Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/arxiv-cornell-university/late-metal-silicate-separation-on-the-iab-parent-asteroid-constraints-0B34QEfX4h
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- ISSN
- 0012-821X
- eISSN
- ARCH-3330
- DOI
- 10.1016/j.epsl.2017.11.034
- Publisher site
- See Article on Publisher Site
Abstract
182 182 Article history: The short-lived Hf– W decay system is a powerful chronometer for constraining the timing of Received 27 June 2017 metal–silicate separation and core formation in planetesimals and planets. Neutron capture effects on Received in revised form 13 November 2017 W isotopes, however, significantly hamper the application of this tool. In order to correct for neutron Accepted 14 November 2017 capture effects, Pt isotopes have emerged as a reliable in-situ neutron dosimeter. This study applies this Available online 27 November 2017 method to IAB iron meteorites, in order to constrain the timing of metal segregation on the IAB parent Editor: F. Moynier body. The ε W values obtained for the IAB iron meteorites range from −3.61 ± 0.10 to −2.73 ± 0.09. Keywords: i 182 182 Correlating ε Pt with ε W data yields a pre-neutron capture ε W of −2.90 ± 0.06. This corresponds IAB iron meteorites Hf–W chronometry to a metal–silicate separation age of 6.0 ± 0.8 Ma after CAI for the IAB parent body, and is interpreted Pt isotopes to represent a body-wide melting event. Later, between 10 and 14 Ma after CAI, an impact led to a multi-collector ICP-MS catastrophic break-up and subsequent reassembly of the parent body. Thermal models of the interior Planetesimals evolution that are consistent with these estimates suggest that the IAB parent body underwent metal– thermal histories silicate separation as a result of internal heating by short-lived radionuclides and accreted at around 1.4 ± 0.1Ma after CAIs with a radius of greater than 60 km. © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction consequently, the timing of metal–silicate separation. This method was successfully applied to magmatic iron meteorites by several 182 182 The short-lived Hf– W chronometer (t = 8.9Ma) is a studies (Kruijer et al., 2014a, 2013a; Wittig et al., 2013). 1/2 valuable tool that can effectively constrain the timing of metal– The IAB meteorites are non-magmatic iron meteorites that silicate separation and core formation in planetary bodies. Hafnium have been classified into numerous sub-groups based on their and W are refractory elements that are strongly fractionated from trace-element geochemistry. These include the Main Group (MG), each other during this process. Hafnium is lithophile and partitions multiple sub-groups including ‘Low-Au Low-Ni’ (sLL) and ‘Low- into the silicate phase, whereas W is siderophile and partitions Au Medium-Ni’ (sLM), and several duos or ungrouped samples 182 182 into metallic phases. The application of the Hf– W chronome- (Wasson and Kallemeyn, 2002). Recently, nucleosynthetic Mo iso- ter to solar system materials, however, is hindered by the effects of tope variations have been used to identify genetic links between neutron capture on W isotopes during exposure to galactic cosmic these sub-groups (Worsham et al., 2017). That study suggests that rays (GCR) (e.g., Leya and Masarik, 2013). Neutron capture reac- while the MG, sLL, sLH and sLM are genetically related, two high- 182 182 182 tions cause burnout of W, leading to incorrect Hf– W ages. Au sub-groups (sHL and sHH), as well as some ungrouped samples, Platinum is an ideal in-situ neutron dose proxy due to the simi- may derive from a distinct parent body or parent body family. lar neutron capture cross-sections of Pt and W (Leya and Masarik, Trace element data, including highly variable concentrations of 2013). Correlating the effects in Pt and W isotopic compositions al- siderophile elements, suggest the IABs did not form by simple frac- lows for an estimation of the pre-GCR exposure ε W value, and tional crystallization in a planetary core, in contrast to magmatic irons (e.g., Wasson and Kallemeyn, 2002; Worsham et al., 2016). Additionally, the IAB irons contain inclusions of chondritic and Corresponding author. E-mail address: alison.hunt@erdw.ethz.ch (A.C. Hunt). non-chondritic silicates, sulfides, graphite, and phosphate-bearing https://doi.org/10.1016/j.epsl.2017.11.034 0012-821X/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 491 Table 1 Sample information and Ir/Pt ratios for analysed IAB meteorites. Samples Source IAB group/ Description of Ir/Pt CRE ages a b subgroup inclusions (Ma) Caddo County ETH un ac, nc 0.40 5 ± 1 Canyon Diablo ETH MG gr, ac 0.40 545 ± 40 Cranbourne ETH MG gr 0.34 45 ± 5 Livingstone (TN) Smithsonian Algarrabo duo gr 0.17 (USNM 1420) Magura ETH MG gr, s 0.37 ∼250 Odessa ETH MG gr 0.41 875 ± 70 Toluca ETH sLL gr, s 0.43 600 ± 150 IAB classifications and Ir/Pt ratios from Wasson and Kallemeyn (2002). IAB group/sub-groups: MG, main group; sLL, subgroup low-Au low-Ni; un, ungrouped. Petrographic descriptions of inclusions: ac, angular chondritic; nc, non-chondritic; gr, graphite-rich; s, silicate (Benedix et al., 2000; Buchwald, 1975). CRE ages from: Takeda et al. (2000). CRE ages from: Schnabel et al. (2001). CRE ages from: Chang and Wänke (1969). CRE ages from: Schulz et al. (2012). CRE ages from: Voshage and Feldmann (1979). inclusions (e.g., Benedix et al., 2000; Buchwald, 1975). The angu- 2. Samples lar texture and the olivine- and pyroxene-rich mineralogy of the Seven IAB meteorites originating from four different IAB groups chondritic silicate inclusions are similar to the winonaite primitive or sub-groups, as defined by Wasson and Kallemeyn (2002), were achondrites (Benedix et al., 1998; Bild, 1977). A genetic relation- chosen for this study. Canyon Diablo, Cranbourne, Magura and ship between the winonaites and MG-IAB iron meteorites is also Odessa are all assigned to the MG, whereas Toluca is a member implied by the identical Mo and O isotope compositions of the of the sLL sub-group, Livingstone (TN) is part of the Algarrabo duo, two groups, suggesting that they originated from a common par- and Caddo County is ungrouped, but a member of the Udei Sta- ent body (Clayton and Mayeda, 1996; Greenwood et al., 2012; tion grouplet. These meteorites contain a variety of inclusion types, Worsham et al., 2017). and cover a range of cosmic ray exposure (CRE) ages (Table 1). All The thermal history and evolution of the IAB parent asteroid, or samples are taken from the ETH collection, except Livingstone (TN) asteroid family, remains unclear. One theory suggests that the IAB (USNM 1420, Smithsonian Institute). irons formed in melt pools created by impacts into a chondritic parent body or bodies (Choi et al., 1995; Wasson and Kallemeyn, 3. Methods 2002; Worsham et al., 2017, 2016). A competing theory argues 26 60 that the decay of short-lived radionuclides such as Al and Fe 3.1. High-precision Pt and W isotope measurements would produce enough heat for incipient partial melting, which was followed by a catastrophic impact. Subsequent reassembly due Between 1.5 and 3.0 g of each IAB meteorite was prepared for to gravity led to the extensive mixing of silicates, sulfides and met- use in this study, and both W and Pt isotope aliquots were taken als (Benedix et al., 2000; Hunt et al., 2017a; Schulz et al., 2012; from the same sample digestion. The sample preparation proce- Theis et al., 2013). This model can apply to groups with the same dures and Pt and W isotope chemical separations employed in this genetic affinities (e.g., the MG, sLL, sLH and sLM) as defined by Mo study are described in the Supplementary Materials. Platinum iso- isotopes (Worsham et al., 2017). tope data for Toluca are taken from Hunt et al. (2017b; Toluca-a). Several studies attempted to constrain the history of the IAB All Pt and W isotope analyses were performed at ETH Zürich 182 182 using a Thermo Scientific Neptune Plus fitted with a Cetac Aridus parent body using the Hf– W chronometer (e.g., Markowski et al., 2006; Qin et al., 2008; Schulz et al., 2009). The metal– II desolvating system and standard H cones. Platinum isotope anal- yses followed the procedure of Hunt et al. (2017c). Analyses were silicate separation ages based on these data and recalculated rela- corrected for instrumental mass bias using the exponential law, tive to the most recent value for calcium aluminum rich inclusions 198 195 and were internally normalized to Pt/ Pt (‘8/5’) = 0.2145 (CAI) (ε W: −3.49 ± 0.07, Burkhardt and Schönbächler, 2015; (Kruijer et al., 2013a). Samples were analysed with a Pt sig- Kruijer et al., 2014b) range from 3.6 ± 2.1 Ma (Schulz et al., 2009) −10 nal of ∼2 × 10 A, equivalent to ∼200 ppb Pt, and utilizing to 11.5 ± 6.5 Ma after CAI (Markowski et al., 2006). This broad ∼200 ng Pt per measurement. Each sample was measured relative range is partly related to difficulties to correct for neutron cap- to the NIST SRM 3140 Pt standard solution, and data are presented ture effects, reasserting the need for a reliable neutron-dose proxy. 