Are there abiotically-precipitated iron-formations on Mars?

14 Pages·2014·11.85 MB·English

MineralSpectroscopy:ATributetoRogerG.Burns
©TheGeochemical Society,SpecialPublicationNo.5, 1996
Editors:M.D.Dyar,C.McCammon andM.W.Schaefer
Are there abiotically-precipitated iron-formations on Mars?
M. W. SCHAEFER
Yale University, Department of Geology and Geophysics, P. O.Box 208109, New Haven, CT06520-8109, U.S.A.
Abstract- In early times, surface conditions on Mars and Earth were probably quite similar,
however, thetwoplanets tookdifferentpathsintheirevolution.Bandediron-formations (BIFs)were
prevalent onEarth inthe Archean, and the likelihood oftheirformation on Mars atthe sametime
isinvestigated.Thereareseveralabioticmechanisms forproducingbandediron-formations onEarth.
These are: (1) oxidation of ferrous iron by free oxygen, (2) photooxidation of ferrous iron, (3)
oxidation of ferrous iron by hydrogen peroxide, and (4) anoxic precipitation of ferrous iron by
changes in pH. Any of these mechanisms could have been active on Mars in its distant past, to
produce BIFs there as well, which should be observable today. The minerals one would expect to
observe on the surface of a BIF on Mars are indeed thought to exist on Mars, but unfortunately
BIFsarenotuniqueincontaining theseminerals, andconfirmationoftheexistence ofBIFsonMars
will have to wait for more data. Proof of the existence of BIFs on Mars would give us valuable
information about the early history of the planet Mars as well as Earth. It would act to confirm
models ofMars that include awarm, wet period, and theexistence ofbodies of standing water.
INTRODUCTION
Itis likely that Earth and Mars were not dissimi-
lar when they were first formed. By studying pro-
INTHE SURFACEof Mars we see arelict of the early
cesses that occurred on the early Earth, we may
evolution of the planet. During the early history
gain insight into what processes might have been
of the solar system, conditions were probably not
important on early Mars. And conversely, as we
dissimilar on the surface of Earth and Mars, but as
obtain more detailed information about the present
time passed, the two planets diverged. Earth, larger
surface ofMars, we will learn more about the early
and with agreater source of internal heating, stayed
history of both planets. BURNS(1993) originally
active and its surface kept evolving. Mars, on the
suggested that banded iron-formations (BIFs)
other hand, smaller and with less internal heating,
could have occurred on Mars. The present paper
froze and its surface almost completely stopped
discusses the likelihood of the occurrence of BIFs
evolving. Therefore, when we look at the surface
on Mars, and what their existence might imply
of Mars, we have the opportunity not only to ob-
about the early environment of Mars.
serve the early history of Mars itself, but also to
observe alandscape perhaps more similar to that of
TERRESTRIAL BANDED IRON-FORMATIONS
early Earth than is the present terrestrial landscape.
The surface of Mars seems to have changed rela- Some of the oldest surviving sedimentary fea-
tively little over the past several billion years. Pro- tures on Earth are the banded iron-formations,
cesses that have acted on its surface in the past chemically-precipitated sedimentary deposits con-
include meteorite bombardment, which was at its taining layers of silica-rich and iron-rich bands.
peak over 3.5 Gy ago; volcanism, which is thought GROSS (1980) classifies banded iron-formations
to have largely ceased by about 1 Gy ago; and into two groups, based on associated types of rock,
fluvial and glacial processes, about which there is lithological features, and interpretation of their
active speculation (BAKERet al., 1991; BAKERet depositional environment. These two groups are
al., 1992; KARGELet al., 1995), but which neither Algoma-type and Lake-Superior-type. Both of
appear to have been active within the last 1 Gy these types exhibit cherty layers interbedded with
(although periglacial processes may be active today iron in four facies: oxides, carbonates, silicates,
(ROSSBACHERand JUDSON,1981). The role of hy- and sulfides (JAMES, 1954). In addition to the
drothermal processes is as yet unknown, however banded iron-formations, there is another type of
they could be very important, both to the chemical iron-formation called an oolitic ironstone. Oolitic
evolution of the regolith and to the isotopic evolu- ironstones are higher in aluminum than BIFs, and
tion of the atmosphere (JAKOSKYand JONES,1996). have a more clastic association. They are found in
Physical and chemical weathering, which is still rocks of avariety of ages from Proterozoic to Plio-
active today, acts only to alter the characteristics cene (MAYNARD,1983).
of the uppermost surface layers (GOODING,1978; Although BIFs are agreed tobe chemical precipi-
GOODINGet al., 1992). tates ofiron, the mechanism by which the precipita-
381
382 M. W. Schaefer
tion occurs is a controversial topic. Both biotic iron in solution. (4) Anoxic precipitation offerrous
(CLOUD, 1973; GARRELSet al., 1973; LABERGE, iron.Under anoxic conditions, itis possible to pre-
1973; LABERGEetal., 1987; ROBBINSetal., 1987) cipitate ferrous iron carbonates and silicates under
and abiotic (BRATERMANand CAIRNS-SMITH,1987; the right conditions of pH and Pco" These mecha-
CAIRNS-SMITH,1978; EUGSTERand CHOU, 1973; nisms are shown schematically in Fig. 1 and will
FRAN<;:OIS,1986; HOLLAND, 1973) mechanisms be discussed in detail below.
