Mineralogical characterization of a unique material having heavy oxygen isotope anomaly in matrix of the primitive carbonaceous chondrite Acfer 094

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 72 (2008) 2723–2734 www.elsevier.com/locate/gca

Mineralogical characterization of a unique material having heavy oxygen isotope anomaly in matrix of the primitive carbonaceous chondrite Acfer 094 Yusuke Seto a,*, Naoya Sakamoto b, Kiyoshi Fujino a, Takashi Kaito c, Tetsuo Oikawa d, Hisayoshi Yurimoto a,b a Department of Natural History Sciences, Hokkaido University, Sapporo 060-0810, Japan Isotope Imaging Laboratory, Creative Research Initiative ‘‘Sousei”, Hokkaido University, Sapporo 001-0021, Japan c SII NanoTechnology Inc., Core Technology Development Department, 36-1 Takenoshita, Oyama-cho, Sunto-gun, Shizuoka 410-1393, Japan d JEOL Ltd., Electron Optics Division, 1-2 Musashino 3 Akishima, Tokyo 196-8558, Japan b

Received 19 September 2007; accepted in revised form 20 March 2008; available online 31 March 2008

Abstract We report the mineral compositions and micro-texture of the isotopically anomalous (d17,18OSMOW  +180&) Fe–S–Ni–O material recently discovered in matrix of the primitive carbonaceous chondrite Acfer 094 [Sakamoto N., Seto Y., Itoh S., Kuramoto K., Fujino K., Nagashima K., Krot A. N. and Yurimoto H. (2007) Oxygen isotope evidence for remnants of the early solar system primordial water. Science 317, 231–233]. Synchrotron radiation X-ray diffraction and transmission electron microscopy studies indicate that this material consists of the symplectitically intergrown magnetite (Fe3O4) and pentlandite (Fe5.7Ni3.3S8) with magnetite/pentlandite volume ratio of 2.3. Magnetite forms column-shaped grains (10–30 nm in diameter and 100–200 nm in length); pentlandite occurs as worm-shaped grains or aggregates of grains 100–300 nm in size between magnetite crystals. Although both the X-ray diffraction and electron energy loss spectra support identification of iron oxide as magnetite, the electron diffraction patterns show that magnetite has a weak 3-fold superstructure, possibly due to ordering of vacancies. We infer that the isotopically anomalous symplectite formed by sulfurization and oxidization of metal grains either in the solar nebula or on an icy planetesimal. The intersite cation distribution of pentlandite suggests that timescale of oxidation was no longer than 1000 years. Ó 2008 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Oxygen is cosmochemically a unique element. It is the most abundant element in solids formed in the solar system and has three isotopes (16, 17, 18) which exhibit massdependent and mass-independent fractionations, providing important constraints on the conditions during formation of solids in the early solar nebula (Clayton, 2006). The isotope fractionation scheme can be divided into two processes; mass-dependent and mass-independent. If the

*

Corresponding author. Fax: +81 11 706 4638. E-mail address: [email protected] (Y. Seto).

0016-7037/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2008.03.010

chemical reactions involving oxygen are controlled by mass-dependent kinetics and equilibrium, d18O variations are about two times larger than d17O variations, where diO is [(iO/16O)sample/(iO/16O)reference  1]  1000 and the reference usually corresponds to the standard mean ocean water (SMOW). Most terrestrial materials follow this fractionation scheme, so that d18O and d17O constitute a line with slope of 0.5 on a three-isotope oxygen diagram (Fig. 1a), which is called the terrestrial fractionation (TF) line. In contrast, oxygen isotopic compositions of meteorites are typically displaced from the TF line (Fig. 1a). Calcium–aluminum-rich inclusions (CAIs) and some chondrules from primitive meteorites are highly 16O-enriched relative to the bulk meteorites, Earth, Moon, and

2724

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

the isotope fractionation due to the conventional thermal processes, e.g., evaporation, condensation, aqueous alteration, and low-temperature chemical reactions. The variations in MIF of oxygen isotopes are generally attributed

Mars and on a three-isotope oxygen diagram plot along slope-1 line (Clayton et al., 1973; Clayton, 1993; Kobayashi et al., 2003; Jones et al., 2004). This mass-independent fractionation (MIF) of oxygen isotopes cannot be explained by

a

200

#17

CO S Sl op e1

150

CC

AM

δO17SMO W (

)

100

TF 50

chd

0

Acfer 094 CAI

-50

a006 -100 -100

-50

0

50

δO

18SMO W

100

(

150

200

)

δO O17 c

b

δO18

10μm

-50

0

50

10 μm

100

δ O17SMO W, δ O18SMO W (

150

200

)

e

d

10μm

300nm

A unique material having heavy oxygen isotope anomaly in Acfer 094

to mixing of two isotopically distinct nebular reservoirs: 16 O-rich and 17,18O-rich (Clayton, 1993). The existence of 16 O-rich reservoir in the early solar system has been inferred from isotopic compositions of high-temperature nebular condensates and unique chondrules (e.g., Krot et al., 2002; Kobayashi et al., 2003). However, little had been known about the 17,18O-rich nebular reservoir (Choi et al., 1998; Yurimoto and Kuramoto, 2004; Krot et al., 2005), until a recent discovery of an isotopically anomalous material, called ‘‘new-PCP” (Sakamoto et al., 2007) in the ungrouped carbonaceous chondrite Acfer 094. The naming of the new-PCP was in line with the rule of conventional nomenclature of materials in meteorites. The term PCP was originally used for a ‘‘poorly characterized phase” composed of Fe, Ni, S, and O in CM carbonaceous chondrites (Fuchs et al., 1973). It was subsequently shown that the PCP is a complex intergrowth of layered minerals of tochilinite and cronstedtite (Tomeoka and Buseck, 1982, 1983; Mackinnon and Zolensky, 1984). As of the discovery of the new-PCP (Sakamoto et al., 2007), the phase was not well characterized, but the chemical and oxygen isotopic compositions were clearly distinct from PCPs in CM chondrites. This is the reason why the phase was referred to as new-PCP in Sakamoto et al. (2007). In this paper, we report mineral chemistry and micro-textures of the new phase reported by Sakamoto et al. (2007) using synchrotron radiation X-ray diffraction analysis (SR-XRD) and transmission electron microscopy (TEM). Hereafter the new phase is referred to as a cosmic symplectite (COS) because the mineralogical characteristics are clarified in this paper and show a unique symplectitic texture. COS grains are typically several to 10-lm sized inclusions and ubiquitously scattered in matrices of Acfer 094 carbonaceous chondrite. The COS grains have unique chemical compositions rich in Fe, O, and S, and contains of minor amounts of Ni. Specific character of the COS grains is an extremely large oxygen isotope anomaly enriched in heavy oxygen isotopes (d17,18O180&). This is the most heavy end member of oxygen isotope anomaly along to the MIF line among chondrite components (Fig. 1). Sakamoto et al. (2007) proposed that COS grains record oxygen isotopic compositions of water in the early solar system, which was extremely enriched in heavy oxygen isotopes ever inferred (e.g., Choi et al., 1998). Thus, the COS provides the evidence that water is a primordial reservoir of heavy oxygen isotope components in the solar system. The mineralogical studies of COS can potentially clarify their formation processes and physico-chemical environments of their formation region.

