JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 98, NO. Ell,
PAGES 20,831-20,853, NOVEMBER
25, 1993
The Nanophase Iron Mineral(s) in Mars Soil
A. BANIN, 1 T. BEN-SHLOMO, AND L. MARGULIES
Department of Soil and Water Sciences, The Hebrew University, Rehovot, Israel
D. F. BLAKE
AND R. L. MANCINELLI
Space Science Division, NASA Ames Research Center, Moffett Field, California
A. U.
GEHRING
Department of Soil Science, University of California, Berkeley
A seriesof surface-modifiedclays containingnanophase(np) iron oxide/oxyhydroxidesof extremely
small particle sizes, with total iron contents as high as found in Mars soil, were prepared by iron
deposition on the clay surface from ferrous chloride solution. Comprehensive studies of the iron
mineralogy in these "Mars-soil analogs" were conducted using chemical extractions, solubility
analyses, pH and redox, x ray and electron diffractometry, electron microscopic imaging, specific
surface area and particle size determinations, differential thermal analyses, magnetic properties
characterization,spectralreflectance,and Viking biology simulationexperiments. The clay matrix and
the procedureused for synthesisproducednanophaseiron oxides containinga certain proportion of
divalent iron, which slowly converts to more stable, fully oxidized iron minerals. The clay acted as an
effective matrix, both chemically and sterically, preventing the major part of the synthesized iron
oxidesfrom ripening,i.e., growingand developinglarger crystals.The precipitatediron oxides appear
as isodiametric or slightly elongatedparticles in the size range 1-10 nm, having large specific surface
area. The noncrystalline nature of the iron compoundsprecipitated on the surface of the clay was
verified by their complete extractability in oxalate. Lepidocrocite (3,-FeOOH) was detected by
selected area electron diffraction. It is formed from a double iron Fe(II)/Fe(III) hydroxy mineral such
as "green rust," or ferrosic hydroxide. Magnetic measurementssuggestedthat lepidocrocite converted to the more stablemaghemite(3,-Fe203) by mild heat treatment and then to nanophasehematite
(a-Fe203) by extensive heat treatment. After mild heating, the iron-enrichedclay became slightly
magnetic, to the extent that it adheresto a hand-held magnet, as was observed with Mars soil. The
chemical reactivity of the iron-enriched clays strongly resembles, and offers a plausible mechanism
for, the somewhatpuzzling observationsof the Viking biology experiments. Their unique chemical
reactivities are attributed to the combined catalytic effects of the iron oxide/oxyhydroxide and silicate
phasesurfaces.The reflectancespectrumof the clay-iron preparationsin the visible range is generally
similar to the reflectance curves of bright regions on Mars. This strengthens the evidence for the
predominanceof nanophaseiron oxides/oxyhydroxidesin Mars soil. The mode of formation of these
nanophaseiron oxideson Mars is still unknown.It is puzzlingthat despitethe longperiod of time since
aqueous weathering took place on Mars, they have not developed from their transitory stage to
well-crystallized end-members.The possibility is suggestedthat these phasesrepresent a continuously
on-going, extremely slow weathering process.
[1992] and Roush et al. [1993]). Primary sources of information regarding the mineralogy of Mars soil include weatherThe surface of Mars is covered with a thin, loosely ing models based on the elemental composition of the soil
packed, fine-particle, weathered soil material which has [Toulmin et al., 1977; Gooding, 1978; Gooding and Kiel,
been studied rather intensively over the last 15 years (see 1978; Zolotov et al., 1983] and direct mineralogical analyses
recent reviews by Arvidson et al. [1989a] and Banin et al.
of the weathered componentsof the SNC meteorites [Good[1992]). The elemental composition of the soil was detering et al., 1991]. Data from Viking Lander experiments,
mined by the XRF analyzers on board the two Viking which measured the magnetic [Hatgraves et al., 1977] and
Landers [Toulmin et al., 1977; Clark et al., 1982], and further
mechanical [Moore et al., 1977] properties and searched for
constrainedby the detailed analysesof the SNC meteorites,
the presenceof life on Mars [Klein, 1979] were also used for
believed to be rocks ejected from Mars [Laul et al., 1986].
inferences on the mineralogical makeup of the soil either
Unlike the chemical composition, the mineralogical constitdirectly or through simulations with terrestrial minerals
uents of the soil have not been analyzed directly yet, and
[Oyama et al., 1978; Banin and Rishpon, 1978, 1979; Humuch controversy surrounds the issue, with numerous canguenin, 1982; Posey-Dowty et al., 1986; Plumb et al., 1989;
didates proposed (e.g., see recent reviews by Banin et al.
Banin et al., 1985, 1988a; J. S. Hubbard, personal communication, 1979]. Additional constraints on the mineralogical
•Alsoat NASA AmesResearch
Center,MoffettField,California.
compositionof Mars soil were achieved from spectralreflecCopyright 1993 by the American Geophysical Union.
tance measurements from Earth, from orbiting and fly-by
spacecrafts,and from the Viking Landers. Especially inforPaper number 93JE02500.
0148-0227/93/93JE-02500505.00
mative are recent telescopic observations, using improved
INTRODUCTION
20,831
20,832
BANIN ET AL.: NANOPHASEIRON MINERALS IN MA•S SOIL
detection systems during the 1988-1990 oppositions [Bell et
al., 1990; Roush et al., 1993; Singer et al., 1990]. Yet the
mineralogical nature of the soil is far from being understood,
nor are its time of formation and weathering processes
known.
Quantitatively, the two major mineral-forming elements in
Mars soil are silicon and iron, constituting 44% and 19% of
the soil as SiO2 and Fe203, respectively. The iron-free
silicate phases have been studied briefly, primarily because
their spectral fingerprinting in the visible (VIS) and nearinfrared (NIR) spectral ranges are limited. Much more
experiments on Mars [Banin and Rishpon, 1979; Banin and
Margulies, 1983; Banin et al., 1988a] and to have spectral
reflectance in the VIS-NIR strongly resembling the bright
regionson Mars [Banin et al., 1985]. The analogy with Mars
soil is based, in a number of aspects, on the nature and
behavior of the iron oxides and oxyhydroxides deposited on
the surface of the clay particles. We wish to summarize our
knowledgeof the properties of these iron phases,present
some recent findings on them, and discuss their mode of
formation and potential relevance to Mars surface processes.
attention was given to the iron minerals in the soil, due to
their pronounced absorption in the VIS and NIR, making
them easily detectable by telescopic observations [Singer,
MATERIALS
AND METHODS
1985; Roush et al., 1993].
Iron, the second most abundant element in Mars soil,
Nanophase Iron Oxide Preparation
appears to be primarily present in its oxidized form (ferric),
strongly suggesting chemical weathering of the basaltic
parent material.
The typical reflectance spectrum of Mars in the visible/
short-wave NIR range bears the strong fingerprinting of
oxidized iron [Singer et al., 1979; McCord et al., 1982a;
Bell et al., 1990]. The spectrum has been reproduced more or
less faithfully in laboratory measurementsof many different
iron-containing phases. These include "amorphous" iron
oxides [Evans and Adams, 1979, 1980], palagonite [Singer,
1982, 1985; Adams et al., 1986; Guinness et al., 1987; Morris
et al., 1990; Bell et al., 1993; Golden et al., 1993], corundum
doped with iron [Morris et al., 1983], jarosite [Burns, 1986],
nanophase hematite deposited in silica matrix [Morris, 1988;
Morris et al., 1989; Morris and Lauer, 1990], and ironenriched smectites [Banin et al., 1985, 1988a]. One conclusion from the abundance of spectral analogs is that
unambiguous identification of the iron mineral(s) in Mars
soil, solely on the basis of reflectance, is not possible at
present. A more important conclusion is that the accumulat-
A smectite clay (SWy-1 montmorillonite) from Crook
County, Wyoming, obtained from the Clay Minerals Society
Repository, was used for the preparation of the series of
iron-enrichedclays. The parent clay was suppliedin powder
form and was converted to the desired ionic compositionby
the quantitative ion exchange method [Banin, 1973; Gerstl
and Banin, 1980]. This method entails titrating a mixture of
freshly prepared hydrogen-saturated montmorillonite and
OH-saturated ion exchange resin in suspension, with the
chloride salt(s) of the desired exchangeable cation. Nitrogen
gas was bubbled in the suspension during the titration to
displace oxygen and prevent or slow down the direct oxida-
tion of Fe2+ in solution.The procedurewas conducted
starting with a large batch (--•200 g) of hydrogen clay. The
clay was titrated at constant rate with a 1 N FeC12 solution
in the presenceof excess strongly basic OH-anion exchange
resin (Amberlite IR-410). The p H and electrical conductivity
were monitored continuously during the titration. Aliquots
of the clay suspensionwere withdrawn from the batch after
predetermined amounts of iron were added. Eh was measured using a Pt/calomel electrode pair; pe values were
calculated by p e = Eh(mV)/59. The clay-iron preparations
were then sedimentedby centrifugation (10,000 rpm for 30
min), freeze dried, and lightly crushed to pass a mesh of
ing observationaland simulation spectral evidence strongly
suggeststhat iron in the weathered component of Mars soil
is mostly present in poorly crystallized clusters of oxyhydroxy ferric iron, or as crystalline minerals having extremely small particle size ("nanophases" (np), or "nanoc- 0.105 mm. The powders were stored under room air for
rystals"). Telescopic observationshave detected the typical further analyses. Aliquots of the suspensionswere kept as
such or in a concentrated paste form. These wet-clay syshematite feature at 860 nm [Morris et al., 1989; Morris and
Lauer, 1990]. However, the feature is very weak, and recent tems were stored at a temperature of 5øC for further analyestimates of the content of hematite and/or other crystalline
iron oxide phasesare in the range of 1-2% or less [Bell et al.,
1990; R. B. Singer, personal communication, 1991]. The
content of the magnetic mineral in the soil is also estimated
to be in the range of 1-7% [Hargraves et al., 1977]. It may be
maghemite, magnetite, or nanophase hematite that is typically superparamagnetic. Still, the bulk of the iron oxide/
oxyhydroxide in the soil appears to be amorphous or shortrange ordered and is characterized by extremely small
particle sizes.
The two most thoroughly studied mineral analogs to the
iron nanophasesin Mars soil are np hematite studied by
Morris et al. [Morris, 1988; Morris et al., 1989, 1990; Morris
and Lauer, 1990] and iron-enriched clays studiedby Banin et
al. [Banin and Rishpon, 1978, 1979; Banin, 1986; Banin et
al., 1985, 1988a].
The iron-enriched clays have been suggestedas important
mineral components of Mars soil. They were shown to
simulate many of the findingsof the Viking Labeled Release
ses.
Total Iron
Content
The total content of iron in the iron-enriched clays was
determined using XRF spectroscopy (XRAL Laboratories,
Ontario, Canada). The iron concentration was calculated to
three different weight bases: the "air dry" basis, the "oven
dry (105øC)" basis, and the "heat stable matrix" basis, i.e.,
after subtracting the weight loss on ignition to 950øC. Results
from all three calculation modes are reported in Table 1.
Surface Ion Composition
The 1% suspensionsprepared from the various clays in the
series were sequentially extracted with 1 N KC1 and 0.5 N
HC1 following Gerstl and Banin [1980]. The ions extracted
by 1 N KC1 were termed the "exchangeable fraction," and
those extracted by 0.5 N HC1 were termed the "acidextractable fraction." Iron, magnesium, and aluminum concentrations
in the combined
extraction
solutions
were deter-
BANIN ET AL.; NANOPHASEIRON MINERALS IN MARS SOIL
TABLE
20,833
1. Representative Series of np Iron Oxide Prepared in Clay Montmorillonite Matrix:
Total
Iron
Concentration
and Fe Oxide
Minerals
Detected
Fe20 3 Content, wt %
Iron
Sample
E6(II)-I
E6(II)-2
E6(II)-3
E6(II)-4
E6(II)-5
E6(II)-6
E6(II)-7
E6(II)-8
E6(II)-9
E6(II)-10
E6(II)-I 1
E6(II)-12
E6(II)-13
E6(II)-14
E6(II)-15
E6(II)-16
E6(III)-I
E6(III)-2
E6(III)-3
E6(III)-4
E6(II1)-5
E6(III)-6
E6(III)-7
E6(III)-8
E6(III)-9
E6(III)-10
E6(III)-I 1
E6(III)-13
E6(III)-15
E7(II)-100%
E7(II)-200%
E7(II)-400%
E7(II)-600%
Oxide
Air Dry
Basis
105øCDry
Basis
Heat Stable
Matrix
Minerals
Detected
2.78
3.38
4.29
3.62
5.78
6.24
6.90
7.23
7.33
7.85
8.86
9.91
9.31
9.64
9.68
9.45
3.65
4.65
5.26
5.66
6.59
7.16
7.44
8.38
8.65
9.21
10.30
12.60
15.40
6.59
9.81
13.50
18.70
ß ßß
ßßß
ß ßß
ß ßß
ß ßß
ßßß
...
-..
...
ßßß
ßßß
ßßß
ßßß
ßßß
..ßßß
4.01
5.19
5.39
6.46
7.39
7.97
8.27
9.52
9.79
10.49
11.50
13.88
17.43
ßßß
ßßß
ßßß
ß ßß
3.88
4.68
5.89
5.61
7.33
9.01
8.40
9.68
9.14
11.14
10.87
13.37
12.89
13.41
13.22
13.62
4.42
5.60
6.31
6.87
7.92
8.66
9.07
10.55
10.42
11.18
12.52
15.33
18.71
7.65
11.54
15.36
21.03
Am
Am
Am
ßß ß
Am
ß- ß
...
.....
Am
ß ßß
Am
ßßß
Lep(e), VW
Lep(e), VW
ßßß
Am
ßßß
ßßß
ß ßß
Am
ßßß
ß ßß
ßßß
ßßß
ßßß
ß ßß
ßß ß
Lep(x), W
ßßß
ßßß
ßßß
ßß ß
Na-montmorillonite
3.79
ßßß
4.24
ßßß
Crude Swy-1 montmorillonite
3.64
..-
4.16
...
See materials and methods section for explanation of mode of calculation. Am, amorphous (no
electron diffraction patterns of iron oxides detected); Lep(e), lepidocrocite, detected by electron
diffraction; Lep(x), lepidocrocite, detected by differential XRD (U. Schwertmann, personal communication, 1989); VW, very weak; and W, weak.
mined using a model 5000 Perkin-Elmer atomic absorption
spectrophotometer.
aluminum sample holder. Scanning was done in the range
40-75ø 20 at 0.5 ø min -• .
Iron Extractability
Transmission Electron Microscopy
Dithionite-extractable iron (Fed) was determined following Mehra and Jackson [1960]. Oxalate-extractable iron
(Feo) was analyzed following the method of McKeague and
Day [1966]. The sampleswere shaken in darkness for 4 hours
with 0.2 N ammonium oxalate, p H = 3. Iron concentration
in the extracts was determined by atomic absorption spectroscopy as indicated above.
