Elephant Moraine 96029, a very mildly aqueously altered and heated CM carbonaceous chondrite: Implications for the drivers of parent body processing

Elephant Moraine (EET) 96029 is a CM carbonaceous chondrite regolith breccia with evidence for unusually mild aqueous alteration, a later phase of heating and terrestrial weathering. The presence of phyllosilicates and carbonates within chondrules and the fine-grained matrix indicates that this meteorite was aqueously altered in its parent body. Features showing that water-mediated processing was arrested at a very early stage include a matrix with a low magnesium/iron ratio, chondrules whose mesostasis contains glass and/or quench crystallites, and a gehlenite-bearing calciumand aluminium-rich inclusion. EET 96029 is also rich in Fe,Ni metal relative to other CM chondrites, and more was present prior to its partial replacement by goethite during Antarctic weathering. In combination, these properties indicate that EET 96029 is one of the least aqueously altered CMs yet described (CM2.7) and so provides new insights into the original composition of its parent body. Following aqueous alteration, and whilst still in the parent body regolith, the meteorite was heated to 400–600 C by impacts or solar radiation. Heating led to the amorphisation and dehydroxylation of serpentine, replacement of tochilinite by magnetite, loss of sulphur from the matrix, and modification to the structure of organic matter that includes organic nanoglobules. Significant differences between samples in oxygen isotope compositions, and water/hydroxyl contents, suggests that the meteorite contains lithologies that have undergone different intensities of heating. EET 96029 may be more representative of the true nature of parent body regoliths than many other CM meteorites, and as such can help interpret results from the forthcoming missions to study and return samples from C-complex asteroids. 2016 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/ licenses/by/4.0/).


INTRODUCTION
CM carbonaceous chondrites are samples of one or more primitive bodies, probably C-complex asteroids, and are valuable sources of information about the early history of the Solar System. These meteorites contain materials that have remained unchanged since being accreted from the protoplanetary disc (crystalline and amorphous silicates, oxides, metals, sulphides, organic matter) together with a suite of secondary minerals that formed by aqueous alteration of parent body interiors (principally phyllosilicates, carbonates, and Fe-rich oxides and sulphides) (e.g., McSween, 1979;Bunch and Chang, 1980). These secondary minerals are a major constituent of the fine-grained meteorite matrix, and also occur within chondrules, calciumand aluminium-rich inclusions (CAIs), and their finegrained rims. The CM meteorites show a range of degrees of aqueous alteration, which can be described using petrographic, mineralogical, chemical and isotopic properties (McSween, 1979;Browning et al., 1996;Rubin et al., 2007;Howard et al., 2009Howard et al., , 2011Howard et al., , 2015Alexander et al., 2012Alexander et al., , 2013. The precursors to the secondary minerals are poorly understood because of the absence from our collections of unaltered CMs (i.e., that would be classified as CM3). By analogy with essentially unaltered members of other carbonaceous chondrite groups, namely the CR3 and CO3 meteorites and the ungrouped meteorite Acfer 094, the matrix and fine-grained rims of a hypothetical CM3 probably contained finely crystalline anhydrous silicates, metals and sulphides in an amorphous silicate groundmass (Brearley, 1993;Greshake, 1997;Abreu and Brearley, 2010). Chondrules and CAIs in a CM3 lithology would have contained only anhydrous minerals and glasses. Results from recent studies of several mildly altered CMs have allowed this analogy to be tested. These meteorites include Yamato (Y-) 791198 (Chizmadia and Brearley, 2008;Maeda et al., 2009), Queen Elizabeth Range (QUE) 97990 (Rubin et al., 2007;Maeda et al., 2009), and Paris (Hewins et al., 2014;Marrocchi et al., 2014;Rubin, 2014Rubin, , 2015. Amorphous silicates occur in the matrices and fine-grained rims of Y-791198 and Paris, but almost all chondrules in the three meteorites contain phyllosilicates that formed by aqueous alteration of an original glass-dominated mesostasis (Richardson, 1981). Therefore even Y-791198, QUE 97990 and Paris do not preserve the full spectrum of the most reactive and ephemeral nebular components. Some CMs were heated within their parent body and to temperatures that were sufficiently high to dehydroxylate, dehydrate and recrystallize alteration products, and recrystallize some of the primary minerals (e.g., Akai, 1990;Nakamura, 2005Nakamura, , 2006Tonui et al., 2014). There are several possible drivers of this thermal processing (Akai, 1988). Meteorite parent bodies are widely believed to have been heated by the decay of short-lived radionuclides, principally 26 Al (e.g., Grimm and McSween, 1989). Sufficient heat was produced to melt accreted ices, and so generated the water that mediated aqueous alteration. If temperatures continued to rise, aqueous alteration would have passed into thermal metamorphism. Impacts have also been proposed as a source of the heat (e.g., Nakamura, 2005). Such a driver is consistent with evidence that the thermal processing was very brief (Nakato et al., 2008), and some of the heated meteorites have been deformed by shock (Nakamura, 2006). A third possibility is that a subset of the CMs experienced radiative heating as their parent body passed close to the Sun.
Here we present evidence that the degree of aqueous alteration of the Antarctic find Elephant Moraine (EET) 96029 is equal to, or even less than, the least altered CM carbonaceous chondrites that have been described to date. This meteorite can therefore bring us closer to understand-ing the nature of the progenitor of the CMs. Our work also, however, shows that EET 96029 was significantly heated after aqueous alteration, and it has evidence for Antarctic weathering. Thus, in order to understand the history of this meteorite, and so ascertain the original composition of its C-complex parent body, it is necessary to disentangle the competing and potentially self-reinforcing effects of mild aqueous processing, later heating, and terrestrial alteration. Owing to this unusual combination of properties, EET 96029 can also provide new insights into the environments and drivers of various agents of parent body processing.

Meteorites studied
EET 96029 was recovered from Antarctica in 1996. It has a mass of 843.3 g, and a weathering grade of A/B (Grossman, 1998). The CM2 classification of this meteorite on petrologic grounds (Righter, 2010) is consistent with its reflectance spectra (Cloutis et al., 2012). According to Grossman (1998) EET 96029 is paired with 10 other Antarctic finds: EET 96005,06,07,11,12,13,14,16,17,19. The present study used three samples of EET 96029, comprising one thin section and two rock chips. The rock chips were sub-sampled for various analyses, and an additional thin section was made from one of them. The sizes of these samples and the techniques that were used to study them are listed in Table 1. Six other CM carbonaceous chondrites were also analysed in order to provide comparative data to aid in the interpretation of results from Raman spectroscopy and thermogravimetric analysis (Table 1).

