Alteration Assemblages in the Nakhlites: Variation with Depth on Mars

Meteoritics & Planetary Science 45, Nr 12, 1847–1867 (2011) doi: 10.1111/j.1945-5100.2010.01123.x

Alteration assemblages in the nakhlites: Variation with depth on Mars

H. G. CHANGELA1 and J. C. BRIDGES1,2*

1Space Research Centre, Department of Physics & Astronomy, University of Leicester, Leicester LE1 7RH, UK 2Scientific Associate, Department of Mineralogy, Natural History Museum, Cromwell Road, London SW7 5BD, UK *Corresponding author. E-mail: [email protected] (Received 12 October 2009; revision accepted 28 August 2010)

Abstract–Secondary mineral assemblages in the nakhlite meteorites, Lafayette, Governador Valadares (GV), Nakhla, Yamato (Y)-000593 ⁄ Y-000749 have been studied using scanning electron microscopy, transmission electron microscopy, and electron probe micro analysis. The different nakhlites have distinctive secondary assemblages in their olivine grains and mesostases, showing compositional fractionation correlated with their relative depths below the Martian surface. Fracture-filled veins in Lafayette at the bottom of the pile consist of a siderite-phyllosilicate-Fe oxide-hydrated silicate gel assemblage. Corresponding veins in Nakhla and GV further up the pile are predominantly a siderite-gel assemblage, with additional evaporites including gypsum. Y-000593 ⁄ Y-000749 veins are dominated by gel. The gel’s Mg ⁄ (Mg + Fe) ratio decreases from Lafayette (0.37) to GV (0.32), Nakhla (0.24), and Y-000593 (0.15). We suggest that hydrothermal fluid flowed up this depth profile, initiated by melting of buried H2O–CO2 ice. Our results show a complex mix of Fe-rich phyllosilicate within the veins and mesostasis of Lafayette with d-spacings of 0.7–1.1 nm suggesting a mixture of smectite and serpentine. The phyllosilicate formed at close to neutral pH, £150 C. We also suggest that water rock ratios (W ⁄ R) of 1–10 occurred in Lafayette with smaller values for the other nakhlites. This is reflected in the volume of alteration minerals: 10% of olivine in Lafayette to 3% in Nakhla. Textural evidence of rapid cooling, together with the W ⁄ R and likely fluid velocities, suggest that the secondary assemblages formed quickly, e.g., within months. A model is proposed in which the secondary assemblages formed in an impact-induced hydrothermal system terminated by precipitation of the gel and evaporation of soluble salts.

INTRODUCTION These are present in fractures not only within olivine but also within the mesostasis (Bridges and Grady The nakhlites are clinopyroxenites thought to have 2000). The preterrestrial, Martian origin of the veins has crystallized as cumulates within a thick basic-ultrabasic been established because Nakhla and Lafayette veins lava flow or shallow (100 m) intrusion on Mars are seen to be truncated by the fusion crust (Gooding 1.3 Ga (e.g., Treiman et al. 1993; Lentz et al. 1999; et al. 1991; Treiman et al. 1993). The secondary mineral Bridges and Warren 2006). Fe-rich olivine grains make assemblages vary in abundance and composition up approximately 10% volume of all but one of them between the nakhlites but include siderite, phyllosilicate, (Miller Range [MIL] 03346, which has no cumulus and salts. They have been dated by K-Ar within the olivine). Within the eight known nakhlites (Nakhla; time range £670 Ma (Swindle et al. 2000) and so this Governador Valadares [GV]; Lafayette; Northwest constrains the timing of the hydrothermal alteration to Africa [NWA] 998; NWA 817; NWA 5790, MIL 03346, a relatively recent time in Mars’ history. and the paired Yamato [Y]-000593 ⁄ Y-000749 ⁄ Y- The secondary mineral assemblage was derived 000802), there are secondary mineral assemblages (e.g., from a brine percolating through the nakhlite parent Ashworth and Hutchinson 1975; Gooding et al. 1991; body (Bridges and Grady 2000; Bridges et al. 2001). Bridges et al. 2001; Gillet et al. 2002; Treiman 2005). This is, for instance, consistent with the varying Ca

1847 The Meteoritical Society, 2011. 1848 H. G. Changela and J. C. Bridges content of the siderite between Lafayette and the other All five nakhlite samples were initially analysed with nakhlites (Bridges and Grady 2000). They formed at scanning electron microscopy (SEM). The Phillips XL30 low temperature (£150 C) and rapidly, to form ESEM and FEI Sirion Field Emission Gun (FEG)-SEM metastable mineral compositions prior to final of UL’s Advanced Microscopy Centre (AMC) were evaporation of the fluid. Some secondary minerals used for this. Alteration products within the olivine within the nakhlites may however, be terrestrial, e.g., were imaged in more detail and further analyzed by sulphates within the Antarctic finds, Y-000593 ⁄ energy dispersive X-ray spectroscopy (EDX). The FEG- Y-000749 (Gooding 1986; Kuebler et al. 2004). SEM and ESEM are equipped with a PGT EDX system In this article, we report the results of a transmission and Oxford Instruments INCA spectrometer. EDX electron microscopy (TEM) and electron microprobe spectra and X-ray maps were made at 15 and 20 kV study of Nakhla, Lafayette, Y-000593, Y-000749, and accelerating voltage, respectively. Montages of five GV to document in more detail how the mineralogy of olivine grains from each of the nakhlites were produced the secondary mineral assemblages within olivine and loaded in Image Pro Plus software to calculate fractures and mesostasis varies between the nakhlites. volume fractions of the secondary assemblages within From this, we aim to further constrain the nature and each of the thin sections. origin of the hydrothermal alteration within the nakhlite Electron probe micro analysis (EPMA) was also parent rocks on Mars. In particular, we have studied the performed on the samples. A Cameca SX100 was used structure and varying composition of nakhlite in the Department of Mineralogy, NHM. The phyllosilicate and gel and show that the gel composition microprobe was operated at 20 kV accelerating voltage varies between the nakhlites. We have also attempted to with a beam current of 20 nA, using appropriate distinguish between a terrestrial overprint and Martian mineral standards to check calibration. Alteration secondary minerals in the Antarctic find Y-000749. products within the olivine veins were analysed in Mikouchi et al. (2003a, 2003b, 2006) suggested the phases with width approximately 5 lm or more in an probable relative original depths of the nakhlites on attempt to constrain interaction volumes within Mars by comparing their mineralogy and groundmass individual phases. textures. Bridges and Warren (2006) added to this After SEM-EDX and EPMA based characteriza- model using the composition of secondary assemblages. tion, cross sections of selected veins were prepared for In these models, MIL 03346 is nearest the surface, the TEM. A dual focussed ion beam (FIB-SEM) technique Yamato meteorites close to the top surface, followed by was used for TEM sample preparation, enabling in situ Nakhla, then GV with Lafayette at the bottom. sectioning of specific micron-sized areas of interest. An Treiman et al. (1993) suggested that in Lafayette the FEI Quanta 200 3D with a Ga+ ion beam was used at water to rock ratio and the temperatures could have UL; milling an approximately 70–110 nm thick wafer been higher compared to the rest of the nakhlites. We containing a section of the secondary assemblage(s). For consider our new results in the light of likely relative precise milling of the sections, the automated TEM depths of the nakhlites and position within a runscript was used with the Ga+ beam at 30 kV to hydrothermal system, perhaps associated with an impact generate a 0.1–0.2 lm thick wafer. The runscript steps event. We aim to provide a model to help explain the down milling currents from 5 nA to <100 pA. Vein formation of some recently identified phyllosilicate and regions chosen for extraction frequently contain some carbonate on the Martian surface (e.g., Ehlmann et al. void space between the olivine and veining assemblages. 2008; Mustard et al. 2008). Electron transparent wafers of these regions are therefore very fragile. To maintain their stability, the SAMPLES AND METHODS produced sections were imaged with snap shots at sequential stages of the milling sequence and the Polished thin sections of Y-000749 and Y-000593 runscript was aborted before the final milling currents have been provided by the National Institute of Polar were employed. Some Y-000593 ⁄ Y-000749 wafers were Research. A thin section of Nakhla (BM 1911, 369) and sectioned at 20 kV in an attempt to help preserve the polished resin blocks of GV (BM 1975, M16) and secondary features with lower voltages and currents. Lafayette (BM 1958, 775) were prepared at the Natural Prior to application of the runscript, the selected areas History Museum (NHM). Resin blocks of GV and in this study were capped with a 30 · 5 lm Pt layer of Lafayette were polished in oil in an attempt to initially approximately 150 nm thickness using the electron preserve soluble phases (Bridges and Grady 2000). A beam. An additional 2 lm thick cap was placed on top nontronite clay standard (Urgeirica Beira Alta, of this during application of the wizard script using the Portugal-MB 1972, 142) was also provided by the Ga+ beam. Both layers were deposited by the Pt NHM. hydrocarbon in the gas injection system (GIS). These Alteration assemblages in the nakhlites 1849