2725

2. SAMPLE AND METHODS 2.1. Sample preparation The COS grains in the ungrouped carbonaceous chondrite Acfer 094 are mainly composed of Fe, Ni, O, and S. Other elements detected (Mg, Al, Si, and Ca) are below 0.1 wt%. Fe and Ni-contents are complementary to each other, keeping O/S atom ratio of about 4 (Sakamoto et al., 2007). The specimen in this study is prepared from the grain #17 described by Sakamoto et al. (2007) embedded in matrix of an Acfer 094 thin section (Fig. 1b–e). This grain has the highest Ni-content among the COS grains studied by Sakamoto et al. (2007) and is characterized by the typical enrichments of 17O and 18O isotopes (d17O = 176 ± 8 [2r]&; d18O = 178 ± 6 [2r]&). The sample was cut out from the Acfer 094 thin section by focused ion beam (FIB) technique for SR-XRD and TEM studies after data acquisition of X-ray spectra under electron microprobe. A field-emission type scanning electron microscope (FE-SEM, JEOL JSM-7000F) equipped with an energy dispersive X-ray spectrometer (EDS, Oxford INCAEnergy) has been used to analyze petrographical texture and chemical compositions on the Acfer 094 thin section. The accelerating voltage, beam current, and measurement time were maintained at 15 kV, 0.9 nA, and 100 s, respectively. For quantitative analysis, we used a ZAF matrix correction routine provided by Oxford Inc. Chemical standards were used as follows: pyrite for S, Fe-metal for Fe, Ni-metal for Ni and fluorite for Ca. For the site-specific sampling in nanometer order, FIB technique is a powerful method (e.g., Heaney et al., 2001; Lee et al., 2003). In the present study, the sample was cut out from the Acfer 094 thin section for SR-XRD and TEM studies using SMI3050TB system of SII Nano Technology. The SMI3050TB is equipped with Ga ion gun for cuttings, scanning electron microscope for in-situ observations, a fine W probe to pick up the sample, and Ar ion gun to remove damaged and contaminated layers produced during the Ga beam bombardment. Also, by spraying carbon-contained compound gas on the sample surface near the ion beam irradiation area, amorphous carbon deposition can be performed locally. The amorphous carbon deposition protects the underlying aimed sample from the Ga beam sputtering. The amorphous carbon deposition is also used as adhesives to connect the W probe and the sample. After lifted the sample out of the thin section by the W probe, the sample was mounted on a TEM grid by the amorphous carbon deposition method. The mounted substrate of TEM grid is made of a silicon single crystal. The

3 Fig. 1. (a) Oxygen isotopic compositions of the COS. The COS grains are plotted on extrapolation of slope-1 line or carbonaceous chondrite anhydrous mineral mixing (CCAM) line. COS and #17: Sakamoto et al. (2007); Acfer 094: bulk isotopic composition of Acfer 094 (Clayton and Mayeda, 1999); chd: typical chondrule of Acfer 094 (Kunihiro et al., 2005); CAI: typical 16O-rich phase of CAIs (Yurimoto et al., 2008); a006: extremely 16O-rich chondrule (Kobayashi et al., 2003). (b) d17OSMOW, (c) d18OSMOW, and (d) backscattered electron images for COS #17 embedded in Acfer 094 matrix. Surrounding materials around the COS have been plucked from the thin section and epoxy resin adhesive is exposed. TEM and SR-XRD samples were lifted out of the left and right dashed line squares, respectively. (e) High-magnification secondary image of the COS surface after 20 keV Cs+ ion bombardment. Characteristic structures consisting of wormy grains (about 200 nm in diameter) are observed.

2726

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

was needed to obtain a clear diffraction pattern (Fig. 2b). The two-dimensional X-ray diffraction patterns were converted to the conventional 2h-intensity profile, and the profiles were analyzed using software developed by Y. S. (Fig. 2c).

mounted sample was thinned by the Ga beam and finished by the Ar beam. Using this FIB system, we prepared high quality and damaged-layer free samples for SR-XRD and TEM studies. 2.2. Synchrotron radiation X-ray diffraction analysis

2.3. Microstructural analysis SR-XRD measurements were carried out in the BL13A beam line at Photon Factory, High-energy Accelerator Research Organization (PF-KEK) in Tsukuba, Japan. For SR-XRD, a 10  10  5 lm3 block was cut out from the left half part of the #17 grain (Fig. 1d) by the FIB technique. The extracted block was mounted on the top of an amorphous carbon pillar of 15-lm long and 5-lm square (Fig. 2a). The amorphous carbon pillar was deposited by the FIB on a silicon single crystal square bar of 20-lm long and 5-lm thick. The incident X-ray beam was monochro˚ and collimated matized to a wavelength of 0.42621(5) A to 15 lm in diameter. The beam size was sufficiently small to avoid X-ray scattering from the silicon square bar. The angle dispersive X-ray diffraction patterns were collected using a flat imaging plate (IP, Rigaku R-Axis4). Because the size of the sample is very small, long exposure (10 h)