have been proposed. The question of whether banded iron-formations
The major idea behind both of these classes of on Earth formed primarily by abiotic or an abiotic
theories is that the atmosphere of the early Earth mechanism is still an open question. Another open
was depleted in O2, compared to the present atmo- question isthe likelihood oflife ever having existed
spheric level (PAL). Under conditions of low oxy- on Mars. The main thrust of the present paper is
gen fugacity, especially in an acidic environment, the discussion of abiotic mechanisms of making
iron exists inthe ferrous form, and is,highly soluble BIFs. Such mechanisms require fewer assump-
in water. Thus any iron input into a lacustrine or tions, when translated to Martian conditions, than
oceanic system by weathering or other process will do the biotic models. However, if life did exist on
stay in solution in the water until it reaches satura- Mars and was instrumental in causing iron precipi-
tion with respect some iron silicate or carbonate, tation in bodies of water there, then the iron forma-
orisoxidized toform ferric iron, which isrelatively tions developed would be alikely place to look for
insoluble in water. Then, for any of a variety of fossil evidence of such life.
reasons, depending on the model, the iron is oxi-
dized, and precipitates out, forming the extensive Dissolved oxygen
deposits that still exist today.
There is geological evidence that free oxygen
In the classic biotic models the factor that causes
was present in the early atmosphere (even as early
this precipitation of iron is the increase in atmo-
as 3.8 Gy ago), although not in quantities ap-
spheric O caused by the development ofphotosyn-
2 proaching PAL (ToWE, 1983). HOLLAND(1973)
thetic organisms. However, it is not necessary to
demonstrated that neither subaerial weathering or
invoke the development of life to achieve oxidation
volcanic emanations were likely to have been the
of iron in Precambrian waters. Also, there is evi-
immediate sources of soluble iron to form the Ham-
dence that very little oxygen was present in the
ersley Basin BIF of Australia. He then proposed
atmosphere during the formation of many of the
that upwelling of bottom waters in the ocean could
deposits. The presence of detrital siderite in Lake-
Superior-type deposits in Labrador and the occur-
rence of detrital grains of uraninite and pebbles of
pyrite in the 2.5-Gy-old Witwatersrand conglomer-
UVradiation
ate indicate a vanishingly low P02 content in the III
atmosphere at that time (GARRELSet al., 1973).
However, biotic models explain this by postulating
procaryotic microorganisms that reduced CO2 to
CH0 in the hydrosphere by taking electrons from surface runoff and
2
ferrous iron (CLOUD,1972, 1973).
ABIOTIC MECHANISMS OF MAKING
IRON-FORMATIONS
There are four abiotic processes that are capable
of forming BIFs. (1) Oxidation offerrous iron by
dissolved oxygen. Although this is also descriptive
of what happens when photosynthesis produces ox-
FIG. 1. Schematic diagram of the four abiotic mecha-
ygen, there are abiotic reactions that may produce nisms of BIF formation. Ferrous iron may be oxidized to
free oxygen. (2) Photooxidation offerrous iron by ferric iron in the mixing layer by means of dissolved
ultraviolet irradiation. Ultraviolet (UV) radiation oxygen orhydrogen peroxide, ornearer the surface byUV
radiation. Anoxic water (not necessarily bottom water, as
is capable of directly oxidizing ferrous iron in solu-
in the figure, if the oxygen content of the atmosphere is
tion. (3) Oxidation of ferrous iron by hydrogen low) may precipitate ferrous iron if the pH is very high
peroxide. Hydrogen peroxide (H202) is a strong (caused by evaporation) or very low (caused by an input
oxidant, and is also able to directly oxidize ferrous of acidic ground or surface water).
Iron-formations onMars?
383
have provided the necessary quantities of dissolved thought (CANUTOetal., 1983). Although the Sun's
iron to produce this deposit, and generalized this total luminosity is thought to have been lower in
mechanism to explain other Lake Superior-type de-
the past, the UV flux was likely much higher (Ta-
posits where there is little clastic debris (HOLLAND, ble 1).
1973). TOWE(1983) then finds a corollary to Hol-
CANUTOet al. (1983) performed a detailed pho-
land's proposition, namely, that free oxygen does
tochemical model of the Earth's early atmosphere
not have tobe absent from the atmosphere to main- to determine the abundance of free oxygen. They
tain the anoxic conditions necessary to transport calculated that whereas the surface O2mixing ratio
large quantities of ferrous iron in solution, instead, for an atmosphere containing 1 PAL CO2 (2.8
that the deeper basins must remain stagnant and X 10-4 bars), irradiated by the present solar UV
anoxygenic for reasonable periods of time. 15
flux, is about 10- , if an atmospheric abundance
In this model, anoxygenic bottom waters, car- of 100 PAL is assumed, irradiated by a solar UV
rying dissolved ferrous iron, circulate upwards to
flux 4 times the present value (as would be appro-
oxygen-rich surface layers in shallower regions of priate at a solar age of about 1Gy), the surface O2
the ocean, where the iron isoxidized and deposited. mixing ratio is a factor of over 103 greater (Fig.
