Molybdenum and Platinum Isotope Anomalies in Iron Meteorites – Constraints on Solar Nebula Heterogeneities and Parent Body Processes

Molybdenum and platinum isotope anomalies in iron meteorites – constraints on solar nebula heterogeneities and parent body processes

Graeme M. Poole

Imperial College London Department of Earth Science and Engineering

A thesis submitted for the degree of Doctor of Philosophy (PhD) 2016

Abstract

Recent investigations revealed systematic nucleosynthetic Mo isotope anomalies in meteorites, affording clear evidence for variable excesses in p- and r-process nuclides, and hence deficits in s-process nuclides. These anomalies were interpreted as reflecting selective destruction/removal of unstable presolar components, within the framework of thermal processing models that also take into account data for other elements (e.g., Ru, Zr, Os). To test such models, this study has undertaken extensive measurements of Mo and Pt isotopes in iron meteorites, providing the most precise data for the broadest range of samples analysed to date. The data presented here are in agreement with previous studies, with all groups analysed (except the IAB/IIICD complex) displaying deficits in s-process Mo nuclides, with the extent varying between groups. This unique dataset allows, for the first time, resolution of decoupled p-process and r-process isotope effects, providing the basis for an updated thermal processing model. Mass-independent Pt isotope anomalies were also observed, but these are interpreted as entirely cosmogenic in origin, resulting from exposure of the meteoroids to galactic cosmic rays. No nucleosynthetic Pt isotope anomalies are resolvable, in accord with predictions from the updated thermal processing model. Systematic variations in the stable isotope compositions of Mo (δ98Mo) and Pt (δ198Pt) within iron meteorite groups were found, reflecting internal processes within the parent bodies. In detail, these result from isotope fractionation during metal–sulphide partitioning of Mo, and solid–liquid metal partitioning of Pt, respectively. Significantly, a previously undetected correlation between the magnitude of the nucleosynthetic Mo isotope anomalies and δ98Mo values of iron meteorite parent bodies provides novel support for the thermal processing model. However, no signatures of elemental processing in the solar nebula are resolvable in δ198Pt, as any such effects were overprinted by the isotopic fractionation that accompanied partitioning of Pt between solid and liquid metal.

Copyright Declaration

The copyright of this thesis rests with the author and is made available under a Creative Commons Attribution Non-Commercial No Derivatives licence. Researchers are free to copy, distribute or transmit the thesis on the condition that they attribute it, that they do not use it for commercial purposes and that they do not alter, transform or build upon it. For any reuse or redistribution, researchers must make clear to others the licence terms of this work.

Declaration of Originality

The contents of this thesis are entirely my own work, except where explicitly indicated by citation or acknowledgement.

Graeme Poole March 2016

Table of Contents

Chapter 1 Introduction 19

1.1 Rationale and aim of study ...... 20 1.2 Solar system formation ...... 20 1.3 Iron meteorites ...... 22 1.4 Mass-independent isotope effects ...... 24 1.4.1 Nucleosynthetic isotope anomalies ...... 24 1.4.2 Cosmogenic isotope effects ...... 29 1.4.3 Measuring mass-independent isotope effects ...... 30 1.5 Mass-dependent isotope effects ...... 31 1.5.1 Measuring mass-dependent isotope effects ...... 32 1.6 Preliminary laboratory work ...... 34 1.6.1 Reagents and materials ...... 34 1.6.2 Iron meteorite sample preparation ...... 34

Chapter 2 Nucleosynthetic molybdenum isotope anomalies in iron meteorites 37 2.1 Introduction ...... 38 2.2 Modelling of nucleosynthetic Mo isotope anomalies ...... 41 2.2.1 Normalisation to 98Mo/96Mo ...... 42 2.2.2 Normalisation to 92Mo/98Mo ...... 43 2.2.3 Normalisation to 97Mo/95Mo ...... 43 2.3 Analytical techniques ...... 44 2.3.1 Ion-exchange chromatography ...... 44 2.3.2 Mass spectrometry ...... 45 2.4 Results ...... 47 2.4.1 Standard solutions and reference materials ...... 47 2.4.2 Iron meteorites ...... 52 2.5 Discussion ...... 63 2.5.1 Molybdenum heterogeneity of solar nebula ...... 63 2.5.2 Correlation with other elements ...... 69 2.5.3 Origin of isotopic heterogeneity ...... 72 2.6 Conclusions ...... 78

Chapter 3 Molybdenum stable isotopes in iron meteorites 79

3.1 Introduction ...... 80 3.2 Analytical techniques ...... 81 3.2.1 Molybdenum double spike preparation and calibration ...... 81 3.2.2 Ion-exchange chromatography ...... 82 3.2.3 Mass spectrometry ...... 82 3.2.4 Resolving mass-dependent and mass-independent Mo isotope effects ...... 84 3.3 Results ...... 86 3.3.1 Standard solutions and reference materials ...... 86 3.3.2 Iron meteorites ...... 88 3.4 Discussion ...... 93 3.4.1 Non-magmatic iron meteorite parent bodies ...... 93 3.4.2 Magmatic iron meteorite parent bodies ...... 93 3.4.3 Bulk δ98Mo of the iron meteorite parent bodies ...... 96 3.4.4 Stable Mo isotope evidence for thermal processing in the solar nebula ...... 98 3.4.5 δ98Mo of Earth ...... 102 3.5 Conclusions ...... 103

Chapter 4 Mass-independent platinum isotope anomalies in iron meteorites 105 4.1 Introduction ...... 106 4.2 Modelling of mass-independent Pt isotope effects ...... 107 4.2.1 Nucleosynthetic Pt isotope anomalies ...... 108 4.2.2 Cosmogenic Pt isotope anomalies ...... 109 4.2.3 Resolving nucleosynthetic from cosmogenic Pt isotope effects ...... 110 4.3 Analytical techniques ...... 112 4.3.1 Ion-exchange chromatography ...... 112 4.3.2 Mass spectrometry ...... 113 4.4 Results ...... 115 4.4.1 Standard solutions and reference materials ...... 115 4.4.2 Iron meteorites ...... 117 4.5 Discussion ...... 122 4.5.1 Cosmogenic Pt isotope anomalies ...... 122 4.5.2 Nucleosynthetic Pt isotope anomalies ...... 126 4.6 Conclusions ...... 128

Chapter 5 Platinum stable isotopes in iron meteorites 129

5.1 Introduction ...... 130 5.2 Analytical techniques ...... 130 5.2.1 Platinum double spike preparation and calibration ...... 131 5.2.2 Ion-exchange chromatography ...... 132 5.2.3 Mass spectrometry ...... 132

5.2.4 Resolving mass-dependent and mass-independent Pt isotope effects ...... 134 5.3 Results ...... 135 5.3.1 Standard solutions and reference materials ...... 135 5.3.2 Iron meteorites ...... 137 5.4 Discussion ...... 139 5.4.1 Magmatic iron meteorite parent bodies ...... 139 5.4.2 Non-magmatic iron meteorite parent bodies ...... 144 5.4.3 Initial δ198Pt of solar nebula ...... 145 5.5 Conclusions ...... 146

Chapter 6 Summary 149

6.1 Implications for solar nebula evolution ...... 150 6.2 Implications for planetary differentiation ...... 151 6.3 Outlook and future work ...... 152

Bibliography 155

Acknowledgements 175

List of Figures

Figure 1.1 A proto-planetary disk surrounding a Sun-like star ...... 22

Figure 1.2 Simplified diagram of the meteorite classification scheme ...... 22

Figure 1.3 Iridium and nickel abundances in iron meteorites ...... 23

Figure 1.4 Nucleosynthetic production pathways of s- and r-process nuclides ...... 25

Figure 1.5 Schematic plot of mass-independent isotope anomalies determined by internal normalisation ...... 31

Figure 1.6 Schematic of the double spike technique ...... 32

Figure 1.7 Effect of mass-independent isotope anomalies on the mass-dependent isotope fractionations determined by the double spike technique ...... 33

Figure 2.1 Nucleosynthetic contributions to the abundances of Mo isotopes ...... 38

Figure 2.2 Molybdenum region of the chart of nuclides ...... 39

Figure 2.3 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 98Mo/96Mo ...... 42

Figure 2.4 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 92Mo/98Mo ...... 43

Figure 2.5 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 97Mo/95Mo ...... 44

Figure 2.6 External reproducibility of the Mo isotope measurements compared to literature ...... 48

Figure 2.7 Molybdenum isotope data for terrestrial standard reference materials ...... 50

Figure 2.8 ε100Mo obtained for NIST SRM 3134 Mo solutions, doped with Ru ...... 51

Figure 2.9 ε94Mo obtained for NIST SRM 3134 Mo solutions, doped with Zr ...... 51

Figure 2.10 Molybdenum isotope data for iron meteorites, normalised to 98Mo/96Mo ...... 52

Figure 2.11 Molybdenum isotope data for iron meteorites, normalised to 92Mo/98Mo ...... 53

Figure 2.12 Molybdenum isotope data for iron meteorites, normalised to 97Mo/95Mo ...... 54

Figure 2.13 Comparison of the Mo data from this study with literature ...... 58 Figure 2.14 Comparison of observed Mo isotope anomalies with those predicted by nuclear field shift theory, using normalisation to 98Mo/96Mo ...... 62

Figure 2.15 Comparison of observed Mo isotope anomalies with those predicted by nuclear field shift theory, using normalisation to 97Mo/95Mo ...... 63

Figure 2.16 Iron meteorite groups fall into two suites ...... 64

Figure 2.17 εiMo vs. εiMo, normalised to 97Mo/95Mo ...... 65

Figure 2.18 εiMo vs. εiMo, normalised to 98Mo/96Mo ...... 66

Figure 2.19 ε100Mo vs. ε95Mo, normalised to 92Mo/98Mo ...... 67

Figure 2.20 ε100Mo vs. ε92Mo, with integrated iron meteorite literature data ...... 68

Figure 2.21 ε100Mo vs. ε92Mo, with integrated chondrite literature data ...... 68

Figure 2.22 Correlation of Mo and Ru nucleosynthetic isotope anomalies in iron meteorites ...... 69

Figure 2.23 Correlation of Mo and Zr nucleosynthetic isotope anomalies ...... 70

Figure 2.24 An updated model of thermal processing of Mo nucleosynthetic components ...... 77

Figure 3.1 Isotope composition of the 100Mo-97Mo double spike ...... 81

Figure 3.2 Schematic of the Mo double spike technique in four-isotope space ...... 84

Figure 3.3 Schematic of the double spike correction procedure, for a hypothetical sample with Mo s-process deficits ...... 85

Figure 3.4 δ98Mo values of NIST SRM 3134 Mo solutions with variable ratios of spike-derived Mo to natural Mo ...... 87

Figure 3.5 δ98Mo for terrestrial standard reference materials ...... 88

98 Figure 3.6 δ Mo(7/5) of iron meteorites, corrected for mass-independent Mo isotope effects ...... 91

Figure 3.7 δ98Mo vs. Mo concentration of iron meteorites ...... 92

Figure 3.8 δ98Mo vs. Ni concentration of IIAB irons ...... 94

Figure 3.9 δ98Mo vs. Ni concentration of IIIAB irons ...... 95

Figure 3.10 δ98Mo vs. Ni concentration of IIIAB irons, showing how the parent 98 body δ Mo(bulk) value was calculated ...... 96

98 Figure 3.11 δ Mo(bulk) vs. initial liquid sulphur content of the IIAB, IIIAB, IVA and IVB parent bodies ...... 98

98 100 Figure 3.12 δ Mo(bulk) vs. ε Mo nucleosynthetic Mo isotope anomalies for parent bodies of magmatic iron meteorites ...... 99

98 Figure 3.13 Plots of δ Mo(bulk) against nucleosynthetic Mo isotope anomalies for parent bodies of magmatic iron meteorites ...... 101

Figure 4.1 Nucleosynthetic contributions to the abundances of Pt isotopes ...... 106

Figure 4.2 Effects on Pt isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Pt, using normalisation to 198Pt/195Pt ...... 108

Figure 4.3 Effects on Pt isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Pt, using normalisation to 196Pt/194Pt ...... 109

Figure 4.4 Modelled cosmogenic effects on Pt isotopes from exposure to GCR ...... 110

Figure 4.5 Integration of models for nucleosynthetic and cosmogenic Pt isotope effects ...... 111

Figure 4.6 Platinum isotope data for IIIAB iron meteorites, normalised to 198Pt/195Pt ...... 117

192 196 Figure 4.7 ε Pt(8/5) vs. ε Pt(8/5) of all iron meteorites analysed ...... 122

192 196 Figure 4.8 ε Pt(8/5) vs. ε Pt(8/5) of the IIIAB iron meteorites ...... 123

194 194 Figure 4.9 ε Pt(8/5) vs. ε Pt(6/5) of all iron meteorites analysed ...... 125

Figure 5.1 Isotope composition of the new 198Pt-196Pt double spike ...... 131

Figure 5.2 Schematic of the Pt double spike technique in four-isotope space ...... 134

Figure 5.3 δ198Pt values of IRMM-010 Pt standard solutions with variable ratios of spike-derived Pt to natural Pt ...... 135

Figure 5.4 δ198Pt of terrestrial standard reference materials ...... 136

198 Figure 5.5 δ Pt(6/4) of iron meteorites, corrected for mass-independent Pt isotope effects ...... 139

Figure 5.6 Platinum vs. iridium concentration of magmatic iron meteorites ...... 141 Figure 5.7 δ198Pt vs. Pt concentration of magmatic iron meteorites ...... 143

Figure 5.8 Platinum vs. iridium concentration of non-magmatic iron meteorites ...... 144

Figure 5.9 δ198Pt vs. Pt concentration of non-magmatic iron meteorites ...... 145

198 Figure 5.10 δ Pt(6/4) vs. Ir concentration of all iron meteorites analysed ...... 146

List of Tables

Table 1.1 List of meteorites analysed in study ...... 35

Table 2.1 Ion-exchange chemistry for separation of Mo from iron meteorites ...... 45

Table 2.2 Faraday cup configuration for Mo isotope measurements...... 46

Table 2.3 Important spectral interferences on Mo isotopes ...... 46

Table 2.4 External reproducibility of Mo isotope measurements...... 47

Table 2.5 Internal precision of Mo isotope measurements...... 47

Table 2.6 Molybdenum isotope data for terrestrial standard reference materials ...... 49

Table 2.7 Ruthenium corrections (ppm) for Mo fractions with Ru/Mo = 3 × 10−5 ...... 50

Table 2.8 Zirconium corrections (ppm) for Mo fractions with Zr/Mo = 4 × 10−5 ...... 50

Table 2.9 Molybdenum isotope compositions of iron meteorites, normalised to 98Mo/96Mo ...... 55

Table 2.10 Molybdenum isotope compositions of iron meteorites, normalised to 92Mo/98Mo ...... 56

Table 2.11 Molybdenum isotope compositions of iron meteorites, normalised to 97Mo/95Mo ...... 57

Table 2.12 Nuclear field shifts (FS) for Mo isotopes, normalised to 98Mo/96Mo ...... 60

Table 2.13 Nuclear field shifts calculated for Mo isotopes to best fit the measured IIAB and IIC data ...... 61

Table 3.1 Ion-exchange chemistry for separation of Mo from iron meteorites ...... 82

Table 3.2 Faraday cup configuration for measurements of stable Mo isotopes ...... 83

Table 3.3 Important spectral interferences on Mo isotopes ...... 84

Table 3.4 Molybdenum concentration and stable isotope compositions of the iron meteorites ...... 90

Table 3.5 Bulk δ98Mo of magmatic iron meteorite parent bodies ...... 97

Table 3.6 The s-process contributions (%) to Mo isotopes for different s-process production models ...... 101 98 Table 3.7 δ Mo(BSE) estimated for various temperatures of core formation ...... 103

Table 4.1 Ion-exchange chemistry for separation of Pt from iron meteorites ...... 113

Table 4.2 Faraday cup configuration for measurement of Pt isotopes ...... 114

Table 4.3 Important interferences on platinum isotopes ...... 114

Table 4.4 External reproducibility (2σ) of Pt isotope measurements for n = 4 bracketing IRMM-010 Pt SRM runs ...... 115

Table 4.5 Typical internal precision (2se) of Pt isotope measurements for single IRMM-010 Pt SRM runs ...... 115

Table 4.6 Platinum isotope data for terrestrial standard reference materials ...... 116

Table 4.7 Platinum isotope compositions of iron meteorites, normalised to 198Pt/195Pt ...... 118

Table 4.8 Platinum isotope compositions of iron meteorites, normalised to 196Pt/194Pt ...... 119

Table 4.9 Platinum isotope compositions of iron meteorites, normalised to 198Pt/194Pt ...... 120

Table 4.10 Platinum isotope compositions of iron meteorites, normalised to 196Pt/195Pt ...... 121

Table 5.1 Ion-exchange chemistry for separation of Pt from iron meteorites ...... 132

Table 5.2 Faraday cup configuration for measurement of stable Pt isotopes ...... 133

Table 5.3 Important spectral interferences on platinum isotopes ...... 133

Table 5.4 Platinum concentrations and stable isotope compositions of the iron meteorites ...... 138

Chapter 1

Introduction

20 Chapter 1

1.1 Rationale and aim of study

Recent advances in analytical, theoretical and observational techniques have opened up new opportunities to expand our knowledge of how the solar system, born in a stellar nursery in a spiral in the Milky Way galaxy, evolved from a giant molecular cloud into a solar nebula, and ultimately through to the planetary system we call home today. This study takes advantage of improvements in the analytical technique of mass spectrometry to focus on the processes in the solar nebula that led to the first solids and subsequent planetary formation, by investigating isotopic signatures from these events recorded in iron meteorites. The combined use of high-precision measurements of mass-independent and mass-dependent isotope effects allow us to peer through the veil of circa 4.6 Ga of cosmic evolution and gain an insight into (i) the physical conditions in the early solar nebula, (ii) the composition and distribution of material in the nebula, and (iii) the internal processes that occurred during formation of the earliest planetary bodies.

1.2 Solar system formation

The solar system is thought to have formed about 4.6 Ga ago when part of a rotating giant molecular cloud of interstellar dust and gas was disturbed and collapsed under its own gravity. The most probable source of this disturbance was a nearby supernova, which sent shockwaves through the cloud and created regions of higher density that collapsed into dense cores (Williams, 2010). Evidence for this is provided by the presence of the stable daughter nuclides of 60Fe in meteorites (e.g., Lee et al., 1998). This short-lived isotope is only created in the cores of massive stars, which end their lives in supernovae, and so its former presence in meteorites indicates one or more supernovae occurred near the solar nebula at around the time it formed. As the cloud collapsed, its rotation significantly increased due to the conservation of angular momentum, with the rate of collapse greatest at the rotational poles, generating a mass-centre (the proto-Sun). The continuing contraction of the cloud transported dust and gas inward towards the proto-Sun, while angular momentum was transported outwards. Consequently, the dust and gas surrounding the proto-Sun collapsed to the plane of rotation of the system to form the solar nebula, a flattened protoplanetary disk (e.g., Yorke et al., 1993; Dullemond et al., 2007). Depending on conditions, including temperature, density, gas/solid ratio and gas turbulence in the solar nebula, it is possible that the material within it would have been well

Introduction 21 mixed and homogenised (e.g., Boss, 2008; Ciesla, 2009). Proposed mechanisms for such mixing include viscous, gravitational and magnetic torques (Boss and Goswami, 2006). However, the detailed pathways and timescales of material transport in a gas rich disk remain poorly constrained. After approximately 100,000 years, the matter in the proto-Sun became compressed and hot enough (107 K) for H-fusion reactions to begin, and the Sun was born. With time, the disk radiated away its energy and started to cool, and eventually temperatures were low enough for the refractory elements to condense from the remaining gases (Boss, 2004). These formed dust grains, which coagulated, first via electrostatic interaction and later through collisions. Terrestrial planetary formation was then able to begin, and the processes are summarised as follows, based on Chambers (2004). As the dust aggregates grew in size, gravitational interactions took over and small bodies of a few km in size were generated. These planetesimals then underwent a period of runaway growth through collisions and gravitational interactions and instabilities. Eventually one large planetary embryo dominated each region as its gravitational influence led it to accrete any remaining planetesimals, in a process referred to as oligarchic growth. Once very few planetesimals remained, the planetary embryos began to migrate towards the centre of the accretion disk, where they collided in giant impacts. This period of chaotic growth is thought to have taken place over 10 to 100 Ma, and eventually formed the four terrestrial planets as we see them today. The asteroid belt, from where most meteorites are believed to originate, is a region between the orbits of Mars and Jupiter that was depleted of material by the gravitational influence of Jupiter, such that no larger planetary embryos were able to grow and survive. As the material never accreted onto planets, meteorites from here present a window in time to the beginning of the solar system. We can see evidence of these processes taking place today in astronomical observations. Shown in Figure 1.1 is an image captured by the Atacama Large Millimetre/sub-millimetre Array (ALMA), which shows a proto-planetary disk about 450 light-years from Earth. Clearly visible are concentric rings separated by well defined gaps, in which newly accreted planets are sweeping their orbits clear of dust and gas.

22 Chapter 1

Figure 1.1 A proto-planetary disk surrounding a Sun-like star (HL Tauri), approximately 450 light-years from Earth and over 35 km in diameter. The gaps between the multiple concentric rings are the orbits of newly formed planets, which are sweeping up remaining dust and gas. Credit: ALMA (NRAO/ESO/NAOJ)

1.3 Iron meteorites

Based on their bulk compositions and textures, meteorites are classified into two categories – differentiated and undifferentiated (Figure 1.2). Undifferentiated meteorites (chondrites) did not experience temperatures high enough to induce melting and metal–silicate segregation. As a result, they represent some of the most primitive material we have from the early solar system. Differentiated meteorites did experience melting and metal–silicate differentiation, primarily due to heat provided by radioactive decay. The achondrites thereby represent the silicate portion of differentiated meteorite parent bodies, while iron meteorites represent the metal part. Some meteorites sample both the silicate and metal portions – these are the stony-irons that in many cases represent material from core–mantle boundaries. The point of the classification scheme is to group together classes of meteorites with similar origins and shared histories, and genetic links to common parent bodies (Krot et al., 2007).

Undifferentiated Differentiated

Chondrites Achondrites Stony-irons Irons

Carbonaceous Enstatite Ordinary IAB IC IIAB IIC IID IIE IIF IIIAB IIICD IIIE IIIF IVA IVB Figure 1.2 Simplified diagram of the meteorite classification scheme, with focus on the iron meteorites. Adapted from Krot et al. (2007).

Introduction 23

Compositionally, iron meteorites consist primarily of Fe-Ni alloys, with minor amounts of Co, P, S and C (Goldstein et al., 2009). The chemical classification into groups is based on compositional trends in plots of the siderophile elements Ni, Ir, Ga and Ge (Figure 1.3). About 85% of all iron meteorites fall into one of these 13 chemical groups: IAB, IC, IIAB, IIC, IID, IIE, IIF, IIIAB, IIICD, IIIE, IIIF, IVA and IVB.

Figure 1.3 Iridium and nickel abundances in iron meteorites. Twelve distinct fields are distinguishable, each representing a group. About 15% of iron meteorites do not fall into any of these groups and remain anomalous. Figure adapted from Norton (2002), using the data of J. T. Wasson.

It is thought that most groups – the magmatic irons – represent the cores of differentiated asteroids. The chemical variations within these groups are consistent with those produced by fractional crystallisation of solid metal during cooling of the cores. Consequently, it is thought that all irons from an individual magmatic group come from a single parent body. There are only three groups of non-magmatic irons (IAB, IIE, IIICD). The IAB and IIICD groups are often termed the IAB/IIICD complex, as they are most likely derived from a common parent body (e.g., Wasson and Kallemeyn, 2002). Whilst the fractional

24 Chapter 1 crystallisation trends exhibited by magmatic irons are in accord with crystallisation of a single core, the elemental compositions of non-magmatic irons are not. Instead, various models have been proposed to explain the origin of the IAB/IIICD meteorites, including crystallisation of a sulphur and carbon-rich core in a partially differentiated object (e.g., Kracher, 1985), breakup and re-assembly of a partially differentiated body (Benedix et al., 2000), or metal segregation in isolated impact melt pools on a porous chondritic body (Wasson and Kallemeyn, 2002). Models for the formation of the silicate-bearing IIE meteorites include impacts on a chondritic body (e.g., Wasson and Wang, 1986) and nascent parent body melting and differentiation (Bogard et al., 2000). It is notable that the magmatic irons have very consistent radiometric ages of within ~2 Ma of solar system formation, whereas the IAB/IIICD complex and IIE meteorites are thought to have ages at least 6 Ma younger (Markowski et al., 2006).

1.4 Mass-independent isotope effects

1.4.1 Nucleosynthetic isotope anomalies

Analyses of bulk meteorites from different parent bodies reveal heterogeneities in isotopic composition that reflect the diverse nucleosynthetic sources of solar system matter. These nucleosynthetic isotope anomalies were first recognised for gaseous elements like O and N, although these may at least partly reflect local processes, such as mass-independent isotope fractionations (e.g., Pepin and Becker, 1982; Lyons and Young, 2005). More recently, other investigations have reported variations in isotope composition between chondrites, differentiated meteorites, and the Earth for a number of refractory metals, including Mo, Ru, Ba and Sm (Andreasen and Sharma, 2006; Ranen and Jacobsen, 2006; Burkhardt et al., 2011; Fischer-Gödde et al., 2015). Elements with atomic mass numbers (A) > 60 are ideally suited for tracing solar nebula processes through nucleosynthetic isotope anomalies because the isotopes are formed by distinct nucleosynthetic reactions (Figure 1.4). The formation of most of the elements heavier than Fe occurred in one of two neutron capture processes: the s-process and the r-process. These two broad divisions are distinguished on the basis of the relative rates of neutron capture and radioactive decay. Contributions to the lighter isotopes of these elements come from the p-process.

Introduction 25

Figure 1.4 Nucleosynthetic production pathways of s- and r-process nuclides in a plot of proton vs. neutron number. Coloured squares represent stable nuclides, white squares represent unstable nuclides. Nuclides that are s-process only are shielded from the r-process by more neutron-rich nuclides. The p-process nuclei are generally the lightest nuclei of an element.

1.4.1.1 s-process nucleosynthesis

The s-process, or slow neutron capture process, involves the irradiation of nuclides at low neutron fluxes, such that the rate of neutron capture by the nuclides is low (~1 per 1000 years). If a short-lived radioactive isotope is formed by neutron capture at such conditions, it will decay before another neutron can be captured. Therefore, the s-process is limited to a path of synthesis through the most stable isotopes of the heavier elements (Figure 1.4). The s-process has two principal components: the weak and main components. The weak s-process component dominates the production of s-process isotopes with A ≤ 90, and occurs primarily during He burning in the cores of massive stars (≥ 15 solar masses) (e.g., Baraffe et al., 1992). The main s-process component synthesises the s-process isotopes with A > 90, terminating with Bi, as no further stable isotopes can be produced after this. The main component is understood to operate primarily in the He-burning shells of low-mass (1–3 solar masses) and intermediate-mass (5–8 solar masses) asymptotic giant branch (AGB) stars (e.g., Gallino et al., 1998; Truran Jr and Heger, 2003).

