Fe isotope composition of bulk chondrules from Murchison (CM2): Constraints for parent body alteration, nebula processes and chondrule-matrix complementarity

Earth and Planetary Science Letters 490 (2018) 31–39

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Fe isotope composition of bulk chondrules from Murchison (CM2): Constraints for parent body alteration, nebula processes and chondrule-matrix complementarity Dominik C. Hezel a,b,c,∗ , Johanna S. Wilden a , Daniel Becker a , Sonja Steinbach d , Frank Wombacher a,c , Markus Harak a a

University of Cologne, Department of Geology and Mineralogy, Zülpicher Str. 49b, 50674 Köln, Germany Natural History Museum, Department of Mineralogy, Cromwell Road, SW7 5BD, London, UK c Steinmann-Institut, Poppelsdorfer Schloss, Meckenheimer Allee 169, 53115 Bonn, Germany d Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Materialphysik im Weltraum, Linder Höhe, 51147 Köln, Germany b

a r t i c l e

i n f o

Article history: Received 25 November 2017 Received in revised form 29 January 2018 Accepted 8 March 2018 Available online xxxx Editor: F. Moynier Keywords: chondrules matrix Fe isotopes tomography chondrule formation complementarity

a b s t r a c t Chondrules are a major constituent of primitive meteorites. The formation of chondrules is one of the most elusive problems in cosmochemistry. We use Fe isotope compositions of chondrules and bulk chondrites to constrain the conditions of chondrule formation. Iron isotope compositions of bulk chondrules are so far only known from few studies on CV and some ordinary chondrites. We studied 37 chondrules from the CM chondrite Murchison. This is particularly challenging, as CM chondrites contain the smallest chondrules of all chondrite groups, except for CH chondrites. Bulk chondrules have δ 56 Fe between −0.62 and +0.24h relative to the IRMM-014 standard. Bulk Murchison has as all chondrites a δ 56 Fe of 0.00h within error. The δ 56 Fe distribution of the Murchison chondrule population is continuous and close to normal. The width of the δ 56 Fe distribution is narrower than that of the Allende chondrule population. Opaque modal abundances in Murchison chondrules is in about 67% of the chondrules close to 0 vol.%, and in 33% typically up to 6.5 vol.%. Chondrule Al/Mg and Fe/Mg ratios are sub-chondritic, while bulk Murchison has chondritic ratios. We suggest that the variable bulk chondrule Fe isotope compositions were established during evaporation and recondensation prior to accretion in the Murchison parent body. This range in isotope composition was likely reduced during aqueous alteration on the parent body. Murchison has a chondritic Fe isotope composition and a number of chondritic element ratios. Chondrules, however, have variable Fe isotope compositions and chondrules and matrix have complementary Al/Mg and Fe/Mg ratios. In combination, this supports the idea that chondrules and matrix formed from a single reservoir and were then accreted in the parent body. The formation in a single region also explains the compositional distribution of the chondrule population in Murchison. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Chondritic meteorites (chondrites) are “cosmic conglomerates” (cf. Brearley and Krot, 2012) and mainly consist of chondrules plus matrix (up to >90 vol.%), and minor amounts of Ca, Al-rich inclusions and opaque phases (mainly metal, sulfide and magnetite). Chondrules are roundish, typically tens of μm to mm in diameter sized, igneous objects that formed during brief high-temperature events in the early solar system. Many models of chondrule formation have been proposed over the past decades. The difficulty

*

Corresponding author at: University of Cologne, Department of Geology and Mineralogy, Zülpicher Str. 49b, 50674 Köln, Germany. E-mail address: [email protected] (D.C. Hezel). https://doi.org/10.1016/j.epsl.2018.03.013 0012-821X/© 2018 Elsevier B.V. All rights reserved.

to identify the chondrule forming process results primarily from insufficient constraints on chondrule formation and contradictions between these constraints. Two of the most critical constraints regarding chondrule formation currently discussed are: (i) how to explain the compositional distributions of chondrule populations in individual meteorites, and (ii) did chondrules and matrix form in the same or in different regions of the protoplanetary disk and were subsequently transported and aggregated into a single parent body. Further, it is important to know whether chondrule formation conditions are similar or different between groups or even individual meteorites of the same group (e.g., Ciesla, 2005; Zanda et al., 2006; Hezel et al., 2006; Hezel and Palme, 2008, 2010; Alexander and Ebel, 2012; Jones, 2012; Palme et al., 2014b, 2015; Becker et al., 2015; Goldberg et al., 2015; Olsen et al., 2016).

32

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

The distribution of stable Fe isotopes in the chondrule population of individual chondrites record the conditions during chondrule formation, as well as information about the region in the protoplanetary disk where chondrules formed. So far, Fe isotope compositions of chondrules and matrix have only been reported for few meteorites: for the CV chondrites Allende, Mokoia and Grosnaja (Mullane et al., 2005; Hezel et al., 2010) and various ordinary chondrites (Kehm et al., 2003; Mullane et al., 2005; Needham et al., 2009). Chondrules have large stable Fe isotope variations, unlike their host bulk chondrites, which are highly homogeneous and indistinguishable from each other (Zhu et al., 2001; Kehm et al., 2003; Poitrasson et al., 2004; Wang et al., 2013; Mullane et al., 2005; Needham et al., 2009; Hezel et al., 2010). Stable isotope variations are expressed in parts per thousand deviation from a standard, and in chondrules the δ 56 Fe ranges from −1.33 to +0.65 h. For comparison, most igneous terrestrial rocks (e.g., peridotites and abyssal basalts) range between −0.43 and +0.18 h (e.g., Weyer and Ionov, 2007; Teng et al., 2013). A number of processes have been suggested to explain the observed δ 56 Fe variations in chondrules: Needham et al. (2009) found that the δ 56 Fe distribution of chondrite chondrule populations becomes narrower with increasing petrologic type of their host chondrites. Hence, parent body metamorphism redistributes Fe, and thereby the chondrule δ 56 Fe compositions decrease towards the host chondrite’s δ 56 Fe compositions. Mullane et al. (2005) suggested that a combination of high-temperature nebula events and subsequent metasomatic processes on the parent body produced the observed δ 56 Fe distribution in Allende and Chainpur chondrule populations. Bouvier et al. (2013) reached a similar conclusion by using stable Mg isotope compositions of bulk chondrules from the CM chondrites Murchison and Murray. Finally, Hezel et al. (2010) concluded that the dominant process for the δ 56 Fe distribution in the Allende chondrule population most likely was evaporation and recondensation in the protoplanetary disk. Other possibilities to explain these isotope distributions could be compositionally heterogeneous chondrule precursor grains or formation of chondrules in compositionally different regions in the protoplanetary disk with subsequent transport and mixing into their parent bodies. Bulk chondrule Fe isotope data from carbonaceous chondrites so far only exist for CV chondrites. Here we study chondrules from the CM chondrite Murchison to (i) expand our knowledge of bulk chondrule Fe isotope distributions to a second carbonaceous chondrite group, (ii) compare our results to the CV chondrite chondrule populations and (iii) to provide constraints for the formation and origin of chondrules. 2. Methods 2.1. Sample selection A total of ∼2 g Murchison was wrapped in weighing paper and gently crushed using a small hammer. Individual chondrules were separated under a binocular with ceramic tweezers to avoid any contamination from metal. In rare cases, matrix adhering to chondrules was removed using a second set of ceramic tweezers. Separated chondrules were stored individually in small, labelled gelatine capsules. 2.2. 3D μ-tomography The method of 3D X-ray computed tomography allows nondestructive, high spatial and contrast resolution imaging of the volume fraction of different chondrule phases in three dimensions. We used different mounts for tomography measurements. In a first round of measurements, the chondrules remained in their gelatine capsules, which were placed in a test tube. This approach was

