Comparisons of the four Miller Range nakhlites, MIL 03346, 090030, 090032 and 090136: Textural and compositional observations of primary and secondary mineral assemblages


Corresponding author. E-mail:


Abstract– Petrological and geochemical analyses of Miller Range (MIL) 03346 indicate that this meteorite originated from the same augitic cumulate layer(s) as the nakhlite Martian meteorites, but underwent rapid cooling prior to complete crystallization. As with the other nakhlites, MIL 03346 contains a secondary alteration assemblage, in this case consisting of iddingsite-like alteration veins in olivine phenocrysts, Fe-oxide alteration veins associated with the mesostasis, and Ca- and K,Fe-sulfate veins. We compared the textural and mineralogical compositions of MIL 090030, 090032, and 090136 with MIL 03346, focusing on the composition and Raman spectra of the alteration assemblages. These observations indicate that the meteorites are paired, and that the preterrestrial olivine-bound alteration assemblages were produced by weakly acidic brine. Although these alteration assemblages resemble similar assemblages in Nakhla, the absence of siderite and halite in the Miller Range nakhlites indicates that the parental alteration brine was comparatively HCO3 depleted, and less concentrated, than that which altered Nakhla. This indicates that the Miller Range nakhlite alteration brine experienced a separate evolutionary pathway to that which altered Nakhla, and therefore represents a separate branch of the Lafayette-Nakhla evaporation sequence. Thin-sections cut from the internal portions of these meteorites (away from any fusion crust or terrestrially exposed edge), contain little Ca-sulfate (identified as gypsum), and no jarosite, whereas thin-sections with terrestrially exposed edges have much higher sulfate abundances. These observations suggest that at least the majority of sulfate within the Miller Range nakhlites is terrestrially derived.


The similarity between the Martian atmospheric composition measured by the Viking landers, and the shock-implanted noble gas isotopic compositions of the Shergotty-Nakhla-Chassigny (SNC) meteorites, indicates that these mafic cumulates originate from Mars (e.g., Bogard and Johnson 1983). There are currently in excess of 40 individually launched SNC meteorites in terrestrial collections, in addition to the ungrouped Martian orthopyroxenite, Allan Hills 84001. The SNC meteorites have a range of compositions from basalts and lherzolites (shergottites), to clinopyroxenites (nakhlites) and dunites (Chassigny and Northwest Africa [NWA] 2737).

The nakhlites are igneous cumulates consisting of phenocrystic augites and minor olivine, titanomagnetites, and siliceous mesostasis (e.g., Treiman 2005). Okazaki et al. (2003), among others, reported their crystallization and ejection ages as 1.3 Ga and 10–12 Ma, respectively. The nakhlites represent the second most abundant group of SNCs, containing seven separate meteorites—Nakhla (an Egyptian fall, 1909), Governador Valadares and Lafayette (well preserved finds), NWA 817, NWA 998, Yamato-000593/749/802 (three launch-paired Antarctic finds), and Miller Range (MIL) 03346 (a heavily weathered Antarctic find).

Miller Range 03346 was collected during the 2003 collaborative expedition between the Antarctic Search for Meteorites program (ANSMET) and the Mars Exploration Program. Petrological and geochemical analyses of MIL 03346 indicate that it originated from the same augitic cumulate layer(s) as the other nakhlites. However, in contrast to the other nakhlites, the presence of a vitrophyric intercumulus matrix in MIL 03346, and the absence of albitic feldspar laths, suggest rapid cooling occurred toward the end of crystallization (Day et al. 2006). Estimates of oxygen fugacity by Hammer and Rutherford (2005) and Dyar et al. (2005) indicate that MIL 03346 may also be the most oxidized SNC meteorite recognized to date.

Both Anand et al. (2005) and Stopar (2007) reported secondary alteration in the form of olivine-bound iddingsite-like alteration veins in MIL 03346; these were found to be similar to preterrestrial alteration veins in other nakhlites. More recently, Changela and Bridges (2011) studied these veins in a group of nakhlites using transmission electron microscopy (TEM), and concluded that, rather than being composed entirely of phyllosilicates, they contain siderite and amorphous silicate gels, the composition of which varies between meteorites. MIL 03346 is the most terrestrially weathered nakhlite, and is one of the only two nakhlites to contain jarosite (along with Yamato-000593/749—Treiman and Steele 2008; Noguchi et al. 2009). Jarosite is a relatively rare hydrated Fe,K-sulfate, that mostly exists in mining-related acid leaching environments on Earth (e.g., Gieré et al. 2003). Sulfates have been detected both from orbit and on the surface of Mars in recent years, and the Mars exploration Rovers have identified Martian jarosite (e.g., Squyres et al. 2004; Bibring et al., 2005; Clark et al. 2005; McLennan et al. 2005; Poulet et al. 2005; Ming et al. 2006; Dohm et al. 2009). Therefore, we must rely upon textural evidence to determine whether jarosite and other sulfate veins in MIL 03346 have terrestrial or Martian origins.

We compared the mineralogical compositions of internal and external thin-sections of MIL 03346 with those of internal thin-sections from the newly collected MIL 090030, 090032, and 090136 meteorites, with particular emphasis on the composition and Raman spectra of the olivine-bound alteration assemblages and sulfate veins. The external thin-sections of MIL 03346 have been exposed to the Antarctic environment along at least one edge, whereas the internal thin-sections were from a central portion of the parent meteorite. The aim of these comparisons was to reveal whether the four Miller Range nakhlites are paired, and if so, whether the secondary alteration phase compositions and abundances are similar to those in other nakhlites. Similarities between preterrestrial alteration phase assemblages in unpaired nakhlite samples would imply the same fluid(s) flowed through and altered the whole nakhlite sequence of rocks on Mars. To assist with these comparisons, we also used the Raman microscope to analyze the olivine-bound alteration assemblages in Nakhla, and compared our data to the Raman spectral data of Kuebler et al. (2004). This study is the first to compare alteration phases in the four Miller Range nakhlites. Differences in secondary alteration phase abundance between the four meteorites may indicate that certain phases have a terrestrial origin.