It is not necessary to invoke photosynthesis here 2). This surface O2 mixing ratio of 3 X 10-12 is
to produce the relatively oxygen-rich surface wa- within the range of O2 values (10-13 to 10-3) re-
ters. In fact, there is evidence against photosynthe- quired to explain the simultaneous existence of oxi-
sis being an important process here. BIF oxide fa-
dized iron and reduced uranium noted earlier in the
cies are remarkably free of organic carbon (TOWE, oldest rocks (CANUTOetal., 1983). Even assuming
1983). Anoxic atmosphere models of BIF forma- the present solar UV flux, and an atmospheric
tion require that photosynthesis in the basin waters partial pressure of 100 PAL (0.028 atm) CO2, the
produce the dissolved oxygen necessary to oxidize
CANUTOet al. (1983) model calculates a surface
the iron. However, if there were large algal com- O2 mixing ratio of about 10-12, still within this
munities living in these basins, one would expect
range. The model of KASTINGet al. (1984), on the
to find evidence of them in the fossil record, which other hand, calculated that the O2 surface mixing
is not seen (ToWE, 1983). This observation is also ratio in an atmosphere containing 1000 PAL CO
supported by the very low contents of phosphorus under the present solar UV flux would be 10-11,2
in BIFs (BEUKESand KLEIN, 1992).
and that an increase in solar UV flux would act to
In the absence of photosynthesis to produce free reduce, not increase, the O2 surface mixing ratio,
oxygen in the atmosphere and dissolved in the sur-
due to the higher H2 densities produced by more
face waters, UV photolysis of water and CO can
2 rapid H20photolysis. Clearly this is a complicated
become important processes:
problem, but we seem close to understanding how
surface O2 levels in the atmosphere before the
(1) development of life could have been in the range
(2)
Table 1: Stellar UV flux as a
When this ferrous iron in solution is exposed to free function of age, from
oxygen, it will oxidize according to the following CANUTO et al. (1983).
reactions (BURNS, 1993)
Age UV
(years) enhancement
+ 21'H20 (3) 110076 510004
Fe3+ + 2H20 "" FeO(OH)(s) ~ +3H+. (4) 5 X 107 100
108 32
Measurements of UV radiation from T-Tauri
stars (thought representative of the Sun at an early 5 x 108 8
stage in its evolution) by the International Ultravio- 109 4
let Explorer (IUE) spacecraft indicate that such
5 X 109 1
stars produce far higher UV fluxes than originally
384 M.W. Schaefer
The above scenario occurs in both saline and
10.9 non-saline solutions, with UV irradiation at X.
~
'5 10·10 < 300 nffi (BRATERMANet al., 1983 and 1984).
(/) At pH values between 6.5 and 7.5, a different
1i) 10.11
~0 set of reactions becomes important (BRATERMAN
ag:> 10.12 et al., 1983 and 1984; BRATERMANand CAIRNS-
:8
::; 10.13 SMITH,1987):
gc" 10.14~ 1PAL Fe2+(aq) "" [Fe(OH)]+(aq) + W(aq) (10)
" 10.15
~
hs/
10.16 [Fe(OH)]+(aq) -+ Fe(0H)3 + W(aq). (11)
0.1 I 1 10 100
SolarUVFlux,Scaled toPresent FluxatEarth The Fe(OH)3 then forms aggregates and eventually
precipitates out as FeO(OH), and the system
FIG.2. Free oxygen mixing ratio at the surface, VS.
solar UV flux,from data in CANUTO et al. (1983). Two evolves gaseous hydrogen. The reaction rate for
cases are given, an assumed Peo, of 1PAL and one of this process is about 10 times greater than at lower
100PAL.ArangeofsolarUVfluxesfrom 1to300times pH, and longer wavelengths (X. >400 nm) are still
the present Earth value are shown. In addition, the two
effective.
bars at about 0.4 and 130present UV fluxrepresent the
Using these reactions and assuming the present
corresponding fluxes at Mars, for this range of fluxes at
Earth. solar UV flux (with 50% loss through scattering
and clouds), in terrestrial Fe2+-containing waters
having a pH > 6.5, a rate of photoprecipitation of
150-750 mg cm yr" at the tropics or 50-250
?
necessary to explain the mineralogy of the mg cm? yr" at the poles can be calculated (BRAT-
sediments. ERMANet al., 1983). These are somewhat larger
than estimates of iron deposition in the Hamersley
BIFs of 9-43 mg cm? yr-1 (TRENDALLand
Ultraviolet irradiation
BLOCKLEY,1970), indicating that photo-oxidation
Another method of oxidizing iron in solution is of iron is a sufficient explanation for their forma-
by means of direct UV irradiation (WEISS, 1935; tion. The Precambrian ocean model of FRAN<;:OIS
JORTNERand STEIN,1962a and b;PAPPand VINCZE, (1986, 1987), which also calculates deposition of
1980). iron formations based on the photo-oxidation of
CAIRNS-SMITH(1978) originally proposed that [Fe(OH)]+ , agrees with these results.
BIFs were formed by the photochemical oxidation The work of ANBARand HOLLAND(1992) on
of iron, according to the reaction the photochemistry of manganese is also consistent
with a photochemical model for the formation of
BIFs. However, DESMARAIS(1994) points out that
the absence of BIF deposition on shallow-water
Archean platforms, even at water depths some dis-
Further studies showed that photo-oxidation of iron
tance below the photic zone, indicates that not only
is an important reaction in ferrous solutions with
was the surface seawater oxidizing, but that a mo-
a wide range of pH values, 7.5 and less (BRATER.