only partly successful, as some of the chondrules moved during the scans. For the next round, we built a sample holder made of X-ray transparent PVC that can take and measure up to 7 chondrules per scan. The chondrules are fixed in small holes with cotton in the top of the holder. With this setup, none of the chondrules moved during the CT-scans. The holders were mounted on the rotating table of a Nanotom® 160NF (Phoenix | X-ray, Germany) at the German Aerospace Center DLR, Cologne, Germany. The measurements were conducted with a tube voltage set to 100 kV and the current set to 100 μA. A total of 1000 projections were acquired, as the sample was rotated through 360◦ in incremental steps of 0.36◦ , with each projection taking 2 s. Dark current noise and spatial heterogeneity of the X-ray beam were corrected by dark current subtraction and division of the direct beam images. The voxel resolution was 7.5 μm. The reconstruction of the 16-bit images was performed with datos | x-reconstruction software (GE Sensing & Inspection GmbH). Quantitative results were extracted from subsets of the tomographic data (image stacks) using ImageJ. 2.3. Sample preparation and dissolution Individual chondrules were dissolved in closed Savillex teflon beakers for at least 12 h at >120 ◦ C in a 2:1 mixture of concentrated HF and concentrated HNO3 . The solutions were dried down at 90 ◦ C and the residues redissolved in concentrated HCl for at least 12 h at >120 ◦ C. This second step was repeated. No residual solids were found at this stage. A total of 7 g Murchison was powdered in an agate mortar, of which ∼50 mg were used for bulk Murchison analyses. Further, two aliquots of a few milligrammes of Jbilet Winselwan (CM) powder (Friend et al., 2018, pers. comm.) as well as two aliquots of a few milligrammes of Smithsonian Allende (CV) reference powder were digested with the same procedure used for the chondrules (cf. Hezel et al., 2015). 2.4. Element analyses A set of 6 elements (Al, Mg, Fe, Ni, Cr, Ca) was measured by means of Optical Emission Spectrometry (ICP-OES), using a Spectro Arcos OES in the University of Cologne facilities. Detection limits for all elements are in the range of 10 ppb or less, and errors are <5 rel.%. As CM chondrules are the smallest – except for CH – we had to preserve as much Fe as possible for isotope analyses, and diluted the solution by at least 1:100. Therefore Cr and other elements with low concentrations are often below or close to the detection limit. This dilution might also introduce some uncertainty to the element concentrations. Only element ratios are given, as we did not determine precise chondrule weights, which would be required to recalculate chondrule element concentrations from the OES measurements. 2.5. Fe separation and mass spectrometry Chromatographic separation of Fe was performed following published methods (Dauphas et al., 2004) using AG1-X4 anion exchange resin. ICP-OES analyses confirmed that complete recovery of Fe was achieved. Sample to procedural blank ratios typically were at least 1000:1, and in rare cases at least 100:1, thereby introducing either no or only a negligible error. The Fe isotope compositions were measured at the joint Cologne–Bonn laboratory, using a Thermo Finnigan Neptune MCICP-MS, in conjunction with a Micromist nebuliser and a dual cyclonic-Scott type glass spray chamber for sample introduction. The analyses were carried out in medium resolution mode. This configuration provided sufficient transmission to allow routine analyses of 1 ppm Fe sample solutions. The δ 56 Fe and δ 57 Fe notation refers to the permil deviation of the 56 Fe/54 Fe and 57 Fe/54 Fe

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

33

The CT-scans did in most cases not allow to unequivocally determine a petrographic chondrule type, i.e., whether it is e.g., a porphyritic or crypto-crystalline chondrule. It was therefore not possible to study a possible correlation between e.g., petrographic type and bulk chondrule isotope composition. 3.2. Element ratios of chondrules and bulk chondrites

Fig. 1. An individual slice of the tomographic stack of chondrule #18, session 2 (6.0 vol.% opaque phases). The bright opaque phases clearly stand out from the silicates.

isotope ratios of the samples from that of the IRMM-014 Fe isotope standard. The external reproducibility of the data was typically ±0.07h (2 sd) for δ 56 Fe and ±0.15h (2 sd) for δ 57 Fe, whilst the internal (within-run) reproducibility (2 se) was generally about a factor of two better. All samples were analysed between one and five times, non-consecutively, during analytical sessions of about 12 hours duration. Each sample measurement was bracketed by the analyses of a standard IRMM-014 Fe solution that was made up to closely match the Fe concentration of the sample. Both, samples and standards were doped with the same Cu standard solution to achieve 1 ppm Cu in the measurement solutions, which was used to correct for mass-bias. More detailed descriptions on sample preparation and mass spectrometry can be found in Hezel et al. (2010, 2015).

The range of bulk chondrule element ratios are identical in both measurement sessions (Fig. 2, Table 1). The Fe/Mg ratios are typically below ∼2 and the Al/Mg ratios are mostly below 0.1. High bulk chondrule Al/Mg ratios indicate high mesostases abundances, as chondrule mesostases are rich in Al. The element ratios of Murchison chondrules (Fig. 2) are similar to element ratios of chondrules reported from the CM chondrites El Quss Abu-Said (Hezel and Palme, 2010) and Jbilet Winselwan (Friend et al., 2018, pers. comm.). The bulk chondrule Fe/Mg ratios of Murchison chondrules appear higher than in El Quss Abu-Said and Jbilet Winselwan, because bulk chondrule compositions of the latter were determined using modal recombination, which omitted the opaque phases. The bulk chondrite element ratios of Murchison and one out of two Jbilet Winselwan samples are similar to reported CM chondrites (e.g., Wolf and Palme, 2001), as well as reported Jbilet Winselwan element ratios (Göpel et al., 2015; Friend et al., 2018, pers. comm.). The bulk element ratios are also similar to CI chondrites (e.g., Palme et al., 2014a and Fig. 2). Only one bulk Jbilet Winselwan sample has a 70% higher Fe/Mg ratio than the other bulks, possibly caused by a large opaque inclusion. All CM matrix element ratios from various literature sources and obtained by different analytical techniques fall in the same range with super-chondritic Fe/Mg and Al/Mg ratios, and are compositionally complementary to chondrules (Fig. 2).

3. Results 3.3. Fe isotope compositions of chondrules and bulk chondrites We studied a total of 37 chondrules from Murchison (CM2). The measurements were split in two separate sessions with almost one year in between. This allowed us to identify or exclude any bias that might have occurred during one of the measurement sessions. We studied 14 chondrules in the first and 23 chondrules in the second session. CM chondrules are the smallest among all chondrite classes, except for CH chondrites (Friedrich et al., 2015; Friend et al., 2018, pers. comm.). 3.1. Modal abundances of opaque phases in chondrules Chondrule opaque phases in CM chondrites are typically metal (e.g., Rubin et al., 2007) and occur as small, round blebs randomly distributed throughout the chondrule (Fig. 1). Opaque modal abundances typically range from 0 to 6.5 vol.% (Table 1). Only chondrule#22 has an exceptionally high opaque abundance of 23.8 vol.%. The results of the two measurement sessions are similar (Table 1). The Fe in the silicate dominates the bulk chondrule Fe, if the opaque abundance in an unaltered type I chondrule (i.e., a chondrule with a bulk FeO < 10 wt.%) is below ∼0.5 vol.% (cf. Hezel et al., 2010). Murchison chondrule mesostases are, however, often replaced with FeO-rich, matrix-like material. Therefore, silicates in CM chondrules will dominate the bulk chondrule Fe even if the amount of opaques in a type I chondrule is up to 5 vol.%, depending on the amount of altered mesostasis. About 80% of the studied chondrules have up to 5 vol.% opaques (Table 1), i.e., in these, the silicates dominate the bulk chondrule Fe. In the remaining about 20%, the opaques dominate the bulk chondrule Fe budget.