We utilized the JEOL JSM-5900LV scanning electron microscope at the University of Hawai’i to produce high-resolution backscatter electron images of four MIL 03346 polished thin-sections (126, 128, 173, and 174), and one thin-section of MIL 090030(23), MIL 090032(24), and MIL 090136(21), prepared at the NASA Johnson Space Center. Elemental X-ray images were also produced to locate areas of interest in each thin-section. These X-ray images were used to determine the mineralogical mode of each sample, via the production of false-color composite images and subsequent pixel counting using Adobe Photoshop software. As each whole thin-section X-ray image was produced at a magnification of ×250, the pixel size was kept constant for each (approximately 20 μm). Specific areas of interest were subsequently X-ray imaged at higher resolutions, with various pixel sizes (approximately 1–5 μm).

The major- and minor-element chemistry of mineralogical phases were analyzed with a JEOL JXA-8500F electron microprobe, also at the University of Hawai’i. The analytical conditions of the electron microprobe included a 10 nA primary beam current, 15 keV accelerating voltage, and 5 μm spot size. The count time for beam-sensitive elements (e.g., Na and K) was set at 15 s, whereas that for less sensitive elements was set at 30 s. Beam-sensitive elements were measured first. Matrix effects were corrected using PAP procedures. The elemental detection limits are 0.03 wt% for SiO2, Al2O3, and MgO; 0.04 wt% for TiO2, CaO, and K2O; 0.06 wt% for Na2O and Cr2O3; 0.07 wt% for MnO; 0.08 wt% for FeO. San Carlos olivine (NMNH 111312), Rockport fayalite (USNM 85276), and Makaopuhi Lake basaltic glass (USNM 113498/1) were utilized as standards, along with USNM 104021 apatite and NMNH 117075 chromite. UH standardized orthoclase Or-1 (5–168), Kakanui augite, Lake County labradorite, and Staunton meteorite troilite were also used as standards.

Primary igneous and secondary alteration phases were defined by Raman spectroscopy at the University of Hawai’i. Spectra were acquired with a confocal Raman microscope (alpha-300, WITec), using the 532 nm excitation wavelength (green) of a frequency-doubled solid-state Nd:YAG laser. The maximum power of the laser was approximately 23–24 mW, but a filter with an optical density of 200 was inserted between the laser and the sample to reduce the laser power accordingly. The laser beam was focused on the samples by a 50× objective, leading to an approximately 2 μm diameter spot. Each spectrum was acquired in a spectral window of −150 to 3750 cm−1 (shift relative to the laser wavelength). The calibration of the spectrometer was checked daily for any shift in spectral peak position, via analysis of a pure silicon wafer with known major Raman peak at 524 cm−1. The acquisition time was fixed at 5 s and each spectrum is an average of five acquisitions.


Petrology and Mineral Chemistry

Primary Phases

Miller Range 090030, 090032, and 090136 have similar mineralogical modes (Table 1) and textures (Fig. 1) to MIL 03346 (Dyar et al. 2005; McKay and Schwandt 2005; Mikouchi et al. 2005; Day et al. 2006; Imae and Ikeda 2007). However, the olivine contents of the three new Miller Range nakhlites extend from 3.0 to 8.3 vol%, compared to 0.8–4 vol% in MIL 03346 (McKay and Schwandt 2005; Mikouchi et al. 2005; Day et al. 2006; Imae and Ikeda 2007). The wide range in olivine content, even between thin-sections of the same meteorite, may be due to the large size of the olivine phenocrysts in relation to thin-section size. However, as previous authors have reported varying olivine content in MIL 03346, this range most likely reflects heterogeneity in the meteorites. Despite the range, these olivine modal abundances are still lower than the reported abundances in other nakhlites (Lentz et al. 1999; Sautter et al. 2002; Imae et al. 2003; Treiman 2005). MIL 090030,23 contains less augite and more mesostasis than in MIL 03346, but aside from this, and the varying olivine content, there does not appear to be any major difference between the mineralogical modes of the Miller Range nakhlites.

Table 1.   Modal mineralogy of the four Miller Range nakhlites (vol%).
SampleTotal PixelsPyroxeneOlivineMesostasisMagnetiteSulfiteSilicaSulfateOlv alt
  1. aAverage values of the four thin-sections MIL 03346,126, 128, 173, and 174, total pixel number is therefore higher.

MIL 03346a1,175,66678.32.917.41.3<<0.1
MIL 090030,23159,16766.48.322.62.0<0.10.3<0.10.2
MIL 090032,2457,74075.<0.10.2<0.10.2
MIL 090136,2165,35075.25.618.11.1<0.10.2<0.10.1
Figure 1.

 False-color X-ray images of MIL 03346,128 (A) and MIL 090030 (B), highlighting the similarity in texture and mineralogy between the two meteorites. These images are composed of elemental X-ray images of Fe (green), Mg (red), Ca (pink), and Si (blue), resulting in yellow-green olivines (olv), pink augites (cpx), blue mesostasis (meso), and bright green magnetites (mag).

All four meteorites are clinopyroxenites, containing cumulus euhedral to subhedral augites, <2.5 mm in length, that commonly show polysynthetic twinning and distinct Fe-rich rims (10–70 μm wide). In places, these phenocrysts appear clustered, and the augite–augite grain boundaries at the center of these clusters commonly have no Fe-rich rims (Fig. 2A). Augite phenocrysts in MIL 090030, 090032, and 090136 have very similar compositions (Table 2), ranging from En39Fs22Wo39 (augite cores) to En5Fs53Wo42 (ferroaugite and ferrohedenbergite rims). Clinopyroxene in MIL 03346 shows a similar range from En37Fs22Wo41 to En17Fs45 Wo38 (Fig. 2B). Day et al. (2006) and Imae and Ikeda (2007) report similar compositions, and expand this range to include hedenbergitic and ferrohedenbergitic intercumulus pyroxenes. Our data indicate that intercumulus pyroxenes in the three new Miller Range nakhlites are of similar compositions. Alongside the clinopyroxene grains are rare subhedral to anhedral olivine phenocrysts, <1.5 mm in diameter. Phenocrystic olivine commonly poikilitically encloses augite; its composition ranges between Fo44–5 in the three new meteorites, and Fo43–14 in MIL 03346 (Fig. 2). In all four meteorites, small anhedral grains of orthopyroxene (10–50 μm) are associated with the olivine phenocrysts, commonly at olivine-clinopyroxene grain boundaries. Orthopyroxene compositions range between En42–29 in MIL 03346, En51–30 in MIL 090030, En47–43 in MIL 090032, and En48–25 in MIL 090136 (Table 2).