MANet al., 1983 and 1984). At pH values less bile oxidant such as dissolved O2 must have pene-
trated below the photic zone to prevent the deposi-
than about 3.5, hydrolysis and precipitation are not
tion of BIFs at those depths.
important, but in the pH range 6.2-4, goethite is
precipitated according to the following reactions:
Hydrogen peroxide
Fe3+(aq) "" [Fe(OH)f+(aq) + H+(aq) (6)
KASTINGet at. (1984, 1985) suggested that hy-
[Fe(0H)]2+(aq) "" -1[Fez(0H)4f+(aq) drogen peroxide (H202) could be formed in the
2 early, carbon dioxide-rich atmosphere of Earth and
then rain out to oxidize and precipitate the iron in
+ W(aq) (7)
the ocean. They determined that hydrogen peroxide
is probably the dominant oxidant in rainwater at
3
free oxygen levels less than 10-2_10- PAL (about
2 X 10-3 to 2 X 10-4 bars), and was probably
Fe(0H)3(aq) -+ FeO(OH)(s) i. (9) an important oxidant even in Oz-free atmospheres,
Iron-formations onMars?
385
provided that the CO2 levels were significantly f02level between that of the hematite + magnetite
higher than today. These results indicate that hy- and magnetite + quartz + fayalite buffers. In their
drogen peroxide is also a viable candidate for an model, evaporative precipitation of carbonates oc-
inorganic oxidizer in the Precambrian ocean, that curs at the margins of a largely spring-fed lake,
could have acted to precipitate the BIFs according or barred lagoon, with magadiite (NaSi0 (OH)3
7 13
to the reactions
X4H20) and iron-silica gel precipitation occurring
inthe deeper parts of the lagoon. Occasional floods,
2Fe2+(aq) + H20z(aq) + 4H20 -> 2Fe(0H)3 ~ which bring in fresh, low-pH waters, are invoked
+4W(aq) (12) to precipitate the magadiite or sodium silica gel,
and ferrous hydroxide (Fe(OH)2), iron-silicate and
Fe(0H)3(aq) -> FeO(OH)(s) L (13) sodium-iron-silicate gels and to wash the carbon-
ates into the lake, producing bands. In addition,
This oxidant, like dissolved oxygen gas, would not
ferric hydroxide (Fe(OH)3) precipitates if free oxy-
have had a problem penetrating below the photic
gen should become available.
zone to prevent the deposition of BIFs at such
This model represents not a mechanism for oxi-
depths.
dizing the iron, but a means of depositing it in
However, subsequent work by KASTING(1990)
the ferrous state. The presently-observed chert and
indicates that when sulfur chemistry is included in
ferric iron deposits are caused by the subsequent
a Iow-O; atmosphere, the dominant photochemi-
diagenesis of the primary precipitates mentioned
cally-produced oxidant is sulfuric acid, and the hy-
above (EUGSTERand CHOU, 1973). This scenario
drogen peroxide abundances are greatly reduced. In
is significantly different than the other models dis-
this case, the sulfuric acid would react with water in
cussed here. The other models assume that the
the atmosphere to form sulfate, which would then
rain out into the ocean (KASTING,private communi- depths of the ocean were low in oxygen, but that
cation). Sulfate is not nearly aspowerful an oxidant ferric iron was oxidized and precipitated in an up-
as hydrogen peroxide, so this mechanism may not per layer, either by dissolved oxygen or by some
be sufficient to oxidize the ferrous iron to form other means. The arguments for oxidization of the
iron formations. iron in solution are strong, and seem to preclude
In summary, there appear to be an abundance of the likelihood of the EUGSTERand CHOU (1973)
adequate abiotic mechanisms toprecipitate out iron mechanism [developed over ten years earlier, be-
in a largely anoxic Precambrian ocean, under a fore the work of Braterman, Cairns-Smith, Canuto,
carbon dioxide-rich atmosphere. At our current Kasting, Towe, and their co-workers (BRATERMAN
level of knowledge, it is difficult to determine un- et al., 1983; BRATERMANet al., 1984; BRATERMAN
equivocally which must have been the dominant and CAIRNS-SMITH, 1987; CAIRNS-SMITH, 1978;
process. Although certainly the development of Iiv-. CANUTOet al., 1983; KASTINGet al., 1984; KAST-
ing organisms radically changed the oxidation state INGet al., 1985)] having been dominant. In addi-
of the early terrestrial environment, however, it is tion, there is evidence from rare earth element
not necessary to invoke the origin of photosynthe- (REE) analysis that Precambrian BIFs are the result
sis to begin the deposition 'of the banded iron of chemical precipitation from solutions that are
formations. mixtures of seawater and hydrothermal input
(KLEINand BEUKES,1992). However, the thermo-
dynamic calculations and predominance diagrams
Anoxic deposition and the banded nature of the
iron-formations that they developed are still useful in determining
what species are stable under given conditions of
In addition, the banded nature of the iron forma- fo2,pH, and fco2,particularly in the anoxic condi-
tions requires explanation, and the most common tions prevalent deep in the ocean.
