebula chemistry:
origin of planetary, satellite, and cometary volatiles. In
MCKINNON AND ZOLENSKY
Origin and Evolution of Planetary and Satellite Atmospheres,
edited by S.K. Atreya, J.B. Pollack, and M.S. Matthews,
University of Arizona Press, Tucson, pp. 78–136.
Prockter, L.M., Head, J.W., Pappalardo, R.T., Sullivan,
R.J., Clifton, A.E., Giese, B., Wagner, R., and Neukum,
G. (2002) Morphology of Europan bands at high resolution: a mid-ocean ridge-type rift mechanism. J. Geophys. Res. 107(E5), 5028, doi: 10.1029/2000JE001458.
Rivkin, A.S., Howell, E.S., Vilas, F., and Lebofsky, L.A.
(2002) Hydrated minerals on asteroids. In Asteroids III,
edited by W.F. Bottke, A. Cellino, P. Paolicchi, and R.P.
Binzel, University of Arizona Press, Tucson, pp. 235–253.
Robie, R.A., Hemingway, B.S., and Fisher, J.R. (1978)
Thermodynamic properties of minerals and related
substances at 298.15 K and 1 bar (105 Pascals) pressure
and at higher temperatures. U.S. Geol. Surv. Bull. 1452,
1–135.
Rosenberg, N.D., Browning, L., and Bourcier, W.L. (2001)
Modeling aqueous alteration of CM carbonaceous
chondrites. Meteor. Planet. Sci. 36, 239–244.
Schenk, P.M. and McKinnon, W.B. (1989) Fault offsets and
lateral crustal movement on Europa: evidence for a mobile ice shell. Icarus 79, 75–100.
Schubert, G., Stevenson, D.J., and Ellsworth, K. (1981) Internal structures of the Galilean satellites. Icarus 47,
46–59.
Schulte, M.D. (1997) Synthesis and processing of aqueous
organic compounds during water/rock reactions
[Ph.D. Thesis], Washington University, St. Louis.
Scott, H.P., Williams, Q., and Ryerson, F.J. (2002) Experimental constraints on the chemical evolution of large
icy satellites. Earth Planet. Sci. Lett. 203, 399–412.
Seward, T.M. and Franck, E.U. (1981) The system hydrogen—water up to 440°C and 2500 bar pressure. Ber.
Bunsenges Phys. Chem. 85, 2–7.
Shock, E.L., Amend, J.P., and Zolotov, M.Y. (2000) The
early Earth vs. the origin of life. In Origin of the Earth
and Moon, edited by R.M. Canup and K. Righter, University of Arizona Press, Tucson, pp. 527–543.
Solomon, S.C., Aharonson, O., Banerdt, W.B., Dombard,
A.J., Frey, H.V., Golombek, M.P., Hauck, S.A., Head,
J.W., Johnson, C.L., McGovern, P.J., Phillips, R.J., Smith,
D.E., and Zuber, M.T. (2003) Why are there so few magnetic anomalies in martian lowlands and basins? [abstract 1382] In Lunar and Planetary Science Conference
XXXIII [book on CD-ROM], Lunar and Planetary Institute, Houston.
Spaun, N.A. and Head, J.W. (2001) A model of Europa’s
crustal structure: recent Galileo results and implications
for an ocean. J. Geophys. Res. 106, 7567–7576.
Stevenson, D.J., Harris, A.W., and Lunine, J.I. (1986) Origins of satellites. In Satellites, edited by J.A. Burns and
M.S. Matthews, University of Arizona Press, Tucson,
pp. 39–88.
Sullivan, R., Greeley, R., Homan, K., Klemaszewski, J.,
Belton, M.J.S., Carr, M.H., Chapman, C.R., Tufts, R.,
Head, J.W., Pappalardo, R., Moore, J., Thomas, P., and
the Galileo Imaging Team (1998) Episodic plate separation and fracture infill on the surface of Europa. Nature 391, 371–373.