A drop of a sample suspension was deposited on carboncoated copper grid, dried, and observed using a JEOL JEM
100CX microscope or, alternatively, deposited on holeycarbon films on copper grids and observed in a Hitachi
H-500H microscope.
Specific Surface Area
The specific surface area (SSA) values were measured by
the ethyleneglycolmonoethylether (EGME) adsorption
method [Ratner-Zohar et al., 1983].
X Ray Diffractometry
X ray diffraction (XRD) patterns were measured with a
Philips PW 1720 x ray diffractometer equipped with Co x ray
tube and Fe filter. Unoriented samples were obtained by
spreading powders on an adhesive tape mounted on an
Selected Area Electron Diffraction
Drops of dilute aqueous suspensions were placed on
holey-carbon film covering a Cu grid, dried at room temperature, and then inspected in an Hitachi H-500H analytical
electron microscope. Electron diffraction patterns (DPs)
were obtained either from single grains or from larger areas
of the specimens. The point-to-point resolution of the microscope in conventional imaging mode was about 0.5 nm.
Selected area electron diffraction (SAD) patterns could be
recorded from particles as small as 100 nm laterally. All
images and analyses were obtained at 100-KeV accelerating
voltage.
20,834
BANIN ET AL..' NANOPHASEIRON MINERALS IN MARS SOIL
a laboratory process [Banin, 1973; Gerstl and Banin, 1980]
which
generally simulates, but considerably enhances, the
Analyses were performed using a DuPont model 1600
natural
reaction sequences of hydrolysis-precipitationhigh-temperature differential thermal analyses (DTA) oven
equipped with a model 910 cell base. Samples were heated oxidation or oxidation-hydrolysis-precipitation occurring
from ambient room temperature (22øC)to 1200øCat a rate of during the oxidative weathering of iron-containing primary
10øC
min-• , withN2 flowingthrough
theovenat a rateof 10 silicate rocks on Earth. The process involves slow titration
mL min-• . A seriesof 20-to 30-mgsamples
of iron-enriched of acidic clay with a solution of Fe(II) salt while the p H is
montmorillonite clay (E6(II) numbers 1-16) was tested controlled by the presence of OH ion exchange resin. A
against a reference of 30-mg aluminum oxide. The areas of mixture of freshly prepared hydrogen-saturated montmorilthe endothermic and exothermic peaks were calculatedusing lonite and OH-saturated ion exchange resin (R) is titrated
Differential Thermal Analyses
the standard DuPont DTA curve-analysis software.
with 1 N FeC12.The addedcationreplacesH + fromtheclay
Spectral Diffuse Reflectance in the Visible Range
can be summarized
Samples were packed by lightly pressing into a metallic
sample holder, and the surface was smoothed. Diffuse reflectance was measured in the wavelength range 380-840 nm
using a Licor Li-1800 spectroradiometer equipped with a
BaSO4-coatedintegrating sphere. Reflectancefrom a clean
BaSO4 surface positioned in place of the sample was recorded also, and the sample reflectance relative to BaSO4
was computed. Digitized spectral data were transferred to a
computer and smoothed by a moving-average approach.
Difference spectra were then computed by subtracting the
spectrum of the pure clay that has undergone all the preparation steps described above, except for the addition of the
iron in solution, from the spectra of the iron-enriched clays.
Magnetic Properties
The bulk susceptibility of the sampleswas measured on a
KLY-1 susceptibility bridge. Isothermal remanent magnetization (IRM) acquisition in fields up to 1 T was used to
identify the ferromagnetic/ferrimagnetic minerals [Lowrie
and Heller, 1982]. The Curie temperature analysis was
performed using a horizontal motion Curie balance [Lebel,
19851.
Simulations of Viking Biology Experiments
The Labeled Release (LR) experiment.
The release of
•4Cfromlabeledformatesolution(0.1 mL of 2.5 x 10-4 M
H•4COONa)addedto 100-to 500-mgmineralpowderswas
followed by the "CO2 adsorption method" suggestedby
surface, while the C1- exchanges with OH- from the resin.
The H + and the OH- combine to form water. The reaction
as follows:
2H+-clay+ 2R-OH- + FeC12-•Fe2+-clay
+ 2R-C1+ 2H20
(1)
This equation describesthe reaction up to consumption of
all the exchangeableH + from the clay [Gerstl and Banin,
1980].
Abovethe cationexchange
capacity,Fe2+ addedto the
suspensionis oxidized and hydrolyzed. Since the oxidation
is controlled by nitrogen bubbling and slowed down due to
the acidity of the suspension, mixed Fe(II)/Fe(III) solids
tend to form. We can represent these compounds by the
hypothetical ferrosic hydroxide [Lindsay, 1979] and describe
the formation by
3Fe2++ 8H2¸ -• Fe3(OH)8
(ferrosic
hydroxide)
+ 8H + + 2e-
(logK ø= -43.75)
(2)
where K ø is the equilibrium constant.
Alternatively, the formation of C1 green-rust I [Bernal et
al., 1959; Taylor and McKenzie, 1980; Taylor, 1984a, b;
Vins et al., 1987] may be taking place, schematically represented by the reaction
Fe2++ 0.2C1- + 2H2¸
-• {[Fe(II)0.
8Fe(III)0.
2(OH)2]
(0'2+)
(C1)(ø'2-)•
0.2
J
(greenrust,reduced) + 2H + + 0.2e-
(3)
Levin and Straat [ 1976] and described in detail by Banin and
Rishpon [1978, 1979] and Banin and Margulies [1983]. The
The Fe(II)/Fe(III) ratio in green rust can vary within a
relatively
wide range (4-0.7), depending on the degree of
rate of the •4C releaseand the percentage
releasedwere
compared to those observed in the LR experiments con- oxidation [Bernal et al., 1959]. In both reactions (2) and (3),
ducted on board the Viking Landers on Mars [Levin and H + partly reacts with OH- releasedfrom the resin by
excess C1 and partly with e- and air-oxygen to give H20.
Straat, 1977].
The Pyrolytic Release (PR) experiment. Simulations The overall reaction for the green rust formation is then
with Ca, Fe, and H montmorillonite were conducted in the
test standard module (TSM) of the Viking Lander by Hubbard [1979, personal communication, 1980] in the "light,
dry, active" mode. The results were compared to those
obtained in the Viking Lander experiments for Mars soil
[Horowitz et al., 1977].
RESULTS AND DISCUSSION
Nanophase Iron Oxide/Oxyhydroxide
on Clay Surfaces
Formation and solubility. The iron phases in the Mars
soil analog clays are deposited on the surface of the clay in
Fe2++ 2C1- + 0.1H2¸ + 0.0502+ 1.8R-OH
-• {[Fe(II)0.8
Fe(III)0.2
(OH)2](ø'2+)(C1)(ø'2-)•
0.2
J
(green rust, reduced) + 1.8R-C1
(4)
Note that green rust has excess positive charge; therefore it
will tend to sorb onto the negatively charged surface of the
clay.
Because of its instability, data for the equilibrium constant
of green rust are lacking, but it is known to be less soluble
than Fe(OH)2 [Arden, 1950]. It was suggestedthat ferrosic
hydroxide, a hypothetical compound not identified yet,
BANIN ET AL.' NANOPHASEIRON MINERALS IN MARS SOIL
SAMPLE
NO.
pH:
2
3
4
5
6
7
8
9
10
14 t5 '16
pH
3.!
$.6 3.9 40 5.2
,:,,: ,o..,,,i-,o., ,,.,
'"' _L
' 7.2 7.47.47.0
,
t ttt
5.1
5.3
EC:
11 12 '13
20,835
5.3
5.3 5.3
5.4 5.4 0.0
3-$ dS/m
TITRATION
TIME
Fig. 1. Synthesisof np iron oxide/oxyhydroxides
on smectiteclay surfacesby the quantitativeion exchange
method[Banin, 1973].The pH and electricalconductivity(EC) were monitoredcontinuously.Eh, the redox potential,
was measuredin samples1-16 periodicallyas the additionof FeC12proceeded,and pe was calculatedfrom the data
(series E6(III)).
should actually be replaced in the stability diagrams of iron
oxides by green rust [Taylor et al., 1985].
Further oxidation with air of green rust produces either
lepidocrocite, goethite, maghemite, or magnetite. In synthesisexperiments, mixtures of these mineralstend to form, the
between various componentsin the iron oxide/oxyhydroxide
system
isto plotthefunction
log[(Fe2+) + 2pH]versusthe
redox parameter (pe + pH) [Schwab and Lindsay, 1983].
The plot (Figure 2) showsthese relationshipsfor the various
iron-enriched clays immediately after the titration and fol-
proportionsprofoundly dependingon the conditions.Lepidocrocite is preferentially formed in neutral conditions, at a
medium rate of oxidation, at lower temperatures, and in the
presence
ofC1- andSO42, whereas
goethite
tendstoformin
Fresh suspension •mmediately
oppositeconditions [Taylor, 1980; Vins et al., 1987; Carlson
and Schwertmann, 1990]. Magnetite forms at higher pH and
lower redox [Taylor, 1980; Tamura et al., 1984a, b]. The
conversionto lepidocrocite can be representedby
11
! after t•trat•on
10-
ß
{[Fe(II)0.8Fe(III)0
ß2(OH)2]
(0'2+)
(C1)(ø'2-)t
0.2
J
8-
(green rust, reduced) -->
after
24 months
•n
suspension
9-
,o'i
7•
y-FeOOH(lepidocrocite)
+ H + + 0.8e- + 0.2C1- (5)
During preparation, p H gradually rose from 3.1 to 5.2
while the exchangeablehydrogen was titrated, then steeply
increased to 6.3, and then slowly rose from 6.3 to 7.4 over
the rest of the procedure(Figure 1). The redox potential,pe,
decreased initially as p H rose, then stabilized. Electrical
conductivity was very low throughout the whole synthesis,
indicating low free salt concentrations in suspension.
The later stages of the process, during which most of the
iron was deposited, occurred within a narrow range of pH
(6.6-7.4), and at practically constant overall redox (pe +
pH = 11.5-12.0). As discussedabove, some form of green
rust tends to precipitate at this stage. Green coloration was
noticed in the reaction mixture during synthesis, especially
in the initial and intermediate steps. The suspensionusually
turned tan to light orange as titration advanced in time and
oxidation proceeded.
A generalized form to present the stability relationships
5
4
3
!o
!1
12
13
14
pe + pH
Fig. 2. Solubilityrelationshipsof iron oxide and oxyhydroxide
mineralsand of np iron oxide/oxyhydroxidesdepositedon smectite
claysurfaces
(series
E6(III)).Measurements
ofpH, Eh,andFe2+in
solution were taken immediately after preparation and after 24
months of storage at 5øC.
20,836
BANINET AL.: NANOPHASE
IRONMINERALSIN MARSSOIL
.8
o
F©, mmol/g
1.2
1.6
2.0
2.4
we obtained montmorillonite with increasing iron content
(up to circa 18-19% total Fe203), as expected. As reported
1.o
earlier, the modification procedure removed almost all the
sodium, calcium, manganese,and phosphorousand some of
the potassium and magnesium [Banin et al., 1988a]. No
L
change occurred in silicon, aluminum, and titanium concentrations. In the parent clay, Na, Mn, P, K, Ca, and Mg
1•1.5
mostly appear as adsorbed ions, soluble salts, and minor
AI
Ill
accessoryminerals such as calcium and magnesium carbon•.4
ates. Magnesium is also a component of the crystal structure
z
FeI•
of the clay and therefore cannot be completely removed
•.2
during the modification. Silicon, aluminum, and titanium are
Mg
part of the crystal lattice of the clay and therefore are not
6.4
9.6
12.8
16.0
19.2
312
influenced by the modification procedure, which is shown to
TOTAL
IRON,
F©203,
%
be primarily a surface reaction.
Iron extractability. Iron extractability and surfacephase
Fig. 3. Exchangeable (KC1 extractable) iron, magnesium, and
aluminum from np iron oxide/oxyhydroxidesdepositedon smectite composition were measured by three methods, and the
results are presented in Figures 3 and 4.
clay surfaces, plotted versus total iron content in the system.
Sequential extraction with 1 N KC1 followed by 0.5 N
HC1 fractionated iron into the electrostatically bound exlowing 24 months of aging in suspension.Solubility lines for changeable fraction and the acid extractable fraction, rethe various minerals are also shown. At the pe range of the spectively [Gerstl and Banin, 1980; Banin et al., 1988a].
experiments,
Fe2+ is thedominant
species.
For thecompu- Exchangeableiron first increases(samples 1--4), stabilizesat
tation the experimental data for total iron in solution were --•0.4 meq/g (5-8), then drops to --•0.2 meq/g (Figure 3). The
correctedfor Fe speciation,and Fe(OH)+ was subtracted. compensatingions are A1 and Mg, together balancing the
Dueto theverylowtotalionicconcentration
(10-4-10-5 M), excess surface charge of the clay. Note that even at the
the activity coefficient was taken as 1. In the fresh suspen- highest loading in the series, a fraction of the iron remained
sion, ferrosic-hydroxide or another compound of similar exchangeable and electrostatically bound to the clay's surface. The HCl-extractable iron increases, as total added iron
solubility seems to have been formed in suspensions7-16.
At lower iron additions the suspension is undersaturated increases,up to sample 8 (--•0.75 mmol Fe added/g clay).
with respect to this mineral due to adsorption of iron as an Beyond that point, decreased extractability with HC1 is seen
exchangeable ion. Aging caused slow transformation into
more stable crystalline iron oxides, involving perhaps comFe, mmol/g
pletion of iron oxidation. After 24 months in suspension,iron
o
.s
'•.o
'•.s
;.o
;.s
2.5
,
,
was controlled by a solid more soluble than lepidocrocite
and maghemite and less soluble than Fe(OH)3. At the lower
iron additions (suspensions 1-4), undersaturation is still
-'0"- Feo
"-0-- Fed
observed and adsorption still controls iron solubility.
2.0
•
Fe(KCl+ HCl)
Evidently, during aging in suspension, continued oxidation of the iron causes slow transformation into thermodynamically more stable mineral form(s). This sequence is
typical for many natural environments on Earth in which
ferrous iron in solution is oxidized and insoluble green rust,
or similar partly oxidized compound, is rapidly precipitated
initially.
Upon
further
oxidation,
these
intermediate
com-
pounds transform to either goethite, lepidocrocite,
maghemite, or ferrihydrite [Taylor, 1987]. Ferrosic hydrox-
ide was suggestedto form in soils under conditions of
cyclicaloxidation-reduction
of iron at relativelyhighpH
values (6-7) and to control iron solubility at (pe + p H) > 8
[Schwab and Lindsay, 1983]. Despite the fact that many
other iron oxide/oxyhydroxide minerals are more stable
thermodynamically, green rust or ferrosic hydroxide are the
first minerals to form, and transformation to the other, more
stable minerals may be rather slow, especially when interfering ions such as A1 and Si are present.
Total
iron content.