Figure 2.

 Backscattered electron images of the Miller Range nakhlites (A), highlight the absence of Fe-rich rims, where augite phenocrysts cluster together in all four meteorites. In addition, clinopyroxene and olivine compositions are similar in the Miller Range nakhlites (B). Data collected from the augitic phenocrysts (both cores and rims) are shown as black symbols, and intercumulus clinopyroxene grain compositions are represented by gray symbols. The previously published data of Day et al. (2006) and Imae and Ikeda (2007) are represented by the white and gray data envelopes, respectively.

Table 2.   Representative major-element mineral chemistry (wt%).
PhasePhenocrystic augiteInt cpxOpxOlivineMagnetiteCristobaliteInt glassOlv alt  
MIL 090030CoreRim  CoreRim     Pyrrhotite 
PhasePhenocrystic augiteInt cpxOpxOlivineMagnetiteCristobaliteInt glassOlv altFe-oxide vein  
MIL 090032CoreRim  CoreRim      Pyrrhotite
PhasePhenocrystic augiteInt cpxOpxOlivineMagnetiteCristobaliteInt glassOlv alt  
MIL 090136CoreRim  CoreRim     Pyrrhotite 
PhasePhenocrystic augiteInt cpxOpxOlivineMagnetiteCristobaliteInt glassOlv altIncip olv altCa-sulfateJarosite 
MIL 090030CoreRim  CoreRim      
  1. Notes: n.d. = not detected; – = not measured.

Fe2O3 41.1543.440.79

A fine-grained felsic mesostasis surrounds the phenocrysts, punctuated with minor skeletal Ti-magnetite, fayalite (<25 μm, Fa91–96), pyrrhotite (<20 μm), cristobalite (as reported by Imae and Ikeda 2007), acicular apatite (<15 μm long), and narrow veins (<5 μm) of Fe-rich alteration (Table 2, Fig. 3). Skeletal Ti-magnetite contains Ti-rich lamellae in all four meteorites, although no pure ilmenite was observed. This phase also contains 4.6–1.2 wt% Al2O3 and minimal MgO (<0.5 wt%). A second, Ti and Al-poor population of magnetite also exists within the mesostasis of these meteorites; it is commonly found near, or within, Fe-rich augite phenocryst rims. This magnetite was described by Day et al. (2006) as angular or blocky rhombohedral Ti-poor grains in MIL 03346. The mesostasis itself has a submicron granular to glassy texture, with compositions ranging from basaltic trachy-andesite to trachyte and dacite (Fig. 4). Our data indicate that MIL 090030 and 090032 have mainly trachy-andesite and trachyte compositions, whereas MIL 090136 and 03346 contain mesostasis with mainly basaltic trachy-andesite and trachy-andesite compositions. However, these variations between samples may be due to a data set limitation, as Day et al. (2006) reported trachyte and dacite mesostasis compositions in MIL 03346. MIL 03346 also contains at least one K-rich area of phonolitic composition.

Figure 3.

 Backscattered electron images of areas of mesostasis in the Miller Range nakhlites. Skeletal Ti-magnetite (Mag), interstitial fayalitic olivine (olv) and hedenbergitic clinopyroxene (cpx), cristobalite (crist), and glassy mesostasis (meso) are common to all four meteorites. Fayalite is altered to laihunite in the mesostasis of MIL 03346 (A), but laihunite was not observed in the other Miller Range nakhlites (B–D). Veins of Fe-oxide, which are probably a mixture of magnetite and hematite, are also visible in all four Miller Range nakhlites, traversing the mesostasis and running through cracks in augite. These veins are commonly observed originating from and/or passing through altered sulfide grains (B).

Figure 4.

 Plot of SiO2 versus total alkalis, illustrating the wide variation in glassy mesostasis compositions in the Miller Range nakhlites, where TB = basaltic tephrite and BTA = basaltic trachy-andesite. The data of Day et al. (2006) are represented by a gray envelope.

Secondary Alteration Phases

The olivine-bound iddingsite-like alteration veins in MIL 03346, as described by Anand et al. (2005) and Stopar (2007), are visible in the olivine phenocrysts of MIL 090030, 090032, and 090136 (Fig. 5). These are up to 50 μm wide and have elemental compositions rich in Si and Fe, with 3.0–7.2 wt% MgO (Table 2). These compositions are consistent with the olivine alteration veins of MIL 03346, which contain 3.0–7.8 wt% MgO (see also Day et al. 2006; Imae and Ikeda 2007). The olivine alteration veins in all four meteorites are also slightly enriched in Al2O3 and SO3 relative to the surrounding olivine (Table 2). Siderite was not observed in any of the Miller Range nakhlite alteration veins. However, these veins contain a second set of narrow veins (1–10 μm wide), which are brighter than the “iddingsite” in backscattered electron images, and are composed of Fe-oxide. These narrow veins are most common in MIL 090032 and 090136, where they are generally present toward the outside of the “iddingsite” olivine alteration veins (Figs. 5C–F).

Figure 5.

 Backscattered electron images of olivine-bound iddingsite-like alteration veins (olv alt veins) in MIL 03346,174 (A), 090030,23 (B), 090032,24 (C and D), and 090136,21 (E and F). Incipient olivine alteration (incip olv alt) is associated with these veins. Where well preserved, Fe-oxide is observed within the iddingsite-like alteration veins, commonly toward the edges.