Scanning electron microscopy (SEM) and X-ray microanalysis
SEM work was undertaken in the Imaging Spectroscopy and Analysis Centre (ISAAC), University of Glasgow. Backscattered electron (BSE) imaging and energydispersive X-ray analysis (EDX) of thin sections used two SEMs, a FEI Quanta and a Zeiss Sigma, both operated at 20 kV. The Quanta is equipped with an EDAX Genesis microanalysis system and the Zeiss with an Oxford Instruments AZTEC system; both acquire X-rays through silicon-drift detectors. The sizes of chondrules and CAIs were determined from BSE images. For each object two measurements were taken, the long axis and the longest dimension normal to the long axis. The size of the object is expressed as the mean of these two measurements. Point counting used the Quanta SEM whereby the stage was moved in a grid pattern and the material in the centre of the field of view was identified at each point. EDX maps with a resolution of $300 nm/pixel were acquired over a period of over $30 min using the Zeiss SEM operated at 3 nA. The protocol for quantitative analysis is described in Appendix A.

Transmission electron microscopy (TEM)
TEM work and allied sample preparation was undertaken in the Kelvin Nanocharacterisation Centre (KNC), University of Glasgow. Electron-transparent foils were manufactured using a FEI Duomill Ò dual-beam focused ion beam (FIB) instrument that is equipped with a fieldemission electron gun and a Ga + ion gun. Following the procedure of Lee et al. (2003), foils were cut from the EET 96029,9 thin section using 30 kV Ga + ions at a range of beam currents. The foils were initially milled to a thickness of $1 lm, then extracted using an in-situ micromanipulator, welded to the tines of a copper holder using electron-and ion-deposited platinum, and finally milled to a thickness of $100 nm. Bright-field images and selected area electron diffraction (SAED) patterns were obtained using a FEI T20 TEM operated at 200 kV. The foils were further ion milled to $60 nm for subsequent scanning TEM (STEM) imaging, EDX analysis and electron energy loss spectroscopy work using a JEOL ARM200cF aberration-corrected instrument that was operated at 200 kV.

Raman spectroscopy
Raman spectroscopy was undertaken at ISAAC. This technique was used to identify minerals in thin section and to determine the structure of carbon in powdered bulk samples. Carbon is useful for reconstructing CM thermal history since it typically increases in structural order with heating . To provide an internal comparison, six CMs were analysed along with EET 96029, some of which have been heated naturally (Tables 1 and  2). Raman spectra were acquired using a Renishaw inVia microscope that is equipped with a 514 nm laser. Focusing of the laser was performed through a petrographic microscope with a 100Â objective. The spectra were accumulated in 12 increments, each measured over two seconds and using 100% of the total laser power of 50 mW. The spot size was typically 1 lm. Calibration was undertaken with respect to wavenumber using a Si standard. Appendix A provides further details of data processing and interpretation.
2.5. Position sensitive detector X-ray diffraction (PSD-XRD) X-ray diffraction (XRD) work was undertaken at the Natural History Museum, London (NHM). For the initial identification of phases within a powdered fragment (8.54 mg) of EET 96029,57,21 (AK-1), an XRD pattern was collected using an INEL X-ray diffractometer with a curved 120°position sensitive detector (PSD) in a static geometry relative to the X-ray beam and sample. The sample was mounted on a low background single crystal sapphire substrate and analysed using Co Ka 1 radiation for 22 h. Silicon and silver behenate were used for calibration, and phases in the meteorite were identified using the International Centre for Diffraction Data (ICDD) database (PDF-2). Modal mineral abundances were obtained from a larger ($50 mg) chip of EET 96029,57,21 (AK-2) that was free of fusion crust. The chip was ground to <35 lm in order to minimise potential grain size artifacts, then loaded into an aluminium sample well using the sharp edge of a spatula. This packing method creates a high degree of randomness in grain orientation and reduces the effects of preferred crystal alignments (Batchelder and Cressey, 1998). For these measurements a second PSD-XRD was used with Cu Ka 1 radiation. The X-ray beam was at an incident angle of 4.2°to the flat top of the rotating sample, and the size of the beam was restricted to 0.24 Â 2.00 mm using post monochromator slits. The XRD pattern for EET 96029,57,21 (AK-2) was collected for 16 h; patterns of mineral standards were acquired under the same experimental conditions for 30 min. Differences in the incident beam flux throughout the experimental run were monitored by analysing a polished block of Fe metal at the start of each day. Details of phase quantification are in Appendix A.

Thermogravimetric analysis (TGA)
The TGA work was undertaken at the NHM using a TA instruments SDT Q600. The whole of EET 96029,57,21 (AK-1, 8.54 mg) and a 12.79 mg aliquot of EET 96029,57,21 (AK-2) were analysed. Each sample was powdered, and then stored in a glass vial within a desiccator. These samples were loaded into an alumina crucible and analysed under a N 2 flow of 100 ml min À1 . The mass of the sample was recorded as it was heated from 25 to 1000°C at a rate of 10°C min À1 . The balance sensitivity of the TGA is 0.1 lg and the overall error on the measured mass loss fraction is $0.1%. DTG curves (wt.% per°C) show the change in mass loss with heating. Peak positions as a function of temperature are characteristic of mineral groups and individual phases. Weakly bonded H 2 O (probably terrestrial contamination) is liberated up to 200°C, dehydration and dehydroxylation of Fe-(oxy) hydroxides occurs from 200 to 400°C, and dehydroxylation of phyllosilicates from 400 to 770°C. Carbonates decompose at 770-900°C King et al., 2015a).

Oxygen isotope analysis
The oxygen isotopic compositions of bulk samples of EET 96029,56-c and EET 96029,57,21 (AK-2) were determined at the Open University (OU) using a modified version of the infrared laser fluorination system described by Miller et al. (1999). Details of the technique are in Appendix A. Published analytical precision (2r) for the OU system, based on replicate analyses of international (NBS-28 quartz, UWG-2 garnet) and internal standards, is approximately ±0.08‰ for d 17 O; ±0.16‰ for d 18 O; ±0.05‰ for D 17 O (Miller et al., 1999). Changes to analytical procedures implemented after the system description in Miller et al. (1999)

Properties of the bulk meteorite
3.1.1. Mineralogy PSD-XRD results show that EET 96029,57,21 (AK-1 and AK-2) contains the following minerals (in vol.%): enstatite (15.7), olivine (12.2), magnetite (3.1), Fe-sulphide (2.4), calcite (1.0), gypsum (1.0), Fe,Ni metal (0.3) (Fig. 1). After subtracting the crystalline phases from the pattern during profile-stripping, the residual did not reduce to zero and the sum of the fit factors was less than one. This result indicates that EET 96029,57,21 has a non-crystalline constituent. The shape and intensity of the residual was in good agreement with the diffuse features of the phyllosilicate standard, so the remaining 64.7 vol.% is attributed to degraded phyllosilicates, which are characteristic of heated CM and CI chondrites (Nakamura, 2005;King et al., 2015b). The presence of amorphous nebular silicates and finely crystalline Fe-(oxy)hydroxides cannot be ruled out, although as the residual could be reduced to zero, their abundance must be low.