All measured samples (Table 1) fall on a single mass-dependent fractionation line (Fig. 3) with a slope of 1.50, close to the theoretical value of 1.48, and a y-axis intercept of −0.01. Bulk chondrite compositions have a narrow δ 56 Fe range of about ±0.05h, indistinguishable from the standard (Fig. 4), and other bulk chondrites (e.g., Craddock and Dauphas, 2010) within error. The δ 56 Fe range of individual chondrules is similar in both measurement sessions, in total ranging from −0.62 to +0.24h (Fig. 4). The δ 56 Fe distribution of the chondrule population is about normal (Fig. 5), except for a peak with 5 chondrules that have very similar compositions (cf. Fig. 4). The majority of chondrules have δ 56 Fe values between −0.20 and +0.20h. 4. Discussion In the following we discuss possible processes how the variable Fe isotope compositions of the chondrules could have been produced. Two principle processes are conceivable: (i) either the variable chondrule Fe isotope compositions were produced prior to agglomerating into the CM parent body (sections 4.1, 4.3), or (ii) all chondrules had similar Fe isotope compositions prior to parent body agglomeration, and the variations were produced subsequently on the CM parent body (section 4.2). As we will show, a combination of both processes seems most likely. We then continue to discuss the origin of chondrules (section 4.4) and a possible chondrule-matrix relationship (section 4.5).

34

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

Table 1 The 38 Murchison (CM2) chondrules measured in this study. Presented are their: Fe isotope compositions (h); element mass-ratios; opaque phase modal abundances; and diameters. Plus bulk chondrite Fe isotope compositions and element mass-ratios. When the isotope composition of a sample was measured multiple times (n), the mean of these measurements is given. In this case 2 sd is the 2 standard deviation of these multiple measurements. For some chondrules diameters are reported, but no opaque phase modal abundances. In these cases, the reconstruction of the μCT was sufficient to determine diameters, but not opaque phase modal abundances. s1: Session 1; s2: Session 2. d: diameter. Sample Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd Chd

1, s1 2, s1 3, s1 4, s1 5, s1 6, s1 7, s1 8, s1 9, s1 10, s1 11, s1 12, s1 13, s1 14, s1 1, s2 2, s2 3, s2 4, s2 5, s2 6, s2 7, s2 8, s2 9, s2 10, s2 11, s2 12, s2 13, s2 14, s2 15, s2 16, s2 17, s2 18, s2 19, s2 21, s2 22, s2 23, s2 24, s2 25, s2

Allende Bulk 1, s2 Allende Bulk 2, s2 Murchison Bulk 1, s2 Murchison Bulk 2, s2 Jbilet Winselwan Bulk 1, s2 Jbilet Winselwan Bulk 2, s2

n

δ 56/54 Fe

1 1 1 1 1 1 1 1 1 1 1 1 1 5 3 4 3 3 3 2 1 3 1 1 1 3 3

−0.21 −0.08 −0.03 −0.19 0.00 −0.19 0.24 0.07 0.10 −0.03 −0.07 0.20 0.12 −0.02 −0.11 0.00 0.22 −0.08 −0.62 0.09 −0.41 0.11 −0.11 0.13 −0.18 0.02 −0.22

1 1 2 1 1 1 3 1 1

−0.19 0.14 0.15 −0.18 0.00 0.10 0.05 −0.19 −0.19

5 5 5 5 4 5

0.00 0.01 −0.02 0.00 −0.02 0.02

2 sd

0.06 0.06 0.08 0.05 0.03 0.07 0.05 0.02

0.06 0.03

0.02

0.03

0.02 0.09 0.08 0.03 0.07 0.04

δ 57/54 Fe

−0.37 −0.12 −0.04 −0.33 0.01 −0.34 0.32 0.07 0.11 −0.15 −0.15 0.31 0.19 −0.02 −0.13 0.01 0.35 −0.08 −0.87 0.14 −0.62 0.17 −0.19 0.18 −0.27 0.05 −0.29 −0.31 0.26 0.22 −0.30 −0.01 0.17 0.08 −0.34 −0.32 −0.04 0.06 −0.01 −0.01 −0.03 −0.01

2 sd

0.06 0.12 0.09 0.09 0.06 0.13 0.06 0.06

0.14 0.04

0.05

0.01

0.03 0.06 0.07 0.09 0.05 0.03

4.1. Chondrule δ 56 Fe variations are most likely not from mixing compositionally variable chondrule precursor grains

The mixing of compositionally variable precursor grains homogenises their initial compositional range. Hence, if the compositional range of chondrule bulk compositions resulted from compositionally variable precursor grains, their compositional range must have been larger. For example, if each chondrule were a mixture of about 10 precursors, the precursor reservoir must have had an initial δ 56 Fe range of −0.5 to +0.5 h to reproduce the observed δ 56 Fe range of CM chondrules between about −0.2 and +0.2 h. If each chondrule were a mixture of about 1000 precursors, the δ 56 Fe of the precursor reservoir must have been between about −4 and +4 h. And if each chondrule were a mixture of about 1,000,000 precursors, the δ 56 Fe of the precursor reservoir must have been between about −120 and +120h (cf. Hezel and Palme, 2007). Unsurprisingly, the compositional range of the precursor reservoir needed to be significantly larger, if chondrules were mixtures of larger numbers of precursors.

Fe/Mg 0.38 0.15 0.25 0.28 0.23 0.24 1.34 0.57 0.13 0.13 0.51 0.76 0.08 0.14 0.50 2.79 2.87 1.05 9.61 2.04 0.21 2.97 0.17 0.80 0.05 7.24 15.95 0.07 0.92 0.41 0.88 0.23 0.29 0.60 6.36 0.10

1.29 1.45 1.64 1.66 1.74 2.96

Al/Mg

Ni/Mg

Cr/Mg

Ca/Mg

0.01 0.01 0.02

0.09 0.01 0.06 0.00 0.01 0.02 0.03 0.02 0.01 0.01 0.03

0.01 0.03

0.02

0.06 0.02 0.12

0.11

0.02

0.12 0.02 0.14 0.02 0.02 0.14 0.05 0.07 0.02 0.10 0.08 0.01 0.02 0.08 0.10 0.09

0.11 0.02 0.03

1.35 0.01 0.01

0.30 0.01 0.02

0.23 0.06 0.05

0.04

0.02

0.02

0.04

0.08

0.00

0.01

0.08

0.01

0.02

0.02 0.02 0.01 0.03 0.02

0.05 0.04 0.03 0.09 0.05

0.02 0.01

0.04 0.03

0.02

0.12 0.10 0.08 0.07 0.06 0.04

0.03 0.03

0.01 0.01 0.00 0.01

vol.%

d (μm)