A more extensive network of Fe-oxide veins is found running through augite grains and the mesostasis, commonly associated with pyrrhotite and magnetite in the latter. One of these veins within an augite phenocryst in MIL 090032,24 is wide enough for quantitative analysis and appears to consist of pure Fe-oxide (Table 2). Although these Fe-oxide veins are not wide, they are commonly extensive, running for up to 500 μm through cracks in augite grains and cross-cutting mesostasis in all four Miller Range nakhlites. In a number of cases, it is possible to trace back the origins of a single Fe-oxide vein to a magnetite or pyrrhotite grain in the mesostasis. For example, MIL 090030,23 contains an Fe-oxide vein that extends from a pyrrhotite grain which has been partially altered to magnetite and hematite (Fig. 3B). The observations of Day et al. (2006) also indicate that some pyrrhotite is altered to hematite in MIL 03346.

Areas of incipient olivine alteration are associated with olivine alteration veins in the Miller Range nakhlites (Fig. 5). Initial alteration takes the form of conical shaped etch pits in the olivine grains, which can be filled with partially altered material. As alteration continues, these etch pits join together to form a whole region of incipiently altered olivine, commonly distinguishable as a slightly darker gray material in backscattered electron images of Miller Range nakhlite olivines. Conical etch pits are a common feature of terrestrial olivine grains, and have been well documented (e.g., Velbel 2009). EMP analyses performed for this study, along with the analyses of Stopar (2007), show that the altered material in MIL 03346 has a similar composition to the surrounding olivine, with the addition of 1–4 wt% SO3, and low analytical totals (approximately 90%)—indicative of an amorphous structure and/or hydration (Table 2).

Ca-sulfate is commonly observed in MIL 03346 as veins along augite grain boundaries and cross-cutting areas of mesostasis (Fig. 6). This phase can be confidently identified as gypsum rather than anhydrite, as Raman spectra show broad H2O-related peaks around 3000 cm−1 and characteristic fluorescence. In addition, major-element compositional analyses (Table 2) show a total of approximately 80%, which increase with subsequent analyses—the EMP beam dehydrates gypsum to anhydrite over time. Gypsum content varies between individual thin-sections, and appears to decrease with increasing distance from external edges. Using the processing “family tree” for MIL 03346 (Meyer 2005; Righter and McBride 2011), in conjunction with the presence or absence of Antarctic varnish (Giorgetti and Baroni 2007), we were able to determine where the four thin-sections analyzed in this study originated from, in the parent meteorite. MIL 03346,173 and MIL 03346,174 were cut from a single, heavily terrestrially weathered, approximately 1 cm-sized chip. Areas of Antarctic varnish confirm that the outer edges of this chip were terrestrially altered. Both thin-sections show abundant gypsum veins at their edges (up to 50 μm wide), which become less common after 2–3 mm, and are narrower (10–20 μm) and sparse toward the center (Fig. 6A). Jarosite has even stronger associations with the edges of these thin-sections—none was found more than 100 μm from an external surface of the parent rock-chip. Jarosite is strongly associated with (but less abundant than) gypsum (Figs. 6B–E). It is also associated with the areas of Antarctic varnish, where veins can be up to 50 μm wide. As with gypsum, jarosite veins are found in cracks in augite grains or cross-cutting the mesostasis, but where the gypsum penetrates deeper into the thin-sections jarosite is absent. In one particular area of MIL 03346,174, a gypsum grain is surrounded by jarosite, which in turn is surrounded by a clay-like material with a composition similar to the preterrestrial olivine-bound iddingsite-like alteration veins (Fig. 6). This material shares a diffuse boundary with the mesostasis, compared to which it is depleted in K, Ca, Al, Si, and Na, but enriched in Fe and Mg.

Figure 6.

 False-color X-ray images of MIL 03346,174, where gyp = gypsum, jaro = jarosite, meso = mesostasis, mag = magnetite, AV = Antarctic varnisha dn, IG = igneous glass. A false-color image of the whole thin-section (A), where pink = Ca and yellow = S, highlights gypsum veins as white and areas of jarosite as bright yellow (relative to the mesostasis, which is a dull yellow). The intergrown nature of gypsum and jarosite is shown on a smaller scale within the area marked by the red box. In this area, backscattered electron images (B) and false-color X-ray images (C—same key as A), show numerous small veins composed of intergrown jarosite and gypsum, associated with Antarctic varnish, as well as an area of jarosite surrounding a gypsum grain (top left). Focusing on this area (D and E) reveals a terrestrially derived iddingsite-like material at the outskirts, where the mesostasis has apparently been leached of K, Ca, Al, Si, and Na. Fe and Mg appear relatively enriched in this area. The high-resolution false-color X-ray image (E) shows Fe in red, S in yellow, and K in cyan.

Only one edge of MIL 03346,126 is covered with patches of Antarctic varnish, indicating that the remainder of the section was internal to the meteorite. The thickest gypsum veins in MIL 03346,126 (approximately 20 μm) are associated with the external edge, but a number of thinner gypsum veins also run through the interior, mostly along cracks at augite grain boundaries. The overall sulfate content of MIL 03346,126 is less than that of MIL 03346,173 or 174, and jarosite is only visible in one area in the mesostasis (a small vein network extending approximately 30 μm), associated with the most extensive section of Antarctic varnish along the one externally exposed edge.

As MIL 03346,128 contains no Antarctic varnish, it appears to be a wholly internal thin-section, although its origin from parent 3 indicates that it was cut from an independent chip, rather than the main meteorite mass (Meyer 2005; Righter and McBride 2011). This thin-section contains no jarosite, and a substantially reduced amount of gypsum relative to MIL 03346, 173, 174, and 126. Gypsum is present as veins no wider than 10 μm, extending along augite cracks and grain boundaries no more than 100 μm for any given vein. These veins are not limited to the edges of the sample, but are distributed evenly throughout. An area of gypsum in MIL 03346,128 appears to have a slightly different texture—rather than a vein it is a 15 μm diameter anhedral grain of gypsum within the mesostasis. This gypsum texture was not observed within the other MIL 03346 thin-sections.