19i 195 in the epsilon notation (i.e., ε Pt/ Pt = deviation in parts per Recently, attempts to correct for the effects of GCR have yielded 10,000 from the average of the bracketing standards). Repeat anal- 182 182 Hf– W ages for the IAB irons of between 3.4 ± 0.7 and 6.9 ± ysis of four aliquots of our in-house reference material, the North 0.4 Ma (Schulz et al., 2012; Worsham et al., 2017). This study also Chile iron meteorite (IIAB), passed through column chemistry in- corrects for the effect of GCR on W isotopes in the IABs, using the dependently give a 2 standard deviation external reproducibility well-established Pt isotope neutron-dose proxy. We report new W 192 194 196 (2 S.D.) of 0.73 for ε Pt, 0.15 for ε Pt, and 0.09 for ε Pt and Pt isotope data obtained on the same sample aliquot for seven (n = 19; Hunt et al., 2017c). Our 2 S.D. external reproducibility for samples. The data provide new insights into the thermal evolution repeat analyses of North Chile is used as the Pt isotope uncertainty of the IAB parent body by constraining the timing of metal–silicate throughout this study. separation. Numerical models of parent body evolution supplement All five W isotopes, along with the interference monitors Hf, 181 188 these results. Additionally, the potential of Pt isotopes as an in-situ Ta, and Os, were measured simultaneously in static mode neutron-dose proxy in iron meteorites is further evaluated, specif- (Cook and Schönbächler, 2016). Single measurements of each sam- ically for non-magmatic irons. ple were bracketed by measurements of the NIST SRM 3163 W 492 A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 Table 2 Platinum and W isotope compositions of the IAB irons. 192 194 196 182 182 184 IAB n ε Pt ±2 ε Pt ±2 ε Pt ±2 n ε W ±2 ε W ±2 ε W ±2 a b b group (Pt) S.D. S.D. S.D. (W) (6/4) S.D. (6/3) S.D. (6/3) S.D. Caddo County un 1 0.09 0.73 0.06 0.15 −0.01 0.09 1 −2.92 0.10 −2.89 0.10 0.04 0.06 Canyon Diablo MG 2 −0.48 0.73 0.04 0.15 0.00 0.09 1 −2.87 0.10 −2.87 0.10 0.00 0.06 Cranbourne MG 1 2.13 0.73 0.34 0.15 0.13 0.09 1 −3.16 0.10 −3.16 0.10 0.06 0.06 Magura MG 2 2.08 0.73 0.21 0.15 0.11 0.09 1 −3.11 0.10 −3.11 0.10 0.01 0.06 Odessa MG 3 12.99 0.73 0.74 0.15 0.56 0.09 1 −3.61 0.10 −3.61 0.10 0.02 0.06 Livingston (TN) Algarrabo duo 2 0.03 0.73 0.24 0.15 0.08 0.09 1 −3.12 0.09 −3.17 0.07 0.04 0.07 c ∗ Toluca sLL 1 −0.17 0.73 0.09 0.15 0.00 0.09 1 −2.73 0.09 −2.73 0.07 0.00 0.07 198 195 All Pt isotope data are normalised to Pt/ Pt = 0.2145 (Kruijer et al., 2013a). Platinum isotope uncertainties are based on repeat analysis of the in-house standard 186 183 186 184 (North Chile IIAB iron meteorite; see section 3.1). Tungsten isotope compositions are normalized relative to W/ W = 1.98594 (6/3) or W/ W = 0.927672 (6/4). Uncertainties are based on the W bracketing standard (NIST SRM 3163) or the long-term reproducibility defined by repeated measurements of SRM 129c (denoted by an asterisk; Cook and Schönbächler, 2016), whichever is larger. See Table 1 for definition of groups. Tungsten isotope ratios corrected for nuclear field shift effects, as described in Cook and Schönbächler (2016). Uncorrected data are given in Table S1 of the Supplementary Appendix. Pt isotope data previously published as Toluca-a (Hunt et al., 2017b). solution standard. Concentrations were adjusted to achieve a signal −10 184 of ∼4.5 × 10 A on W. Instrumental mass bias was corrected 186 183 using the exponential law with either W/ W = 1.98594 (‘6/3’) 186 184 or W/ W = 0.927672 (‘6/4’) (Völkening et al., 1991) and data are presented in the epsilon notation. Additionally, data are cor- rected for nuclear field shift (NFS) effects using the method out- lined in Cook and Schönbächler (2016). Uncertainties on W isotope ratios are mainly based on the deviation (2 S.D.) of the bracket- ing standard analyses during each analytical session (Table 2). Five replicates of the NIST Fe–Ni steel, SRM 129c, were also analyzed in each session; this material was used as an external standard to define the long-term reproducibility (2 S.D.) of the isotopic measurements and to validate their accuracy during the analyti- cal campaign. These data are published in Cook and Schönbächler (2016). Epsilon values based on SRM 129c have the following ex- 184 182 ternal precisions: ± 0.05 for ε W (6/3), ± 0.07 for ε W (6/3), and ± 0.08 for ε W (6/4) (Cook and Schönbächler, 2016). The external precision of SRM 129c is similar to that derived using the bracketing standard NIST SRM 3163. However, for each W isotope ratio presented, the larger of the two values is taken as the uncer- tainty (Table 2). 3.2. Numerical modelling The thermal evolution of the IAB parent body was modelled using both 2D and 3D fluid dynamics simulations employing the 196 192 194 Fig. 1. Co-variation of ε Pt vs. (a) ε Pt and (b) ε Pt for the IAB samples. Also I2/I3ELVIS code family (Gerya and Yuen, 2007, 2003). The code shown are the model calculations of Leya and Masarik (2013) for Ir/Pt ratios of uses a finite-differences fully-staggered grid and solves for the 0.17 to 0.41, an exposure time of 1000 Ma and all shielding depths across pre- conservation equations for mass, momentum and energy includ- atmospheric radii of 5 to 120 cm (dashed lines). The model lines for the different Ir/Pt ratios illustrate the influence of varying Ir/Pt on ε Pt; see Table 1 for Ir/Pt ra- ing material self-gravity. The numerical model accounts for heat- tios of the studied IABs. Uncertainties (2 S.D.) are based on the North Chile in-house ing by radiogenic, shear and latent heat production source terms. standard. 26 27 Each model was assigned an initial Al/ Al ratio and planetesi- mal radius, which together determined the internal evolution. The 26 27 −5 −5 icate body (Schmeling et al., 2008; Crameri et al., 2012). For an Al/ Al varied from 2.5 × 10 to 0.7 × 10 (corresponding in-depth discussion of the simulation method see Golabek et al. to formation times t = 0.75 to 2.0 Ma after CAI, using the form 26 (2014). The thermomechanical evolution regimes of early solar sys- assumption of an initially homogeneous Al distribution in the 26 27 tem planetesimals are described in Lichtenberg et al. (2016). protoplanetary disk with the canonical initial Al/ Al of 5.25 × −5 10 ; Kita et al., 2013) and planetesimal radii R from 20 to 4. Results 200 km. The silicate melt fraction was linearly parameterized tak- ing into account both consumption and release of latent heat. For 4.1. Platinum isotope data models reaching high melt fractions above 0.4, the convective heat flux was approximated following soft turbulence scaling relations The three samples Caddo County, Canyon Diablo and Toluca dis- (Kraichnan, 1962; Siggia, 1994). The model incorporates an initial macroporosity of 0.3, with sintering and compaction effects param- play Pt isotope compositions that overlap within uncertainty with the terrestrial value (Table 2; Fig. 1). In contrast, Cranbourne and eterized according to laboratory experiments (Henke et al., 2012; Gail et al., 2015). An ambient and starting temperature of 290 K Magura both exhibit small but resolvable positive shifts (ε Pt = was employed. Cooling at the surface was handled via a low- 2.13 and 2.08 ± 0.73; ε Pt = 0.13 and 0.11 ± 0.09, respectively), density and low-viscosity layer of ‘sticky air’ surrounding the sil- whereas Odessa shows large excesses (i.e., ε Pt = 12.99 ± 0.73; A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 493 ε Pt = 0.56 ± 0.09) relative to the terrestrial standard. Living- The Pt isotope compositions of the IAB irons lie within un- stone falls within uncertainty of the terrestrial value in the ε Pt certainty of the neutron capture model trends for their Ir/Pt ra- 196 194 196 vs. ε Pt diagram, but not when considering ε Pt vs. ε Pt tios (Fig. 1). The Pt isotope deviations of Cranbourne, Livingstone, space. Magura and Odessa from the terrestrial value fit the model predic- tions well and can be accounted for solely by GCR exposure, with 4.2. Tungsten isotope data Odessa being the most strongly exposed of the meteorites studied here. The data for Caddo County, Canyon Diablo and Toluca also fall The IAB irons show variable W isotope compositions (Table 2). on the model trend but overlap with the terrestrial