Abundance of water/OH
TGA was undertaken on two sub-samples of EET 96029,57,21 (AK-1 and AK-2). Over the 25-900°C range they lost 23.1% and 19.5% of their original weight (Table 3). Approximately equal amounts of water were released in the 25-200°C (the range ascribed to removal of terrestrial water) and 400-770°C steps (the range ascribed to phyllosilicate breakdown). In Table 3 these data are compared with TGA results from EET 96029 in Garenne et al. (2014), who recorded a significantly lower weight loss of 14.0% over the 25-900°C temperature range.

Oxygen isotope composition
Two replicate analyses of EET 96029,56-c and EET 96029,57,21 (AK-2) were obtained (Table 4). In an oxygen three-isotope plot these analyses lie at almost opposite ends of the range of CM compositions (Fig. 2). EET 96029 was also analysed by Tyra et al. (2007), and their data plot close to EET 96029,57,21 (AK-2) at the isotopically light end of the CM field (Fig. 2).

Properties of the matrix and fine-grained rims
Quantitative analysis of cation oxides in the fine-grained matrix gives a mean total of 83.55 ± 3.32 wt.% (Table 6; Fig. 4). These results indicate the presence of minerals containing water/OH, although other factors may contribute to lowering the totals (e.g., the presence of micropores, and the lower density of amorphous materials relative to the analytical standards). The elemental composition of this fine-grained material lies close to serpentine (Fig. 5), and is consistent with the identification of Fe 2+bearing serpentine in reflectance spectra of EET 96029 (Cloutis et al., 2012). As regards the MgO/FeO ratio, which in CM matrices varies with the degree of alteration, EET 96029,9 has a value of 0.35, as compared with 0.39 for Paris (Table 6). Another diagnostic ratio is S/SiO 2 , which is significantly lower in EET 96029,9 than Paris (0.08 versus 0.15, respectively; Table 6). TEM shows that fine grained rims contain patches of a compact material that are a few micrometres in size. These patches are composed of nanoscale fibres within an amorphous groundmass ( Fig. 6a and b). The fine grained rim material between     Clayton and Mayeda (1999). Heated CMs are from Tonui et al. (2003Tonui et al. ( , 2014, and Paris data are from Hewins et al. (2014). The oval shaded area is the CV envelope was constructed using data in Clayton and Mayeda (1999). TFL denotes the terrestrial fractionation line and CCAM the carbonaceous chondrite anhydrous mineral line.
patches contains a high density of sub-micrometre sized pores, at the edges of some of which are organic nanoglobules ( Fig. 6a and c).
The only carbonate mineral indentified is calcite, which occurs mainly as scarce grains in the matrix (Lee et al., 2014). Tyra et al. (2007) also found small quantities of Corresponding false coloured X-ray map made by blending maps for Mg (red), Si (blue) and Fe (green). In the two X-ray maps the matrix is turquoise whereas Mg-rich olivine and pyroxene grains are purple. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 5 The abundance of constituents of EET 96029,9 matrix and chondrules as determined by point counting. calcite in EET 96029. Organic carbon was identified by Raman spectroscopy, and results from EET 96029,56-b along with data from six CMs are plotted in Fig. 7. PCA 91008 has been naturally heated, and has a higher D/G (intensity) and a lower D (FWHM) than the other CMs. EET 96029,56-b has a slightly higher D/G (intensity) than five of the analysed meteorites.

Fe,Ni metal and sulphides in the matrix and chondrules
Fe,Ni metal comprises 0.33 vol.% of EET 96029,9 (Table 5). It is more abundant in chondrules than the matrix, and grains range in size from $1 to 200 lm. All of the grains that were qualitatively chemically analysed are Ni-poor (kamacite) and internally homogeneous. More Fe,Ni metal was formerly present in the chondrules but has been altered to Fe-(oxy)hydroxides, which comprise 1.40 vol.% of EET 96029,9 (Table 5). The chondrule in Fig. 8a provides an example of such alteration. It contains two grains with cores of Fe,Ni metal and concentrically layered rims that are rich in O and Fe and contain goethite, as identified by Raman spectroscopy (Fig. 8a-c). EET 96029,9 has 1.14 vol.% Fe-rich sulphide, most of which occurs in the matrix as grains $16-84 lm in size ( Table 5). The majority of these grains are composed of pyrrhotite with micrometre-sized lamellae or blebs of pentlandite ( Fig. 8d and e). Individual grains of pentlandite are absent, although two sulphide grains have average atomic Ni/(Ni + Fe) values of 0.39 and 0.41, which lie at the boundary between intermediate sulphide and pentlandite (Fig. 9).
The matrix contains rounded objects that are typically a few tens of micrometres in diameter, often contain calcite grains, and are rich in Fe, S and O (Lee et al., 2014) (Fig. 10). These 'Fe-S-O clumps' constitute 5 vol.% of EET 96029,9. Raster analyses yield element ratios that are similar to those of 'type II PCP' (serpentine intergrown with tochilinite; Tomeoka and Buseck, 1985) (Fig. 11). The clumps comprise laths that lie in a lower Z groundmass. Individual laths contain O, S, Fe and Ni (Fig. 11), and their SAED patterns index as magnetite ( Fig. 10b and c); no sulphide minerals were identified by TEM. The groundmass is rich in O, Mg, Si and Fe, and yields SAED patterns with indistinct reflections that indicate a poor crystallinity.