A specimen for transmission electron microscopy (TEM) of 8lm  8lm  50 nm film was also cut out from the right half part of the #17 grain by the FIB technique (Fig. 1d), and was finished on both sides by argon ion milling (accelerating voltage <1 kV) to remove damaged and contaminated layers produced during the FIB cutting using the same FIB system (Fig. 3a). Microstructural and microchemical analyses were performed using three different TEMs. Microstructural observations and electron diffraction analyses were undertaken using a conventional TEM (JEOL JEM-2010) equipped with a LaB6 cathode and an EDS system (Thermo Electron Noran system SIX) of the Mineralogy Laboratory of Hokkaido University. In order to examine the distribution of elemental concentration in nanometer scale, elemental maps were acquired with a scan-

b

a

COS

carbon pillar 5 μm 044

113

c

333 115

022 004

448 008 139

246

135

226 335

135

Intensity

555 157 444 226 335008 355 137 026

224

133

024 133

004

222

022

002

333 115

222

355 137

044

113

111

111

pentlandite magnetite 4

6

8

10

12

14

16

18

20

22

2θ (degree) Fig. 2. Synchrotron radiation X-ray diffraction experiments. (a) A secondary electron image of a 10  10  5lm3 block cut from the COS #17. The block is mounted on an amorphous carbon pillar. (b) A Debye–Scherrer ring pattern. The exposure time was 10 hours. (c) An X-ray ˚ ] and pentlandite diffraction pattern converted from Fig. 2(b) The profile shows diffraction peaks of magnetite (blue) [Fd 3m, a = 8.3870(5) A ˚ ]. Each peak is assigned a lattice plane with the corresponding mineral. (For interpretation of the references to (red) [Fm 3m, a = 10.156(2) A color in this figure legend, the reader is referred to the web version of this paper.)

A unique material having heavy oxygen isotope anomaly in Acfer 094

a

2727

b

amorphous carbon

COS

matrix

2000nm

400nm

c Intensity

Fe

O

Ni

0

S

Fe

Fe Ni

Si 2

4

6

8

10

Energy(keV) Fig. 3. TEM micrographs (bright field image) and characteristic X-ray spectrum of COS. (a) Whole image of the TEM specimen made by the FIB. (b) The magnified image shows that the COS is an aggregate consisting of wormy grains (e.g., enclosed area with dashed line). The grains are elongated or oval, depending on the crosssection with the size of 100–300 nm in diameter. (c) The characteristic X-ray spectrum of a wormy grain.

ning TEM (STEM, HITACHI HD2000) and an EDS system (EDAX Genesis series) of the Open Facility of Hokkaido University. The STEM is equipped with a cold type field emission gun, where the probe size of the electron beam is below 0.3 nm. In addition, electron energy loss spectroscopy (EELS) measurements were performed using JEOL JEM-2100F with a field emission gun, and postcolumn energy filter (Gatan GIF Tridiem) of the JEOL R&D Laboratory. The EELS provides information about the atomic environment and electronic state. In this system, the full width at half maximum (FWHM) of the obtained zero-loss peak is 1.06 eV. The collected EELS spectra were deconvoluted by the zero-loss peak using the Fourier-Ratio technique (Egerton, 1996) after the background subtraction. Accelerating voltage of TEM investigations were 200 kV. 3. RESULTS AND DISCUSSION 3.1. Characterization and chemistry of the mineral assemblage Using characteristic X-ray intensity and ZAF correction method of FE-SEM-EDS analysis, it is difficult to determine accurately oxygen content of the COS phase. Moreover, because Fe, Ni, and S have different valences, the chemical composition of the COS phase cannot be uniquely

determined. Therefore, we first determined the mineral phases of the COS by the SR-XRD, and then analyzed the EDS spectra to determine the elemental composition. The X-ray diffraction pattern shows clear Debye–Scherrer rings, indicating that the COS is composed of randomly oriented aggregates of nano-sized crystals (Fig. 2b). There are no peaks arising from the amorphous carbon pillar, silicon crystal bar, or any other artificial materials in the pattern. The diffraction peaks (Fig. 2c) can be assigned to two phases: a spinel-like oxide phase [space group Fd  3m, ˚ ] and pentlandite (Fe,Ni)9S8 [space group a = 8.3870(5) A ˚ ]. It is known that spinel-structured Fm 3m, a = 10.156(2) A iron oxides exist as a magnetite (Fe3O4)-maghemite (cFe2O3) solid solution, and the cell parameter of the spinel-like phase is very close to that of pure magnetite (e.g., ˚ after Haavik et al., 2000), and deviates from a = 8.3965 A ˚ after Pecharroma´n pure maghemite (e.g., a = 8.33 A et al., 1995). On the other hand, the cell parameter of the present pentlandite is relatively larger than those reported in previous works (e.g., Hall and Stewart, 1973; Rajamani and Prewitt, 1975; Tsukimura et al., 1992). Because the studied specimen is composed of magnetite (Fe3O4) and pentlandite (Fe,Ni)9S8, we could calculate the oxygen contents and the chemical compositions of each mineral by the EDS analysis as shown in Table 1. The total weight value (100.2%) is very close to 100%. The obtained chemical formula of the magnetite and pentlandite are

2728

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

Table 1 Chemical compositions of COS #17 Weight % FeOa Fe2O3a CaO Feb Ni S Total Oxidized atom #c Fe Ca O Sulfurized atom #b,c Fe Ni S a b c

22.0 48.9 0.1 12.1 7.3 9.7 100.2

24.24 0.07 32.39 5.70 3.30 8

Assumed as molar FeO:Fe2O3 = 1:1. Assumed as (Fe,Ni)9S8. Sum of cations of oxidized atoms when S is normalized to 8.

Fe3O4 and Fe5.7Ni3.3S8, respectively. The molar ratio of magnetite to pentlandite is calculated to be 8–1, and the volume ratio is 7–3. Using the composition of each mineral, we also evaluated the ratio of the two phases based on the analysis of the XRD intensities. Structure data of magnetite and pentlandite used are from Haavik et al. (2000) and Tsukimura et al. (1992), respectively. Although the composition of the present pentlandite is slightly different from that in Tsukimura et al. (1992), the effect of the difference seems to be negligible because Fe and Ni have similar scattering factors. The inferred molar ratio of magnetite to pentlandite from the XRD data is about 12–1, which is higher than the value from the EDS data. This discrepancy may be due to crystalline states of the pentlandite; if a portion of the pentlandite exists in amorphous or as poorly crystalline state, the diffraction intensity of the pentlandite would become weaker. Thus, the volume of the pentlandite may be underestimated by X-ray diffraction. The large cell parameter of the pentlandite probably corresponds to the ordering state of Ni/Fe in the structure. The structure of pentlandite is based on a pseudo-cubic-closest packing of sulfur atom, and Fe and Ni are distributed in 4 octahedral and 32 tetrahedral sites. In nature, Fe:Ni atomic ratios of pentlandite are restricted from 1:3 to 3:1 with most natural compositions being close to 1:1 (Riley, 1977). Tsukimura et al. (1992) proposed that the Fe/ (Fe + Ni) ratio in the octahedral sites could be expressed as 3.3157z  6.7216a0 + 66.654, where z is the chemical composition expressed in Fe/(Fe + Ni) and a0 is the cell parameter in Angstroms. Using this equation, we calculated the Fe/(Fe + Ni) ratio in the octahedral sites of 0.49 for the pentlandite which is smaller than the value for natural terrestrial pentlandite (0.7) (Tsukimura et al., 1992). The intersite distributions of Fe and Ni between the octahedral and tetrahedral sites in pentlandite are sensitive to temperature and pressure (Tsukimura et al., 1992). According to the chemical composition and the Fe/