standard value, Toluca displays the least negative ε W (6/4) value among the signifying that these samples were only weakly-exposed to GCR. IABs (−2.73 ± 0.09). The ε W (6/4) values of Caddo County The good fit of our Pt isotope data to the neutron capture model and Canyon Diablo are −2.92 and −2.87 ± 0.10, respectively, also confirms that nucleosynthetic Pt isotope variations are absent, whereas Cranbourne, Livingstone and Magura yield identical val- in agreement with previous studies (Hunt et al., 2017b; Kruijer ues within uncertainty (Table 2). Odessa displays the most negative et al., 2014a, 2013a; Peters et al., 2015; Wittig et al., 2013). This ε W (−3.61 ± 0.10). Tungsten isotope ratios may be affected demonstrates that s- and r-process Pt isotopes were distributed by NFS effects that arise during chemical separation of W, partic- homogeneously in the solar nebula in the region sampled by iron ularly ratios that include and/or are normalized using W; all meteorites and the Earth. W isotopic ratios have been corrected for these effects (Supple- mentary Tables S1 and S2; Cook and Schönbächler, 2016). After 5.2. Pre-exposure ε W using combined Pt–W systematics correction for NFS effects, the ε W (6/3) values overlap with 186 184 182 196 those obtained with the W/ W normalisation scheme (Ta- In an ε W (6/4) vs. ε Pt diagram, the data fall on a well- ble 2). Furthermore, all ε W (6/3) data are within uncertainty defined linear regression with a slope of −1.46 ± 0.34 (Fig. 2a). 182 192 of the terrestrial bracketing standard (see Supplementary Mate- The correlation between ε W (6/4) and ε Pt is dependent on rial) and indicate that the IAB irons are not characterized by nu- the Ir/Pt of the samples, and is therefore subject to scatter (Fig. 2b). cleosynthetic variations in W isotopes (e.g., Kruijer et al., 2013a; Normalisation of all samples to Ir/Pt = 0.4 (ε Pt ) reduces norm Kruijer et al., 2017). Thus, no corrections are needed for such ef- this scatter and yields a correlation with a slope of −0.06 ± fects on ε W. 0.03. These trends are predicted by the neutron capture models of Leya and Masarik (2013). The neutron capture model trends and 5. Discussion the regression through our data mostly overlap within uncertainty (Fig. 2). The model trends, however, feature steeper slopes such 5.1. Origin of Pt isotope variations that they overestimate the GCR-induced effects on ε W at more extreme ε Pt values (Fig. 2a). This was previously observed for Platinum isotope variations can have two origins. Firstly, vari- IVB irons (Kruijer et al., 2013a; Wittig et al., 2013), and may be due ations could reflect the heterogeneous distribution of isotopes to variations in the compositions of the samples, which slightly al- formed in specific stellar environments (nucleosynthetic varia- ters the neutron energy spectrum and is not accounted for in the tions) and secondly, isotope ratios can be altered due to expo- modelling (Kruijer et al., 2013a, 2013b). Alternatively, it may re- sure to GCR. Nucleosynthetic isotope variations are reported in flect the under-production of Pt relative to W in the GCR model iron meteorites for elements including Mo (Burkhardt et al., 2011; (Hunt et al., 2017b). Regardless, the empirically determined slope 182 196 Dauphas et al., 2002; Worsham et al., 2017), Ru (Chen et al., 2010; between ε W (6/4) and ε Pt identified here (−1.46 ± 0.34) Fischer-Gödde et al., 2015) and Pd (Ek et al., 2017; Mayer et al., agrees well with those defined previously (−1.32 ± 0.11, n = 5; 2015), but have not previously been recognised for Pt (Hunt et Kruijer et al., 2014a), and further demonstrates the robustness and al., 2017b; Kruijer et al., 2014a, 2013a; Peters et al., 2015; Wittig usefulness of Pt isotopes as a neutron dose proxy. et al., 2013). The low-abundance isotope Pt (0.79%) is the only Following the approach adopted previously (Kruijer et al., major s-process nuclide of Pt, whereas the more abundant iso- 2014a, 2013a; Wittig et al., 2013), we estimated the ε W values 194 195 196 198 topes Pt, Pt, Pt and Pt are all almost uniquely produced of our samples before GCR exposure, which in turn can be used to by the r-process (Bisterzo et al., 2011). Therefore, variable mix- constrain the timing of the last metal–silicate equilibration on the ing of different nucleosynthetic components should result in Pt IAB parent body. Using this method, the intercepts of the regres- variations relative to the other Pt isotopes. Four of the IAB irons 192 196 sions through the IAB samples where both ε Pt and ε Pt norm (Cranbourne, Livingstone, Magura, Odessa) display well-resolved Pt 192 are 0 yield the ε W value before GCR exposure. These pre-GCR isotope excesses (Table 2; Fig. 1). However, these ε Pt excesses 194 196 ε W (6/4) values are −2.93 ± 0.15 and −2.89 ± 0.07 when cor- correlate with positive shifts in ε Pt and ε Pt and therefore 192 196 related against ε Pt and ε Pt, respectively (Fig. 2). Pre-GCR norm the Pt isotope variations do not follow the predicted nucleosyn- 182 192 196 ε W (6/3) calculated using ε Pt and ε Pt yields −2.93 ± thetic pattern. Additionally, since some samples are identical to the norm 0.19 and −2.89 ± 0.15, respectively. A weighted average of the terrestrial standard for all Pt isotope ratios, this provides further strong evidence for the absence of nucleosynthetic variations. four pre-GCR ε W values combines the intercepts defined by the Irradiation of a meteoroid by GCR can readily alter Pt iso- four regressions and yields a single pre-GCR exposure ε W of tope ratios and expected neutron capture effects were modelled −2.90 ± 0.06 for the investigated IAB irons. Based on this value, by Leya and Masarik (2013). Excesses in ε Pt are generated by the timing of metal–silicate differentiation can be calculated rela- 195 196 neutron capture on Pt, creating Pt, and are dependent on tive to CAI formation using the following equation: exposure time, radius, depth within the meteoroid and matrix 182 182 composition (Leya and Masarik, 2013). Platinum-192 is produced 1 ε W − ε W pre-GCR chondrite 191 t =− ln (1) CAI by neutron capture on the relatively abundant Ir (37.3%), fol- 182 182 λ ε W − ε W CAI chondrite lowed by β-decay. Therefore, ε Pt shifts due to GCR exposure 182 182 are also highly dependent on the Ir/Pt ratio of the meteorite (Leya where ε W is the pre-GCR exposure ε W calculated pre-GCR and Masarik, 2013). The IAB irons studied here have Ir/Pt ratios of for the IAB irons, ε W is the composition of carbona- chondrite ∼0.4, except Livingstone (Ir/Pt = 0.17; Table 1; Wasson and Kalle- ceous chondrites (−1.9 ± 0.1; Kleine et al., 2004), and ε W CAI meyn, 2002). is the solar system initial determined from CAIs (−3.49 ± 0.07; 494 A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 gested to have formed separately through independent impacts into the parent body, forming distinct melt pools (Choi et al., 1995; Wasson and Kallemeyn, 2002; Worsham et al., 2016, 2017). There- fore, these melt pools may have different metal–silicate separation ages, depending on the timing of the melt-generating impact. Sam- ples from all four groups, however, fall on the same regression 182 19i for ε W vs. ε Pt. A regression through the MG-IAB irons alone yields a pre-exposure ε W of −2.95 ± 0.04, which is identical within uncertainty to the pre-exposure ε W value obtained from the regression of all seven IAB samples (ε W =−2.90 ± 0.06). Moreover, Caddo County, Toluca and Livingstone are not resolvable from the members of the main group IAB irons (Fig. 2). This fur- ther corroborates that the MG irons and sub-groups investigated here experienced metal–silicate differentiation contemporaneously. However, there is some scatter within our dataset. A study with higher precision would be needed to clarify whether this is solely due to analytical scatter or represents temporally distinct melting events on the IAB parent body, particularly between the MG irons and sLL sub-group. Nucleosynthetic Mo isotope variations indicate that the IAB sub-groups may originate from numerous parent bodies (Worsham et al., 2017). Nucleosynthetic variations have been described be- tween some iron meteorite groups for ε W (e.g., Burkhardt et al., 2012; Kruijer et al., 2013a, 2017) and this may allow multiple parent bodies for the IAB complex to be distinguished. However, no W nucleosynthetic variations are present between the sam- ples measured here (Table 1), consistent with a single parent body origin. Furthermore, a lack of nucleosynthetic variations in Pd iso- topes between all groups studied here (Ek et al., 2016), and a homogeneous Mo isotope composition between the MG