As MIL 090030,23, MIL 090032,24, and MIL 090136,21 are all internal thin-sections, they do not contain fusion crust or Antarctic varnish at their edges. They contain sulfate abundances similar to MIL 03346,128, with gypsum present as veins less than 10 μm wide and 100 μm long, extending mostly along cracks in augite grains. Jarosite was not found in these thin-sections.

Raman Spectroscopy of Alteration Phases

Olivine Alteration

One of the main Raman shift characteristics for the Miller Range nakhlite olivine alteration veins is a broad peak, or series of peaks, between 150 and 750 cm−1 (Figs. 7A–D). Peaks in the region <600 cm−1 probably correspond to clay mineral translations, although the broad nature of the 150–750 cm−1 feature makes it difficult to distinguish any specific phase. A common peak within the 150–750 cm−1 feature of all four Miller Range nakhlites occurs at 299 cm−1, and is possibly indicative of göethite. However, göethite is difficult to conclusively identify, as all its other characteristic Raman peaks also fall within the 150–750 cm−1 region, and hence, may be masked by smectite translation features. Hematite and iddingsite also have their most prominent Raman peaks within this area (Kuebler et al. 2003), which probably explains the broad and unbroken nature of the 150–750 cm−1 feature in the Miller Range nakhlite olivine alteration veins. A shoulder peak at 692 cm−1 is also common to all four Miller Range samples; this peak is indicative of the Si-O-Si bond in phyllosilicates (Wang et al. 2002). The remaining two peaks at 983 cm−1 and 1135 cm−1 are within the stretching region of nonbridging oxygen in phyllosilicate clay minerals.

Figure 7.

 Raman spectra, 200–2000 cm−1 range, of olivine alteration veins from the four Miller Range nakhlites (A–D), compared with similar olivine alteration from Nakhla (E) and Lafayette (F, adapted from Kuebler et al. 2004). Diagnostic peaks include a Si-O-Si bond peak at 692 cm−1, characteristic of phyllosilicates, two peaks at 983 cm−1 and 1135 cm−1 which are within the stretching region of nonbridging oxygen in phyllosilicate clay minerals, and a possible göethite peak at 299 cm−1.

The last feature common to all the samples analyzed, as well as previously published Raman analyses of Lafayette (Fig. 7F), (Kuebler et al. 2004), is a peak at or above 3000 cm−1. These peaks suggest that H2O is present in the alteration veins. The observed variability in the position of the H2O peaks, both within and between samples, could imply poor crystallinity in the olivine alteration veins (Kuebler et al. 2004). During Raman analysis, these veins showed variable resistance to the laser, and different sized laser pits were created in different regions of the veins. Low laser power ensured that most areas of the veins were resistant to damage, but some areas were much less resistant and large pits were created. This variable resistance indicates variable density, and hence, the heterogeneous nature of the veins and possible poor crystallinity of some regions. MIL 03346 shows a much larger H2O feature than the other nakhlites, but this may be because of the age and condition of the thin-section. MIL 03346,174 was prepared using water in 2006, and since this time, it has been at atmospheric pressure and temperature. Therefore, even if sample preparation did not add a significant amount of water to the alteration veins, they may have absorbed water from the atmosphere over the ensuing 5 yr.

At first glance the Raman spectra of olivine alteration veins in Nakhla appear similar to those in the Miller Range nakhlites (Fig. 7E). However, there are a number of differences. First, a remnant olivine double peak is still visible at 819 and 844 cm−1. In addition, the 983 and 1135 cm−1 peaks appear to be suppressed in Nakhla olivine alteration veins relative to those of the Miller Range nakhlites. The data of Kuebler et al. (2004) indicate that Lafayette has some more obvious differences in its olivine alteration veins, including a trough between the 185 cm−1 and 555 cm−1 peaks. Beyond 555 cm−1, a broad peak similar to that observed within this study is visible, although this feature has no shoulder peak at 692 cm−1. Possible nonbridging oxygen peaks are present at 964 and approximately 1100 cm−1, and a H2O peak is present at 3610 cm−1. Kuebler et al. (2004) indicated that these peaks are characteristic of a mixture of a polymerized silicate phase, göethite, and a fine-grained trioctahedral phyllosilicate (possibly vermiculite).

Raman spectra of incipiently altered olivine (Figs. 8A and 8B) show the characteristic double peak at 819 and 846 cm−1, which reduces in intensity relative to the other peaks as alteration progresses. The opposite effect is observed for the other peaks in the spectra, (e.g., at 185, 299, 602, and 935 cm−1)—these increase in intensity with increasing alteration. Therefore, there appears to be a continuum from unaltered olivine, through various stages of incipiently altered olivine, to the iddingsite-like olivine vein material. This olivine alteration progression is the same in all four Miller Range nakhlites.

Figure 8.

 Representative Raman spectra of (both lightly and heavily) incipiently altered phenocrystic olivine (A–B), gypsum (C) and jarosite (D) veins, remobilized vein (E) and primary skeletal (F) magnetite, and laihunite (G) and fayalite (H), in the Miller Range nakhlites. Gypsum shows a characteristic broad fluorescence at approximately 3000 cm−1, while jarosite’s H2O feature is more defined. The spectra of the two magnetite types have no significant differences.

Analyses of fayalite in the mesostasis of MIL 03346 revealed that the majority of these grains have been altered to the ferriolivine laihunite (Figs. 3A, 8G, and 8H). This phase has a distinctly different Raman spectra to fayalite—it shows a characteristic doublet at 572 cm−1 (rather than 819 cm−1), and other prominent peaks at 319, 788, and 902 cm−1 (see also Kuebler et al. 2006, 2011). Small cores of fayalite are still present within the laihunite grains of MIL 03346 (<10 μm diameter), but laihunite is always the most abundant of the two phases. MIL 090030,23, MIL 090032,24, and MIL 090136,21 do not contain laihunite—fayalite grains appear unaltered in these internally cut thin-sections.

Fe-Oxide Veins

These veins show the same Raman spectra as the skeletal magnetite veins, with a major peak at 672 cm−1, and minor peaks at 305 cm−1 and 554 cm−1 (Figs. 8E and 8F). Characteristic göethite peaks are not visible in the Raman spectra of these veins.