1.4.1.2 r-process nucleosynthesis

Neutron-rich nuclides (r-nuclides) are predominantly synthesised by the r-process, or rapid neutron capture process, involving chain reactions in which the capture rate is so rapid that even an unstable nucleus will capture a neutron before it has the opportunity to decay. Eventually, the nuclei capture enough neutrons so that they become highly unstable with extremely short half-lives. At this point, they β− decay to new nuclides, which are more stable

26 Chapter 1 and capable of capturing even more neutrons. When the neutron flux ends, the unstable isotopes transform by β− decay through intermediate steps to form stable nuclei. For isobars with more than one stable nuclei, only the most neutron-rich isotope is produced by the main r-process – the others are shielded (Figure 1.4). However, it has been demonstrated by meteoritic studies and astrophysical observations that there are at least two distinct r-processes, which are thought to be decoupled at Z = 56 (A ~ 140) (e.g., Cowan et al., 1995; Wasserburg et al., 1996; Akram et al., 2013). The ‘main r-process’ dominates the production of the heavy r-nuclides (Z > 56), and also contributes to the abundances of the light r-nuclides (Z ≤ 56). However, additional processes must also contribute to the production of the lighter r-nuclides in order to produce the observed abundances. These processes have been suggested to include charged particle reactions (CPRs) and the weak r-process, though these are not as well understood as the main r-process. Charged particle reactions typically involve neutron-induced, proton-induced, and alpha-particle-induced reactions, which lock up the neutrons into heavier nuclei up to A ≈ 130, such that there are no free neutrons available for neutron-capture and the CPRs dominate. The weak r-process differs from the main r-process in that the neutron flux is lower, with a neutron/seed ratio intermediate between the low and the high neutron/seed ratios of the s-process and the main r-process, respectively. Consequently, the weak r-process is less productive than the main r-process. Traditionally, the very high neutron fluxes required by the r-process were thought to be prevalent during Type II (core-collapse) supernova events. When the core of a star with a mass greater than 8 solar masses has been converted to Fe, the hydrodynamic equilibrium between thermal expansion and gravitational compression is upset, and this results in a Type II supernova. It begins with the collapse of the core, and once the matter in the centre of the core is compressed beyond the density of nuclear matter (3 x 1014 g/cm3), it rebounds, sending out a massive shock wave that travels outward through the core (White, 2013). The compression produces temperature increases, which result in the breakdown of nuclei by photodisintegration and these reactions provide the extremely high neutron fluxes required. The r-process is thought to have a duration of 1 to 100 seconds during the peak of the supernova explosion. It has also been suggested that the different r-processes occur at different mass levels in the neutrino driven winds from Type II supernovae ejecta (Kratz et al., 2008; Farouqi et al., 2010). It follows that CPRs dominate in the higher-density (low-entropy) mass shells close to

Introduction 27 the proto-neutron star, while the main and weak r-processes operate in the outer, lower density (higher entropy) mass shells of the neutrino driven winds. Alternatively, Wasserburg and Qian (2009) proposed that the processes operate in different masses of Type II supernovae events. In their model, the main r-process takes place in low mass (8–11 solar masses) Type II supernovae, while the weak r-process occurs in those of normal mass (12–25 solar masses). Charged particle reactions synthesize lighter r-nuclides in both settings. However, the results of various recent studies have led to a growing realisation that the source of the main r-process cannot be Type II supernovae, and there is mounting evidence that the stellar sources of the different r-processes are decoupled. Numerical simulations of Type II supernovae reveal the neutrino-driven winds are not as neutron-rich as previously assumed because of neutrino interactions (Nishimura et al., 2012; Wanajo, 2013). While the weak r-process can operate here and produce the light r-nuclides, the revised estimates of neutron fluxes make it unsuitable for main r-process nucleosynthesis; Wanajo (2013) demonstrated the winds can be the origin of no more than 10% of the observed abundances of the heavy r-nuclides. Instead, consensus now favours neutron star mergers in binary systems as the dominant site of the main r-process. Neutron stars are the remnants of core collapse supernovae, typically measuring 10 km in size and 1.3 to 2 solar masses, consisting almost entirely of neutrons and compressed to the density of nuclear matter (3 x 1014 g/cm3). Binary neutrons stars gradually lose energy through gravitational waves as they circle and fall towards one another. When they finally collide and merge into a single object, a massive flux of neutrons is released and the r-process occurs in the ejected matter. Numerical simulations of neutron star mergers have successfully produced the heavy r-process nuclides (e.g., Bauswein et al., 2013). These are backed up by recent near-infrared observations of a short-duration gamma ray burst (GRB 130603B), which provide evidence for the ejection of almost pure r-process nuclides from a neutron star merger (e.g., Tanvir et al., 2013). Neutron star mergers were previously not thought to be a suitable r-process site due to their rarity and high yield of r-process nuclides per event, which would lead to r-process enrichment that is not consistent with observations of very low metallicity stars (e.g., Argast et al., 2004). However, more recent investigations have produced models capable of satisfying these conditions (e.g., Tsujimoto and Shigeyama, 2014a; Hirai et al., 2015). The examination of dwarf galaxies, which have much fewer stars than the Milky Way and thus probably

28 Chapter 1 experienced few or no neutron star mergers, were found to have stars which show no increase in r-process elements, despite increasing metallicities. Since the metallicity is linked to nearby supernovae, this suggests that supernovae do not produce the heavy r-nuclides. Conversely, in larger galaxies, where neutron star mergers are more likely to have occurred, younger stars were found to have greater r-process contents, suggesting a progressive enrichment of such material. It was calculated that the neutron star merger event rate needed to explain this is about 1/1000th the expected supernova rate – this is comparable to the estimated rarity of neutron star mergers, thus supporting neutron star mergers as the origin of the heavy r-process nuclides. Furthermore, Tsujimoto and Shigeyama (2014a) demonstrated that while supernovae ejecta propagate through the interstellar medium (ISM) for about 100 light years, this is only a localised effect on the galactic scale. Conversely, they suggest neutron star merger ejecta propagate distances more than 1000 times further due to the higher velocities involved (about 10 to 30% of the speed of light). Therefore, the concentrations of r-process nuclides are diluted over a greater area, preventing any one region from becoming too rich in r-process nuclides, thus explaining the previously problematic lack of observed stars very rich in r-nuclides. Leading on from this, Tsujimoto and Shigeyama (2014b) concluded that Type II supernovae generate the light r-nuclides, with neutron star mergers producing the heavy r-nuclides. This view is not completely univocal: other models have suggested that it is conceivable that some neutron star mergers could produce all the r-nuclides with the correct abundances (Wanajo et al., 2014; Just et al., 2015). Therefore, no definitive consensus for identifying the overall sites of the r-process, particularly the production of the light r-nuclei, has been achieved to date.

1.4.1.3 p-process nucleosynthesis

The nucleosynthetic p-process produces proton-rich (or neutron-deficient) nuclei, which are bypassed by the s- and r-neutron capture processes. The p-nuclides are commonly 10 to 1000 times less abundant than the s- and r-nuclides (Travaglio et al., 2011). The exact nature of the p-process pathway has been a controversy since it was first proposed by Burbidge et al. (1957). The dominant p-process pathway, known as the γ-process, involves photodisintegration reactions and β+ decay in the O/Ne layers of core collapse (Type II) supernovae. However, while calculations indicate that these can produce the observed solar

Introduction 29 system abundances of the bulk of p-nuclei, it significantly under-produces those with A < 110 (e.g., Hayakawa et al., 2008; Fisker et al., 2009; Rauscher, 2010). Alternatively, it has been demonstrated that the γ-process in Type Ia supernovae can contribute at least 50% to the solar abundance of all p-nuclei (i.e., both A < 110 and A ≥ 110) (e.g., Lambert, 1992; Arnould and Goriely, 2003; Travaglio et al., 2015). Type Ia supernovae represent the violent explosion of a white dwarf star in a binary star system. They occur once the white dwarf has reached the Chandrasekhar mass limit following accretion of material from the companion star. As such, there is an ongoing debate about the source of p-nuclides, but it seems likely that both Type II and Type Ia supernovae contribute to the abundances. Additionally, two other p-process pathways, the rp- and vp-processes, are thought to contribute to the abundances of the light p-nuclei. The rp-process, or rapid p-process, is a + sequential reaction of successive proton captures and β decays. Since (γ,p) reactions are faster than proton captures, the rp-process requires very proton-rich environments – suggested sites include explosive H- and He-burning on the surface of a mass-accreting neutron star (X-ray bursts) (e.g., Schatz et al., 1998; Schatz et al., 2001). The rp-process ends in a closed SnSbTe cycle, preventing the synthesis of elements heavier than Te. The vp-process is a specific type of rp-process, occurring in the neutrino-driven winds of Type II supernovae, where intense neutrino fluxes create a proton-rich environment (Pruet et al., 2006; Wanajo, 2006; Wanajo et al., 2011).

1.4.2 Cosmogenic isotope effects

Irradiation of meteoroids by galactic cosmic rays (GCR) can cause small shifts in isotopic compositions that are resolvable by mass spectrometry. Galactic cosmic rays consist mainly of protons and alpha-particles, with very high energies of ~0.1 to 10 GeV. When these primary cosmic ray particles interact with a meteoroid surface, a cascade of nuclear reactions is generated, producing secondary protons and neutrons, with high energies of ~1 to 300 MeV. The high-energy primary and secondary particles produce cosmogenic nuclides via spallation reactions. Commonly produced nuclides include cosmogenic noble gas isotopes, which are used as a monitor to determine the cosmic-ray exposure ages of meteorites (e.g., Herzog, 2007; Beyersdorf-Kuis et al., 2015). In the nuclear reactions, positively-charged secondary protons have to overcome the Coulomb barrier to leave the nucleus, whereas non-charged neutrons can easily leave the nucleus. Consequently, the flux densities of secondary neutrons are much larger than the flux densities of secondary protons. The flux densities also depend on the radius of the meteoroid

30 Chapter 1 and the shielding depth of primary and secondary particles, as well the chemical composition of the irradiated object (e.g., Masarik and Reedy, 1994). The speed and energy of the secondary neutrons are moderated and reduced by elastic scattering with increasing depth from the meteoroid surface. The probability of a secondary neutron being captured is related to the neutron capture cross-section of the target nuclide, for which two energy regions arise where neutron capture probabilities are high: 1–10 keV (epithermal energies) and ~0.025 eV (thermal energies) (Leya and Masarik, 2013). At epithermal energies, heavy elements (including W and Pt) have high resonance integrals, whereby depending on the nuclear shell structure, the capture of neutrons in certain small energy ranges can be enhanced by orders of magnitude. In iron meteorites, the large resonance of 56Fe means that most secondary neutrons are captured before reaching thermal energies, such that most of the neutron capture reactions in iron meteorites are with epithermal neutrons (Leya and Masarik, 2013).

1.4.3 Measuring mass-independent isotope effects

In order to obtain high-precision isotope ratio measurements using mass spectrometry, instrumental mass fractionation must be accurately corrected. This is best achieved by internal normalisation. Hereby, two isotopes that are not affected by radioactive decay or ingrowth are selected to define an isotope ratio for normalisation. The ‘raw’ value that is measured for this ratio by mass spectrometry is corrected to a consensus terrestrial reference value (Equation 1.1). The β factor required for this correction is then used to correct all other isotope ratios for instrumental mass fractionation (Equation 1.2).

� ln ! � ! Equation 1.1 β = � ln ! �!

RA is the reference isotope ratio of N2/N1, where Ni is the abundance of isotope i with mass mi. rA is the fractionated ratio of n2/n1 where ni refers to the measured flux of isotope i.

! �! �! = �! Equation 1.2 �!

RB is the fractionation-corrected isotope ratio of N3/N1, where Ni is the abundance of isotope i with mass mi. rB is the fractionated ratio of n3/n1, where ni refers to the measured flux of isotope i. Wombacher and Rehkämper (2003)

However, if the normalising isotopes themselves have mass-independent isotope anomalies, then the correction procedure itself can create apparent anomalies (Figure 1.5). To evaluate

Introduction 31 and minimise these analytical artefacts, it is important to examine the isotopic pattern that results, not just the individual isotopic anomalies.

Figure 1.5 Schematic plot of mass-independent isotope anomalies determined by internal normalisation. The instrumental mass fractionation curve is shown in red. Black circles represent the results of isotope analyses following correction for instrumental mass fractionation using internal normalisation to m4/m2, whilst the red circles are the ‘raw’ uncorrected data. (a) The isotopes with masses m1, m3, m5, m6 have mass-independent isotope anomalies, which offset them from the mass fractionation curve (red circles). Correcting for instrumental mass fractionation results in the black pattern, with the anomalies of these isotopes visible. (b) Only nuclide m4 has a mass-independent isotope anomaly, offsetting it from the mass fractionation curve. However, this isotope is used in the normalisation procedure and the resulting black curve appears to show that isotopes m1, m3, m5, m6 all have positive mass-independent isotope anomalies (black circles), when in fact they do not. It must also be noted that the anomaly pattern shown in (b) can be achieved if m1, m3, m5, m6 all have true mass-independent anomalies and m4 doesn’t. Therefore, several different normalisation schemes must be used and compared, in order to determine the presence of any true mass-independent isotope anomalies.

1.5 Mass-dependent isotope effects

Measurements of stable isotope compositions can be used to determine the extent of natural mass-dependent isotope fractionation recorded by a sample relative to a standard reference material. From the 1950s to the 1990s, the most common elements that were investigated for stable isotope variations were ‘light’ elements, particularly H, O, C and S. In the last 20 years, however, advances in mass spectrometry have enabled the increasing emergence of ‘non-traditional’ stable isotope systems, with many studies of heavier (trace) metals such as Fe, Cu, Zn, and Cd. There are two principle routes to stable isotope fractionation: an equilibrium and a kinetic pathway. Equilibrium isotope effects arise from reversible isotope exchange reactions that run to completion such that full equilibrium is established. Such fractionations are primarily controlled by the quantum mechanical effects on molecular vibrations (for a more detailed discussion see Chacko et al., 2001). As a rule of thumb, the heavy isotopes tend to occupy sites with the stronger bond at equilibrium (e.g., Criss, 1999).

32 Chapter 1

Kinetic isotope fractionations result from mass-dependent differences in the reaction or diffusion rates of different isotopes. Such effects are normally associated with irreversible, incomplete or unidirectional processes, such as evaporation into a vacuum or biological uptake. The reaction product is hereby often enriched in the lighter isotopes because these move or react faster than heavier isotopes (e.g., Cole and Chakraborty, 2001). Measurements of stable isotope fractionations for non-traditional isotope systems are becoming an increasingly common and important tool in cosmochemistry. Recent studies of various elements in meteorites and their components have identified and interpreted stable isotope fractionations as reflecting primarily (i) condensation and evaporation processes in the solar nebula and on meteorite parent bodies (e.g., Moynier et al., 2010), (ii) core formation and associated metal–silicate differentiation (e.g., Armytage et al., 2011), and (iii) localised fractionation processes within parent bodies (e.g., Bridgestock et al., 2014).

1.5.1 Measuring mass-dependent isotope effects

The process of internal normalisation corrects for both instrumental and natural mass fractionation. However, if one wishes to measure the natural mass-dependent isotope fractionation of a sample, a different method must be used to correct for instrumental mass fractionation. Such a method is the double spike technique, a graphical representation of which is provided in Figure 1.6 and described below.

Figure 1.6 Schematic of the double spike technique for determining mass-dependent isotope fractionations in four-isotope space. Adapted from Galer (1999). For details see text.

Introduction 33

The double spike composition, S, is known from the double spike calibration, while n is the composition of the δ = 0 standard reference material (SRM). The isotope composition of the sample, N, differs from n as a result of the natural mass fractionation fn. The isotope composition of a mixture of sample and double spike, m, is determined by a mass spectrometric measurement. This differs from the true, unfractionated, isotope composition of the mixture, M, as a result of the instrumental mass discrimination or mass bias, fm.

In four-isotope space (Figure 1.6), the points N, n and S lie on the same plane, Pn, while M, m and S lie on another plane, Pm. These two planes intersect along a line formed by

N-M-S. From this mathematical framework, N can be calculated, and then used to determine fn relative to the isotope composition n of the SRM (for more details see Rudge et al., 2009). A key characteristic of the double spike data reduction for stable isotope analyses is, therefore, the assumption that the isotopic difference between the sample and the SRM is solely the result of mass-dependent isotope fractionation. If, however, the isotope composition of a sample is also altered by mass-independent effects, this needs to be explicitly considered in the double spike data reduction process (Figure 1.7). An additional correction is required, which takes into account the mass-independent isotope effects and thus allows selective quantification of only the mass-dependent isotope fractionation.

Figure 1.7 Effect of mass-independent isotope anomalies on the mass-dependent isotope fractionations determined by the double spike technique. For a hypothetical sample, the red line represents the mass-independent isotope effects, as measured by mass spectrometry with internal normalisation to m4/m2. The blue line represents the apparent mass-dependent isotope fractionation as determined by the double spike technique. However, this consists of both mass-dependent and mass-independent effects. Therefore, the blue line must be corrected for the effects of the red line. The true mass-dependent isotope fractionation, represented by the green line, is the difference between the blue and the red line.

34 Chapter 1

1.6 Preliminary laboratory work

1.6.1 Reagents and materials

All sample preparation and processing was performed in Class 10 laminar flow hoods in a

Class 1000 Clean Room Laboratory at Imperial College London. HCl and HNO3 were purified in-house by sub-boiling distillation in quartz stills, at concentrations of 11 M and

15 M respectively. Trace element grade Optima HBr (8 M) and Optima HClO4 (12 M) were purchased from Fischer Scientific, while trace element grade Plasma Pure Plus HF (28 M) was purchased from Qmx. All necessary dilutions were made using Millipore

18 MΩ cm H2O. Savillex PFA vials were used for sample handling, and were cleaned sequentially in

8 M HNO3, 6 M HCl and dilute distilled acids (HNO3 and HCl). Prior to use, the vials received a final cleaning by refluxing with a solution of distilled 4 M HNO3 and 0.05 M HF for Mo studies, and distilled 6 M HCl for Pt analyses.

1.6.2 Iron meteorite sample preparation

Specimens of 53 iron meteorites from 11 different groups were obtained from the Natural History Museum, London, and private collectors (Table 1.1). Following initial abrasion with silicon carbide paper (80–220 grades) to remove any visible effects of weathering and saw marks, the samples were sequentially leached in 0.5 M HBr and 6 M HCl for 1 hour in an ultrasonic bath, followed by leaching with modified aqua regia (prepared from 6 M HCl and

15 M HNO3, 3:1) for 30 minutes, to remove terrestrial contamination. The leached samples were then submersed in distilled ethanol, dried and weighed, before being digested in modified aqua regia on a hotplate at 100–140 °C.

Introduction 35

Table 1.1 List of meteorites analysed in study

Group Sample Mass (g) a Source b IAB Bitburg 1.6 NHM: BM 1985,M278 Campo Del Cielo 1 5.7 Private Collector Campo Del Cielo 2 7.5 Private Collector Campo Del Cielo 3 7.2 Private Collector Canyon Diablo 1 1.1 Private Collector Canyon Diablo 2 1.2 Private Collector Cosby's Creek 6.0 NHM: BM 1985,M291 Odessa 3.8 NHM: BM 1966,270 Toluca 7.0 Private Collector

IC Arispe 1 1.1 NHM: BM. 2005,M11 Arispe 2 0.8 NHM: BM. 2005,M11 Arispe 3 0.6 NHM: BM. 2005,M11 Arispe 4 0.7 NHM: BM. 2005,M11 Bendego 1 3.4 NHM: BM. 1938,2656 Bendego 2 1.4 NHM: BM. 1938,2656 Santa Rosa 4.4 NHM: BM 1985,M291

IIAB Coahuila 2.6 NHM: BM. 41031 Murphy 2.3 NHM: BM 84552 Negrillos 1.0 NHM: BM. 1938,2656 North Chile 3.4 NHM: BM. 1959,917 Sikhote Alin 1 7.9 Private Collector Sikhote Alin 2 2.0 NHM: BM 1992,M39

IIC Ballinoo 0.7 NHM: BM. 1920,285 Kumerina 7.2 NHM: BM 1938,220 Salt River 0.6 NHM: BM. 1985,M202

IIE Kodaikanal 2.2 NHM: BM 1920,310 Verkhne Dnieprovsk 1.6 NHM: BM 51183 Weekeroo Station 3.8 NHM: BM 1929,196

IIIAB Bear Creek 4.6 NHM: BM. 1959,973 Cape York 1.9 NHM: BM 2005,M66 Charcas 1 2.4 NHM: BM 85075 Charcas 2 1.4 NHM: BM 85075 Henbury 1 7.3 NHM: BM. 1932,1405 Henbury 2 6.9 NHM: BM 1934,337 Lenarto 2.7 NHM: BM 2005,M154 Santa Apolonia 1 0.8 NHM: BM. 1959,971 Santa Apolonia 2 2.1 NHM: BM 1959,971 Verkhne Udinsk 5.6 NHM: BM. 36012 Williamette 4.6 NHM: BM. 1977,M8

IIICD Carlton 1 3.4 NHM: BM. 65790 Carlton 2 1.7 NHM: BM. 65790 Nantan 1 5.4 Private Collector Nantan 2 6.7 Private Collector

IIIE Staunton 3.2 NHM: BM. 44761

IIIF Clark County 2.5 NHM: BM 1959,949

IVA Gibeon 1 4.1 Private Collector Gibeon 2 1.9 NHM: BM. 2005,M115 Muonionalusta 1.1 Private Collector Obernkirchen 2.2 NHM: BM 36056

IVB Cape of Good Hope 2.1 NHM: BM 1985,M246 Santa Clara 1.1 NHM: BM 1983,M27 Tlacotepec 1 0.8 NHM: BM 1959,913 Tlacotepec 2 1.0 NHM: BM 1959,913 a Mass available for analysis, after leaching; b NHM = Natural History Museum, London, United Kingdom.

Chapter 2

Nucleosynthetic molybdenum isotope anomalies in iron meteorites

38 Chapter 2

2.1 Introduction

The solar system formed from the collapse of a molecular cloud of interstellar dust and gas featuring isotopically diverse material produced by nuclear reactions in various pre-existing stellar sources. Whilst presolar grains in primitive meteorites show this cloud was isotopically heterogeneous at the grain-size level (Zinner, 2007), it was initially thought planetary bodies evolved from a hot and well-mixed solar nebula. However, various studies identified isotopic variations in meteorites for a number of refractory elements, which were interpreted to reflect planetary-scale heterogeneities in presolar matter. These include Ba (Ranen and Jacobsen, 2006), Ca (Simon et al., 2009; Chen et al., 2011; Dauphas et al., 2014), Cr (Trinquier et al., 2007; Dauphas et al., 2010), Mo (Dauphas et al., 2002; Burkhardt et al., 2011), Nd (Andreasen and Sharma, 2006), Ni (Regelous et al., 2008; Steele et al., 2012), Ru (Chen et al., 2010; Fischer-Gödde et al., 2015), Sm (Andreasen and Sharma, 2006), Ti (Trinquier et al., 2009; Zhang et al., 2012) and Zr (Schönbächler et al., 2003; Akram et al., 2015). Significantly, such findings can place critical constraints on the physical conditions within the solar nebula. Molybdenum is ideally suited as a tracer of planetary-scale isotopic heterogeneity as its seven isotopes were produced by distinct nucleosynthetic processes (Figure 2.1).

Figure 2.1 Nucleosynthetic contributions to the abundances of Mo isotopes. Data from Lu and Masuda (1994) and Arlandini et al. (1999).

The Mo isotopes with an s-process contribution (94Mo, 95Mo, 96Mo, 97Mo, 98Mo and 100Mo) were formed in the main s-process in asymptotic giant branch (AGB) stars. The atomic number (Z) of Mo is 42, i.e., < 56. Therefore, the neutron-rich isotopes with r-process contributions (95Mo, 97Mo and 100Mo) were produced, at least partly, by the weak r-process

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 39 and charged particle reactions in Type II supernovae, as well as the main r-process (e.g., Qian and Wasserburg, 2007; Farouqi et al., 2010). The s-only isotope 96Mo receives no r-process contribution due to the shielding effects of the 96Zr isobar preventing the r-process β− decay path reaching 96Mo, while 100Mo is mostly an r-process nuclide (> 95%) because the short 99 decay timescale of Mo (t½ ~ 66 hours) largely prevents the s-process path from reaching 100Mo (except via minor branching; see Figure 2.2).

Figure 2.2 Molybdenum region of the chart of nuclides. Stable isotopes are represented by black squares, with their nucleosynthetic processes of production also shown (where multiple processes involved, listed in descending order of relative contributions). Unstable isotopes and their half-lives are shown as yellow (β+ decay) and blue (β− decay) squares. The s-process path is represented by the green arrows, while the red arrows denote the direction of the r-process paths.

While p-nuclei are typically 10–1000 times less abundant than the s- and r-nuclides (Travaglio et al., 2011), this is not true in the case of Mo – the p-only 92Mo has a terrestrial abundance of 14.7% and the predominantly-p (> 99% p) 94Mo has an abundance of 9.2%, both of which are no less than a third of the most abundant s- or r-nuclei (98Mo: 24.2%). The particularly high abundance of 92Mo arises from its neutron number (N = 50) – the neutron shells are full (closed) such that the binding energies are particularly strong. As a consequence, the photodisintegration timescales significantly increase and abundances of 92Mo build up. Traditionally, most models of p-process nucleosynthesis resulted in an under- production of 92Mo and 94Mo. However, recent studies have found certain conditions under which these nuclei can be formed in the correct abundances. In the neutrino-driven winds of Type II supernovae, charged particle reactions and rapid proton captures have been found to produce 92Mo and 94Mo in the correct abundances (Wanajo, 2006; Wanajo et al., 2011), though this is contested by other models that do not produce the correct 92Mo/94Mo ratio (e.g., Fisker et al., 2009). Alternatively, it has been demonstrated that photodisintegration reactions

40 Chapter 2

(the γ-process) in Type Ia supernovae are capable of producing significant 92Mo and 94Mo abundances (Travaglio et al., 2011). It thus appears the 92Mo/94Mo ratio is highly sensitive to supernova dynamics. Despite its suitability as a tracer of isotopic heterogeneity in the solar nebula, previously obtained Mo isotopic evidence has been questioned by other investigations that presented conflicting or even contradictory data. For instance, Yin et al. (2002) used N-TIMS to determine the bulk Mo isotope compositions of two carbonaceous chondrites (Murchison and Allende), while Dauphas et al. (2002) used MC-ICP-MS to measure the same for Allende. Both these studies claimed to find s-process deficits. Conversely, Becker and Walker (2003) also examined the Mo isotope compositions in Allende (and the ordinary chondrite Forest Vail) using N-TIMS, but found no resolvable Mo isotope anomalies, and argued for efficient mixing in the solar nebula leading to isotopic homogeneity. Unlike in the chondrites, Yin et al. (2002) found no Mo nucleosynthetic isotope anomalies in iron meteorites, nor did Becker and Walker (2003). Dauphas et al. (2002), on the other hand, found anomalies in various iron groups, as well as two pallasites and a mesosiderite, which correlated with those found in Allende (but of a smaller magnitude). Possible reasons for the discrepancies in the findings of these studies include (i) the incomplete dissolution of meteorite samples (particularly their presolar grains) (Becker and Walker, 2003); (ii) analytical artefacts (Schönbächler et al., 2002; Schönbächler et al., 2003; Andreasen and Sharma, 2007); (iii) spallation or mass-independent isotope effects (Fujii et al., 2006a); and (iv) the procedure that is used to correct isotopic measurements for the effects of instrumental mass bias; this procedure involves normalisation of isotope data to an isotopic ratio that is assumed to be constant for all samples, but this assumption is questionable in some cases (e.g., Niederer et al., 1985). A recent study (Burkhardt et al., 2011) sought to solve this using a newer generation MC-ICP-MS with analytical precision several times better than early studies. By using improved digestion techniques (including CO2 laser fusion), they were able to minimise any effects of incomplete digestion of presolar grains in chondritic meteorites. They found all chondrites exhibited clear s-process deficits. Analyses of iron meteorites also yielded s-process deficits (except for the non-magmatic IAB/IIICD group), as did two pallasites measured. Additionally, analyses of Martian meteorites and an angrite revealed a terrestrial Mo isotope composition. The authors were therefore able to conclude that Earth, Mars, and the parent bodies of angrites and non-magmatic IAB/IIICD irons, accreted from material with

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 41 a higher proportion of s-process Mo than the rest of the iron meteorite, chondrite and pallasite parent bodies. Generally, the findings of Burkhardt et al. (2011) are in agreement with those of Dauphas et al. (2002). The discrepancies between these and the earlier studies most likely arise from differing levels of analytical precision – Yin et al. (2002) and Becker and Walker (2003) did not detect the smaller anomalies (i.e., in iron meteorites) because of their larger analytical uncertainties (Yokoyama and Walker, 2016). While various recent studies have uncovered nucleosynthetic isotope anomalies for a range of refractory elements (including Mo), other investigations have yielded isotopic homogeneity for different elements, such as Hf and Os (e.g., Sprung et al., 2010; Walker, 2012). Therefore, any models proposed to account for nucleosynthetic anomalies must also account for any observed isotopic homogeneity, where relevant. A number of such scenarios have been posited, including inefficient/incomplete mixing in the solar nebula (e.g., Jacobsen and Ranen, 2006; Schiller et al., 2015), late injection coupled with grain size sorting (Dauphas et al., 2010), thermal processing (e.g., Burkhardt et al., 2011), and parent body processes (e.g., aqueous alteration; Yokoyama et al., 2011). Results of the most recent studies have led to a growing consensus for the most promising model to be thermal processing of material in the solar nebula, which imparts isotopic fractionation by affecting components with certain properties linked to nucleosynthetic sources of origin. Yet there exists some controversy and debate as to what specifically controls which elements in these components will be affected, with two main views emerging. The first reasons that the exact volatile/refractory nature of the element is the primary control (e.g., Burkhardt et al., 2012b), while the second argues that the decoupling of nucleosynthesis for light vs. heavy refractory elements is responsible (e.g., Akram et al., 2015). To further elucidate the validity of the thermal processing concept, this study has obtained precise Mo isotopic data for 11 groups of iron meteorites. With better precision than previous studies, the results provide constraints on planetary-scale processes in the solar nebula and help resolve the origin of nebula-wide nucleosynthetic isotope anomalies.