and sLL sub-group (Worsham et al., 2017)permit derivation from a single parent body. 5.3. Comparison to previous studies 182 196 Fig. 2. (a) Correlation of ε W (6/4) with ε Pt, resulting in a pre-GCR exposure 182 182 192 ε W of −2.89 ± 0.07 and (b) correlation of ε W (6/4) with ε Pt, resulting a The age of 6.0 ± 0.8 Ma after CAI formation is more precise, but 182 192 pre-GCR exposure ε W of −2.93 ± 0.15 against ε Pt . Data in Fig. 2b were norm in good agreement with several other studies (e.g., Markowski et normalized to Ir/Pt = 0.4 (closed symbols). Where normalisation resulted in a dif- al., 2006; Qin et al., 2008; Schulz et al., 2009). Moreover, ametal– ference to ε Pt, the original data are shown as open symbols. Regression lines silicate separation age of 6.0 ± 0.8 Ma after CAI agrees well and error envelopes were calculated using ISOPLOT (Ludwig, 2003). Also displayed are GCR model calculations from Leya and Masarik (2013) for an exposure time of with a postulated silicate melting event recorded in the winon- 1000 Ma, Re/W = 0.25, Os/W = 3.125 and Ir/Pt = 0.4 (dashed line). The GCR model aites at 6.6 ± 2.9 Ma (Schulz et al., 2010; recalculated relative to lies within error of the regression lines, however, Fig. 2a shows that the model tends the latest CAI value, Burkhardt and Schönbächler, 2015; Kruijer et 182 196 i towards more negative ε W for higher ε Pt. Uncertainties on ε Pt are based on i al., 2014b), further corroborating the close relationship between the North Chile in-house standard (2 S.D.); uncertainties on ε W are based on the W bracketing standard (NIST SRM 3163) or the long-term reproducibility defined by IABs and winonaites that is suggested by geochemical and textu- repeated measurements of SRM 129c (Cook and Schönbächler, 2016; see Table 2 for ral similarities, and Mo and O isotopes (e.g., Benedix et al., 2000; further details), whichever is larger. Clayton and Mayeda, 1996; Worsham et al., 2017). Our new age also agrees with the value of 6.9 ± 0.4 Ma (recal- Burkhardt and Schönbächler, 2015; Kruijer et al., 2014b). This culated relative to the newest CAI estimate; Burkhardt and Schön- equation yields a model age for Hf–W fractionation, assuming a bächler, 2015; Kruijer et al., 2014b) determined by Schulz et al. previous uniform reservoir with a chondritic Hf/W ratio. To facili- (2012) for the majority of IAB irons studied therein (Fig. 3). How- tate comparison and consistency with previous studies (i.e., Kruijer ever, Schulz et al. (2012) do not obtain a single pre-exposure age et al., 2014a, Schulz et al., 2012, 2009; Worsham et al., 2017), the for the IAB samples also studied here (e.g., Caddo County). Those uncertainties on the CAI and CI values are not taken into account, authors used a correlation of ε W with CRE ages (determined both for our new data and for the data from previous studies recal- using cosmogenic noble gases) defined by the majority of their culated relative to the latest value for CAI. Using our new weighted samples to obtain a pre-exposure ε W. This approach works to average (ε W =−2.90 ± 0.06), equation (1) yields a time of a first order, but our data demonstrate that longer CRE ages do metal–silicate separation in the IAB parent body of 6.0 ± 0.8 Ma not necessarily lead to a linear increase in ε Pt (Fig. 4a), and after CAI formation. Alternatively, our best single regression (−2.89 by extension ε W. This is because cosmogenic noble gases are 182 196 ± 0.07 for ε W (6/4) vs. ε Pt) defines a nearly-identical metal– generally produced by high energy reactions at shallow depths in silicate separation event at 6.1 ± 0.9 Ma after CAI formation. a meteoroid, whereas W and Pt isotopes undergo lower energy A single metal–silicate separation age for all IABs investigated (thermal–epithermal) neutron capture reactions at greater depths here may be surprising for the following reason. Of the seven (e.g., Kruijer et al., 2013a; Leya and Masarik, 2013; Wittig et al., samples analyzed, only four (Canyon Diablo, Cranbourne, Magura 2013). Furthermore, the (epi-)thermal neutron fluence also de- and Odessa) belong to the IAB main group. Toluca belongs to pends on the pre-atmospheric radius and depth of the sample the sLL sub-group, Livingstone (TN) is part of the Algarrabo duo, within the meteoroid (Fig. 4b), and this effect is not accounted and Caddo County is ungrouped. These different groups are sug- for by CRE ages. Therefore, CRE ages are suitable to identify unex- A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 495 posed samples, but do not provide a means of accurate correction of the GCR effects on ε W. An age of 6.0 ± 0.8 Ma is also consistent with that determined for the sLL and sLM sub-groups (5.0 ± 1.0 and 5.1 ± 0.6 Ma, re- spectively; Fig. 3) by Worsham et al. (2017). The segregation of the sLM sub-group at this time adds further evidence for a body-wide event. However, Worsham et al. (2017) determined an earlier age of metal–silicate segregation for the MG (3.4 ± 0.7 Ma). Combin- ing our dataset with the MG of Worsham et al. (2017) yields a pre-exposure ε W of −3.01 ± 0.08, which is similar to the value deduced from our data only and corresponds to metal–silicate seg- regation at 4.6 ± 0.9 Myr after CAI. For the further discussion we will use the age derived from our data only because the dis- crepancy between the two datasets may be partly a result of the method of GCR correction employed by Worsham et al. (2017). For 3 out of the 4 MG samples analysed in that study, the neutron- dose proxy (Os isotopes) and W isotopes were not determined on the same sample aliquot, but on separate digestions of the sam- ples. Models of GCR effects in iron meteorites indicate that moving 2–3 cm deeper into an iron meteorite results in a ε W (6/4) shift which is double that modelled for ε Os (Leya and Masarik, 2013). This is supported by correlated Os and W isotope ratios for Fig. 3. Overview of pre-exposure ε W for the IAB irons from this study (dark-grey the IID iron meteorite Carbo and the IVB irons, where series of band), compared to data from Schulz et al. (2012) and Worsham et al. (2017), rela- variably exposed samples display offsets induced by GCR that are tive to the CAI value (light-grey band; Kruijer et al., 2014b). Our study demonstrates 182 182 189 that the wide range of scatter in ε W for the IABs is partly related to neutron cap- greater for ε W than for ε Os (Qin et al., 2015; Wittig et al., ture and not distinct timings for individual metal–silicate separation events on the 2013). An inaccurate GCR correction may therefore induce errors parent body. of ∼2–4 ppm to ε W, assuming aliquots for W and Os analy- ses are sampled within 3 cm of each other. However, this effect is very unlikely to fully account for the discrepancy between our study and the MG data of Worsham et al. (2017). Nevertheless, it demonstrates the importance of determining W isotopes and a neutron-dose proxy such as Pt or Os isotopes on the same sample aliquot, particularly when high-precision data are required. Using our approach, all IAB samples define a single value for pre-GCR exposure ε W. This demonstrates that, within the reso- lution of this study, the analysed IAB samples experienced metal– silicate separation contemporaneously and several distinct melting events, as previously suggested, cannot be resolved (Schulz et al., 2012; Worsham et al., 2017). 5.4. Thermal evolution of the IAB parent asteroid 5.4.1. Hypotheses for IAB parent asteroid evolution Two hypotheses have been proposed to explain the evolution of the IAB parent body. Local impact-generated melt pools were pro- posed based on element concentration data of IAB metals, which display only limited fractional crystallisation trends caused by modest separation of solid metal from the remaining metal melt. At the same time, the IAB concentration data show significant scatter and this was accommodated by invoking various impact- related melt pools (Choi et al., 1995; Wasson and Kallemeyn, 2002; Worsham et al., 2016). However, our new data show that the main group IAB irons, multiple sub-groups (sLL and sLM), plus one duo and an ungrouped sample, experienced metal–silicate differenti- ation contemporaneously. This uniform age argues for a single global-scale metal–silicate separation event. Our data cannot ex- clude the scenario in which the five groups formed in separate melt pools induced by impacts. Nonetheless, in order to recon- Fig. 4. (a) Plot of ε Pt vs. cosmic ray exposure age. This graph shows no lin- 196 196 cile the coeval metal–silicate separation time of the MG-IAB irons ear increase of ε Pt with CRE age because ε Pt is strongly dependent on the and sub-groups with impact-related heat sources, multiple impacts depth of the sample below the surface. CRE ages are therefore not suitable to cor- rect for neutron capture effects in iron meteorites. See Table 1 for sources of CRE large enough to induce significant melting would have to occur al- ages. (b) Expected GCR-induced shift in ε Pt for meteoroids with pre-atmospheric most simultaneously, or a single impact generated several different radii between 25 and 120 cm and an exposure time of 875 Ma (Leya and Masarik, large melt pools. Moreover, it has been argued that impacts alone 2013). The offset in ε Pt is strongly dependent on both the pre-atmospheric ra- cannot provide enough heat to melt significant fractions of a par- dius and the depth of the sample below the surface. Dashed black line and grey ent body, thereby excluding them as a global heat source (Ciesla et shaded area show ε Pt for Odessa (exposure time = 875 ± 70 Ma; Voshage and Feldmann, 1979). al., 2013; Davison et al., 2012). 