idea is that the iron supply is periodically variable
A more recent model for explaining the banded
(BUTTON et al., 1982; DREVER, 1974; EUGSTER
nature of the BIFs was developed by MORRIS
and CHOU, 1973; HOLLAND,1973; MORRIS, 1993;
(1993). This detailed model of the Hamersley BIF
TOWE, 1983).
in Australia relies on the interaction of two major
EUGSTERand CHOU(1973) proposed that banded
oceanic supply systems to produce bands of iron
iron-formations were deposited in a playa-lake
and silica: (1) surface currents and (2) convective
complex. They assumed that the atmosphere was
upwelling from mid-ocean ridges or hot spots
free of oxygen, and that precipitation of iron oc-
(MORRIS, 1993). The surface currents were satu-
curred in response to changes in pH and fco2at an rated in silica, but relatively free of iron, either
M. W. Schaefer
386
because the system was oxidizing or because of values of surface albedo, latitude, and orbital posi-
photoprecipitation. Precipitation from these surface tion, pressures of from 0.75 to 5bars are necessary.
waters, possibly driven by evaporative concentra- Work by KASTING,however (KASTING,1991), indi-
tion, formed the chert-rich layers. Periodically, up- cates that CO2 alone may not be sufficient to warm
the surface of Mars above the freezing point of
welling from iron-rich deeper waters, driven by
water. He finds that under Martian conditions car-
convective overturn, overwhelmed the deposition
bon dioxide cannot be present in the atmosphere
of silica to produce the iron-rich layers. This model
at such high pressures, but instead will condense
includes a low-oxygen to anoxic atmosphere,
out. There is recent work from FANALEand POST-
higher heat flow under the oceans than at present,
AWKO(1995) and from YUNGet al. (1995) indicat-
and a modest level of photosynthetic activity, but
ing that adding SOz to the models may eliminate
none of these conditions are essential to the model
the problem of CO condensation. Even if acarbon
(MORRIS, 1993). Another factor considered in this 2
dioxide atmosphere alone was not enough to cause
model of the formation of the Hamersley deposits
awarm early Mars, there is still sufficient evidence
was the effects of surface volcanic activity.
that Mars outgassed large amounts of carbon diox-
The bulk of both ferrous and ferric iron presently
ide (POLLACKet al., 1987), and there is much geo-
in the Hamersley BIF is therefore attributed to oxi-
logical evidence of flowing water upon the surface,
dation in the surface zone of the ocean. The
implying warm temperatures. In addition, PARKER
FeO(OH) precipitate eventually is transformed to
and coworkers (PARKERet al., 1987; PARKERand
(1) hematite, if oxidizing conditions are maintained
GORSLINE,1992; PARKERet al., 1993) have seen
in the sediments, or (2) magnetite, if partial reduc-
geomorphic evidence for ancient shorelines in the
tion occurred due to the action of organic matter,
northern plains of Mars.
or (3) siderite/silicate, if substantial reduction oc-
The existence of lakes or oceans on early Mars
curred in the photic zone (increasing its iron con-
will therefore be assumed as aworking hypothesis
tent) or in the sediment itself (MORRIS, 1993).
for this paper, but I will show that if BIFs do exist
on Mars, their existence offers proof of the exis-
EARLY GEOCHEMISTRY OF EARTH tence of large bodies of water on Mars at some
AND MARS
time.
Mars is farther from the Sun, which means that
When Mars and Earth were intheir infancy, they
it receives less radiation than Earth, and would
were far more alike than they are at present. Both
have been cooler. More importantly, it is smaller
were rocky bodies, heated by their accretion, and
than Earth, so that its internal heat was not enough
with atmospheres dominated by carbon dioxide,
to support plate tectonics, or to keep the crust from
nitrogen, and water vapor. Volcanic activity pro-
thickening and cutting off volcanic activity to the
duced vast basalt plains, as well as volcanic edi-
surface. As it lost volatiles from the atmosphere,
fices. The bombardment of the crust with meteor-
both to the surface and to space, the surface cooled
ites also kept hydrothermal circulation systems
past the freezing point of water, and its geochemi-
active. In topographically low areas, lakes and
cal evolution slowed and nearly ceased.
oceans formed. Active weathering of the silicate
But before about 3500 million years ago, the two
crust by carbonic acid-rich waters provided a rich
planets were relatively similar, and geochemical
source of iron, silica, and other cations to supply
processes, such as the formation of BIFs, that we
the early oceans, and left behind itaclay and oxide-
observe on Earth may have occurred on Mars as
rich weathering profile, which might still be seen
well. The second part of this paper investigates
today on Mars (TOWE, 1983).
how one must modify the mechanisms discussed
The existence of lakes or oceans on early Mars
above to apply toMartian conditions, and discusses
is'still controversial. Values of total outgassed CO2 what implications the existence of BIFs on Mars
from several to about 10 bars are consistent with
would have for the history of the planet (such as
present knowledge (POLLACKet al., 1987), and this confirmation ofthe warm, wet, early Mars hypothe-
amount of COz implies an amount of water out- sis), and how we might determine their existence
gassed at least equal to an equivalent depth of 500
there today.
to 1000 meters (CARR, 1986). POLLACK et at.