(Fe + Ni) ratio in the octahedral site, the partition coefficient (Fe/Ni)octahedral site/(Fe/Ni)tetrahedral site is calculated to be 0.513. Using the dependence of the partition coefficient on temperature and pressure (Tsukimura et al., 1992), we calculate the equilibrium temperature as 455 ± 3 K under the assumption of 0 GPa. This assumption seems to be plausible because the pressure dependence is effective under variations of GPa scale and there is no evidence of such high-pressure conditions in the Acfer 094 chondrite. The error value of 3 K, however, may be underestimated because the equation of temperature dependence was derived from the thermo-chemical data around 450 K. The realistic uncertainty could be a few tens degrees. Under constant annealing temperature of 450 K, the intersite cation distribution would achieve in equilibrium within a few days (Tsukimura et al., 1992). 3.2. Micro-texture of the COS TEM observations show that the COS has complex micro-textures. The COS is an aggregate of wormy-shaped grains 100–300 nm in diameter (Figs. 1e and 3b). Analytical TEM studies show that characteristic X-ray spectra from respective wormy grains are nearly identical to each other (Fig. 3c). An STEM–EDS mapping shows the micro-structures in individual wormy grains correspond to chemical heterogeneities (Fig. 4), where Ni and Fe distribute with S and O, respectively, and the Ni and S distributions show a strong inverse correlation with the Fe and O distributions. The Ni–S-rich and Fe–O-rich parts form a symplectitic texture. High resolution TEM (HRTEM) observations also show the characteristic texture corresponding to the elemental distributions (Fig. 5). In order to comprehend the three-dimensional texture, we observed the specimen from many directions by tilting of the specimen. The magnetite has a column like shape with 10–30 nm width and 100– 200 nm length. The lattice image indicates that the magnetite grains are arranged in the same direction. The interstices of these magnetite columns are filled with pentlandite. Because the thickness of the TEM specimen is 50 nm, these columns are overlapped with each other. Therefore, in the TEM photographs, the volume ratio of the magnetite to pentlandite seems to be larger than the value of 7–3 found by the EDS analysis (previous section). The selected area electron diffraction (SAED) of the COS is shown in Fig. 6. The size of the selected area is about 120 nm in diameter (Fig. 6a), so that a number of crystalline columns (magnetite and pentlandite) contribute to the electron diffraction pattern. Nonetheless, the pattern basically shows a single crystal pattern, rather than a polycrystalline one (Fig. 6b–d). Namely, the magnetites in each wormy grain are oriented in the same crystallographic direction, consistent with HRTEM observations. The diffraction pattern consists of strong and weak reflections. The strong reflections are arranged in the manner of facecentered cubic symmetry (i.e., indices h, k, and l are all odd or all even) and diagonal glide plane symmetry (i.e., hk0, h + k = 4n, n is an integer). Consequently, the pattern of the main strong spots is similar to that of magnetite (space group Fd  3m) and the corresponding cell parameter

A unique material having heavy oxygen isotope anomaly in Acfer 094

2729

Fig. 4. Symplectitic texture in wormy grains of the COS. (a) A bright field transmission electron image, (b) high angle annular dark-field (HAADF) image, and (c) chemical distribution images of Ni, Fe, S, and O.

˚ . But there are weak extra spots between the is a = 8.4(1) A main spots, indicating a 3-fold superstructure along all ˚ . The positions of the extra spots axes; a = b = c = 25.2 A do not agree with the spots belonging to pentlandite, or the spots produced by the multiple diffraction. The periodicity corresponding to the superstructure also appears in the HRTEM image (Fig. 5). The periodicity of 14.5 and ˚ in Fig. 5 cannot be indexed according to the normal 8.4 A ˚ seems to correspond to magnetite unit cell. (Note: The 8.4 A g = 1 0 0, which is, however, excluded by the extinction rule of the face-centered lattice.) The periodicity of 14.5 and ˚ are 3-fold as large as (1 1 1) and (2 2 0) of the normal 8.4 A unit cell, and thereby the indices are expressed as (1/3 1/3 1/ 3) and (2/3 2/3 0) in Fig. 5, respectively. We will discuss the superstructure in the next section. No electron diffraction pattern from pentlandite has been observed, although a clear XRD pattern from pentlandite has been identified (Fig. 2). The reason for the disappearance of the pentlandite diffraction spots might be due to alteration to the amorphous state induced by highenergy electron radiation. We should also note that the

present pentlandite might be partly amorphous before electron radiation as noted in Section 3.1. The fragility of pentlandite crystal structure seems to be a peculiarity of the COS. Such fragility is not usually observed for normal pentlandite crystals. Fig. 7 shows the EELS fingerprinting in the part of the iron oxide. The EELS pattern with energy resolution of <1 eV is available by the deconvolution by the zero-loss peak. Colliex et al. (1991) reported that there are four distinct peaks in the O K-edge spectra of Fe oxides, labeled A, B, C, and D, which will be present or absent depending on which of the phases are present; FeO, a-Fe2O3, c-Fe2O3, and Fe3O4. The EELS pattern of the COS shows clear peaks of A, B, C, and D. Because FeO (wu¨stite) and a-Fe2O3 (hematite) do not show a peak in position C, these phases can be excluded as being present in the COS (Fig. 7a and c). The Fe L3- and L2-edge show broad and single peaks, which are characteristic of that of magnetite (Paterson and Krivanek, 1990) as shown in Fig. 7b and d. Therefore, the EELS patterns indicate that the iron oxide in the COS shows characteristics of magnetite, consistent with the XRD results.