496 A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 Fig. 5. Overview of pre-GCR exposure ε W and corresponding metal–silicate sepa- ration ages, determined by correlation with Pt isotopes for magmatic iron meteorite Fig. 6. Time evolution of the interior region experiencing temperatures in agreement groups (Kruijer et al., 2014a, 2013a; Wittig et al., 2013) and IAB irons (this study), with temperature constraints (1240 K – upper bound of each band, and 1470 K – plus ε W for weakly-exposed ureilites (Budde et al., 2015). This illustrates the lower bound of each band) until 14 Ma after CAI. Any model eligible as the IAB par- earlier metal–silicate separation age of the magmatic irons compared to primitive ent asteroid configuration must exhibit the above temperatures from the moment achondrites and related parent bodies (ureilites, IAB irons). Grey shaded area de- of metal–silicate separation (6.0 ± 0.8 Ma) until the time of catastrophic break- notes CAI and associated uncertainty (Kruijer et al., 2014b). up (10–14 Ma after CAI). The radii and formation times for models (A), (B) and (C) are given in the figure. Model (A) cooled too fast to sustain sufficiently high Benedix et al. (2000) proposed an alternative model, which ar- temperatures until break-up, while model (B) reached temperatures above 1470 K gues that the decay of short-lived radionuclides produced enough before metal–silicate separation and was thus discarded. Model (C) and all 2D mod- els formed at 1.5 Ma with radii larger than 60 km experienced internal temperatures heat for incipient partial melting. This was followed by a catas- within the correct temperature bracket between the estimated times for metal– trophic impact causing break-up and reassembly of the parent silicate separation and catastrophic break-up, and featured up to 60% as the eligible body, which led to the juxtaposition of lithologies such as un- volume. The dashed line, (C) V , illustrates this volume for model (C). IAB melted, chondritic silicates mixed with molten metal. In the con- text of their model, they put forward that metamorphism, par- perature boundary. Additionally, the majority of the body reached tial melting and incomplete separation of metallic melt from the a maximum temperature of ∼1470 K, based on peak tempera- residue occurred to explain the element concentrations of the met- tures obtained from both silicate clasts and winonaite meteorites als and silicate inclusions. At the same time, it is likely that some (Benedix et al., 1998, 2000) and winonaite trace-element geochem- of the metal, in particular incompletely extracted metal closer to istry (Hunt et al., 2017a). A body in this temperature range can- the surface, had already solidified by the time of impact and re- not fully differentiate, thus allowing for the preservation of the assembly. These metal pools could have been locally remelted by range of compositions observed in IAB irons and the winonaites. the disrupting impact, adding to the scatter in the element concen- This includes both the residues of silicate melting and the melts tration data of the IABs. Theis et al. (2013) proposed a time interval themselves, plus abundant chondritic lithologies (e.g., Benedix et for the catastrophic break-up and reassembly of the parent body 107 107 al., 2000; Hunt et al., 2017a; Ruzicka and Hutson, 2010). How- based on the Pd– Ag system and other chronometers that all ever, partial core formation could set in via incomplete percolative record a heating event around 10 to 14 Ma after CAI. A model of flow through the silicate matrix, which would leave portions of internal heating plus later break-up and reassembly may provide the metal in the mantle (Bagdassarov et al., 2009; Cerantola et an explanation for the differentiation event recorded by the IAB al., 2015). An additional constraint is provided by the mixing of groups studied here. However, the age of the IABs is significantly molten metal with cool silicates during the parent body break-up younger than those of the magmatic iron meteorites, which likely and reassembly at 10 to 14 Ma after CAI (Benedix et al., 2000; differentiated within the first 1–4 Ma after CAI formation as a re- Theis et al., 2013). This implies that molten metal is still present sult of internal heating (Fig. 5; Kruijer et al., 2014a, 2013a; Wittig at this time, at some depth in the pre-impact IAB parent body and et al., 2013). These different timescales merit an explanation. requires a portion of the body to remain above ∼1240 K. Our new study dates metal segregation to 6.0 ± 0.8 Ma, and hence, metal 5.4.2. Numerical models of IAB asteroid thermal evolution must stay molten for at least another 4 Ma after metal–silicate The new and precise metal–silicate separation age (6.0 ± separation. We used these constraints in our thermal models to 0.8 Ma after CAI) provides constraints on the accretion time and identify planetesimals that satisfy these temperature criteria at the subsequent thermal evolution of the IAB parent body. Thermal time of metal–silicate separation and break-up (Fig. 6). models incorporating internal heating by radioactive nuclides ( Al Depending on initial radius and formation time, and adopting and Fe) can help to constrain whether it is possible to model the assumption of an initial homogeneous distribution of Al in the features described for the IAB parent body via internal heat- the protoplanetary disk, the numerical models display different ing, and if so, to constrain the relationship between the timing of accretion and metal–silicate separation as well as parent body size. evolutions. Larger radii and earlier formation times lead to higher internal temperatures. The radius primarily controls the cooling Brecciated silicate clasts are present in many IAB samples (Ta- ble 1; e.g., Benedix et al., 2000, 2005; Buchwald, 1975; Ruzicka and timescale, such that small bodies cool quicker than larger ones (Fig. 6). Therefore, larger bodies sustained specific temperature Hutson, 2010) and these textures, combined with peak tempera- tures obtained from both silicate clasts and the related winonaite regimes at a specific radius for longer than smaller bodies. Devia- tions between 2D and 3D models resulted from changed surface- meteorites set upper and lower bound temperatures for the timing of metal–silicate separation and break-up. Textures in both the IAB to-volume ratios in the 3D runs, which enabled these 3D models to cool more efficiently. Modelling a large number of bodies with irons and winonaites indicate the presence of molten metal, which implies that large portions of the parent body reached at least the varying radii and formation times, combined with the constraints cotectic melting point for the Fe, Ni–FeS system (∼1220–1261 K; from metal–silicate separation age and break-up time (see above), Kubaschewski, 1982; Kullerud, 1963) and sets a minimum tem- limits the possible formation time and radius of the IAB asteroid A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 497 Fig. 7. Evaluated parameter space showing 2D (circles) and 3D (squares) numeri- cal simulations of the interior thermal evolution of potential IAB parent asteroid bodies. Models marked with a diamond were run using both geometries, with the upper part of the diamond indicating the obtained value for the 2D runs and the lower part representing the 3D results. Gray intensity scales display the fraction of eligible volume at 10 Ma for each model (which for 2D models was interpolated to a 3D volume) satisfying the temperature constraints at the estimated times for metal–silicate separation and catastrophic break-up (cf., Fig. 6 