MECHANISMS OF MAKING IRON-
(1987) have made extensive calculations deriving
FORMATIONS ON MARS
the amount of carbon dioxide necessary to achieve
Dissolved oxygen
a strong enough greenhouse effect to raise the sur-
face temperature of Mars above the freezing point At present, the Martian atmosphere contains, on
of water, and have determined that, for different average, 0.13% O2, in an atmosphere of about 7.5
Iron-formations on Mars? 387
X 10-3bars CO2, or about 10-5bars of O2 (OWEN, on these two terrestrial models. This is more than
1992). This oxygen is produced by the photolysis enough oxygen to oxidize ferrous iron in solution
of water and CO2, as it was described above for so that it will precipitate out as ferric iron. [Note:
the prebiotic Earth. Ifthere could be liquid water There does appear to be a small stability field for
on the surface of Mars today, the upper layers, at siderite at f02equal to 10-13and below, and pH "",6
least, would be oxygenated, and ferrous iron in (see Fig. 3, discussed in more detail later).] The
solution could not exist. In the past, when there problem with forming BIFs under these conditions,
was more carbon dioxide and water vapor in the conversely, it that there must have been some
atmosphere of Mars, there might have been corre- mechanism for maintaining the ferrous iron in solu-
spondingly more free oxygen as well. However, tion so that it could be transported. Possibly, any
the oxygen produced in the atmosphere of Mars bodies of water on early Mars could have been
today is in contact with a highly-oxidized surface, deep enough and still enough (Mars, after all, can
whereas in the distant past it would not have been, have had no tides) so that conditions near the bot-
and the oxygen in the atmosphere would have been tom could have been anoxic.
decreased by its reactions with the surface materi-
als and the cations in solution in water. The amount Ultraviolet irradiation
of oxygen in the atmosphere of Mars today is
thought to be controlled by redox balance, such Because Mars is only about 1.5times as far from
that the H escape rate is equal to twice the 0escape the Sun as is the Earth, the flux of UV (and other
rate (McELROY and DONAHUE,1972) plus surface solar) radiation at its surface is only about afactor
oxidation. Active volcanism in the past would have of 2.3 times less than that at the surface of the
released H2 and other reduced gases, which also Earth. The yield of a photochemical process is di-
would have lowered the P02 relative to present
levels.
If an early Earth-type atmospheric model, such
as that of CANUTOet al. (1983) or KASTINGet (It.
(1984), is used to calculate the amount of oxygen
in the atmosphere of Mars, it produces a much Fe(OH)3
lower mixing ratio of oxygen than ispresently seen.
The CANUTOetal. (1983) model produces asurface
oxygen mixing ratio for a solar UV flux of 1.7
times the present solar flux at Earth (appropriate
for about 3.5 Gy ago, when the solar UV flux at
Earth would have been about 4 times the present
flux) and an atmosphere of 100 terrestrial PAL (2.8
X 10-2bars) of approximately 10-13 (Fig. 2). If §
this model is further extrapolated to 10,000 terres- eo
.9
trial PAL, or 2.8 bars (a value that may have been
necessary toproduce enough of agreenhouse effect
on Mars toraise the temperature above the freezing
point of water), a mixing ratio of about 10-10 may
be produced.
The KASTINGet al. (1984) model, extrapolated
in asimilar fashion, also produces asurface oxygen
mixing ratio of about 10-10to 10-11• It is not really
avalid approximation to extrapolate either of these
two models in this fashion, because there is a good
deal of Earth-specific atmospheric chemistry in
pH
each of them, but it is useful to give us a vague
approximation of the surface oxygen mixing ratio FIG. 3. Solubilities of Fe(OH)3, Fe(OH)2, greenalite
(dot-dashes, in equilibrium with amorphous silica), and
that might have been expected on early Mars.
siderite (dashes, with a log Pco, of 0 and -3.5), at 25°C
Therefore, let us assume that the surface oxygen
and alog total iron content of -4.The magnetite-hematite
mixing ratio in the warm, wet period of early Mars
(MH) and Iayalite-magnetite-quartz (FMQ) boundaries
may have been approximately 10-10to 10-13, based are also shown. Adapted from EUGSTER and CHOU (1973).
388 M. W. Schaefer
rectly proportional to the number of photons ab- curve at pH = 8.6 (EUGSTERand CHOU, 1973).
sorbed, so the efficiency of iron photoprecipitation This implies that magadiite can be precipitated by
on Mars at a given time should be about 2.3 times either evaporative concentration (raising the pH),
less than at the same time on Earth, if scattering or by flooding of silica-rich, high-pH brines by
and loss of UV flux through the atmosphere is fresher, more acid waters. This mechanism is not
disregarded. The range of iron photoprecipitation dependent on any factors that are different on Mars
rates calculated by BRATERMANet at. (1983) for than on Earth, and would therefore work equally
Earth may be thus scaled to Mars, to get 65-320 well on Mars as it does on Earth.
mg cm-2 yr-1 at the tropics or 22-110 mg cm" The deposition of the iron-rich layers, in this
yr-1 at the poles. These rates are still larger than anoxic model, is more complicated. The stabilities
the rate estimates for the terrestrial Hamersley BIF of various iron oxides, hydroxides, and iron car-
deposition (9-43 mg cm" yr-1 (TRENDALLand bonate relative to one another must be determined,
BLaCKLEY, 1979)), indicating that photoprecipita- to derive which species are stable under likely Mar-
tion of iron is a sufficient mechanism to produce tian conditions. The solubility of iron hydroxides
BIFs on Mars as well. and hydroxides depends on the fo, and the pH, and
that of siderite (FeC03) depends as well on Pco,.