2.2 Modelling of nucleosynthetic Mo isotope anomalies

The effects of increasing/decreasing the proportions of p-, s- and r-process components on the Mo isotopic patterns can be modelled using different normalisation schemes. Here,

42 Chapter 2 normalisations to 98Mo/96Mo, 92Mo/98Mo and 97Mo/95Mo are modelled, and the potential advantages and disadvantages of one over another are discussed. The different normalisation schemes effectively rotate the isotope patterns to produce correlated effects in different Mo isotopes. For consistency, all terrestrial Mo isotope abundances are obtained from the same data set (Lu and Masuda, 1994), as are all nucleosynthetic contributions (Arlandini et al., 1999). For each normalisation scheme, the following steps are taken to model each process excess/deficit:

(1) The Mo isotope abundances of the ‘sample’ are generated by adding the deficit/excess to the appropriate terrestrial isotope abundances. (2) Isotope ratios for the ‘sample’ are then calculated using these abundances. (3) The isotope ratios of the ‘sample’ from Step (2) are internally normalised using the procedure outlined in Chapter 1, Section 1.4.3. (4) Using Equation 2.1, nucleosynthetic isotope anomalies for each ratio from Step (3) are calculated relative to the terrestrial ratios. Plots of εiMo vs. iMo are then generated as models.

i x Mo Mo εiMo = 'sample' − 1 × 104 Equation 2.1 iMo xMo terrestrial where xMo is the same isotope used in the denominator of the normalising ratio

2.2.1 Normalisation to 98Mo/96Mo

Figure 2.3 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 98Mo/96Mo = 1.453174. Nucleosynthetic processes contributions from Arlandini et al. (1999) are used to calculate isotope anomalies (εiMo) relative to terrestrial Mo. Anomalies are scaled so the absolute value of largest anomaly in each panel = 1ε.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 43

Normalisation to 98Mo/96Mo = 1.453174 produces the distinctive Mo isotope patterns displayed in Figure 2.3. A p-process excess (a) results in positive ε92Mo and ε94Mo values. An s-process excess (b) produces a characteristic m-shaped pattern, while an r-process excess (c) produces a w-shape but with an extra kink at 94Mo, in agreement with Burkhardt et al. (2011). The excesses and deficits are mirror images of each other. It is noteworthy that 96Mo and 98Mo are both partially or wholly s-process isotopes, and so any s-process excess/deficit should not dramatically alter the 98Mo/96Mo ratio (Becker and Walker, 2003).

2.2.2 Normalisation to 92Mo/98Mo

Figure 2.4 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 92Mo/98Mo = 0.607898. Nucleosynthetic processes contributions from Arlandini et al. (1999) are used to calculate isotope anomalies (εiMo) relative to terrestrial Mo. Each panel is scaled relative to the respective panel in Figure 2.3.

Normalising to 92Mo/98Mo = 0.607898 produces less distinctive patterns for s- and r-process excesses and deficits (Figure 2.4), with smaller and less easy to resolve anomalies than those produced by 98Mo/96Mo normalisation (except ε100Mo). Since 92Mo is produced entirely by the p-process, and 98Mo is an s- and r-process nuclide, any excess or deficit in the p-, s- or r-process contributions will significantly affect the 92Mo/98Mo ratio, leading to apparent anomalies that may actually be analytical artefacts. Consequently, the isotope pattern as a whole must be considered rather than individual isotopes.

2.2.3 Normalisation to 97Mo/95Mo

Normalising to 97Mo/95Mo = 0.602083 produces more distinctive patterns (Figure 2.5) than the 92Mo/98Mo normalisation but not as obvious as the 98Mo/96Mo normalisation scheme for

44 Chapter 2 s-process effects. The 97Mo/95Mo normalisation does, however, have the advantage that both 95Mo and 97Mo do not have any isobaric interferences, whereas all isotopes used in the previous normalisation schemes do (92Mo = 92Zr; 96Mo = 96Ru and 96Zr; 98Mo = 98Ru). Critically, 95Mo and 97Mo have very similar s- and r-process contributions (97Mo: 59% s- and 41% r-; 95Mo: 55% and 45%). Therefore, variations in these components will not significantly alter the 97Mo/95Mo ratio.

Figure 2.5 Effects on Mo isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Mo, using normalisation to 97Mo/95Mo = 0.602083. Nucleosynthetic processes contributions from Arlandini et al. (1999) are used to calculate isotope anomalies (εiMo) relative to terrestrial Mo. Each panel is scaled relative to the respective panel in Figure 2.3.

2.3 Analytical techniques

Aliquots of the iron meteorite sample solutions that were prepared as described in Chapter 1 were employed for the analyses that are detailed below. All materials and reagents used, and the laboratory conditions, were the same as those outlined in Chapter 1.

2.3.1 Ion-exchange chromatography

Separation of Mo was achieved by a two-stage procedure (Table 2.1), which applies anion exchange chromatography with Bio-Rad AG1-X8 resin (200–400 mesh, chloride form). All chemistry was undertaken in a Class 1000 clean laboratory. The first stage was adapted from Pearce et al. (2009). The second stage was adapted from Schönbächler et al. (2004), and repeated two or three times depending on the Ru content of the samples. For each sample, 100–300 mg of meteorite (~2000 ng Mo) was loaded onto the columns, and Mo yields were typically 75–90%. Full procedural blanks were commonly

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 45

< 2 ng – the blank contribution to the total Mo analysed per sample was < 1‰ and thus negligible. Following the chemistry, the purified Mo fractions typically had Ru/Mo and Zr/Mo ratios of lower than 3 × 10−5 and 4 × 10−5, respectively.

Table 2.1 Ion-exchange chemistry for separation of Mo from iron meteorites

Stage 1: Bio-Rad column, 2 ml resin reservoir, 8 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (2 ml) Step Resin volumes Acid

Cleaning 15 3 M HNO3 Pre-condition resin 2 1 M HF / 0.5 M HCl Load Sample 5–10 1 M HF / 0.5 M HCl Rinse matrix (inc. Fe, Ni) 4 1 M HF / 0.5 M HCl Rinse matrix (Ru) 5 1 M HF

Collect Mo 5 3 M HNO3

Stage 2: Teflon column, 0.15 ml resin reservoir, 3 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (0.15 ml) Step Resin volumes Acid

Cleaning 16 3 M HNO3 Pre-condition resin 9 4 M HF Load Sample 5 4 M HF Rinse matrix (Ru) 12 4 M HF Elute Zr & W 7 6 M HCl / 1 M HF

Collect Mo 5 3 M HNO3

2.3.2 Mass spectrometry

2.3.2.1 Instrumentation and data collection protocol

The isotope measurements were performed using the Nu Instruments Nu Plasma HR MC-ICP-MS at Imperial College London and samples were delivered via a Nu Instruments DSN-100 desolvating nebuliser at an uptake rate of ~140 µl/min. Typical sensitivity for Mo was 150–180 V/ppm for solutions with ~200 ppb Mo. The data were acquired in a two-sequence routine. The Faraday cup configuration and positions of the masses are indicated in Table 2.2. A simultaneous measurement of 92Mo, 94Mo, 95Mo, 96Mo, 97Mo, 98Mo, 100Mo and 99Ru ion beams (using 1011 Ω amplifiers) was performed in the first sequence. The second sequence, which immediately followed the first,

46 Chapter 2 collected the 90Zr/95Mo ratio used for Zr interference corrections. The analyses utilised 2 blocks with 20 integrations of 8 seconds each for the first sequence, and of 2 seconds each for the second sequence. Each block was preceded by a 45 second on-peak baseline measurement while the ion beam was deflected by the electrostatic analyser.

Table 2.2 Faraday cup configuration for Mo isotope measurements.

Faraday Cup L5 L4 L3 L2 L1 Ax H1 H2 H3 H4 H5 H6

Sequence 1 92 94 95 96 97 98 99 100

Sequence 2 90 92 94 95 96 97 98

2.3.2.2 Interference and mass bias corrections

Possible spectral interferences on the Mo isotopes are shown in Table 2.3. Zirconium interferences were corrected using 90Zr as the interference monitor and the Zr isotopic abundances of Schönbächler et al. (2004). Ruthenium interferences were corrected using 99Ru as the interference monitor and the Ru isotopic abundances of Becker et al. (2002).

Table 2.3 Important spectral interferences on Mo isotopes

Interference 92Mo 94Mo 95Mo 96Mo 97Mo 98Mo 100Mo

92 94 96 98 100 Isobaric Zr Zr Zr Ru Ru 96Ru

Doubly charged ions 184W2+ 188Os2+ 190Os2+ 192Os2+ 194Pt2+ 196Pt2+ 200Hg2+ 184Os2+ 190Pt2+ 192Pt2+ 196Hg2+

Argides 52Cr40Ar 54Cr40Ar 55Mn40Ar 56Fe40Ar 57Fe40Ar 58Fe40Ar 60Ni40Ar 54Fe40Ar

Oxides 76Se16O 78Se16O 79Br16O 80Se16O 81Br16O 82Se16O 84Se16O 76Ge16O 78Kr16O 80Kr16O 82Kr16O 84Kr16O

Instrumental mass bias was corrected by normalisation to 98Mo/96Mo = 1.453174, 92Mo/98Mo = 0.607898 and 97Mo/95Mo = 0.602083, using the exponential law (Lu and Masuda, 1994; Young et al., 2002; Wombacher and Rehkämper, 2003). Since some of these ratios have isobaric interferences (Zr on 92Mo and 96Mo; Ru on 96Mo and 98Mo), an iterative procedure was used to subtract these effects from the normalising ratios.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 47

Data are reported in εiMo notation (Equation 2.2), calculated relative to the mean of several bracketing runs of NIST SRM 3134 Mo made up to closely match the Mo concentration of the samples (~200 ppb).

i x Mo Mo εiMo = sample − 1 × 104 Equation 2.2 iMo xMo standard where xMo is the same isotope used in the denominator of the normalising ratio

2.4 Results

2.4.1 Standard solutions and reference materials

Typical external reproducibility (2σ = 2sd) and internal precisions (2se) of the bracketing NIST SRM 3134 standards measurements (typically n = 4), for all normalisations, are shown in Table 2.4 and Table 2.5 respectively. The fact that the external reproducibility is very similar to the internal precision demonstrates the excellent precision of the procedures.

Table 2.4 External reproducibility of Mo isotope measurements. Given by the 2σ uncertainty of n = 4 bracketing NIST SRM 3134 solutions

Normalising Ratio ε92Mo ε94Mo ε95Mo ε96Mo ε97Mo ε98Mo ε100Mo

98 96 Mo/ Mo ±0.39 ±0.26 ±0.20 - ±0.14 - ±0.23 92Mo/98Mo - ±0.16 ±0.16 ±0.13 ±0.14 - ±0.23 97Mo/95Mo ±0.39 ±0.22 - ±0.15 - ±0.20 ±0.38

Table 2.5 Internal precision of Mo isotope measurements. Given by the typical 2se uncertainty for single run measurements of NIST SRM 3134 solutions

Normalising Ratio ε92Mo ε94Mo ε95Mo ε96Mo ε97Mo ε98Mo ε100Mo

98 96 Mo/ Mo ±0.28 ±0.22 ±0.16 - ±0.11 - ±0.16 92Mo/98Mo - ±0.12 ±0.09 ±0.09 ±0.10 - ±0.12 97Mo/95Mo ±0.24 ±0.16 - ±0.12 - ±0.14 ±0.25

Normalising to 92Mo/98Mo gives a better precision for most of the Mo ratios than normalising to 98Mo/96Mo, because it spans most of the large mass range of Mo, whereas normalising to 98Mo/96Mo appears to give less adequately corrected fractionation at the mass extremes, since it spans less of the Mo mass range. It has also been suggested that this reflects the higher abundance of 96Ru than 98Ru (Burkhardt et al., 2009). However, as explained in Section 2.2.2,

48 Chapter 2

Mo isotope patterns for normalisation to 92Mo/98Mo are far less characteristic and any anomalies less easily resolved. Moreover, the magnitude of anomalies expected from previous studies (Dauphas et al., 2002; Burkhardt et al., 2011) are much smaller for 92Mo/98Mo and are closer to the reproducibility of the standard measurements. Consequently, 98Mo/96Mo and 97Mo/95Mo are more suitable for resolving anomalies. Normalising to 97Mo/95Mo generally gives a similar precision as 98Mo/96Mo, except for ε100Mo where it is worse. This can be explained by 100 amu being further away from the normalising masses. One might expect 97Mo/95Mo to give better precision for ε96Mo since it straddles it, but in fact 92Mo/98Mo gives a better precision, probably due to the lower abundance of 97Mo than 98Mo (9.6 and 24.4% respectively) resulting in lower beam intensities for 97Mo in the mass spectrometer and thus less-precise measurements. The precision achieved here is generally a factor of ~1.5 better than the most precise previous study of Burkhardt et al. (2011) – Figure 2.6 compares the external 2σ reproducibilities (both 2σ). Primarily, the improved precision arises from measuring with ~200 ppb Mo solutions rather than the ~100 ppb used by Burkhardt et al. (2011), thus generating greater ion beam intensities.

Figure 2.6 External reproducibility of the Mo isotope measurements compared to literature. Here the 2σ uncertainty of the NIST SRM 3134 standard solutions were generally a factor of ~1.5 more precise for all normalisations than that of Burkhardt et al. (2011).

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 49

The robustness and reproducibility of the Mo separation procedure and MC-ICP-MS Mo isotope measurements were routinely evaluated by repeated analyses of terrestrial standard reference materials. NIST SRM 129c (High-Sulphur Steel) was chosen as an analogy of iron meteorite matrix due to its high (2%) sulphur content. NIST SRM 361 (AISI 4340 Steel) was similarly chosen to ensure slight matrix variations in Fe-Ni alloys do not affect the results. Aliquots of NIST SRM 3134 were also processed through the ion-exchange chromatography and analysed to ensure no analytical artefacts perturb the isotope measurements. Finally, a Mo standard solution (ICL-Mo, Specpure 35758), which is used in-house as a stable isotope reference material (Goldberg et al., 2013), was run against the NIST SRM 3134 standard solutions to ensure constant plasma behaviour. All of these standard reference material measurements routinely yielded terrestrial Mo isotope compositions (Table 2.6 and Figure 2.7).

Table 2.6 Molybdenum isotope data for terrestrial standard reference materials Normalised to 98Mo/96Mo Reference Material N a ε92Mo ε94Mo ε95Mo ε97Mo ε100Mo

NIST SRM 129c 9 −0.13 ± 0.09 −0.05 ± 0.07 −0.05 ± 0.06 −0.03 ± 0.04 −0.08 ± 0.09 NIST SRM 361 8 0.11 ± 0.16 0.08 ± 0.08 0.07 ± 0.05 −0.01 ± 0.06 0.01 ± 0.12 NIST SRM 3134 b 11 −0.02 ± 0.10 0.05 ± 0.08 0.00 ± 0.03 −0.03 ± 0.05 −0.01 ± 0.09 ICL-Mo 12 0.07 ± 0.13 0.06 ± 0.08 0.03 ± 0.06 0.02 ± 0.04 0.04 ± 0.07 MEAN −0.01 ± 0.14 0.02 ± 0.08 0.00 ± 0.07 −0.02 ± 0.02 −0.03 ± 0.05

Normalised to 92Mo/98Mo Reference Material N a ε94Mo ε95Mo ε96Mo ε97Mo ε100Mo

NIST SRM 129c 9 0.00 ± 0.07 −0.02 ± 0.07 0.04 ± 0.03 −0.02 ± 0.04 −0.09 ± 0.10 NIST SRM 361 8 0.02 ± 0.11 0.02 ± 0.08 −0.04 ± 0.05 −0.03 ± 0.07 0.03 ± 0.10 NIST SRM 3134 b 11 0.01 ± 0.07 0.00 ± 0.05 0.00 ± 0.04 −0.04 ± 0.06 −0.03 ± 0.06 ICL-Mo 12 −0.02 ± 0.05 −0.02 ± 0.06 −0.02 ± 0.04 0.00 ± 0.05 0.05 ± 0.05 MEAN 0.01 ± 0.01 0.00 ± 0.03 0.00 ± 0.04 −0.03 ± 0.01 −0.03 ± 0.07

Normalised to 97Mo/95Mo Reference Material N a ε92Mo ε94Mo ε96Mo ε98Mo ε100Mo NIST SRM 129c 9 −0.03 ± 0.18 0.02 ± 0.10 0.06 ± 0.04 0.05 ± 0.07 0.01 ± 0.17 NIST SRM 361 8 −0.08 ± 0.16 −0.02 ± 0.08 −0.03 ± 0.04 0.07 ± 0.10 0.16 ± 0.21 NIST SRM 3134 b 11 −0.08 ± 0.10 −0.01 ± 0.05 0.01 ± 0.05 0.05 ± 0.09 0.12 ± 0.09 ICL-Mo 12 0.04 ± 0.12 0.04 ± 0.08 −0.02 ± 0.02 0.00 ± 0.06 0.00 ± 0.13 MEAN −0.06 ± 0.03 0.00 ± 0.03 0.02 ± 0.05 0.06 ± 0.01 0.10 ± 0.09

a Number of times sample analysed; b Aliquots of NIST SRM 3134 run through Mo separation chemistry. All uncertainties are 2se = 2σ/√n.

50 Chapter 2

Figure 2.7 Molybdenum isotope data for terrestrial standard reference materials, relative to NIST SRM 3134. Uncertainties are 2se. Grey-shaded area represents reproducibility (2σ) of the bracketing NIST SRM 3134 runs. a Aliquots of NIST SRM 3134 run through Mo separation chemistry.

2.4.1.1 Interferences

Ruthenium and zirconium interference corrections for the typical Ru/Mo and Zr/Mo ratios of the purified iron meteorite Mo fractions are shown in Table 2.7 and Table 2.8. In all normalisations, the corrections for Ru are greatest on ε100Mo, and for Zr on ε94Mo.

Table 2.7 Ruthenium corrections (ppm) for Mo fractions with Ru/Mo = 3 × 10−5

Normalising Ratio ε92Mo ε94Mo ε95Mo ε96Mo ε97Mo ε98Mo ε100Mo 98Mo/96Mo 30 20 15 - 10 - 45 92Mo/98Mo - 1 1 10 2 - 35 97Mo/95Mo 0 0 - 10 - 2 40

Table 2.8 Zirconium corrections (ppm) for Mo fractions with Zr/Mo = 4 × 10−5 Normalising Ratio ε92Mo ε94Mo ε95Mo ε96Mo ε97Mo ε98Mo ε100Mo 98Mo/96Mo 30 70 10 - 5 - 5 92Mo/98Mo - 50 25 10 10 - 15 97Mo/95Mo 60 100 - 10 - 0 0

To test the accuracy of the corrections, NIST 3134 standard solutions were doped with Zr and Ru. Corrections were found to hold up to Ru/Mo = 3.5 × 10−4 (equivalent to corrections of

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 51

~500 ppm) and Zr/Mo = 1.7 × 10−3 (~2900 ppm) (Figure 2.8 and Figure 2.9). These ratios are well above the typical Ru/Mo and Zr/Mo ratios of the purified Mo fractions.

Figure 2.8 ε100Mo obtained for NIST SRM 3134 Mo solutions, doped with Ru SRM. Grey bars represent 2σ external reproducibility of the NIST SRM 3134 Mo solutions. Uncertainties on data points are 2se. Normalising to 98Mo/96Mo is the most sensitive to Ru as both normalising isotopes have Ru isobaric interferences (a). An accurate correction for ε100Mo is possible up to Ru/Mo of 3.5 × 10−4.

Figure 2.9 ε94Mo obtained for NIST SRM 3134 Mo solutions, doped with Zr SRM. Grey bars represent 2σ external reproducibility of the NIST SRM 3134 Mo solutions. Uncertainties on data points are 2se in-run precisions. The Zr interference on ε94Mo be adequately corrected up to Zr/Mo of 1.7 × 10−3. Above this ratio, the uncertainties become too significant.

52 Chapter 2

Spectral interferences from molecular species and doubly charged ions (Table 2.3) were also monitored but the separation chemistry for Mo was efficient enough that no corrections were needed. Additionally, while large amounts of 93Nb may generate peak-tailing effects on 92Mo and 94Mo (Schönbächler et al., 2004), the chemical separation produces Mo fractions that have insignificant Nb contents for this to occur.

2.4.2 Iron meteorites

All meteorites belonging to the same group display identical (within uncertainty) Mo isotopic anomalies. As such, the Mo isotope composition of any meteorite is representative of the complete metal core of its parent body.

2.4.2.1 Iron meteorites normalised to 98Mo/96Mo

Molybdenum isotope data internally normalised to 98Mo/96Mo = 1.453174 are shown in Figure 2.10 and Table 2.9. Meteorites of all groups (except IAB/IIICD) exhibit offsets from εiMo = 0: the largest anomalies are in ε92Mo, with the magnitude of anomalies decreasing in the order ε92Mo > ε94Mo > ε95Mo > ε100Mo > ε97Mo. This gives rise to the characteristic w-shaped pattern of s-process deficits, as predicted by the modelling Figure 2.3(e).

Figure 2.10 Molybdenum isotope data for iron meteorites, normalised to 98Mo/96Mo = 1.453174. Shown are the group means with 2se uncertainties. Measured relative to NIST SRM 3134, which represents the terrestrial Mo isotope composition. Grey-shaded area represents reproducibility (2σ) of the bracketing NIST SRM 3134 runs.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 53

Whilst the positive ε97Mo anomaly could potentially represent a radiogenic contribution from 97 the decay of extinct Tc (t1/2 ~ 3.8 Myr), Dauphas et al. (2002) and Burkhardt et al. (2011) showed that any such contribution will be insignificant and not reflected in the 97Mo isotope measurements. The extent of the anomalies displayed by the different magmatic iron groups are variable. IIC irons display by far the most extreme anomalies with ε92Mo = +3.12 ± 0.27, while the IVAs are the least anomalous of the magmatic iron groups (ε92Mo = +0.95 ± 0.19). The non-magmatic IAB and IIICD groups, often termed the IAB/IIICD complex (e.g., Wasson and Kallemeyn, 2002), present Mo isotope compositions indistinguishable from terrestrial Mo, while the non-magmatic IIE group is slightly anomalous with ε92Mo = +0.86 ± 0.32.

2.4.2.2 Iron meteorites normalised to 92Mo/98Mo

Results for this normalisation are shown in Figure 2.11 and Table 2.10. As predicted by modelling in Section 2.2, the Mo isotope patterns for normalisation to 92Mo/98Mo are far less characteristic than 98Mo/96Mo, and the anomalies at ε94Mo, ε95Mo and ε97Mo are closer to the reproducibility of the standard measurements and less easily resolved. Only ε100Mo, and to a lesser extent ε96Mo, are easily resolvable using this normalisation.

Figure 2.11 Molybdenum isotope data for iron meteorites, normalised to 92Mo/98Mo = 0.607898. Shown are the group means with 2se uncertainties. Measured relative to NIST SRM 3134, which represents the terrestrial Mo isotope composition. Grey-shaded area represents reproducibility (2σ) of the bracketing NIST SRM 3134 runs.

54 Chapter 2

The isotope pattern is again suggestive of variable s-process deficits (see Figure 2.4e), with the IIC irons showing the greatest anomalies (ε100Mo = +1.96 ± 0.10). Similarly, the IVAs have the smallest anomalies of the magmatic irons (ε100Mo = +0.57 ± 0.12), while the non- magmatic IAB/IIICD complex have terrestrial compositions but IIEs do not (ε100Mo = +0.51 ± 0.04).

2.4.2.3 Iron meteorites normalised to 97Mo/95Mo

Results for normalisation to 97Mo/95Mo = 0.602083 are shown in Figure 2.12 and Table 2.11. Just as in the previous normalisations, the IICs display the largest anomalies (ε100Mo = +1.33 ± 0.14) and the IVAs the smallest (ε100Mo = +0.34 ± 0.04). The isotope patterns most closely resemble the s-process deficit model Figure 2.5(e).