+ capping layers protected the sample area from the Ga (a) ions during the milling and deposition processes. Wafers of length 15 and 4.5 lm width were produced. They were extracted and attached to an Omniprobe copper grid. Welding of the microneedle to the wafer and the wafer to the copper grid was performed with the FIB- SEM carbon GIS. Final ‘‘polishing’’ currents were employed manually after attachment to the grid for optimal electron transparency. The NHM nontronite standard was prepared by taking flakes scraped onto a carbon sticky pad, coated with 15 nm carbon and then inserted directly into the FIB-SEM for extraction and subsequent attachment to an Omniprobe grid. (b) A Jeol 1200 TEM with a PGT EDS system was used for analyses of the extracted veins and standards. Its LaB6 source was operated at 200 kV and 109–111 lA emission current. Bright field (BF) imaging, high resolution TEM (HRTEM), TEM-EDS, and selected area electron diffraction (SAED) were performed on the TEM samples. d-Spacing calibration was performed with an Agar graphitized standard with 0.23 nm lattice planes. Due to contortion and irregularities of fringes, HRTEM measurements were made directly from the HRTEM BF images, using a histogram intensity profile tool with the Gatan Digital Micrograph Camera Fig. 1. a) Average maximum width of 25 veins in each of the software. Errors based on the Agar standard were all studied nakhlite thin sections. b) Total volume fraction of <5% and we take this as a conservative upper limit for secondary minerals (i.e., veining volume) from the sum of five olivine grains in each of the nakhlites. our analyses. The d-spacing measurements were also calibrated with an evaporated Al-oxide diffraction standard at desired camera lengths. Nanobeam broadly similar across the nakhlites, with the Yamato diffraction was also generated on phases too small for meteorites averaging 4 lm, and Nakhla and GV SAED (less than approximately 50 nm). Qualitative averaging 5–6 lm. However, Nakhla and GV do EDX was typically performed for 200 s using a 25 nm contain a few large veins (>15 lm wide). The Yamato spot size. Some spectra were acquired for longer meteorites also contain a higher total volume of veins times on smaller phases, with spot sizes such as 15 and (4% of olivine) compared to Nakhla and GV (3%), 10 nm. and they have a larger volume of the gel compared to Nakhla and GV (Fig. 1b). RESULTS Vein Mineralogy Vein Mineralogy and Morphology Lafayette. Ca-Mn-Mg siderite is present at the outer margins of sawtooth-shaped, fracture-controlled All of the nakhlites studied—Lafayette, Nakhla, veins in Lafayette (Figs. 2 and 3). This siderite is GV, Y-000593, Y-000749—have fracture-filling veins in partially corroded and replaced by phyllosilicate, which the olivine which consist largely of amorphous hydrated in turn has an Fe-oxide phase (possibly ferrihydrite, see silicate—a ‘‘gel.’’ Phyllosilicate is also present in the bottom of this section) at its inner edge (Fig. 2). We veins of Lafayette and siderite is found in varying have only found phyllosilicate in Lafayette. One to two amounts in Nakhla, GV, and Lafayette. lm diameter rosettes of the same phyllosilicate (seen to be spherical in FIB sections and subsequent TEM Vein Sizes and Densities images) are also present within the surrounding The fracture fillings are limited to within the Ca-siderite. The gel occupies the centre of the Lafayette boundaries of the olivine grains and the average vein veins. Thus, the crystallization sequence was olivine sizes are shown in Fig. 1a. In Lafayette, veins can be up fracturing, siderite crystallization, siderite corrosion, and to 50 lm wide and occupy 9.5% of the sampled olivines the formation of phyllosilicate as zones within the veins (Fig. 1). Apart from Lafayette, the vein widths are and within spherical rosettes. This crystallization seems 1850 H. G. Changela and J. C. Bridges

Fig. 2. SEM-EDX mapping of siderite-phyllosilicate-gel and ‘‘rosette’’ assemblage in Lafayette vein. Fe is concentrated in the zone between the gel and phyllosilicate, and also in the core of the rosettes. Si is concentrated in the gel. Al is richer in the phyllosilicate regions and Ca is concentrated within the carbonate. Textures show corroded siderite being replaced by the phyllosilicate. Scale bar is 10 lm. Fr, Fe oxide (ferrihydrite?); Phyl, phyllosilicate; Ol, olivine; Sd, siderite. Top left images backscattered electrons. to have been terminated by the formation of the of Na are also found within all of the siderite grains. Fe-oxide phase. Lastly, silicate gel was precipitated in BF TEM images show the phyllosilicate textures to vary the centre of the veins. The phyllosilicate has a sheet from coarse 100–200 nm width grains to gradually finer texture with the mineral grains arranged perpendicular grained £5 nm and then the amorphous gel in the centre to the fracture sides. This may be a result of the of the veins (Fig. 3c). The lattice fringe spacings of the direction of mineral growth from the vein walls phyllosilicate vary for different regions of the sample inwards. The BSE image in Fig. 2 shows that the (Tables 1 and 2). We found 0.7–0.9 nm fringes in rosettes of radial phyllosilicate grains have Fe-oxide rich the veining phyllosilicate (Fig. 3d). The veining central regions like the inner margins of the vein phyllosilicate from which EPMA analyses were made phyllosilicate. The contact between the Lafayette (see the Gel and Phyllosilicate Compositions section) phyllosilicate and the gel is a complex intermixed zone contained 0.9 nm fringes. The phyllosilicate in the of phyllosilicate and fine and ultra fine-grained rosettes also have 0.9 nm basal fringes that are Fe-oxides from <5 to 10 nm size (Fig. 3a). In the ‘‘sandwiched’’ between minor 0.7 nm fringes (Fig. 4a). carbonate regions surrounding the gel core and Another rosette was also extracted and found to have phyllosilicate, TEM shows precipitates of Mn-rich lattice fringes from 0.9 to 1.1 nm (Fig. 4b). Regular siderite—identified by qualitative EDS—that are from rings are also observed within the coarse grained 50 to 100 nm in diameter (Fig. 3b) which are enclosed areas by both SAED and nanobeam diffraction, within domains of more Ca-rich siderite. Trace amounts indicating multiple crystallographic orientations within Alteration assemblages in the nakhlites 1851

Fig. 3. Lafayette. a) BSE image of extracted region from Lafayette vein. b) Bright field image of Ca-Mn-siderite region surrounding the phyllosilicate and gel (i.e., region of intense Ca Ka from X-ray map in Fig. 2). Grains of Fe-Mn rich siderite show high scattering (they are dark grains) indicating a different crystallographic orientation to the Ca-siderite. c) Phyllosilicate– gel transition in (a). Transition zones in Lafayette veins start from a coarse phyllosilicate (4), becoming fine grained (3), then finer grained (2) and leading to the amorphous silicate gel (1) in the centre of the veins. d) HRTEM image of veining phyllosilicate from which EMPA analyses were made, showing the presence of 0.9 nm lattice fringes that we suggest to be Fe-smectite. Phyl, phyllosilicate; Ol, olivine; Sd, siderite (Fe-Mn rich); Ca-Sd, calcium-rich siderite.

Table 1. Lafayette phyllosilicate d-spacings from Table 2. d-Spacing measurements of phyllosilicate diffraction patterns. basal fringes. 1st 2nd 3rd Textural setting d-Spacing ˚ ˚ ˚ Textural setting ring (A) ring (A) ring (A) Veins in olivine 0.79 (10), 0.72 (7), 0.91 (5) Phyllosilicate in olivine 2.60 1.55 1.31 Rosettes 0.85 (9), 0.85 (9), 0.89 (9), 0.9 (6), Phyllosilicate in mesostasis 2.47 1.57 1.36 0.93 (7), 0.93 (5), 1.06 (6), 1.1 (6) Nontronitea 2.62 1.62 1.40 Mesostasis 0.68 (7), 0.67 (7), 0.68 (5), 1.1 (4) Note: Errors in diffraction measurements are £5%. Note: d-Spacings calculated from average number of fringes, aNontronite standard BM 1972, 142. number in brackets. Errors in diffraction measurements are £5%. the phyllosilicate. Comparison with the nontronite the relatively coarse (100–200 nm width) grains. Our standard prepared in the same way for TEM yielded nontronite standard d-spacing measurements (Table 2) fairly similar three-ring polycrystalline SAED patterns are generally larger than our Lafayette phyllosilicate (Figs. 5a and 5b). The coexistence of the polycrystalline ring patterns. Nanobeam diffraction of a fine-grained patterns may reflect a natural variation in grain size iron oxide shows d-spacings of symmetrical diffraction within the samples, including nanoscale in addition to spots at 1.56 ± 0.05 A˚ with a hexagonal symmetry 1852 H. G. Changela and J. C. Bridges

Fig. 4. HRTEM bright image of the phyllosilicate found in the rosettes. a) The main fringes at approximately 0.9 nm spacing are sandwiched by 0.7 nm fringes. b) Lattice fringes from another rosette containing similar fringes but also including 1.1 nm spacing.

(a) (b)

Ol Ol

Gel Sd Sd

Ol 1 µm 0.1 µm Gel

(c) (d) α

b* α a*

α β α α α

Fig. 5. Lafayette. a) SAED of Lafayette phyllosilicate. ˚ ˚ ˚ d1 = 2.60 A, d2 = 1.55 A, d3 = 0.131 A. b) SAED of Fig. 6. a) Low magnification bright field TEM of vein section ˚ ˚ ˚ nontronite standard. d1 = 2.62 A, d2 = 1.62 A, d3 = 1.40 A. from Nakhla. b) Higher magnification of siderite zone with c) Five nanometre nanobeam diffraction of weakly crystalline grain sizes typically 5–10 nm. Large rarer grains (50 nm) can dark grains (Fe oxide, ferrihydrite?) in the phyllosilicate. also be found. c) Characteristic TEM-EDX signature from Sd ˚ d1 = 1.56 A. Errors £ 5%. d) HRTEM bright field of zone in (b). d) SAED diffraction of a large grain in the Sd nontronite (smectite) standard. Average lattice spacing is zone. The zone axis shown by a* and b*is[)1 0 0]. Sd, 0.93 nm. siderite; Ol, olivine.

(Fig. 5c). The symmetry suggests that this nanoscale however, span across the entire depth (4 lm) of the FIB mineral might be ferrihydrite (Michel et al. 2007). sections (Fig. 6a). BF TEM micrographs show that the Nakhla and GV. In Nakhla and GV many veins siderite-bearing zones consist of grains less than contain crystalline siderite along their margins with the approximately 50 nm in size (Fig. 6b). These zones have silicate gel in the centres of the veins. No phyllosilicates complex fine-grained textures but TEM-EDX of many have been identified. However, the siderite-bearing zones grains across entire zones yield qualitative identification have an irregular texture on the submicron scale of Mg-Ca-Mn siderite. Siderite compositions can, compared to the more regular vein textures seen in however, vary within Nakhla, with qualitative analyses Lafayette. The crystalline siderite regions are commonly showing some zones to be rich in Ca and Mn, while found along one of the margins and do not follow the others are Mg-rich with trace amounts of Ca and Mn, as whole length of the vein across the olivine. They do, shown in Fig. 6c. The textures are finer grained than the Alteration assemblages in the nakhlites 1853