Chondrule mineralogy
EET 96029,9 chondrules have a mean size of 400 ± 220 lm (1r; n = 18). Of the 18 chondrules that were studied in detail, 12 are type I and six are type II. Olivine phenocrysts in the type I and II chondrules have mean compositions of Fo 98±2% (1r, n = 57 grains) and Fo 67±6% (1r, n = 67 grains), respectively. Phyllosilicates are present in the mesostasis of all type I chondrules and in all but one of the type II chondrules. The phyllosilicate-free chondrule (EETC-4) has a glass mesostasis and is described first.
3.2.3.1. Chondrule EETC-4. EETC-4 is a type IIA chondrule ( Fig. 12a and b). It contains euhedral crystals of olivine (Fo $69 ), between which is a mesostasis that has a dominant Si-rich component and a subordinate Fe-rich component. The Si-rich mesostasis principally contains O, Al, Si and Fe (Table 7; Fig. 5), and its analytical totals of 97 wt.% may reflect the presence of water. The main constituent of the Si-rich mesostasis is a glass with quench crystallites, grains of Fe,Ni metal and spherules of Fe-sulphide ( Fig. 12c and d). EDX maps show that the glass is rich in O, Na, Al and Si, whereas the quench crystallites contain O, Mg, Si, Ca and Fe. The crystallites were too small to individually analyse by EDX, but assuming that the glass has an Al/Si ratio (by weight) of 2.25, the quench crystallites have a composition of Mg 0.3 Ca 1.0 Fe 0.7 Si 2 O 6 (i.e., hedenbergite). This mineralogy was confirmed by indexing their SAED patterns. Fe-sulphide spherules range in diameter from $50 to 200 nm. One of the largest, which is 200 nm in diameter, contains O and Fe, but is free of S.
The Fe-rich mesostasis forms patches up to $10 lm in size that may enclose quench crystallites ( Fig. 12a and e). Relative to Si-rich mesostasis it is depleted in Na, Al, Si, and Ca, and enriched in Mg, S, Fe and Ni (Fig. 5, Table 7). SAED patterns of the Fe-rich mesostasis contain rings with d-spacings of 0.252, 0.205, 0.170, 0.160 and 0.145 nm, which are consistent with 6-line ferrihydrite (Janney et al., 2000). The compositional range of these patches therefore reflects the presence within the analysed volume of ferrihydrite, glass, and quench crystallites.
3.2.3.2. Phyllosilicate-bearing chondrules. In a (Si + Al)-Mg-Fe ternary diagram the phyllosilicate mesostasis of types I and II chondrules plot close to the meteorite's fine-grained matrix, thus suggesting that serpentine is abundant (Fig. 5). Mean analytical totals of 87.81 wt.% are consistent with the presence of water/OH (Fig. 4). The mesostasis contains grains with a layered microstructure that are $2-3 lm in size and lie in a finer grained groundmass (Fig. 13a). SAED patterns show that these grains are poorly crystalline and so their mineralogy could not be determined (Fig. 13a). Out of the 18 chondrules that were studied in detail, nine contain quench crystallites ( Fig. 13b and c) or pores of a similar size and shape (i.e., moulds of crystallites; Fig. 13d). These crystallites are equally as abundant in type I and type II chondrules. Narrow phyllosilicate veins cross-cut olivine phenocrysts in a small proportion of the type I chondrules. Olivine grains in two of the EET 96029,9 type II chondrules have etched margins, and their interiors contain veins of fine-grained alteration products with serrated walls. Grains of clinopyroxene in type I and II chondrules commonly contain abundant micropores and narrow cracks. These intracrystalline structures are most likely to have formed during the transformation of protoenstatite to clinoenstatite (Hanowski and Brearley, 2001).

Abundance and mineralogy of CAIs
CAIs comprise 1.8 vol.% of EET 96029,9. The 23 that were studied in detail range in size from $40 to $200 lm (mean 100 ± 48 lm, 1r; value excludes the fine-grained rim). Three morphological types of CAIs (Rubin, 2007) can be recognised: banded (61%), nodular (35%), and simple (4%). Of these 48% are intact. All of them contain aluminous spinel, 70% have pyroxene, 43% perovskite, 26% hibonite, 4% corundum and 4% contain gehlenite ( Fig. 14a and b). Chemical analysis of the gehlenite yielded a mineral formula of Ca 2.02 Al 0.74 Fe 0.02 Mg 0.13 (Al 1.00 Si 1.11 ) O 7 , and its presence is of note given the scarcity of this mineral in CAIs from other CMs Rubin, 2007). Fe-rich phyllosilicates occur in 52% of the CAIs, and occur most commonly as a layer within banded inclusions. In one of these CAIs the phyllosilicate has a hexagonal pattern of fractures that is reminiscent of desiccation cracks in terrestrial sediments (Fig. 14c). The CAIhosted phyllosilicates have been previously interpreted to have formed by aqueous alteration of melilite (e.g., gehlenite) Lee and Greenwood, 1994).

DISCUSSION
The mineralogy and composition of EET 96029 are consistent with it having undergone unusually mild aqueous alteration. This meteorite therefore has the potential to Fig. 4. Analytical totals obtained by X-ray microanalysis of various components of CM carbonaceous chondrites. The upper chart shows the mean totals obtained from fine grained rims (FGRs) of Murchison (unheated and experimentally heated to two temperatures) and the naturally heated Antarctic find B 7904 (Nakato et al., 2008). Also plotted are the matrices of Paris (Hewins et al., 2014) and Y-793321 (Nakamura, 2006), and phyllosilicate mesostases of ALH 81002 (Hanowski and Brearley, 2001), Y791198 and QUE 97990 (Maeda et al., 2009), and Paris (Hewins et al., 2014). The lower chart shows the totals obtained from EET 96029,9 fine-grained matrix (average 83.55 ± 3.32 (1r; n = 28)) and altered chondrule mesostasis (average 87.81 ± 7.18 (1r; n = 87 analyses in 16 chondrules)). For reference the analytical totals obtained from chondrule olivine grains are also plotted. reveal the nature of materials that were accreted to form its C-complex parent body but have been lost from most other CMs. Some of the meteorite's original constituents and products of aqueous alteration have also been modified by parent body heating and terrestrial weathering. The nature of changes to the original constituents of EET 96029 are summarised in Fig. 15 and discussed below.

Evaluating the degree of aqueous alteration of EET 96029
Several schemes have been proposed to quantify the degree of aqueous alteration of CM carbonaceous chondrites. Here we discuss results from EET 96029 relative to three of them with the aim of providing new insights into the nature of its parent body prior to processing by liquid water.

Phyllosilicate fraction
The phyllosilicate fraction metric of Howard et al. (2009Howard et al. ( , 2011Howard et al. ( , 2015 expresses the degree of aqueous alteration of CMs as the volume of phyllosilicates divided by phyllosilicates plus anhydrous silicates. A phyllosilicate fraction of <0.05 classifies the meteorite as type 3.0 (i.e., unequilibrated), and >0.95 as type 1.0 (i.e., completely altered) ( Table 2). Of the 25 CMs listed in Howard et al. (2015), the least altered are both type 1.6. Assuming that the 64.7 vol.% non-crystalline component of EET 96029,57,21 was originally phyllosilicate, then this meteorite had a phyllosilicate fraction of 0.69, corresponding to type 1.6. This metric therefore indicates that EET 96029 has been very mildly altered, and so it could contain nebular materials that have been lost from most other CMs.

Wt.% H in water/OH
The scheme of Alexander et al. (2013) uses chemical analyses of bulk meteorite samples and is based on wt.% H (hosted by water and OH). Their alteration scale also ranges from type 3.0 (unequilibrated) to 1.0 (completely altered). The more altered CMs have higher contents of H in water/OH owing to greater abundances of phyllosilicates. After correction for H in carbonaceous material, the analysis of EET 96029 in Alexander et al. (2012) corresponds to a type 2.0, which would make EET 96029 less aqueously altered than all but one of the 54 CM meteorites listed in Alexander et al. (2013). A caveat to this conclusion is that if EET 96029 has been heated, the associated loss of water and OH could make the meteorite appear to be less aqueously altered than it was prior to thermal processing. The impact of heating on bulk H content can be estimated from the bulk dD value of EET 96029 and by making two assumptions: (i) bulk dD does not change during heating; (ii) there is a linear correlation between bulk H and dD in unheated CMs (Alexander et al., 2013). This correction suggests that before thermal processing EET 96029 may have been a type 1.4, which is again consistent with mild aqueous alteration.