Figure 2.12 Molybdenum isotope data for iron meteorites, normalised to 97Mo/95Mo = 0.602083. Shown are the group means with 2se uncertainties. Measured relative to NIST SRM 3134, which represents the terrestrial Mo isotope composition. Grey-shaded area represents reproducibility (2σ) of the bracketing NIST SRM 3134 runs.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 55

Table 2.9 Molybdenum isotope compositions of iron meteorites, normalised to 98Mo/96Mo

Group Sample N a ε92Mo b ε94Mo b ε95Mo b ε97Mo b ε100Mo b IAB Bitburg 10 − 0.10 ± 0.07 − 0.01 ± 0.05 − 0.01 ± 0.05 − 0.05 ± 0.03 0.00 ± 0.06 Campo del Cielo 1 11 0.32 ± 0.11 0.22 ± 0.04 −0.05 ± 0.05 −0.05 ± 0.05 0.06 ± 0.08 Campo del Cielo 2 9 0.13 ± 0.13 0.15 ± 0.04 0.01 ± 0.06 −0.04 ± 0.02 −0.04 ± 0.07 Campo del Cielo 3 9 0.10 ± 0.09 0.05 ± 0.06 −0.11 ± 0.09 0.03 ± 0.06 0.05 ± 0.11 Canyon Diablo 1 6 −0.06 ± 0.23 −0.01 ± 0.09 −0.14 ± 0.03 −0.04 ± 0.08 0.24 ± 0.18 Canyon Diablo 2 6 −0.11 ± 0.24 −0.05 ± 0.07 −0.19 ± 0.06 −0.06 ± 0.06 0.10 ± 0.08 Cosby's Creek 13 −0.01 ± 0.07 −0.01 ± 0.06 0.02 ± 0.04 −0.06 ± 0.03 −0.08 ± 0.06 Odessa 10 0.01 ± 0.11 0.05 ± 0.08 −0.16 ± 0.01 0.01 ± 0.06 0.18 ± 0.05 Toluca 12 0.00 ± 0.08 0.14 ± 0.06 −0.06 ± 0.04 −0.05 ± 0.03 0.02 ± 0.06 IAB MEAN 0.00 ± 0.08 0.05 ± 0.06 −0.07 ± 0.06 −0.04 ± 0.02 0.05 ± 0.08

IC Arispe 1 8 1.04 ± 0.10 0.87 ± 0.06 0.25 ± 0.07 0.22 ± 0.08 0.38 ± 0.07 Arispe 2 7 0.98 ± 0.14 0.85 ± 0.11 0.28 ± 0.06 0.16 ± 0.06 0.24 ± 0.10 Arispe 3 4 1.20 ± 0.13 0.95 ± 0.18 0.32 ± 0.07 0.21 ± 0.06 0.16 ± 0.07 Arispe 4 8 1.03 ± 0.13 0.87 ± 0.08 0.31 ± 0.13 0.20 ± 0.05 0.25 ± 0.15 Bendego 1 9 1.16 ± 0.18 0.94 ± 0.07 0.28 ± 0.06 0.20 ± 0.08 0.35 ± 0.11 Bendego 2 7 0.93 ± 0.10 0.80 ± 0.08 0.23 ± 0.05 0.12 ± 0.07 0.22 ± 0.08 Santa Rosa 8 1.14 ± 0.15 0.99 ± 0.08 0.34 ± 0.08 0.18 ± 0.05 0.34 ± 0.03 IC MEAN 1.08 ± 0.06 0.92 ± 0.08 0.30 ± 0.05 0.18 ± 0.02 0.30 ± 0.05

IIAB Coahuila 12 1.48 ± 0.09 1.18 ± 0.06 0.49 ± 0.08 0.27 ± 0.04 0.44 ± 0.06 Murphy 11 1.26 ± 0.14 1.08 ± 0.10 0.47 ± 0.07 0.25 ± 0.05 0.41 ± 0.08 Negrillos 10 1.53 ± 0.12 1.23 ± 0.12 0.47 ± 0.05 0.27 ± 0.06 0.25 ± 0.09 North Chile 16 1.52 ± 0.14 1.25 ± 0.09 0.57 ± 0.06 0.28 ± 0.03 0.47 ± 0.08 Sikhote Alin 1 13 1.58 ± 0.07 1.23 ± 0.07 0.50 ± 0.05 0.22 ± 0.07 0.42 ± 0.09 Sikhote Alin 2 8 1.38 ± 0.14 1.18 ± 0.04 0.58 ± 0.09 0.25 ± 0.07 0.26 ± 0.06 IIAB MEAN 1.45 ± 0.10 1.19 ± 0.06 0.51 ± 0.04 0.26 ± 0.02 0.38 ± 0.08

IIC Ballinoo 7 3.07 ± 0.18 2.35 ± 0.15 1.62 ± 0.11 0.81 ± 0.06 1.05 ± 0.12 Kumerina 10 2.92 ± 0.09 2.28 ± 0.08 1.53 ± 0.07 0.79 ± 0.06 1.12 ± 0.08 Salt River 6 3.38 ± 0.14 2.56 ± 0.06 1.68 ± 0.10 0.87 ± 0.09 0.84 ± 0.09 IIC MEAN 3.12 ± 0.27 2.40 ± 0.16 1.61 ± 0.09 0.82 ± 0.05 1.00 ± 0.17

IIE Kodaikanal 6 1.04 ± 0.16 0.79 ± 0.06 0.34 ± 0.06 0.10 ± 0.03 0.24 ± 0.12 Verkhne Dnieprovsk 8 0.99 ± 0.16 0.78 ± 0.10 0.37 ± 0.04 0.19 ± 0.06 0.17 ± 0.08 Weekeroo Station 5 0.54 ± 0.11 0.57 ± 0.15 0.16 ± 0.05 0.08 ± 0.03 0.29 ± 0.11 IIE MEAN 0.86 ± 0.32 0.71 ± 0.15 0.29 ± 0.13 0.12 ± 0.07 0.24 ± 0.07

IIIAB Bear Creek 10 1.16 ± 0.18 1.00 ± 0.14 0.44 ± 0.08 0.19 ± 0.05 0.41 ± 0.06 Cape York 8 1.31 ± 0.09 1.09 ± 0.06 0.44 ± 0.07 0.16 ± 0.08 0.36 ± 0.04 Charcas 1 6 1.10 ± 0.07 0.89 ± 0.08 0.32 ± 0.10 0.12 ± 0.06 0.35 ± 0.16 Charcas 2 9 1.02 ± 0.13 0.92 ± 0.09 0.30 ± 0.10 0.22 ± 0.03 0.41 ± 0.07 Henbury 1 6 1.12 ± 0.14 0.97 ± 0.06 0.33 ± 0.07 0.16 ± 0.08 0.29 ± 0.06 Henbury 2 10 1.20 ± 0.19 1.06 ± 0.09 0.44 ± 0.07 0.22 ± 0.07 0.24 ± 0.05 Lenarto 10 1.39 ± 0.10 1.13 ± 0.08 0.51 ± 0.05 0.17 ± 0.05 0.30 ± 0.08 Santa Apolonia 1 6 1.25 ± 0.27 0.99 ± 0.18 0.42 ± 0.10 0.21 ± 0.06 0.33 ± 0.13 Santa Apolonia 2 8 1.18 ± 0.14 0.95 ± 0.07 0.35 ± 0.05 0.16 ± 0.05 0.33 ± 0.13 Verkhne Udinsk 7 1.44 ± 0.14 1.11 ± 0.08 0.42 ± 0.08 0.18 ± 0.05 0.49 ± 0.09 Williamette 8 1.40 ± 0.19 1.06 ± 0.20 0.51 ± 0.11 0.21 ± 0.06 0.34 ± 0.03 IIIAB MEAN 1.27 ± 0.10 1.04 ± 0.05 0.43 ± 0.05 0.18 ± 0.01 0.36 ± 0.05

IIICD Carlton 1 5 0.37 ± 0.18 0.27 ± 0.11 0.02 ± 0.08 −0.02 ± 0.03 0.00 ± 0.09 Carlton 2 2 0.32 ± 0.20 0.08 ± 0.04 −0.10 ± 0.08 −0.05 ± 0.20 0.08 ± 0.17 Nantan 1 10 0.14 ± 0.17 0.21 ± 0.13 0.09 ± 0.10 0.02 ± 0.07 0.08 ± 0.07 Nantan 2 11 −0.03 ± 0.11 0.08 ± 0.11 −0.06 ± 0.08 −0.04 ± 0.05 0.05 ± 0.07 IIICD MEAN 0.20 ± 0.29 0.16 ± 0.03 −0.01 ± 0.06 −0.02 ± 0.02 0.05 ± 0.03

IIIE Staunton 8 1.40 ± 0.19 1.07 ± 0.13 0.45 ± 0.09 0.25 ± 0.07 0.40 ± 0.15

IIIF Clark County 7 1.70 ± 0.09 1.24 ± 0.05 0.92 ± 0.05 0.39 ± 0.06 0.59 ± 0.04

IVA Gibeon 1 5 1.11 ± 0.25 0.85 ± 0.25 0.42 ± 0.24 0.18 ± 0.10 0.27 ± 0.18 Gibeon 2 4 1.02 ± 0.15 0.80 ± 0.19 0.37 ± 0.10 0.20 ± 0.04 0.17 ± 0.19 Muonionalusta 7 0.76 ± 0.28 0.64 ± 0.14 0.30 ± 0.11 0.20 ± 0.08 0.46 ± 0.04 Obernkirchen 6 1.01 ± 0.10 0.82 ± 0.05 0.33 ± 0.09 0.17 ± 0.06 0.22 ± 0.09 IVA MEAN 0.95 ± 0.19 0.76 ± 0.12 0.34 ± 0.05 0.18 ± 0.02 0.30 ± 0.16

IVB Cape of Good Hope 10 2.04 ± 0.14 1.38 ± 0.10 0.98 ± 0.06 0.49 ± 0.06 0.72 ± 0.05 Santa Clara 12 1.93 ± 0.14 1.37 ± 0.10 0.97 ± 0.06 0.51 ± 0.07 0.84 ± 0.09 Tlacotopec 1 9 1.86 ± 0.17 1.34 ± 0.11 0.84 ± 0.07 0.47 ± 0.05 0.80 ± 0.08 Tlacotopec 2 12 1.47 ± 0.11 1.14 ± 0.09 0.65 ± 0.06 0.46 ± 0.04 0.92 ± 0.10 IVB MEAN 1.88 ± 0.22 1.33 ± 0.09 0.90 ± 0.15 0.49 ± 0.03 0.81 ± 0.09 a Number of times sample was analysed; b Normalised to 98Mo/96Mo = 1.453174 using the exponential law, i i 96 i 96 4 ε Mo = [( Mo/ Mo)smp/( Mo/ Mo)std −1] × 10 ; Uncertainties for samples are 2se = 2σ/√n, where n is the number of times sample analysed. Uncertainties for group means are 2se = 2σ/√n, where n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of the specimens is taken to represent that sample when calculating the group mean).

56 Chapter 2

Table 2.10 Molybdenum isotope compositions of iron meteorites, normalised to 92Mo/98Mo

Group Sample N a ε94Mo b ε95Mo b ε96Mo b ε97Mo b ε100Mo b IAB Bitburg 10 0.03 ± 0.04 0.04 ± 0.04 0.03 ± 0.02 − 0.01 ± 0.03 − 0.05 ± 0.05 Campo del Cielo 1 11 −0.04 ± 0.04 −0.21 ± 0.02 −0.10 ± 0.04 −0.11 ± 0.06 0.19 ± 0.10 Campo del Cielo 2 9 0.09 ± 0.05 −0.03 ± 0.05 −0.04 ± 0.04 −0.04 ± 0.03 0.03 ± 0.02 Campo del Cielo 3 9 −0.01 ± 0.07 −0.12 ± 0.06 −0.03 ± 0.03 0.00 ± 0.07 0.08 ± 0.11 Canyon Diablo 1 6 0.00 ± 0.12 −0.15 ± 0.10 0.02 ± 0.08 −0.05 ± 0.09 0.18 ± 0.19 Canyon Diablo 2 6 0.04 ± 0.13 −0.17 ± 0.14 0.04 ± 0.08 −0.02 ± 0.10 0.08 ± 0.09 Cosby's Creek 13 0.02 ± 0.05 0.07 ± 0.05 0.01 ± 0.03 −0.04 ± 0.03 −0.11 ± 0.06 Odessa 10 0.06 ± 0.07 −0.14 ± 0.07 −0.03 ± 0.02 −0.01 ± 0.06 0.14 ± 0.04 Toluca 12 0.09 ± 0.05 −0.07 ± 0.04 0.00 ± 0.03 −0.05 ± 0.04 0.03 ± 0.05 IAB MEAN 0.04 ± 0.03 −0.06 ± 0.08 0.00 ± 0.03 −0.03 ± 0.02 0.04 ± 0.08

IC Arispe 1 8 0.19 ± 0.03 −0.29 ± 0.06 −0.32 ± 0.04 0.04 ± 0.05 0.63 ± 0.11 Arispe 2 7 0.19 ± 0.04 −0.22 ± 0.06 −0.31 ± 0.04 0.00 ± 0.05 0.51 ± 0.07 Arispe 3 4 0.11 ± 0.14 −0.26 ± 0.02 −0.39 ± 0.04 0.01 ± 0.06 0.52 ± 0.06 Arispe 4 8 0.20 ± 0.08 −0.20 ± 0.08 −0.34 ± 0.04 0.04 ± 0.06 0.56 ± 0.12 Bendego 1 9 0.20 ± 0.07 −0.28 ± 0.05 −0.37 ± 0.06 0.00 ± 0.07 0.69 ± 0.07 Bendego 2 7 0.19 ± 0.04 −0.24 ± 0.05 −0.30 ± 0.03 −0.05 ± 0.05 0.52 ± 0.09 Santa Rosa 8 0.20 ± 0.06 −0.27 ± 0.04 −0.36 ± 0.05 −0.04 ± 0.04 0.64 ± 0.08 IC MEAN 0.19 ± 0.02 −0.26 ± 0.01 −0.35 ± 0.02 −0.01 ± 0.04 0.60 ± 0.05

IIAB Coahuila 12 0.20 ± 0.04 −0.20 ± 0.03 −0.46 ± 0.04 0.01 ± 0.04 0.83 ± 0.04 Murphy 11 0.26 ± 0.06 −0.13 ± 0.06 −0.40 ± 0.04 0.04 ± 0.05 0.77 ± 0.06 Negrillos 10 0.23 ± 0.07 −0.20 ± 0.05 −0.49 ± 0.04 0.02 ± 0.05 0.75 ± 0.07 North Chile 16 0.23 ± 0.06 −0.21 ± 0.05 −0.49 ± 0.04 0.01 ± 0.04 0.93 ± 0.09 Sikhote Alin 1 13 0.24 ± 0.05 −0.25 ± 0.04 −0.52 ± 0.02 −0.01 ± 0.04 0.84 ± 0.08 Sikhote Alin 2 8 0.24 ± 0.06 −0.12 ± 0.02 −0.44 ± 0.05 −0.01 ± 0.09 0.69 ± 0.01 IIAB MEAN 0.23 ± 0.02 −0.18 ± 0.03 −0.46 ± 0.03 0.01 ± 0.02 0.81 ± 0.07

IIC Ballinoo 7 0.27 ± 0.07 0.11 ± 0.04 −0.99 ± 0.06 0.31 ± 0.06 1.99 ± 0.08 Kumerina 10 0.27 ± 0.05 0.04 ± 0.05 −0.94 ± 0.03 0.25 ± 0.05 2.04 ± 0.07 Salt River 6 0.32 ± 0.07 0.06 ± 0.04 −1.08 ± 0.06 0.35 ± 0.07 1.87 ± 0.07 IIC MEAN 0.28 ± 0.03 0.07 ± 0.05 −1.00 ± 0.08 0.31 ± 0.06 1.96 ± 0.10

IIE Kodaikanal 6 0.13 ± 0.06 −0.17 ± 0.06 −0.34 ± 0.05 −0.07 ± 0.03 0.55 ± 0.08 Verkhne Dnieprovsk 8 0.12 ± 0.04 −0.10 ± 0.05 −0.32 ± 0.05 0.00 ± 0.05 0.50 ± 0.07 Weekeroo Station 5 0.20 ± 0.12 −0.11 ± 0.07 −0.17 ± 0.03 −0.01 ± 0.05 0.47 ± 0.13 IIE MEAN 0.15 ± 0.05 −0.12 ± 0.04 −0.27 ± 0.10 −0.03 ± 0.05 0.51 ± 0.04

IIIAB Bear Creek 10 0.25 ± 0.04 −0.18 ± 0.06 −0.38 ± 0.06 0.01 ± 0.03 0.77 ± 0.07 Cape York 8 0.21 ± 0.05 −0.17 ± 0.03 −0.42 ± 0.03 −0.06 ± 0.07 0.74 ± 0.04 Charcas 1 6 0.19 ± 0.12 −0.25 ± 0.13 −0.36 ± 0.02 −0.05 ± 0.06 0.71 ± 0.13 Charcas 2 9 0.25 ± 0.04 −0.24 ± 0.05 −0.35 ± 0.06 0.03 ± 0.04 0.74 ± 0.03 Henbury 1 6 0.23 ± 0.07 −0.23 ± 0.06 −0.36 ± 0.05 −0.03 ± 0.09 0.68 ± 0.06 Henbury 2 10 0.26 ± 0.05 −0.18 ± 0.05 −0.39 ± 0.06 0.02 ± 0.05 0.69 ± 0.04 Lenarto 10 0.17 ± 0.04 −0.21 ± 0.03 −0.45 ± 0.03 −0.04 ± 0.04 0.77 ± 0.06 Santa Apolonia 1 6 0.20 ± 0.08 −0.19 ± 0.04 −0.40 ± 0.09 −0.02 ± 0.06 0.67 ± 0.04 Santa Apolonia 2 8 0.19 ± 0.05 −0.23 ± 0.07 −0.38 ± 0.05 −0.03 ± 0.06 0.69 ± 0.10 Verkhne Udinsk 7 0.14 ± 0.06 −0.24 ± 0.03 −0.46 ± 0.04 −0.05 ± 0.06 0.93 ± 0.09 Williamette 8 0.17 ± 0.04 −0.17 ± 0.06 −0.45 ± 0.06 −0.03 ± 0.05 0.79 ± 0.04 IIIAB MEAN 0.20 ± 0.03 −0.21 ± 0.02 −0.41 ± 0.03 −0.03 ± 0.02 0.76 ± 0.06

IIICD Carlton 1 5 0.07 ± 0.09 −0.15 ± 0.08 −0.12 ± 0.06 −0.08 ± 0.02 0.05 ± 0.11 Carlton 2 2 −0.05 ± 0.15 −0.21 ± 0.02 −0.10 ± 0.06 −0.12 ± 0.33 0.17 ± 0.11 Nantan 1 10 0.01 ± 0.07 −0.09 ± 0.05 −0.04 ± 0.04 −0.05 ± 0.06 0.12 ± 0.09 Nantan 2 11 0.06 ± 0.06 −0.08 ± 0.05 0.02 ± 0.03 −0.07 ± 0.04 0.01 ± 0.06 IIICD MEAN 0.02 ± 0.03 −0.14 ± 0.10 −0.06 ± 0.10 −0.08 ± 0.05 0.09 ± 0.04

IIIE Staunton 8 0.14 ± 0.09 −0.26 ± 0.09 −0.45 ± 0.06 0.00 ± 0.06 0.83 ± 0.18

IIIF Clark County 7 0.10 ± 0.04 0.11 ± 0.06 −0.55 ± 0.03 0.12 ± 0.04 1.12 ± 0.06

IVA Gibeon 1 5 0.08 ± 0.05 −0.14 ± 0.09 −0.35 ± 0.08 −0.05 ± 0.05 0.55 ± 0.17 Gibeon 2 4 0.11 ± 0.12 −0.14 ± 0.06 −0.31 ± 0.05 0.00 ± 0.05 0.41 ± 0.11 Muonionalusta 7 0.11 ± 0.08 −0.06 ± 0.08 −0.24 ± 0.09 0.05 ± 0.06 0.69 ± 0.09 Obernkirchen 6 0.11 ± 0.06 −0.18 ± 0.05 −0.33 ± 0.03 −0.01 ± 0.06 0.54 ± 0.08 IVA MEAN 0.10 ± 0.01 −0.13 ± 0.07 −0.30 ± 0.06 0.01 ± 0.05 0.57 ± 0.12

IVB Cape of Good Hope 10 0.05 ± 0.05 0.01 ± 0.06 −0.67 ± 0.05 0.17 ± 0.06 1.34 ± 0.05 Santa Clara 12 0.09 ± 0.02 0.01 ± 0.04 −0.62 ± 0.04 0.19 ± 0.08 1.42 ± 0.08 Tlacotopec 1 9 0.10 ± 0.04 −0.10 ± 0.06 −0.60 ± 0.05 0.15 ± 0.04 1.37 ± 0.03 Tlacotopec 2 12 0.16 ± 0.04 −0.04 ± 0.03 −0.48 ± 0.04 0.19 ± 0.05 1.39 ± 0.07 IVB MEAN 0.09 ± 0.04 −0.02 ± 0.05 −0.61 ± 0.07 0.18 ± 0.01 1.38 ± 0.05

a Number of times sample was analysed; b Normalised to 92Mo/98Mo = 0.607898 using the exponential law, i i 98 i 98 4 ε Mo = [( Mo/ Mo)smp/( Mo/ Mo)std −1] × 10 ; Uncertainties for samples are 2se = 2σ/√n, where n is the number of times sample analysed. Uncertainties for group means are 2se = 2σ/√n, where n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of specimens is taken to represent that sample when calculating the group mean).

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 57

Table 2.11 Molybdenum isotope compositions of iron meteorites, normalised to 97Mo/95Mo

Group Sample N a ε92Mo b ε94Mo b ε96Mo b ε98Mo b ε100Mo b IAB Bitburg 10 − 0.08 ± 0.12 − 0.04 ± 0.05 0.02 ± 0.03 0.02 ± 0.06 0.04 ± 0.11 Campo del Cielo 1 11 0.38 ± 0.05 0.23 ± 0.07 0.04 ± 0.05 0.06 ± 0.08 0.14 ± 0.18 Campo del Cielo 2 9 0.04 ± 0.09 0.14 ± 0.08 0.01 ± 0.04 0.05 ± 0.04 0.09 ± 0.13 Campo del Cielo 3 9 0.27 ± 0.14 0.19 ± 0.09 0.03 ± 0.06 −0.03 ± 0.12 −0.06 ± 0.18 Canyon Diablo 1 6 0.30 ± 0.15 0.20 ± 0.10 0.11 ± 0.05 −0.03 ± 0.09 0.10 ± 0.28 Canyon Diablo 2 6 0.32 ± 0.25 0.24 ± 0.10 0.14 ± 0.04 −0.04 ± 0.08 −0.07 ± 0.15 Cosby's Creek 13 −0.24 ± 0.10 −0.08 ± 0.04 0.01 ± 0.03 0.10 ± 0.05 0.08 ± 0.13 Odessa 10 0.43 ± 0.10 0.29 ± 0.06 0.10 ± 0.05 −0.09 ± 0.06 −0.12 ± 0.07 Toluca 12 0.12 ± 0.06 0.22 ± 0.05 0.06 ± 0.03 0.05 ± 0.05 0.03 ± 0.10 IAB MEAN 0.13 ± 0.21 0.13 ± 0.12 0.06 ± 0.04 0.01 ± 0.05 0.02 ± 0.06

IC Arispe 1 8 0.81 ± 0.10 0.61 ± 0.02 −0.20 ± 0.05 −0.15 ± 0.06 0.18 ± 0.16 Arispe 2 7 0.54 ± 0.11 0.56 ± 0.09 −0.20 ± 0.03 −0.10 ± 0.06 0.21 ± 0.15 Arispe 3 4 0.71 ± 0.14 0.49 ± 0.22 −0.24 ± 0.04 −0.15 ± 0.09 0.18 ± 0.18 Arispe 4 8 0.51 ± 0.20 0.51 ± 0.12 −0.22 ± 0.07 −0.16 ± 0.08 0.23 ± 0.07 Bendego 1 9 0.62 ± 0.08 0.63 ± 0.03 −0.22 ± 0.05 −0.14 ± 0.10 0.29 ± 0.18 Bendego 2 7 0.54 ± 0.11 0.51 ± 0.07 −0.17 ± 0.06 −0.06 ± 0.07 0.33 ± 0.12 Santa Rosa 8 0.51 ± 0.12 0.53 ± 0.09 −0.24 ± 0.07 −0.06 ± 0.05 0.39 ± 0.07 IC MEAN 0.58 ± 0.08 0.55 ± 0.02 −0.21 ± 0.03 −0.10 ± 0.05 0.30 ± 0.11

IIAB Coahuila 12 0.61 ± 0.08 0.56 ± 0.05 −0.37 ± 0.05 −0.14 ± 0.05 0.44 ± 0.10 Murphy 11 0.47 ± 0.18 0.54 ± 0.07 −0.36 ± 0.04 −0.15 ± 0.08 0.41 ± 0.14 Negrillos 10 0.44 ± 0.10 0.48 ± 0.05 −0.36 ± 0.04 −0.14 ± 0.05 0.50 ± 0.04 North Chile 16 0.52 ± 0.12 0.57 ± 0.06 −0.41 ± 0.04 −0.15 ± 0.05 0.57 ± 0.13 Sikhote Alin 1 13 0.52 ± 0.08 0.60 ± 0.05 −0.35 ± 0.06 −0.09 ± 0.07 0.56 ± 0.10 Sikhote Alin 2 8 0.34 ± 0.12 0.48 ± 0.03 −0.38 ± 0.05 −0.04 ± 0.12 0.52 ± 0.11 IIAB MEAN 0.50 ± 0.07 0.54 ± 0.03 −0.38 ± 0.02 −0.13 ± 0.03 0.49 ± 0.06

IIC Ballinoo 7 0.30 ± 0.16 0.30 ± 0.06 −1.16 ± 0.06 −0.40 ± 0.10 1.38 ± 0.16 Kumerina 10 0.36 ± 0.10 0.37 ± 0.07 −1.14 ± 0.05 −0.38 ± 0.07 1.41 ± 0.11 Salt River 6 0.28 ± 0.19 0.35 ± 0.13 −1.26 ± 0.07 −0.47 ± 0.12 1.19 ± 0.20 IIC MEAN 0.31 ± 0.05 0.34 ± 0.04 −1.19 ± 0.07 −0.42 ± 0.05 1.33 ± 0.14

IIE Kodaikanal 6 0.32 ± 0.14 0.37 ± 0.09 −0.22 ± 0.04 0.01 ± 0.08 0.40 ± 0.10 Verkhne Dnieprovsk 8 0.29 ± 0.18 0.30 ± 0.07 −0.29 ± 0.04 −0.08 ± 0.09 0.29 ± 0.16 Weekeroo Station 5 0.24 ± 0.13 0.32 ± 0.10 −0.12 ± 0.04 −0.01 ± 0.07 0.36 ± 0.13 IIE MEAN 0.29 ± 0.05 0.33 ± 0.04 −0.21 ± 0.10 −0.02 ± 0.05 0.35 ± 0.07

IIIAB Bear Creek 10 0.49 ± 0.11 0.51 ± 0.05 −0.29 ± 0.09 −0.10 ± 0.05 0.49 ± 0.05 Cape York 8 0.48 ± 0.03 0.50 ± 0.05 −0.29 ± 0.06 −0.02 ± 0.09 0.57 ± 0.15 Charcas 1 6 0.50 ± 0.24 0.54 ± 0.10 −0.19 ± 0.06 −0.02 ± 0.07 0.53 ± 0.16 Charcas 2 9 0.74 ± 0.05 0.64 ± 0.09 −0.26 ± 0.06 −0.16 ± 0.04 0.28 ± 0.14 Henbury 1 6 0.52 ± 0.18 0.54 ± 0.04 −0.25 ± 0.06 −0.10 ± 0.10 0.36 ± 0.18 Henbury 2 10 0.49 ± 0.16 0.55 ± 0.10 −0.31 ± 0.07 −0.11 ± 0.04 0.34 ± 0.09 Lenarto 10 0.44 ± 0.06 0.51 ± 0.06 −0.35 ± 0.04 −0.06 ± 0.08 0.55 ± 0.19 Santa Apolonia 1 6 0.44 ± 0.12 0.48 ± 0.11 −0.28 ± 0.08 −0.08 ± 0.07 0.44 ± 0.16 Santa Apolonia 2 8 0.46 ± 0.13 0.50 ± 0.06 −0.23 ± 0.04 −0.06 ± 0.07 0.48 ± 0.22 Verkhne Udinsk 7 0.59 ± 0.08 0.54 ± 0.07 −0.30 ± 0.06 −0.07 ± 0.07 0.61 ± 0.14 Williamette 8 0.39 ± 0.07 0.41 ± 0.10 −0.35 ± 0.08 −0.04 ± 0.05 0.60 ± 0.12 IIIAB MEAN 0.50 ± 0.05 0.51 ± 0.04 −0.29 ± 0.03 −0.07 ± 0.02 0.51 ± 0.07

IIICD Carlton 1 5 0.20 ± 0.18 0.27 ± 0.14 0.03 ± 0.04 0.08 ± 0.07 0.03 ± 0.22 Carlton 2 2 0.39 ± 0.36 0.19 ± 0.04 0.05 ± 0.05 0.03 ± 0.49 0.05 ± 0.35 Nantan 1 10 0.16 ± 0.14 0.14 ± 0.03 0.02 ± 0.07 0.04 ± 0.06 0.12 ± 0.12 Nantan 2 11 0.04 ± 0.08 0.13 ± 0.07 0.06 ± 0.06 0.07 ± 0.06 0.06 ± 0.10 IIICD MEAN 0.20 ± 0.20 0.18 ± 0.09 0.04 ± 0.00 0.05 ± 0.00 0.07 ± 0.05

IIIE Staunton 8 0.61 ± 0.12 0.53 ± 0.03 −0.32 ± 0.06 −0.10 ± 0.06 0.50 ± 0.20

IIIF Clark County 7 −0.08 ± 0.12 0.02 ± 0.07 −0.65 ± 0.02 −0.14 ± 0.06 0.94 ± 0.11

IVA Gibeon 1 5 0.30 ± 0.20 0.30 ± 0.10 −0.27 ± 0.15 0.01 ± 0.11 0.50 ± 0.11 Gibeon 2 4 0.39 ± 0.17 0.36 ± 0.10 −0.25 ± 0.07 −0.04 ± 0.06 0.22 ± 0.05 Muonionalusta 7 0.28 ± 0.19 0.29 ± 0.09 −0.24 ± 0.09 −0.13 ± 0.09 0.37 ± 0.18 Obernkirchen 6 0.45 ± 0.19 0.38 ± 0.09 −0.24 ± 0.07 −0.09 ± 0.07 0.30 ± 0.13 IVA MEAN 0.36 ± 0.10 0.33 ± 0.05 −0.25 ± 0.01 −0.08 ± 0.07 0.34 ± 0.04

IVB Cape of Good Hope 10 0.20 ± 0.17 0.15 ± 0.11 −0.74 ± 0.04 −0.26 ± 0.08 0.94 ± 0.20 Santa Clara 12 0.25 ± 0.08 0.15 ± 0.05 −0.73 ± 0.06 −0.29 ± 0.09 0.99 ± 0.12 Tlacotopec1 9 0.38 ± 0.10 0.33 ± 0.07 −0.63 ± 0.05 −0.26 ± 0.06 0.90 ± 0.11 Tlacotopec 2 12 0.50 ± 0.12 0.37 ± 0.06 −0.57 ± 0.03 −0.32 ± 0.08 0.72 ± 0.09 IVB MEAN 0.30 ± 0.15 0.21 ± 0.13 −0.69 ± 0.09 −0.28 ± 0.02 0.91 ± 0.11 a Number of times sample was analysed; b Normalised to 97Mo/95Mo = 0.602083 using the exponential law, i i 95 i 95 4 ε Mo = [( Mo/ Mo)smp/( Mo/ Mo)std −1] × 10 ; Uncertainties for samples are 2se = 2σ/√n, where n is the number of times sample analysed. Uncertainties for group means are 2se = 2σ/√n, where n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of specimens is taken to represent that sample when calculating the group mean).