Raman analyses of jarosite and gypsum (Figs. 8C and 8D) show no evidence that these phases are mixed with each other or with any other phase on a micrometer scale, although they are associated with each other. Jarosite shows typical peaks at 233, 299, 357, 436, 569, 624, 1008, 1105, and 1154 cm−1, and a H2O peak at around 3411 cm−1. Gypsum was proved to be hydrated by the presence of a major peak at 1008 rather than 1018 (indicative of anhydrite). Characteristic fluorescence under the 532 nm (green) laser was also observed for gypsum.


The lack of bulk rock trace element and isotopic data for the three new Miller Range meteorites make it impossible to say for certain whether they are paired with the nakhlite MIL 03346. However, overwhelming textural and mineralogical similarities between these four meteorites (Figs. 1–4) indicate that they probably are all paired, and that they originated from the same basaltic cumulate on the surface of Mars. This conclusion is in agreement with the mineralogical observations of Corrigan et al. (2011), and Udry and McSween (2011).

Crystallization History of the Miller Range Nakhlites

The absence of a compositional gradient between contacting augite grains in any of the four Miller Range nakhlites (Fig. 2) indicates that these crystals were probably in contact prior to the overgrowth of later hedenbergitic rims. This implies the construction of an early clustered crystal framework in all four meteorites, as suggested by Day et al. (2006) for MIL 03346. Based on textural observations, and their relatively fayalitic composition, olivine phenocrysts in these four nakhlites appear to have crystallized after the beginning of augite crystallization, but, as they only poikilitically enclose augite grains without Fe-rich rims, they began to form prior to this later overgrowth. Where olivine and augite phenocrysts share a grain boundary, neither has an Fe-rich rim, suggesting the rims of both sets of phenocrysts formed at a similar time. As both populations of magnetite are associated with, and penetrate, these phenocryst rims (Fig. 3), they also appear to have crystallized at this time. Ti-rich skeletal magnetite is more abundant in the mesostasis, indicating Ti-enrichment of the residual melt in all four Miller Range nakhlites. The skeletal texture of these magnetite grains indicates quench cooling, after a period of slow cooling and crystallization. The high abundance of mesostasis and low SiO2 concentrations in residual glasses relative to the other nakhlites (Fig. 4), as well as the absence of albitic feldspar and relatively low abundance of unequilibrated olivine phenocrysts, also points toward fast cooling rates for the Miller Range nakhlites (see also Day et al. 2006).

Hammer and Rutherford (2005) calculated a cooling rate of 3−6 °C h−1 for MIL 03346, based on experimental crystallization of synthetic materials. These observations led Day et al. (2006) to suggest that MIL 03346 originated from above the other nakhlites in the lava flow—if indeed the nakhlites all originate from a single approximately 30 m thick flow (e.g., Lentz et al. 1999)—closest to the Martian atmosphere and just below an unsampled quenched upper crust. Alternatively, a similar cooling history could have occurred if the Miller Range nakhlites were the product of a break-out event at the front or side of the/a nakhlite lava flow, after a period of crystallization within the flow. This would explain the development of a phenocrystic crystal framework prior to quench cooling of the remaining melt (see Hon et al. (1994), and references therein, for terrestrial field studies of lava morphology and cooling rates).

Secondary Alteration in the Miller Range Nakhlites

Preterrestrial Alteration

Secondary alteration of the Miller Range nakhlites falls into two categories. The first contains the iddingsite-like olivine alteration veins and incipient olivine alteration (Fig. 5). These appear to have a preterrestrial origin, based on the previously reported cross-cutting relationships with cracks and fusion crust in other nakhlites (e.g., Gooding et al. 1991; Treiman et al. 1993; Bridges and Grady 2000; Imae et al. 2003). Our Raman data show the continuum from incipient olivine alteration to iddingsite-like olivine vein alteration (Figs. 7 and 8); therefore, the former appears to be a necessary precursor to the latter. The narrow Fe-oxide veins related to pyrrhotite and magnetite alteration also appear to be preterrestrial, based on the association of these veins with areas of iddingsite-like alteration, their lack of association with the external edges, and their similar abundance in all four meteorites. Raman spectral analyses show that these narrow veins consist of remobilized magnetite, with no indication of a göethite component (Fig. 8). A low analytical total for the sole major-element compositional analysis may indicate that these veins have a hematite component (hematite does not have a typical Raman spectrum, it simply fluoresces), but a larger data set is needed to confirm this observation. Day et al. (2006) also suggested the presence of hematite in relation to pyrrhotite alteration in the mesostasis. Pyrrhotite and magnetite, along with the olivine phenocrysts, are in disequilibrium with the mesostasis. Therefore, these phases would have been a target for any fluid-derived alteration on Mars.