for three showcase 26 27 examples). Only bodies formed at 1.4 ± 0.1 Ma (or equivalently, Al/ Al ∼(1.35 ± −5 0.13) × 10 ) robustly satisfy the temperature criteria derived for the IAB parent asteroid body. parent body (Fig. 7). Many modelled parent bodies show only very narrow regions (eligible volume, V ) within the correct temper- IAB ature regime or do not sustain temperatures above 1240 K until Fig. 8. Overview of the temporal evolution of the IAB parent body. (a) Accretion 10 Ma after CAI or later (Figs. 6 and 7); these bodies were dis- of chondritic material at ∼1.4 Ma after CAI. Parent body has a radius of 60 km or greater. (b) Metal–silicate differentiation at 6.0 ± 0.8 Ma. This is a result of internal carded as potential IAB parent bodies. heating by short-lived radionuclides. Molten metal was not efficiently extracted into Model planetesimals with radii larger than 60 km and forma- a core, and also existed as pools and veins. (c) Prior to the catastrophic impact, the tion times t = 1.4 ± 0.1 Ma after CAI satisfy the temperature form body began to cool and metal pools and veins in the outer parts of the body may constraints for a significant portion of their interior, which qual- have begun to solidify. However, liquid metal was still present. (d) Impact-related ifies them as eligible parent asteroids for the IAB meteorites. In- break-up and reassembly of the parent body. This event occurred between 10 and 14 Ma after CAI (Theis et al., 2013)and led to extensive mixing of lithologies in the ternal heating is therefore a viable mechanism for the formation parent body. of the IAB irons. These models reached the appropriate temper- ature regime throughout major parts of their interiors for more 182 182 than 10 Ma after their formation. Additionally, some 2D mod- of percolation, the Hf– W system may not have fully closed els with an earlier accretion time (t = 1.25 Ma) and radius to re-equilibration, thereby explaining the time gap between the form 60 km ≤ R ≤ 120 km showed minor eligible volume fractions models (Fig. 6) and the recorded metal–silicate separation age. which could be the source of IAB meteorites. However, these ex- Variable closure times for the metal pools may also explain the perienced temperatures above 1470 K before the recorded time scatter between the MG and sLL noted in our dataset (section 5.2). 182 182 for metal–silicate separation, and thus should have undergone However, the Hf– W system must essentially be closed by complete metal–silicate separation accompanied by closure of the ∼6 Ma in order to preserve the measured W isotope ratios re- 182 182 Hf– W system. These models were therefore discarded as vi- ported here. This may happen as the liquid metal coalesces into able solutions. larger ponds and diffusion from silicates becomes negligible. Addi- For all models, the time interval between Fe, Ni–FeS melting tional diffusion of W may have occurred during the break-up and 182 182 within the body and the closure of the Hf– W system was reassembly stage when the silicate inclusions were mixed into the approximately τ = 2–3 Ma (Fig. 6). Liquid metal can perco- metal. However, the initial cooling to below the Fe, Ni–FeS solidus seg late approximately x = τ k ρ g/μ ≈ 500–1000 m during (∼1220 K) must have been very fast to keep the inclusions in sus- seg this time interval assuming the following: perfect Darcy flow with pension in the liquid metal (e.g., Wasson and Kallemeyn, 2002). −17 2 −1 constant melt fraction, no compaction (for parameter references Additionally, W diffusion in silicates is very slow (10 cm s −15 2 see Bagdassarov et al., 2009), a permeability of k = 10 m at at ∼1423 K; Quitté et al., 2005). Hence, diffusion may have led −2 182 182 10% melt fraction, Fe–S shear viscosity μ of 10 Pa s, and FeS to some small-scale metamorphic resetting of Hf– W ages in −3 melt to solid peridotite density contrast of ρ = 1900 kg m at silicate inclusions (Schulz et al., 2009), but was not sufficient to 50 km distance from the center of the planetesimal. This model reset the bulk metal ages. Furthermore, Schulz et al. (2012) noted and the resulting liquid metal percolation speed allows metal– that both bulk metal and metal entrained in silicate inclusions in silicate segregation via interconnected percolation networks in the the Landes meteorite (IAB-MG) showed the same W isotope signa- 182 182 appropriate time frame of 2–3 Ma until closure of the Hf– W tures. chronometer. Furthermore, W can diffuse from silicates to metal To summarize, in an environment with homogeneous, canoni- 26 27 during cooling with diffusion length scales per time unit depend- cal Al/ Al, the IAB parent body accreted around ∼1.4 Ma and ing on temperature (Quitté et al., 2005). Therefore, if metal existed likely had a radius ≥60 km (Fig. 8a), as constrained by our nu- as fine veins and small pools in the mantle in the initial stage merical thermal models. Members of the IAB main group and the 498 A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 sub-groups investigated here underwent metal–silicate separation 6. Conclusions at 6.0 ± 0.8 Ma (Fig. 8b). This metal–silicate separation event is late in comparison to the magmatic iron meteorite parent bodies Tungsten isotopes are affected by neutron capture when ex- but still early enough to be the result of internal heating by means posed to GCR, altering their isotopic ratios. This hampers the use 182 182 of the Hf– W chronometer and can lead to erroneous ages for of the decay of short-lived radionuclides. As shown by previous modelling, a primordial chondritic shell can survive in the upper metal–silicate separation. Here, we use Pt isotopes as a neutron- dose proxy to correct for W isotope shifts in IAB iron meteorites. layers of a planetesimal (Lichtenberg et al., 2016). Therefore a body consisting of metamorphosed but broadly chondritic (winonaite- Platinum and W isotope data for the IAB irons fall along the trends of model calculations for expected neutron capture effects, and Pt like) material, with pools and veins of molten metal inefficiently extracted into a core, is feasible. Lithologies present in the IAB isotope variations are wholly due to exposure to GCR. By correlat- ing Pt and W isotope ratios, a single pre-exposure ε W value of irons show unambiguous evidence of mixing by impact, which can −2.90 ± 0.06 can be derived for the IAB irons. This corresponds be explained by a later catastrophic break-up and reassembly of the IAB parent body (Benedix et al., 2000). Thermal models in- to a metal–silicate separation age of 6.0 ± 0.8 Ma after CAI for the IAB parent body. dicate that certain parts of the parent body sustain temperatures high enough to keep liquid metal until the time it is mixed with We propose that this event is still early enough such that short-lived radionuclides caused metal–silicate separation. Numer- solid silicates during the reassembly event (Fig. 8c), although metal pools closer to the surface may have been crystallised or partially ical modelling indicates that this likely occurred in a parent body with a radius of 60 km or greater, provided that accretion occurred crystallised. This catastrophic break-up event is dated to 10–14 Ma around 1.4 ± 0.1 Ma after CAI formation. This resulted in a par- after CAI formation (Theis et al., 2013). Once reassembled, the in- tially differentiated body where metal was not fully segregated into terior portions of the body were scrambled, mixing cooler silicates the core, but formed metal veins and pools. A later impact caused into molten metal (Fig. 8d). a major disruption on the parent body between 10 and 14 Ma after CAI formation. This event fragmented the IAB parent body, 5.4.3. Implications for the timing of core formation in planetesimals which was subsequently reassembled, leading to the mixing of dif- There are only a few other parent bodies known to have ex- ferent lithologies. Our results confirm that the parent bodies of perienced delayed metal–silicate differentiation. Among these are non-magmatic iron meteorites and primitive achondrites accreted ureilites, with a postulated accretion at ∼1.6 Ma and a metal– and differentiated later than those of magmatic iron meteorites. silicate separation age of 3.3 ± 0.7 Ma after CAI (Budde et al., Alternatively, in the scenario where Al was distributed hetero- 2015), which is slightly older than that calculated here for the IAB geneously in our solar system, our models indicate that primitive irons (Fig. 5). Additionally, the parent body for the acapulcoite– achondrite and iron meteorite parent bodies may have accreted at lodranites underwent metal–silicate separation at 4.7 ± 1.1 Ma af- the same time. However, the primitive achondrite parent bodies ter CAI formation (Schulz et al., 2010; recalculated to the latest CAI 26 27 were not able to fully differentiate due to a lower initial Al/ Al estimate, Burkhardt and Schönbächler, 2015; Kruijer et al., 2014b), ratio and therefore a reduced radiogenic heat budget. thus in the same time period as the IAB parent body. Both the ure- ilite and acapulcoite–lodranite parent bodies are also interpreted to Acknowledgements have undergone incomplete partial melting and core-mantle dif- ferentiation as a result of subdued internal heating by radioactive This work was supported by the European Research Coun- nuclides (Budde et al., 2015; Golabek et al., 2014). cil under the European Union’s Seventh Framework Programme The thermal models used to predict the timing of accretion as- 26 27 −5 (FP7/2007–2013)/ERC Grant agreement No. [279779]. T.L. was sup- sume a homogeneous, canonical Al/ Al (∼5.25 × 10 ; Kita et ported by ETH Research Grant ETH-17 13-1. The numerical simula- al., 2013) at the time of planetesimal formation. Some evidence tions in this work were performed on the EULER high performance suggests that this assumption may be incorrect and that Al was computing cluster of ETH Zürich. T.L. and G.J.G. thank Taras V. heterogeneously distributed in the early solar system (Larsen et Gerya for providing the I2/I3MART code family. Parts of this work al., 2011; Schiller et al., 2015). The IAB parent asteroid may still have been carried out within the framework of the National Center form and partially differentiate via internal heating in a low Al for Competence in Research PlanetS supported by the Swiss Na- environment, but the timing of accretion cannot be as accurately tional Science Foundation. We wish to thank Frederic Moynier for constrained. Under such conditions, the parent bodies displayed 26 27 editorial handling of this manuscript, and Peter Sprung and a sec- in Fig. 6 (which are associated with specific initial Al/ Al that ond reviewer for their helpful insights. A.C.H. additionally thanks determines the temperature history of the body; see section 3.2) Gretchen Benedix for many discussions on the IABs and winon- can no longer be linked to specific formation times. Generally, the aites. We would like to thank the Smithsonian Institution (Julie bodies are shifted to earlier formation times, and the IAB parent Hoskin) for the loan of Livingstone (TN). body had to accrete before ∼1.4 Ma (as in previous models, see section 5.4.2) otherwise formation in an environment with supra- 26 Appendix A. Supplementary material canonical Al is implied. One consequence of earlier formation is the lower eligible volume fraction (V ) that remains in the cor- IAB Supplementary material related to this article can be found on- rect temperature range at ∼10 Ma after CAI for each planetesimal line at https://doi.org/10.1016/j.epsl.2017.11.034. model, because of the prolonged cooling experienced by the bod- ies. However, this does not qualitatively change our conclusions. 26 References For the case of heterogeneous Al distribution, comparisons with other early-formed bodies are not straightforward; however, one Bagdassarov, N., Solferino, G., Golabek, G.J., Schmidt, M.W., 2009. Centrifuge assisted implication would be that if Al was heterogeneously distributed percolation of Fe–S melts in partially molten peridotite: time constraints for in the early solar system, then the primitive achondrites could planetary core formation. Earth Planet. Sci. Lett. 288, 84–95. Benedix, G.K., Lauretta, D.S., McCoy, T.J., 2005. Thermodynamic constraints on the form at the same time as the magmatic iron parent bodies, but formation conditions of winonaites and silicate-bearing IAB irons. Geochim. Cos- in a lower Al environment. In this scenario, the lower radiogenic mochim. Acta 69, 5123–5131. heat budget available implies that the IAB-winonaite, acapulcoite– Benedix, G., McCoy, T., Keil, K., Bogard, D., Garrison, D., 1998. A petrologic and iso- lodranite, and ureilite parent bodies were never able to fully dif- topic study of winonaites: evidence for early partial melting, brecciation, and ferentiate, despite their postulated earlier accretion times. metamorphism. Geochim. Cosmochim. Acta 62, 2535–2553. A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 499 Benedix, G., McCoy, T., Keil, K., Love, S., 2000. A petrologic study of the IAB iron Hunt, A.C., Ek, M., Schönbächler, M., 2017c. Separation of platinum from pal- meteorites: constraints on the formation of the IAB-Winonaite parent body. Me- ladium and iridium in iron meteorites and accurate high precision analy- teorit. Planet. Sci. 35, 1127–1141. ses of platinum isotopes by multi-collector ICP-MS. Geostand. Geoanal. Res. Bild, R., 1977. Silicate inclusions in group IAB irons and a relation to the anomalous https://doi.org/10.1111/ggr.12176. stones Winona and Mt Morris (Wis). Geochim. Cosmochim. Acta 41, 1439–1456. Kita, N.T., Yin, Q.-Z., MacPherson, G.J., Ushikubo, T., Jacobsen, B., Nagashima, K., Ku- 26 26 Bisterzo, S., Gallino, R., Straniero, O., Cristallo, S., Käppeler, F., 2011. The s-process rahashi, E., Krot, A.N., Jacobsen, S.B., 2013. Al– Mg isotope systematics of the in low-metallicity stars – II. Interpretation of high-resolution spectroscopic ob- first solids in the early solar system. Meteorit. Planet. Sci. 48, 1383–1400. 182 182 servations with asymptotic giant branch models. Mon. Not. R. Astron. Soc. 418, Kleine, T., Mezger, K., Münker, C., Palme, H., Bischoff, A., 2004. Hf– W isotope 284–319. systematics of chondrites, eucrites, and martian meteorites: chronology of core Buchwald, V.F., 1975. Handbook of Iron Meteorites. Univ. California Press, Berkeley, formation and early mantle differentiation in Vesta and Mars. Geochim. Cos- California, USA. mochim. Acta 68, 2935–2946. Budde, G., Kruijer, T.S., Fischer-Gödde, M., Irving, A.J., Kleine, T., 2015. Planetesimal Kraichnan, R.H., 1962. Turbulent thermal convection at arbitrary Prandtl number. differentiation revealed by the Hf–W systematics of ureilites. Earth Planet. Sci. Phys. Fluids 5, 1374–1389. Lett. 430, 316–325. Kruijer, T.S., Burkhardt, C., Budde, G., Kleine, T., 2017. Age of Jupiter inferred from the Burkhardt, C., Kleine, T., Dauphas, N., Wieler, R., 2012. Origin of isotopic heterogene- distinct genetics and formation times of meteorites. Proc. Natl. Acad. Sci. 114, ity in the solar nebula by thermal processing and mixing of nebular dust. Earth 6712–6716. Planet. Sci. Lett. 357–358, 298–307. Kruijer, T., Fischer-Gödde, M., Kleine, T., Sprung, P., Leya, I., Wieler, R., 2013a. Neu- Burkhardt, C., Kleine, T., Oberli, F., Pack, A., Bourdon, B., Wieler, R., 2011. Molybde- tron capture on Pt isotopes in iron meteorites and the Hf–W chronology of core num isotope anomalies in meteorites: constraints on solar nebula evolution and formation in planetesimals. Earth Planet. Sci. Lett. 361, 162–172. origin of the Earth. Earth Planet. Sci. Lett. 312, 390–400. Kruijer, T.S., Sprung, P., Kleine, T., Leya, I., Wieler, R., 2013b. The abundance and iso- Burkhardt, C., Schönbächler, M., 2015. Intrinsic W nucleosynthetic isotope variations topic composition of Cd in iron meteorites. Meteorit. Planet. Sci. 48, 2597–2607. in carbonaceous chondrites: implications for W nucleosynthesis and nebular vs. Kruijer, T., Touboul, M., Fischer-Gödde, M., Bermingham, K., Walker, R., Kleine, parent body processing of presolar materials. Geochim. Cosmochim. Acta 165, T., 2014a. Protracted core formation and rapid accretion of protoplanets. Sci- 361–375. ence 344, 1150–1154. Cerantola, V., Walte, N.P., Rubie, D.C., 2015. Deformation of a crystalline olivine ag- Kruijer, T., Kleine, T., Fischer-Gödde, M., Burkhardt, C., Wieler, R., 2014b. Nucleosyn- gregate containing two immiscible liquids: implications for early core-mantle thetic W isotope anomalies and the Hf–W chronometry of Ca–Al-rich inclusions. differentiation. Earth Planet. Sci. Lett. 417, 67–77. Earth Planet. Sci. Lett. 403, 317–327. Ciesla, F., Davison, T., Collins, G., O’Brien, D., 2013. Thermal consequences of impacts Kubaschewski, O., 1982. Iron-Binary Phase Diagram. Springer. 