2730

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

Fig. 5. HRTEM images of a wormy grains in the COS. Columnarshape magnetites (10–30 nm in diameter and 50–200 nm in length) are shown (a) parallel and (b) normal to the axis of the column. Although the interstitial pentlandite between the magnetites does not contribute to diffraction contrast, the magnetites show the periodicity corresponding to superstructure of magnetite (see Fig. 6 and context).

These results indicate that column-shaped magnetites in the respective wormy grains are almost oriented in the same crystallographic direction, and pentlandite between magnetite columns seems to be amorphous for electron diffraction, whereas the X-ray diffraction shows clear diffraction peaks of pentlandite. 3.3. Superstructure of magnetite The HRTEM image (Fig. 5) and the electron diffraction patterns (Fig. 6) show that the magnetite has the 3-fold

superstructure although the X-ray diffraction (Fig. 2c) and electron energy loss spectra (Fig. 7) show that the iron oxide is not different from conventional magnetite. Therefore the superstructure might be due to the ordering of the Fe vacancies. Maghemite (c-Fe2O3), defect spinel, can be described as a spinel structure with an 8/3 defect vacancy of octahedral site per formula unit cell. Maghemite–magnetite series forms a solid solution (Fukasawa et al., 1993; Iwatsuki and Fukasawa, 1993). If the vacancy distribution is totally random, the symmetry can be treated as space group Fd  3m, which is the same with spinel. In fact, a long range ordering of the vacancies often occurs in this series, and the structure is modulated. It has been reported that the ordering of the vacancies causes 3-fold or higher order superstructure (Pecharroma´n et al., 1995; Shmakov et al., 1995; Kelm and Mader, 2005). Well-ordered maghemite shows distinct extra peaks corresponding to its lower symmetry by the conventional X-ray diffraction. Since the superstructure of the magnetite in this study was not detected by the SR-XRD, the degree of the structural modulation is likely to be small. The basic unit cell parameter of the present magnetite is closer to that of pure magnetite rather than to that of pure maghemite, suggesting that the amount of the vacancies is relatively small and thereby the contribution to the elastic scattering of the extra spots is also small. The symmetries of maghemite are considered to be progressively changed in steps as Fd  3m, P4132 and P43212 (Pecharroma´n et al., 1995). Although the structure of the P43212 maghemite is modulated with a unit cell three times as large in the c direction as that of an original spinel structure, the magnetite in the COS shows the extra spots corresponding to the 3-fold periodicity along the all a, b, c axes, not one axis. Such modulated magnetite has not been reported so far. According to the group theory, an ordering of crystallographic sites causes a change of a space group to its subgroups. In the present case, one of the possible subgroups for the modulated magnetite is Fd  3m with a 3  3  3 unit cell, which is a maximal isomorphic subgroup of Fd  3m. Thorough the transition from Fd  3m to 3  3  3 Fd  3m, one crystallographic equivalent octahedral site splits into six different sites. In other words, the 3  3  3 Fd  3m structure will appear when the vacancies are present in one or some (not all) of the six different sites. Kinematical electron diffraction simulations (EMS On Line, http://cecm.insa-lyon.fr/CIOL/) suggest that the 3  3  3 Fd  3m model is qualitatively consistent with the extra spots (Fig. 6c and d). We should note that this model is one of many possible models, and we cannot determine the exact structure without higher quality of the diffraction data than those in this study. This unique ordering may provide a novel constraint for the environments of COS formation in the future. 3.4. Formation process of the COS Mineralogical and thermodynamic analyses suggest that the COS resulted from oxidation of iron metal and/or iron sulfide by nebular water (Sakamoto et al., 2007). According to the oxygen fugacity expected in the solar nebula with so-

A unique material having heavy oxygen isotope anomaly in Acfer 094

2731

Fig. 6. Electron diffraction patterns of a wormy grain of the COS. (a) The grain in the center of the image was used for selected area diffraction. A dashed circle corresponds to the size of the selected area. (b) The electron diffraction patterns from the three directions, [1 1 1], [1 0 1], and [0 0 1]. All indices are labeled according to the normal magnetite symmetry. There are weak extra spots between strong spots, indicating a 3-fold superstructure of magnetite along all axes. (c) Simulated diffraction patterns for normal magnetite symmetry. (d) Simulated diffraction patterns for the superstructured magnetite proposed in the present study. Diffraction intensities (corresponding to the spot size) are calculated by kinematical diffraction theory (not dynamical diffraction theory).

lar composition, the stability field of magnetite is below 360 K (Herndon et al., 1975; Sakamoto et al., 2007). Using the compositions of the COS and of the constituents, we can derive the net chemical reactions as follows: 33ðFe0:90 Ni0:10 Þ½metal þ 32H2 O½gas þ 8H2 S½gas ¼ 8Fe3 O4½magnetite þ ðFe0:63 Ni0:37 Þ9 S8½pentlandite þ 40H2½gas Ni is preferably distributed to the sulfide phase, while Fe is preferably distributed to the oxide phase. Such preference

of Ni distribution have been reported in Herndon et al. (1975); as oxidation progresses, the sulfide mineral becomes iron deficient, and nickel remains in the sulfide. Thus, the metal sulfide becomes increasingly rich in nickel, while the magnetite has an almost pure composition of Fe3O4. This reaction, however, may not fully correspond to each step of the COS formation process because of the complex submicron to nano-scale textures observed in the COS. The texture of COS is divided broadly into two hierarchies: (1) texture composed of wormy grains having symplectitic

2732

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

a

b

O K-edge

Fe L-edge

B A C

D

L3 500

520 540 560 Energy Loss (eV)

L2

705 710 715 720 725 Energy Loss (eV)

580

d

c O K-edge

Fe L-edge

D B C

A

FeO

α-Fe2O3

D C

B

α-Fe2O3

A

D

γ-Fe2O3

C

γ -Fe2O3

B A B

Fe3O4 500

A

D

Fe3O4

C

520 540 560 Energy Loss (eV)

L3 580

L2

705 710 715 720 725 Energy Loss (eV)

Fig. 7. EELS spectra. (a) An O K-edge spectrum of the COS. (b) An Fe L-edge spectrum of the COS. (c) O K-edge spectra for several major iron oxides (after Colliex et al., 1991). (d) Fe L-edge spectra for several major iron oxides (after Paterson and Krivanek, 1990).