Changela and Bridges (2011) reported TEM data suggesting that olivine alteration veins in Lafayette are composed of a siderite-phyllosilicate-Fe oxide-hydrated silicate gel assemblage, whereas the veins in Nakhla and Governador Valadares olivine grains predominantly contain siderite and silicate gel. This is in agreement with an earlier study by Treiman and Lindstrom (1997), suggesting that the phyllosilicate in Lafayette iddingsite formed from a silica gel. Abundant SiO2 and Fe2O3, moderate MgO, and minor Al2O3 and SO3, in the olivine alteration veins, suggest that the assemblage in the Miller Range nakhlites is composed of a heterogeneous mixture of Fe-rich phyllosilicates and fine-grained Fe-oxides. Low analytical totals—even when all compositional iron is assumed to be Fe2O3—indicate the presence of water. Therefore, Fe-oxides within this “iddingsite” phase are likely to be at least partially hydrous—probably a mixture of göethite and hematite. In anoxic environments, Fe is dissolved from olivine more readily than Mg and other alkaline metals (Hausrath and Brantley 2010). Therefore, the anoxic conditions on the Martian surface resulted in the dominance of Fe over Mg in the preterrestrial olivine alteration veins, and the depletion of Ca in the veins relative to the primary olivine grains. In addition, ferric iron phases (e.g., Fe-oxyhydroxide) can incorporate Al into their crystal structure (Hurowitz et al. 2005). Therefore, these phases are probably responsible for the slight Al-enrichment in the olivine alteration veins (Table 2). As mobilization of Al only occurs if a fluid has a low pH (Stumm and Morgan 1996), the olivine alteration veins must have formed as a result of basaltic alteration by an acidic fluid. The main source of Al was probably low-Si glass in the mesostasis of the Miller Range nakhlites, as this phase is susceptible to dissolution at low pHs. Basaltic glass dissolution at low pH would release cations in proportions approximately equal to their stoichiometric proportions in the glass (Hurowitz and McLennan 2007), and the basaltic glasses in the Miller Range nakhlites are Al-rich (Table 2). However, even very low-Si basaltic glasses do not dissolve as readily as olivine (Wolff-Boenisch et al. 2004; Hurowitz and McLennan 2007), which explains the dominance of Fe, Si, and Mg over Al in the alteration vein compositions. Elevated amounts of SO3 in the “iddingsite” olivine alteration veins and incipient olivine alteration (relative to the surrounding olivine) are probably the result of fluid enrichment via pyrrhotite dissolution. Changela and Bridges (2011) calculated that the pH of the nakhlite parent fluid would have been close to neutral/weakly acidic when it altered Lafayette (see below). Based on these calculations, and the assumption that the evaporation sequence model for nakhlite alteration assemblages is correct, this fluid could not have contained substantial SO3 prior to its evolution through the nakhlites, as SO3 content in water decreases pH.

The high SiO2 content of the Miller Range olivine alteration veins (Table 2), along with variability in the position of the H2O Raman peaks (Fig. 7), and the heterogeneous density of individual veins detected during Raman analyses, may indicate the presence of an amorphous silica gel component. Allen and Conca (1991) analyzed the chemical and mineralogical composition of etch-pit residues in basalt samples from Victoria Land, Antarctica—a dry valley with low annual surface moisture flux. They report that rapid and complete evaporation (and therefore precipitation of alteration residues) is common, due to the very small amount of water involved, and the residual heat capacity of the rocks. This rapid precipitation, coupled with low temperature, affects the crystallinity of the evaporation products—after complete evaporation the salts are crystalline, but Si, K, and Al, along with minor Mg, are left in a gel. In Antarctica, these gels are eventually converted to a mixture of illite and quartz with the aid of solar illumination (Allen and Conca 1991). However, the weaker solar radiant flux on the surface of Mars, and the fact that the nakhlites were probably located subsurface, may have made this conversion impossible, or produced only a partial conversion.

Alteration veins within Lafayette olivine phenocrysts are much wider (≤50 μm) than those found in Nakhla (≤15 μm), but are comparable to the width of the Miller Range nakhlite alteration veins (≤50 μm). Therefore, the absence of visible structure in the majority of the latter cannot be the result of sample preparation (narrower veins containing finer structures would be more susceptible to homogenization by thin-section polishing). Instead, the differing mineralogical structure and composition of olivine alteration veins in Lafayette and the Miller Range nakhlites suggest chemical differences in the fluids that altered these rocks on Mars. In addition, the presence of Fe-oxide veinlets toward the outside of the olivine alteration veins (Figs. 5C–F) indicates that the alteration fluid became progressively more oxidizing (Chevrier et al. 2007; Changela and Bridges 2011). These observations support the data of Bridges and Grady (2000), which indicate that the nakhlites were altered via a fluid evaporation sequence, where Nakhla was altered by concentrated brine that evolved from the Lafayette alteration fluid.

Using the model of Chevrier et al. (2007), Changela and Bridges (2011) calculated that the pH of the Lafayette parent fluid would have been close to neutral/weakly acidic, based on suggested partial pressures in the Amazonian (Manning et al. 2006), and the average Mg# for the phyllosilicate in Lafayette. The average Mg# for the olivine alteration veins in the internal thin-sections of MIL 090030, 090032, and 090136 is 0.2, indicating a parent fluid pH of between 3.5 and 5, using the same model. As the Miller Range olivine alteration veins contain other phases which are not phyllosilicates (e.g., Fe-oxides), the precision of this calculation remains unknown. However, the suggestion of a low-pH parental fluid is in agreement with the mobilization of Al. These observations, and the similarity between the Raman spectra of the alteration assemblages in Nakhla and the Miller Range nakhlites, indicate the latter also formed from concentrated brine. However, the probable absence of siderite, and lack of halite, suggest the parental fluid, which formed the alteration assemblages in these meteorites evolved along a slightly different pathway to the alteration fluid of Nakhla. The Miller Range parental fluid appears to have been depleted in HCO3 relative to the Nakhla parental fluid, and less concentrated (as halite only forms at the very end of the evaporation sequence). Changela and Bridges (2011) noted that a similar parental fluid formed the alteration assemblages in Yamato (Y-) 000593, which may imply that this meteorite was located close to the Miller Range nakhlites in the nakhlite parent flow(s) on Mars.

Terrestrial Alteration

The second category of alteration in the Miller Range nakhlites consists of gypsum and jarosite veins. The strong association of these veins with terrestrially exposed surfaces of MIL 03346 (Fig. 6) and their rarity (or absence in the case of jarosite) in thin-sections cut from the internal regions of all four Miller Range nakhlites, indicates that gypsum and jarosite are terrestrially derived phases in these meteorites.