Iron content
in the various
series of
preparations is given in Table 1. The crude SWy-1 montmorillonite contains 3.6-3.8% Fe20 3 in its lattice, mostly as
isomorphous substitution in the octahedral sites [van Olphen
and Fripiat, 1979]. Iron content in the modified clays was
always higher than in the crude clay, showing that the
structural iron was not removed during modification and that
0
4
8
12
1•6
20
TOTAL
IRON,
F©20
3,%
Fig. 4. Extractable iron from np iron oxide/oxyhydroxidesdeposited on smectite clay surfaces, plotted against the total iron
content. Feo is oxalate extractable (oxalate at pH = 3.0) (amorphous oxides); Fed is DCB extractable (dithionite at pH = 7.0)
(crystallineoxides); FeKCi+HC
! is sum of iron extracted by sequential extractions with 1 N KCl (exchangeable) and 0.5 N HCl (acid
extractable) (series E6(III)). Bars represent the analytical standard
deviation, three replicates.
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
20,837
Fig. 5. Nanophase iron oxide/oxyhydroxides deposited on smectite clay surfaces. Transmission electron micrographstaken at x 100,000magnification.SeriesEt(III); samplenumbersand total iron as % Fe203 (105øCbasis)are (a)
sample 1 (4.01%; clay alone); (b) 5 (7.39%); (c) 7 (8.27%); (d) 8 (9.52%); (e) 9 (9.79%); (f) 10 (10.5%); (g) 12 (13.3%);
(h) 13 (13.9%); (i) 15 (17.4%); and (j) 16 (19.2%).
and easy to bend and wrinkle. When iron oxide content
reached 7.39%, the clay particles showed increased electron
density and appeared to be more rigid, suggestinga fine
coatingof iron oxide/oxyhydroxide.Above 13.3% Fe20 3 the
clay plates appear again to become less opaque with the
discreteiron oxide particles spreadover the basal surfacesof
the clay plates (Figures 5f-5j).
(Fea) procedures(Figure 4). The Feo/Fea ratio throughout
We estimatedthe particle size of the np iron precipitatesin
the series was close to 1 (1.07 _+0.16). This indicated that all the different preparationswhere they could be discerned. At
the added iron formed poorly crystalline phases. Lowered 9.79% Fe203 the iron oxide particleswere few, were spherextractability of added iron at the highest iron contents ical in shape, and their size was less than 3 nm. At 13.3%
(Figure 4) suggestsperhaps the formation of more stable Fe20 3 the particlesbecamemoredenseandtheir sizeranged
phase(s).
between 6 and 9 nm. At 16.5% Fe203, needle-shaped
(Figures 3 and 4). Extraction with oxalate [McKeague and
Day, 1966] is a procedure for estimating amorphous, noncrystalline or poorly ordered oxides, while extraction by the
dithionite-citrate-bicarbonate (DCB) procedure [Mehra and
Jackson, 1960] dissolves also crystalline oxides of iron (but
not structural iron of alumosilicates). In series E6(III), all
addediron was extractedby both the oxalate (Feo) and DCB
Particle Size of Iron Phases
As previously stated, a critical characteristic of the iron
phases in Mars soil analogs is their extremely small particle
sizes, assumed to be in the nanometer range. We employed
two approaches to measure or estimate the particle sizes of
the deposited iron oxides: direct transmission electron microscopy and specific surface area measurement by polar
compound sorption. Both methods have shown that the iron
phases formed on the clay surfaces are indeed nanophases.
Transmissionelectron microscopy. No discrete particles
of iron oxides were observed at low iron additions, though
the clay particles became electron dense, suggestinga fine
coating with the added iron (Figures 5a-5e). Above 7-8%
Fe203, small, discrete electron-dense particles were observed, while needle-shaped particles appeared above 12-
13% Fe20 3. The particles seemedto precipitateon or to be
adsorbedto the clay plates and have never been observed as
independent entities (Figures 5 and 6). Subtle changeswere
observed in the appearance of the clay plates themselves.
The plates of the pure hydrogen clay and after low additions
of iron (up to 5.89% Fe20 3) appearedrelatively transparent
particles are observed also. Their width-to-length ratio is
7/100 nm, while the spherical particles remain in the size
range6-9 nm. At 19.2% Fe203 the sizesare about 3-5/100
nm and 7-25 nm for the needle-shaped and the spherical
particles, respectively (Figure 6). These measurementsrefer
to the well-definedvisible particles. Many (probably most) of
the iron oxide particles in the system have smaller particle
sizes, at the limit of resolution of the TEM we used, and
therefore were not measured. Even so, the observed particle
sizes are generally smaller than those usually reported for
crystalline iron oxides and oxyhydroxides. Spherical particles were ferrihydrite 35-75 nm [Eggleton and Fitzpatrick,
1988], hematite 100-150 nm [Sidhu, 1988], and maghemite
220-260 nm [Taylor and Schwertmann, 1974]. Needleshaped particles were lepidocrocite 100 nm [Lewis and
Farmer, 1986] and goethite 200-800 nm [Schwertmannet al.,
1985].
Specific surface area. No change occurred in the SSA
values with increased iron additions, and the average value
for the whole serieswas 816 +- 17 m2/g (Table 2). This
observation is puzzling at first. Theoretical calculations
(following Dyal and Hendriks [1950]) of the expected SSA,
20,838
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
Fig. 6.
Nanophaseiron oxide/oxyhydroxidesdepositedon smectiteclay surfacesobservedat increasingmagnifica-
tionsin theTEM (x 104to x2.5 x 105magnification).
Sample
E6(III)-16;totalironcontent
19.2%asFe203.
according to unit cell dimensionsand formula weight, sug-
Assuming additivity of clay and oxide contributions, the
gestthatit shouldhavedecreased
from797m2/gfor the H measuredspecificsurfacearea (Sm) is given by
clay to 762 m2/g for the Fe clay. A decreaseof this
magnitude should be easily detected with the method we
employed. Furthermore, assumingthat the clay is the exclusive component influencing SSA, a decrease in SSA should
take place due to "dilution" of the clay with the added iron
oxides. Yet, as indicated in Table 2, SSA remained practically constant throughout the series. This leads to the
conclusionthat the iron phase contributed to the SSA of the
mixture to the same extent as the clay, i.e., that it had
extremely small particle size.
We calculatedthe fractions of clay (Fc) and oxide (F o) in
each sample by
Fc = (100 - T)/(100 - 4.01)
(6)
and
F o = 1.00-
Fc
(7)
respectively,where T is the measuredtotal Fe20 3 contentin
percent (4.01% is the content of the structural iron in the H
clay). The resultsare given in Table 2, convertedto percentages (F x 100) of added iron oxide and net clay.
S m = ScFc + SoFo
(8)
We now solve (8) under two alternative sets of assumptions.
1. So = 0; only the clay contributesto S m, and thus
Sc = Sm/Fc
(9)
2. Sc - constat 800m2/g;the precipitated
oxidealso
contributesto S m and its contribution is given by
So = (Sm- 800Fc)/Fo
(10)
The calculations by (9) predict (Table 2) an unreasonable
systematicincrease in the SSA of the clay if it was the sole
contributorto Sm; assumption1 is thus unacceptable.The
contribution of the iron oxides to S m calculated using
assumption2 (10) gives very high So at low additionsbut
converges
to a fairlyconstant
valueof 700-900m2/gabove
4-5% added Fe203 (Table 2). These values are higher than
those reported in the literature for most of the crystalline
ironoxidesandoxyhydroxides
(50-200m2/g[Schwertmann,
1987a]), exceptfor ferrihydrite that reportedly has SSA of up
BANIN ET AL.: NANOPHASE
IRONMINERALSIN MARSSOIL
TABLE
2.
20,839
Specific Surface Area of np Iron Oxide/Oxyhydroxides Formed in a Montmorillonite Clay Matrix (Series E6(III))
SpecificSurfaceArea,m2/g
Calculated (Total Contribution)
Iron Oxide Fe203, %
Measured
Sample
Total
Added*
Net Clay,?
%
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
4.01
5.19
5.89
6.46
7.39
7.97
8.27
9.52
9.79
10.50
11.50
13.20
14.00
16.00
17.40
19.20
0.00
1.23
1.95
2.55
3.52
4.13
4.44
5.73
6.02
6.75
7.80
9.61
10.40
12.50
13.90
15.80
100.0
98.8
98.0
97.4
96.5
95.9
95.6
94.3
94.0
93.2
92.2
90.4
89.5
87.5
86.0
84.2
Corrected$
Clay{}
Iron Oxideô
Avg.
s.d.
Avg.
s.d.
Avg.
s.d.
Avg.
s.d.
698
703
693
698
698
711
690
673
686
682
675
673
652
701
677
724
0.4
4.2
3.2
9.6
3.5
7.1
3.5
4.4
10.4
2.8
3.5
7.6
3.2
3.3
8.5
12.0
829
834
823
829
828
843
819
799
814
809
802
807
774
832
804
859
0.5
5.0
3.8
11.4
4.2
8.5
4.2
5.2
12.3
3.3
4.2
7.0
3.9
3.3
10.0
14.2
828
844
839
850
858
880
856
847
856
868
869
870
863
940
918
998
0.4
5.0
3.9
11.7
4.3
8.8
4.3
5.5
13.1
3.5
4.6
7.7
4.3
2.4
8.6
9.8
......
......
......
......
......
......
1221
779
1027
937
915
661
548
975
825
1172
94
91
201
49
38
69
35
77
65
90
Avg., average; s.d., standard deviation; three replicates.
*Calculated by (7).
?Calculated by (6).
$Corrected for pressure and temperature effects accordingto a standard Wyoming montmorillonite clay included in the series analyses,
andassuming
thatits specific
surface
areais 800m2/g.
õCalculated by (9).
ôCalculated by (10).
to 600m2/g.However,Oades[1984]hasestimated
a contri- routine powder XRD was not successful. The basal spacing
butionof 300-900m2/gfromiron-hydroxy
polycations
added for the clay could not be seen in the DPs because the clay
to montmorillonite.
was preferentially oriented parallel to the support film and
perpendicular to the beam. The (020) reflection of lepidocrocite, which is the most intense line of the mineral, was
not seen either. This suggests that the lepidocrocite was
exposeperunitmassis larger.To havean SSAof 750m2/g, oriented parallel to (0k0) on the clay surface. This orientaprecipitated iron oxide particles should have an average tion may be the result of either mechanical orientation or a
diameter of 2 nm, asstimingthey are sphereshaving a density topotactic crystal growth. Note that even when lepidocrocite
of 4.0 g/cm3, or a sideof 2 nmif theyarecubesof the same formed, most of the precipitated iron particles in the system
density. These values are somewhat lower than the range of remained as amorphous oxide/oxyhydroxides.
sizeswe estimatedfrom the transmissionelectron micrographs
The iron oxide/oxyhydroxide particles are seen to cover
(discussedabove). As pointed out previously, a wide size evenly the basal surfaces of the clay particles in samples
distribution exists, out of which we visually tend to detect and produced at the first stages of the titration (Figures 8a and
measure the larger particles seen in the electron micrographs. 8b), but preferentially binding to the edges at the later stages
(Figure 8c and, particularly, 8d). These observations may be
Mineral Identification
explained by the varying net surface charge of lepidocrocite
around its isoelectric point. At pH below the isoelectric
Direct
mineral
identification
and characterization
was
conducted by XRD and selected area electron diffraction and point (iep) of lepidocrocite, or of the nanocrystals of the
precipitated iron phase (such as the green rust discussed
assisted by electron microscopy, thermal analysis (DTA),
above), their particles will be positively charged and will be
magnetic measurements, and spectroscopic studies.
Diffraction studies. X ray diffractograms of all the sam- electrostatically attracted to the negatively charged sites of
ples in series E6(III), containingup to ---19% total Fe20 3 the clay. Because of the large basal-to-edge charge ratio of
were similar to that of the crude SWy-1 clay, and none of the the clay, they will be mostly bound to the basal surfaces. As
characteristic features of well-crystallized iron oxides or pH rises toward and beyond the iep (6.7 +_ 0.2 for lepidocrocite [Posey-Dowty et al., 1986]), the np iron oxide
oxyhydroxides was detected (Figure 7).
Electron DPs of selected areas revealed lepidocrocite crystals will acquire increasingly negative charge and will be
(3,-FeOOH) as the sole crystalline iron oxide/oxyhydroxide repelled from the negatively charged sites and adsorb to the
mineral present (Figures 8 and 9; Table 1). Identification was positive edges of the clay. The sorbed nanocrystals may act
based on lines at d-spacingsof 3.26, 2.34, 1.92, 1.83, and 1.51 as nuclei for further precipitation of iron and a site for the
3,, corresponding
to the (120),(111),(051)and(200),(220), growing crystallites. It is possible that, due to their anchorand (231) planes of lepidocrocite, respectively (ASTM card ing to the surface, further Ostwald ripening and crystal
8-98). Due to the small domain size, the identification by growth is slowed down or retarded.
We conclude that the particle size of the iron oxides in the
clay matrix is smaller than that of the crystalline iron
oxide/oxyhydroxides and therefore the surface area they
20,840
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
Q
3.37A
TOTAL
M
2.47A
2.59A3.12A
Q
1.49A
Q 1.e•A1.6gA
1.82A
IRON
F©203,%
13.75A
M
3.94 (SWy-1)
5.00•
/
2.11A
75
7o
es
eo
se
2.17A 2.33A
so
46
,m
•
•o
•
•o
16
lO
4.01
(1)
5.89
(3)
7.39
(5)
7.97
(6)
9.52
(8)
9.79
(g)
10.5
(10)
13.3
(12)
16.5
(14)
19.2
(16)
s
28,Co Ka
Fig. 7. X ray diffractogramsof iron-enrichedsmectiteclay powders.Total iron, as % Fe20 3 is calculatedto the
basis of the 105øCdry weight of the system and includes clay-crystal lattice iron (constant at 4.01%) and increasing
amounts of external iron oxide/oxyhydroxides. M, smectite-montmorilonite;Q, quartz; no peak for crystalline iron
oxides can be seen (series E6(III)).
Differential thermal analyses. The DTA curves of the
acid-treated montmorillonite clay and three iron-enriched
preparations containing increasing levels of added iron are
depicted in Figure 10. The first thermal event observed is
assignedto two overlapping endotherms, both due to water
vaporization. The first endotherm (---100-115øC) is due to the
vaporization of water adsorbed to the clay, while the second
(---125-130øC) is due to the vaporization of more tightly
bound water that is probably associated with the iron oxyhydroxide.