196 pp. in the early solar system. Meteorit. Planet. Sci. 48, 2559–2576. Kullerud, G., 1963. The Fe–Ni–S system. Ann. Rep. Geophys. Lab. 1412, 175–189. Chang, C., Wänke, H., 1969. Beryllium-10 in Iron meteorites, their cosmic-ray ex- Leya, I., Masarik, J., 2013. Thermal neutron capture effects in radioactive and stable posure and terrestrial ages. In: Millman, P. (Ed.), Meteorite Research. Springer, nuclide systems. Meteorit. Planet. Sci. 48, 665–685. Netherlands, pp. 397–406. Larsen, K.K., Trinquier, A., Paton, C., Schiller, M., Wielandt, D., Ivanova, M.A., Con- Chen, J.H., Papanastassiou, D.A., Wasserburg, G.J., 2010. Ruthenium endemic iso- nelly, J.N., Nordlund, Å., Krot, A.N., Bizzarro, M., 2011. Evidence for magnesium tope effects in chondrites and differentiated meteorites. Geochim. Cosmochim. isotope heterogeneity in the solar protoplanetary disk. Astrophys. J. Lett. 735, Acta 74, 3851–3862. L37. Choi, B., Ouyang, X., Wasson, J., 1995. Classification and origin of IAB and IIICD iron Lichtenberg, T., Golabek, G.J., Gerya, T.V., Meyer, M.R., 2016. The effects of short- meteorites. Geochim. Cosmochim. Acta 59, 593–612. lived radionuclides and porosity on the early thermo-mechanical evolution of Clayton, R.N., Mayeda, T., 1996. Oxygen isotope studies of achondrites. Geochim. planetesimals. Icarus 274, 350–365. Cosmochim. Acta 60, 1999–2017. Ludwig, K.R., 2003. ISOPLOT 3.00. A geochronological toolkit for Microsoft Excel. Cook, D.L., Schönbächler, M., 2016. High-precision measurement of W isotopes in Berkeley Geochronological Center Spec. Publ. No. 4, 70 pp. Fe–Ni alloy and the effects from the nuclear field shift. J. Anal. At. Spectrom. 31, Markowski, A., Quitté, G., Halliday, A., Kleine, T., 2006. Tungsten isotopic composi- 1400–1405. tions of iron meteorites: chronological constraints vs. cosmogenic effects. Earth Crameri, F., Schmeling, H., Golabek, G.J., Duretz, T., Orendt, R., Buiter, S.J.H., May, Planet. Sci. Lett. 242, 1–15. D.A., Kaus, B.J.P., Gerya, T.V., Tackley, P.J., 2012. A comparison of numerical sur- Mayer, B., Wittig, N., Humayun, M., Leya, I., 2015. Palladium isotopic evidence for face topography calculations in geodynamic modelling: an evaluation of the nucleosynthetic and cosmogenic isotope anomalies in IVB meteorites. Astrophys. ‘sticky air’ method. Geophys. J. Int. 189, 38–54. J. Lett. 809, 180–188. Dauphas, N., Marty, B., Reisberg, L., 2002. Molybdenum evidence for inherited plane- Peters, S.T.M., Münker, C., Wombacher, F., Elfers, B.-M., 2015. Precise determination tary scale isotope heterogeneity of the Protosolar Nebula. Astrophys. J. 565, 640. 174 180 190 of low abundance isotopes ( Hf, W and Pt) in terrestrial materials and Davison, T., Ciesla, F., Collins, G., 2012. Post-impact thermal evolution of porous meteorites using multiple collector ICP-MS equipped with 10 Faraday am- planetesimals. Geochim. Cosmochim. Acta 95, 252–269. plifiers. Chem. Geol. 413, 132–145. Ek, M., Hunt, A.C., Schönbächler, M., 2016. Nucleosynthetic Pd isotope variations in Qin, L., Dauphas, N., Horan, M.F., Leya, I., Carlson, R.W., 2015. Correlated cosmogenic iron meteorites. Goldschmidt Conf. Abstr. 26, 750. Japan. W and Os isotopic variations in Carbo and implications for Hf–W chronology. Ek, M., Hunt, A.C., Schönbächler, M., 2017. A new method for high-precision palla- Geochim. Cosmochim. Acta 153, 91–104. dium isotope analyses of iron meteorites and other metal samples. J. Anal. At. Qin, L., Dauphas, N., Wadhwa, M., Masarik, J., Janney, P., 2008. Rapid accretion Spectrom. 32, 647–656. 182 182 and differentiation of iron meteorite parent bodies inferred from Hf– W Fischer-Gödde, M., Burkhardt, C., Kruijer, T., Kleine, T., 2015. Ru isotope heterogene- chronometry and thermal modeling. Earth Planet. Sci. Lett. 273, 94–104. ity in the solar protoplanetary disk. Geochim. Cosmochim. Acta 168, 151–171. Quitté, G., Birck, J.-L., Allègre, C.J., 2005. Stony-iron meteorites: history of the Gail, H.-P., Henke, S., Trieloff, M., 2015. Thermal evolution and sintering of chondritic metal phase according to tungsten isotopes. Geochim. Cosmochim. Acta 69, planetesimals. II. Improved treatment of the compaction process. A&A 576, A60. 1321–1332. Gerya, T.V., Yuen, D.A., 2003. Characteristics-based marker-in-cell method with con- Ruzicka, A., Hutson, M., 2010. Comparative petrology of silicates in the Udei Station servative finite-differences schemes for modeling geological flows with strongly (IAB) and Miles (IIE) iron meteorites: implications for the origin of silicate- variable transport properties. Phys. Earth Planet. Inter. 140, 293–318. bearing irons. Geochim. Cosmochim. Acta 74, 394–433. Gerya, T.V., Yuen, D.A., 2007. Robust characteristics method for modelling multi- Schmeling, H., Babeyko, A.Y., Enns, A., Faccenna, C., Funiciello, F., Gerya, T., Golabek, phase visco-elasto-plastic thermo-mechanical problems. Phys. Earth Planet. In- G.J., Grigull, S., Kaus, B.J.P., Morra, G., Schmalholz, S.M., van Hunen, J., 2008. A ter. 163, 83–105. benchmark comparison of spontaneous subduction models-towards a free sur- Golabek, G.J., Bourdon, B., Gerya, T.V., 2014. Numerical models of the thermome- face. Phys. Earth Planet. Inter. 171, 198–223. chanical evolution of planetesimals: application to the acapulcoite–lodranite Schiller, M., Connelly, J.N., Glad, A.C., Mikouchi, T., Bizzarro, M., 2015. Early accretion parent body. Meteorit. Planet. Sci. 49, 1083–1099. of protoplanets inferred from a reduced inner solar system Al inventory. Earth Greenwood, R.C., Franchi, I.A., Gibson, J.M., Benedix, G.K., 2012. Oxygen isotope Planet. Sci. Lett. 420, 45–54. variation in primitive achondrites: the influence of primordial, asteroidal and Schnabel, C., Ma, P., Herzog, G.F., di Tada, M.L., Hausladen, P.A., Fifield, L.K., 2001. terrestrial processes. Geochim. Cosmochim. Acta 94, 146–163. Terrestrial ages of Canyon Diablo meteorites. Meteorit. Planet. Sci. 36, A184. Henke, S., Gail, H.-P., Trieloff, M., Schwarz, W.H., Kleine, T., 2012. Thermal history Schulz, T., Münker, C., Mezger, K., Palme, H., 2010. Hf–W chronometry of primitive modelling of the H chondrite parent body. A&A 545, A135. Hunt, A.C., Benedix, G.K., Hammond, S.J., Bland, P.A., Rehkämper, M., Kreissig, K., achondrites. Geochim. Cosmochim. Acta 74, 1706–1718. Schulz, T., Münker, C., Palme, H., Mezger, K., 2009. Hf–W chronometry of the IAB Strekopytov, S., 2017a. A geochemical study of the winonaites: evidence for limited partial melting and constraints on the precursor composition. Geochim. iron meteorite parent body. Earth Planet. Sci. Lett. 280, 185–193. Schulz, T., Upadhyay, D., Munker, C., Mezger, K., 2012. Formation and exposure his- Cosmochim. Acta 199, 13–30. Hunt, A.C., Ek, M., Schönbächler, M., 2017b. Platinum isotopes in iron meteorites: tory of non-magmatic iron meteorites and winonaites: clues from Sm and W galactic cosmic ray effects and nucleosynthetic homogeneity in the p-process isotopes. Geochim. Cosmochim. Acta 85, 200–212. isotope Pt and the other platinum isotopes. Geochim. Cosmochim. Acta. Siggia, E.D., 1994. High Rayleigh number convection. Annu. Rev. Fluid Mech. 26, https://doi.org/10.1016/j.gca.2017.05.026. 137–168. 500 A.C. Hunt et al. / Earth and Planetary Science Letters 482 (2018) 490–500 Takeda, H., Bogard, D., Mittlefehldt, D., Garrison, D., 2000. Mineralogy, petrology, Wasson, J., Kallemeyn, G., 2002. The IAB iron-meteorite complex: a group, five sub- 39 40 chemistry, and Ar– Ar and exposure ages of the Caddo County IAB iron: groups, numerous grouplets, closely related, mainly formed by crystal segrega- evidence for early partial melt segregation of a gabbro area rich in plagioclase- tion in rapidly cooling melts. Geochim. Cosmochim. Acta 66, 2445–2473. diopside. Geochim. Cosmochim. Acta 64, 1311–1327. Wittig, N., Humayun, M., Brandon, A.D., Huang, S., Leya, I., 2013. Coupled W–Os–Pt Theis, K., Schönbächler, M., Benedix, G., Rehkämper, M., Andreasen, R., Davies, C., isotope systematics in IVB iron meteorites: in situ neutron dosimetry for W 2013. Palladium–silver chronology of IAB iron meteorites. Earth Planet. Sci. isotope chronology. Earth Planet. Sci. Lett. 361, 152–161. Lett. 380, 402–411. Worsham, E.A., Bermingham, K.R., Walker, R.J., 2016. Siderophile element system- Völkening, J., Köppe, M., Heumann, K.G., 1991. Tungsten isotope ratio determinations atics of IAB complex iron meteorites: new insights into the formation of an by negative thermal ionization mass spectrometry. Int. J. Mass Spectrom. Ion enigmatic group. Geochim. Cosmochim. Acta 188, 261–283. Process. 107, 361–368. Worsham, E.A., Bermingham, K.R., Walker, R.J., 2017. Characterizing cosmochemical Voshage, H., Feldmann, H., 1979. Investigations on cosmic-ray-produced nuclides materials with genetic affinities to the Earth: genetic and chronological diversity in iron meteorites, 3. Exposure ages, meteoroid sizes and sample depths de- within the IAB iron meteorite complex. Earth Planet. Sci. Lett. 467, 157–166. termined by mass spectrometric analyses of potassium and rare gases. Earth Planet. Sci. Lett. 45, 293–308.
Journal
Astrophysics – arXiv (Cornell University)
Published: Feb 28, 2018