58 Chapter 2

2.4.2.4 Comparison to literature

The data presented here are in excellent agreement with Burkhardt et al. (2011) and of greater precision and quantity (Figure 2.13). Their study revealed the iron meteorite groups IC, IIAB, IID, IIE, IIIAB, IIIE, IIIF, IVA and IVB possess s-process deficit signatures in the ε range, while the IAB/IIICD group appears to have terrestrial Mo isotope composition. Intriguingly, the observation that the previously unmeasured IICs are significantly more anomalous than all iron meteorites measured to date, opens up the opportunity to further explore rare or ungrouped irons.

Figure 2.13 Comparison of the Mo data from this study with literature. (a) Means for the IIAB meteorites measured here with the group mean and uncertainty of Burkhardt et al. (2011) shown in grey. (b) The group mean of IIIAB (with 2se uncertainty) shown in blue field is about a factor of ~2 more precise than that of Burkhardt et al. (2011) shown in the grey field. All data are normalised to 98Mo/96Mo.

However, the data do not agree with Yin et al. (2002) or Becker and Walker (2003), who argued for no resolvable Mo isotope anomalies in iron meteorites. It is notable however, that apart from the IIC irons which they did not analyse, all other anomalies presented here are at the limit of the analytical resolution of their nTIMS studies, a point previously stated by Burkhardt et al. (2011). Notably, the results reveal the IAB and IIICD irons have Mo compositions indistinguishable from terrestrial Mo, with no difference between the groups, in contrast to Dauphas et al. (2002). The data here thus support the previously proposed concept of a common genetic link, with a IAB/IIICD complex consisting of a chemical main group and several subgroups (Wasson and Kallemeyn, 2002). A recent investigation into the different IAB/IIICD complex subgroups found the sHH and sHL subgroups were isotopically anomalous from the main group, while the sLL and sLM subgroups were not, interpreted as evidence against a single parent body for the complex (Worsham and Walker, 2015). In the study presented here, only meteorites from the main group and sLL and sLM subgroups were analysed, all of which display consistent isotope

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 59 compositions, in agreement with Worsham and Walker (2015). Since the anomalous sHH and sHL subgroups were not sampled here, no additional evidence can be provided with respect to Mo isotope variation within the IAB/IIICD complex. Another recently presented study argues the Mo isotope compositions of iron meteorites have been modified by exposure to galactic cosmic rays (GCR), based on meteorites from within the same group displaying variations in ε95Mo (normalised to 98Mo/96Mo) (Worsham et al., 2015). This follows the premise that since 95Mo has the largest neutron capture cross section and resonance integral of the Mo isotopes, it has the highest susceptibility to modification by GCR, most likely by the reaction 95Mo(n,γ)96Mo. Worsham et al. (2015) report deviations of 30 ppm in the 95Mo/96Mo ratio for Tlacotopec compared to the rest of the IVBs they measured. In this study, Cape of Good Hope and Santa Clara display ε95Mo = +0.98 ± 0.06 and +0.97 ± 0.06 respectively, whereas Tlacotopec 1 is slightly lower with ε95Mo = +0.84 ± 0.07 and Tlacotopec 2 even lower with ε95Mo = +0.65 ± 0.06. The offset of Tlacotopec 2 would seem to agree with the 30 ppm offset in 95Mo/96Mo reported by Worsham et al. (2015) for the same meteorite. Previous studies have shown Tlacotopec to be very irradiated, with a long history of cosmic ray exposure (Halliday and Kleine, 2006; Kruijer et al., 2013; Wittig et al., 2013); indeed, the Pt isotope data presented later in Chapter 4 also show it is by far the most irradiated sample used here. In this study, specimens Tlacotopec 1 and Tlacotopec 2 were originally adjacent pieces. Therefore, the difference between Tlacotopec 1 and 2, if due to GCR, suggests these affects can act on the mm- to cm-scale. The only other group to host a meteorite which has ε95Mo offset from the others in the group is the non-magmatic IIE: Weekeroo Station has ε95Mo = +0.54 ± 0.11, compared to ε95Mo = +1.04 ± 0.16 and +0.99 ± 0.16 for Kodaikanal and Verkhne Dnieprovsk, respectively. Again, the data for Weekeroo Station presented in Chapter 4 show its Pt isotope composition is more affected by GCR exposure than the other IIEs measured here, and so its slightly different ε95Mo may also be due to exposure to GCR. However, apart from Tlacotopec and Weekeroo Station, all other meteorites examined here exhibit identical ε95Mo (within uncertainties) to the rest of the meteorites from the same group. These uniform compositions are a clear indication that GCR exposure has not dramatically affected the Mo isotope compositions – only the most irradiated are affected, and even these display no more than 30 ppm deviations in 95Mo/96Mo. If GCR exposure were to have a dramatic affect, one would not expect the Mo isotope compositions to show such pronounced within-group homogeneity, and thus it appears that the isotope data presented

60 Chapter 2 here are dominated by nucleosynthetic processes and barely (if at all) by exposure to GCR. However, more investigation into the effects of GCR exposure on the other Mo isotopes and the isotope patterns it can generate are needed before any such effects can be completely disregarded. Finally, mass-independent Mo isotope effects observed in meteorites have previously been argued to arise from the nuclear volume field shift (Fujii et al., 2006b). A nucleus undergoing a neutron addition will also experience changes in the nuclear volume and electrostatic charge distribution, resulting in the altering of the atomic energy transitions of the inner s and p shell electrons (Bigeleisen, 1996). Fujii et al. (2006a) predict such variations in nuclear volumes can induce isotopic fractionation as a function of differing bonding energies. These authors formulated a method to calculate the nuclear field shift to an isotope ratio (i/x), internally normalised to the isotope ratio of (y/x) to correct for instrumental mass fractionation, in terms of ε-units (Equation 2.3).

! !! !!!!! ! ε! = δ � !!,!! − δ � !!,!! × α Equation 2.3 !! !!!!! 2 where mi is the atomic mass of nuclide i; δ⟨r ⟩x,i and is the change in the mean square radius of two nuclei x and i; α is an adjustable parameter

The nuclear volume field shifts for Mo isotopes are determined for internal normalisation to 98Mo/96Mo (Table 2.12), using Equation 2.3 and employing the nuclear masses from Wang et al. (2012) and the mean square nuclear radii from Angeli (2004).

Table 2.12 Nuclear field shifts (FS) for Mo isotopes, normalised to 98Mo/96Mo

εiMo/ε95Mo εiMo FS × α FS Model IIAB Data IIC Data (1) (2) (3) (4) (5) ε92Mo +0.134 +1.59 +2.84 +1.94 ε94Mo +0.056 +0.66 +2.33 +1.49 ε95Mo +0.084 - - - ε97Mo +0.078 +0.93 +0.51 +0.51 ε100Mo −0.119 −1.42 +0.75 +0.62 Masses of nuclides from Wang et al. (2012), mean square nuclear radii from Angeli (2004)

In Table 2.12, column 2, the field shift in terms of the parameter α is given. The model predicts that the resulting anomaly for ε94Mo is lower than both ε95Mo and ε97Mo – this is in contrast to the data presented in Table 2.9, where all meteorites (except those in the

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 61

IAB/IIICD complex) display ε94Mo anomalies greater than ε95Mo and ε97Mo. Furthermore, the model predicts a negative ε100Mo anomaly, whereas all meteorites exhibit positive ε100Mo anomalies (excluding the IAB/IIICD complex). In column 3, the dependence on α is removed by calculating the ratio of one field shift (εiMo) over another (ε95Mo). For comparison, the same ratios are calculated for the IIAB and IIC group means. Once again, the predicted nuclear field shifts fail to match the ratios observed here. The ratios ε92Mo/ε95Mo and ε94Mo/ε95Mo are lower than the data, while ε97Mo/ε95Mo are too high. Additionally, the ε100Mo/ε95Mo for the field shift is negative, while for the IIAB and IIC groups it is positive. To further assess the possibility of the observed mass-independent Mo isotope anomalies arising from nuclear volume field shifts, the parameter α has been optimised to best fit the IIAB and IIC data (Table 2.13). For the IIAB group means, the optimised parameter of α = 9.5 is used to calculate the field shift (column 2), while the best-fit of the field shift to the IIC group means is found using α = 22.0 (column 5). The effect of the nuclear field shift can be subtracted from the measured isotope anomalies (columns 3 and 6), to leave the residual anomalies (columns 4 and 7). If the nuclear volume field shift were entirely responsible for the mass-independent Mo isotope anomalies, then the residual anomalies in columns 4 and 7 should all be εiMo = 0: yet, it is clear they are not.

Table 2.13 Nuclear field shifts calculated for Mo isotopes to best fit the measured IIAB and IIC data

IIAB IIC εiMo a Field Shift b Measured Residual Field Shift c Measured Residual

(1) (2) (3) (4) (5) (6) (7) ε92Mo +1.27 +1.45 ± 0.10 +0.18 ± 0.10 +2.94 +3.12 ± 0.27 +0.18 ± 0.27 ε94Mo +0.53 +1.19 ± 0.06 +0.66 ± 0.06 +1.23 +2.40 ± 0.16 +1.17 ± 0.16 ε95Mo +0.80 +0.51 ± 0.04 −0.29 ± 0.04 +1.85 +1.61 ± 0.09 −0.24 ± 0.09 ε97Mo +0.74 +0.26 ± 0.02 −0.48 ± 0.02 +1.72 +0.82 ± 0.05 −0.90 ± 0.05 ε100Mo −1.13 +0.38 ± 0.08 +1.51 ± 0.08 −2.61 +1.00 ± 0.17 +3.61 ± 0.17 a Normalised to 98Mo/96Mo; b α optimised to 9.5; c α optimised to 22.0

The data from Table 2.13 are also plotted in Figure 2.14. For the IIAB group, only the ε92Mo anomaly is fully removed – there are residual positive ε94Mo and ε100Mo anomalies, and negative ε95Mo and ε97Mo anomalies. While the field shift model appears to be able to account for the IIC ε92Mo and ε95Mo anomalies, there exist residual positive ε94Mo and ε100Mo anomalies, and a negative ε97Mo anomaly.

62 Chapter 2

Figure 2.14 Comparison of observed Mo isotope anomalies with those predicted by nuclear field shift theory, using normalisation to 98Mo/96Mo. Panels on the left show the measured data for IIAB and IIC (blue lines; group means ± 2se), and the predicted isotope variations resulting from the nuclear field shifts (dashed green line), calculated using Equation 2.3. The parameter α was optimised to 9.5 for the IIAB group, and 22.0 for the IIC. Panels on the right show the residual isotope anomalies (purple lines) after correction of the nuclear field shift – the existence of such anomalies shows the nuclear field shifts are not responsible for the observed Mo isotope anomalies in iron meteorites.

The same modelling has also been applied to the 97Mo/95Mo internal normalisation scheme (Figure 2.15), with residual anomalies persisting for ε94Mo and ε100Mo for IIAB and IIC (as well as for ε92Mo for IIC). These results therefore indicate nuclear field shifts cannot be solely responsible for the Mo isotope anomalies measured in iron meteorites. Such conclusions have previously been made for Ba, Nd, Sm and Zr isotopes (Brennecka et al., 2011; Akram et al., 2013). Whilst Fujii et al. (2006b) argued all mass-independent Mo anomalies (except ε100Mo) in Allende could be explained by the nuclear field shift, it is important to note that the Mo data used, from Dauphas et al. (2002) and Yin et al. (2002) are less precise than the data here by a factor of ~ 5 to 10. With these greater uncertainties, it is easier to correct the anomalies for the nuclear volume field shifts to εiMo = 0. However, with the much more precise data made available by this study, the nuclear field shift can now be critically evaluated and deemed insufficient to create the observed Mo isotope anomalies.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 63

Figure 2.15 Comparison of observed Mo isotope anomalies with those predicted by nuclear field shift theory, using normalisation to 97Mo/95Mo. Panels on the left show the measured data for IIAB and IIC (blue lines; group means ± 2se), and the predicted isotope variations resulting from the nuclear field shifts (dashed green line), calculated using Equation 2.3. The parameter α was optimised to 3.7 for the IIAB group, and 10.7 for the IIC. Panels on the right show the residual isotope anomalies (purple lines) after correction of the nuclear field shift, demonstrating that nuclear field shifts are not responsible for the observed Mo isotope anomalies in iron meteorites.

2.5 Discussion

2.5.1 Molybdenum heterogeneity of solar nebula

While the Mo isotope compositions fit most closely with the s-process deficit models than any others in Section 2.2, the data do not fit the models perfectly – there are some discrepancies. This is most notable in the 97Mo/95Mo normalisation: while the IICs have the highest ε100Mo, they do not have the highest ε92Mo, which one would expect to be the case should the groups all have pure s-process deficits. Instead, IC, IIAB, IIE, IIIAB, IIIE and IVA have higher ε92Mo than IIC, IIIF and IVB, but the opposite is true for ε100Mo. This is summarised in Figure 2.16 – comparison with Figure 2.5(e) shows the yellow suite (IC, IIAB, IIE, IIIAB, IIIE and IVA) corresponds with the pure s-process deficit pattern (i.e., ε92Mo ≈ ε100Mo), but the brown suite (IIC, IIIF and IVB) is not in accord (ε92Mo << ε100Mo). Since 97Mo and 95Mo have similar s- and r-process contributions (59 and 55% s-process, respectively), the 97Mo/95Mo ratio is not significantly affected by s-deficits. Therefore, the discrepancies seen in Figure 2.16 at ε92Mo and ε100Mo indicate variations in

64 Chapter 2 p- and r-process components. For the IIC, IIIF and IVB groups, there are excesses in r-process components relative to p-process (but still overall p- and r-excesses relative to s-process).

Figure 2.16 Iron meteorite groups fall into two suites. One where ε92Mo is higher than ε100Mo (yellow shaded area) and one where ε100Mo is higher than ε92Mo (brown shaded field). The yellow area shows the mean of the IC, IIAB, IIE, IIIAB, IIIE and IVA groups with 2se uncertainties. The brown area shows the mean of the IIC, IIIF and IVB groups with 2se uncertainties. All data are normalised to 97Mo/95Mo.

To further assess the observed nucleosynthetic isotope effects, plots of εiMo vs. εiMo are presented (Figure 2.17 and Figure 2.18). Shown in these are the mixing lines for a p-excess, r- excess and s-deficit (solid orange, purple and green lines, respectively), calculated using the nucleosynthetic contributions from Arlandini et al. (1999). For comparison, alternative s-deficit mixing lines from the s-process production models of Bisterzo et al. (2011) (dashed green line) and Bisterzo et al. (2015) (dotted-dashed green line) are also shown. Just as the Mo isotope patterns of the iron meteorites were found to fit more closely with the s-process deficit models than any others from Section 2.2, the εiMo vs. εiMo plots generally fit closer to the s-deficit mixing lines that those for p- and r-excesses. Of particular significance, when an isotope with a p-process nucleosynthetic contribution is plotted against one with an r-process contribution (e.g., 92Mo vs. 95Mo), it is apparent there exists the same two suites as observed in Figure 2.16. This is demonstrated in Figure 2.17(a & b) – IIC, IIIF and IVB plot between the s-deficit and r-excess mixing lines, whereas the IC, IIAB, IIE, IIIAB, IIIE and IVA group means (enclosed in black circle) create a best-fit (solid black) line which is in close agreement to the s-deficit mixing lines.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 65

Crucially, in Figure 2.17(c & d), where the εiMo vs. εiMo plots use no p-process isotopes, no such distinctions are observed. Here, the black best-fit lines through the IC, IIAB, IIE, IIIAB, IIIE and IVA group means, generally also pass through the IIC, IIIF and IVB group means. This provides further evidence that variations in p- and r-process components create the two suites, meaning the observed nucleosynthetic anomalies do not reflect pure s-deficits.

Figure 2.17 εiMo vs. εiMo, normalised to 97Mo/95Mo = 0.602083. Shown are group means with 2se uncertainties. Theoretical p-excess, r-excess and s-deficit mixing lines (solid orange, purple and green lines, respectively) calculated using the nucleosynthetic s-process production model of Arlandini et al. (1999) are shown. For comparison, s-deficit mixing lines using the models of Bisterzo et al. (2011) (dashed green line) and Bisterzo et al. (2015) (dotted-dashed green line) are also shown. In (a) and (b), the data fall into two suites: the IC, IIAB, IIE, IIIAB, IIIE and IVA groups (enclosed by black circles) plot around a best-fit line (black line) which does not pass through the IIIF, IVB or IIC groups. In (c) and (d), there are no separate suites. Of the different s-process mixing lines, the black best-fit lines generally fall closest to those of Arlandini et al. (1999).

The same discrepancies from pure s-process deficits are also found with other normalisations. In εiMo vs. εiMo plots using normalisation to 98Mo/96Mo, the two suites can be observed where p- and r-process isotopes are plotted against each other – Figure 2.18(a & b) – but not where only r- and s-process isotopes are plotted against each other – Figure 2.18(c & d). The

66 Chapter 2 discrepancies are less pronounced than those found using normalisation to 97Mo/95Mo because the s-process contributions to 98Mo and 96Mo are less similar (76 and 100%, respectively) than 97Mo and 95Mo (59 and 55%).

Figure 2.18 εiMo vs. εiMo, normalised to 98Mo/96Mo = 1.453174. Shown are group means with 2se uncertainties. Theoretical p-excess, r-excess and s-deficit mixing lines (solid orange, purple and green lines, respectively) calculated using the nucleosynthetic s-process production model of Arlandini et al. (1999) are shown. For comparison, s-deficit mixing lines using the models of Bisterzo et al. (2011) (dashed green line) and Bisterzo et al. (2015) (dotted-dashed green line) are also shown. In (a) and (b), the data fall into two suites: the IC, IIAB, IIE, IIIAB, IIIE and IVA groups (enclosed by black circles) plot around a best-fit line (black line) which does not pass through the IIIF, IVB or IIC groups. In (c) and (d), there are no separate suites. Of the different s-process mixing lines, the black best-fit lines generally fall closest to those of Arlandini et al. (1999).

From herein, the pure s-process suite (groups IC, IIAB, IIE, IIIAB, IIIE and IVA) is denoted as the r=p suite, while the other s-process suite, with r-excess relative to p-excess, (IIC, IIIF and IVB) is denoted as the r>p suite. Notably, the best-fit lines through the groups in the r=p suite in Figure 2.17 and Figure 2.18 generally fall closest to the s-deficit mixing lines calculated from the models of Arlandini et al. (1999).

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 67

It is harder to resolve the r=p and r>p suites in the 92Mo/98Mo normalisation for two reasons. Firstly, the magnitudes of anomalies are much smaller and thus any variations from pure s-deficit patterns are difficult to resolve. Secondly, the normalisation scheme uses 92Mo, which is pure p-process, and 98Mo, which is s- and r-process (76 and 24%). This reduces the magnitude of any effects. However, at ε95Mo the anomalies can be resolved well enough to see the two fields when plotting ε100Mo vs. ε95Mo.

Figure 2.19 ε100Mo vs. ε95Mo, normalised to 92Mo/98Mo = 0.607898. Shown are group means with 2se uncertainties. Theoretical p-excess, r-excess and s-deficit mixing lines (solid orange, purple and green lines, respectively) calculated using the nucleosynthetic s-process production model of Arlandini et al. (1999) are shown. For comparison, s-deficit mixing lines using the models of Bisterzo et al. (2011) (dashed green line) and Bisterzo et al. (2015) (dotted-dashed green line) are also shown. The IC, IIAB, IIE, IIIAB, IIIE and IVA groups plot around the s-deficit line, while the IIC, IIIF and IVB groups do not, but rather plot in the field between the r-excess and s-deficit lines.

Whilst Burkhardt et al. (2011) argue for a pure s-process deficit, the precision of their data is not enough to resolve any p- or r-process heterogeneities. However, if their data for irons is included in the ε100Mo vs. ε92Mo plot with samples from this study (normalised to 97Mo/95Mo), the distribution into two fields still holds (Figure 2.20) – their IIIF and IVB irons plot in the r>p suite, while their IC, IIAB, IIIAB, IIIE and IVA irons fall in the r=p suite. Furthermore, including their chondrite data yields a tantalising observation – ordinary and enstatite chondrites plot on the s-deficit line (and thus fall into the r=p suite), whereas carbonaceous chondrites, which exhibit greater s-deficit anomalies, fall into the r>p suite (Figure 2.21). This significance of this will be discussed in Section 2.5.3.

68 Chapter 2

Figure 2.20 ε100Mo vs. ε92Mo, with integrated iron meteorite literature data. All data normalised to 97Mo/95Mo. Shown are group means from this study with 2se uncertainties. Theoretical p-excess, r-excess and s-deficit mixing lines (solid orange, purple and green lines, respectively) calculated using the nucleosynthetic s-process production model of Arlandini et al. (1999) are shown. Also shown are group means from Burkhardt et al. (2011) – their IC, IIAB, IIIAB, IIIE and IVA groups fall around the s-deficit line, but the IIIF and IVB irons fall in the r>p suite.

Figure 2.21 ε100Mo vs. ε92Mo, with integrated chondrite literature data. All data normalised to 97Mo/95Mo. Shown are group means for the irons from this study, with 2se uncertainties. Theoretical p-excess, r-excess and s-deficit mixing lines (solid orange, purple and green lines, respectively) calculated using the nucleosynthetic s-process production model of Arlandini et al. (1999) are shown. Also shown are group means, with 2se uncertainties, for chondrites from Burkhardt et al. (2011) – carbonaceous chondrites fall in the r>p suite, while ordinary and enstatite chondrites fall on the s-deficit line. Note CM data are not included, as this suffers from incomplete sampling of bulk rock material.

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 69

The heterogeneities observed in p- and r-process components indicate that the Mo isotope compositions of iron meteorites (and chondrites) are not just a result of a deficit of s-process components relative to the Earth, but also a decoupling of the p- and r-process components. Consequently, any models that account for the Mo isotope composition of the solar nebula must consider more than one process of heterogeneity.

2.5.2 Correlation with other elements

It has been well documented that neighbouring elements Mo and Ru correlate with s-process deficits (Dauphas et al., 2004; Burkhardt et al., 2012b; Fischer-Gödde et al., 2012; Bermingham et al., 2015). Using the Ru data of Fischer-Gödde et al. (2015), a linear regression of ε100Ru vs. ε92Mo from this study yields a slope of −0.45, Figure 2.22(a). This in agreement with the predicted slope of −0.44 calculated by Dauphas et al. (2004) from the s-process abundances of Mo and Ru from SiC grains. Likewise, Burkhardt et al. (2011) obtained a slope of −0.46, using the Ru data of Chen et al. (2010). Therefore, the data concur with the hypothesis that the Mo and Ru isotope anomalies reflect a heterogeneous distribution of a common s-process carrier. The variations and scattering around this trend, notably IVB, probably represent the variation in p- and r-process components. Support for this comes from Figure 2.22(b), where the correlation of ε100Ru vs. ε100Mo is tighter and shows less scattering, as no Mo p-isotopes are involved.

Figure 2.22 Correlation of Mo and Ru nucleosynthetic isotope anomalies in iron meteorites. Mo data from this study, normalised to 98Mo/96Mo = 1.453174. Ru data from Fischer-Gödde et al. (2015), normalised to 99Ru/101Ru = 0.7450754, using the exponential law. Ru and Mo error bars are 2se of the group means. (a) Plot of ε100Ru vs. ε92Mo. Regression of the data yields a slope of −0.45 (black line), symptomatic of a correlation of s-process Mo and Ru isotopes as calculated by Dauphas et al. (2004) (grey dashed line). (b) Plot of ε100Ru vs. ε100Mo. Less scattering is seen here due to absence of p-isotopes in plot.

Additionally, it has been demonstrated that Mo correlates well with its other neighbouring element Zr (Akram et al., 2015). While Zr isotopic analyses of iron meteorites are unavailable due to the lithophile behaviour of Zr, such data do exist for undifferentiated meteorites, which

70 Chapter 2 also have corresponding Mo isotope data. Using this, Akram et al. (2015) found a correlation between Mo and Zr isotope compositions in carbonaceous, ordinary and enstatite chondrites (Figure 2.23).

Figure 2.23 Correlation of Mo and Zr nucleosynthetic isotope anomalies in Allende CAIs, and carbonaceous, ordinary and enstatite chondrites, demonstrated in a plot of ε100Mo vs. ε96Zr .The Mo isotope data, normalised to 98Mo/96Mo, are from Burkhardt et al. (2011); Zr isotope data, normalised to 94Zr/90Zr, are from Akram et al. (2013) and Akram et al. (2015). Also shown are modelled s-process deficits using the nucleosynthetic models of Arlandini et al. (1999) (solid green line, A99), and Bisterzo et al. (2011) (dashed green line, B11).