0.0

0.0

160 370 230 250 360 280 340 510 340 320 290 220 620 290 623 403 165 543 371 217 193 280 298 200 322 210 252 315

0.6 0.3 6.0 0.0 0.0 23.8 0.0 0.0 1.6

228 263 245 200 221 252 161 140 252

0.0 0.0 0.7 0.6 2.1

1.1 6.5 5.9 0.0 0.0 0.0 0.1 0.0 5.1 0.0 0.0 0.0 1.0 3.0

0.03 0.04 0.03 0.03 0.07 0.04

0.02 0.04 0.01 0.01

0.02 0.01 0.12 0.11 0.11 0.09 0.10 0.09

0.08 0.07 0.07 0.09 0.10 0.09

CM chondrules have an average diameter of around 200 μm (Friedrich et al., 2015; Friend et al., 2018, pers. comm., Table 1). If each chondrule was mixed from μm-sized precursor grains, about 4 million grains were required for each chondrule. If precursor grains were larger, e.g., about 10 μm in diameter, still about 4000 grains would have been required for each chondrule. It is, however, generally assumed that chondrules formed from finegrained, probably μm sized, precursor grains (e.g., Alexander, 2004; Ciesla, 2005). Hence, each chondrule was made from millions of precursors. In this case, and if the compositional range of chondrules were to be explained by compositionally variable precursor grains, their compositional range of δ 56 Fe must have spanned >250 h. Even with occasionally larger grains included in the chondrule precursor aggregate, the compositional range of the precursors would have had to span tens of permil. If chondrules were mixtures of compositionally variable precursors, smaller chondrules would be made of smaller numbers of precursor grains than larger chondrules, and smaller chondrules should therefore show larger variations in δ 56 Fe. Yet such a relationship is not seen (Table 1). Further, mixing of compositionally

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

35

Fig. 2. Element ratios of the studied Murchison bulk chondrules and bulk chondrite compositions. Literature data of chondrule, matrix and bulk compositions from CM chondrites are added for comparison and to put the results of this study into context. Green arrows indicate Fe/Mg ratios of chondrules for which no Al/Mg data are available. chd: chondrule; mtx: matrix; El Quss: El-Quss Abu Said; JW: Jbilet Winselwan; References: 1: Hezel and Palme (2010); 2: Friend et al. (2018); 3: Zolensky et al. (1993), McSween and Richardson (1977), Fuchs et al. (1973); 4: Barrat et al. (2012). (For interpretation of the colours in the figure(s), the reader is referred to the web version of this article.)

Fig. 3. All studied samples fall on a mass-fractionation line. No error bars are given for samples that were analysed only once. The typical 2 sd for these samples is given in the lower right corner.

Fig. 5. The Murchison chondrule population has an about normal distribution of δ 56 Fe.

variable precursor grains would have required a mechanism to produce compositionally heterogeneous precursors. Such a mechanism is not known, and no reports of individual small grains within meteorites with largely different δ 56 Fe exist. We therefore exclude compositionally heterogeneous precursor grains as a likely explanation for the δ 56 Fe distribution in the Murchison chondrule population. 4.2. Chondrule δ 56 Fe variations were partly established during aqueous parent body alteration

Fig. 4. Murchison chondrule and bulk chondrite Fe isotope analyses from this study. Number of analysed chondrules in brackets. No error bars are given for samples that were analysed only once. The typical 2 sd for these samples is given in the upper left corner.

Murchison is of petrologic type 2.5 and one of the least altered and least hydrated CM chondrites (Rubin et al., 2007; Howard et al., 2011). Parent body aqueous alteration largely replaced chondrule mesostases with FeO-rich, matrix-like, hydrated, phases. Further, chondrule olivines contain narrow veinlets of Fe-rich serpentine. However, chondrule metal remained largely unaffected from aqueous parent body alteration (e.g., Rubin et al., 2007; Lee and Lindgren, 2016). Experimentally derived equilibrium Fe isotope fractionation factors for δ 56 Fe between an aqueous fluid and e.g., hematite, FeS, or magnetite strongly depend on ambient conditions such as temperature or Fe aqueous speciation, and range from 0 to about 0.5h, and up to 1.6h in case of magnetite (Guilbaud et al., 2011; Saunier et al., 2011; Frierdich et al., 2014; Scott et al., 2017). The FeO-rich phyllosilicates in the chondrule mesostases are clear evidence for Fe-redistribution on the parent body. The origin of the Fe in the fluids is, however, less clear. Metal in Murchison chondrules are not rimmed by magnetite, as observed in CM chondrites

36

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

Fig. 6. The bulk chondrule Fe isotope variations decrease with increasing bulk chondrule Fe/Mg ratios. The chondrule compositions clearly do not follow a hypothetical trend for aqueous alteration on the parent body only, but rather suggests a combination of evaporation & recondensation plus subsequent parent body aqueous alteration.

Chondrules in which opaques dominate the bulk chondrule Fe composition have variable, i.e., heavy and light δ 56 Fe (Fig. 7). The chondrule metal is most likely of pre-accretionary origin and remained largely unaffected by parent body alteration, i.e., should have retained its original isotope composition. Hence, parent body alteration cannot account for the observed δ 56 Fe variation of metal rich chondrules. Even if some of the opaque phases were secondary magnetite, which is not expected for CM chondrules (Rubin et al., 2007), the formation of this magnetite during alteration would make the chondrule only heavier in Fe. The chondrules with light Fe could not be explained this way. In summary, while aqueous parent body alteration cannot explain the full extent of observed bulk chondrule Fe isotope variations, it most likely had an effect on the final bulk chondrule isotope compositions, as we will demonstrate in the next section. 4.3. Chondrule δ 56 Fe variations were produced during pre-accretionary processes

Fig. 7. Chondrules, in which opaques dominate the Fe budget of the chondrule can have light and heavy Fe isotope compositions. The grey field where silicates dominate the Fe budget of the chondrules refers to an average range, as it strongly depends on the amount of chondrule mesostasis, and portion of mesostasis replaced by FeO-rich, matrix-like material during parent body alteration. This grey field might range from less than 1 to up to 5 vol.%.

of lower petrologic types (Rubin et al., 2007), hence, metal is not a likely source for the Fe in the fluid. As most chondrules are of FeOpoor type I, chondrules in general cannot have been the source of the Fe in the fluid. Hence, chondrules did not change their initial Fe-isotope compositions due to loss of Fe during aqueous alteration. Chondrite matrices are typically Fe-rich (e.g., Palme et al., 2015 and references therein; cf. Fig. 2) and the most likely source of the Fe in the fluids. Fluids that formed in equilibrium with Fe oxides and sulfides would have been isotopically light (see above), and transport of light Fe into chondrules would have lowered their initial δ 56 Fe to variable degrees, depending on their initial Fe contents and amount of mesostasis that was replaced. The added Fe would have also increased the Fe/Mg ratio of the chondrules. Together, a compositional change could be expected that should follow the hypothetical blue trend of aqueous alteration shown in Fig. 6. The studied chondrules, however, do not follow such a trend. If all chondrules had the same initial δ 56 Fe of around 0h, aqueous alteration could have only shifted the δ 56 Fe of the chondrules to light isotope compositions, i.e., negative values. Therefore, aqueous alteration cannot explain the heavy δ 56 Fe, and, consequently, aqueous alteration cannot explain the full range of observed bulk chondrule Fe isotope compositions. Studies of terrestrial weathering showed that bulk meteorite δ 56 Fe is shifted by no more than about +0.1h (Saunier et al., 2010; Hezel et al., 2011, 2015). This is significantly less than the total δ 56 Fe range of >0.4h in CM chondrules, further indicating that aqueous alteration cannot explain the full extent of bulk chondrule Fe isotope compositions.