Hydrogen peroxide Figure 4 shows a predominance diagram (from
EUGSTERand CHOU,1973) for iron species in solu-
The same photochemical reactions that could tion at 2Ye. At low fo, the diagram is dominated
have formed hydrogen peroxide in the early atmo- by ferrous species, with Fe2+ being the most im-
sphere of Earth could have acted in the early atmo- portant at pH below 8. The pH of an early Martian
sphere of Mars. As a first approximation, one may ocean islikely tohave varied between 5.4 (initially)
look at the terrestrial model of KASTINGet at. and 10(as the carbon dioxide in the atmosphere
(1985), which determined that hydrogen peroxide is reduced), over a time period of about 0.5 Gy,
is probably the dominant oxidant in rainwater at according to the model of SCHAEFER(1993). There-
free oxygen levels less than 10-2_10-3 PAL (about fore, at least at earlier times, ferrous iron should
2 X 10-3to 2 X 10-4 bars). This model was also
run with a mixing ratio of CO of 100 PAL, and
2
the H0 rain-out rate was found to decrease by
2 2
only 5-30%. This implies that the crossover point
between free oxygen or hydrogen peroxide as the
major oxidant was probably the same, within an 0
order of magnitude or so, on Mars as it was on
Earth. The calculations for Mars above indicate
that there may have been between 10-11 and 10-3
O2 surface mixing ratio, or between 10-11 and 10-3 I
bars O2, assuming a total atmospheric pressure of Fe(OH)4-
1bar. This range is from far less than the amount I "\
required to make free oxygen the dominant oxi-
N
dant, toroughly the borderline between oxygen and ..9 4
00 - 0
hydrogen peroxide at the dominant oxidant. Based ...9
on this, hydrogen peroxide also was capable of I
precipitating BIFs on Mars. Fe2+
I
-60
Anoxic deposition
The anoxic model of terrestrial BIF formation
developed by EUGSTERand CHOU(1973) attempts
to explain not only iron-rich bands, but also the
silica-rich chert bands. Their explanation of the
formation of the chert bands is that these were
2 6 8 10 1
originally deposited in the form of magadiite
pH
(NaSi0 (OH)3 X 4H0). Magadiite, which is ob-
7 13 2
served as the precursor to chert beds in Lake Ma- FIG. 4. Predominance diagram (adapted from EUGSTER
gadi, in Africa, has a minimum in its solubility and CHOU (1973) for iron species in solution at 2Ye.
Iron-formations on Mars? 389
in Fig. 6. Under conditions of high Pco, siderite
will precipitate out at amuch lower pH than calcite
\
\ (CaC03). Siderite begins to precipitate at an iron
\
\ concentration of only 0.5-0.05 ppm (log aFe~oo~f
\
\\ 10-5_10-6) at low pH, and even lower at high pH.
\\ __ F.."~°!'l22__ Figure 3isalog f02-pHplot showing the compar-
ative solubilities of Fe(OH)3, Fe(OH)2, greenalite
(Fe3Si205(OH)4), and siderite, at 2SOCand a log
total iron content of -4, corresponding to about 5
ppm iron (after EUGSTERand CHOU(1973)). Under
present terrestrial Peo, conditions, greenalite will
precipitate preferentially to siderite at conditions
oflow f02and moderate pH. However, in the higher
pH regime thought to exist on early Mars, siderite
will precipitate preferentially. If the log fo, is
higher, above about -15 at apH of6,then Fe(OH)3
is the preferential phase of iron in this system.
In summary, while the terrestrial anoxic model
of EUGSTERand CHOU(1973) produces iron-rich
layers that are dominated by calcite + greenalite,
with some siderite present, the Martian version of
this model strongly favors the precipitation of sid-
-40 -80 erite. Both models produce chert bands, originally
o logf02 deposited as magadiite.
FIG.5.The solubility ofmagnetite, hematite, siderite, DISCUSSION
and some iron hydroxides at a pH of 7 and 2SOC,
adapted from EUGSTERand CHOU(1973). The solubili- The application of abiotic models of producing
ties ofFe(OH)3, magnetite, and hematite are shown as terrestrial BIFs to the conditions likely on early
solid lines, Fe(OH)2 as dashes, and siderite (at log
fco,of-3.5 and0)asdots. Theboundary between Fe2+
and Fe(OH)i is shown as dash-dot.
o
have been highly mobile in relatively anoxic waters
(log f02 < ~ -20) on Mars. As was mentioned
above, the atmosphere of Mars is now, and may -2
always have been, oxidizing, but still waters may
have been anoxic at depth. Also, groundwater that ~('j -4
is not in direct contact with the atmosphere (i.e., it ~
bJ)
has a long residence time in the aquifer) can be .9 -6
extremely anoxic, particularly if it is buffered by
+E
a redox buffer (such as Fe2+-Fe(OH)3 or Fe2+ N& -8
-Fe203) (DREVER,1988). '"
eo
Figure 5 (after EUGSTERand CHOU,1973) shows .9
the solubility of magnetite, hematite, siderite, and -1
some iron hydroxides at a pH of 7 and 2SOC.The
chief feature of interest here is the change in solu- -121
bility of siderite with change in Peo" it being far
more stable under conditions of high Pco" for any
2 4 8 10 12
value of oxygen fugacity. At log fo, values of less pH
than about -50, and a Peo2 of 1 bar, siderite is
FIG.6. Comparison of the solubilities of calcite and
preferentially precipitated over Fe(OH)3'
siderite, atlowfo,and25°C,adapted fromEUGSTERand
The solubility of siderite relative to calcite for
CHOU(1973). Siderite curves are drawn for log fco,of
low fo, (after EUGSTERand CHOU(1973)) is seen -3.5 and O.