Some differences were observed though – Burkhardt et al. (2012b) observed that whole rock and acid leachates of Mo isotopes fall on the same mixing lines in εiMo vs εiMo space, yet this is not the case for Zr isotopes (Akram et al., 2015). Crucially, Zr isotopes have lower neutron-capture cross sections and a neutron-capture pathway that is more sensitive to branching, and as such the isotopic compositions of s-process Zr are far more sensitive to AGB masses than Mo isotopes are. Since solar system Zr received significant s-process contributions from both low mass (1–3 solar masses) and intermediate-mass (5–8 solar masses) AGB stars (Travaglio et al., 2004), there are variations in s-process Zr isotope compositions. On the other hand, low mass AGB stars have been shown to account for the bulk (> 85%) of the s-process Mo in the solar system. Combined with the lower sensitivity of Mo isotopes to AGB star mass variations, this results in very limited variation in s-process Mo isotope compositions, in contrast to Zr. Such differences have been employed to explain why whole-rock and leachate Zr data reveal disparate isotope signatures, while Mo isotopes do not (Burkhardt et al., 2012b; Akram et al., 2015). Furthermore, this can also explain some

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 71 scattering on the Zr vs. Mo plot (Figure 2.23). Such differences mean that Zr and Mo s-process isotopes cannot be hosted in a single common carrier, as multiple carriers are required for the Zr isotopes. Furthermore, excesses in ε96Zr (due to deficits in s-process components) in bulk meteorites vary proportionately with excesses in the neutron-rich ε50Ti and ε54Cr (Trinquier et al., 2009; Schönbächler et al., 2011). The excesses in ε54Cr are thought to be caused by the heterogeneous distribution of nm-sized spinels (Dauphas et al., 2010). However, these spinels are low in Ti and are not the carrier of the ε50Ti excesses (Qin et al., 2011). Therefore, the ε54Cr excesses are hosted in a different carrier to ε96Zr and ε50Ti. By extension, the Mo s-process heterogeneity, which correlates extensively with that of Zr, must also arise from a different carrier to Cr and Ti neutron-rich isotope excesses (Akram et al., 2015). However, the heterogeneity of s-process Mo (and Ru, Zr) contrasts with the homogenous distribution of s-process Hf and Os (Yokoyama et al., 2007; Sprung et al., 2010; Yokoyama et al., 2010; Walker, 2012). Likewise, it is argued that Sm and Nd have a homogenous distribution of s-process and r-process isotopes, but do exhibit p-process deficits (Andreasen and Sharma, 2006). That is partly disputed though by claims in Carlson et al. (2007) of a Nd and Sm s-process deficiency in carbonaceous chondrites. It has thus been proposed, and demonstrated via modelling, that there may be several distinct carriers responsible for planetary-scale heterogeneities in the solar nebula (e.g., Burkhardt et al., 2011). For instance, a common carrier(s) may be responsible for the correlated Mo, Ru and Zr s-process deficits, with this carrier(s) being heterogeneously distributed in the solar nebula at the time of meteorite parent body formation. A different carrier(s) hosting the Hf and Os isotopes, uniformly distributed in the nebula, could explain the isotopic homogeneity of these elements, in contrast with the heterogeneity of Mo, Ru and Ba. Alternatively, a heterogeneously distributed single carrier could give rise to both the isotopic homogeneity and heterogeneity observed. If the carrier had low concentrations of Os and Hf (compared to Mo, Ru and Ba) and/or less extreme Hf and Os isotope compositions that are closer to the solar system average, then the affect of a heterogeneous distribution of Hf and Os isotopes would be far less than that for Mo, Ru and Ba, and may not be currently resolvable, such that isotopic homogeneity pertains. Leaching experiments of carbonaceous chondrites provide further support for these carrier relationships. For instance, in different leachates of the Murchison meteorite, Burkhardt et al. (2012b) found correlated s-process signatures, consistent with these isotope

72 Chapter 2 anomalies being hosted in common carriers. This is based on the premise that different phases are selectively attacked by certain acids during each leaching step, such that the different leaching steps release elements from particular phases. On the other hand, no obvious correlation was observed for Mo with the iron-peak elements (Ti, Cr, Ca) in the leaching steps, indicating that these reside in different carriers.

2.5.3 Origin of isotopic heterogeneity

Our current understanding of solar system formation posits large-scale mixing and transport of material in the solar nebula. This entails, over time, a progressive homogenisation of isotopically diverse components. Yet it has been shown above that the nebula was not completely homogenised with respect to Mo isotopes at the time of parent body formation. A number of scenarios have been proposed to explain the origin of the isotopic heterogeneities and the apparent disagreement between elements. One such explanation is that the heterogeneity still pertained at the time of parent body formation because the process of homogenisation of the nebula was not yet complete, giving rise to the traditional idea of a “cosmic chemical memory”, as first proposed by Clayton (1982). In this model, isotope abundances were heterogeneously distributed in the presolar molecular cloud. Following the collapse of the cloud, chemical processing and mixing did not fully erase the memory of the initial composition, such that early solids inherited the heterogeneities. The discovery of more isotopic anomalies as analytical techniques and precision improved gave impetus to this idea. However, an alternative approach has been proposed, in which the solar nebula was initially homogenous but received a late injection of isotopically anomalous material to impart heterogeneity during the period of early solids formation (Trinquier et al., 2007; Lugaro et al., 2014). Such a late injection of s-process rich material may explain the Mo isotope anomalies observed in the iron meteorites. It could also explain why the IAB/IIICD complex has a Mo isotope composition similar to Earth – these bodies formed later than the rest of the iron meteorite parent bodies, such that the injected material may have spatially homogenised by the time these younger bodies formed. However, a significant obstacle to the late injection hypothesises is provided by the bulk-rock Zr heterogeneity observed by Akram et al. (2015). They demonstrate that the Zr s-process material is a homogenous mixture of material from multiple stellar sources. This mixture is variably distributed in the solar system (with respect to p- and r-process components), which cannot be the result of an injection of material from a single event. It

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 73 therefore seems unlikely that multiple late injections of Zr s-process material from multiple sources into a homogeneous solar nebula can mix together but not with the rest of the nebula components. Consequently, it is becoming increasingly apparent that an alternative mechanism is required for the isotopic heterogeneity observed for many different elements and isotope systems, and consensus is starting to lean towards more localised processes, either in the solar nebula or on parent bodies, being responsible for the isotopic anomalies in bulk meteorites. One such possibility is the physical sorting of different dust grain types, which carry different isotopic signatures. Regelous et al. (2008) argues that Ni isotope anomalies reflect physical sorting of different phases (silicates, sulphides and metals) within the solar nebula. Subsequent localised and preferential removal of particular phases will thus result in isotopic variations in the remaining material, from which the meteorite parent bodies formed. Similarly, grain-size sorting has been posited as a mechanism for selectively excluding or enriching certain host phases and thus isotopic compositions (Dauphas et al., 2010). However, the previously discussed correlation of neutron-rich Zr, Ti and Cr, and the need for Cr isotopes to be hosted in a different carrier to those of Ti and Zr isotopes casts doubt on this. The Cr anomalies are hosted in nm-sized spinels, but it is very unlikely the stellar winds, which sort the spinels, will also have the same interactions with the host phases of the neutron-rich Zr and Ti isotopes (Akram et al., 2015). It therefore appears that the anomalies in Zr and Ti (and thus Mo, since Mo and Zr correlate) are not a result of grain-size sorting. Yokoyama et al. (2011) has suggested aqueous alteration on parent bodies may have resulted in selective modification and redistribution of isotopically anomalous presolar components. For Mo isotopes this appears unlikely, again due to the correlation of Mo and Zr isotope anomalies. While Zr is an immobile element, Mo is sensitive to oxidation and more readily interacts with fluid, allowing it to be transported more easily (Akram et al., 2015). Such a process would result in a decoupling of Mo and Zr isotopes, yet this is not observed. Consequently, any alteration must have operated only on a local scale and not affected bulk rock compositions, and is not responsible for the Mo isotope anomalies observed by this study. Increasingly, the most promising models for producing the observed isotopic variations in bulk meteorites involve the destruction of thermally unstable presolar grains in the nebula. For instance, Trinquier et al. (2009) report correlated mass-independent variations of ε46Ti and ε50Ti, despite these isotopes having different nucleosynthetic origins. They argue the correlation suggests the presolar dust inherited from the molecular cloud was well mixed

74 Chapter 2 when the earliest solids were forming in the solar nebula, but subsequent thermal processing resulted in selective destruction of thermally unstable, isotopically anomalous components, producing the residual isotopic variability observed at the planetary scale in bulk meteorites. A similar mechanism has been proposed to explain the decoupling of Mo and W isotopes in bulk meteorites and acid leachates (Burkhardt et al., 2012a, b). Leachates of Murchison show broadly correlated Mo and W isotope anomalies associated with a variable distribution of s-process isotopes, such that it can be demonstrated that anomalous Mo and W are hosted in the same carriers. However, such a correlation does not exist for bulk meteorites. For W, only IVB and IID iron meteorites display nucleosynthetic isotope anomalies, with IIAB, IIIAB and IVA having W isotopes indistinguishable from terrestrial composition (once corrected for radiogenic and cosmogenic effects). Yet for Mo, this study has shown all magmatic iron meteorite groups, as well as non-magmatic IIE, display variable s-process deficits. Therefore, while the leachate data suggest Mo and W isotope anomalies are hosted in the same carriers, these isotope anomalies are decoupled at the bulk meteorite scale. The thermal processing model of Burkhardt et al. (2012b), is based on the premise that during evaporation under oxidising conditions, Mo becomes more volatile than W and could have been preferentially removed from the nebular dust and entered into the gas phase. Subsequently, the Mo isotope composition of the processed dust is modified because isotopically anomalous Mo is removed as volatile oxides, while the W isotope composition remains largely unaffected. The generally uniform W isotope compositions of bulk meteorites reflect the initially homogenous distribution of presolar material in the solar nebula, while the nucleosynthetic Mo isotope anomalies reflect the isotopic heterogeneity imparted on the initially homogenous mixture of presolar dust. Fundamental to the thermal processing model of Burkhardt et al. (2012b) is the basis that elements characterised by isotopic homogeneity (e.g. Os, Hf, W) are all highly refractory with 50% condensation temperatures of ~1700–1800 K, whereas elements that display bulk- meteorite isotopic heterogeneities are more moderately refractory and have lower condensation temperatures (e.g., Mo = 1587 K) (Lodders, 2003). However, the observed Zr isotope heterogeneity casts significant doubt on this model. Since Mo isotope variations for primitive and differentiated meteorites correlate well with s-process Zr deficits in bulk rocks (see Section 2.5.2), it is most likely the same processes that led to the variable distribution of s-process Mo in the solar system also led to related s-process effects in Zr isotopes (Akram and Schönbächler, 2013). Given the Zr isotope

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 75 heterogeneity cannot be explained by an injection of material from a single stellar source, a more likely scenario for the Zr s-process heterogeneity is thermal processing. However, the 50% condensation temperature of Zr is 1741 K: this is higher than that of Hf (1684 K) and close to W (1789 K) (Lodders, 2003). Therefore, the volatility of the elements cannot be the determining factor in whether a particular element has nucleosynthetic isotope anomalies generated by thermal processing. Instead, the most likely cause is the decoupling of ‘light’ (Z ≤ 56) and ‘heavy’ (Z > 56) r-process nuclides (e.g., Wasserburg et al., 1996; Akram et al., 2013). These have been demonstrated to have different nucleosynthetic sites of production. The heavy r-process isotopes are formed by the main r-process, with the preferred stellar site of neutron star mergers (Tsujimoto and Shigeyama, 2014a). Conversely, the light r-process isotopes are thought to be formed predominantly by the weak r-process and charged particle reactions (CPRs), as well as small contributions from the main r-process, in Type II supernovae and their neutrino driven winds (Wanajo, 2013). It thereby entails that the host phases of ‘light’ and ‘heavy’ r-process isotopes have different susceptibilities to thermal processing, perhaps due to different grain sizes and/or chemical compositions of the phases. All elements with observed s- or r-process heterogeneity in bulk meteorites (Mo, Ba, Ca, Cr, Ni, Ru, Ti and Zr) have Z ≤ 56, while those homogeneously distributed in the solar nebula have Z > 56 (Hf, Os, Pt and W). Therefore, it appears that the host phases of ‘light’ r-process isotopes are more susceptible to thermal processing than those phases hosting ‘heavy’ r-process, and indeed s-process, isotopes. In the context of such thermal processing models, the Mo s-deficits observed in iron meteorites in this study are created by the thermal destruction and removal of unstable phases hosting p- and r-process nuclides, as outlined in the following. The solar nebula was initially isotopically homogenous, as supported by the homogeneity of elements like Os and W, and the apparent homogenisation of Zr s-process material from multiple stellar sources. Then thermal processing of solar nebula dust preferentially vaporised and removed Mo in p- and r-process host phases, which was then separated from the dust before it could completely re-condense. As such, the residual processed dust, from which the planets and meteorite parent bodies formed, has an s-process excess, while the vaporised component contains p- and r-process excesses (relative to the previously homogenised nebula). Since the material closer to the proto-Sun would have been more thermally processed, it stands that the Earth is formed from material which was more processed than the material from which meteorite parent bodies accreted (except IAB/IIICD,

76 Chapter 2 which may have formed close to the Earth and sampled the same material). Consequently, the Earth has a greater proportion of s-process material and meteorites have an s-deficit relative to the Earth. Under the same reasoning, the meteorite parent bodies which formed furthest out from the proto-Sun would have accreted from the least processed material, and should thus have the greatest s-deficit relative to the Earth. The Mo isotope data therefore suggests that the IIC parent body formed the furthest out from the Earth and Sun. This is supported by the data of Burkhardt et al. (2011), who observe that carbonaceous chondrites have greater s-deficits than ordinary and enstatite chondrites. Carbonaceous chondrites are thought to have formed further out than ordinary and enstatite chondrites (Chambers, 2006), and as such would have formed from less processed material. Furthermore, enstatite chondrites are thought to have formed further towards the Sun than ordinary chondrites, and this is also reflected by the Mo isotope composition for the enstatites being more terrestrial-like (i.e., lower s-deficit) than the ordinary chondrites. Significantly, this study updates the thermal processing model for Mo isotopes with the evidence of p- and r-process decoupling and heterogeneities therein. Whilst p- and r-process Mo isotopes are both thought to have formed during explosive stellar nucleosynthesis they do not originate entirely from the same type of supernova. The p-process Mo isotopes are thought to form in both Type Ia and Type II supernovae, while r-process isotopes form by the weak r-process and CPRs in Type II. Therefore, phases hosting p- and r-process Mo isotopes formed in Type II supernovae are likely to have very similar properties and susceptibilities to thermal processing. However, the p-process components formed in Type Ia supernovae likely have slightly different properties, including susceptibilities to thermal processing, than those p-process components formed in Type II supernovae. The data from this study reveal that iron meteorites from parent bodies closer into the Sun (as indicated by their lower s-deficits) have pure s-process deficits. This would indicate that under the conditions in this region, denoted the r=p region in the following, r-process host phases are similarly susceptible to thermal processing as p-process host phases. As we move out further from the Sun, and as the s-deficits increase, there comes a point where the iron meteorites have excesses in r-process components relative to p-process. In this region, denoted as the r>p region, p-process host phases are now more thermally unstable than r-excess phases. This perhaps reflects the different properties of the Type Ia and Type II p-process host phases. Indeed, Type Ia supernovae dust is thought to be very fine-grained, more so than Type II dust (Yu et al., 2013; Meyer and Clayton, 2015). This means it is

Nucleosynthetic molybdenum isotope anomalies in iron meteorites 77 probably more thermally susceptible, so that at greater heliocentric distances, a greater proportion of p-process Mo is removed, resulting in an excess of r-process isotopes relative to p-process, giving rise to the r>p region. The updated thermal processing model for Mo isotopes now relates both the size of the s-deficits, and the decoupling associated with the r>p region, with heliocentric distance, and these systematics are summarised in Figure 2.24.

Figure 2.24 An updated model of thermal processing of Mo nucleosynthetic components. Excesses and deficits of p-, s- and r-processes are relative to the initially homogenous nebula (dashed red line). With increasing heliocentric distance, the size of the s-excess (solid green line) decreases, while the p- and r-deficits also decrease (orange and purple lines respectively). Closest to the Earth, in the r=p region, the r-deficit is equal to the p-deficit, and thus there is a pure s-excess. In the r>p region, the p-deficit is greater than the r-deficit, resulting in an excess of r- relative to p-process nuclides, and thus an impure s-excess. Note the diagram is not to scale, and the listed order of meteorites within each region are in numerical order only – many groups overlap within uncertainty and the entire order of increasing s-deficits cannot be resolved.

Notably, the chondrite data from Burkhardt et al. (2011) provides support for this interpretation. The carbonaceous chondrites belong to the r>p suite (see Section 2.5.1 and Figure 2.21), while the ordinary chondrites and enstatite chondrites are in the r=p suite. These observations are in accord with the model and their respective placing into the r=p and r>p regions, as carbonaceous chondrites are thought to have formed further out from the Sun than ordinary and enstatite chondrites (Chambers, 2006), and in accord with the size of their respective s-process deficits.

78 Chapter 2

2.6 Conclusions

Analyses of Mo nucleosynthetic isotope anomalies in magmatic iron meteorites yield clear evidence of variable deficits in s-process nuclides, while the non-magmatic IAB/IIICD complex is observed to have terrestrial Mo isotopic compositions. These findings are in agreement with several recent high-precision studies. The extensive dataset presented here provides the most precise data and covers the broadest range of iron meteorites analysed yet, and facilitates the resolution of a decoupling between p-process and r-process Mo nuclides. In accord with the size of the s-process deficits, the extent of decoupling is dependent on heliocentric distance. Comparison of the Mo isotope data with studies of other elements (Ru, Ba, W, Zr, Os, Hf, Ti, Cr, Ni) suggest the most likely cause of this Mo isotopic variability is thermal processing and selective destruction/removal of unstable presolar components, for which an updated model is presented.

Chapter 3

Molybdenum stable isotopes in iron meteorites

80 Chapter 3

3.1 Introduction

Measurements of stable isotope compositions can be used to determine the extent of mass-dependent isotope fractionation recorded by a sample relative to a standard reference material. Studies of such effects for various elements in meteorites and their components have previously been interpreted as reflecting primarily (i) condensation and evaporation processes in the solar nebula and on meteorite parent bodies (e.g., Sr; Moynier et al., 2010), (ii) core formation and associated metal–silicate differentiation (e.g., Si; Armytage et al., 2011), and (iii) localised fractionation processes within parent bodies (e.g., Zn; Bridgestock et al., 2014). Mass-dependent fractionation of Mo isotopes has been a useful tool in geochemistry for some time (e.g., Siebert et al., 2003; Yang et al., 2015). However, very little is currently known about Mo isotope fractionation in solar system materials and the solar nebula. Results of some previous investigations into Mo isotope variations in iron meteorites were questioned due to their very large analytical uncertainty (e.g., Wieser and De Laeter, 2009). More recently, Burkhardt et al. (2014) reported the first Mo stable isotope data for chondrites, achondrites and iron meteorites, obtained by the more precise double spike technique. With the improved precision of the available data, these authors suggested that Mo stable isotope compositions could be used as a tracer of core formation conditions, based on the partitioning of Mo isotopes between silicate and metal during differentiation of planetary bodies. A key part of their study was to determine how to resolve mass-independent and mass-dependent isotope signatures – only by removing the nucleosynthetic isotope anomalies can a representative stable isotope composition be achieved. Following on from this, presented here is a more comprehensive study of Mo stable isotopes in iron meteorites to further investigate the potential they offer to resolve parent body processes. Taking advantage of the high-precision nucleosynthetic isotope anomalies reported in Chapter 2, this study strives to achieve the most accurate measurements of stable Mo isotopes yet accomplished. Moreover, the prospective application of Mo stable isotope data to study mass-dependent fractionation events in the solar nebula is explored. Of particular interest is the potential for stable Mo isotopes in iron meteorites to record mass-dependent isotope fractionations associated with the thermal processing that led to the nucleosynthetic Mo isotope anomalies. Such a correlation has yet to be observed, but with the improved precision this study affords, a relationship may, if indeed present, finally be resolved.

Molybdenum stable isotopes in iron meteorites 81

3.2 Analytical techniques

All materials and reagents used, and the laboratory conditions, were the same as those outlined in Chapter 1. Aliquots of the iron meteorite sample solutions that were prepared as described in Chapter 1 were employed for the analyses that are detailed below. Additionally, solutions of troilite nodules (FeS) from the IAB Toluca meteorite were obtained from the study of Bridgestock et al. (2014) – see paper for a detailed description of their separation and digestion methods. The latter samples were chosen for investigation because mass-dependent isotope variations have previously been observed for other elements in troilite nodules (e.g., Cu – Williams and Archer, 2011), and may also be present for Mo.

3.2.1 Molybdenum double spike preparation and calibration

This study used a Mo double spike previously prepared at the Imperial College London MAGIC Laboratories by Goldberg et al. (2013), and their preparation work is briefly described in the following. Enriched single spikes of 97Mo and 100Mo were obtained from Isoflex USA as Mo-oxides and mixed in a ratio of 1 to 1.6, based on the double spike toolbox calculations of Rudge et al. (2009), to produce the 100Mo-97Mo double spike (Figure 3.1). The method of Siebert et al. (2001) was applied to calibrate the double spike composition, using external normalisation to admixed Pd for correction of instrumental mass bias.

Figure 3.1 Isotope composition of the 100Mo-97Mo double spike. Shown are the double spike Mo isotope abundances (blue bars) in comparison to the terrestrial Mo isotope abundances (red bars) from Berglund and Wieser (2011).

82 Chapter 3

3.2.2 Ion-exchange chromatography

Separation of Mo was achieved by the same two-stage procedure as described in Chapter 2, with 50–200 mg of meteorite (~800 ng natural Mo) processed for each sample (Table 3.1). Prior to separation chemistry, the samples were mixed and equilibrated with the Mo double spike by refluxing on a hotplate for at least 24 hours.

Table 3.1 Ion-exchange chemistry for separation of Mo from iron meteorites

Stage 1: Bio-Rad column, 2 ml resin reservoir, 8 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (2 ml) Step Resin volumes Acid

Cleaning 15 3 M HNO3 Pre-condition resin 2 1 M HF / 0.5 M HCl Load Sample 5–10 1 M HF / 0.5 M HCl Rinse matrix (inc. Fe, Ni) 4 1 M HF / 0.5 M HCl Rinse matrix (Ru) 5 1 M HF

Collect Mo 5 3 M HNO3

Stage 2: Teflon column, 0.15 ml resin reservoir, 3 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (0.15 ml) Step Resin volumes Acid

Cleaning 16 3 M HNO3 Pre-condition resin 9 4 M HF Load Sample 5 4 M HF Rinse matrix (Ru) 12 4 M HF Elute Zr & W 7 6 M HCl / 1 M HF

Collect Mo 5 3 M HNO3

3.2.3 Mass spectrometry

3.2.3.1 Instrumentation and data collection protocol

The isotope measurements were performed using a Nu Instruments Nu Plasma HR MC-ICP-MS at the Imperial College London MAGIC Laboratories. Sample introduction utilised a Nu Instruments DSN-100 desolvating nebuliser that was used with a glass nebuliser at an uptake rate of ~140 µl/min. Typical sensitivity for Mo was 150–180 V/ppm for solutions with ~200 ppb Mo.

Molybdenum stable isotopes in iron meteorites 83

The data were acquired by static multiple collection with the Faraday cups of the instrument, with simultaneous measurement of the 95Mo, 97Mo, 98Mo, 100Mo and 99Ru ion beams (using 1011 Ω amplifiers) in a single cycle. The Faraday cup configuration and positions of the masses are indicated in Table 3.2. The measurements utilised 3 blocks with 20 integrations of 5 seconds each. Each block was preceded by a 30 second on-peak baseline measurement while the ion beam was deflected by the electrostatic analyser.

Table 3.2 Faraday cup configuration for measurements of stable Mo isotopes

Faraday Cup L5 L4 L3 L2 L1 Ax H1 H2 H3 H4 H5 H6

amu 95 97 98 99 100

Data are reported in δ98Mo notation (Equation 3.1), calculated relative to the mean of several bracketing runs of spiked NIST SRM 3134 Mo solutions. The mixed double spike – SRM 3134 solutions were thereby made up to closely match both the Mo concentrations (of ~200 ppb) and molar Mospike/Monatural ratios of the samples. All data reduction took place offline, applying the measured isotope ratios of 95Mo/98Mo, 97Mo/98Mo and 100Mo/98Mo using the iterative approach of Siebert et al. (2001), implemented in a Microsoft Excel spreadsheet (Ripperger and Rehkämper, 2007; Xue et al., 2013).

98Mo 95Mo 98 sample 3 δ Mo = − 1 × 10 Equation 3.1 98Mo 95Mo standard

3.2.3.2 Interferences

The possible spectral interferences on Mo isotopes that were used in the double spike data reduction are shown in Table 3.3. In part, this particular choice of isotopes reflects the desire to avoid isobaric interferences from Zr. This is readily possible as the only Mo isotopes with Zr isobars are 92Mo, 94Mo and 96Mo, and none of these are used in the double spike data reduction scheme. Isobaric interferences from Ru on 98Mo and 100Mo were corrected using 99Ru as the interference monitor and the Ru isotopic abundances of Becker et al. (2002). Spectral interferences from molecular species and doubly charged ions (Table 3.3) were also monitored but the efficiency of the separation chemistry for Mo was sufficient so that all relevant interferences were negligible.

84 Chapter 3

Table 3.3 Important spectral interferences on Mo isotopes

Interference 95Mo 97Mo 98Mo 100Mo

98 100 Isobars Ru Ru

Doubly charged ions 190Os2+ 194Pt2+ 196Pt2+ 200Hg2+ 190Pt2+ 196Hg2+ Argides 55Mn40Ar 57Fe40Ar 58Fe40Ar 60Ni40Ar

Oxides 79Br16O 81Br16O 82Se16O 84Se16O 82Kr16O 84Kr16O

3.2.4 Resolving mass-dependent and mass-independent Mo isotope effects

Figure 3.2 Schematic of the Mo double spike technique in four-isotope space. Adapted from Galer (1999). For details see Rudge et al. (2009).

A graphical representation of the Mo double spike technique is provided in Figure 3.2 and the method is further described in brief below. The double spike composition, S, is known from the double spike calibration, while n is the composition of the δ = 0 standard reference material (SRM). The isotope composition of the sample, N, differs from n as a result of the natural mass fractionation fn. The isotope composition of a mixture of sample and double spike, m, is determined by a mass spectrometric measurement. This differs from the true, unfractionated, isotope composition of the mixture, M, as a result of the instrumental mass

Molybdenum stable isotopes in iron meteorites 85

discrimination or mass bias, fm. In four-isotope space (Figure 3.2), the points N, n and S lie on the same plane, Pn, while M, m and S lie on another plane, Pm. These two planes intersect along a line formed by N-M-S. From this mathematical framework, N can be calculated, and then used to determine fn relative to the isotope composition n of the SRM (for more details see Rudge et al., 2009). The double spike data reduction assumes that the isotopic difference between the sample and the SRM is solely the result of mass-dependent isotope fractionation. Yet, the results presented in Chapter 2 show iron meteorites clearly host resolvable nucleosynthetic Mo isotope anomalies, and so the measured δ98Mo values for the irons need to be corrected for the apparent mass fractionation that arises solely from these. The correction process employed here utilised an approach based on that developed by Burkhardt et al. (2014), and this is summarised in Figure 3.3 and explained below.

Figure 3.3 Schematic of the double spike correction procedure, for a hypothetical sample with Mo s-process deficits. The composition of NIST SRM 3134 defines δiMo = 0. (a) The red line represents the effect of nucleosynthetic isotope anomalies (internally normalised to 97Mo/95Mo) on the δiMo values. The blue line represents the δiMo values consisting of both mass-dependent isotope fractionation and apparent isotope fractionation from mass-independent effects. The green line is the true mass-dependent isotope fractionation of the sample, obtained by subtracting the red line from the blue line. (b) The process is shown with respect to δ98Mo only, with the procedural steps in numerical order (see text for details). NB. The diagrams are not to scale. Adapted from Burkhardt et al. (2014).

