The nanophase iron mineral(s) in Mars soil

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.)