Y-000593 and Y-000749. Y-000593 and Y-000749 alteration veins within olivine are dominated by silicate gel (Fig. 9). This amorphous signature is characteristic of the hydrated silicate found across all of the nakhlites, with the exception of the coarse phyllosilicate in Lafayette. In Y-000593 and Y-000749, traces of S are concentrated along the edges of silicate gel veins, sometimes with minor Cl and K. These are regions where it seems that the olivine has been altered to a partially crystalline silicate. This partially broken down olivine is also present as patches within the gel (Fig. 9a). Recent studies have suggested this type of phase to 2+ 3+ be laihunite Fe Fe 2(SiO4)2—a nonstoichiometric olivine-type mineral (Noguchi et al. 2009). Laihunite is an oxidation product of olivine that has a distorted olivine structure with Fe3+ charge balanced by vacancies, e.g., Kondoh et al. (1985); Bandfield et al. (1990). In Y-000593, more irregular veins have been found, lacking any zonation but they include amorphous islands distinguished by having larger amounts of S and K (Fig. 9c). No carbonates have yet been identified through TEM in the Yamato meteorites. Both Y-000593 and Y-000749 seem therefore to have a much lower proportion of crystalline secondary Martian minerals than the other nakhlites in this study. Carbonate EDX signatures have been tentatively identified in previous studies of other thin sections (e.g., Spencer et al. 2008) but the mineralogical evidence for this has not yet Fig. 7. Veins identified in Nakhla and GV. a) BSE image of large gel vein in Nakhla. A zone of gypsum is associated with been established. Iron oxides (grains approximately the gel. b) BSE image of GV vein similar to Nakhla vein but 100 nm in size) are also occasionally found in Y-000593 with an outer zone of siderite. Typical veins in Nakhla also (Fig. 9c). The paired meteorite Y-000749 contains some have the same siderite-gel assemblage. Ol, olivine; Sd, siderite; veins entirely composed of the sulphate jarosite Gp, gypsum. 3+ KFe 3(OH)6(SO4)2 (Fig. 10a). Jarosite veins have been identified which crosscut part of the fusion crust and are, therefore terrestrial in origin (Fig. 10d). Much of siderite in Lafayette, e.g., grains of 10–25 nm diameter. the sulphate veining was found in close proximity to this The Lafayette siderite also differs in having low MgCO3 region of the thin section and crosscutting the gel. compositions and traces of Na. Previous, quantitative Jarosite has not been identified in the less terrestrially EPMA analyses of Lafayette siderite have shown it to altered meteorite pair Y-000593. contain 22–37 mole% CaCO3, 4–35 mole% MnCO3, 0–2 mole% MgCO3 with Nakhla having 0–6 mole% Mesostasis Phyllosilicate CaCO3, 1–40 mole% MnCO3, 2–41 mole% MgCO3 (Bridges and Grady 2000). SAED of the Ca-Mn-siderite Some of the mesostasis of Lafayette has been zone in Nakhla is shown in Fig. 6d. The carbonate zones replaced by phyllosilicate (Fig. 11). It has a more in Nakhla and GV frequently exist only on one side of irregular texture than the vein phyllosilicate where the the veins (Fig. 7b) although zones on both sides of veins sheets have crystallized perpendicular to the fracture in Nakhla have occasionally been identified. walls. Small patches of gel (Fig. 11) were also found Veins in Nakhla have also been found to contain within one region of mesostasis phyllosilicate making up zones of calcium sulphate, identified as gypsum from approximately 10 vol% of it. Qualitative EDX suggests SAED (Fig. 8). Finer grains of gypsum (which co-exist that other regions of mesostasis contain the same with the coarser gypsum) are approximately 5 nm in phyllosilicate. The mesostasis phyllosilicate and gel diameter and have multiple crystallographic orientations, occupies approximately 2% of the polished section that shown by their ring diffraction patterns (Fig. 8b). we studied. As well as this alteration phase, this region, 1854 H. G. Changela and J. C. Bridges

(a) (b)

(c) (d)

α

α

α

β α

Fig. 8. Gypsum vein found in Nakhla (SEM image of other part of vein is shown in Fig. 6a). a) Low magnification bright field image of wafer. b) SAED of gypsum region. Monocrystalline pattern from an assumed large grain is superimposed by a ring pattern due to the majority fine-grained material with multiple orientations. d1 = 0.247 nm, d2 = 0.214 nm, and d3 = 0.135 nm. The d-spacings have been matched to Schofield et al. (1996) to verify gypsum. c) HRTEM bright field in gypsum zone. d) TEM- EDX of gypsum zone. Ol, olivine; Gyp, gypsum. shown in Fig. 11a, contains mainly plagioclase feldspar studies of Nakhla and Lafayette (Gooding et al. 1991; with Fe-Ti-oxides, Ca-Cl phosphate and Fe-sulphide. Treiman et al. 1993). However, the Fe versus Si lines HRTEM shows most of the basal fringes to be plot with different y-intercepts in the order of Lafayette approximately 0.7 nm that is characteristic of a 1:1 phyllosilicate (lowest) Lafayette gel–GV gel–Nakhla gel– phyllosilicate, e.g., serpentine. However, some parts of Y-000593 gel (highest). The normalized Fe versus Si this mesostasis phyllosilicate have fringes of 1.1 nm wt% ratios are broadly similar across the nakhlites with representing a 2:1 structure. HRTEM (e.g., Fig. 11b) the gradient of their slopes being )1.2 with only the shows its lattice fringe spacing to vary from Y-000593 gel (gradient )1.6) being different. Previously approximately 0.7 to 1.1 nm, which is inconsistent with reported terrestrial Fe-rich smectite compositions such a pure smectite identity that is characterized by as saponite and nontronite (e.g., Brigatti 1983) plot basal spacings of approximately 1 nm (Tables 1 and 2). within the same field of the gel and phyllosilicate Representative EDX and average compositions of this analyses. Mg ⁄ (Mg + Fe) atomic ratios decrease in the phyllosilicate and gel are given in Tables 3 and 4. order of assumed decreasing depth (Fig. 12b), ranging from a consistent average of 0.37–0.39 (in the Lafayette Gel and Phyllosilicate Compositions gel veining, mesostasis clay and the clay veining), 0.32 in the gel of GV, 0.24 in Nakhla to 0.15 in Y-000593. The composition of the gel is that of a hydrated The highest alkali elements Na + K concentration are Fe-Mg-Al silicate. Microprobe data of the gel and within GV gel (0.74 wt%), then Nakhla gel (average phyllosilicate in the nakhlite samples are plotted in 0.59 wt%), followed by Lafayette gel (average Fig. 12 and also shown in Tables 3 and 4. For both the 0.34 wt%) with the lowest amount in Y-000593 gel (see gel and phyllosilicate, Fe against Si abundance defines also Fig. 12c). The Lafayette gel, however, is quite distinct lines, with negative slopes, for each nakhlite variable in Na + K content, ranging from 0.13 to (Fig. 12a). This has already been established in previous 0.75 wt%. Thus, there is a tentative indication of some Alteration assemblages in the nakhlites 1855

also analysed by EPMA (Table 3). The ternary diagram (a) (b) α in Fig. 13 shows that Lafayette phyllosilicate plots from the serpentine to saponite and nontronite

α α composition solid solution lines. These lines are defined

α by the atomic ratios of Si, Al: Fe, Mg (8:4) for nontronite 3+ β (CaO0.5,Na)0.3Fe 2(Si,Al)4O10(OH)2 ÆnH2O, Si,Al: +2 Fe,Mg (8:6) for saponite (½Ca,Na)0.33 (Mg,Fe )3 (Si,Al)4O10(OH)2Æ4H2O, and Si,Al:Fe,Mg (3:2) for serpentine. Previous studies have suggested an Fe-rich (c) (d) smectite identity in both Nakhla and Lafayette (e.g., Burns and Martinez 1991; Gooding et al. 1991; Treiman et al. 1993). Treiman et al.’s coarse smectite in Lafayette identified chemically as saponite also matches our composition. The ferric iron-rich smectite— nontronite—has higher Si + Al abundances than the Lafayette phyllosilicate. The closest smectite to the Lafayette phyllosilicate composition is the Fe-rich saponite, griffithite, which has an unusually high abundance of Fe3+ in its octahedral sites (Komadel et al. Fig. 9. Y-000593. a) Low magnification bright field of Y- 2000). However, site allocations of cations based on 22 000593 wafer. Partially broken down olivine (PBO) is oxygens for a smectite structure (Table 4) indicate a associated with the gel. b) TEM-EDX of the gel (dotted line) slight excess of octahedral cations (average 6.2) in the and the Si-poor PBO region. c) Low magnification bright field image of an irregular vein in Y-000593. 1 = Fe-oxide grain, vein phyllosilicate. Due to even higher calculated excesses 2 = ‘‘island’’ of amorphous Fe-Mg-K-Na sulphate ⁄ silicate, in octahedral sites for some phyllosilicate (6.7 based on 3 = amorphous Fe-Mg-K-Al-Cl sulphate ⁄ silicate, 4 = PBO 22 oxygens) and the presence of low basal fringe region as in (a). d) SAED of gel phase showing the measurements, i.e., approximately 0.7 nm, which are amorphous structure of the gel. Ol, olivine. Pt is the protective inconsistent with pure 2:1, smectite phyllosilicate, Table 4 Pt bar deposited during the FIB-SEM extraction process. also presents the stoichiometry based on seven oxygens for a serpentine area within the mesostasis. Thus, our alkali enrichment within the Nakhla and GV gel d-spacing and chemical analyses show the Lafayette vein compared to the gel and phyllosilicate within Lafayette. and mesostasis phyllosilicate has a range of compositions The olivine veining phyllosilicate has a higher alkali consistent with a mixture of serpentine and smectite content than the mesostasis phyllosilicate with the (saponite to serpentine-type) phyllosilicates. Our data so phases averaging Na + K = 0.74 and 0.36 wt% far raise the possibility that there is a higher proportion respectively. Calcium contents are highest in Lafayette of serpentine in the mesostasis phyllosilicate. The (gel average 0.90 wt%), closely followed by Nakhla phyllosilicate identities are considered further in the (average 0.88 wt%) and then Y-000593 (average Phyllosilicate section. 0.22 wt%) suggesting that this element shows a fractionation trend in the gel previously identified by DISCUSSION siderite analyses between Lafayette, Nakhla, GV (Bridges and Grady 2000). Aluminium is most abundant In the following sections, we describe the nakhlite in Lafayette gel and phyllosilicate, followed by Y-000593 parent fluid and the formation conditions of the (Fig. 12d). Nakhla and GV have very low abundances secondary assemblages. The relative depths of these (0.05–0.22 wt%) of Al2O3. The phyllosilicate in nakhlites have already been proposed with the Yamato Lafayette olivine (Table 3) has Al2O3 contents lower by meteorites at the top, followed by Nakhla, then GV and 3–5 wt% than that in the mesostasis. Sulphur is most Lafayette at the bottom (Mikouchi et al. 2003a, 2003b, abundant in the gel of Yamato meteorites, especially in 2006; Bridges and Warren 2006). The nakhlites’ Y-000749, averaging 5.6 wt% SO3 (Table 3) compared petrologies are similar to some terrestrial PreCambrian to approximately 1.3 wt% or less in the other nakhlites ultrabasic lava flows which formed at depths of (Fig. 12f). Due to the suspected terrestrial alteration of approximately 10–100 m (e.g., Lentz et al. 1999). We Y-000749, we have not plotted the other gel oxide have interpreted our results to suggest that the fluid abundances of Y-000749 in Fig. 12. moved from deep (Lafayette) towards the other, Large regions of coarse phyllosilicate in Lafayette— shallower nakhlites. We propose that the fluid was within olivine fractures and in the mesostasis—were derived from a subsurface source such as the melting of 1856 H. G. Changela and J. C. Bridges