Chondrules were molten during their brief, high temperature formation events. Many studies demonstrated that chondrule melts exchanged material with the ambient gas during this high temperature stage (Tissandier et al., 2002; Hezel et al., 2003, 2006, 2010; Krot et al., 2004; Libourel et al., 2006; Sears et al., 1996; Alexander et al., 2008; Friend et al., 2016). One mechanism for material exchange might have been evaporation and recondensation. We propose such a scenario of evaporation and recondensation to explain the observed range of heavy and light bulk chondrule isotope compositions: As discussed above, chondrules appear to have formed from millions of precursor grains, hence, had the average composition of the precursor reservoir, which most likely was the same as the bulk chondrite of δ 56 Fe = 0h. When chondrule melts partly evaporated, the residual chondrule melts became enriched in heavy isotopes – i.e., had positive δ 56 Fe. The extent of isotope fractionation in each chondrule depended on a number of parameters. For each chondrule, these include e.g., the fraction of evaporated Fe, which depends on its surface/volume ratio, duration and peak temperature of chondrule formation, which might have slightly varied depending on e.g., bulk chondrule composition, or the chondrule forming process, etc. This means, after evaporation, each chondrule had a slightly different, but always positive, δ 56 Fe. Such an effect of evaporation is known from cosmic spherules (e.g., Alexander et al., 2002; Taylor et al., 2005), meteorite fusion crusts (Hezel et al., 2015) and evaporation experiments (cf. Davis and Richter, 2014). While all chondrules became enriched in heavy isotopes, i.e., positive δ 56 Fe, the evaporated gas phase became enriched in light isotopes, i.e., negative δ 56 Fe (e.g., Alexander, 2001, 2004; Wang et al., 2001). The isotope composition of the gas might have quickly homogenised and had a uniform, negative δ 56 Fe. When the gas started to recondense onto the chondrules, the δ 56 Fe of each chondrule would have been the mix of the δ 56 Fe of the chondrule and the δ 56 Fe of the gas. In addition, kinetic condensation could have further enriched the chondrules in light isotopes (e.g. Richter, 2004). The mass balance of the mix of Fe condensed onto chondrules again depended on a number of parameters such as Fe concentration and δ 56 Fe of each chondrule, amount of Fe condensed onto each chondrule, etc. Importantly, if a chondrule had a very low Fe concentration, and maybe even also small positive δ 56 Fe, the addition of large amounts of Fe from the gas could have shifted the δ 56 Fe of the chondrule to negative values. Hence, this proposed process of evaporation and recondensation can explain both, positive and negative δ 56 Fe of chondrules. In this scenario, mass balance predicts that chondrules with lower Fe/Mg would be affected more than chondrules with high

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

37

4.4. Origin and formation of chondrules in the protoplanetary disk

Fig. 8. Comparison of Fe isotope data from Murchison CM chondrules (this study), CV chondrules (Mullane et al., 2005; Hezel et al., 2010), CV matrices (Hezel et al., 2010) and CV bulk chondrites (Mullane et al., 2005; Needham et al., 2009; Hezel et al., 2010). No additional Fe isotope chondrule compositions from other carbonaceous chondrites classes have been reported so far.

Fe/Mg ratios. This relationship is seen in the plot of Fig. 6, where chondrules with low Fe/Mg have generally larger δ 56 Fe variations than chondrules with high Fe/Mg. It could also be expected that chondrules of various sizes, i.e., various surface/volume ratios would show a relationship with δ 56 Fe. Yet, no such relationship can be seen (Table 1). This might be related to difficulties of determining correct chondrule diameters from non-roundish objects. The proposed scenario of evaporation and recondensation is similar to what has been proposed earlier for CV chondrules (Mullane et al., 2005; Hezel et al., 2010), indicating similar conditions and a similar process of chondrule formation for chondrules of the CV and CM chondrite groups. Loss of opaque phases from chondrules as suggested by Pringle et al. (2017) to explain Zn isotope variations in CV chondrules most likely cannot explain the observed bulk chondrule δ 56 Fe variations, as no correlation between Fe/Mg and Fe isotopes is observed. The addition of Fe to the chondrules during aqueous alteration on the parent body would have lowered their heavy δ 56 Fe and increased their light chondrule δ 56 Fe compositions, while at the same time increased the chondrule Fe/Mg ratios. These compositional changes would have decreased the initial spread of bulk chondrule δ 56 Fe. The combination of pre-accretionary and parent body alteration would follow a hypothetical trend as indicated by the orange lines of evaporation & recondensation + aqueous alteration in Fig. 6. A general decrease in bulk chondrule δ 56 Fe while their Fe/Mg ratios increase is clearly visible. The Murchison chondrule population has a narrower δ 56 Fe distribution compared to the Allende chondrule population (Fig. 8). This agrees with the observation of Bouvier et al. (2013), who found a narrower δ 25 Mg distribution in the Murchison chondrule population compared to the Allende chondrule population. This similarity of narrower δ 56 Fe and δ 25 Mg in the Murchison chondrule population compared to Allende likely indicates a more extensive parent body alteration and element redistribution in Murchison compared to Allende. The observed bulk chondrule isotope compositions are much smaller than what is calculated using the Rayleigh equation with a kinetic isotope fractionation factor of α kin ≈ (54/56)0.5 for evaporation into vacuum and as observed in cosmic spherules, which can produce δ 56 Fe of up to several tens of permil (e.g., Alexander et al., 2002). The comparatively small δ 56 Fe variations support previous suggestions of high pressure and/or high chondrule densities during chondrule formation (e.g., Cuzzi and Alexander, 2006; Alexander et al., 2008; Hezel et al., 2010, 2015), which is expected to suppress large isotope fractionations.