SULFATE CONTENT OF EUROPA’S OCEAN
Tonui, E.K., Zolensky, M.E., Lipschutz, M.E., Wang,
M.-S., and Nakamura, T. (2003) Yamato 86029: aqueously altered and thermally metamorphosed CI-like
chondrite with unusual textures. Meteor. Planet. Sci. 38,
269–292.
Turcotte, D.L. and Schubert, G. (2002) Geodynamics, 2nd
edit., John Wiley & Sons, New York.
Vaughan, D.J. and Craig, J.R. (1978) Mineral Chemistry of
Metal Sulfides, Cambridge University Press, Cambridge,
UK.
Von Damm, K.L. (1990) Seafloor hydrothermal activity;
black smoker chemistry and chimneys. Annu. Rev. Earth
Planet. Sci. 18, 173–204.
Ward, W.R. (1997) Protoplanet migration by nebula tides.
Icarus 126, 261–281.
Wegner, W. and Ernst, W.G. (1983) Experimentally determined hydration and dehydration reaction rates in
the system MgO–SiO2–H2O. Am. J. Sci. 283–A, 151–180.
Wilson, L., Keil, K., Browning, L.B., Krot, A.N., and
Bourcier, W. (1999) Early aqueous alteration, explosive
disruption, and reprocessing of asteroids. Meteor.
Planet. Sci. 34, 541–557.
Young, E.D., Ash, R.D., England, P., and Rumble, D.
(1999) Fluid flow in chondritic parent bodies: deciphering the compositions of planetesimals. Science 286,
1331–1335.
Zimmer, C., Khurana, K.K., and Kivelson, M.G. (2000)
Subsurface oceans on Europa and Callisto: constraints
from Galileo magnetometer observations. Icarus 147,
329–347.
Zolensky, M.E. and McSween, H.Y. (1988) Aqueous alteration. In Meteorites and the Early Solar System, edited
by J.F. Kerridge and M.S. Matthews, University of Arizona Press, Tucson, pp. 114–143.
Zolensky, M.E., Bourcier, W.L., and Gooding, J.L. (1989)
Aqueous alteration of the hydrous asteroids: results of
EQ3/6 computer simulations. Icarus 78, 411–425.
Zolensky, M., Barrett, R., and Browning, L. (1993) Miner-
897
alogy and composition of matrix and chondrule rims
in carbonaceous chondrites. Geochim. Cosmochim. Acta
57, 2123–3148.
Zolensky, M.E., Nakamura, K., Gounelle, M., Mikouchi,
T., Kasama, T., Tachikawa, O., and Tonoi, E. (2002) Mineralogy of Tagish Lake: a grouped type 2 carbonaceous
chondrite. Meteor. Planet. Sci. 37, 737–761.
Zolotov, M.Yu. and Fegley, B. (2000) Eruption conditions
of Pele volcano on Io inferred from chemistry of its volcanic plume. Geophys. Res. Lett. 27, 2789–2792.
Zolotov, M.Yu. and Shock, E.L. (2001a) A hydrothermal
origin for the sulfate-rich ocean of Europa [abstract
1990]. In Lunar and Planetary Science Conference XXXII
[book on CD-ROM], Lunar and Planetary Institute,
Houston.
Zolotov, M.Y. and Shock, E.L. (2001b) Composition and
stability of salts on the surface of Europa and their
oceanic origin. J. Geophys. Res. 106, 32815–32827.
Zolotov, M.Yu. and Shock, E.L. (2001c) Geochemical constraints on the oxidation states of the europan mantle
and ocean [abstract 2025]. In Lunar and Planetary Science
Conference XXXII [book on CD-ROM], Lunar and Planetary Institute, Houston.
Zolotov, M.Y. and Shock, E.L. (2003) Energy for biologic
sulfate reduction in a hydrothermally formed ocean on
Europa. J. Geophys. Res. 108(E4), 5022, doi:10.1029/2002
JE001966.
Address reprint requests to:
Dr. William B. McKinnon
Department of Earth and Planetary Sciences and
McDonnell Center for the Space Sciences
Washington University
St. Louis, MO 63130
E-mail: [email protected]