(a) (b)

Ol Gel Ja

PBO Ol

10 µm 10 µm

(c) (d) Fusion Ja PBO crust Ja Ja 5 µm Gel Fusion crust Ol

30 µm 100 µm

Fig. 10. BSE images of a variety of veins found in Y-000593 and Y-000749. a) Jarosite vein. Jarosite veins are only found in Y-000749. A milled trench for a TEM section is shown across the vein. b) Typical alteration vein found in Y-000593. Areas of PBO are less well defined than the PBO in the veins of Y-000749. c) Y-000749 vein containing a large PBO zone along its margin which is adjacent to terrestrial jarosite. d) Y-000749 fusion crust with adjacent fracture partially filled by jarosite in Y-000749. Inset shows jarosite veining crosscutting the fusion crust. Ja, jarosite; Ol, olivine; PBO, partially broken down olivine. a buried ice reservoir such as permafrost (Demidov and conditions on early Mars, it maybe also applicable to Gilichinsky 2009) rather than formation at the Martian the mineral assemblages in the Amazonian nakhlites. surface followed by migration of the fluid to deeper They demonstrated how smectite can be precipitated levels. at a lower pCO2 than previously thought (e.g., at £1–10 mbar). Their phase equilibrium calculations for Fluid pH, Temperature, and Redox Conditions the carbonate and smectite are expressed as a function of Mg ⁄ (Mg + Fe) atomic ratio (Fig. 14). Based on Our analyses show a change from the coarse suggested partial pressures in the Amazonian (Manning phyllosilicate and siderite at the margins of the et al. 2006), ranging from 10 to 100 mbar which we Lafayette veins to the amorphous gel in the centre of take to be equivalent to the amount of CO2 in a the veins. This may be because the amorphous gel in suggested ice reservoir, and our average Mg number of the nakhlite veins cooled more rapidly than the coarse 0.4 for the phyllosilicate in Lafayette, the pH of the Lafayette assemblages which formed at an earlier stage nakhlite parent fluid would have been close to in the alteration sequence. Corrosion of the Ca-siderite neutral ⁄ weakly acidic (Fig. 14). Reduction in pCO2 due in Lafayette prior to the Fe-smectite ⁄ serpentine to carbonate precipitation would then have led to the crystallization suggests that during the hydrothermal formation of phyllosilicate. If a higher proportion of event the fluid stopped being in equilibrium with CO2 was stored in the ice from which the hydrothermal ) carbonate, presumably due to exhaustion of HCO3 and fluid was derived, pH conditions would be shifted to a lack of CO2 being dissolved and recharged into the more alkaline values. fluid. Furthermore, the Fe oxide (ferrihydrite?) rims to Temperatures for the formation of Fe-smectite can the phyllosilicate are evidence that the fluid became vary from up to 250 C (Tosca et al. 2008), and would progressively more oxidizing. have been less than or equal to the temperatures Chevrier et al. (2007) described a thermodynamic governing the precipitation of the Ca-siderite, unless an model to explain the stability of smectite-carbonate additional heat source had emerged (Vicenzi and assemblages on Mars. Although their model addressed Heaney 2000). However, the order of phase formation Alteration assemblages in the nakhlites 1857

precipitation of gel that we have described. Treiman et al. (1993) identified the presence of ferri iron oxide phases that could be the same Fe-oxide phases we have identified. Reducing conditions leading to the formation of saponite or a trioctahedral phyllosilicate from a pre-existing nontronite phase (Manceau et al. 2000) or other dioctahedral phyllosilicate seems unlikely in the nakhlites due to our evidence of late stage precipitation of some ferric iron oxides. Fe oxide grains within smectite have also been suggested to crystallize as the weathering of silicate minerals progresses (Sherman et al. 1962) and this is consistent with the Lafayette phyllosilicate assemblage. By our model, subsurface phyllosilicate formed without thermodynamic equilibrium with the hydrosphere. An assumption in the model of Chevrier et al. is thermodynamic equilibrium between the atmosphere and subsequent stable minerals. However, the mineral sequences found in Lafayette (e.g., phyllosilicate after carbonate and amorphous silica with Fe oxide after phyllosilicate oxidation) are very similar to the sequence described by Chevrier et al. (2007). Replacement of carbonate by phyllosilicate in a progressively cooling fluid implies an upper limit approximately 150 C for phyllosilicate formation. To produce this hydrothermal activity, a thermal event must have occurred: The association with sawtooth fractures that have been subsequently filled with secondary phases and Mars’ heavily cratered past could imply that impact was the heat source. Saw-toothed Fig. 11. Phyllosilicate found in the mesostasis of Lafayette. a) fracturing within other nakhlites has been demonstrated BSE image of phyllosilicate region. Small patches of assumed gel are also found (arrowed). b) HRTEM Bright Field of the by previous authors (Treiman et al. 1993; Gillet et al. wafer extracted on the top right of the image in (a). d- 2002). Bridges et al. (2001) also suggested short-lived Spacings (0.67 nm) that we have found are mostly smaller pulses of fluid in the parent rocks of the SNCs and the than those of the smectite in olivine and our nontronite formation of metastable phases over a period of days to standard and suggest the presence of serpentine. Fp, feldspar; months rather than the extended timeframes over Phyl, phyllosilicate; Gel, amorphous silicate gel. thousands of years such as those in long-lived hydrothermal systems. from siderite to clay followed by amorphous gel we have identified does not show any evidence of a second Hydrous Silicate Gel thermal event. In terms of carbonate, this is consistent with crystallization of metastable carbonate occurring We suggest on the basis of our SAED results rapidly at low temperature £150 C (e.g., Baker et al. (Fig. 9d) that the ‘‘gel’’ parts of the veins are amorphous. 2000; Golden et al. 2000). Isotopic measurements of The softness of this phase does raise the possibility of d18O show heavy enrichment in the Nakhla carbonate crystallographic structure having been destroyed during (Saxton et al. 2000) and also imply low temperatures for sample preparation with the FIB. This phase common to siderite deposition (80–170 C). Chevrier et al. also all of the nakhlites studied is clearly distinct from the showed the stability fields for nontronite, Fe3+ and phyllosilicates and Fe-oxides found in Lafayette. Its ferrihydrite as a function of dissolved silica activity composition also crosses smectite and serpentine solid and pH. At neutral pH, silica saturation leads to solution lines. the formation of amorphous silica and Fe oxide The amorphous nature of some material within the (ferrihydrite) stability. This is consistent with the veins has also been suggested by previous authors (e.g., assemblage of Fe-oxides (although we do not have a Ashworth and Hutchison 1975; Vicenzi and Heaney conclusive identification of ferrihydrite) followed by the 1999). Ashworth and Hutchison (1975) preliminary 88H .Cagl n .C Bridges C. J. and Changela G. H. 1858

Table 3. Gel compositions (wt%). Y593 Y749 Nakhla GV Laf. Laf. Laf. meso. Y593 gel av. Y749 gel av. Nakhla gel av. GV gel av. Laf. gel av. meso. gel av. gel (n =9) gel (n = 17) gel (n = 13) gel (n =4) gel (n =8) gel (n = 8) Y593a Y749b Nakhlac Laf.d

Na2O 0.12 0.15 0.14 0.28 0.16 0.24 0.43 0.37 0.29 0.11 0.3 0.3 n.d. 0.13 1.16 0.30 MgO 3.84 3.77 4.12 4.84 6.56 6.20 8.14 8.23 10.80 9.29 8.5 10.8 3.80 7.49 6.82 13.90 Al2O3 0.90 0.72 1.58 1.42 0.15 0.12 0.06 0.07 5.20 3.02 6.7 4.8 0.56 3.49 0.74 4.77 SiO2 47.1 44.8 39.9 37.4 43.5 39.5 43.9 43.5 40.2 41.0 40.7 41.6 51.4 49.1 40.2 49.1 CaO 0.22 0.31 0.46 0.54 0.42 1.24 0.22 0.20 1.00 1.26 1.7 1.1 0.08 0.45 1.14 0.70 K2O 0.16 0.24 1.12 1.55 0.58 0.49 0.68 0.57 0.58 0.30 0.4 0.3 0.52 0.55 0.60 0.41 TiO2 0.06 0.04 0.06 0.04 0.04 0.02 0.03 0.03 0.02 0.06 b.d. b.d. n.d. 0.02 0.20 0.01 Cr2O3 0.05 0.03 0.02 0.03 0.02 b.d. b.d. b.d. b.d. b.d. b.d. b.d. n.d. n.d. n.d. n.d. MnO 0.50 0.61 0.22 0.44 0.46 0.79 0.54 0.52 0.53 0.91 0.5 0.5 0.22 0.15 0.63 0.48 FeO 33.4 37.8 32.7 33.9 33.1 34.5 31.0 31.2 27.7 27.7 26.9 24.2 25.7 28.8 34.1 29.0 SO3 0.70 1.08 3.34 5.63 0.05 0.05 0.15 0.16 b.d. 0.15 0.4 0.09 2.79 n.d. n.d. n.d. Cl 0.03 0.07 0.06 0.08 0.24 0.25 0.94 0.74 0.02 0.13 b.d. b.d. n.d. n.d. n.d. n.d. Total 87.1 89.6 83.7 86.2 85.3 83.4 86.1 85.6 86.3 83.9 86.0 83.6 79.1 90.1 85.6 97.9 Note: b.d. = below detection limits; n.d. = no data. aIddingsite Y-000593; Noguchi et al. (2009). bIddingsite Y-000749; Treiman and Goodrich (2002). cIddingsite Nakhla; Gooding et al. (1991). dIddingsite Lafayette; Treiman et al. (1993). Alteration assemblages in the nakhlites 1859

Table 4. Composition of Lafayette olivine and from a precursor gel similar to smectite composition. mesostasis phyllosilicate. Given that the Lafayette gel is chemically close to the Olivine Olivine Meso. Meso. composition of the Lafayette phyllosilicate (Fig. 13), but (smectite) (smect. av) (serpentine) (phyl. av) has a lower Mg number than the phyllosilicate (see next