The second thermal event, an exotherm occurring at
260-280øC, is observed in the clays containing the higher
iron enrichments (Figure 10, curves c and d). This endothermic event is due to the dehydroxylation of •,-FeOOH (lepi-
docrocite) to form •,-Fe203 (maghemite). The intensity of
the endotherm increases with increasing amounts of added
iron. Thermomagnetic (TM) measurements (discussed below) reveal the formation of a magnetic phase at the same
temperature range. The temperature at which the 7-FeOOH
--• •,-Fe203 dehydroxylation conversion takes place in a
synthetic mineral was reported to peak at 308øC[Mackenzie,
1970]. Lower temperatures were reported for less crystalline
samples [Schwertmann and Taylor, 1979; Gehring et al.,
1990]. Even in syntheses under seemingly identical laboratory conditions, it was found [Gomez-Villacieros et al., 1984]
that the dehydroxylation, measured by differential scanning
colorimetry, peaked at 297øCand 270øCin two lepidocrocite
samples, which were apparently synthesized under similar
conditions. Schwertmann and Wolska [1990] reported that
dehydroxylation temperature increased from 270øC for a
pure synthetic lepidocrocite to 300øCfor one containing 12.5
mol % substitutedA1. From these observations,it is apparent that the dehydroxylation conversion of lepidocrocite to
maghemite may take place over a range of temperatures
(250-310øC), depending on the chemical composition, the
crystallinity, and the particle size. Another endothermic
reaction in this temperature range may be the oxidation of
magnetite(Fe304) to maghemite. This reaction peaks, however, at lower temperatures (less than 200øC [Sidhu, 1988].
Furthermore, no magnetite was detected in the electron
diffraction patterns of the iron-enriched clays, whereas lepidocrocite
was.
This transformation
is thus ruled
out as a
possibility.
The exothermic peaks observed at 320øC and 425øC are
clearly correlated with increasing levels of iron in the clay.
The 320øC exotherm
is less intense than the 425øC exotherm
and tends to shift to lower temperatures (from 335øC to
321øC) as the iron content of the clay increases. While earlier
studiesof lepidocrocite did not report an exothermic peak at
320øC [e.g., Mackenzie, 1957], it was recently observed by
Gomez-Villacieros et al. [1984]. Since no change in the XRD
peaks of maghemite was detected, the authors rejected the
possibility that their exotherm represents early conversion
of maghemite to hematite. N2 adsorption isotherms, analyzed according to the Brunauer-Emmett-Teller
(BET)
method, revealed profound changes in the porosity (volume
as well as shape) after the 325øC exotherm. The energy
released in the process of internal sintering and reduction of
BET surface area was suggested to be the cause of the
exotherm. However, Schwertmann [ 1987a, p. 233] seemsto
disagree with this interpretation and attributes the exothermic peak to early/partial conversion of maghemite to hema-
BANINET AL..'NANOPHASE
IRONMINERALSIN MARSSOIL
20,841
Fig. 8. Electron micrographsand selectedarea electron-diffractionpatterns of np iron oxide/oxyhydroxides
depositedon smectiteclay surfaces.SeriesE6(II); samplenumbersandtotal iron contentas % Fe203 (air dry basis)
are (a) sample I (2.78%; clay alone); (b) 5 (5.78%); (c) 9 (7.33%); and (d) 15 (9.68%). Diffraction lines for the
montmorillonite-smectite
substrateare observedin all samples.Lepidocrocite(L) is detectedin (d) the samplewith the
highest iron addition.
tite. It may be noted that a weak exothermicpeak at ---350øC lower temperatures. Gomez-Villacieros et al. [1984] obwas observed in various maghemites [Sidhu, 1988] and served the conversion in two synthetic lepidocrocites betentatively attributed to the removal of combined water from
within the crystals. This may supportthe interpretationof
Gomez-Villacieros et al. [1984].
The major exothermic peak at 425øC does not exhibit a
temperature shift with changesin iron content of th• clay.
This exothermmay be due to the transitionof •'-Fe203
(maghemite)-• a-Fe203 (hematite). In a synthetic lepidocrocite this conversion was observed at 508øC [Mackenzie, 1970], but in natural samplesof lepidocrociteit occursat
tween 415øC and 490øC, peaking at 450-460øC. The TM
studies (discussedbelow) show that the relative magnetization (Ms/Mo) in the first heating cycle decreased abruptly
starting at ---440øC, and the curve returned to show a
paramagnetic thermal decay behavior. This suggests that
this exothermic feature is indeed the maghemite-• hematite
conversionoccurringat a lower temperature due to the small
particle size of the iron oxides. On the other hand, another
exothermic peak, broader and less pronouncedthan the one
20,842
BANIN ET AL.' NANOPHASEIRON MINERALSIN MARS SOIL
....
:,•:::::---------•-•.•.:•.::::::,::
============================
::::..::::.:::::::•
'•:::::•
•;•:::•::::::::::-::•-•
.....
..... ;,,.::::
:::::::::::::::::::::::
.....
:•:.:•..:-.::•...•.
::::•.-,•-•-..•::½:::::::':':::'
............
•:-:•'::.:::,.
C..'-•:...:•
................
:•
..........
--':•-:••
.....
:,:::.
::::::::::::::::::::::::::::
ß
%:•.:..
............
.-...:•::....::.
:•:::
...•::
...........
:.:<.:::.•
.:..•:'
...-:::.-..:.:..::;.
-'::•'•:
............
'•::½::------•E:L
' ::'::•:". ::.:,.:
:•:--::"':•""½::::<•-:-'-•:<
.......................
•:/.:-:-:•:•;•:---------•-.:•-:..•::::::•:.:
:.;:•.:::::..•'•'::::.•
....
::::--•:.•::•½•::.:.
::::::::::::::::::::::::
•::::-:<::::::-•:--•4.•.:•;•:
'•:'
'•:
-:
,,.'•':•:•::::
....
••••<•,,...•.::•••
.......
:..............
...,-:..::
...............
?•<
..•...•..-......•:....•::••:•••••:
••.....•....••••<.••••••::•
........... .......................
::..-;•<-*
.......
::-.--:...-•.
•
:....
.......
Fig. 9. Nanophaseiron oxidesformingon smectiteclay surfaces,at an intermediateiron load, viewed at (a-c)
variousmagnifications
in the transmission
electronmicroscope.Few lepidocrociteparticlescan be morphologically
identifiedbut are not detectedin the electrondiffractionpattern.Here, lep, lepidocrociteneedles;np, nanophase
iron
-oxide/oxyhydroxide
particles(sampleE6(II)-14; 9.64% Fe203, air dry basis).
at 420øC, but nevertheless well established through the
quantitativeanalysisof the thermograms,is seento peak at
550-580øC. This peak is clearly related to the np iron
crystallites(and not the clay) in the system.It increaseswith
iron loading and becomes quite discernible in the more
iron-rich clays of the series. It is probably due to the
transition of maghemite--> hematite. Pure maghemite was
zation curve also suggests,as pointed out above, that the
maghemite --• hematite conversion was completed at
--•490øC.The 550øCpeak may be causedby crystal growth,
recrystallization,and/or sinteringof the np hematite to form
larger particles of better crystallinity. Such changes were
observed in thermal studies of maghemites [Mackenzie,
1970; Sidhu, 1988]. Since the particle size of the iron oxides
reported to convert to hematite at 500øC, while substituted residing on the clay surface is, originally, very small, the
maghemites containing traces (<1%) of transition metals systemmay have excesssurfaceenergy which is dissipated
converted at higher temperatures (up to 650øC) [Sidhu, at temperatureshigh enough to enable sinteringand crystal
1988]. It is usually observedthat maghemitesproducedby growth of the np hematite.
oxidation of magnetite are more stable and convert to
The broadendothermoccurringat --•675øCis accompanied
hematiteat highertemperaturesthan maghemitesproduced by the formation of water vapor and is due to dehydroxylaby dehydroxylationof lepidocrocite. The thermal magneti- tion of the clay. The sharp exotherm at --•950øCis due to the
high-temperaturetransition typical of montmorilloniteand is
distinct from other smectite clays.
E
Magnetic properties. We have recently found [Banin et
al., 1991] that heating of the iron-enriched clays, in which
lepidocrocite has been identified, to 250-300øC, renders
them magnetic to the extent that they adhere to a hand-held
magnet. This is in agreement with the observations that the
Mars soil as a whole adheredto insertedmagnets[Hargraves
et al., [1977]. The heat treatment of the iron-enriched clays
convertedthe antiferromagneticlepidocrociteto the strongly
ferrimagnetic maghemite, a well-known conversion. Hargraves et al. [1977] originally suggestedmaghemite as the
magneticmineral in Mars soil but have qualifiedthis suggestion, since a formation scenario on Mars could not be offered
at the time. Our present findings suggestthat the magnetic
propertiesof Mars soil can be attributed to partial conver0
2•)0
400
600
800
1000
1200 sion of np lepidocrocite to maghemite by relatively brief
heatingto 300øC,such as by meteoritic impacts or shallow
Temperature (oC)
magmatic intrusions. Detailed measurements of the magFig. 10. DTA curves of acid-treated montmorillonite and three
netic propertiesof the np amorphousiron oxide/oxyhydroxiron-enrichedclay samples(seriesE6(II)); samplenumbersand total
iron content as % Fe20 3 (air dry basis) are (curve a) sample 1 ide + np lepidocrocite on clay surfaces have addressed
(2.78%; clay alone), (curve b) 5 (5.78%), (curve c) 9 (7.33%), and various characteristicsof these samples.
(curve d) 15 (9.45%). Curves are offsetalongthe y axis for clarity.
Bulk susceptibility. Loading the smectite surfaces with
BANIN ET AL.' NANOPHASE
IRON MINERALSIN MARS SOIL
20,843
Samples with high-Fe content (e.g., E6(II)-13 and E6(II)-15)
show monotonous thermal decay up to around 280øC, followed by a sharp increase in magnetization due to the
Magnetic
SusceI?tibility,formation of a new magnetic phase (Figure 11). Above
m•/kg
Sample
Treatment
300øC, thermal magnetic decay of this phase and the paramagnetic phases are superimposed. A kink in the thermoUnheated Samples
magnetic curve at 560øCcan be attributed to the Curie point
E6(II)-I
0.54 x 10-8
of the magnetic phase formed during heating. No further
E6(II)-2
1.05x 10-8
decay is observed up to 700øC. Upon cooling, the increase of
E6(II)-9
8.0 x 10-8
E6(II)-13
12.6x 10-8
the magnetization starts around 520øC (Figure 11), i.e.,
E6(II)-15
15.6x 10-8
--•40øCbelow the Curie point found during heating. The lack
of a singular Curie temperature can be explained by the low
Heated Samples
E6(II)-13
heatedto 700øCat 10ø/min
28.9 x 10-8
concentration of ferromagnetic phases and the dominance of
E6(II)-15
heatedat 300øCfor 12hours
2.8 X 10-6
paramagnetic phases in the samples [Lowtie, 1990].
E6(II)-15
heatedto 300øCat 10ø/min
2.9 x 10-6
Isothermal reinanent magnetization. IRM of the unE6(II)-15
heatedto 700øCat 10ø/min
4.8 X 10-6
heated sample E6(II)-13 increased rapidly below 0.1 T and
reached saturation between 0.2 and 0.25 T (Figure 12a). No
further increase of the magnetization is observed in the
increasing amounts of Fe led to an increase of the bulk
maximum field (B = 1 T) applied. This suggeststhe occursusceptibility (Table 3). The increase is in the range expected rence of a low-coercivity phase at the surface of the Fefor paramagneticiron if one assumesthat 1 wt % of Fe203 enriched smectite, which probably formed during synthesis.
TABLE 3. Magnetic Susceptibility of Unheated and Heated
Samples of np Iron Oxides Deposited on Clay Surfaces
addsa valueof 2.07 x 10-8 m3/kgto theinitialsusceptibility
[Collinson, 1986]. Upon heating, bulk susceptibility of the
sampleswith higher iron content in the series increased, the
increase being more pronounced in sample E6(II)-15 than in
sample E6(II)-13 (Table 3). For sample E6(II)-15 most of the
increase took place between room temperature and 300øC
and is probably related to the lepidocrocite-to-maghemite
dehydroxylation conversion. The magnetic susceptibility of
the heated sample E6(II)- 15 brings it to the range of magnetic
susceptibilities estimated for the Mars soil, if total iron was
the same. The heat treatment converted the antiferromagnetic lepidocrocite to the strongly ferrimagnetic maghemite,
a well-known
conversion.
However,
the formation
of a
heat-stable phase (e.g., magnetite) may have also taken
place at some point during heating, as evidenced by the
residual susceptibility measured in the samples heated to
700øC after cooling back to room temperature (Table 3).
Thermomagnetization. The thermomagnetic curve of the
untreated sample E6(II)-I in a 0.25 T field between ambient
and 700øC is only weakly defined and essentially shows
paramagnetic behavior according to the Curie law [Collinson, 1986]. In the Fe-enriched samplesa significantparamagnetic contribution to the thermomagnetic curve is observed.
1.0
0.8
0.6
0.4JlE6(11)-13
.....
0.2
0.0
0.0
0.2
0.4
0.6
0.8
1.0
Applied Field [T]
1.0
1.00
0.80
•.,,,% ßheating
curve
ß,,,,,.,.-• o cooling
curve
(b)
0.8
0.6
0.60
0.4
0.40
E6(11)-13 Heated to 700oC
Mmax.- 3.58E-02 [A/m]
0.2
0.20
0.00
100
200
300
400
500
600
700
0.0
0.0
Temperature [øC]
Fig. 11. Thermomagnetization
curve of np iron oxide/
oxyhydroxides deposited on smectite clay surfaces. The thermomagnetization event at 280ø-300øCis attributed to the dehydroxylation conversion of lepidocrocite -• maghemite.
0.2
0.4
0.6
0.8
1.0
Applied Field [T]
Fig. 12. IRM of iron-enriched smectite clay containing amorphous np iron oxide/oxyhydroxides (sample E6(II)-13) (a) unheated
and (b) heated to 700øC and cooled back to room temperature.
20,844
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
1.1
TOTAL
field absorption modes of iron in an oxide-oxyhydroxide
matrix [Sherman et al., 1982; Sherman, 1985; Banin et al.,
1985]. The clay substrate has no absorption in this wavelength range, and the spectral characteristics of the np iron
oxides dominate the absorption curve. In laboratory studies
it was shown that the addition of just 1% by weight of
crystalline hematite to montmorillonite was sufficient to
cause a pronounced and detectable absorption feature at
450-550 nm [Singer, 1982]. The lack of pronounced diagnostic absorption peaks of crystalline iron oxides, even at high
deposition rates of iron oxides in the clay matrix, suggests
the lack of iron oxide crystallinity in our analogs, even
though direct quantitative comparison with Singer's data is
IRON
- Fe203, %
- 4.01
MARS
BRIGHT
REGIONS
ditficult
7.39
10.5/J
19.2
300
I
I
500
700
900
WAVELENGTH, nm
Fig. 13. Reflectancespectrain the VIS to shortNIR rangeof tip
iron oxide/oxyhydroxides deposited on smectite clay surfaces,
compared to the Mars bright regions. Top curve is for the clay
substratealone. All spectra are normalized to Mars spectrumat 0.75
/am (selected samples from series E6(III)).
The IRM is extremely sensitive to the presence of traces of
magnetic phases [Hirt and Gehring, 1991]. If the phase
present on the smectite surface is magnetite, its estimated
content is of the order of 0.005-0.01% (as Fe203), namely,
only some 0.05-0.1% of the total iron present in the system.