Jarosite has previously been reported as an alteration mineral on the surface of micrometeorites from the Yamato ice field region in Antarctica (Osawa et al. 2003). In addition, jarosite has been detected in the Y-000593/749-paired nakhlites (Noguchi et al. 2009; Changela and Bridges 2011), and the two shergottites, Roberts Massif (RBT)-04262 (Greenwood et al. 2009) and Queen Alexandra Range 94201 (Ross et al. 2010). Textural and hydrogen isotopic data from these studies suggest that jarosite has a terrestrial origin in these meteorites. Noguchi et al. (2009) did report jarosite in the “iddingsite” olivine alteration veins, but could not determine whether this was the result of terrestrial contamination or preterrestrial alteration. Conversely, Vicenzi et al. (2007) reported a deuterium enrichment in MIL 03346 jarosite veins (δD = 2498 ± 480‰ to 2956 ± 955‰), and argued that a number of these veins were truncated by the fusion crust (although at least one vein was observed to cut through the fusion crust). In addition, McCubbin et al. (2009) analyzed a reportedly unbreached, amphibole-bearing, augite-bound melt inclusion, which they reported to contain a separate generation of preterrestrial hydrothermal jarosite. However, the cleavable structure of augite, and the lack of three-dimensional information in any thin-section indicate that the unbreached nature of this melt inclusion cannot be guaranteed. If terrestrial meteoric water did penetrate this pyrrhotite, titanomagnetite, and Fe-rich glass bearing melt inclusion, alteration of these minerals could have provided the necessary ingredients for jarosite (and hematite/göethite) formation. Hydrogen isotopic analysis may be the only way to determine the true origin of this jarosite.

Chevrier et al. (2006) studied the Ferrar dolerite of Antarctica as an analog for Martian weathering, and reported that while the presence of sulfates is probable as a product of pyrrhotite weathering, none was actually observed. (Oxy)hydroxides were the most common form of weathering observed in this lithology. Berkley (1982) reported anhydrite in basanite flows and tuffs from Ross Island, Antarctica, but did not observe sulfates in hawaiites (which are similar in composition to the nakhlites) from the same area. In fact, hawaiites were reported to contain mostly zeolite secondary alteration phases, along with some K-feldspar, calcite, and phyllosilicates. In contrast, Allen and Conca (1991) reported secondary alteration assemblages in basaltic rocks from Victoria Land, Antarctica, and found mostly illite and poorly crystalline quartz (95%), thought to be converted from silica gel (as discussed above). The remaining 5% of these assemblages consisted of sulfate salts, hematite and leucoxene particles (a poorly defined, microcrystalline form of TiO2) in varying proportions. The variability of alteration phases in different Antarctic basaltic lithologies, even those from the same locality, suggests that the alteration assemblage produced is highly dependent on rock composition. As pyrrhotite in the Miller Range nakhlites appears to have already been partially altered before it reached Earth, it follows that sulfates should form readily with the addition and subsequent evaporation of terrestrial water. Sulfur in the olivine alteration veins, and incipiently altered olivine (Table 2), may also have contributed to sulfate formation, if terrestrial water penetrated these veins prior to evaporation. Reports of phyllosilicates as a common terrestrial weathering product in Antarctic basalts suggest that the iddingsite-like material observed close to an external edge in MIL 03346,174 (Fig. 6D and 6E) has a terrestrial origin. The location of this material, as well as its association with jarosite and gypsum, also suggest that it is terrestrially derived.

Some narrow veins of gypsum are present in the internal thin-sections from the Miller Range nakhlites, and it is possible that these have a preterrestrial origin. Gypsum has been widely reported on Mars, both from orbit and surface-based exploration Rovers (e.g., Squyres et al. 2004; Bibring et al., 2005; Clark et al. 2005; McLennan et al. 2005; Poulet et al. 2005; Ming et al. 2006; Dohm et al. 2009). In addition, Nakhla (a fall meteorite) contains gypsum, at least some of which can be assumed to be preterrestrial, based on the presence of coexisting, preterrestrial halite (Bridges and Grady 1999). However, Nakhla gypsum is much less abundant than MIL 03346 gypsum, and the gypsum texture in Nakhla is dissimilar to that mostly found in the Miller Range nakhlites—it occurs mainly as anhedral grains in the mesostasis rather than veins. This may indicate that the sole gypsum grain observed in internal thin-section MIL 03346,128 is a relic preterrestrial grain, but only isotopic analysis could prove this.


Our textural and mineralogical analyses imply that the three new Miller Range nakhlites—MIL 090030, MIL 090032, and MIL 090136 are paired with each other, and with MIL 03346. These meteorites appear to have originated from the same lithological unit on the Martian surface—an augitic cumulate with rare, unequilibrated, fayalitic olivine phenocrysts, skeletal Ti-magnetites, and a quenched mesostasis. This lithology appears to have initially cooled slowly, probably within a thick lava flow (approximately 30 m based on terrestrial flow analogs, e.g., Lentz et al. 1999), encouraging the formation of a clustered crystal framework. Subsequently, an event occurred to cause greater exposure of this section of the lava to the Martian atmosphere—possibly a laval breakout event.

The olivine-bound vein assemblages in the Miller Range nakhlites appear to have been produced by an acidic Martian parental fluid, probably with a pH between 3.5 and 5. This fluid broke down the olivine structure and effectively mobilized a high proportion of the Fe, Si, and Al cations, while leaving most of the Mg and Ca behind in the residual olivine structure. This same fluid also added measurable amounts of S to the vein deposits, probably via the alteration of pyrrhotite grains. These observations indicate that the Miller Range nakhlite alteration fluid was a concentrated brine. However, the lack of halite and siderite in the alteration assemblages suggests that these meteorites were altered by a slightly different fluid to that which altered Nakhla, and may be more closely related to the Y-000593 nakhlite in terms of alteration.

Preterrestrial gypsum may also have been present in the meteorites, but was probably not abundant based on the low gypsum content of Nakhla. Once the meteorites were resident in Antarctica, terrestrial fluids produced gypsum and jarosite veins along the exposed edges of the stones, associated with areas of Antarctic varnish. These terrestrial sulfates would have overprinted any pre-existing gypsum, which was possibly redissolved by the terrestrial fluid, or is now simply indistinguishable without isotopic analysis. Preterrestrial gypsum may have survived in the central portions of the meteorites, which do not appear to have been extensively affected by terrestrial weathering.


Acknowledgments–– We thank the NASA Johnson Space Center for allocation of the Miller Range nakhlite thin-sections, and Eric Hellebrand for his assistance with our EMP analyses. We also thank the NASA astrobiology institute for funding this research. The comments and suggestions of Prof. Allan Treiman and two anonymous reviewers, as well as Associate Editor Prof. Christine Floss, were invaluable in improving this manuscript.

Editorial Handling–– Dr. Christine Floss