Na2O 0.3 0.2 0.1 0.1 section) a final pulse of fluid may have cooled too quickly MgO 11.4 10.5 10.6 10.4 to crystallize into phyllosilicate and instead precipitated Al2O3 4.5 4.1 8.3 7.9 as a metastable gel. Cooling may have been enhanced by SiO2 45.1 42.6 38.1 38.1 the decreasing volumes that were available to be filled CaO 1.13 1.4 1.0 1.1 within the fractures during the last fluid pulse. The K2O 0.6 0.6 0.3 0.3 progressively finer grained textures we have identified at TiO2 b.d. b.d. b.d. b.d. the inner margins of some Lafayette phyllosilicate veins Cr2O3 b.d. b.d. b.d. b.d. are also consistent with this cooling. MnO 0.6 0.7 0.6 0.6 FeO 26.1 27.6 30.0 29.6 Compositional Fractionation of Gel SO3 0.05 0.08 b.d. b.d. Cl 0.02 0.05 b.d. b.d. We see that Fe + Si in the gel increases and Total 89.8 87.8 89.0 88.3 Mg ⁄ (Mg + Fe) decreases up the nakhlite pile. The gel 22 Oxygens 7 Oxygens in the nakhlites from near the surface has a greater Si content than the coarser crystalline secondary silicates Tetrahedral Si 7.10 7.02 2 2.01 at the base of the known nakhlite pile. The variation in Al 0.88 0.82 – – Fe3+ 0.03 0.16 – – Fe + Si abundances between the different nakhlites Sum 8 8 2 2.01 studied here, together with the phyllosilicate to gel 4) Octahedral Fetot 3.41 3.51 1.31 1.31 precipitation order may be the result of SiO4 Mg 2.74 2.59 0.83 0.82 enrichment and Fe fractionation during the lifetime of Mn 0.06 0.09 0.02 0.03 the fluid. The fluid precipitated much of its less soluble Al – – 0.51 0.49 cations (e.g., Ca) in Lafayette, followed by decreasing Sum 6.21 6.19 2.67 2.65 proportions in GV then Nakhla and also Y-000593. Interlayer Ca 0.19 0.24 0.06 0.02 This process could have led to silica enrichment in the K 0.15 0.12 b.d. 0.01 fluid and hence facilitated precipitation of the gel phase. Na 0.05 0.06 b.d. b.d. Neither the Mg number of the nakhlites’ olivines, S 0.007 b.d. b.d. b.d. pyroxenes (Bridges and Warren 2006) nor the Cl 0.007 b.d. b.d. b.d. Sum 0.41 0.42 0.06 0.03 meteorites’ bulk Mg numbers (Meyer 2006) follow the Note: b.d. = below detection limits. Mg ⁄ (Mg + Fe) fractionation trend we have identified in the alteration phases. This suggests that the fluid passing through these fractures brought in the majority study of Nakhla using microtome sections indicated the of its ionic constituents through dissolution of presence of a colloidal gel. Weakly crystalline minerals surrounding rock before encountering the nakhlite with compositions similar to our gel, can form from the parent rocks that we have available for study, although weathering of iron-rich saponites (Kohyama and Sudo the hydrothermal cell as a whole, may have been solely 1975; Romero et al. 1992). The amorphous gel in the within a nakhlite lithology. A similar conclusion was nakhlites has a variable composition but closely reached by Gillet et al. (2002) in their study of the resembles the range of phyllosilicate composition we NWA 817 nakhlite. Evidence that some localized determined. However, our textural evidence of late dissolution and remobilization into the fluid also took stage precipitation of the gel from the hydrothermal place is shown by the mesostasis phyllosilicate of fluid could suggest that the gel is not the result of Lafayette because it is enriched in Al. The low Al phyllosilicate alteration. abundances in the gel of Nakhla and GV imply that the Fiore et al. (1995) demonstrated that gels can also local dissolution of Al-rich mesostasis did not, however, act as an intermediate stage or a precursor for the contribute to the gel composition in the nakhlite olivine formation of spherical phyllosilicate particles. In an veins (Fig. 12e). experimental synthesis of saponite ⁄ nontronite by Grauby et al. (1994), gel precursors were made which were then Phyllosilicate formed into clay via mixing with distilled water and gradual heating (30 days) at 200 C under alkaline Our chemical analyses show the Lafayette conditions (pH 9–10). Treiman and Lindstrom (1997) phyllosilicate lie between the serpentine, saponite, and also proposed that the Lafayette ‘‘iddingsite’’ formed nontronite solution lines (Fig. 13) but with unusually 1860 H. G. Changela and J. C. Bridges

50 (a) 0.5 (b) 40 0.4

30 0.3 Mg# Y000593 Gel Fe (wt%) 20 Y000749 Gel 0.2 Nakhla Gel GV Gel 10 Lafayette Gel 0.1 Lafayette Ol. Phyl. Lafayette Meso. Phyl. 0 0.0 12 14 16 18 20 22 24 26 28 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 Si (wt%) Fe/Si 3.0 1.4 (c) (d) 1.2 2.5 1.0 2.0

0.8 1.5 0.6 1.0 Ca (wt %) Ca (wt Na + K (wt %) Na + K (wt 0.4 0.2 0.5 0.0 0.0 12 14 16 18 20 22 24 26 28 12 14 16 18 20 22 24 26 28 Si (wt %) Si (wt %) 6.0 5.0 (e) (f) 5.0 4.0 4.0 3.0 3.0 S (wt%) Al (wt %) Al (wt 2.0 2.0

1.0 1.0

0.0 0.0 12 14 16 18 20 22 24 26 28 12 14 16 18 20 22 24 26 28 30 Si (wt %) Si (wt%) Fig. 12. EPMA Microprobe analysis of the gel in the nakhlites and Lafayette phyllosilicate. The Lafayette mesostasis gel analysis is SEM-EDX. a) Fe versus Si wt%. Y-000593 gel slope = )1.6, Nakhla gel slope = )1.2, GV gel slope = )1.2, Lafayette gel slope = )1.2, and Lafayette phyllosilicate slope = )1.2. b) Mg number (Mg#) versus Fe ⁄ Si (wt). c) Al versus Si wt%. d) Ca versus Si wt%. e) Volatiles (Na + K) versus Si wt%. f) S versus Si wt%. Compositions in (a–f) were normalized to anhydrous 100%. The difference between the Y-000593 gel slope and that of the other nakhlites suggest that Y-000593 was not on exactly the same fluid fractionation path. high total Fe concentrations compared to terrestrial they do not contain as much Fe as in the clay of samples. Table 4 shows the atomic proportions of the Lafayette. Octahedral site substitutions of Fe2+ and cations and their respective sites based on 22 oxygen Fe3+ in saponite to nontronite solid solutions have been atoms and 8 cations in the tetrahedral sites for the experimentally demonstrated by Grauby et al. (1994). veining phyllosilicate. The deficit of tetrahedral cations In a previous study of Nakhla, Burns and Martinez (Si, Al) indicates that there is some Fe3+ substitution (1991) reported the presence of ferric iron in veining into the tetrahedral sites and Fe3+ with Fe2+ in the ‘‘iddingsite.’’ Based on our study, this would have most octahedral sites. This feature is relatively common in likely been the veins’ amorphous gel. Gooding et al.’s natural trioctahedral saponites (Nimis et al. 2004), (1991) Nakhla ‘‘rust’’ classified as a ferric smectite also where all octahedral sites are filled compared to every underlines the ferric nature of the hydrous silicates in third site being empty in dioctahedral structures such as the nakhlites. nontronite. Fe-rich saponites are also frequently found The 1.1 nm d-spacings we have also identified are in many terrestrial soils (Brigatti et al. 1999). However, consistent with a Fe-smectite identity for some of the Alteration assemblages in the nakhlites 1861

phase based on the 0.9 nm fringes also identified in our nontronite standard. Previous TEM studies of the Lafayette phyllosilicate (Treiman et al. 1993) by ultra microtomy also identified 1 nm d-spacings. However, the low basal spacings 0.7 nm which we have found are also present, particularly in the mesostasis, are not consistent with smectites but rather of kaolins and serpentines. The stoichiometry and d-spacing of the phyllosilicate in the mesostasis of Lafayette is consistent with the presence of a berthierine-type serpentine 2+ 3+ composition (Fe ,Fe ,Mg,Al)2-3(Si,Al)2(OH)4 with vacancies in the octahedral sheet (Table 4). Thus, the phyllosilicate is a fairly complex mixture of Fe-smectite and serpentine. In the light of this complexity, it is possible that future work will identify further phyllosilicates within the nakhlites. The presence of ferric iron within the smectite component (as also suggested by Gooding et al. 1991), Fig. 13. Atomic ratios of the nakhlites’ gel and Lafayette the serpentine and also potentially the gel indicate a phyllosilicate. Iron is total Fe2+ and Fe3+. Clay analysis of Lafayette from Treiman et al. (1993) is also included. Similar more oxidizing environment than would be associated terrestrial saponites are from: Deer et al. (1966); Kodama with a ferroan trioctahedral endmember such as a pure et al. (1998); Brigatti et al. (1999); Komadel et al. (2000). saponite or ferroan serpentine. The oxidizing nature of Terrestrial nontronites are from: Brigatti (1983); Ko¨ ster et al. the Martian near surface is well known. Progressive (1999); Gates et al. (2002). Horizontal lines represent oxidation of the fluid associated with the enrichment of stoichiometric compositions of nontronite, saponite, and serpentine. ferric iron is suggested by the crystallization of some iron oxide at the inner margins of the vein phyllosilicate. The oxidizing nature of the fluid is also consistent with the presence of laihunite, e.g., Bandfield 1 et al. (1990). 0 Siderite We have only identified phyllosilicate in Lafayette and not the other nakhlites. By contrast, Gooding et al. -1 Amazonian (1991) suggested that there were also occasional flakes (Bar)] 2 -2 } of ferric smectite in Nakhla. The TEM results of Gillet CO

p et al. (2002) for NWA 817 also suggest that some -3 Nontronite poorly crystalline smectite with d-spacing approximately Log[ -4 Saponite 1 nm did form in that meteorite. The abundance of -5 phyllosilicate amongst the hydrous silicate phases in the 0462 8 10 12 14 nakhlites is uncertain but, apart from in Lafayette, the pH gel is clearly the dominant nakhlite secondary silicate Fig. 14. Thermodynamic model showing the stability of phase. phyllosilicates on Mars (Chevrier et al. 2007). The phase equilibrium diagram for the carbonate and nontronite ⁄ Saline Zones (Siderite and Gypsum) saponite is expressed as a function of Mg ⁄ (Mg + Fe) atomic ratio. The dotted line represents the Mg# of Lafayette The soluble salts, e.g., sulphates found in Nakhla phyllosilicate (0.4). The Amazonian log pCO2 ranging from )1to)2 corresponds to weakly acidic to neutral pH and GV in this study and previously (e.g., Bridges and condition for nontronite ⁄ saponite stability. Grady 2000) are further up the hydrothermal cell and hence closer to the Martian surface than the Lafayette assemblage. Their presence suggests that evaporation phyllosilicate. Fully collapsed smectites have fringes of occurred in the nakhlites’ parent rocks, with the most 1 nm and can be a result of electron beam exposure and soluble sulphates and chlorides precipitating from the containment under high vacuum conditions (Klimentidis final portions of the hydrothermal fluid. and Mackinnon 1986). The 0.9 nm fringes of similar The lack of any identified zones of preterrestrial stacking sequence are slightly low for smectites but we crystalline sulphates or carbonate in Y-000593 ⁄ believe this to also be a fully collapsed Fe-smectite Y-000749 implies that soluble salts were not precipitated 1862 H. G. Changela and J. C. Bridges