Measurement of IRM of the E6(II)-13 sample after it was
heated to 700øC again reveals a low-coercivity phase which
saturates at the 0.1 T field (Figure 12b). The magnetization
increasesby a factor of--• 10 due to the heating. The increase
of the magnetic intensity suggeststhat only a small amount
of the adsorbed iron formed a magnetic phase. Furthermore,
it is very likely that most of the amorphous Fe eventually
converts to hematite upon heating to high temperatures
(Figure 9). The IRM curve showed no indication of a
high-coercivity phase. Therefore it can be proposedthat the
hematite formed mostly had fine particle size showing superparamagnetic behavior.
Optical Properties
Mars reflectance spectra are characterized by almost
monotonous but significant decreasesin reflectance from the
NIR toward the UV [McCord et al., 1982b; Singer, 1982,
1985]. It has been shown that the reflectance spectra of
iron-enriched clays in the VIS strongly resemble the reflectance curves of the bright regions on Mars [Banin et al.,
1985, 1988a; Orenberg and Handy, 1992]. Neither the Mars
spectra nor the iron-enriched clay spectra show any pronounced, typical, and well-defined absorptionfeatures of the
crystalline iron oxides in the VIS (Figure 13).
The deposition of iron oxides on the clay surface caused
nonlinearly increased absorption at any given wavelength,
with a tendency to saturation (Figure 14). The subtle feature
around 450 nm was assigned to intraconfigurational crystal
due to differences
in the mode of iron oxide addition
used for the preparation of the two systems (chemical
depositionin the present study versus mechanical mixture in
Singer's study).
The peculiar nature of the spectral curve of Mars has been
generally attributed to the presence of np iron oxides. It was
recently specifically attributed to superparamagneticnp hematite [Morris et al., 1989, 1990; Morris and Lauer, 1990;
Bell et al., 1990]. Already in their 1989 paper, Morris et al.
raised the possibility that the spectral (and magnetic) properties of all np iron oxides are similar, so that unambiguous
identification of the Mars minerals may not be possible on
the basis of the spectral data. Our current data show that
indeed this may be the case. We find that amorphous iron
oxides and np lepidocrocite also agree quite well with the
martian spectra in the VIS-NIR range, as shown by the data
in Figure 13. It appearsthat on the basisof currently available
spectral data for Mars, it is still impossible to determine
unambiguouslyand singularly the exact mineralogical nature
of the iron oxide phases of Mars soil, beyond establishing
that they are nanophasesof low-crystallinity nature.
1.4
1.3
TOTAL
IRON
Fe203, %
1.2
1.1
EB .7
O
(/) .6
•
.5
.2
.1
5OO
7OO
WAVELENGTH, nm
Fig. 14. Computed absorbance spectra of np iron oxide/
oxyhydroxide deposited on smectite clay surfaces. Note the monotonous decrease
in absorbance
from the UV
toward
the short NIR
with only subtle features seen at --•450 nm, and the increased
absorbance,at any given wavelength, with increased iron content
(series E6(III)).
BANIN ET AL.' NANOPHASEIRON MINERALS IN MARS SOIL
t00
20,845
H-MONT.
VL-1,
CYCLE
FeMONT.
L..........
PALAP
......
,.',*,,2,
+-HAWAIIAN PALA
PAHALA
ASH (t6/15)
"'-BASALT ROCK (PULVERIZED)
x-BRITISH
"'
VL-1, CYCLE
I
o
COLUMBIA
v-BRITISHCOLUMBIA
PALA ACIDIFIED
=" ,-CONTROL
40
• 0•••••1
- t
0
I00
200
T I M E
o
PALA
•
I
300
400
(HOURS)
PALAGONITES
20
CONTROL
ACIDIFIED
PALAGONITE
0
o
100
200
T I M E
$00
( HOURS
400
500
)
Fig. 15. Kineticsof 14Creleasecaused
by Marssoilin twocyclesof theVikingLR experiment
(VL-1), compared
with simulations using Fe montmorillonite and H montmorillonite and various palagonites. The VL-1 curves are
redrawn from the results of Levin and Straat [1977].
Maximal rate of decomposition was observed at a total
Chemical Reactivity
Although not the original intention, the results of the three
Viking biology experiments are a valuable and unique source
of information on the chemical and mineralogical nature of
the Mars soil. These experiments, in essence, studied the
interaction of the soil with gases, water, salts, and organic
compounds [Klein, 1979]. By comparing the results of simulation experiments to the Viking biology results, we may
deduce, albeit speculatively, additional constraints on the
probable composition and properties of the martian soil and
its mineral components. Such simulation experiments have
shown that the chemical reactivity of the iron-enriched clays
strongly resembles, and offers a plausible mechanism for,
the somewhat puzzling observations of the Viking biology
experiments.
Carbon
14 was released
from the medium
solution
iron contentof 7-7.5% as Fe203. At that stage,iron addition
was equivalent to the cation exchange capacity of the clay.
Increased iron loading on the surface of the clay diminished
the decompositionrate somewhat (Figure 16). The reduction
in the activity at total iron content of 17.4% Fe203 was
similar to that observed when ferrihydrite, low-crystallinity,
fully oxidized iron mineral, was mixed with the most active
Fe clay. This suggests that the decomposition reaction,
involving decarboxylation and/or oxidation of organic acids,
requires the presence of both np iron phases and the extensive surface area of the clay silicate, at a balanced combination.
The main features of the Viking PR experiment were also
used in
the Viking LR experiment when reacted with the clays, at
rates and quantities similar to those measured by Viking on
Mars (Figure 15) [Banin and Rishpon, 1978, 1979]. The
comparative studies with palagonites have failed to mimic
the behavior of Mars soil even after they were acidified
(Figure 15) [Banin and Margulies, 1983; Banin et al.,
lOO
•
8o
j
6o
Fe-MONT.
1988b].
The component in the medium-solution that decomposes
most readily was formate, followed by lactate. Glycine, also
a component of the medium, did not decompose at a
measurable rate. The decomposition curve of the complete
medium was essentially a summation of the decomposition
of the formate and lactate, accounting for the slow-rate
decomposition reactions measured by Viking on Mars [Banin et al., 1981]. The reaction rate and its extent were
affected by the pH, by the type of exchangeable ion, by salt
additives, and by preheating. Heating of the active clay
(mixed with soluble salts) to 160øC in a CO2 atmosphere
showed, in some cases, a decreasein the decompositionrate
and activity, similar to the Viking results on Mars. However,
heating in air or an N2 atmosphere usually enhanced the
decomposition rate [Banin and Rishpon, 1979].
Fe-MONT.
•
+ FERRIHYDRITE
Fe-MONT. (17.4%Fe203)
20
SWy-1 MONT. + FERRIHYDRITE
CONTROL
0
100
200
300
TIME, hr
Fig. 16. Kineticsof I4C releasein the VikingLR simulation
experiment.
Percentreleaseof •4Cfromlabeledformatesolution
reacted with various iron-containing mineral analogs to Mars soil.
Fe-mont. is sample E6(III)-5 (7.39% total iron content); Fe-mont. +
ferrihydrite is sample E6(III)-5 to which 12% by weight of ferrihydrite was added; and SWy-1 mont. + ferrihydrite is crude montmorillonite to which 12% by weight of ferrihydrite was added.
20,846
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
900
leaching of the more soluble Mg-rich phases leaves behind
VIKING
8OO
an iron-rich
TESTS:
eC=Chr¾se Ptanitia (VL-t)
eU=Utopia Ptanitia (VL-2)
700
TSM
SIMULATION
• MT=
600
(various
•Fe- MT
TESTS:
Montmoril[onite
adsorbed
ions
H-MT
"o 500
I) Fe-MT
cu
400
I
<1: 3oo
'eC4
-eC5
ß C6
I)Fe- MT
•)Ca-
•oo
MT
Fe- MT
eC2
too
_
Background
.
.•-.---1. u 2
0
tO
20
PEAK-
$0
t,
40
50
60 70 x t04
dprn
Fig. 17. Comparison of Viking PR results with TSM simulations
using H-, Fe- and Ca-smectite-montmorillonite. TSM simulations
were conducted in the light, dry, active mode (J. S. Hubbard,
personal communication, 1980). Data for the Mars soil were taken
from Horowitz et al. [1977].
simulated using the PR TSM with iron-enriched and hydrogen-enriched clays (Figure 17) [Hubbard, 1979, personal
communication, 1980]. The Fe clay and H clay have both
simulatedthe typical results obtained on Mars, i.e., low CO2
sorption ("peak 1" readings) and significant abovebackground synthesis activity ("peak 2" readings). The Ca
clay showed excessiveCO2 sorptiondue to its higherpH. It
was thus shown that the typical chemical reactivity of the
martian soil can be simulated, by two independent methods,
using the iron-enriched smectite clays.
Modes of Formation and Relevance to Mars
Formation of iron oxides in weathering environments. Iron-containing primary silicate rocks readily weather
under the combined effects of humidity and oxidizing conditions upon exposure to atmospheric or hydrothermal environments. Fe(II) is the dominant iron species in the
primary rock-forming silicate minerals, and its oxidation to
Fe(III) is one of the major chemical changes, as is the
hydrolysis of iron.
Two major pathways of weathering may be grossly distinguished. One pathway is taking place in situ, within the
primary rock. The nucleation of the secondary minerals
starts at sites with localized higher iron atom concentrations
in the Mg-Fe or Ca-Fe matrix. The orientation and crystal
habit of the product minerals are strongly related to the
atomic orientation of the parent mineral [Eggleton, 1987].
An example is the weathering of olivine to iddingsite, a
yellow-orange oriented aggregate of goethite and saponite
[Eggleton, 1984]. As the alteration advances, increased
mineral.
The second major pathway involves complete dissolution
of the primary rock, release of its ionic components into
solution, and precipitation of new minerals, as dictated by
their solubility and the kinetics of the precipitation reaction.
Because of its higher solubility, Fe(II) is involved in most of
these transformations rather than Fe(III), and reaction sequences of oxidation-hydrolysis-precipitation or hydrolysisprecipitation-oxidationare typical, resulting in the formation
of various iron oxides and oxyhydroxides. Which mineral is
precipitated is profoundly dependent on many solution parameters. Although strict thermodynamic stability considerations predict goethite as the ultimate mineral form, because
it has the lowest solubility product, kinetic effects lead to the
nucleation and rapid precipitation of the various other iron
oxide minerals, as shown in the iron oxide formation scheme
(Figure 18). The resulting metastable system may evolve
only very slowly to a more stable one (Ostwald step rule).
Synthesis experiments in pure solution systems have
shown that the parameters that affect the precipitation of a
specific iron mineral from solution are numerous and complex. They include the Fe(II)/Fe(III) ratio, p H, redox potential, rate of oxidation, and concentration of anionic ligands
(C1, SO4, HCO3) and of cationic species(Si, A1, Mg, Mn)
(see reviews by Schwertmann [ 1987a, b] and Taylor [1987]).
In natural environments, such as in weathering rock and soil
systems, additional parameters include the water activity,
the concentration and composition of organic matter, the
presence of other minerals causing surface effects and retarding attainment of thermodynamic equilibrium, and the
presence of various trace elements [e.g., Schwertmann,
1987b].
Although it is qualitatively clear that the basic parameters
determining precipitation of an iron oxide should be the
activities of Fe(III) and H/OH in solution, no generalized
understandingof the effects of the various parameters listed
above on the precipitation products has been developed yet.
In other words, an overall understandingof the "selection
rule" for the formation of a given iron mineral is lacking. Let
us review the available information emphasizing lepidocrocite formation.
Green rusts: Important intermediateproductsin the formation processof lepidocrociteare the doublemetal ferrous-ferric
(or ferrous-metal) hydroxy (or hydroxy-anionic) salts of the
generalformula[M(II)xM(III)y(OH)(2
x+3y_zn)][mz-]nkH20
(Figure 18). These salts/mineralstypically have an octahedral
layer structureformed by the metallic cations and the hydroxy
groups,with the nonhydroxyl anion and the water residingin
the interlayers, in a structure characteristicof the pyroaurite
group of minerals [Allmann, 1968; Brindly and Bish, 1976].
These double iron-hydroxy minerals appear as dark green or
blue precipitates (green rust) during the oxidation of Fe(II)
solutionsat neutral or slightly acidic pH, and under terrestrial
conditions rapidly oxidize, changing their color to yellowbrown [Bernal et al., 1959;Taylor and McKenzie, 1980;Vins et
al., 1987;Hansen, 1989; Hansen and Taylor, 1990]. One such
typical product was formed by Taylor [1980] by oxidizing
Fe(HCO3)2 in solutionat pH 7. The dark greenprecipitatehad
a structure
analogous
to that of hydrotalcite
(Mg6A12(OH)16(CO3)4H 2O).
Interestingly, hydrotalcite is found to be the first mineral
rapidly forming upon the hydrothermal weathering of volca-
BANINET AL.' NANOPHASE
IRONMINERALSIN MARSSOIL
20,847
Fe2SiO4;(Mg, Fe) SiO4
(fayelite)
(olivine)
F•.
/'•-.-,-.--/•
''"-..
Oxidation
-F•
Rapid/complete
oxidation •
I Slow/partial'"-.. andhydrolysis
/' Fe(")-"•I oxidation
and
'"-..
i
i
i
i
i
\,
•
•
\
i
i
i
i
'\
i
i
i
i
i
•
•
',
•
"'
].AtpH5-7 • AtpH8-9
t
'
•
_
•'
Aggregation, I
i
I\
I (ferrihydrite)
I \
IX
\
In
'\
solution
i
i
\
i
dehydration
I
Dissolution\ \
I
and •
\
,
.
' \
Co.m.p!.ete
I
"'-..
\
OXlClaIlon
•
'"-•.
•
Fe304
(magnetite)
H
' ,
[Fe"(1-x)Fe"•x(OH)2]X*(Ax/n'yH20)xI•
(green
rust)
[ \
7-FeOOH
(lepidocrocite)
...,.IFe
i
i
i
i
I reprecipitation
\ \\
I
athigh
pH •
\
•
•1•
•
•/-Fe203
(maghemite)
c•-Fe203
(hematite)
i
i
i
i
i
c•-FeOOH
(goethite)
i
In dry
IHating
(200-250øC)
•II Hating
(600-700øC)
•
Ss;l•
Solid
phase
Major
•, Possible
oxidation
Heating (200-300øC)
dehydration
Legend:
structural rearrangement
mineral
pathway
state
pathway
'•'
Fig. 18. A schemedescribingthe formationpathwaysof the variousiron oxidesand oxyhydroxidesfrom primary
igneousminerals.Note the role, as intermediates,
of greenrustandferrihydritein the two majorpathwaysstartingwith
Fe 2+ '
nic glass and volcanic ash [Thomassin and Touray, 1982;
Jakobssonand Moore, 1986]. This is also seen during the
experimental alteration of basalt in seawater at 50øC [Cro-
visier et al., 1983, 1987]. Recently, mineralsfrom this group
have been identified in the alteration
crust of submarine lava
in the Red Sea [Ramanaidou and Noack, 1987]. These
studies suggestthat the mentioned precipitation products
constitute a short-lived, transitory phase and evolve toward
phyllosilicates of the 2:1 type [Crovisier et al., 1983] or
serpentine[Thomassinand Touray, 1982].It appearsthat the
group of double metal hydroxy minerals/salts of iron, aluminum, magnesium, etc., readily and rapidly nucleate at the
solution/rockinterface and are the first mineral phasesto
form near a weatheringvolcanicrock underlow-temperature
hydrothermal weathering conditions.