Instead of using the SRM composition for n in the double spike data reduction, the relevant nucleosynthetic isotope anomalies from Chapter 2 are added to the Mo isotope composition of the SRM. In essence, this yields a modified SRM composition n' that is equivalent to the isotope composition of the sample but devoid of any mass-dependent isotope effects. From this, the true isotope composition of the sample, N, is then calculated in the double spike data

86 Chapter 3 reduction and a δ98Mo value relative to the SRM isotope composition (which defines δ98Mo = 0) can then be calculated. However, this δ98Mo value quantifies the effects of both mass-dependent isotope fractionation and any apparent mass fractionation that is generated as a result of nucleosynthetic isotope anomalies (as represented by the blue line in Figure 3.3). To obtain only the mass-dependent isotope fractionation, the effect of the nucleosynthetic isotope anomalies on δ98Mo is determined from the internally normalised data obtained in Chapter 2 (red line) and then subtracted from the δ98Mo results obtained for 98 98 the sample runs (blue line). The resulting δ Mo values, reported as δ Mo(7/5), quantify solely the mass-dependent isotope fractionation (green line) relative to the SRM isotope composition. To assess the magnitude and importance of the corrections, the δ98Mo data that result 98 if no such corrections are applied are also calculated, and reported as δ Mo(SRM). These results are obtained when the SRM composition is applied as n in the double spike data reduction and no nucleosynthetic isotope anomalies are subtracted from the δ98Mo values generated. In this study, the data reduction employs the nucleosynthetic isotope data internally normalised to 97Mo/95Mo, on the premise that 97Mo and 95Mo have very similar s-process contributions of 59 and 55%, respectively. Therefore any variations in the proportions of s-process Mo will not significantly affect the 97Mo/95Mo ratio. Additionally, since this is the same scheme used by Burkhardt et al. (2014), this allows direct comparison of δ98Mo values.

3.3 Results

3.3.1 Standard solutions and reference materials

To ascertain the precision and accuracy of the data, several NIST SRM 3134 solutions with varying proportions of double spike Mo to natural Mo were analysed (Figure 3.4). The various mixtures were measured relative to a solution with a ratio of spike-derived Mo to natural Mo of S/N = 1. This is close to the optimum S/N = 0.91, as calculated using the double spike toolbox of Rudge et al. (2009). These analyses yielded two main results. (1) The S/N ratio does not significantly affect the internal precision of the measurements between values of about 0.5 and 4. (2) Ratios of S/N < 1 produce negative δ98Mo values, whereas ratios of > 2 produce positive δ98Mo; in contrast S/N values between 1 and 2 yield δ98Mo of 0.000 ± 0.010‰.

Molybdenum stable isotopes in iron meteorites 87

Based on these results, the samples were spiked accordingly, aiming for S/N ≈1.5. In this case, if the actual spiking was slightly off because the Mo concentration of a sample was not characterised well, this will have no significant effect on data quality as long as the actual S/N is between about 1 and 2. In any cases where the S/N of a spiked sample was outside this range, a matched bracketing standard solution was produced that featured the same S/N value as the sample.

Figure 3.4 δ98Mo values of NIST SRM 3134 Mo solutions with variable ratios of spike-derived Mo to natural Mo (S/N). Calculated relative to a solution with S/N = 1. The grey shaded area represents the typical external reproducibility obtained for runs of the bracketing SRM solutions. For S/N ratios between 1 and 2, δ98Mo varies by < 0.01‰.

The typical external reproducibility obtained for the bracketing NIST SRM 3134 standard solutions with S/N ≈ 1.5 was δ98Mo = ±0.032‰. This is slightly better than the ±0.05‰ and ±0.056‰ previously reported by Goldberg et al. (2013) and Burkhardt et al. (2014) respectively. Primarily, this reflects that our analyses were conducted with more concentrated Mo solutions, which yielded higher beam intensities. The typical internal precision for the SRM runs was δ98Mo = ±0.018‰. The robustness and reproducibility of the Mo separation procedure and MC-ICP-MS Mo isotope measurements were routinely evaluated by repeated analyses of a Mo standard solution, ICL-Mo (Specpure 35758), which is used in-house as a stable isotope reference material (Goldberg et al., 2013). Aliquots of ICL-Mo that had been processed through the ion- exchange chromatography were routinely analysed, as were aliquots without any processing to monitor mass spectrometer performance. In both cases, the measurements yielded the same

88 Chapter 3 offset from NIST SRM 3134 (−0.170 ± 0.032‰ and −0.168 ± 0.023‰, 2σ, respectively), in agreement with the previously reported value of −0.15 ± 0.05‰ by Goldberg et al. (2013), confirming the procedures yielded consistent and accurate results (Figure 3.5). Finally, aliquots of NIST SRM 3134 were processed through the ion-exchange chromatography and analysed to further ensure that no analytical artefacts perturbed the isotope measurements. These yielded δ98Mo = +0.003 ± 0.032‰ (2σ, n = 13) (Figure 3.5).

Figure 3.5 δ98Mo for terrestrial standard reference materials. Measured relative to NIST SRM 3134. a Aliquots of standard solutions processed through Mo separation chemistry. Error bars denote the 2se in-run precision; uncertainties of reported means are 2σ. Grey-shaded area represents the typical ±0.032‰ reproducibility of the bracketing NIST SRM 3134 runs.

3.3.2 Iron meteorites

The Mo stable isotope compositions and concentrations of the iron meteorites are shown in Table 3.4 and Figure 3.6.

3.3.2.1 Corrections to δ98Mo values

The δ98Mo values corrected for mass-independent isotope effects, as well as the δ98Mo data that would result if no such corrections are applied, are shown in Table 3.4, labelled 98 98 ‘δ Mo(7/5)’ and ‘δ Mo(SRM)’, respectively. 98 98 The largest differences between δ Mo(7/5) and δ Mo(SRM) values are apparent for the 98 IIC irons – Ballinoo, Kumerina and Salt River are corrected to higher δ Mo(7/5) values by

Molybdenum stable isotopes in iron meteorites 89

0.055, 0.058 and 0.067‰, respectively. These larger corrections reflect that the IICs have the largest nucleosynthetic isotope anomalies of all meteorites analysed in Chapter 2. Like the other irons, the IAB/IIICD complex samples were also corrected for their mass-independent isotope effects, even though these meteorites display no nucleosynthetic isotope anomalies (see Chapter 2), with internally normalised Mo isotope compositions that 98 98 are identical, within uncertainty, to terrestrial Mo. Importantly, the δ Mo(SRM) and δ Mo(7/5) values for these samples are also identical within uncertainty. This further demonstrates that the correction procedure produces no analytical artefact but accurate stable isotope data. While no mass-independent isotope data are available for the Toluca troilite nodules, 98 it can be stated with high confidence that the δ Mo(SRM) values reported here for the troilites are completely dominated by stable isotope effects, and are not an artefact from incomplete correction of mass-independent isotope effects. Firstly, all meteorites from the IAB/IIICD complex were found to have terrestrial Mo isotope compositions (see Chapter 2). Since troilites are late-stage segregates from liquid metal, they must inherit the nucleosynthetic isotope signature of the metal and thus cannot have a distinct composition. One would therefore expect any troilites from the IAB/IIICD complex to also have terrestrial Mo isotope compositions. 98 Secondly, the δ Mo(SRM) values of the Toluca troilites are the most extreme reported here and very large mass-independent isotope effects would be required to generate the large measured stable isotope fractionations as analytical artefacts. Toluca Troilite 1 was found to 98 have δ Mo(SRM) = –0.561 ± 0.011‰. In order to correct this value towards the IAB/IIICD 98 average δ Mo(7/5), mass-independent isotope anomalies of greater than 25 ε-units would be needed. Clearly, such an extreme nucleosynthetic signature is unexpected – the highest observed anomalies were ε92Mo ≈ 3 for the IIC meteorites. Furthermore, Toluca Troilite 2 has 98 a very positive δ Mo(SRM) of +0.417 ± 0.013‰. For this to be produced by unresolved mass-independent isotope effects, s-process excesses of greater than 20 ε-units would be required, whereas no s-excesses, even in the 1 ε-unit range, were found for any iron meteorite. 98 Consequently, the reported δ Mo(SRM) values for the Toluca troilites are likely to predominantly, if not exclusively, reflect mass-dependent isotope effects. 98 98 From herein, any discussion of δ Mo values refers to only the corrected δ Mo(7/5) 98 values for all meteorites, with the exception of the troilites, where only the δ Mo(SRM) values are available.

90 Chapter 3

Table 3.4 Molybdenum concentration and stable isotope compositions of the iron meteorites

a b 98 c 98 d Group Sample N Mo (µg/g) δ Mo(SRM) (‰) δ Mo(7/5) (‰) IAB Toluca Troilite 1 5 2.9 0.417 ± 0.013 - Toluca Troilite 2 4 9.5 –0.561 ± 0.011 - Toluca Troilite 3 18 12.5 0.086 ± 0.010 - IAB TROILITES MEAN –0.019 ± 0.574 -

IAB Bitburg 15 27.0 0.370 ± 0.009 0.368 ± 0.010 Campo del Cielo 1 20 8.5 –0.088 ± 0.005 –0.091 ± 0.010 Campo del Cielo 2 11 8.1 –0.140 ± 0.008 –0.144 ± 0.008 Campo del Cielo 3 10 3.0 –0.097 ± 0.009 –0.095 ± 0.015 Canyon Diablo 1 12 6.4 –0.038 ± 0.012 –0.034 ± 0.013 Canyon Diablo 2 9 6.6 –0.059 ± 0.007 –0.056 ± 0.008 Cosby's Creek 13 35.4 0.452 ± 0.003 0.443 ± 0.004 Odessa 11 6.0 0.213 ± 0.012 0.222 ± 0.012 Toluca 25 6.6 –0.096 ± 0.009 –0.101 ± 0.010 IAB MEAN 6.6 0.130 ± 0.203 0.129 ± 0.201

IC Arispe 1 9 6.6 –0.166 ± 0.011 –0.149 ± 0.012 Arispe 2 10 6.5 –0.158 ± 0.009 –0.145 ± 0.009 Arispe 3 6 6.8 –0.050 ± 0.009 –0.033 ± 0.009 Arispe 4 8 6.9 –0.216 ± 0.012 –0.195 ± 0.013 Bendego 1 15 7.1 –0.174 ± 0.024 –0.154 ± 0.024 Bendego 2 11 6.9 –0.153 ± 0.007 –0.143 ± 0.008 Santa Rosa 10 7.5 –0.187 ± 0.007 –0.177 ± 0.008 IC MEAN –0.166 ± 0.023 –0.152 ± 0.027

IIAB Coahuila 11 6.9 –0.198 ± 0.012 –0.177 ± 0.012 Murphy 15 6.1 –0.182 ± 0.009 –0.162 ± 0.010 North Chile 21 7.0 –0.201 ± 0.011 –0.172 ± 0.012 Sikhote Alin 1 13 14.2 –0.420 ± 0.010 –0.403 ± 0.011 Sikhote Alin 2 10 6.7 0.009 ± 0.014 0.020 ± 0.014 IIAB MEAN –0.196 ± 0.010 –0.175 ± 0.012

IIC Ballinoo 10 8.6 –0.276 ± 0.011 –0.222 ± 0.012 Kumerina 11 8.2 –0.271 ± 0.010 –0.213 ± 0.010 Salt River 10 8.6 –0.263 ± 0.008 –0.196 ± 0.010 IIC MEAN –0.270 ± 0.004 –0.210 ± 0.015

IIE Kodaikanal 11 7.3 –0.196 ± 0.010 –0.193 ± 0.011 Verkhne Dneiprovsk 11 7.8 –0.195 ± 0.010 –0.184 ± 0.010 Weekeroo Station 11 6.2 –0.168 ± 0.011 –0.163 ± 0.012 IIE MEAN –0.187 ± 0.018 –0.180 ± 0.017

IIIAB Bear Creek 11 9.6 –0.313 ± 0.022 –0.295 ± 0.023 Cape York 10 6.5 –0.347 ± 0.008 –0.339 ± 0.009 Charcas 1 6 7.1 –0.167 ± 0.011 –0.160 ± 0.011 Charcas 2 9 6.9 –0.165 ± 0.014 –0.146 ± 0.014 Henbury 1 10 5.8 –0.161 ± 0.011 –0.143 ± 0.012 Henbury 2 10 5.8 –0.150 ± 0.007 –0.134 ± 0.007 Lenarto 10 8.4 –0.128 ± 0.008 –0.115 ± 0.009 Santa Apolonia 1 10 6.0 –0.173 ± 0.014 –0.159 ± 0.014 Santa Apolonia 2 13 6.2 –0.149 ± 0.006 –0.138 ± 0.007 Verkhne Udinsk 10 6.6 –0.183 ± 0.015 –0.168 ± 0.016 Williamette 8 4.2 –0.174 ± 0.013 –0.165 ± 0.013 IIIAB MEAN –0.203 ± 0.057 –0.190 ± 0.057

IIICD Carlton 1 7 2.3 –0.033 ± 0.012 –0.041 ± 0.010 Carlton 2 7 3.1 –0.048 ± 0.015 –0.051 ± 0.017 Nantan 1 5 8.4 –0.030 ± 0.011 –0.032 ± 0.011 Nantan 2 9 9.5 –0.026 ± 0.018 –0.033 ± 0.019 IIICD MEAN –0.034 ± 0.013 –0.039 ± 0.013

IIIE Staunton 11 8.3 –0.197 ± 0.008 –0.169 ± 0.012

IIIF Clark County 8 4.8 –0.223 ± 0.009 –0.198 ± 0.010

IVA Gibeon 1 13 5.3 –0.180 ± 0.010 –0.170 ± 0.013 Gibeon 2 11 5.5 –0.185 ± 0.010 –0.178 ± 0.010 Muonionalusta 11 5.5 –0.182 ± 0.005 –0.163 ± 0.006 Obernkirchen 8 5.2 –0.175 ± 0.007 –0.162 ± 0.008 IVA MEAN –0.180 ± 0.005 –0.166 ± 0.008

IVB Cape of Good Hope 9 22.7 –0.255 ± 0.012 –0.214 ± 0.012 Santa Clara 9 29.5 –0.232 ± 0.007 –0.190 ± 0.007 Tlacotopec 1 8 22.3 –0.198 ± 0.007 –0.158 ± 0.008 Tlacotopec 2 9 22.6 –0.193 ± 0.008 –0.153 ± 0.008 IVB MEAN –0.228 ± 0.035 –0.187 ± 0.034

a Number of analyses; b Mo concentrations determined by isotope dilution, typical uncertainties = 1–2%; c 98 ‘δ Mo(SRM)’ values obtained if no corrections are applied to the data for mass-independent isotope effects; d 98 ‘δ Mo(7/5)’ values corrected for mass-independent isotope effects, with uncertainties of mass-independent effects propagated; Uncertainties are 2se = 2σ/√n, where for samples n is the number of sample analyses, and for group means n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of specimens is taken to represent that sample when calculating the group mean).

Molybdenum stable isotopes in iron meteorites 91

98 Figure 3.6 δ Mo(7/5) of iron meteorites, corrected for mass-independent Mo isotope effects, from this study (circle symbols). Meteorites within the same group are plotted alphabetically. Also shown are δ98Mo values for iron meteorites (corrected for mass-independent effects) as measured by Burkhardt et al. (2014) (triangle symbols). The plotted uncertainties are 2se = 2σ/√n, where n is the 98 number of times a sample was analysed. The overall mean δ Mo(bulk) for the magmatic iron meteorite parent bodies is −0.177 ± 0.039‰ (2σ) (see Section 3.4.3 for details), shown here by the grey shaded area.

3.3.2.2 Mo concentrations and δ98Mo

Excluding IVBs, the magmatic iron meteorites have Mo concentrations ranging from 4.2 to 14.2 ppm. In contrast, the IVB irons have much higher Mo concentrations of 22.3 to 29.5 ppm, in accord with previous studies (Burkhardt et al., 2014). Non-magmatic irons have a greater variation of Mo concentrations, from 2.3 to 35.4 ppm. The troilites separated from Toluca IAB also fall within this range of concentrations. The δ98Mo values for the magmatic iron meteorites vary within a limited range, from –0.403 to +0.020‰. The non-magmatic irons show more variation (–0.144 to +0.443‰),

92 Chapter 3 whilst troilites (FeS) from the Toluca IAB iron display the greatest range (–0.561 to +0.417‰) and exhibit the lightest and nearly the heaviest δ98Mo values measured here. For all non-magmatic and magmatic groups analysed, there is no clear relationship between the Mo concentrations and δ98Mo values of the meteorites (Figure 3.7).

Figure 3.7 δ98Mo vs. Mo concentration of iron meteorites. Uncertainties on δ98Mo are propagated 2se. Uncertainties on Mo concentrations not shown, but were typically 1–2%.

Molybdenum stable isotopes in iron meteorites 93

3.4 Discussion

3.4.1 Non-magmatic iron meteorite parent bodies

98 The large variation in δ Mo values observed for the Toluca IAB troilites is similar to the extensive range observed for Cu isotopes in various troilite and metal samples (Williams and Archer, 2011). In this work, it is suggested the variations are most likely produced by disequilibrium processes, such as kinetic Cu isotope fractionation from preferential diffusion of the lighter isotopes from metal to sulphide. Based on this, it is proposed that the δ98Mo variations observed here for the Toluca IAB troilites may arise from similar processes which generated Mo isotope fractionation during diffusion of Mo between metal and sulphide phases. However, it is also conceivable that the δ98Mo variations of the troilites arise from equilibrium isotope fractionation during (equilibrium) metal–sulphide partitioning of Mo, or an intermediate process. Similarly, the IAB/IIICD complex irons show highly variable δ98Mo with both isotopically heavy and light Mo. This stands in contrast to the results of Burkhardt et al. (2014), who found only heavy Mo isotope signatures for such samples, which were interpreted to reflect evaporative loss of isotopically light Mo during energetic impacts. Based on the new data, and the Mo isotope fractionation observed for the troilites, the large range of δ98Mo values observed for IAB/IIICD irons is most likely due to the segregation and crystallisation of S-rich melts on the parent body (Benedix et al., 2000) and associated equilibrium partitioning or (non-equilibrium) diffusion between metal and sulphide phases. One could also argue that the wide variations in Mo concentrations in IAB/IIICD meteorites and IAB troilites may also arise from metal–sulphide partitioning. However, given that this is a non-magmatic body, the exact state of homogeneity of Mo in the segregated S-rich melts is unknown, and so no firm conclusions can be made regarding the impact of metal–sulphide partitioning on the concentration of Mo in the IAB/IIICD iron meteorites.

3.4.2 Magmatic iron meteorite parent bodies

Within the magmatic iron meteorite groups, the variations in δ98Mo are much smaller than those exhibited by the troilites and the IAB/IIICD complex irons. However, small but resolvable and systematic variations of δ98Mo within magmatic iron groups are found. For instance, in the IIAB group, Coahuila, Murphy and North Chile all have identical δ98Mo within uncertainty (−0.170 ± 0.009‰), but the Sikhote Alin samples 1 and 2 have markedly different δ98Mo values of −0.403 ± 0.011‰ and +0.020 ± 0.014‰, respectively. A

94 Chapter 3 plot of Mo stable isotope compositions versus Ni concentration suggests that the greatest variability in δ98Mo is seen at higher Ni concentrations (Figure 3.8). Like Ni, S is also incompatible in solid metal, and becomes enriched in the liquid metal during fractional crystallisation of cores (e.g., Wasson, 1970). This implies that iron meteorites, which form towards the end of the fractional crystallisation sequence, incorporate metal with higher Ni and S contents compared to irons that crystallised earlier. It is therefore posited that the δ98Mo variations seen within magmatic iron meteorite groups at the evolved, Ni-rich end of crystallisation sequences reflect processes that are associated with the increased liquid metal S content. Hence, these processes may be similar to those invoked for the troilites and the IAB/IIIICD complex irons, most likely involving metal–sulphide partitioning of Mo and associated isotope fractionation. In this context it is notable that the samples Sikhote Alin 1 and 2 are specimens from the same meteorite that clearly differ in δ98Mo (Table 3.4). This implies that the metal–sulphide fractionation processes and associated isotope fractionation can operate on the sub-metre scale.

Figure 3.8 δ98Mo vs. Ni concentration of IIAB irons. The greatest variability in δ98Mo is observed at higher Ni concentrations. Uncertainties in δ98Mo are 2se = 2σ/√n, where n is the number of sample analyses. Nickel data taken from Goldstein et al. (2014), Wasson (1969), Wasson (1970), Wasson (1974) and Buchwald (1975).

As for δ98Mo, some variability is seen in Mo concentration for iron meteorites from the same group. For instance, Sikhote Alin 1 has a Mo concentration of 14.2 ppm, whereas the other IIAB irons (including Sikhote Alin 2) have a limited range of Mo concentrations that vary between 6.1 and 7.0 ppm. Previous investigations have concluded that the IIAB parent body had a S-rich core and that at some point during crystallisation, the S content of the liquid metal exceeded a critical level, such that two immiscible liquids segregated, one S-rich and

Molybdenum stable isotopes in iron meteorites 95 the other P-rich (Chabot and Drake, 1999; Goldstein et al., 2009). The partitioning of Mo between solid and liquid metal phases is known to be dependent on the sulphur content of the liquid metal (Liu and Fleet, 2001). Therefore, one might expect variations in Mo concentrations to arise when solid metal crystallises from particularly S-rich melts, and this will occur for meteorites that fall near the Ni-rich end of a fractional crystallisation sequence. Notably, this is exactly what is seen for Sikhote Alin 1, indicating that both δ98Mo and Mo concentrations are affected by metal–sulphide partitioning during core crystallisation. The same correlation of δ98Mo with Ni concentration is seen for the IIIAB group irons (Figure 3.9). Bear Creek has the highest Ni concentration and δ98Mo is significantly lighter than for most other meteorites in the group. Lenarto has the second highest Ni concentration, and also has the heaviest δ98Mo value. Additionally, it is clear that Cape York has an unusually light δ98Mo value, especially given its low Ni concentration. Different masses of the Cape York meteorite are known to have variations in sulphur content by a factor of ~10 (Esbensen et al., 1982), which have been interpreted as being caused by trapped melt pools during the dendritic crystallisation of the IIIAB core (Haack and Scott, 1992; Wasson, 1999). Therefore, one would expect a wide range of δ98Mo signatures for the different masses of Cape York; indeed the specimen of Cape York analysed here has a δ98Mo value that differs clearly from the rest of the IIIAB irons with similar Ni concentrations. Hence it most likely reflects the presence of sulphide-rich metal and associated metal–sulphide partitioning and isotope fractionation.

Figure 3.9 δ98Mo vs. Ni concentration of IIIAB irons. Uncertainties in δ98Mo are 2se = 2σ/√n, where n is the number of times meteorite was analysed. Nickel concentrations from Scott et al. (1973).

96 Chapter 3

3.4.3 Bulk δ98Mo of the iron meteorite parent bodies

Table 3.4 shows the mean δ98Mo values that were calculated for the magmatic iron meteorite parent bodies using all available individual sample results. Hence, these averages are 98 significantly affected by meteorites with unusual δ Mo that result from localised metal– sulphide partitioning and isotope fractionation. By omitting these atypical results, a more representative average δ98Mo values for the parent cores can be obtained. Furthermore, as Mo is moderately siderophile with D(metal–silicate) ≈ 90 to 140 (McDonough, 2003; Palme and O'Neill, 2003), these results can essentially be considered as the δ98Mo value of the bulk 98 parent body, and denoted δ Mo(bulk) from herein. For the IIIAB parent body, this involves omitting the δ98Mo results for Bear Creek, Cape York and Lenarto from the calculation of the mean parent body δ98Mo value (Figure 3.10). Similarly, Sikhote Alin 1 and Sikhote 2 are omitted from the dataset that is used to 98 98 determine δ Mo(bulk) for the IIAB parent body. For the δ Mo(bulk) of the IC parent body, Arispe 3 is omitted because this sample is anomalously heavy compared to Arispe 1, 2 and 4, 98 and indeed the rest of the IC group. This observation suggests that the δ Mo of Arispe 3 was also influenced by small-scale metal–sulphide fractionation processes. Alternatively, it is also conceivable that any of the anomalous isotope compositions of the omitted specimens reflect the presence of small troilite nodules that were not visible in the hand specimens.

98 98 Figure 3.10 δ Mo vs. Ni concentration of IIIAB irons, showing how the parent body δ Mo(bulk) value was calculated. Bear Creek, Cape York and Lenarto are omitted from the calculation of the mean value due to the probable effect of metal–sulphide partitioning on the isotope compositions (see main text for details). Uncertainties in δ98Mo are 2se = 2σ/√n, where n is the number of times the meteorite was analysed. Nickel data taken from (Scott et al., 1973).

Molybdenum stable isotopes in iron meteorites 97

98 The δ Mo(bulk) values that were calculated following the above procedure are summarised in Table 3.5. Notably, the IIIE and IIIF parent bodies are only represented by one meteorite 98 each, Staunton and Clark County, respectively. However, the δ Mo values of these 98 meteorites are very similar to the δ Mo(bulk) values of the other parent bodies and this suggests that the stable Mo isotope compositions of Staunton and Clark County were probably not overprinted by late metal–sulphide fractionation processes. Hence, these values are considered as representative for the bulk parent bodies in the following. Overall, the iron meteorite parent body averages are remarkably well defined. Among 98 the results, the IVB irons display the greatest uncertainty in δ Mo(bulk) (Table 3.5). Given the similar Ni concentrations (16.9, 17.9 and 15.9%) of the three meteorites that were used to define the IVB bulk value, none of the samples can be omitted as anomalous or atypical, for example as a result of localised metal–sulphide processes. Hence, the IVB bulk value has a relatively large uncertainty.

Table 3.5 Bulk δ98Mo of magmatic iron meteorite parent bodies

a b 98 c Group Samples Omitted N δ Mo(bulk) (‰) IC Arispe 3 3 –0.163 ± 0.016 IIAB Sikhote Alin 1 3 –0.170 ± 0.009 Sikhote Alin 2

IIC - 3 –0.210 ± 0.015 IIIAB Bear Creek 5 –0.155 ± 0.011 Cape York Lenarto

IIIE - 1 –0.169 ± 0.012 IIIF - 1 –0.198 ± 0.010 IVA - 3 –0.166 ± 0.008 IVB - 3 –0.187 ± 0.034 MEAN d –0.177 ± 0.038 a 98 b Samples omitted from calculation of δ Mo(bulk), see text for details; number of remaining meteorites analysed in that group; c uncertainties for groups are 2se = 2σ/√n, where for groups with more than one meteorite analysed n is number of meteorites in group, but for groups with only one meteorite analysed n is number of times meteorite was analysed; d uncertainty of 98 mean δ Mo(bulk) is 2σ.

98 The overall mean of the δ Mo(bulk) for the magmatic iron meteorite parent bodies is −0.177 ± 0.039‰ (2σ) (Table 3.5, grey shaded area in Figure 3.6). Notably, this result is identical, within uncertainty, to a previously reported average of –0.16 ± 0.02‰ for iron meteorites and chondrites but excluding IAB irons and CK, CM chondrites (Burkhardt et al.,

98 Chapter 3

2014). These authors interpreted their result to reflect the Mo isotope composition of the inner 98 solar system. However, no such interpretation is suggested here, as the δ Mo(bulk) values of the parent bodies appear to vary spatially within the inner solar system, as discussed in more detail in the following sections. 98 There appears to be no clear correlation between the δ Mo(bulk) results and estimates for the initial liquid sulphur content of the parent body cores by Chabot (2004). This suggests 98 that differentiation and core formation on the parent bodies did not affect the δ Mo(bulk) value of the cores. During differentiation, S partitions into the core, and the presence of such S may affect the partitioning of Mo and Mo isotopes between the silicate and metal. This could lead to variations in stable Mo isotope compositions between the cores of bodies with different sulphur contents. However, despite the estimated differences between S contents, the bulk 98 δ Mo(bulk) values of the IIAB, IIIAB, IVA and IVB do not appear to be affected by this (Figure 3.11).

98 Figure 3.11 δ Mo(bulk) vs. initial liquid sulphur content of the IIAB, IIIAB, IVA and IVB parent bodies. Estimates of the initial liquid sulphur content of parent bodies can only be made for those magmatic iron meteorite groups that are well sampled by a large number of meteorites.