Permafrost Permafrost Impact-induced fractures

Fe+Si Mg2+ Ca2+ YAMATO

NAKHLA Na+K

GV 10-100 m INCREASING IN GEL INCREASING IN DECREASING IN GEL IN DECREASING LAFAYETTE INCREASING IN GEL IN INCREASING

Carbonate Evaporites Gel Phyllosilicate

Fig. 15. Diagram illustrating the variation in the types of veining found between the Lafayette, GV, Nakhla, and Yamato nakhlites in an impact-induced hydrothermal system. An impact on the Martian surface initiated the melting of buried H2O–CO2 ice (permafrost). The fluid (indicated by curved arrows) flows in the direction from Lafayette to Nakhla through impact-induced fractures. The enlarged region shows the different assemblages associated with the olivine grains of the different nakhlites. The Yamato meteorites experienced a parallel pathway of fluid but still experienced the same fluid fractionation sequence in Mg# and Ca. Other nakhlites not studied here, e.g., NWA 817 may also have been on parallel fluid pathways. from an evaporating fluid in the part of the nakhlite MIL 03346, NWA 998, and NWA 817 parent rocks sampled by the Y-000593 ⁄ Y-000749 meteorites. Alternatively, some soluble salts (e.g., Secondary assemblages have also been reported in chlorides, sulphates) might have been lost during their the other nakhlites MIL 03346, NWA 998, NWA 817 terrestrial residence or had leached out from the inner that have not been included in this study (e.g., Gillet parts to regions closer to the fusion crust (Wentworth et al. 2002; Imae and Ikeda 2007; Treiman and Irving and Gooding 1994). The fluid passing upwards through 2008). Imae and Ikeda (2007) tentatively identified the nakhlites on Mars could have been exhausted in serpentine and saponite within olivine fractures of sulphate and carbonate once it reached the Yamato MIL 03346. However, in the absence of structural nakhlites. This suggests that Y-000593 ⁄ Y-000749 was on information this could have been our identified ‘‘gel’’ as a similar (i.e., parallel) but not identical fluid pathway shown by the gel analyses of nakhlites in our study on to that which the other nakhlites studied here were the ternary diagram of Fig. 13, which can lie between linked by. Although the Yamato meteorites would have the serpentine and saponite solid solution lines. MIL been even closer to the cold Martian surface than the 03346 vein analyses record low Mg numbers similar to other nakhlites (Mikouchi et al. 2003a, 2003b, 2006), those of the Yamato meteorites (Imae and Ikeda 2007). they may have experienced a slightly higher water to Treiman and Irving’s (2008) study of the gel or rock ratio than Nakhla or GV due to the marginally phyllosilicate in NWA 998 also reported a saponite higher volume of veins in Y-000593. The difference composition. The Mg numbers of their quantitative between the Y-000593 Fe versus Si gradient (Fig. 12) analyses are similar to those of Nakhla. Ferroan and that shown by the other nakhlites in our study carbonate and Ca-sulphates, thought to be of Martian supports this slightly different fluid history. Similarly, origin, have also been identified. This resembles the NWA 817 may have experienced a similar but not types of assemblages identified in Nakhla and GV. identical fluid fractionation and cooling history. However, structural information about the hydrous Terrestrial overprints have also occurred in the silicate phases in NWA 998 is missing and thus we do Antarctic finds such as the sulphur rich islands within not know whether they are crystalline or the amorphous the gel of some Y-000593 veins (Fig. 9c) and jarosite in gel phase. The similarity between the secondary Y-000749 (Fig. 10). assemblages in NWA 998 and Nakhla and GV is Alteration assemblages in the nakhlites 1863 consistent with an origin for NWA 998 in the upper been suggested to represent a lower limit for a crater parts of nakhlite pile (Bridges and Warren 2006). This size associated with an impact-induced hydrothermal is, however, in contrast to Mikouchi et al.’s (2006) system (Hagerty and Newsom 2003). Based on this, the model where based on inferred cooling rates and lower limit for the nakhlite impact crater associated ground mass textures, NWA 998 is at the bottom of the with the hydrothermal fluid is approximately 2 km. nakhlite pile. Schwenzer and Kring (2009) modelled the mineral Gillet et al.’s (2002) study of NWA 817 did report abundances for smectites and other low temperature crystalline Fe-smectite in some of that meteorite’s olivine minerals in impact-derived hydrothermal cells. They veining. The Mg number of that alteration material is have shown that the secondary assemblages are largely closest to that of GV in our study. However, Mikouchi controlled by W ⁄ R ratios and temperature. et al. (2006) proposed that NWA 817 lay at one of the Experimental simulations by Baker et al. (2000) with shallowest depths of the nakhlites along with MIL 03346. CO2–H2O fluids interacting with basaltic rock show Thus, although our results suggest that phyllosilicate is siderite to form at 200 C with W ⁄ R ratios of 1–10 and concentrated in the lower parts of the nakhlite pile, where they proposed that this was consistent with the the W ⁄ R was highest and cooling rate slowest there is nakhlites’ hydrothermal system. Taking Lafayette as our evidence from other workers that occasionally some most altered rock and the volume of secondary crystalline smectite formed in higher parts, perhaps as a assemblages as approximately 10% of the olivine, which results of localized variations in W ⁄ R (see below) and is 10% of the bulk rock (Treiman 2005), the volume of cooling rate along parallel fluid pathways. water equivalent to the total volume of secondary assemblages is 1%. With conversion to a mass ratio Nakhlite Hydrothermal Cell W ⁄ R, based on a nakhlite having a specific gravity of approximately 3.5 (Lange and Carmichael 1987), our The fracturing seen in the nakhlites suggest a shock W ⁄ R to produce this volume of secondary assemblages event predating the formation of secondary assemblages at one instant in time would be 0.003. Naumov (2005) that filled them. This suspected impact may have been suggested that relatively high flow rates of fluid (10)4– the heat source for melting permafrost under the 10)3 ms)1) are likely in impact-induced hydrothermal Martian surface and the hydrothermal system that the systems. Assuming this flow rate and the length of the nakhlites have recorded. Hoffman (2000) showed that Lafayette mass to be 1 m, W ⁄ R ratios from 1 to 10 H2O-dominated subsurface ice extended to depths from (Baker et al. 2000) imply a short-lived hydrothermal 5 to 10 km in Amazonian terrains, far below the system of 1–10 months. The W ⁄ R value, within this maximum approximately 100 m depth of the nakhlites. range, would have been highest in Lafayette and Impact-induced melt sheets have been shown to sustain decreased towards the margins of the hydrothermal cell. large temperatures (>100 C) for long periods of time, A combination of more rapid cooling and lower W ⁄ R 104–105 yr, in 100 km sized craters (Daubar and Kring may therefore have generally prohibited the formation 2001; Abramov and Kring 2005). Neighbouring rocks to of phyllosilicates in the other meteorites. This is such large craters should experience higher degrees of consistent with the impact model of Schwenzer and shock compared to the relatively small amounts Kring (2009) in which smectites are favourably experienced by the nakhlites (Fritz et al. 2005). However, precipitated when the W ⁄ R ratio increases, although we do not know what radial distance the nakhlites were their W ⁄ R ratios were much higher (1000) and from the assumed impact source of the hydrothermal modelled for larger hydrothermal systems. Some cell. We start with the assumption of a relatively small variation in cooling rates and W ⁄ R may explain the crater impact (<10 km) inducing a hydrothermal presence of some crystalline smectite in NWA 817 at a system. The Amazonian age of the nakhlites (1.3 Ga) shallow depth in the nakhlite rocks (Gillet et al. 2002). shows that they were derived from a relatively young Martian surface. Areas of the northern lowlands that A Model of the Nakhlites’ Alteration Process have similar ages to the nakhlites, show 10)3– 10)4 craters km)2 for crater diameters ranging from 1 to Carbonate and Clay Identified on the Martian Surface 10 km (Hartmann and Neukum 2001). Large craters Carbonates coupled with the presence of (>100 km) are much rarer (densities approximately phyllosilicates have recently been discovered through 10)7 craters km)2) and are therefore less likely to be the orbiter observations of Noachian or Hesperian terrains impact heat source of this hydrothermal cell. on Mars (Bibring et al. 2005; Poulet et al. 2005; Ehlmann The Lonar crater, India (diameter 1.8 km), contains et al. 2008; Mustard et al. 2008). These secondary evidence of extensive hydrothermal alteration in the minerals are similar to those found in the nakhlites (i.e., form of phyllosilicates due to the impact event and has siderite and Fe-smectite, serpentine). The formation 1864 H. G. Changela and J. C. Bridges conditions for smectites identified in this way have been 4. The nakhlite parent fluid seeped upwards from suggested to require significant water reservoirs with Lafayette towards GV, Nakhla, and the Yamato moderate to alkaline pHs and warm temperatures, either meteorites—depositing the gel together with some maintained on the surface or in the shallow crust of Mars siderite in GV and Nakhla veins and mesostasis. (Poulet et al. 2005; Mustard et al. 2008). Similar Only a little smectite-serpentine (e.g., other workers conditions would also favour carbonates on the Martian have mentioned some in Nakhla and NWA 817) surface such as those found in the Nili Fossae region was precipitated beyond that in Lafayette because where surface fluvial activity and possibly shallow lake of the faster cooling rate in the hydrothermal deposits could have been the setting for carbonate system once the gel precipitation had commenced formation (Ehlmann et al. 2008). The phyllosilicates and also because of the decreasing W ⁄ R in the which have been recorded in the walls, ejecta and central latter stages. As the fluid migrated, its composition peaks of the ancient highlands, however, also show a fractionated with a decrease in siderite Ca ⁄ Mg strong association with more localized impact-induced ratio, and the gel’s Mg ⁄ (Mg + Fe) ratio and an hydrothermal activity (Mustard et al. 2008). On the basis increase in Fe + Si contents. Some incorporation of the evidence from our nakhlite study, although of of Al from dissolved mesostasis in Lafayette into different age to the minerals identified from orbit, we also the phyllosilicate that replaced it also occurred in suggest that impact is a likely origin for many of the addition to this fluid fractionation between the Martian phyllosilicate and carbonates. nakhlites. 5. The Yamato meteorites (and NWA 817) may have Impact Driven Hydrothermal Alteration of the Nakhlites encountered separate fluid paths, due to their In a hydrothermal cell initiated by an impact on the slightly higher W ⁄ R ratio than Nakhla and GV, Martian surface, the possible stages of fluid activity different Fe ⁄ Si ratio, and higher Al contents in the recorded in the nakhlites (Fig. 15) were as follows: Y-000593 gel. The Yamato meteorites would have, 1. An impact (we suggest 1–10 km Dc) created however, experienced a similar fluid fractionation fracturing in the nakhlites (as most clearly shown sequence as Lafayette to Nakhla except the fluid in ) by the sawtooth fractures in Lafayette, Nakhla, Y-000593 had been exhausted of HCO3 . and GV) and melted H2O–CO2 ice within 6. Once the fluid reached the near surface, e.g., in approximately 100 m of the surface. The icy fluid Nakhla, progressive evaporation of the fluid led to source had a stored CO2 depository equivalent to the formation of soluble salts such as gypsum and approximately 10–100 mbar pCO2. This CO2 halite. There is some evidence the most soluble ) + dissolved in the fluid to form HCO3 . cations were also adsorbed into the gel (e.g., Na 2. The fluid, with a W ⁄ R ratio of 1–10 in Lafayette, and K+) in higher abundances at this point. The dissolved olivine and mesostasis within the nakhlite deposition process of the secondary assemblages parent body and migrated upwards from Lafayette in the hydrothermal cell lasted approximately at the deepest level towards Nakhla near the 1–10 months following the impact. surface. The more insoluble cations Fe, Ca, Mn were the main constituents of the first secondary CONCLUSIONS mineral siderite to precipitate out along the edges of the fracture walls within the olivine. This was A variation in the mineralogy of the Lafayette, GV, followed by the crystallization of a mixture of Nakhla, and Y-000593 nakhlite secondary assemblages serpentine and saponite ⁄ nontronite smectite in the is found that follows a trend associated with their veins followed by Fe oxide as the fluid became relative formation depths approximately 10–100 m progressively more oxidizing in Lafayette. The fluid under the surface of Mars. The secondary assemblages at this stage was £150 C with near neutral pH. consist of siderite, Fe smectite-serpentine, minor Fe 3. Depletion in Ca, Fe, Mn, Mg, and CO2 exhaustion oxide, an amorphous hydrated silicate phase ‘‘gel’’ of led to silica enrichment in the fluid. Amorphous or smectite-serpentine like composition and soluble salts very weakly crystalline silica-rich hydrous gel was such as gypsum. These minerals are mainly present not then deposited in the centre of veins as the fluid only within olivine fractures but also within the cooled. The fluid also dissolved parts of the mesostasis. All of the nakhlites studied are dominated mesostasis in Lafayette and precipitated some by the amorphous gel phase. We note that both intermixed serpentine and smectite with gel there as structural (TEM-SAED) and chemical analyses are well. Weakly crystalline, partially broken down required to distinguish between the phyllosilicates and olivine is also found at the margins of many veins the gel phase, rather than solely quantitative chemical in the Yamato meteorites. analysis. Alteration assemblages in the nakhlites 1865