œepidocrociteformation: Thermodynamically, the double iron-hydroxy minerals are metastableand quite rapidly
"ripen" and transform into more stablephases.Green rusts
may evolve into lepidocrocite, goethite, maghemite,ferrihydrite, or mixtures of them, dependingon the environmental
conditions. Slight differences in the rate of oxidation, temperature, iron concentration,p H, total ionic strength, and
ionic compositionof the solution (Si, A1, anionic species)
profoundly affect the resulting mineral(s) [Taylor, 1987].
Full understanding of the mechanisms of formation and
effects of the environmental factors is still lacking, but
experimental observations have shown that the formation of
lepidocrocitefrom a hydroxy carbonategreen rust is favored
in comparisonto goethite by lower initial Fe(II) concentra-
tions,thepresence
of excess
S042-or CI-, andhigherrates
of oxidation [Taylor and Schwertmann, 1978; Taylor,
1984a, b; Schwertmann and Wolska, 1990]. At much higher
aeration rates, ferrihydrite is formed, and at limited aeration/
oxidation, magnetite becomes dominant.
Generally, lepidocrocite in terrestrial soils is believed to
be formed by the following mechanism. Under hydromorphic conditions and as a result of anaerobiosis, some Fe(III)
from oxides is reduced and solubilized. The Fe(II) hydroxy
is hydrolyzed and partly reoxidized to Fe(III)-hydroxy in
solution, and the two hydroxy iron speciesreadily form the
insolubledouble metal hydroxy salts (green rusts). These, in
turn, further "ripen" by oxidation to the more stable lepidocrocite or to goethite, the most stable oxyhydroxy iron
mineral. Thus most occurrences of lepidocrocite on Earth
have been documented in soils developed under water
loggingor generally hydromorphic conditions.
Apparently, these observations have led Fuller and Hargraves [1978] to suggestthat on Mars lepidocrocite may have
formed at fluctuating water tables. They further suggested
that this mineral may have formed a "mega-duricrust," the
remainders
of which are seen as the vesicle-rich
blocks on
the martian surface. In view of the ample evidence for the
volcanic
nature of the rocks and blocks on the surface of
Mars, and the lack of evidence for extensive near-surface
groundwater fluctuations, this suggestionwas not met with
widespreadacceptance.But the presenceof lepidocrocitein
the weathered soil on the surface of Mars may not be
completely rejected. Few recent reports have detected lepidocrocite in nonhydromorphic soils on Earth [Tarzi and
Protz, 1978; Ross and Wang, 1982]. Tarzi and Protz [1978]
20,848
BANIN ET AL.: NANOPHASEIRON MINERALS IN MARS SOIL
have identified lepidocrocite in a calcareous soil. They
suggested that it formed by direct incorporation of Fe(II)
from primary minerals (e.g., mica, chlorite) into the crystal
structure of the forming lepidocrocite and its later oxidation,
thus eliminating the solubilization step.
It is plausible that a similar mode of formation may take
place on the surface of primary minerals such as part of a
"surface-weathering"
mechanism. Because the major
known mechanisms of weathering on Earth involve excess
liquid water, they may not be relevant under Mars conditions [Gooding et al., 1991]. However, surface weathering
taking place at the atmosphere-rock microscopic interface
may. Such surface weathering has been observed by Allen
and Conca [1991] to affect rocks in the Dry Valleys of
Antarctica. They proposed that the pitting of the rocks is due
to short periods of time when the surface of the rocks sorbs
and accumulates water in small-scale dips and troughs,
causing rock weathering and secondary mineral formation.
At more extreme conditions, surface weathering may involve just the unfrozen water (one to three molecular layers)
present at the interface of any rock or solid phase by virtue
of the surface potential exerted by the solid [Banin and
Anderson, 1975]. These weathering reactions are, in general,
driven by the same driving forces that act on rocks in the
more
humid
environments
on
Earth
with
which
we
are
familiar. Overall, these weathering processescan be viewed
as reequilibration of minerals in volcanic or basaltic rocks,
formed under conditions of high temperature and low redox,
with their new conditions of lower temperature, higher redox
potential, higher water activity, and frequently, under soil
conditions having a more varied and less fractionated ionic
environment. So, it is expected that the products of these
weathering processes that may be proceeding on Mars may
be similar
to those
found
in the more
water-rich
environ-
ments of Earth. A major difference in rate is expected,
however. The lack of excess liquid water will tend to cause
chemical saturation of the aqueous phase and reduce the
overall rate of weathering. Furthermore, the low temperature prevailing in the extremely cold environment of Mars
slows down
the intrinsic
rate of the reactions.
In a set of
sequential reactions, such as is observed frequently during
weathering, it is expected that this decrease in rates will not
only affect the quantity of the weathering products but their
very nature and relative composition. An accumulation of
the initial and intermediate products such as hydrotalcite,
green rust, and less crystalline silicates is presumed at these
circumstances.
Nanophase iron oxides: Synthetic analogs to the iron
oxides in Mars soil. Many iron oxide minerals have been
proposed as candidates for the iron mineralogy of Mars soil.
In recent years, attention focused on np iron oxides as
satisfying most of the constraints imposed by the available
information on the soil. The two most studied analogsare np
hematite produced in confining matrixes and np iron oxides/
oxyhydroxide +np lepidocrocite depositedon clay surfaces.
Which of these two is the "true" Mars soil analog, if either?
Morris and coworkers [Morris et al., 1989; Morris and
Lauer, 1990] have simulated the spectral details of Mars
reflectance using various proportions of np hematite, (<10
nm size) contained in either silica gel or activated alumina
matrixes, mixed with bulk phase hematite. They concluded
that "... hematite is very likely present on Mars and may
even be the dominant ferric mineralogy" [Morris and Lauer,
1990].
The production of np hematite by Morris et al. [1989]
followed a common recipe for hematite production, namely,
prolonged calcination at high temperatures (550øC) of ferric
solutions predeposited and dryed (400øC) in the matrix
cavities. Samples with varying iron contents were obtained
by different numbers of cycles of ferric solution impregnation followed by drying and calcination steps. This procedure is clearly conducive to the formation of hematite, and
hematite only, since this mineral is the most thermally stable
iron oxide form. The procedure does not represent a naturally occurring alteration environment where hematite is
likely to form naturally, at least not as a monomineralic
product. It can be seen, however, as a method to produce np
iron oxides using the narrow cavities of the silica gel as a
"mechanical mold" which limits the sintering and growth of
the precipitated iron oxide crystals. The end product is thus
a valid mineral analog to Mars iron oxide nanophases,but its
formation scenario is highly artificial (as Morris and Lauer
[1990] point out).
In contrast, crystal growth of the np iron oxides deposited
on the surface of iron-enriched clays (present study) is
limited by the thermodynamic and electrostatic conditions
during precipitation and not mechanically. Furthermore,
precipitation is done at low ferrous iron concentrations,
involving in situ oxidation at acidic to near-neutral pH and at
low temperatures, conditions that represent potentially realistic weathering environments on Earth, or on Mars.
Another open question is whether np hematite is spectrally unique. This question was raised by Morris et al.
[1989] when they suggested np hematite as a Mars soil
component. The rather monotonous drop off in light absorption from the NIR to the UV, modulated only by very weak
features attributed to crystal field transitions at 450 nm, is
generally similar to that measured for the bright regions of
Mars [Singer, 1982, 1985; McCord et al., 1982b]. The
experimental results reported by Morris and Lauer [1990]
show, however, that the matrix (silica gel or alumina) in
which the np hematite is produced has a large influence on
the reflectance spectra of the np hematite and that the
relationship between Fe203 concentration and the reflectance properties is not unique and singular. For the same
Fe203 concentration, significantdifferencesin the position
and intensity of the absorption features is found, depending
on the matrix-scattering properties. Furthermore, the data
presented here show that np iron oxides/oxyhydroxides
formed on smectite clay surfaces also depict similar spectra
to Mars, as do a number of other noncrystalline or np iron
oxides. The accumulating data thus prove that this spectral
behavior is not unique for np hematite, suggestinglack of
uniquenessof the spectral reflectance curve (in the VIS) for
the individual np iron oxides. Continued laboratory studies
of various np iron oxides are necessary in order to better
define their specific spectral characteristics and find if there
are diagnostic spectral features that may be useful for
mineral-type identification. Until then, it is premature to
categorically assign to the iron oxide phases in Mars soil any
definitive
mineral
identification.
The diffraction, thermal, and magnetization studies of the
np oxides deposited in the clay matrix offer additional insight
into the possible nature and properties of the np iron
oxides/hydroxides in Mars soil. These studies, schematically
BANINET AL.: NANOPHASE
IRONMINERALS
IN MARSSOIL
np-Amorphous
np-Amorphous iron
oxyhydroxide
Aggregates of amorphous
iron oxide
iron oxides
ß
_•©•©•227
•©©o
ß
Go
•
© .(.•
280øC
ß
•
©o•©
©,,•
• IiDehydroxy
' ß;EZ•ß
N
•
lation
Polynuclear
• • [•
ferric ions
np-Lepidocrocite
(y-FeOOH)
\
ß
:•
ß
ß
325øC
280oci
1) Loss of hydroxyl/
water
(complete)
Hematite
np-Hematite
(•-Fe203)
Exotherm (small)
425oc
I
Non-magnetic
Polynuclear ferric cations
np-Amorphous iron oxyhydroxide
np-Amorphous iron oxide
np-Lepidocrocite(y-FeOOH)
:Z•
np-Maghemite
(•-Fe203)
•
•
Aggregates of amorphous iron oxide
np-Hematite (o•-Fe203)
•
Exotherm
(•
Exotherm (large)
Magnetic
•
Aggregation/
crystallization;
sintering
conversion
2)T--->Ot
conversion
(partial)
o
©
ß
570øC
425øC
•
np-Maghemite 3) Crystal growth
(7-Fe203)
Endotherm
Non-magnetic
ß•
20,849
Bulkhematite
(o(-Fe203)
Fig. 19. Structural and magnetic transformationsof np iron oxide/oxyhydroxides deposited on clay surfaces, as a
function of heating. The scheme is based on the results reported and discussedin the text.
summarized in Figure 19, suggestthat at low iron additions,
the Fe phases adsorbed onto the smectite surface are mainly
soil convert to maghemite (via lepidocrocite) in order to
develop the saturation magnetization observed in the Mars
amorphous. Samples with more than 8% Fe203 contain
soil.
some lepidocrocite and traces (not detectable by diffractometry) of ferrimagnetic oxides such as magnetite or
maghemite. The comparison of the IRM data and the bulk
susceptibility suggeststhat traces of magnetite or maghemite
can dominate the magnetic properties of the unheated samples. Both DTA and thermomagnetic data show that the
ferric phases at the smectite surface do not change upon
heating up to -280øC. Around 280øC, magnetization increases considerably. Such a change in the magnetization
has been observed for the transformation of lepidocrocite
into maghemite. Upon further heating, a decrease of magnetization due to the conversion of the newly formed ferrimagnetic maghemite to hematite is expected. A specific temperature at which this happens is not seen in the
thermomagnetic measurements of sample E6(II)-13 (Figure
12). This sample did not contain much lepidocrocite (no clear
electron diffraction patterns seen). It is likely that in samples
with higher lepidocrocite content, this conversion can be
noticed magnetically. An alternative explanation for the lack
of a noticeable transition is that another magnetic phase is
formed simultaneously with the transition maghemite-•
hematite. More data are needed to clarify these points.
Mode of formation of np iron oxides on Mars: An open
question. The accumulating evidence leads us to the conclusion that much of the iron oxide in Mars soil is noncrystalline or has crystals in the nanometer size range. This
conclusion represents, so we believe, a consensus (explicit
in some cases and only implicit in others) among the re-
The saturationmagnetization(Js) values for np hematite
are about 10 times higher than those of bulk hematite. To
account for the observations in the magnetic experiment on
Mars, all the iron in the soil has to be present as np hematite
[Morris et al., 1989]. Because Js of maghemite is about 3
orders of magnitude higher than that of hematite (and about
2 orders of magnitude higher than np hematite), it will be
sufficientthat just a small fraction of the np iron oxides in the
searchers
in the
field.
It
was
arrived
at on the
basis
of
accumulating excellent telescopic observations during the
late 1980s (1988-1990 Mars-Earth oppositions) and detailed
laboratory studies on soil analogs.
If we accept this conclusion, a new and intriguing question
is opened: what is the peculiar mode of weathering that has
produced these "amorphous," noncrystalline iron oxides?
More specifically, what is the cause for the limited growth of
the crystals of the iron oxide minerals in the weathered soil
on Mars? The currently most accepted scenario for soil
formation on Mars is that it took place during earlier epochs
(3.5-4.0 b.y. ago) when Mars may have been "warm and
wet" [e.g., Banin et al., 1992; Gooding et al., 1992]. It is
therefore puzzling that apparently only a minor portion of
the iron oxides has crystallized and developed a more
thermodynamically stable mineralogical composition and
particle size distribution over the long periods of time since
their formation.
Similar iron oxide entities are found on Earth only as
transitory phaseswhich ripen rapidly (geologically speaking)
to more stable minerals. Palagonites represent such a transitory assemblage of minerals produced as the initial weathering product of volcanic glass. The maximal age of palago-
20,850
BANIN ET AL.: NANOPHASE
IRONMINERALSIN MARSSOIL
nite depositson Earth is in the rangeof 106-107years tively) attested also to the noncrystalline nature of the iron
[Grambow et al., 1985; Crovisier et al., 1987].
compoundsprecipitated on the surface of the clay.
It is concluded that the clay acted as an effective matrix,
dust and soil is due to a relatively "young" weathering both chemically and sterically, preventing the major part of
product that has formed, and continues to form, at an the synthesized iron oxides from ripening, i.e., growing and
extremely slow rate, over the last several hundreds of developing larger crystals. The iron oxides were preserved
million to a billion years? It is then plausible that this as nanophase, short-range ordered grains. On the basis of
"recent" (in Mars time scale of changes) and nonevolved the data presented in this paper, we suggest that as such,
weathering product is coating or burying ancient, more they are appropriate models for the iron phases in Mars soil.
evolved weathered mineral assemblages that formed in the
The one crystalline component detected by selected area
earlier warm and wet epochs of Mars. This suggestion
electron diffraction was lepidocrocite (3,-FeOOH). The preapparently goes against the "common wisdom" on weathcursor for lepidocrocite is believed to be a double iron
ering. The low temperature and arid conditions on Mars
Fe(II)/Fe(III) hydroxy mineral such as green rust or ferrosic
seem to preclude hydrolytic weathering as a significant
hydroxide. In turn, lepidocrocite was converted to the more
process on its surface [Gooding et al., 1992]. However,
surface weathering, involving gas phase/solid phase reac- stable maghemite (?-Fe203) by mild heat treatment and
tions taking place at the atmosphere-rock microscopic inter- finally to np hematite(a-Fe203) by extensiveheat treatment.