inner structure and (2) the symplectitic texture. These textural hierarchies should correspond to the formation process. Iron sulfide in the solar nebula is believed to be formed by sulfurization of pre-existed iron metal (Lauretta et al., 1996). The iron metal had been condensed in hot gas of the solar composition as Fe alloy (Fegley, 2000). If cooling of the iron metal is fast, the metal transforms to martensite. Kimura et al. (2006) suggest that most iron metal in Acfer 094 is martensite indicating rapid cooling from high-temperature, higher than 900 K. The wormy cell texture of COS might correspond originally to martensite texture because martensite shows acicular texture. The iron metal was corroded to form iron sulfide by H2S gas below 700 K in the solar nebula (Lauretta et al., 1996). Kinetics and mechanisms of sulfurization of iron in the solar nebula have been experimentally investigated (Lauretta et al., 1998). The sulfurization starts with nucleation of iron sulfide on the metal surface and forms the sulfide layer having a preferred orientation of the crystals. This texture may be different from the precursor texture inferred from the martensite texture. However, because each crystal sizes of the martensite and COS sizes are much smaller than the crystal sizes used in the experimental study (Lauretta et al., 1998), textures after sulfurization may depend on the sizes of precursor metal grains. Multiple heating as to cause evaporation, melting, condensation, and solidification of materials commonly occurred in the solar nebula (Davis and Richter, 2003). When iron sulfide evaporates in the solar composition gas, iron metals with spongy texture were formed as an evaporate residue resulting from incongruent evaporation (Tsuchiyama et al., 1997; Tachibana and Tsuchiyama, 1998). If such spongy iron metal was corroded to form iron

sulfide, the wormy texture may be developed by the sulfurization. Under high-temperature sulfurization, the kinetics proposed by Lauretta et al. (1998) would control the texture. As a result, preferred oriented monosulfide solid solution (mss, [Fe,Ni]1xS) with exsolved pentlandite is developed. However, such high-temperature condition was not continued for the long time during the COS formation because the size of mss grains should be smaller than that of the wormy grains (200 nm in diameter). Assuming experimental grain growth rates on iron sulfurization (Lauretta et al., 1998), the high-temperature annealing duration would be shorter than days. The Ni solubility in mss decreases with temperature decreases (Mirsa and Fleet, 1973; Naldrett, 2004), and the excessive Ni is exsolved as pentlandite. The micro-structure in wormy grains is formed as a lamellar texture under this stage. As a sequence, the mineral assemblage by 360 K is troilite (pyrrhotite), pentlandite, and ±metals. Most pentlandite grains observed in COS (Fig. 5) would form by this stage. Below 360 K, metals and sulfides were oxidized to magnetite by water (Urey, 1952; Fegley, 2000; Sakamoto et al., 2007). The pentlandite survived the oxidation. As a result, the exsolution textures of troilite–pentlandite assemblage in the wormy grain changed to a symplectitic texture of magnetite–pentlandite assemblage. The sulfur fugacity is controlled by the magnetite–pentlandite buffer. Kinetics of intersite cation distribution of pentlandite constrain the time scale of the cooling (Tsukimura et al., 1992). The intersite distribution of the pentlandite suggests that the closure temperature of the sulfurization was 450 K indicating that the sulfurization period at the temperature was longer than several days. If the oxidization occurred after the sulfurization in the single cooling event, the oxidization would be completed within thousands years, because the typical equilibrium time scale of the intersite cation distribution at 360 K is 103 years (Tsukimura et al., 1992). 3.5. Astrophysical setting of COS formation The processes discussed in Section 3.4 may be realized in the solar nebula. Some of the iron metal grains formed by condensation from gas or by evaporation of iron sulfide in the high-temperature region inferred in the inner nebula (Cameron, 1995; Willacy et al., 1998) would be transported outwards by turbulent radial mixing (Bockele´e-Morvan et al., 2002; Ciesla, 2007). The temperature becomes gradually lower as the transportation and the metal grains are sulfurized to form iron sulfides. The time scale of turbulent radial mixing (104 years) is sufficient for the sulfurization of metal grains in the solar nebula (Lauretta et al., 1996) and consistent to the sulfurization period inferred by the intersite cation distribution kinetics of pentlandite. The iron sulfides are subsequently oxidized to form magnetite by water vapor that is migrating inwards from the outer solar nebula. The oxidization temperature suggests that the oxidization place would be inside the snow line of the solar nebula. The time scale of turbulent radial mixing is sufficient to oxidize the sulfide and the metal inferred

A unique material having heavy oxygen isotope anomaly in Acfer 094

by intersite cation distribution of pentlandite (<103 years). The turbulent mixing time may be comparable to a time scale inferred by magnetite formation kinetics from iron determined experimentally (Hong and Fegley, 1998). However, under the solar nebular conditions, the rate-limiting step may be supply of water molecules to the sulfide and metal. A simple collision theory modeling suggests that the magnetite formation time in the solar nebula is much longer than the turbulent radial mixing time (Fegley, 1988). In order to increase the oxidation rates, temporal pressure increases by shock waves in the solar nebula may be effective (Ciesla et al., 2003). Further experimental and theoretical studies are needed to assess the COS formation in the solar nebula. Alternatively the oxidization place to form COS may be on the parent planetesimal. Aqueous alteration on planetesimals is widely observed in carbonaceous chondrites although the degree is highly variable (Brearley, 2003). The oxygen isotopic exchange between water and rock is relatively fast, while the secondary minerals take time to grow (Young, 2001, 2002). Thus, the secondary hydrous minerals tend to grow from waters that had already exchanged oxygen isotopes with primary minerals. The observed oxygen isotopic compositions of secondary minerals of CM chondrites (e.g., Choi et al., 1998; Clayton and Mayeda, 1999) correspond to the products from such exchanged water. CM chondrites show intense aqueous alteration. The alteration temperature inferred by oxygen isotope fractionations of minerals ranges from <290 to 320 K (Clayton and Mayeda, 1984; Leshin et al., 1997; Clayton and Mayeda, 1999; Benedix et al., 2003; Guo and Eiler, 2007). The temperature is comparable to or slightly lower than the oxidation temperature of COS formation temperature inferred in this study (<360 K). Therefore, if the COS formed on the parent planetesimal, the aqueous alteration had completed before hydrous mineral formation. There are no obvious mineralogical and petrographical evidence of aqueous alteration of Acfer 094 (Greshake, 1997), suggesting that water in Acfer 094 had been lost before evolving aqueous alteration, i.e., the aqueous alteration of Acfer 094 stopped at the very initial stage. Such environments would be plausible conditions for COS formation and be generally achieved at the shallow part of mildly heated small icy planetesimal. However, further studies are needed to assess the COS formation on the parent planetesimal because it is unclear whether chemical reaction rates in aqueous solution or aqueous vapor are faster for the iron sulfide and the iron metal than for the primary silicates on mildly heated small icy planetesimal. ACKNOWLEDGMENTS We thank I. Nakatani for assistance of FIB sample preparation, A. Yasuhara for assistance of EELS experiments, and A.N. Krot, S. Itoh, K. Kuramoto, and A. Meibom for fruitful discussion. We also thank L. Nittler and T. Zega for constructive review and A.N. Krot for editorial handling. X-ray observations were conducted at Photon Factory, Japan (BL13A, Proposal No. 2005G143). This research was supported by a Monka-sho Grant to Y.H., and by a Monka-sho Grant for the 21st Century COE