A number of element and isotope compositions differ among bulk chondrites. It has therefore been suggested that chondrites might have formed in compositionally different regions, i.e., the protoplanetary disk was probably not entirely homogeneous (e.g., Budde et al., 2016a; Van Kooten et al., 2016; Gerber et al., 2017). It has also been suggested that material was transported between various regions, and e.g., chondrules and matrix from compositionally different regions were mixed into a single parent body (e.g., Anders, 1964; Grossman and Wasson, 1982; Olsen et al., 2016; Zanda et al., 2017). These suggestions might be tested using the Fe isotope compositions of chondrite components and the chondritic bulk. Here we provide 3 arguments that are incompatible with multiple parental reservoirs for chondrules and matrix from individual chondrites: (i) Bulk δ 56 Fe compositions of all chondrites are the same within error. In contrast, chondrules have variable δ 56 Fe compositions. If we assume that chondrules of a single chondrite chondrule population originated in compositionally different regions, then chondrules of these regions always needed to mix in the correct proportions to produce the same bulk chondrite δ 56 Fe composition. This seems unlikely. (ii) Chondrule populations have about normal δ 56 Fe distributions as seen in Fig. 5 and previously reported by e.g., Mullane et al. (2005), Needham et al. (2009), Hezel et al. (2010). Yet, if chondrules of an individual chondrite came from multiple reservoirs with distinct δ 56 Fe, their distribution could be expected to be multi-modal, which is not the case. (iii) The Fe isotope data of CV chondrules and matrix reported by Hezel et al. (2010) indicate a complementary relationship in mean chondrule and matrix δ 56 Fe (Fig. 8), which is supported by complementary δ 88 Sr isotope compositions in CV chondrites reported by Moynier et al. (2010). Such complementary relationships require mixing chondrules and matrix in precise proportions. Should chondrules from individual chondrites further have originated from multiple parental reservoirs, also these needed to mix in precise proportions to achieve the bulk chondrite mean. This means, all chondrules from different regions, plus matrix from different regions needed to produce exactly the chondritic bulk δ 56 Fe. This seems highly unlikely, and therefore also the formation and transport of chondrules from different localities in the protoplanetary disk to the accretion region of the parent body seems highly unlikely. We therefore suggest that all, or almost all chondrules and matrix of individual chondrites formed in the same reservoir and locality in the protoplanetary disk. 4.5. Complementary chondrule-matrix element ratios Fig. 2 displays literature element ratios of CM chondrite matrices and bulk CM and CI chondrites. Virtually all matrices have higher than bulk Al/Mg and Fe/Mg ratios, and almost all chondrule compositions have low Al/Mg and Fe/Mg ratios. At the same time, bulk Murchison has CI chondritic, i.e., solar Al/Mg and Fe/Mg ratios. This kind of complementary element compositions in the major components chondrules and matrix, while the host chondrite has CI chondritic compositions, has been described before for other chondrite groups and elements (Bland et al., 2005; Hezel and Palme, 2008, 2010; Palme et al., 2014b, 2015; Becker et al., 2015; Budde et al., 2016a, 2016b; Ebel et al., 2008, 2016; Kadlag and Becker, 2016; Friend et al., 2017). The elemental complementarity displayed in Fig. 2 has not been reported in CM chondrites before, and, importantly, the chondrule bulk data were obtained with a different method (solution based OES) than the matrix (electron microprobe). Further, bulk chondrule data from the literature obtained with electron microprobe (Hezel and Palme, 2010) cover

38

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

identical ranges than the data obtained here with ICP-OES. Chondrule matrix complementarity is, hence, found independent of the method used to determine the component compositions. Parent body alteration cannot produce this complementarity. Redistribution of Fe and Mg would produce chondrules and matrix of more similar, not different compositions. Further, Al is an immobile element, and significant redistribution during parent body alteration is not expected. 5. Conclusions The range of bulk chondrule δ 56 Fe is very similar in CM and CV chondrite chondrule populations. Further, CM and CV chondrule populations have similar Fe isotope distributions as those from ordinary chondrites. It is therefore likely that the Fe isotope compositions of chondrule populations in all chondrite groups are similar. This is supported by stable Mg isotope distributions, which are similar to Fe isotope distributions in CM and CV chondrites, as shown by Bouvier et al. (2013). The Fe isotope compositions of all Murchison chondrules are best explained by evaporation and recondensation processes in a common region before incorporation into the Murchison parent body. This was followed by Fe redistribution during aqueous alteration on the parent body, which narrowed the initial bulk chondrule Fe isotope variations. Absolute δ 56 Fe values are small when compared to kinetic evaporation into vacuum. Evaporation and recondensation must have occurred in a high pressure and/or high dust/gas ratio environment (cf. Alexander et al., 2008). The exchange of material between chondrules with the surrounding gas during evaporation and recondensation, i.e., chondrules acting as open systems is in line with a large body of previous studies suggesting chondrules acted as open systems (Friend et al., 2016 and references therein). The results also support the conclusions from studies on chondrule-matrix compositional complementarities that require the formation of chondrules and matrix from the same parental reservoir (Bland et al., 2005; Hezel and Palme, 2008, 2010; Palme et al., 2014b, 2015; Becker et al., 2015; Budde et al., 2016a, 2016b; Ebel et al., 2008, 2016; Kadlag and Becker, 2016; Friend et al., 2017), and also with the conclusion from Jones (2012) that each chondrite represents a unique reservoir. Chondrule-matrix complementary, determined by solution based OES combined with electron microprobe data from the literature exists for Al/Mg and Fe/Mg, while bulk Murchison has CI chondritic ratios. This is in line with the above conclusion that chondrules and matrix formed from the same parental reservoir, with no significant addition or removal of chondrule and/or matrix material from or to another region of the protoplanetary disk. Acknowledgements We thank three anonymous reviewers, who helped making this a better contribution. Jochen Scheldt and Michael Staubwasser are thanked for access to and assistance in the clean and ICP-OES laboratories. We are grateful for Stefan-Frank Richter’s assistance at the CT scanner. We thank Eric Strubb from the Institute for Nuclear Chemistry, University of Cologne for the sample of Murchison used in this study. References Alexander, C.M.O’D., 2001. Exploration of quantitative kinetic models for the evaporation of silicate melts in vacuum and in hydrogen. Meteorit. Planet. Sci. 36, 255–283. Alexander, C.M.O’D., 2004. Chemical equilibrium and kinetic constraints for chondrule and CAI formation conditions. Geochim. Cosmochim. Acta 68, 3943–3969. Alexander, C.M.O’D., Ebel, D.S., 2012. Questions, questions: can the contradictions between the petrologic, isotopic, thermodynamic, and astrophysical constraints on chondrule formation be resolved? Meteorit. Planet. Sci. 47, 1157–1175.