Petrologic subtype
The discussions above suggest that prior to heating EET 96029 was phyllosilicate-poor relative to many other CMs owing to unusually mild aqueous alteration. However, these conclusions are tentative because those phyllosilicates that had formed were subsequently degraded by heating. A clearer understanding of the degree of aqueous alteration of EET 96029 may come from determining its petrologic subtype (Rubin et al., 2007) (Rubin et al., 2007;Hewins et al., 2014). The former presence of serpentine in EET 96029,9 matrix is therefore consistent with a petrologic subtype of 6CM2.7, although meteorites with a subtype of >CM2.7 could still contain phyllosilicates (Rubin et al., 2007). The MgO/FeO ratio of the EET 96029,9 matrix ($0.35) is low relative to that of other mildly altered CMs. For example, Rubin (2015) found that Paris matrix has a MgO/FeO ratio of 0.39, and combining data from two other studies gave 0.42 ± 0.02. As MgO/FeO ratios should increase in step with the degree of alteration, the low value of EET 96029,9 matrix suggests that it is more primitive than that of Paris. Sulphur concentrations also vary between meteorites, with the least altered CMs having higher values. EET 96029,9 matrix has 2.2 wt.%, in comparison to Paris with 4.2 wt.%. Whilst these differences may suggest that EET 96029,9 has been more highly altered than Paris, evidence discussed below suggests that its S was lost from EET 96029,9 matrix owing to heating of sulphides.
The fine-grained rims of EET 96029,9 contain patches of a very compact and amorphous/nanocrystalline material that may be equivalent to the amorphous material of a probable nebular origin that occurs in the matrices of the mildly altered CMs Paris and Y-791198 (Chizmadia and Brearley, 2008;Hewins et al., 2014) (Table 2). The preservation of such primitive material suggests that some parts of EET 96029 have escaped aqueous alteration, and that the meteorite has a petrologic subtype equal to or greater than Y-791198 (i.e., PCM2.4). The presence within EET 96029 of calcite, but absence of dolomite, indicates a petrologic subtype of PCM2.3 (Lee et al., 2014). The characteristics of the EET 96029 matrix therefore confirm that it has been less intensely aqueously altered than many other CMs.

Nature of metal and sulphide.
Metal is a very useful indicator of the intensity of aqueous alteration of a CM carbonaceous chondrite because it is one of the most highly reactive of the primary constituents (e.g., Tomeoka and Fig. 8. Images of metal and sulphide grains in EET 96029,9. (a) BSE image of a type IAB chondrule. On the edge of the chondrule are two grains of Fe,Ni metal (white), both with concentrically layered rims. Narrow Fe-rich veins extend from these composite grains into the finegrained rim. (b) and (c) X-ray maps of one of the grains in (a) demonstrating that its rim is rich in O and Fe. Raman spectroscopy shows that the rim is composed principally of goethite. (d) and (e) BSE image and X-ray map of a sulphide grain that contains pyrrhotite (yellow) and pentlandite (purple). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Buseck, 1985). EET 96029,9 contains 0.33 vol.% Fe,Ni metal, but also has 1.4 vol.% of Fe-(oxy)hydroxide, which is interpreted to have formed by Antarctic weathering of the metal. Thus, prior to its fall to Earth, EET 96029,9 contained as much as 1.73 vol.% Fe,Ni metal. By comparison, Paris (CM2.7) has 1.2 vol.% Fe,Ni metal (Rubin, 2014). Hence, a metal abundance of 1.73 vol.% would suggest that the aqueous alteration of EET 96029,9 has been very mild, and equivalent to a petrologic subtype of PCM2.7. Although metal can also be formed by heating of carbonaceous chondrites, the chemical composition and microtexture of EET 96029,9 metal is inconsistent with such an origin.

Preservation of chondrule components.
The olivine phenocrysts and glass mesostasis of chondrules in carbonaceous chondrites are highly susceptible to aqueous alteration (e.g., Richardson, 1981). Olivine is typically replaced by veins of serpentine (Velbel et al., 2012), and such veins have been reported in meteorites of subtypes up to and including CM2.5 (Rubin et al., 2007;Lee and Lindgren, 2016). Thus, the presence of phyllosilicate veins in EET 96029 chondrule phenocrysts does not necessarily indicate a high degree of aqueous alteration.
The mesostasis of almost all chondrules in meteorites of subtypes CM2.0-2.7 contains phyllosilicates that have formed by the aqueous alteration of original glass (Rubin et al., 2007;Hewins et al., 2014). Given that chondrule EETC-4 contains glass, EET 96029 could be the first meteorite with a petrologic subtype of >CM2.7. However, EETC-4 is not unique because glass-bearing chondrules have been described from five other CMs: Murchison, Murray, Niger (I), Paris, Y-790123, and Y-75293 (Fuchs et al., 1973;Olsen and Grossman, 1978;Desnoyers, 1980;Ikeda, 1983;Metzler et al., 1992;Rubin, 2015). The compositions of some of these glasses are in Table 7 and Fig. 5a. Of these six CMs, the degree of alteration of Murchison, Murray and Paris has been studied, and all are mildly processed (Table 2). Therefore, the presence of one chondrule containing mesostasis glass is not in itself sufficient evidence to classify EET 96029 as >CM2.7, but nonetheless shows that it has been mildly altered, and at least to an equivalent level as Murchison, Murray and Paris. Glass in all of the other EET 96029,9 chondrules has been altered to phyllosilicate, and in 50% of these chondrules the mesostasis also contains quench crystallites or moulds of them. As quench crystallites also occur in the phyllosilicate mesostasis of chondrules in QUE 97990 (CM2.6; Maeda et al., 2009), their preservation may be another reliable indicator of mild aqueous alteration (i.e., a subtype of PCM2.6).
Glass and quench crystallites are likely to have been preserved in EET 96029,9 owing to the low porosity and permeability of the matrix and fine-grained rims . Porosity and permeability were probably also inhomogeneous on the sub-millimetre scale so that during early stages of aqueous alteration some chondrules were Fig. 9. Compositional range of Fe-Ni sulphides in EET 96029,9. The EDX data are plotted as atomic Ni/(Ni + Fe) ratios after Chokai et al. (2004). The fields of pyrrhotite (Py) and pentlandite (Pn) correspond to atomic Ni/(Ni + Fe) values of 0-0.10 and 0.40-0.55, respectively. Light-grey columns show raster analyses of sulphide grains containing a mixture of Ni-rich and Ni-poor areas. Medium-grey columns show spot analyses of the Ni-rich and Ni-poor areas. Dark-grey columns show spot analyses of homogeneous sulphides. exposed to much less water than others. Within these objects the components that are most susceptible to aqueous alteration (i.e., glass) have been preserved.