2733

Program on ‘‘Neo-Science of Natural History” at Hokkaido University (Y.S. and K.F.). Y.S. and N.S. are supported by JSPS research fellowship.

REFERENCES Benedix G. K., Leshin L. A., Farquar J., Jackson T. and Theimens M. H. (2003) Carbonates in CM2 chondrites: constraints on alteration conditions from oxygen isotopic compositions and petrographic observations. Geochim. Cosmochim. Acta 67, 1577–1588. Bockele´e-Morvan D., Gautier D., Hersant F., Hure J.-M. and Robert F. (2002) Turbulent radial mixing in the solar nebula as the source of crystalline silicates in comets. Astron. Astrophys. 384, 1107–1118. Brearley A. J. (2003) Nebular versus parent-body processing. In Meteorites, Comets and Planets, vol. 1 (ed. A. M. Davis) (eds. H. D. Holland and K. K. Turekian). Elsevier-Pergamon, Oxford, pp. 247–268. Cameron A. G. W. (1995) The first 10-million years in the solar nebula. Meteoritics 30, 133–161. Ciesla F. J. (2007) Outward transport of high-temperature materials around the midplane of the solar nebula. Science 318, 613– 615. Ciesla F. J., Lauretta D. S., Cohen B. A. and Hood L. L. (2003) A nebular origin for chondritic fine-grained phyllosilicates. Science 299, 549–552. Choi B. G., McKeegan K. D., Krot A. N. and Wasson J. T. (1998) Extreme oxygen-isotope compositions in magnetite from unequilibrated ordinary chondrites. Nature 392, 577–579. Clayton R. N. (1993) Oxygen isotopes in meteorites. Ann. Rev. Earth Planet. Sci. 21, 115–149. Clayton R. N. (2006) Meteorites and their parent asteroids. Science 313, 1743–1744. Clayton R. N., Grossmann L. and Mayeda T. K. (1973) A component of primitive nuclear composition in carbonaceous chondrite. Science 182, 485–488. Clayton R. N. and Mayeda T. K. (1984) The oxygen isotope record in Murchison and other carbonaceous chondrites. Earth Planet. Sci. Lett. 67, 151–161. Clayton R. N. and Mayeda T. K. (1999) Oxygen isotope studies of carbonaceous chondrites. Geochim. Cosmochim. Acta 63, 2089– 2104. Colliex C., Manoubi T. and Ortiz C. (1991) Electron-energy-lossspectroscopy near-edge fine structures in the iron–oxygen. Phys. Rev. Lett. B 44, 11402–11411. Davis A. M. and Richter F. M. (2003) Condensation and evaporation of solar system materials. In Meteorites, Comets and Planets, vol. 1 (ed. A. M. Davis) (eds. H. D. Holland and K. K. Turekian). Elsevier-Pergamon, Oxford, pp. 407–431. Egerton R. F. (1996) Electron Energy-Loss Spectroscopy in the Electron Microscope. Plenum Press, New York. Fegley B., Jr. (1988) Cosmochemical trends of volatile elements in the solar system. In Workshop on the Origins of Solar Systems (eds. J. A. Nuth and P. Sylvester). L PI Technical Report No. 88-04, Lunar and Planetary Inst., Houston, Texas. Fegley, Jr., B. (2000) Kinetics of gas-grain reactions in the solar nebula. Space Sci. Rev. 92, 177–200. Fuchs L. H., Olsen E. and Jensen K. J. (1973) Mineralogy, mineralchemistry, and composition of the Murchison (C2) meteorite. Smithson. Contrib. Earth Sci. 10, 1–39. Fukasawa T., Iwatsuki M. and Furukawa M. (1993) State analysis and relationship between lattice constants and compositions including minor elements of synthetic magnetite and maghemite. Anal. Chim. Acta 281, 413–419.