Alexander, C.M.O’D., Grossman, J.N., Ebel, D.S., Ciesla, F.J., 2008. The formation conditions of chondrules and chondrites. Science 320, 1617–1619. Alexander, C.M.O’D., Taylor, S., Delaney, J.S., Peixue, M.A., Herzog, G.F., 2002. Mass-dependent fractionation of Mg, Si, and Fe isotopes in five stony cosmic spherules. Geochim. Cosmochim. Acta 66, 173–183. Anders, 1964. Origin, age and composition of meteorites. Space Sci. Rev. 3, 583–714. Barrat, J.A., Zanda, B., Moynier, F., Bollinger, C., Liorzou, C., Bayon, G., 2012. Geochemistry of CI chondrites: major and trace elements, and Cu and Zn isotopes. Geochim. Cosmochim. Acta 83, 79–92. Becker, M., Hezel, D.C., Schulz, T., Elfers, B.-M., Münker, C., 2015. The age of CV chondrites from component specific Hf–W systematics. Earth Planet. Sci. Lett. 432, 472–482. Bland, P.A., Alard, O., Benedix, G.K., Kearsley, A.T., Menzies, O.N., Watt, L.E., Rogers, N.W., 2005. Volatile fractionation in the early solar system and chondrule/matrix complementarity. Proc. Natl. Acad. Sci. 103, 13755–13760. Bouvier, A., Wadhwa, M., Simon, S.B., Grossman, L., 2013. Magnesium isotopic fractionation in chondrules from the Murchison and Murray CM2 carbonaceous chondrites. Meteorit. Planet. Sci. 48, 339–353. Brearley, A.J., Krot, A.N., 2012. Metasomatism in the early solar system: the record from chondritic meteorites. In: Metasomatism and the Chemical Transformation of Rock. In: Lecture Notes in Earth System Sciences, pp. 659–789. Budde, G., Burkhardt, C., Brennecka, G.A., Fischer-Gödde, M., Kruijer, T.S., Kleine, T., 2016a. Molybdenum isotopic evidence for the origin of chondrules and a distinct genetic heritage of carbonaceous and non-carbonaceous meteorites. Earth Planet. Sci. Lett. 454, 293–303. Budde, G., Kleine, T., Kruijer, T.S., Burkhardt, C., Metzler, K., 2016b. Tungsten isotopic constraints on the age and origin of chondrules. Proc. Natl. Acad. Sci. USA 113, 2886–2891. Ciesla, F.J., 2005. Chondrule-forming processes – an overview. In: Krot, A.N., Scott, E.R.D., Reipurth, B. (Eds.), Chondrites and the Protoplanetary Disk. In: Astronomical Society of the Pacific Conference Series, vol. 341, pp. 811–820. Craddock, P.R., Dauphas, N., 2010. Iron isotopic compositions of geological reference materials and chondrites. Geostand. Geoanal. Res. 35, 101–123. Cuzzi, J.N., Alexander, C.M.O’.D., 2006. Chondrule formation in particle-rich nebular regions at least hundreds of kilometres across. Nature 441, 483–485. Dauphas, N., Janney, P.E., Mendybaev, R.A., Wadhwa, M., Richter, F.M., Davis, A.M., Van Zuilen, M., Hines, R., Foley, C.N., 2004. Chromatographic separation and multi-collection-ICPMS analysis of iron. Investigating mass-dependent and independent isotope effects. Anal. Chem. 76, 5855–5863. Davis, A.M., Richter, F.M., 2014. Condensation and evaporation of Solar System materials. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry, vol. 2, second edition. Elsevier, Oxford, pp. 335–360. Ebel, D.S., Brunner, C., Konrad, K., Leftwich, K., Erb, I., Lu, M., Rodriguez, H., CrapsterPregont, E., Friedrich, J.M., Weisberg, M.K., 2016. Abundance, major element composition and size of components and matrix in CV, CO and Acfer 094 chondrites. Geochim. Cosmochim. Acta 172, 322–356. Ebel, D.S., Weisberg, M.K., Hertz, J., Campbell, A.J., 2008. Shape, metal abundance, chemistry, and origin of chondrules in the Renazzo (CR) chondrite. Meteorit. Planet. Sci. 43, 1725–1740. Friedrich, J.M., Weisberg, M.K., Ebel, D.S., Biltz, A.E., Corbett, B.M., Wajih, I.V.I., Matthew, S.K., Wolman, D., 2015. Chondrule size and related physical properties: a compilation and evaluation of current data across all meteorite groups. Chem. Erde 75, 419–443. Friend, P., Hezel, D.C., Barrat, J.-A., Zipfel, J., Palme, H., Metzler, K., 2018. Composition, petrology and chondrule-matrix complementarity of the recently discovered Jbilet Winselwan CM2 chondrite. Meteorit. Planet. Sci. (pers. comm.). Friend, P., Hezel, D.C., Mucerschi, D., 2016. The conditions of chondrule formation, Part II: open system. Geochim. Cosmochim. Acta 173, 198–209. Friend, P., Hezel, D.C., Palme, H., Bischoff, A., Gellissen, M., 2017. Complementary element relationships between chondrules and matrix in Rumuruti chondrites. Earth Planet. Sci. Lett. 480, 87–96. Frierdich, A.H., Beard, B.L., Scherer, M.M., Johnson, C.M., 2014. Determination of the Fe(II)aq–magnetite equilibrium iron isotope fractionation factor using the threeisotope method and a multi-direction approach to equilibrium. Earth Planet. Sci. Lett. 391, 77–86. Fuchs, L.H., Olsen, E., Jensen, K.J., 1973. Mineralogy, mineral-chemistry, and composition of the Murchison (C2) meteorite. Smiths. Contrib. Earth Sci., 10, 39 pp. Gerber, S., Burkhardt, C., Budde, G., Metzler, K., Kleine, T., 2017. Mixing and transport of dust in the early solar nebula as inferred from titanium isotope variations among chondrules. Astrophys. J. Lett. 841, L17. Goldberg, A.Z., Owen, J.E., Jacquet, E., 2015. Chondrule transport in protoplanetary discs. Mon. Not. R. Astron. Soc. 452, 4054–4069. Göpel, C., Birck, J.-L., Alber, G., Barrat, J.-A., Zanda, B., 2015. Mn–Cr systematics in primitive meteorites: insights from mineral separation and partial dissolution. Geochim. Cosmochim. Acta 156, 1–24. Grossman, J.N., Wasson, J.T., 1982. Evidence for primitive nebular components in chondrules from the Chainpur chondrite. Geochim. Cosmochim. Acta 46, 1081–1099. Guilbaud, R., Butler, I.B., Ellam, R.M., Rickard, D., Oldroyd, A., 2011. Experimental determination of the equilibrium Fe isotope fractionation between Fe2+ aq and FeSm (Mackinawite) at 25 and 2 ◦ C. Geochim. Cosmochim. Acta 75, 2721–2734.

D.C. Hezel et al. / Earth and Planetary Science Letters 490 (2018) 31–39

Hezel, D.C., Palme, H., 2007. The conditions of chondrule formation, Part I: closed system. Geochim. Cosmochim. Acta 71, 4092–4107. Hezel, D.C., Palme, H., 2008. Constraints for chondrule formation from Ca–Al distribution in carbonaceous chondrites. Earth Planet. Sci. Lett. 265, 716–725. Hezel, D.C., Palme, H., 2010. The chemical relationship between chondrules and matrix and the chondrule-matrix complementarity. Earth Planet. Sci. Lett. 294, 85–93. Hezel, D.C., Palme, H., Brenker, F.E., Nasdala, L., 2003. Evidence for fractional condensation and reprocessing at high temperatures in CH-chondrites. Meteorit. Planet. Sci. 38, 1199–1216. Hezel, D.C., Palme, H., Nasdala, L., Brenker, F.E., 2006. Origin of SiO2 -rich components in ordinary chondrites. Geochim. Cosmochim. Acta 70, 1548–1564. Hezel, D.C., Needham, A.W., Armytage, R., Georg, B., Abel, R., Kurahashi, E., Coles, B.J., Rehkämper, M., Russell, S.S., 2010. A nebula setting as the origin for bulk chondrule Fe isotope variations in CV chondrites. Earth Planet. Sci. Lett. 296, 423–433. Hezel, D.C., Schlüter, J., Kallweit, H., Jull, A.J.T., Al Fakeer, O.Y., Al Shamsi, M., Strekopytov, S., 2011. Meteorites from the United Arab Emirates: description, weathering and terrestrial ages. Meteorit. Planet. Sci. 46, 327–336. Hezel, D.C., Poole, G., Hoyes, J., Coles, B.J., Unsworth, C., Albrecht, N., Smith, C., Rehkämper, M., Pack, A., Genge, M., Russell, S.S., 2015. Fe and O isotope composition of meteorite fusion crusts: possible natural analogues to chondrule formation? Meteorit. Planet. Sci. 50, 229–242. Howard, K.T., Benedix, G.K., Bland, P.A., Cressey, G., 2011. Modal mineralogy of CM chondrites by X-ray diffraction (PSD-XRD): Part 2. Degree, nature and settings of aqueous alteration. Geochim. Cosmochim. Acta 75, 2735–2751. Jones, R.H., 2012. Petrographic constraints on the diversity of chondrule reservoirs in the protoplanetary disk. Meteorit. Planet. Sci. 47, 1176–1190. Kadlag, Y., Becker, H., 2016. Highly siderophile and chalkogen element constraints on the origin of components of the Allende and Murchison meteorites. Meteorit. Planet. Sci. 51, 1136–1152. Kehm, K., Hauri, E.H., Alexander, C.M.O’D., Carlson, R.W., 2003. High precision iron isotope measurements of meteoritic material by cold plasma ICP-MS. Geochim. Cosmochim. Acta 67, 2879–2891. Krot, A.N., Libourel, G., Goodrich, C., Petaev, M.I., 2004. Silica-igneous rims around magnesian chondrules in CR carbonaceous chondrites: evidence for fractional condensation during chondrule formation. Meteorit. Planet. Sci. 39, 1931–1955. Lee, M.R., Lindgren, P., 2016. Aqueous alteration of chondrules from the Murchison CM carbonaceous chondrite: replacement, pore filling, and the genesis of polyhedral serpentine. Meteorit. Planet. Sci. 51, 1003–1021. Libourel, G., Krot, A., Tissandier, L., 2006. Role of gas–melt interaction during chondrule formation. Earth Planet. Sci. Lett. 251, 232–240. McSween Jr., H.Y., Richardson, S.M., 1977. The composition of carbonaceous chondrite matrix. Geochim. Cosmochim. Acta 41, 1145–1161. Moynier, F., Agranier, A., Hezel, D.C., Bouvier, A., 2010. Sr stable isotope composition of Earth, the Moon, Mars, Vesta and meteorites. Earth Planet. Sci. Lett. 300, 359–366. Mullane, E., Russell, S.S., Gounelle, M., 2005. Nebular and asteroidal modification of the iron isotope composition. Earth Planet. Sci. Lett. 239, 203–218. Needham, A.W., Porcelli, D., Russell, S.S., 2009. An Fe isotope study of ordinary chondrites. Geochim. Cosmochim. Acta 73, 7399–7413. Olsen, M.B., Wielandt, D., Schiller, M., Van Kooten, E.M.M.E., Bizzarro, M., 2016. Magnesium and 54 Cr isotope compositions of carbonaceous chondrite chondrules – insights into early disk processes. Geochim. Cosmochim. Acta 191, 118–138. Palme, H., Hezel, D.C., Ebel, D.S., 2015. The origin of chondrules: constraints from matrix composition and matrix-chondrule complementarity. Earth Planet. Sci. Lett. 411, 11–19. Palme, H., Lodders, K., Jones, A., 2014a. Solar system abundances of the elements. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry, vol. 2, second edition. Elsevier, Oxford, pp. 15–36.