The first conversion step results in increased magnetization
face may proceed, albeit at a very slow rate, even there.
In Marslike environments on Earth (e.g., Dry Valleys of of the mineral and renders the iron-enriched clay slightly
Antarctica), where weathering processes involving excess magnetic, to the extent that it adheres to a hand-held
water are limited or nonexistent, the products of the surface- magnet, as was observed with Mars soil.
The reflectance spectrum in the VIS of the clay-iron
weathering mechanisms become quantitatively important,
preparations
is generally similar to the reflectance spectra of
given enough time to proceed. The data available for Antarctic soils only cover a period of--•3 m.y. [Allen and Conca, bright regions on Mars. Both the Mars and the clay-iron
1991]. The Mars surface may be much older, perhaps up to spectra are characterized by almost monotonous but signif1 b.y. old [Arvidson et al., 1989b]. The global surface icant decreases in reflectance from the NIR toward the UV
stability (lack of plate tectonics) facilitates the accumulation and do not show any well-defined absorption feature of the
of weathering products over long periods. Even at an ex- crystalline iron oxides in the VIS. This spectral behavior has
tremely slow linear rate of rock-surface weathering of 0.3 been generally attributed to np iron oxides and was recently
nm/yr (about one molecular layer per year) an average layer specifically attributed to superparamagnetic np hematite.
of 30cmcouldbeweathered
onMarsoverthelast109years. This assignmentdoes not appear to be unique and singular,
Wind abrasion rates on Mars are estimated at 1 /xm/yr since as we have shown in this work, amorphous iron oxides
[Arvidson et al., 1979], i.e., 3-4 orders of magnitude faster and np lepidocrocite also agree quite well with the martian
than the assumed hydration front penetration rate; thus the spectra in the VIS-NIR range.
assumption of a linear rate of advancement of the weathering
The chemical reactivity of the iron-enriched clays strongly
front (rather than a square-root-of-time dependence) is warresembles, and offers a plausible mechanism to explain, the
ranted. This leads to higher overall rate of weathering than
somewhat puzzling observations of the Viking biology exestimated under diffusion-limited conditions [Gooding et al.,
periments.
The analogs
decomposed
14Cformate,the most
1992]. Particularly susceptible to such weathering are the
labile organic component in the media used in the LR
more recently erupted volcanic materials [Plescia, 1990].
experiment, at a rate and to the extent measured on Mars,
Estimatedat 26.4 x 106 km3 duringAmazonian
[Greelyand
and simulated the sorption and synthesis activities of the
Schneid, 1991], these materials may have supplied the highly
Mars soil measured during the PR experiments. These
unstable minerals whose weathering produced the thin reunique chemical reactivities are attributed to the combined
cent coating remotely sensed on much of the planet's surcatalytic effects of the iron oxide/oxyhydroxide and silicate
face.
phase surfaces.
The increasing realization that most of the secondary iron
SUMMARY AND CONCLUSIONS
minerals in Mars soil either have low crystallinity and/or
A series of surface-modified clays was prepared, contain- extremely small particles leaves us with an intriguing set of
ing np iron oxide/oxyhydroxides of extremely small particle questions. Why are the major minerals in Mars soil poorly
sizes, with total iron contents as high as found in Mars soil. crystalline? What are the weathering pathways? Why did the
The clay matrix and the procedure used for synthesispro- secondary iron oxides in Mars soil not "ripen" and develop
duced np iron oxides/oxyhydroxides containing a certain into well-crystallized phases despite the long period of time
proportion of divalent iron which slowly converts to more since aqueous weathering took place? Is this a result of slow
stable fully oxidized iron minerals with no increase in kinetics? Antarctic weathering analogs do not support a
kinetics argument. Alternatively, is it possible that the
particle size.
Transmission electron micrographs showed that the pre- minerals in the loose top soil, which is what we have
cipitated iron oxides appear as isodiametric or slightly elon- sampled and analyzed, are the accumulated products of
gated particles in the size range 1-10 nm. The SSA of the continuously on-going, extremely slow weathering proclay/iron oxide preparationsrevealed high specificcontribu- cessesof volcanic materials on the surface of Mars? Clearly,
tions from the iron oxide phases, commensurate with their to conclusively answer all these intriguing questions, more
extremely small particle sizes. Complete extractability of mineralogic work must be done on Mars soil and rocks and
added iron by both oxalate and DCB (Feo and Fed, respec- on their terrestrial analogs.
Is it possible then that the peculiar nature of the martian
BANIN ET AL.' NANOPHASEIRON MINERALS IN MARS SOIL
Acknowledgments. The work reported here was supported in
part by NASA Exobiology Research Program and by grants from
the National Research Council and Hebrew University. We are
indebted and grateful to many who helped during the years in the
studies of the iron-enriched clays: C. Sherban and J. Yableovitch
prepared the clays by the quantitative ion exchange method. D.
Hirsch and E. Ben-Dor helped in their chemical analyses. Y.
Rishpon has adapted the "CO2 absorption method" for the LR
simulation, and conducted the first LR simulation studies with the
clays. J. S. Hubbard supplied the data of the PR-TSM simulations
with three clays. U. Schwertmann was the first to detect lepidocrocite in the samplesand alerted us to its presence. A. Hirt helped in the
magnetic analyses, and M. R. White in the DTA study. M. Shen
provided considerable help in the technical preparation of the
manuscript. Much of the summary and write-up was undertaken by
A.B. while on an NRC Senior Associateshipin collaboration with G.
Carle at the Solar Exploration Branch, NASA, Ames Research
Center, Moffett Field, California.
REFERENCES
Adams, J. B., M. O. Smith, and P. E. Johnson, Spectral mixture
modelling--A new analysis of rock and soil types at the Viking
Lander site, J. Geophys. Res., 91, 8098-8112, 1986.
Allen, C. C., and J. L. Conca, Weathering of basaltic rocks under
cold, arid conditions--Antarctica and Mars, Proc. Lunar Planet.
Sci. Conf., 21,711-717, 1991.
Allmann, R., The crystal structure of pyroaurite, Acta Crystallogr.,
Sect. B, 24, 972-977, 1968.
Arden, T. V., The solubility products of ferrous and ferrosic
hydroxides, J. Chem. Soc., 75, 882-885, 1950.
Arvidson, R. E., E. A. Guinness, and S. Lee, Differential aeolian
redistribution rates on Mars, Nature, 278, 533-535, 1979.
Arvidson, R. E., J. L. Gooding, and H. J. Moore, The martian
surface as imaged, sampled, and analyzed by the Viking Landers,
Rev. Geophys., 27, 39-60, 1989a.
Arvidson, R. E., E. A. Guinness, M. A. Dale-Bannister, J. Adams,
M. Smith, P. R. Christensen, and R. B. Singer, Nature and
distribution of surficial deposits in Chryse Planitia and vicinity,
Mars, J. Geophys. Res., 94, 1573-1587, 1989b.
Banin, A., Quantitative ion exchangeprocessfor clays, U.S. patent
3,725,528, 1973.
Banin, A., Clays on Mars, in Clays and the Origin of Life, edited by
H. Hartman and A. G. Cairns-Smith, pp. 106-115, Cambridge
University Press, New York, 1986.
Banin, A., and D. M. Anderson, A similar law may govern freezing
in minerals and living organisms, Nature, 255, 261-262, 1975.
Banin, A., and L. Margulies, Simulation of Viking biology experiments suggestssmectites not palagonite, as martian soil analogs,
Nature, 305,523-525, 1983.
Banin, A., and J. Rishpon, Experimental simulation of the Viking
Labeled Release (LR) results with iron-adsorbed smectite clay
minerals, Life Sci. Space Res., 17, 59-64, 1978.
Banin, A., and J. Rishpon, Smectite clays in Mars soil: Evidence for
their presence and role in Viking biology experimental results, J.
Mol. Evol., 14, 133-152, 1979.
Banin, A., J. Rishpon, and L. Margulies, Composition and properties of the martian soil as inferred from Viking biology data and
simulation experiments with smectite clays, Third International
Colloquium on Mars, pp. 16-18, Lunar and Planet. Inst., Houston, Texas, 1981.
Banin, A., L. Margulies, and Y. Chen, Iron-montmorillonite: A
spectral analog of martian soil, Proc. Lunar Planet. Sci. Conf.
15th, Part 2, J. Geophys. Res., 90, suppl., C771-C774, 1985.
Banin, A., G. C. Carle, S. Chang, L. M. Coyne, J. B. Orenberg, and
T. W. Scattergood, Laboratory investigations of Mars: Chemical
and spectroscopic characteristics of a suite of clays as Mars soil
analogs, Origin Life Evol. Biosphere, 18, 239-265, 1988a.
Banin, A., L. Margulies, T. Ben-Shlomo, G. C. Carle, L. M. Coyne,
J. B. Orenberg, and T. W. Scattergood, Constraining Mars soils
mineralogical composition: Palagonite versus iron enriched smectite clays (abstract), Lunar Planet. Sci., XIX, 27-28, 1988b.
Banin, A., D. F. Blake, and T. Ben-Shlomo, Detection of nanophase
lepidocrocite (•/-FeOOH) in iron-smectite Mars soil analog materials (MarSAM) (abstract), Lunar Planet. Sci., XXII, 49-50, 1991.
Banin, A., B. C. Clark, and H. Wanke, Surface chemistry and
20,851
mineralogy, in Mars, edited by H. H. Keifer, et al., pp. 594-625,
University of Arizona Press, Tucson, 1992.
Bell, J. F., III, T. B. McCord, and P. D. Owensby, Observational
evidence of iron oxides on Mars, J. Geophys. Res., 95, 14,44714,461, 1990.
Bell, J. F., III, R. V. Morris, and J. B. Adams, Thermally altered
palagonitic tephra: A spectral and process analog to the soil and
dust of Mars, J. Geophys. Res., 98, 3373-3385, 1993.
Bernal, J. D., D. R. Dasgupta, and A. L. Mackay, The oxides and
hydroxides of iron and their structural interrelationships, Clay
Miner. Bull., 4, 15-30, 1959.
Brindley, G. W., and D. Bish, Green rust: A pyroaurite type
structure, Nature, 263,353, 1976.
Burns, R. G., Terrestrial analogs of the surface rocks on Mars?,
Nature, 320, 55-56, 1986.
Carlson, L., and U. Schwertmann,The effect of CO2 and oxidation
rate on the formation of goethite versus lepidocrocite from an
Fe(II) system at pH 6 and pH 7, Clay Miner., 25, 65-71, 1990.
Clark, B. C., A. K. Baird, R. J. Weldon, D. M. Tsusaki, L.
Schnabel, and M.P. Candelaria, Chemical composition of martian
fines, J. Geophys. Res., 87, 10,059-10,067, 1982.
Collinson, D. W., Magnetic properties of antarctic Shergottite
meteorites
EETA
79001 and ALHA
77005:
Possible
relevance
to
a martian magnetic field, Earth Planet. Sci. Lett., 77, 159-164,
1986.
Crovisier, J. L., J. H. Thomassin, T. Juteau, J.P. Eberhart, J. C.
Tourayl, and P. Baillif, Experimental seawater basaltic glass
interaction at 50øC: Study of early developed phases by electron
microscopy and X-ray photoelectron spectrometry, Geochim.
Cosmochim. Acta., 47, 377-387, 1983.
Crovisier, J. L., J. Honnorez, and J.P. Eberhart, Dissolution of
basaltic glass in seawater: Mechanism and rate, Geochim. Cosmochim. Acta., 51, 2977-2990, 1987.
Dyal, R. S., and S. B. Hendricks, Total surface of clays in polar
liquid as characteristic index, Soil Sci., 69,421-432, 1950.
Eggleton, R. A., The formation of iddingsite rims on olivine, Clays
Clay Miner., 32, 1-11, 1984.
Eggleton, R. A., Noncrystalline Fe-Si-Al-oxyhydroxides, Clays
Clay Miner., 35, 29-37, 1987.
Eggleton, R. A., and R. W. Fitzpatrick, New data and revised
structural model for ferrihydrite, Clays Clay Miner., 36, 111-124,
1988.
Evans, D. L., and J. B. Adams, Comparison of Viking lander
multispectral images and laboratory reflectance spectra of terrestrial samples, Proc. Lunar Planet. Sci. Conf., loth, 1829-1834,
1979.
Evans, D. L., and J. B. Adams, Amorphous gels as possible analogs
to martian weathering products, Proc. Lunar Planet. Sci. Conf.,
11th, 757-763, 1980.
Fuller, A. O., and R. B. Hargraves, Some consequencesof a liquid
water saturated regolith in early martian history, Icarus, 34,
614-621, 1978.
Gehring, A. U., R. Karthein, and A. Reller, Activated state in the
lepidocrocite structure during thermal treatment, Naturwissenschaften, 77, 177-179, 1990.
Gerstl,Z., and A. Banin,Fe2+-Fe3+ transformations
in clay and
resin ion-exchange systems, Clays Clay Miner., 28, 335-345,
1980.
Golden, D.C., R. V. Morris, D. W. Ming, H. V. Lauer, Jr., and
S. R. Yang, Mineralogy of three slightly palagonitized basaltic
tephra samples from the summit of Mauna Kea, Hawaii, J.
Geophys. Res., 98, 3401-3411, 1993.
Gomez-Villacieros, R., L. Hernan, J. Morales, and J. L. Tirado,
Textural evolution of synthetic 3/-FeOOH during thermal treatment by differential scanning calorimetry, J. Colloid Interface
Sci., 101,392-400, 1984.
Gooding, J. L., Chemical weathering on Mars: Thermodynamic
stabilities of primary minerals (and their alteration products) from
mafic igneous rocks, Icarus, 33,483-513, 1978.
Gooding, J. L., and K. Keil, Alteration of glassas a possible source
of clay minerals on Mars, Geophys. Res. Lett., 5,727-730, 1978.
Gooding, J. L., S. J. Wentworth, and M. E. Zolensky, Aqueous
alteration of the Nakhla meteorite, Meteoritics, 26,135-143, 1991.
Gooding, J. L., R. E. Arvidson, and M. Yu. Zolotov, Physical and
chemical weathering, in Mars, edited by H. H. Keifer, et al., pp.
626-651, University of Arizona Press, Tucson, 1992.
20,852
BANIN ET AL.: NANOPHASE
IRON MINERALSIN MARS SOIL
Grambow, B., M. J. Jercinovic, R. C. Ewing, and C. D. Byers,
Weathered basalt glass: A natural analogue for the effects of
reaction progress on nuclear waste glass alteration, in Scientific
Basis for Nuclear Waste Management, Proceedings of the
M.R.S. Symposium, edited by L. Werme, 50, 263-272, 1985.