Our textural observations suggest that within the analyses in this study. M.Phys. students Miss Nishad olivine fractures of Lafayette the crystallization Karim and Mr Sherali Walji of the Department of sequence was Ca-rich siderite that was corroded and Physics and Astronomy, University of Leicester are also partially replaced by phyllosilicate followed by Fe oxide thanked for assisting with this study. Mr John Critchell crystallization. Then, amorphous silicate gel was of Jeol UK Ltd is thanked for his advice on TEM of precipitated in the fractures. Finally, soluble salts phyllosilicates and also Dr David Bish for his assistance crystallized from the fluid. Partially broken down with phyllosilicate analyses in this study. The STFC is olivine (probably laihunite) is also present at the acknowledged for funding of Hitesh Changela’s Ph.D. margins of some olivine fractures within Y-000593. All studentship. T. Mikouchi and two anonymous referees of the secondary minerals apart from the final soluble are thanked for reviews which very much helped salts precipitated in Lafayette. The other, shallower improve this article. depth nakhlites we have studied contain the gel, siderite, and salts (apart from Y-000593 ⁄ Y-000749 which Editorial Handling—Dr. Alan Treiman contains only the gel). Other workers have shown that occasionally phyllosilicate crystallized in other nakhlites, REFERENCES which we suggest may be the result of variations in Abramov O. and Kring D. A. 2005. Impact-induced W ⁄ R and ⁄ or temperature. hydrothermal activity on early Mars. Journal of The phyllosilicate in Lafayette is a complex mixture Geophysical Research 110:E12S09. of 1:1 and 2:1 phyllosilicate that we suggest to be Ashworth J. R. and Hutchison R. 1975. Water in non- Fe-serpentine and Fe-smectite, although the relative carbonaceous stony meteorites. Nature 256:714–715. complexity of phyllosilicate means that we do not rule Baker L. L., Agenbroad D. J., and Wood S. A. 2000. Experimental hydrothermal alteration of a Martian out the possibility of other phyllosilicates being analogue basalt implications for Martian meteorites. identified in future. There may be a higher proportion Meteoritics & Planetary Science 35:31–38. of serpentine within the mesostasis phyllosilicate. Bandfield J. F., Veblen D. R., and Jones B. F. 1990. A model has been proposed to describe the nakhlite Transmission electron microscopy of subsolidus oxidation hydrothermal setting. We suggest that the aqueous fluid and weathering of olivine. Contributions to Mineralogy and Petrology 106:110–113. was near neutral pH, £150 C and became progressively Bibring J.-P., Langevin Y., Gendrin A., Gondet B., Poulet F., more oxidized, silica-rich, and, ultimately, when near Berthe´ M., Soufflot A., Arvidson R., Mangold N., the surface, evaporite-salt rich. An association with Mustard J., and Drossart P. 2005. Mars surface diversity sawtooth fracturing and upwards fluid migration has as revealed by the OMEGA ⁄ Mars Express observations. led us to propose that the fluid was initiated by melting Science 307:1576–1581. Bridges J. C. and Grady M. M. 2000. Evaporite mineral of buried H2O–CO2 ice after a local impact. The fluid assemblages in the nakhlite (Martian) meteorites. Earth migrated through the nakhlites towards the surface. We and Planetary Science Letters 176:267–279. suggest on the basis of likely W ⁄ R ratios (1–10; Bridges J. C. and Warren P. H. 2006. The SNC meteorites: experimental work of Baker et al. 2000) and fracture- Basaltic igneous processes on Mars. Journal of the filling times based on fluid velocities (Naumov 2005) Geological Society, London 163:229–251. Bridges J. C., Catling D. C., Saxton J. M., Swindle T. D., that the hydrothermal system lasted for 1–10 months, Lyon I. C., and Grady M. M. 2001. Alteration cooling relatively rapidly. The fluid fractionation is assemblages in the Martian meteorites: Implications for most clearly shown by variation in the Mg ⁄ (Mg + Fe) near-surface processes. In Chronology and evolution of ratio and Fe + Si abundances between the gel of the Mars, edited by Kallenbach R., Geiss J., and Hartmann nakhlites and ultimately precipitation of soluble salts. W. K. Kluwer: Dordrecht. pp. 365–392. Brigatti M. F. 1983. Relationships between composition and Veining in Y-000749 contains a terrestrial overprint structure in Fe-rich smectites. Clay Minerals 18:177–186. of jarosite and other sulphates. More information on the Brigatti M. F., Lugli C., Poppi L., and Venturelli G. 1999. nakhlite hydrothermal cell may be obtained by further Iron-rich saponite: Dissolution reactions and Cr uptake. analysis of the other nakhlites (MIL 03346, NWA 5790, Clay Minerals 34:637–645. NWA 817, and NWA 998) in the light of our study. Burns R. G. and Martinez S. L. 1991. Mossbauer spectra of olivine rich achondrites: Evidence of preterrestrial redox reactions. Proceedings, 21st Lunar and Planetary Science Acknowledgments— We thank the Department of Conference. pp. 331–340. Mineralogy, NHM and Dr Caroline Smith and Mr Chevrier V., Poulet F., and Bibring J. P. 2007. Early Peter Tandy for providing the nakhlite samples and clay geochemical environment of Mars as determined from standard. We also thank Mr Graham Clark of the thermodynamics of phyllosilicates. Nature 448:60–63. Daubar I. J. and Kring D. A. 2001. Impact-induced Advanced Microscopy Centre, Leicester and Mr John hydrothermal systems: Heat sources and lifetimes (abstract Spratt of the Department of Mineralogy, NHM for #1727). 32nd Lunar and Planetary Science Conference. assistance in the electron microscopy and EPMA CD-ROM. 1866 H. G. Changela and J. C. Bridges