Mineralogy and abundance of CAIs
Rubin (2007) predicted that CAIs should be abundant in the most primitive CMs, and that some of these inclusions should also contain gehlenite. The 1.8 vol.% of CAIs in EET 96029,9 is equal to their abundance in QUE 97990 (Rubin, 2007), and considerably greater than in Paris (0.76 vol.%; Rubin, 2015), and Mighei and Cold Bokkeveld ($0.6-0.3 and $0.01 vol.%, respectively; Rubin, 2007). The abundance of CAIs is therefore consistent with a low degree of aqueous alteration for EET 96029,9. Gehlenite is highly susceptible to aqueous alteration (Nomura and Miyamoto, 1998), and is correspondingly rare in the CMs Lee and Greenwood, 1994;Rubin, 2007). This mineral has been described from Murchison (Armstrong et al., 1982;MacPherson et al., 1983;Simon et al., 2006) and Paris (Hewins et al., 2014;Marrocchi et al., 2014). The occurrence of gehlenite in EET 96029,9 therefore suggests that this sample of the meteorite has undergone an equivalent degree of alteration to the other two CMs (i.e., CM2.5-2.7) (Fig. 16).

Evidence for mild aqueous alteration of EET 96029
The properties of EET 96029 have been compared to three schemes for describing the degree of aqueous alteration of CMs, and results of these comparisons confirm that the meteorite has been very mildly processed. A caveat to this conclusion is that understanding the degree of aqueous alteration of EET 96029 is complicated by the effects of post-alteration heating and terrestrial weathering (Fig. 15). Those properties that should have been least affected by these two processes are the abundance of Fe,Ni metal, gehlenite-bearing CAIs, mesostasis glass, and mesostasis quench crystallites. Together, these properties indicate that EET 96029 has a petrologic subtype of CM2.7 and so is one of the least aqueously altered CMs yet described (Fig. 16).
Contrasts between CMs in their degree of aqueous alteration may be due to heterogeneities in the volume of water that was available. EET 96029 could have been sourced from a parent body region that had accreted relatively little water Fig. 10. A Fe-S-O clump in EET 96029,9. (a) BSE image of the clump, which has a calcite grain (grey) in its lower part. The clump contains high-Z lath-shaped grains (white) in a lower Z groundmass. (b) Bright-field TEM image of part of a foil that was extracted from the clump in (a). The midplane of this foil is delineated by ''FIB". The inset SAED pattern was acquired from the centre of the lath, using a 1 lm diameter selected area aperture, and indexes as magnetite. (c) On the left hand side is part of a SAED pattern acquired from the whole lath in (b) using a 5 lm diameter selected area aperture. The right hand side shows the spacings of reflections for polycrystalline magnetite. The good correspondence between the SAED pattern and the simulation shows that the lath contains many small crystals of magnetite. ice, or that had lost liquid water early, for example by fluid flow. Regarding the second possibility, the flow of water within meteorite parent bodies has been predicted by numerical simulations of their early evolution (Young et al., 2003;Travis and Schubert, 2005;Palguta et al., 2010). Until recently there has been little evidence from the meteorite record to support these models; work on the most primitive carbonaceous chondrites had suggested that the intergranular permeability was so low that fluid flow could not have taken place under the necessary timescales ). This constraint does not however negate fluid flow via fractures, and good evidence for the movement of aqueous solutions along such high permeability conduits has been described by Lee et al. (2013). The very mild aqueous alteration of EET 96029 therefore confirms the heterogeneous availability of liquid water within the CM parent body(ies).

Heating of EET 96029
Here we explore the effects of thermal processing on the primary and secondary constituents of a mildly aqueously altered CM (Fig. 15), constrain the temperatures attained, and speculate on the heating mechanisms.

Dehydroxylation and amorphisation of phyllosilicates
Phyllosilicates are a sensitive indicator of the intensity of heating of an aqueously altered meteorite through the loss of water and hydroxyls, and concomitant degradation of their crystal structure. Totals obtained from the EET 96029,9 fine-grained matrix and phyllosilicate-rich chondrule mesostasis ($83.6 wt.% and $87.8 wt.%, respectively) are high relative to analyses of the same components of CMs with no evidence for thermal processing (Fig. 4). Whilst recognising that analytical totals from X-ray microanalysis can reflect many properties other than just the content of water/OH, these findings are consistent with partial dehydroxylation upon heating. Temperatures can be constrained with reference to previous studies. For example, the loss of OH from phyllosilicates in Murchison finegrained rims was experimentally quantified by Nakato et al. (2008). They found that pre-heating analytical totals of 80.7 wt.% increased to 87.4 wt.% after 1 h at 600°C, and reached 93.1 wt.% after 96 h. Nakamura (2006) recorded mean totals of 92 wt.% from the matrix of Y-793321, which was inferred to have been briefly heated naturally to $500°C. By analogy with these experimental studies, EET 96029,9 was likely heated to <$500-600°C.  Table 7 Chemical compositions of the Si-rich and Fe-rich mesostasis of EETC-4, and pristine mesostasis of chondrules in other CMs.  Olsen and Grossman (1978), mean of eight analyses. c Ikeda (1983), mean of two analyses. Evidence for loss of water/OH from EET 96029 is also provided by the TGA results in Garenne et al. (2014). Their sample lost 4.5 wt.% in the 400-770°C step, which mainly represents the dehydroxylation of phyllosilicates. This value is considerably lower than the mass loss recorded over the same temperature range for meteorites with no natural heating and a similar degree of aqueous alteration . TGA results from two sub-samples of EET 96029,56,21 showed a much a greater mass loss in the 400-770°C step (mean of 8.7 wt.%, Table 3) than was recorded by Garenne et al. (2014). There are two possible explanations for these contrasting results: (i) differences between laboratories in equipment or methodology; (ii) the NHM samples had a higher degree of aqueous alteration and/or a lower intensity of heating. In order to test the first possibility five samples of Murchison were analysed at the NHM, and in Table 3 these TGA results are compared with data from Murchison in Garenne et al. (2014). The two datasets are broadly similar and so interlaboratory differences are unlikely to explain the contrasts between EET 96029 samples in weight loss over the 400-770°C range. With regards to point (ii), the same sample that was used for TGA at the NHM was also studied by PSD-XRD, and results show that it has been both mildly aqueously altered and heated. The TGA data thus show that samples of EET 96029 differ in their degree of parent body processing, probably heating, owing to the presence of clasts of different lithologies.