Greely, R., and B. D. Schneid, Magma generation on Mars:
Amounts, rates, and comparisons with Earth, Moon, and Venus,
Science, 254, 996-998, 1991.
Guinness, E. A., R. E. Arvidson, M. A. Dale-Bannister, R. B.
Singer, and E. A. Bruckenthal, On the spectral reflectance
properties of materials exposed at the Viking lander sites, Proc.
Lunar Planet. $ci. Conf. 17th, Part 2, J. Geophys. Res., 92,
suppl., E575-E587, 1987.
Hansen, H. C. B., Composition, stabilization and light absorption of
Fe(II)-Fe(III) hydroxycarbonates ("green rusts"), Clay Miner.,
24, 663-669, 1989.
Hansen, H. C. B., and R. M. Taylor, Formation of synthetic
analogues of double metal-hydroxy carbonate minerals under
controlled pH conditions, I, The synthesis of pyroaurite and
reevesite, Clay Miner., 25, 161-179, 1990.
Hargraves, R. B., D. W. Collinson, R. E. Arvidson, and C. R.
Spitzer, The Viking magnetic properties experiment: Primary
mission results, J. Geophys. Res., 82, 4547-4558, 1977.
Hirt, A.M., and A. U. Gehring, Thermal alteration of the magnetic
mineralogy in ferruginousrocks, J. Geophys. Res., 96, 9947-9953,
1991.
Horowitz, N., G. Hobby, and J. Hubbard, Viking on Mars: The
carbon assimilation experiments, J. Geophys. Res., 82, 46594662, 1977.
Hubbard, J. S., Laboratory simulations of the pyrolitic release
experiments: An interim report, J. Mol. Evol., 14, 211-221, 1979.
Huguenin, R. L., Chemical weathering and the Viking biology
experiments on Mars, J. Geophys. Res., 87, 10,069-10,082, 1982.
Jakobsson, S. P., and J. G. Moore, Hydrothermal minerals and
alteration rates at Surtsey volcano, Iceland, Geol. Soc. Am. Bull.,
97, 648-659, 1986.
Klein, H. P., The Viking mission and the search for life on Mars,
Rev. Geophys., 17, 1655-1662, 1979.
Laul, J. C., M. R. Smith, H. Wanke, E. Jagoutz, G. Dreibus, H.
Palme, B. Spettel, A. Burghele, M. E. Lipschutz, and R. M.
Verkouteren, Chemical systematics of the Shergotty metrorite
and the composition of its parent body (Mars), Geochim. Cosmochim. Acta., 50, 909-926, 1986.
Lebel, P., Inbetriebnahme und Verwendung einer Curiewaage,
Diploma thesis, 73 pp., ETH-Zurich, 1985.
Levin, G. V., and P. A. Straat, Labeled release--An experiment in
radiochemistry, Origins Life, 7, 293-311, 1976.
Levin, G. V., and P. A. Straat, Recent results from the Viking
labeled release experiments on Mars, J. Geophys. Res., 82,
4663-4667, 1977.
Lewis, D. G., and V. C. Farmer, Infrared absorption of surface
hydroxyl groups and lattice vibrations in lepidocrocite (¾FeOOH) and boehmite (¾-AIOOH), Clay Miner., 21, 93-100,
1986.
Lindsay, W. L., Chemical Equilibria in Soils, Wiley Interscience,
New York, 1979.
Lowrie, W., Identification of ferromagnetic minerals in a rock by
coercivity and unblocking temperature properties, Geophys. Res.
Lett., 17, 159-162, 1990.
Lowrie, W., and T. Heller, Magnetic properties of marine limestones, Rev. Geophys., 20, 171-192, 1982.
Mackenzie, R. C., Oxides and hydroxides of higher-valency elements, in Differential Thermal Analysis, vol. 1, Fundamental
Aspects, edited by R. C. Mackenzie, pp. 272-303, Academic, San
Diego, Calif., 1970.
McCord, T. B., R. N. Clark, and R. B. Singer, Mars: Near-infrared
reflectance spectra of surface regions and compositionalimplications, J. Geophys. Res., 87, 3021-3032, 1982a.
McCord, T. B., R. B. Singer, B. R. Hawke, J. B. Adams, D. L.
Evans, J. W. Head, P. J. Mouginis-Mark, C. M. Pieters, R. L.
Huguenin, and S. H. Zisk, Mars: Definition and characterization
of global surfaceunits with emphasison composition,J. Geophys.
Res., 87, 10,129-10,148, 1982b.
McKeague, J. A., and J. H. Day, Dithionite and oxalate extractable
Fe and AI as aids in differentiating various classesof soils, Can. J.
Soil $ci., 46, 13-22, 1966.
Mehra, O. P., and M. L. Jackson, Iron oxide removal from soils and
clays by a dithionite-citrate system buffered with sodium bicarbonate, Clays Clay Miner., 7, 317-327, 1960.
Moore, H. J., R. E. Hutton, R. F. Scott, C. R. Spitzer, and R. W.
Shorthill, Surface materials of the Viking landing sites, J. Geophys. Res., 82, 4497-4523, 1977.
Morris, R. V., What ferric oxide/oxyhydroxide phases are present
on Mars?, Mars Sample Return Science Workshop, LPI Tech.
Rep. 88-07, pp. 126-127, Lunar and Planet. Inst., Houston,
Texas, 1988.
Morris, R. V., and H. V. Lauer, Matrix effects for reflectivity
spectra of dispersed nanophase (superparamagnetic) hematite
with application to martian spectral data, J. Geophys. Res., 95,
5101-5109, 1990.
Morris, R. V., H. V. Lauer, Jr., J. L. Gooding, and W. W. Mendell,
Spectral evidence and implications for the occurrence of aluminous iron oxides on Mars (abstract), Lunar Planet. $ci., XIV,
526-527, 1983.
Morris, R. V., D. G. Agresti, H. V. Lauer, Jr., J. A. Newcomb,
T. D. Shelfer, and A. V. Murali, Evidence for pigmentary
hematite on Mars based on optical, magnetic, and Mossbauer
studiesof superparamagnetic(nanocrystalline) hematite, J. Geophys Res., 94, 2760-2778, 1989.
Morris, R. V., J. L. Gooding, H. V. Lauer, Jr., and R. B. Singer,
Origins of the Mars-like spectral and magnetic properties of a
Hawaiian palagonitic soil, J. Geophys. Res., 95, 14,427-14,434,
1990.
Oades, J. M., Interactions of polycations of aluminum and iron with
clays, Clays Clay Miner., 32, 44-57, 1984.
Orenberg, J., and J. Handy, Reflectance spectroscopyof palagonite
and iron-rich montmorillonite clay mixtures: Implications for the
surface composition of Mars, Icarus, 96, 214-225, 1992.
Oyama, V. I., B. J. Berdahl, F. Woeller, and M. Lehwalt, The
chemical activities of the Viking biology experiments and the
argumentsfor the presence of superoxides, peroxides, ,/-Fe203
and carbon suboxide polymer in the martian soil, Life $ci. Space
Res., 16, 3-8, 1978.
Plescia, P., Recent flood lavas in the Elysium region of Mars,
Icarus, 88, 465-490, 1990.
Plumb, R. C., R. Tantayanon, M. Libby, and W. W. Xu, Chemical
model for Viking biology experiments: Implications for the composition of the martian regolith, Nature, 338, 633-635, 1989.
Posey-Dowty, J., B. Moskowitz, D. Crerar, R. Hargraves, L.
Tanenbaum, and E. Dowty, Iron oxide and hydroxide precipitation from ferrous
solutions
and its relevance
to martian
surface
mineralogy, Icarus, 66, 105-116, 1986.
Ramanaidou, E., and Y. Noack, Palgonites of the Red Sea; A new
occurrence of hydroxysulfate, Miner. Mag., 51, 139-143, 1987.
Ratner-Zohar, Y., A. Banin, and Y. Chen, Oven drying as a
pretreatment for surface area determinations of soils and clays,
Soil $ci. $oc. Am. J., 7, 1056-1057, 1983.
Ross, G. J., and C. Wang, Lepidocrocite in a calcareous, welldrained soil, Clays Clay Miner., 30, 394-396, 1982.
Roush, T. L., D. L. Blaney, and R. B. Singer, The surface
compositionof Mars as inferred from spectroscopicobservations,
in Remote Geochemical Analysis: Elemental and Mineralogical
Composition, edited by C. M. Pieters and P. A. J. Englert, pp.
367-393, Cambridge University Press, New York, 1993.
Schwab, A. P., and W. L. Lindsay, Effect of redox on the solubility
and availability of iron, Soil $ci. $oc. Am. J., 47, 201-205, 1983.
Schwertmann, U., Some properties of soil and synthetic iron
oxides, in Iron in Soil and Clay Minerals, NATO Adv. Stud. Inst.
$er., edited by J. W. Stucki, B. A. Goodman, and U. Schwertmann, pp. 203-250, D. Reidel, Norwell, Mass., 1987a.
Schwertmann, U., Occurrence and formation of iron oxide in
various pedoenvironments, in Iron in Soil and Clay Minerals,
NATO Adv. Stud. Inst. $er., edited by J. W. Stucki, B. A.
Goodman, and U. Schwertmann, pp. 267-308, D. Reidel, Norwell, Mass., 1987b.
Schwertmann, U., and R. M. Taylor, Natural and synthetic poorly
crystallized lepidocrocite, Clay Miner., 14, 285-293, 1979.
Schwertmann, U., and E. Wolska, The influence of aluminum on
iron oxides, XV, Al-for-Fe substitution in synthetic lepidocrocite,
Clays Clay Miner., 38, 209-212, 1990.
Schwertmann, U., P. Cambier, and E. Murad, Properties of goet-
BANIN ET AL..' NANOPHASEIRON MINERALS IN MARS SOIL
hites of varying crystallinity, Clays Clay Miner., 33, 369-378,
1985.
Sherman,
D. M., Electronicstructures
of Fe3+ coordination
sitesin
iron oxides: Applications to spectra, bonding, and magnetism,
Phys. Chem. Miner., 12, 161-165, 1985.
Sherman, D. M., R. G. Burns, and V. M. Burns, Spectral characteristics of the iron oxides with application to the martian bright
region mineralogy, J. Geophys. Res., 87, 10,169-10,180, 1982.
Sidhu, P.S.,
Transformation of trace element-substituted
maghemite to hematite, Clays Clay Miner., 36, 31-38, 1988.
Singer, R. B., Spectral evidence for the mineralogy of high-albedo
soils and dust on Mars, J. Geophys. Res., 87, 10,159-10,168, 1982.
Singer, R. B., Spectroscopic observations of Mars, Adv. Space
Res., 5, 59-68, 1985.
Singer, R. B., T. B. McCord, R. N. Clark, J. B. Adams, and R. L.
Huguenin, Mars surface composition from reflectance spectroscopy: A summary, J. Geophys. Res., 84, 8415-8426, 1979.
Singer, R. B., J. S. Miller, K. W. Wells, and E. S. Bus, Visible and
near-IR spectral imaging of Mars during the 1988 opposition
(abstract), Lunar Planet. Sci., XXI, 1164-1165, 1990.
Tamaura, Y., T. Yoshida, and T. Katsura, The synthesis of green
rust II (Fe•III•-ve2II,) and its spontaneous transformation into
Fe30 4, Bull. Chern. Soc. Jpn., 57, 2411-2416, 1984a.
Tamaura, Y., M. Saturno, K. Yamada, and T. Katsura, The
transformation 3,-FeO(OH) to Fe30 4 and green rust II in an
aqueous solution, Bull. Chem. Soc. Jpn., 57, 2417-2421, 1984b.
Tarzi, J. G., and R. Protz, The occurrence of lepidocrocite in two
well-drained Ontario soils, Clays Clay Miner., 26, 448-451, 1978.
Taylor, R. M., Formation and properties of Fe(II)Fe(III) hydroxycarbonate and its possible significance in soil formation, Clay
Miner., 15,369-382, 1980.
Taylor, R. M., Influence of chloride on the formation of iron oxides
from Fe(II) chloride, I, Effect of [C1]/[Fe] on the formation of
magnetite, Clays Clay Miner., 32, 167-174, 1984a.
Taylor, R. M., Influence of chloride on the formation of iron oxides
from Fe(II) chloride, II, Effect of [C1] on the formation of
lepidocrocite and its crystallinity, Clays Clay Miner., 32,175-180,
Taylor, R. M., and R. M. McKenzie, The influence of aluminum on
iron oxides, VI, The formation of Fe(II)-AI(III) hydroxy,-chlorides, -sulfates, and -carbonates as new members of the pyroaurite group and their significancein soils, Clay Miner., 26, 179-187,
1980.
Taylor, R. M., and U. Schwertmann, Maghemite in soils and its
origin, II, Maghemite synthesesat ambient temperature and p H 7,
Clay Miner., 10, 299-310, 1974.
Taylor, R. M., and U. Schwertmann, The influence of AI on iron
oxides, I, The influence of AI on Fe oxide formation from the
Fe(II) system, Clays Clay Miner., 26, 373-381, 1978.
Thomassin, J. H., and J. C. Touray, Hydrotalcite, a transitory
hydroxycarbonate rapidly developed during the basaltic glassseawater interaction, Bull. Mineral., 105,312-319, 1982.
Toulmin, P., III, A. K. Baird, B. C. Clark, K. Keil, H. J. Rose, Jr.,
R. P. Christian, P. H. Evans, and W. C. Kelliher, Geochemical
and mineralogical interpretation of the Viking inorganic chemical
results, J. Geophys. Res., 82, 4625-4634, 1977.
van Olphen, H., and J. J. Fripiat (Eds.), Data Handbook for Clays
and Other Nonmetallic Materials, p. 346, Pergamon, New York,
1979.
Vins, J., J. Subrt, V. Zapletal, and F. Hanousek, Preparation and
properties of green rust type substances, Colloq. Czech. Chem.
Commun., 52, 93-102, 1987.
Zolotov, M. Yu, Yu. I. Sidorov, V. P. Volkov, M. V. Borisov, and
I. L. Khodakovsky, Mineral composition of martian regolith:
Thermodynamic assessment (abstract), Lunar Planet. Sci., XIV,
883-884, 1983.
A. Banin, T. Ben-Shlomo, and L. Margulies, Department of Soil
and Water Sciences, The Hebrew University, P.O. Box 12, Rehovot, 76100 Israel.
D. F. Blake and R. L. Mancinelli, Space Science Division, NASA
Ames Research Center, Moffett Field, CA 94035-1000.
A. U. Gehring, Department of Soil Science, University of California, Berkeley, CA 94720.
1984b.
Taylor, R. M., Non-silicate oxides and hydroxides, in Chemistry of
Clays and Clay Minerals, edited by A. C. D. Newman, Mineralogical Society, London, 1987.
20,853
(Received August 28, 1992;
revised August 24, 1993;
accepted August 31, 1993.)