390 M.W. Schaefer
Mars indicates that these deposits could have As a result, in order to determine the existence
formed there, as they did on Earth. How can we of BIFs on Mars, we must see below the surface.
determine if BIFs do indeed exist on Mars? And A lander capable of penetrating past the uppermost
if they do, what implications does the existence of weathered layer of the surface could, perhaps even
BIFs on Mars have for the evolution of the Martian by visual inspection alone, determine the existence
atmosphere and hydrosphere? of BIFs on Mars. Terrestrial BIFs form deposits of
iron oxides and silica, with other associated miner-
Observations
als, which are quite distinctive in hand specimens.
The three mechanisms of BIF formation involv- Adding to this simple visual inspection any spec-
ing free oxygen, photoprecipitation, or hydrogen toscopic measurements would enable easier identi-
peroxide all produce deposits of FeO(OH). Under fication, especially if the proposed Martian iron-
different formation conditions, FeO(OH) will pre- formations do not have the same banded structure
cipitate in different crystal phases: a-FeO(OH) as the terrestrial ones. On Mars, one would wish
(goethite), ,B-FeO(OH) (akaganeite),. y-FeO(OH) to examine shallowly-buried impact debris in the
(lepidocrocite), o-FeO(OH), or 0'-FeO(OH) (fer- northern plains or in large basins, such as Hellas
oxyhyte) (BURNS, 1993). Akaganeite may sub- or Argyre. It is in these localities that one might
sequently transform to either goethite or hematite expect to find evidence of BIFs on Mars, should
(a-Fe203); lepidocrocite may revert to maghemite they exist. By examining impact debris we increase
(y-Fe203); feroxyhyte may transform to goethite, greatly the depth of sampling into the solid bedrock
hematite, or mixtures of the two (BURNS, 1993). of the planet over what any feasible lander might
Many of these phases have been suggested as ex- be able to sample on its own. And by looking at
isting on the surface of Mars (e.g., BURNS, 1980; shallowly-buried debris, such as might still be ac-
HARGRAVESet al., 1979; MORRIS et al., 1989; cessible to a lander, we allow the possiblilty of
MORRISand LAUER, 1990; SINGER,1982), but in finding some material that has not entirely been
the present Martian environment, hematite may be altered to hematite.
the only stable iron oxide (GOODING,1978). The Mars Pathfinder, to be launched in 1996,
Under Martian conditions, the anoxic mecha- is not likely to find evidence for iron-formations,
nism of BIF formation produces deposits of sider- because itwill not be looking beneath the surface.
ite. Siderite is unstable under present Martian The Mars Volatiles and Climate Surveyor, pro-
conditions, and is converted via gas-solid decom- posed for launch in 1998,has the necessary instu-
position into hematite (GOODING,1978).Nanocrys- mentation to search for iron-formations, but will
talline hematite is a strong candidate to exist on be visiting the south polar cap region, which is not
the Martian surface (MORRISet al., 1989;MORRIS a likely place to find iron-formations on Mars. We
and LAUER,1990),and a nanophase product could will have to wait for one of the later Mars Surveyor
be expected as aresult of gas-solid decomposition missions to search for iron-formations on Mars.
of siderite.
BURNS(1993) suggested that the bright deposits
Implications
littering the Argyre and Hellas basins could be com-
posed of dust derived from dessicated hydrous iron If iron-formations are found on Mars, this will
oxide-silica deposits that remained there after the be confirmation of those models of the early envi-
water had evaporated from these deep depressions. ronment of Mars that include the existence of a
There is evidence, therefore, that one source for warm, wet atmosphere and bodies of liquid water.
the iron oxides that are observed on Mars may Their existence implies the existence of liquid wa-
be iron-formations. However, this is not the only ter basins, probably of significant depth, and of a
possible source, because hematite and other iron active weathering profile on the land, supported by
oxides could have formed as weathering rinds on a groundwater system and possibly rain as well. If
the rocks of the surface, without having been trans- it is possible to determine which of the different
ported in solution, or could have formed by hydro- formation mechanisms produced the BIFs, at least
thermal processes, in the presence of groundwater between the oxidizing mechanisms (producing
alone. FeO(OH)) and the anoxic mechanism (producing
Spectroscopic data alone, whether ground-based siderite), then we can place limits on the oxygen
or from aMars-orbiting spacecraft, are not enough fugacity of the early Martian atmosphere, or on the
to determine the existence of BIFs on Mars, be- necessary depth of ocean basins.
cause the only surface expression of BIFs we ex- In all the abiotic mechanisms of BIF formation
pect is hematite. discussed above, gaseous hydrogen is produced.