2734

Y. Seto et al. / Geochimica et Cosmochimica Acta 72 (2008) 2723–2734

Greshake A. (1997) The primitive matrix components of the unique carbonaceous chondrite Acfer 094: a TEM study. Geochim. Cosmochim. Acta 61, 437–452. Guo W. and Eiler J. M. (2007) Temperatures of aqueous alteration and evidence for methane generation on the parent bodies of the CM chondrites. Geochim. Cosmochim. Acta 71, 5565–5575. Haavik C., Stolen S., Fjellvag H., Hanfland M. and Hausermann D. (2000) Equation of state of magnetite and its high-pressure modification: thermodynamics of the Fe–O system at high pressure. Am. Mineral. 85, 514–523. Hall S. R. and Stewart J. M. (1973) The crystal structure of argentian pentlandite (Fe,Ni)8AgS8, compared with the refined structure of pentlandite (Fe,Ni)9S8. Can. Mineral. 12, 169–177. Heaney P. J., Vicenzi E. P., Giannuzzi L. A. and Livi K. J. T. (2001) Focused ion beam milling: a method of site-specific sample extraction for microanalysis of Earth and planetary materials. Am. Mineral. 86, 1094–1099. Herndon J. M., Rowe M. W., Larson E. E. and Watson D. E. (1975) Origin of magnetite and pyrrhotite in carbonaceous chondrite. Nature 253, 516–518. Hong Y. and Fegley, Jr., B. (1998) Experimental studies of magnetite formation in the solar nebula. Meteor. Planet. Sci. 33, 1101–1112. Iwatsuki M. and Fukasawa T. (1993) Nondestructive and complete analysis of magnetite–maghemite solid solutions by a combined X-ray diffraction/fluorescence method. Anal. Sci. 9, 95–98. Jones R. H., Leshin L. A., Guan Y., Sharp Z. D., Durakiewicz T. and Schilk A. J. (2004) Oxygen isotope heterogeneity in chondrules from the Mokoia CV3 carbonaceous chondrite. Geochim. Cosmochim. Acta 68, 3423–3438. Kelm K. and Mader W. (2005) The symmetry of ordered cubic cFe2O3 investigated by TEM. Z. Naturforsch. B 61, 665–671. Kimura M., Weisberg M. K. and Grossman J. N. (2006) Fe–Ni metal and sulfides in Acfer 094: thermal history of the most primitive chondrite. Meteor. Planet. Sci. 41, A96. Kobayashi S., Imai H. and Yurimoto H. (2003) New extreme 16Orich reservoir in the early solar system. Geochem. J. 37, 663– 669. Krot A. N., Hutcheon I. D., Yurimoto H., Cuzzi J. N., McKeegan K. D., Scott E. R. D., Libourel G., Chaussidon M., Aleon J. and Petaev M. I. (2005) Evolution of oxygen isotopic composition in the inner solar nebula. Astrophys. J. 622, 1333–1342. Krot A. N., McKeegan K. D., Leshin L. A., MacPherson G. J. and Scott E. R. D. (2002) Existence of an 16O-rich gaseous reservoir in the solar nebula. Science 295, 1051–1054. Kunihiro T., Rubin A. E. and Wasson J. T. (2005) Oxygen-isotopic compositions of low-FeO relicts in high-FeO host chondrules in Acfer 094, a type 3.0 carbonaceous chondrite closely related to CM. Geochim. Cosmochim. Acta 69, 3831–3840. Lauretta D. S., Kremser D. T. and Fegley, Jr., B. (1996) The rate of iron sulfide formation in the solar nebula. Icarus 122, 288–315. Lauretta D. S., Lodders K. and Fegley, Jr., B. (1998) Kamacite sulfurization in the solar nebula. Meteor. Planet. Sci. 33, 821– 833. Lee M. R., Bland P. A. and Graham G. (2003) Preparation of TEM samples by focused ion beam (FIB) techniques: applications to the study of clays and phyllosilicates in meteorites. Mineral. Mag. 67, 581–592. Leshin L. A., Rubin A. E. and McKeegan K. D. (1997) The oxygen isotopic composition of olivine and pyroxene from CI chondrites. Geochim. Cosmochim. Acta 61, 835–845.

Mackinnon I. D. R. and Zolensky M. E. (1984) Proposed structures for poorly characterized phases in C2M carbonaceous chondrite meteorites. Nature 309, 240–242. Mirsa K. C. and Fleet M. E. (1973) The chemical composition of synthetic and natural pentlandite assemblage. Econ. Geol. 68, 518–539. Naldrett A. J. (2004) Magmatic Sulfide Deposits: Geology, Geochemistry and Exploration. Springer Verlag, Heidelberg, Berlin. Paterson J. H. and Krivanek O. L. (1990) ELNES of 3d transition metal oxides II. Variations with oxidation state and crystal structure. Ultramicroscopy 32, 319–325. Pecharroma´n C., Gonza´lez-Carren˜o T. and Iglesias J. E. (1995) The infrared dielectric properties of maghemite, c-Fe2O3, from reflectance measurement on pressed powders. Phys. Chem. Miner. 22, 21–29. Rajamani V. and Prewitt C. T. (1975) Thermal expansion of the pentlandite structure. Am. Mineral. 60, 39–48. Riley J. F. (1977) The pentlandite group (Fe,Ni,Co)9S8; new data and an appraisal of structure–composition relationships. Mineral. Mag. 41, 345–349. Sakamoto N., Seto Y., Itoh S., Kuramoto K., Fujino K., Nagashima K., Krot A. N. and Yurimoto H. (2007) Oxygen isotope evidence for remnants of the early solar system primordial water. Science 317, 231–233. Shmakov A. N., Kryukova G. N., Tsybulya S. V., Chuvilin A. L. and Solovyeva L. P. (1995) Vacancy ordering in c-Fe2O3: synchrotron X-ray powder diffraction and high-resolution electron microscopy studies. J. Appl. Crystallogr. 28, 141–145. Tachibana S. and Tsuchiyama A. (1998) Incongruent evaporation of troilite (FeS) in the primordial solar nebula: an experimental study. Geochim. Cosmochim. Acta 62, 2005–2022. Tomeoka K. and Buseck P. R. (1982) An unusual layered mineral in chondrules and aggregates of the Allende carbonaceous chondrite. Nature 299, 327–329. Tomeoka K. and Buseck P. R. (1983) A new layered mineral from the mighei carbonaceous chondrite. Nature 306, 354–356. Tsuchiyama A., Uyeda C. and Makoshi Y. (1997) Incongruent evaporation experiments on iron sulfide (Fe1dS) under H2-rich (at 1 atm) and evacuated conditions. Geochem. J. 31, 289–302. Tsukimura K., Nakazawa H., Endo T. and Fukunaga O. (1992) Cation distribution in pentlandites (Fe,Ni)9S8: dependence of pressure and temperature and kinetics of the cation exchange reaction. Phys. Chem. Miner. 19, 203–212. Urey H. C. (1952) The Planets: Their Origin and Development. Yale University Press, New Haven, p. 245. Willacy K., Klahr H. H., Millar T. J. and Henning T. (1998) Gas and grain chemistry in a protoplanetary disk. Astron. Astrophys. 338, 995–1005. Young E. D. (2001) The hydrology of carbonaceous chondrite parent bodies and the evolution of planet progenitors. Phil. Trans. R. Soc. Lond. A359, 2095–2110. Young E. D. (2002) Oxygen isotopic compositions of water ice within carbonaceous chondrite parent bodies are irretrievable. Meteor. Planet. Sci. 37, A153. Yurimoto H., Krot A. N., Choi B.-G., Aleon J., Kunihiro T. and Brearley A. J. (2008) Oxygen isotopes of chondritic components. Rev. Miner. Geochem. 68, 141–186. Yurimoto H. and Kuramoto K. (2004) Molecular cloud origin for the oxygen isotope heterogeneity in the solar system. Science 305, 1763–1766. Associate editor: Alexander N. Krot