Deer W. A., Howie R. A., and Zussman J. 1966. An introduction Komadel P., Madejova J., Laird D. A., Xia Y., and Stucki J. to the rock forming minerals. London: Longman. 528 p. W. 2000. Reduction of Fe (III) in griffithite. Clay Minerals Demidov N. E. and Gilichinsky D. A. 2009. Terrestrial 35:625–634. permafrost models and analogues of Martian habitats and Kondoh S., Kitamura M., and Morimoto N. 1985. Synthetic 2+ 3+ inhabitants. Soil Biology 16:323–341. laihunite (ClassXFe-2 ⁄ )-3XFe-2X( )SiO4), an oxidation- Ehlmann B. L., Mustard J. F., Murchie S. L., Poulet F., product of olivine. American Mineralogist 70:737–746. Bishop J. L., Brown A. J., Calvin W. M., Clark R. N., Ko¨ ster H. M., Ehrlicher U., Gilgh A., Jordan R., Murad E., Des Marais D. J., Milliken R. E., Roach L. H., Roush T. and Onnich K. 1999. Mineralogical and chemical L., Swayze G. A., and Wray J. J. 2008. Orbital characteristics of five nontronites and Fe-rich smectites. identification of carbonate-bearing rocks on Mars. Science Clay Minerals 34:579–599. 322:1828–1832. Kuebler K., Jolliff B. L., and Treiman A. 2004. A survey of Fiore S., Huertas F. J., Huertas F., and Linares J. 1995. alteration products and other secondary minerals Morphology of kaolinite crystals synthesized under recovered from Martian meteorites in Antarctica (abstract hydrothermal conditions. Clays and Clay Minerals 43:353– #2228). 35th Lunar and Planetary Science Conference. 360. CD-ROM. Fritz J., Artemieva N., and Greshake A. 2005. Ejection of Lange R. A. and Carmichael I. S. E. 1987. Densities of Na2O- Martian meteorites. Meteoritics & Planetary Science K2O-CaO-MgO-FeO-Fe2O3-Al2O3-TiO2-SiO2 40:1393–1411. liquids—New measurements and derived partial molar Gates W. P., Slade P. G., Manceau A., and Lanson B. 2002. properties. Geochimica et Cosmochimica Acta 51:2931– Site occupancies by iron in nontronites. Clay and Clay 2946. Minerals 50:223–239. Lentz R. C. F., Taylor G. J., and Treiman A. H. 1999. Gillet P., Barrat J. A., Deloule E., Wadhwa M., Jambon A., Formation of a Martian pyroxenite: A comparative study Sautter V., Devouard B., Neuville D., Benzerara K., and of the nakhlites and Theo’s Flow. Meteoritics & Planetary Lesourd M. 2002. Aqueous alteration in the Northwest Science 34:919–932. Africa 817 (NWA 817) Martian meteorite. Earth and Manceau A., Drits V. A., Lanson B., Chateigner D., Wu J., Planetary Science Letters 203:431–444. Huo D., Gates W. P., and Stucki J. W. 2000. Oxidation- Golden D. C., Ming D. W., Schwandt C. S., Morris R. V., reduction mechanism of iron in dioctahedral smectites: II. Yang S. V., and Lofgren G. E. 2000. An experimental study Crystal chemistry of reduced Garfield nontronite. on kinetically-driven precipitation of calcium, magnesium, American Mineralogist 85:153–172. and iron carbonates from solution: Implications for the Manning C. V., McKay C. P., and Zahnle K. J. 2006. Thick low-temperature formation of carbonates in Martian and thin models of the evolution of carbon dioxide on meteorite Allan Hills 84001. Meteoritics & Planetary Science Mars. Icarus 180:38–59. 35:457–465. Meyer C. 2006. Mars Meteorite Compendium. http://curator. Gooding J. L. 1986. Clay-mineraloid weathering products in jsc.nasa.gov/antmet/mmc/index.cfm. Accessed October 30, Antarctic meteorites. Geochimica et Cosmochimica Acta 2010. 50:2215–2223. Michel F. M., Ehm L., Antao S. M., Lee P. L., Chupas P. J., Gooding J. L., Wentworth S. J., and Zolensky M. E. 1991. Liu G., Strongin D. R., Schoonen M. A. A., Phillips B. Aqueous alteration of the Nakhla meteorite. Meteoritics L., and Parise J. B. 2007. The structure of ferrihydrite, 326:135–143. a nanocrystalline material. Science 316:1726–1729, Grauby O., Petit S., Decarreau A., and Baronnet A. 1994. doi:10.1126/science.1142525. The nontronite-saponite series: An experimental approach. Mikouchi T., Koizumi E., Monkawa A., Ueda Y., and European Journal of Mineralogy 6:99–112. Miyamoto M. 2003a. Mineralogy and petrology of Yamato Hagerty J. J. and Newsom H. E. 2003. Hydrothermal 000593: Comparison with other Martian nakhlite alteration at the Lonar Lake impact structure, India: meteorites. Antarctic Meteorite Research (NIPR) 16:34–57. Implications for impact cratering on Mars. Meteoritics & Mikouchi T., Koizuni E., Monkawa A., Ueda Y., and Planetary Science 38:365–381. Miyamoto M. 2003b. Mineralogical comparison of Hartmann W. K. and Neukum G. 2001. Evolution of Mars. Y000593 with other nakhlites: Implications for relative In Cratering chronology and the evolution of Mars, edited burial depths of nakhlites (abstract #1944). 34th Lunar by Kallenbach R., Geiss J., and Hartmann W. K. Kluwer: and Planetary Science Conference. CD-ROM. Dordrecht. pp. 365–392. Mikouchi T., Miyamoto M., Koizumi E., Makishima J., and Hoffman N. 2000. White Mars: A new model for Mars’ Mckay G. 2006. Relative burial depths of nakhlites: An surface and atmosphere based on CO2. Icarus 146:326–342. update (abstract #1865). 37th Lunar and Planetary Science Imae N. and Ikeda Y. 2007. Petrology of the Miller Range Conference. CD-ROM. 03346 nakhlite in comparison with the Yamato-000593 Mustard J. F., Murchie S. L., Pelkey S. M., Ehlmann B. L., nakhlite. Meteoritics & Planetary Science 42:171–184. Milliken R. E., Grant J. A., Bibring J.-P., Poulet F., Klimentidis R. and Mackinnon I. D. R. 1986. High resolution Bishop J., Noe Dobrea E., Roach L., Seelos F., Arvidson imaging of ordered mixed layer clays. Clay and Clay R. E., Wiseman S., Green R., Hash C., Humm D., Minerals 34:155–164. Malaret E., McGovern J. A., Seelos K., Clancy T., Clark Kodama H., de Kimpe I. C. R., and Dejou J. 1998. Ferrian R., Des Marais D., Izenberg N., Knudson A., Langevin saponite in a gabbro saprolite at Mont Me´ gantic, Quebec. Y., Martin T., McGuire P., Morris R., Robinson M., Clays and Clay Minerals 36:102–110. Roush T., Smith M., Swayze G., Taylor H., Titus T., and Kohyama N. and Sudo T. 1975. Hisingerite occurring as a Wolff M. 2008. Hydrated silicate minerals on Mars weathering product of iron-rich saponite. Clays and Clay observed by the Mars Reconnaissance Orbiter CRISM Minerals 23:215–218. instrument. Nature 454:305–309. Alteration assemblages in the nakhlites 1867

Naumov M. V. 2005. Principal features of impact-generated Swindle T. D., Treiman A. H., Lindstrom D. J., Burkland M. hydrothermal circulation systems: Mineralogical and K., Cohen B. A., Grier J., Li B., and Olson E. K. 2000. geochemical evidence. Geofluids 5:165–184. Noble gases in iddingsite from the Lafayette meteorite: Nimis P., Tesalina S. G., Omenetto P., Tartarotti P., and Evidence of liquid water on Mars in the last few hundred Lerouge C. 2004. Phyllosilicate minerals in the million years. Meteoritics & Planetary Science 37:107– hydrothermal mafic–ultramafic-hosted massive-sulfide 115. deposit of Ivanovka (southern Urals): Comparison with Tosca N. J., Milliken R. E., and Michel F. M. 2008. Smectite modern ocean seafloor analogues. Contributions to formation on early Mars: Experimental constraints (abstract Mineralogy and Petrology 147:363–383. #7030). LPI Workshop on Martian Phyllosilicates: Noguchi T., Nakamura T., Misawa K., Imae N., Aoki T., and Recorders of Aqueous Processes. Houston, Texas: Lunar Toh S. 2009. Laihunite and jarosite in the Yamato 00 and Planetary Institute. nakhlites: Alteration products on Mars? Journal of Treiman A. H. 2005. The nakhlite meteorites: Augite-rich Geophysical Research 114:E10004. igneous rocks from Mars. Chemie der Erde 65:203–270. Poulet F., Bibring J.-P., Mustard J. F., Gendrin A., Mangold Treiman A. H. and Goodrich C. A. 2002. Pre-terrestrial N., Langevin Y., Arvidson R. E., Gondet B., and Gomez aqueous alteration of the Y-000593 and Y-000749 nakhlite C. 2005. Phyllosilicates on Mars and implications for early meteorites. Proceedings of the NIPR Symposium on Martian climate. Nature 438:623–627. Antarctic Meteorites 27:166–167. Romero R., Robert M., Elsass F., and Garcia C. 1992. Treiman A. H. and Irving A. J. 2008. Petrology of the Evidence by transmission electron microscopy of nakhlite (Martian) meteorite Northwest Africa (NWA) weathering microsystems in soils developed from 998. Meteoritics & Planetary Science 43:829–854. crystalline rocks. Clay Minerals 27:21–33. Treiman A. H. and Lindstrom D. J. 1997. Trace element Saxton J. M., Lyon I. C., Chatzitheodoridis E., and Turner G. geochemistry of Martian iddingsite in the Lafayette 2000. Oxygen isotopic composition of carbonate in the meteorite. Journal of Geophysical Research 102:9153– Nakhla meteorite: Implications for the hydrosphere and 9163. atmosphere of Mars. Geochimica et Cosmochimica Acta Treiman A. H., Barret R. A., and Gooding J. L. 1993. 64:1299–1309. Preterrestrial aqueous alteration of the Lafayette (SNC) Schofield P. F., Knight K. S., and Stretton I. C. 1996. meteorite. Meteoritics & Planetary Science 28:86–97. Thermal expansion of gypsum investigated by neutron Vicenzi E. P. and Heaney P. J. 1999. Examining Martian powder diffraction. American Mineralogist 81:847–851. alteration products using in situ TEM sectioning: A novel Schwenzer S. P. and Kring D. A. 2009. Impact-generated application of the focused ion beam (fib) for the study of hydrothermal systems capable of forming phyllosilicates extraterrestrial materials (abstract #2005). 30th Lunar and on Noachian Mars. Geology 37:1091–1094. Planetary Science Conference. CD-ROM. Sherman G. D., Ikawa H., Uehara G., and Okazaki E. 1962. Vicenzi E. P. and Heaney P. J. 2000. Multiple Martian Types of occurrence of nontronite and nontronite-like fluids: The alteration sequence in the Lafayette SNC minerals in soils. Pacific Science 16:57–62. meteorite (abstract). Meteoritics & Planetary Science Spencer L. M., Mckay D. S., Gibson E. K., Thomas-Keprta 35:A164–A165. K. L., Wentworth S. J., and Clemett S. J. 2008. Wentworth S. J. and Gooding J. L. 1994. Carbonates and Observation and analysis of Martian meteorite Y000593: sulfates in the Chassigny meteorite: Further evidence of Evidence for biosignatures (abstract #1680). 39th Lunar aqueous chemistry on the SNC parent planet. Meteoritics and Planetary Science Conference. CD-ROM. 29:860–863.