39

Palme, H., Spettel, B., Hezel, D.C., 2014b. Siderophile elements in chondrules of CVchondrites. Chem. Erde 74, 507–516. Poitrasson, F., Halliday, A.N., Lee, D.-C., Levasseur, S., Teutscha, N., 2004. Iron isotope differences between Earth, Moon, Mars and Vesta as possible records of contrasted accretion mechanisms. Earth Planet. Sci. Lett. 223, 253–266. Pringle, E.A., Moynier, F., Beck, P., Paniello, R., Hezel, D.C., 2017. The origin of volatile element depletion in early solar system material: clues from Zn isotopes in chondrules. Earth Planet. Sci. Lett. 468, 62–71. Richter, F., 2004. Timescales determining the degree of kinetic isotope fractionation by evaporation and condensation. Geochim. Cosmochim. Acta 68, 4971–4992. Rubin, A.E., Trigo-Rodríguez, J.M., Huber, H., Wasson, J.T., 2007. Progressive aqueous alteration of CM carbonaceous chondrites. Geochim. Cosmochim. Acta 71, 2361–2382. Saunier, G., Poitrasson, F., Moine, B., Gregoire, M., Seddiki, A., 2010. Effect of hot desert weathering on the bulk-rock iron isotope composition of L6 and H5 ordinary chondrites. Meteorit. Planet. Sci. 45, 195–209. Saunier, G., Pokrovski, G.S., Poitrasson, F., 2011. First experimental determination of iron isotope fractionation between hematite and aqueous solution at hydrothermal conditions. Geochim. Cosmochim. Acta 75, 6629–6654. Scott, S.R., Sims, K.W.W., Frost, B.R., Kelemen, P.B., Evans, K.A., Swapp, S.M., 2017. On the hydration of olivine in ultramafic rocks: implications from Fe isotopes in serpentinites. Geochim. Cosmochim. Acta 215, 105–121. Sears, D.W.G., Huang, S., Benoit, P.H., 1996. Open-system behavior during chondrule formation. In: Hewins, R.H., Jones, R.H., Scott, E.R.D. (Eds.), Chondrules and the Protoplanetary Disk. Cambridge University Press, Cambridge, pp. 221–231. Taylor, S., Alexander, C.M.O’D., Delaney, J., Ma, P., Herzog, G.F., Engrand, C., 2005. Isotopic fractionation of iron, potassium, and oxygen in stony cosmic spherules: implications for heating histories and sources. Geochim. Cosmochim. Acta 69, 2647–2662. Teng, F.-Z., Dauphas, N., Huang, S., Marty, B., 2013. Iron isotopic systematics of oceanic basalts. Geochim. Cosmochim. Acta 107, 12–26. Tissandier, L., Libourel, G., Robert, F., 2002. Gas–melt interactions and their bearing on chondrule formation. Meteorit. Planet. Sci. 37, 1377–1389. Van Kooten, E.M.M.E., Wielandt, D., Schiller, M., Nagashima, K., Thomen, A., Larsen, K.K., Olsen, M.B., Nordlund, Å., Krot, A.N., Bizzarro, M., 2016. Isotopic evidence for primordial molecular cloud material in metal-rich carbonaceous chondrites. Proc. Natl. Acad. Sci. USA 113, 2011–2016. Wang, J., Davis, A.M., Clayton, R.N., Mayeda, T.K., Hashimoto, A., 2001. Chemical and isotopic fractionation during the evaporation of the FeO–MgO–SiO2 – CaO–Al2 O3 –TiO2 rare Earth element melt system. Geochim. Cosmochim. Acta 65, 479–494. Wang, K., Moynier, F., Barrat, J.-A., Zanda, B., Paniello, R.C., Savage, P.S., 2013. Homogeneous distribution of Fe isotopes in the early solar nebula. Meteorit. Planet. Sci. 48, 354–364. Weyer, S., Ionov, D.A., 2007. Partial melting and melt percolation in the mantle: the message from Fe isotopes. Earth Planet. Sci. Lett. 259, 119–133. Wolf, D., Palme, H., 2001. The solar system abundances of phosphorus and titanium and the nebular volatility of phosphorus. Meteoritics 36, 559–571. Zanda, B., Hewins, R.H., Bourot-Denise, M., Bland, P.A., Albarède, F., 2006. Formation of solar nebula reservoirs by mixing chondritic components. Earth Planet. Sci. Lett. 248, 650–660. Zanda, B., Zanetta, P.M., Leroux, H., Guillou, C., Le Lewin, É., Pont, S., Deldicque, D., Hewins, R.H., 2017. The chondritic assemblage. In: Workshop on Chondrules and the Protoplanetary Disk, 2035 (abstract). Zhu, X.K., Guo, Y., O’Nions, R.K., Young, E.R.D., Ash, R.D., 2001. Isotopic homogeneity of iron in the early solar nebula. Nature 412, 311–313. Zolensky, M., Barrett, R., Browning, L., 1993. Mineralogy and composition of matrix and chondrule rims in carbonaceous chondrites. Geochim. Cosmochim. Acta 57, 3123–3148.