EETC-4 mesostasis a Mesostasis of other CMs
Petrographic evidence for dehydroxylation upon heating includes the hexagonal fracture pattern within phyllosilicates from a banded CAI, and amorphisation of phyllosilicates in the chondrule mesostasis. Estimates of the temperature at which the crystal structure of serpentine breaks down range from $400°C (Tonui et al., 2014) to 600°C (Nakato et al., 2008). Calcite grains show no evidence for degradation owing to heating (which for example may be recognised by formation of vesicles owing to CO 2 loss; Nakamura, 2006). Thus the matrix of EET 96029,9 did not reach the temperature at which calcite decomposes ($890°C, Nakato et al., 2008).

Recrystallization of sulphides
The mineralogy and microstructure of sulphides can be used to constrain heating temperatures. Tochilinite is particularly sensitive, and estimates of its breakdown temperature range from 245°C (Fuchs et al., 1973) to $300°C (Tonui et al., 2014) and 6600°C (Nakato et al., 2008). The Fe-S-O clumps in the EET 96029 matrix resemble objects composed of intergrown serpentine and tochilinite that are commonplace in unheated CMs (e.g., Lee et al., 2014). However, tochilinite is absent from Fe-S-O clumps. Instead, they comprise lath-shaped aggregates of nanoscale magnetite crystals between which are poorly crystalline/ amorphous grains of Mg-Fe silicate. Thus, the Fe-S-O clumps are interpreted to have formed by heating of serpentine-tochilinite aggregates; serpentine was amorphised to leave the Mg-Fe silicate grains, whereas tochilinite recrystallized to magnetite. Tomioka et al. (2007) described objects in experimentally shock-heated samples of Murchison that are petrographically and chemically very similar to the Fe-S-O clumps, and Tonui et al. (2014) confirmed that tochilinite breaks down to magnetite during artificial heating. The matrix of EET 96029,9 is depleted in S relative to Paris (Table 6), which is consistent with mobility of S during heating.
The properties of coarsely crystalline sulphide grains in the matrix are also consistent with thermal modification  (Kimura et al., 2011). Seven of the nine sulphides in EET 96029,9 that were analysed quantitatively are composed of pyrrhotite with lm-sized lamellae or blebs of pentlandite, and this microtexture is consistent with temperatures of 300-750°C (i.e., Category B of Kimura et al., 2011).

Structural modification of organic matter
Raman analysis of organic carbon has been used successfully to reconstruct the thermal history of terrestrial rocks (Wopenka and Pasteris, 1993;Beyssac et al., 2002;Lahfid et al., 2010) and extraterrestrial samples (Busemann et al., 2007;Beck et al., 2014;Quirico et al., 2014). This work assumes that the ordering of organic carbon increases with temperature (Beyssac et al., 2002). CM chondrites contain highly disordered carbon derived from both soluble and insoluble organic matter, and these components could have had different origins and thermal histories (Remusat et al., 2005;Quirico et al., 2014). In the present study powered samples were analysed without any prior extraction and so the Raman spectra reveal the structure of carbon in both soluble and insoluble organic matter.
Results reveal highly disordered carbon with broad D and G bands overlaying a fluorescing background. Three of the other six CMs that were studied for comparison have been heated: EET 96029, PCA 91008, QUE 93005 (Quirico et al., , 2015. Results from the present study show that PCA 91008 has been most highly heated, with a high I D /I G and a low FWHM (D) (Fig. 7). The stronger intensity of the D band (D1) relative to the G band (D2 + G) has also been recorded in low-grade metamorphic rocks from the Swiss Alps (Lahfid et al., 2010). Here increased thermal alteration led to higher I D /I G and lower FWHM (D) values, in line with results of the present study. EET 96029,56-b has the second highest I D /I G value, thus indicating that it has been heated more than the other five CMs (Fig. 7). Although these data are consistent with heating of EET 96029,56-b, they cannot be used to constrain the timing of thermal processing relative to aqueous alteration.
The data in Fig. 7 shows that the two sub-samples of LON 94101 plot in different parts of the diagram. One subsample has a relatively high I D /I G and low FWHM (D) (more heated) and the other, with a much higher FWHM (D), plots away from the main group of samples. This range is consistent with the abundance and petrological diversity of clasts within LON 94101  and underscores the point that clasts in CMs can have very different thermal histories.

Maximum temperature attained during heating
Various petrographic and mineralogical indicators enable the maximum temperature experienced by EET 96029 to be constrained. The partial dehydroxylation and amorphisation of serpentine suggests heating to >$400°C, whereas the preservation of calcite shows that it did not reach $890°C. Tonui et al. (2014) suggested that chondrules begin to integrate with the matrix so that their edges become blurred if temperatures exceed 600°C, but no such textures have been observed in the present study. These petrographic and mineralogical constraints suggest that EET 96029 was heated to between 400 and 600°C, and probably at the lower end of that range (i.e., <500°C). Nakamura (2005) divided heated carbonaceous chondrites into four stages (I to IV). Results from EET 96029 are consistent with Stage II, where serpentine and tochilinite Fig. 15. Summary of changes to the original constituents of EET 96029 in response to aqueous alteration, heating, and terrestrial weathering. Solid lines indicate complete transformation of the component whereas dashed lines denote a partial transformation. Blue lines denote aqueous alteration, red lines heating, and black lines terrestrial weathering. 1 The matrix and fine-grained rims originally contained finely crystalline anhydrous silicates, metals, sulphides, and amorphous silicates. 2 Occurs in the matrix, chondrules and fine-grained rims. 3 Occurs in CAIs. Other constituents of CAIs are not included in the diagram as they have remained unchanged (i.e., spinel, hibonite, perovskite, diopside, corundum), and chondrule pyroxene is also unaltered. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 16. The petrologic subtype of EET 96029,9. Only those properties that were most resistant to modification by heating are shown. have been amorphised. Stage II corresponds to a brief period of heating to temperatures of 300-500°C, or a longer period of heating to $200-400°C. Tonui et al. (2014) proposed that this four stage notation should be qualified using a prefix to describe the classification of the meteorite. Therefore EET 96029 would be a CM2TII (where ''T" denotes thermal metamorphism), or if the petrologic subtype were included, a CM2.7TII.

The environment and drivers of heating
Despite the considerable literature on the effects of heating on CM carbonaceous chondrites (e.g., Nakamura, 2005;Tonui et al., 2014), the drivers of this process are poorly understood. The most common suggestion is that a brief episode of heating accompanied shock, and some CMs have evidence for a commensurately short timescale of thermal processing (Nakato et al., 2008;Yabuta et al., 2010;Orthous-Daunay et al., 2013). Shock heating is also consistent with evidence for impact deformation of some of the affected CMs; for example, Y-793321 has evidence for heating, ductile deformation and brecciation (Nakamura, 2006). As Y-793321 contains high concentrations of solar wind derived 4 He, 20 Ne and 36 Ar, Nakamura (2006) concluded that the meteorite had spent a considerable period of time exposed to the impactors in a parent body regolith. The heated CM MET 01072 has also undergone impact deformation, as demonstrated by the presence of flattened chondrules and a strong petrofabric (Lindgren et al., 2015).
The lack of evidence for shock deformation of other heated CMs, including PCA 91008 (Tonui et al., 2014) and EET 96029 (the present study) may be explained by the heterogeneous passage of shock waves through parent body regoliths. The alternative is to appeal to other heat sources, namely the decay of 26 Al or solar radiation. The former possibility is inconsistent with a short duration of thermal processing identified in other heated CMs. Radiative heating is capable of producing temperatures of >550 K over periods of hours to days in those carbonaceous asteroids with perihelia of <0.15 AU (Chaumard et al., 2012). The fact that EET 96029 is a breccia containing lithologies with contrasting hydroxyl contents would be consistent with heterogeneous heating during impacts, or with impact-mixing of parts of the regolith that had been exposed to different intensities of radiative heating. The present study provides insufficient information to discriminate between these two possibilities.

Tracking aqueous alteration and heating using oxygen isotopes
EET 96029 has good evidence for having been mildly aqueously altered and subsequently heated to $400-600°C. As both processes will leave their imprint on the bulk oxygen isotope composition of a carbonaceous chondrite, EET 96029 provides a good opportunity to examine and constrain their relative roles in the evolution of the oxygen isotope system.
The three EET 96029 samples analysed span almost the entire range in oxygen isotopic compositions of the CM carbonaceous chondrites. Two of the samples lie at the isotopically light end, and within the CO-CV field (Fig. 2). Hewins et al. (2014) showed that a line regressed through d 18 O and d 17 O values of multiple sub-samples of Paris passes through other CM2 falls and the CO3 falls, thus suggesting a genetic link between the two carbonaceous chondrite groups. They also noted that analyses of Paris lie close to the CO3s, thus implying that it is quite pristine. The oxygen isotope composition of EET 96029 in Tyra et al. (2007) is very similar to that of the least altered lithology of Paris. It falls precisely on the regression line through the Paris dataset, and even closer to the CO3 falls. Therefore by using the same reasoning as Hewins et al. (2014), at least one part of EET 96029 is more primitive than Paris.
One explanation for the wide range in oxygen isotope compositions of the three EET 96029 samples is intrameteorite heterogeneity in their degree of alteration. The isotopically heavy values of EET 96029,56-c could reflect a greater degree of aqueous alteration, although the petrographic properties of the sample do not support this explanation. An alternative explanation is that EET 96029,56-c had experienced a greater intensity of heating than the other two samples. Dehydroxylation of its phyllosilicates would have been accompanied by mass-dependent isotope fractionation and loss of isotopically light water. Mayeda and Clayton (1998) showed that experimental heating of Murchison to temperatures of up to 700°C increases its d 17 O and d 18 O values, with a maximum change in d 18 O of $4‰. The 8‰ range in d 18 O values between samples of EET 96029 therefore suggests heating to a higher temperature and/or for a greater duration. We conclude that differences in isotopic compositions between the three samples of EET 96029 primarily reflect contrasts in temperatures and/or timescales of heating. Therefore the sample of EET 96029 that was analysed by Tyra et al. (2007) was pristine (i.e., with a low degree of aqueous alteration and low intensity of heating).

Pairing of EET 96029 and inter-meteorite heterogeneity
If EET 96029 is paired with the EET 96005 group (Grossman, 1998), then these other meteorites may help to provide a broader picture of the nature of their parent meteoroid. This pairing was questioned by Tyra et al. (2007) because EET 96029 differs to some of the other finds in its carbonate abundance and bulk oxygen isotope composition (Table 8). The oxygen isotope compositions of the two samples analysed for the present study show that EET 96029 is highly heterogeneous (Fig. 2), and this range is greater than the difference in oxygen isotope compositions between the four finds analysed by Tyra et al. (2007). Thus, oxygen isotope data cannot be used as evidence that EET 96029 is a different fall to the other ten meteorites, and we agree with Grossman (1998) that it is likely to be a member of the EET 96005 group.
If all of the finds listed in Table 8 are from the same meteoroid, their differences in oxygen isotope compositions underscore its internal heterogeneity. Analyses of EET 96006 and EET 96016 in Alexander et al. (2012) show that these two meteorites had not been heated after aqueous alteration. Thus, the meteoroid that delivered the EET 96005 group would have contained both heated and unheated lithologies. At least some of the heated lithologies had been very mildly aqueously altered (e.g., samples analysed for the present study). EET 96029 and its pairs can therefore provide new insights into the nature of the regolith of C-complex asteroids. In this sense the EET 96005 group would be comparable to Sutters Mill, a CM carbonaceous chondrite is a regolith breccia that contains clasts with evidence for different degrees of aqueous alteration and heating (Jenniskens et al., 2012;Beck et al., 2014).

CONCLUSIONS
1. EET 96029 is a regolith breccia. Different samples of this CM carbonaceous chondrite have contrasting water/ hydroxyl contents and bulk oxygen isotope compositions. Such intra-meteorite heterogeneity indicates that the 843 g find contains lithologies that differ in their degree of aqueous alteration and the extent to which they have lost isotopically light oxygen due to heating. Despite the lack of evidence for shock, these different lithologies are likely to have been mixed by impacts. 2. If EET 96029 is a member of the EET 96005 paired group, then it is even more apparent that the meteoroid from which all of the finds were derived contained lithologies with contrasting degrees of aqueous alteration and intensities of heating. 3. Owing to heating, the degree of aqueous alteration of EET 96029 is difficult to quantify using schemes employing water/hydroxyl contents, and phyllosilicate abundances as measured using XRD. Those properties that were most resistant to thermal modification (e.g., occurrence of Fe,Ni metal, mesostasis glass, gehlenite) indicate that EET 96029 contains very mildly aqueously altered lithologies that have a petrologic subtype of CM2.7. This conclusion is consistent with the unusually light bulk oxygen isotope compositions of two samples of EET 96029. 4. Petrographic evidence for heating of EET 96029 after aqueous alteration includes the partial dehydroxylation and amorphisation of serpentine, replacement of tochilinite by magnetite and concomitant loss of sulphur from the matrix. The maximum temperatures attained were between 400 and 600°C, and probably at the lower end of that range. EET 96029 can be classified as CM2.7TII. It remains unknown whether the agent of heating was impacts or solar radiation. 5. EET 96029 has provided new insights into the nature and history of carbonaceous chondrite parent body regoliths, and such information will be crucial in the evaluation of results from future missions to return samples of C-complex asteroids.