3.4.4 Stable Mo isotope evidence for thermal processing in the solar nebula

Nucleosynthetic Mo isotope anomalies in iron meteorites, as demonstrated in Chapter 2 most likely resulted from thermal processing of material in the solar nebula. This involved the destruction and removal of unstable phases hosting p- and r-process Mo isotopes. In this model, the material closest to the Sun experienced more thermal processing and hence lost a higher proportion of p- and r-nuclides. As a consequence, the Earth, which formed closer to the Sun than the iron meteorite parent bodies, was accreted from more material that

Molybdenum stable isotopes in iron meteorites 99 experienced more thermal processing in comparison to the material that accreted to form the iron meteorite parent bodies. One might expect mass-dependent fractionation of Mo isotopes to accompany the nucleosynthetic isotope effects – given that the temperatures within the solar nebula decreased with increasing distance from the Sun, any mass-dependent isotope fractionation associated with the thermal processing of p- and r-process host phases should also decrease with heliocentric distance. If these assumptions are correct, they would imply that the parent bodies with the greatest s-process deficits (i.e., those accreted from the least processed material) should feature lighter Mo stable isotope compositions (or lower δ98Mo) in comparison to bodies with smaller s-process deficits, which accreted from material that received more thermal processing in closer proximity to the Sun. 98 100 The Mo isotope data plotted in Figure 3.12, which displays δ Mo(bulk) vs. ε Mo (normalised to 97Mo/95Mo) for the iron meteorites, provide the first evidence for such a correlation. It must be noted that the calculated regressions assume that the mass-dependent and nucleosynthetic Mo isotope effects are linearly related due to variations in heliocentric distance. In reality, this relationship may differ in detail but this has no consequence for the interpretation of the observation.

98 100 Figure 3.12 δ Mo(bulk) vs. ε Mo nucleosynthetic Mo isotope anomalies for parent bodies of magmatic iron meteorites. Nucleosynthetic isotope data (ε100Mo) from Chapter 2, internally normalised to 97Mo/95Mo = 0.602083. Error bars are 2se uncertainties. A correlation exists between 98 increasing nucleosynthetic isotope anomalies and more negative δ Mo(bulk), in accord with the thermal processing model for the origin of nucleosynthetic Mo isotope anomalies. Regressions shows the 98 98 98 correlation exists when the δ Mo(bulk) are calculated using either the δ Mo(7/5) or δ Mo(SRM) data reported in Table 3.4 (solid and dashed black lines, respectively).

100 Chapter 3

It is notable that the correlation exists for Mo stable isotope data from all stages of the correction procedure for mass-independent isotope effects on δ98Mo values. In particular, the 98 correlation seen in Figure 3.12 is equally apparent if the δ Mo(bulk) values are calculated 98 98 using the uncorrected (δ Mo(SRM)) data rather than the corrected (δ Mo(7/5)) values reported in Table 3.4. To further scrutinise the significance and robustness of the correlation, any potential artefacts from data reduction must be examined. The corrections in the double spike data reduction for mass-independent isotope effects applied data internally normalised to 97Mo/95Mo because of the very similar s-process contributions to these two isotopes (Table 3.6). However, whilst the contributions are very similar, they are not exactly identical, and this could possibly generate data reduction artefacts. An exact 1:1 ratio of s-process contributions to 97Mo and 95Mo would give a true mass-independent correction, but any 98 departure from 1:1 could potentially result in corrected δ Mo(7/5) values having some residual 98 mass-independent artefacts. To assess this, the effects on δ Mo(bulk) were modelled using the s-process contribution models (Table 3.6) of Arlandini et al. (1999), Bisterzo et al. (2011) and Bisterzo et al. (2015), as described in the following.

(1) The Mo isotope abundances of a ‘sample’ were generated by adding s-nuclide deficits, representative of those presented in Chapter 2, to the SRM isotope abundances. (2) The isotope composition of the ‘sample’ was then mixed with that of the double spike. (3) A realistic instrumental mass fractionation was applied to the sample-double spike mixture. (4) The data from Step (3) were processed through the double spike data reduction. (5) The resulting δ98Mo value, if not 0, represents the remaining mass-independent artefact that persists from the small variations in s-process contributions to 97Mo and 95Mo.

The same modelling was also applied using a ‘sample’ that was mass-fractionated prior to double spike mixing, to replicate natural mass-dependent isotope fractionation. After double 98 98 spike data reduction, the differences between the calculated δ Mo(7/5) and the expected δ Mo values, based on the natural mass-fractionation that was applied to the ‘samples’, were identical to the offsets, which were obtained for ‘samples’ that featured no natural mass

Molybdenum stable isotopes in iron meteorites 101 fractionation. This demonstrates that the modelling procedure produces consistent and accurate results. From this modelling, the persisting artefacts on δ98Mo from the s-nuclide deficits exhibited by the meteorites (presented in Chapter 2) were estimated, and corrections for such 98 effects were applied to the δ Mo(bulk) values. The results are shown in Figure 3.13, whereby the observed slope of the correlation between nucleosynthetic and stable isotope anomalies is corrected for the effects of disparate s-process contributions to 97Mo and 95Mo, according to different s-process production models (green lines).

Table 3.6 The s-process contributions (%) to Mo isotopes for different s-process production models

92Mo 94Mo 95Mo 96Mo 97Mo 98Mo 100Mo Arlandini et al. (1999) 0 1 55 100 59 76 4 Bisterzo et al. (2011) 0 1 70 100 64 82 5 Bisterzo et al. (2015) 0 1 60 100 53 74 5

98 Figure 3.13 Plots of δ Mo(bulk) against nucleosynthetic Mo isotope anomalies for parent bodies of magmatic iron meteorites. Nucleosynthetic isotope data (ε100Mo) from Chapter 2, internally normalised to 97Mo/95Mo = 0.602083. Error bars are 2se uncertainties. A correlation exists between increasing 98 nucleosynthetic isotope anomalies and more negative δ Mo(bulk) values from Table 3.5 (solid black line). The correlation is further corrected for persisting mass-independent artefacts due to differences in s-process contributions to 97Mo and 95Mo, using the models of Arlandini et al. (1999) (solid green line), Bisterzo et al. (2011) (dashed green line) and Bisterzo et al. (2015) (dotted-dashed green line).

Crucially, the calculated corrections to the correlation are minimal since s-process contributions to 97Mo and 95Mo are so similar for all s-process production models. Hence, the corrections only affect the slopes of the correlation (green lines) in Figure 3.13 in a minor way. Notably, in order to affect the slope enough to remove any correlation, the ratio of

102 Chapter 3 s-process contributions to 97Mo and 95Mo would need to be approximately 3:1. Clearly, such a mismatch is not in accord with any model of s-process nucleosynthesis (Table 3.6). 98 100 Therefore, it can be stated with confidence that the δ Mo(bulk vs. ε Mo correlation of the iron meteorite parent bodies is not an artefact of data reduction and processing. Rather, 98 our precise isotope analyses were able to resolve variations in δ Mo(bulk) for iron meteorite parent bodies, which correlate with nucleosynthetic isotope anomalies and this correlation provides support for thermal processing models that were previously outlined in Chapter 2.

3.4.5 δ98Mo of Earth

Since the Earth accreted closer to the Sun than the iron meteorite parent bodies and thus experienced greater thermal processing, it is likely to be characterised by a slightly heavier 98 Mo isotope composition. In fact, an estimate of δ Mo(bulk) for the Earth can be made from the 98 correlation of δ Mo(bulk) with nucleosynthetic isotope anomalies. In Figure 3.12, as εiMo = 0 represents the terrestrial composition, the y-intercept of the 98 regression provides an estimate for the Earth’s δ Mo(bulk) value of −0.144 ± 0.009‰ (2σ). Notably, since the core contains most of the terrestrial Mo budget, this value is also representative of the stable Mo isotope composition of the Earth’s core. Of course, it should be noted that this result is based on the simplifying assumption of a linear relationship between nucleosynthetic isotope anomalies and mass-dependent Mo isotope fractionation, which may require refinement when considered in detail. 98 In the following, the estimated δ Mo(bulk) for the bulk Earth and its core is applied in the context of core formation to estimate the Mo stable isotope composition of the bulk 98 silicate Earth (δ Mo(BSE)). In particular, Hin et al. (2013) investigated Mo isotope fractionation during metal–silicate differentiation and derived an empirical formula for Mo isotope fractionation as a function of temperature (Equation 3.2). Using the predicted δ98Mo 98 for the Earth’s core of (δ Mo(bulk)) = −0.144 ± 0.009‰, and a range of core formation temperatures in accord with literature data (e.g., Wade and Wood, 2005; Cottrell et al., 2009), 98 estimates can be obtained for δ Mo(BSE) by re-arranging Equation 3.2 to give Equation 3.3.

98 !4.70 ± 0.59 ∆(δ Mo)Metal–silicate = Equation 3.2 (105 / T2)

!4.70 ± 0.59 δ98Mo = δ98Mo − Equation 3.3 (BSE) (bulk) (105 / T2)

Molybdenum stable isotopes in iron meteorites 103

As shown in Table 3.7, for core formation temperatures ranging from 1750 to 4000 °C, 98 δ Mo(BSE) values of between +0.019 and −0.105‰ are obtained.

Table 3.7 98 δ Mo(BSE) estimated for various temperatures of core formation

Temperature of core δ98Mo (‰) a formation (°C) (BSE) 1750 +0.020 ± 0.028 2000 −0.016 ± 0.023 2500 −0.058 ± 0.018 3000 −0.081 ± 0.015 3500 −0.095 ± 0.013 4000 −0.104 ± 0.012

a 98 98 δ Mo(BSE) calculated using Equation 3.2, with Earth δ Mo(bulk) = −0.144 ± 0.009‰ and 2σ uncertainty propagated through calculation.

98 While these values are in accord with some previously reported data for δ Mo(BSE) (e.g., Burkhardt et al., 2014), there is currently a great deal of debate about Mo isotope fractionation 98 in terrestrial igneous rocks. Reported δ Mo(BSE) values, re-normalised to NIST SRM 3134, vary by over 0.5‰, from +0.38 to −0.19‰ (e.g., Liang et al., 2013; Greber et al., 2015). Once 98 this debate is resolved, such that a confident estimate of δ Mo(BSE) is possible, this would allow Equation 3.3 to be used to better constrain the temperature of core formation.

3.5 Conclusions

The data presented here indicate that the δ98Mo values of iron meteorites can be influenced by isotope fractionation processes associated with metal–sulphide partitioning. In particular, troilite (FeS) nodules were found to exhibit a wide range of δ98Mo values with both heavy and light Mo stable isotope compositions. A similar but somewhat smaller range of δ98Mo was observed for iron meteorites of the IAB/IIICD complex. These variations are posited to result from the segregation and crystallisation of S-rich melts on the parent body. Such processes are associated with metal–sulphide partitioning of Mo and concomitant Mo isotope fractionation. This model provides an alternative explanation for the observed Mo isotope variations of IAB/IIICD complex irons that were previously ascribed to evaporative loss of isotopically light Mo during energetic impacts (Burkhardt et al., 2014). The magmatic iron meteorite groups display much more consistent δ98Mo values than those exhibited by the troilites and the IAB/IIICD complex irons. However, small but

104 Chapter 3 resolvable and systematic variations of δ98Mo were found within magmatic iron groups, particularly in meteorites towards the Ni-rich end of the fractional crystallisation sequences. These variations are most probably also associated with metal–sulphide partitioning. Finally, a correlation between δ98Mo and the magnitude of nucleosynthetic isotope anomalies in magmatic iron meteorites was observed. With increasing heliocentric distance, as inferred from the magnitude of the nucleosynthetic isotope anomalies, the bulk δ98Mo values of parent bodies shift to lighter Mo isotope compositions. This is a crucial observation, as it provides additional evidence for thermal processing as the cause of planetary-scale nucleosynthetic isotope heterogeneity in the solar system.

Chapter 4

Mass-independent platinum isotope anomalies in iron meteorites

106 Chapter 4

4.1 Introduction

Platinum isotopes potentially provide a fantastic opportunity to further examine thermal processing models of nucleosynthetic isotope anomalies. The six isotopes of Pt were produced by distinct nucleosynthetic processes: the p-, s- and r-processes (Figure 4.1). The s-process contributions to 192Pt, 194Pt, 195Pt and 196Pt were derived from the main s-process in asymptotic giant branch (AGB) stars (e.g., Gallino et al., 1998). Since the atomic number (Z) of Pt is > 56, the r-process contributions to 192Pt, 194Pt, 195Pt, 196Pt and 198Pt originated from main r-process nucleosynthesis, thought to operate in neutron star mergers (e.g., Tsujimoto and Shigeyama, 2014b). The r-process only isotope 198Pt receives no s-process contribution, 197 because the short decay timescale of Pt (t½ < 20 hours) prevents the s-process path from reaching 198Pt. The p-process only isotope 190Pt is thought to be formed by photodisintegration reactions (the γ-process) in Type Ia supernovae, as well as in neutrino- driven winds of Type II supernovae via charged particle reactions (e.g., Lambert, 1992).

Figure 4.1 Nucleosynthetic contributions to the abundances of Pt isotopes. Note the terrestrial abundances of 190Pt and 192Pt are very small (0.01 and 0.78% respectively). Data from Bisterzo et al. (2011) and Creech et al. (2013).

Until recently, very little was known about any mass-independent Pt isotope anomalies in iron meteorites, and to date, no nucleosynthetic isotope anomalies have been confirmed for Pt. An absence of nucleosynthetic Pt isotope anomalies would be in accord with the thermal processing model presented in Chapter 2, where it was argued that decoupling of ‘light vs. heavy’ r-process nucleosynthesis is the primary control. Conversely, it would be in discord with those thermal processing models that favour the ‘moderately vs. highly’ refractory nature

Mass-independent platinum isotope anomalies in iron meteorites 107 as being the critical factor. Under the premise of these models, the lower 50% condensation temperature of Pt, in comparison to Mo (1403 vs. 1587 K), implies the former should be more susceptible to thermal processing, unlike elements with higher 50% condensation temperatures, such Os and Hf, for which no anomalies have been observed (Lodders, 2003; Yokoyama et al., 2007; Sprung et al., 2010). Platinum is, therefore, a promising element to test these hypotheses for the creation of nucleosynthetic isotope anomalies in bulk asteroidal bodies. The few previous studies of mass-independent Pt isotope anomalies in iron meteorites focused on their application as a neutron dosimeter for the correction of cosmogenic isotope effects on the short-lived Hf-W and Pd-Ag decay systems (e.g., Kruijer et al., 2014; Hunt et al., 2015). Exposure to galactic cosmic rays results in the burnout of 195Pt with production of 196Pt, as well as burnout of 191Ir and 193Ir on to 192Ir and 194Ir, and subsequent β− decay to 192Pt and 194Pt, respectively (Leya and Masarik, 2013). Here, the current models of cosmogenic Pt isotope effects in iron meteorites are tested to determine if mass-independent Pt isotope anomalies are entirely cosmogenic in origin. This study involved acquisition of precise Pt isotopic data for 11 groups of iron meteorites. The new results provide constraints on mass-independent isotope anomalies produced by exposure to galactic cosmic rays and planetary-scale mixing processes in the solar nebula, and help inform the debate about the origin of nucleosynthetic isotope anomalies in bulk asteroids.

4.2 Modelling of mass-independent Pt isotope effects

Measurements of mass-independent isotope anomalies can be obtained by internal normalisation to correct for the instrumental mass bias, as previously outlined in Chapter 1. This involves selecting two isotopes and correcting the measured ratio of these to the terrestrial reference value (Equation 1.1). This yields the mass fractionation coefficient β, which can then be used to correct all other isotope ratios for instrumental mass fractionation (Equation 1.2). As the abundances of 190Pt and 192Pt are very small (0.01 and 0.78%, respectively), ratios involving these isotopes cannot be measured at optimal precision and hence these nuclides are best not applied in the normalisation ratio. Therefore, only combinations of 194Pt, 195Pt, 196Pt and 198Pt have been used here for the internal normalisation i of instrumental mass bias. For brevity, the nomenclature ε Pt(y/x) is adopted, where the ratio

108 Chapter 4

i 19x 19y 19x 194 194 195 198 195 Pt/ Pt is normalised to Pt/ Pt (e.g., ε Pt(8/5) = Pt/ Pt normalised to Pt/ Pt and given relative to terrestrial Pt, in parts per ten thousand).

4.2.1 Nucleosynthetic Pt isotope anomalies

Platinum isotope patterns resulting from nucleosynthetic isotope anomalies can be modelled by adding or subtracting nucleosynthetic components from the terrestrial Pt isotope abundances, calculated in the same way as demonstrated in Chapter 2, Section 2.2. All of the isotopes used in the normalising ratios (194Pt, 195Pt, 196Pt and 198Pt) are predominantly r-process (Figure 4.1), and so use of different combinations of these isotopes in normalising ratios produces very similar isotope patterns. This differs from Mo, where the normalisations to different ratios give rise to different isotopic patterns. In detail, this is demonstrated in Figure 4.2 (normalisation to 198Pt/195Pt) and Figure 4.3 (normalisation to 196Pt/194Pt), which display very similar isotope patterns. Clearly apparent for both normalisations are anomalies in ε190Pt for a p-excess/deficit, in ε192Pt for an s-excess/deficit, and in ε190Pt and ε192Pt for an r-excess/deficit, whilst other isotope ratios are essentially unaltered with εiPt ≈ 0. Notably, 190Pt and 192Pt are the p- and s-process nuclides, respectively, whilst the ratios used for normalisation utilise two isotopes that are predominantly of r-process provenance (Figure 4.1). Therefore, normalisation to 198Pt/194Pt and 196Pt/195Pt also yields isotopic patterns that are very similar to those shown below.

Figure 4.2 Effects on Pt isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Pt, using normalisation to 198Pt/195Pt = 0.216277. Calculated using the nucleosynthetic contributions in Bisterzo et al. (2011) and scaled so absolute value of largest anomaly in each panel = 1ε.

Mass-independent platinum isotope anomalies in iron meteorites 109

Figure 4.3 Effects on Pt isotope patterns of p-, s- and r-process excesses/deficits relative to terrestrial Pt, using normalisation to 196Pt/194Pt = 0.770802. Calculated using the nucleosynthetic contributions in Bisterzo et al. (2011) and scaled so absolute value of largest anomaly in each panel is = 1ε.

4.2.2 Cosmogenic Pt isotope anomalies

The effects that galactic cosmic rays (GCR) have on Pt isotopes have been recently modelled by Kruijer et al. (2013) and Leya and Masarik (2013). Exposure to GCR produces an excess in 196Pt relative to 195Pt and 198Pt. This is because 195Pt captures thermal, epithermal and faster neutrons more efficiently than 196Pt and 198Pt, and so the burnout of 195Pt to produce 196Pt, shown by the reaction 195Pt(n,γ)196Pt, dominates the neutron capture effects in ratios between these three isotopes. Additionally, both 191Ir and 193Ir have high neutron capture cross sections and resonance integrals and can efficiently capture neutrons at epithermal energies of ~1 to 10 keV (Leya and Masarik, 2013). When 191Ir captures a neutron it produces 192Ir, whilst 193Ir produces 194Ir following neutron capture. However, 192Ir and 194Ir are unstable and β− decay to 192Pt and 194Pt, respectively. These processes are summarised by the reactions 191Ir(n,γ)192Ir(β−)192Pt and 193Ir(n,γ)194Ir(β−)194Pt (Kruijer et al., 2013) that will generate excesses in 192Pt and 194Pt, whereby the magnitude critically depends on the Ir/Pt ratio. For internal normalisation to 198Pt/195Pt, the neutron capture reactions result in 192 196 192 excesses of ε Pt(8/5) and ε Pt(8/5) (Figure 4.4a). The extent of the ε Pt (8/5) anomaly relative 196 192 196 to ε Pt(8/5) (i.e., the slope in plots of ε Pt(8/5) vs. ε Pt(8/5)) depends on the Ir/Pt ratio. At 192 196 higher Ir/Pt ratios, ε Pt(8/5) increases relative to ε Pt(8/5), and a steeper correlation is 192 obtained. From the modelling of Leya and Masarik (2013), the relationship between ε Pt(8/5) 196 and ε Pt(8/5) is defined by Equation 4.1, and the slope of the line that defines the cosmogenic effects for a particular meteorite, denoted as Ir/Pt(n,γ), by Equation 4.2. The Ir/Pt(n,γ)

110 Chapter 4 cosmogenic trends for meteorites with different Ir/Pt ratios can thus be predicted (Figure 4.4a). Ir ε192Pt = 70 × × ε196Pt Equation 4.1 (8/5) Pt (8/5) Ir Ir/Pt(n,γ) = 70 × Equation 4.2 Pt

Figure 4.4 Modelled cosmogenic effects on Pt isotopes from exposure to GCR. Adapted from Leya and Masarik (2013) and Kruijer et al. (2013). Plotted are cosmogenic isotope trends for Ir/Pt ratios of 1 (red line), 192 196 194 194 0.5 (blue line) and 0.25 (green line), in (a) ε Pt(8/5) vs. ε Pt(8/5) space and (b) ε Pt(8/5) vs. ε Pt(6/5) space.

As 194Pt is far more abundant than 192Pt (the terrestrial abundances are 32.9 and 0.78%, respectively), the excess in 194Pt from the reaction 193Ir(n,γ)194Ir(β−)194Pt is far smaller than the excess in 192Pt from reaction 191Ir(n,γ)192Ir(β−)192Pt. Therefore, ε194Pt is much less sensitive to 194 differences in Ir/Pt ratios. This is demonstrated in Figure 4.4b, where ε Pt(8/5) is plotted 194 195 196 versus ε Pt(6/5). Here, the reaction Pt(n,γ) Pt plays the most dominant role in the slopes generated by GCR exposure, and these vary less with variable Ir/Pt ratios than those shown in Figure 4.4a. Notably, all slopes of Figure 4.4 pass through the origin, which represents the composition of terrestrial Pt, which is assumed to be unaltered by cosmogenic isotope effects.

4.2.3 Resolving nucleosynthetic from cosmogenic Pt isotope effects

Given that previous studies found extensive GCR effects on Pt isotopes in iron meteorites (e.g., Kruijer et al., 2013; Hunt et al., 2014), one would expect any nucleosynthetic signatures, if present, to be at least partially overwritten. Yet the different isotope effects can potentially be resolved by integrating models for nucleosynthetic isotope anomalies with models of cosmogenic isotope effects. Using this approach, mixing lines in εiPt vs. εiPt space between

Mass-independent platinum isotope anomalies in iron meteorites 111 various nucleosynthetic excesses/deficits and terrestrial Pt can be predicted from the nucleosynthetic models generated in Section 4.2.1. These can then be applied to the plots of Pt isotope compositions of iron meteorites (Figure 4.5). On plotting the measured Pt isotope composition of a meteorite, the calculated Ir/Pt(n,γ) trend for that meteorite (see Section 4.2.2) can be drawn so that it passes through this data point. If there are no nucleosynthetic isotope anomalies for Pt, this cosmogenic trendline should pass through the origin. However, if the meteorite has nucleosynthetic Pt isotope anomalies it will instead intercept the nucleosynthetic production trend at a different composition. Figure 4.5(a) demonstrates these systematics for three different meteorites from a single parent body (denoted by the filled red, blue and green stars), which all feature both nucleosynthetic and cosmogenic isotope effects. The open red star denotes where one would expect a meteorite to plot if the composition was only altered by cosmogenic isotope effects. In this case, the cosmogenic trend line (red dashed line) passes through the origin.

Figure 4.5 Integration of models for nucleosynthetic and cosmogenic Pt isotope effects. The isotope compositions of three meteorites, from one group, are plotted with Ir/Pt ratios of 1, 0.5 and 0.25 (red, blue and green stars, respectively). The cosmogenic production trend for the Ir/Pt ratio of each meteorite is plotted through the respective meteorite data. Notably, the cosmogenic trends should all converge and intercept the calculated nucleosynthetic isotope trend at the same point if a nucleosynthetic isotope anomaly is present (black star) – this is the composition of the parent body prior to GCR exposure. The open red star and dashed red line denote where the solid red star and its cosmogenic production trend would plot if its isotope composition was only altered by cosmogenic and not nucleosynthetic isotope effects (see text).

Furthermore, as the cores of magmatic iron meteorite parent bodies underwent complete melting and homogenisation, all meteorites from the same parent body should display essentially identical nucleosynthetic isotope signatures, as was observed for Mo (Chapter 2). Therefore, the predicted Ir/Pt(n,γ) trendlines for meteorites from the same group should all converge and intersect the nucleosynthetic trend at the same point. These systematics are

112 Chapter 4 illustrated for the three hypothetical meteorites in Figure 4.5a. Importantly, the isotopic composition of the intercept (given by the black star) denotes the composition of the parent body before exposure to GCR, and its position relative to the terrestrial value (denoted by εiPt = 0) quantifies the nucleosynthetic isotope anomaly for the group. 194 194 Figure 4.5(b) shows the same schematics but in ε Pt(6/5) vs. ε Pt(8/5) space. Here the slopes of the cosmogenic Ir/Pt(n,γ) trends show much less variability than in Figure 4.5(a). This is expected as any excess of 194Pt from the reaction of 193Ir(n,γ)194Ir(β−)194Pt is far smaller than the excess in 192Pt from the reaction 191Ir(n,γ)192Ir(β−)192Pt. This reflects the much higher natural abundance of 194Pt in comparison to 192Pt. Notably, the slopes of the cosmogenic Ir/Pt(n,γ) trendlines for realistic Ir/Pt ratios of 1, 0.5 and 0.25 all differ significantly from the nucleosynthetic trend that is expected for an excess of s-process nuclides. In summary, model calculations indicate that it should be relatively straightforward to differentiate between mass-independent isotope anomalies that are predominantly of cosmogenic and nucleosynthetic origin.

4.3 Analytical techniques

Aliquots of the iron meteorite sample solutions that were prepared as described in Chapter 1 were employed for the analyses that are detailed below. All materials and reagents used, and the laboratory conditions, were the same as those outlined in Chapter 1.

4.3.1 Ion-exchange chromatography

Separation of Pt was achieved by an ion-exchange chromatography procedure (Table 4.1), adapted from Rehkämper and Halliday (1997) and Creech et al. (2014a). For each sample, 30–100 mg of meteorite (~1000 ng Pt) was loaded onto the columns. Following the full separation and clean-up chemistry, with yields of 60–80%, the purified Pt fractions typically had Os/Pt and Hg/Pt ratios lower than 1×10−5. The isotope 191Ir was also monitored due to its potential to produce tailing effects on 192Pt (e.g., Kruijer et al., 2013). For the samples analysed here, the Ir/Pt ratios in Pt fractions were typically ~0.05. The procedural blank of the full sample preparation procedure was routinely monitored and was generally < 1 ng Pt. Considering that samples with ~500 ng Pt were generally processed, the blank contribution to the total Pt analysed was < 2‰ and thus negligible.

Mass-independent platinum isotope anomalies in iron meteorites 113

Table 4.1 Ion-exchange chemistry for separation of Pt from iron meteorites Quartz column, 3 ml resin reservoir, 50 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (1 ml) Step Resin volumes Acid a Cleaning 1.5 8 M HNO3 1.5 11 M HCl 1 1 M HCl Pre-condition resin 4 1 M HCl Load Sample 2–4 1 M HCl Elute matrix (inc. Fe, Ni) 40 0.5 M HCl Elute Ir, Os, Re, Pd 50 11 M HCl a

Elute and collect Pt 14 13.5 M HNO3 a Step followed by rinse with 1 ml H2O to prevent aqua regia forming in resin during next step

Elution of Pt with such concentrated (13.5 M) HNO3 poses a potential problem because it is known that organic compounds can be released from ion exchange resin at such conditions, and these organic contaminants can have a detrimental impact of the quality of isotope analyses by MC-ICP-MS (e.g., Gault-Ringold and Stirling, 2012). To mitigate any such effects, the method of Murphy et al. (2016) was followed, which employs liquid-liquid extraction with n-heptane, to remove any such organic compounds from the eluate prior to MC-ICP-MS analyses. Finally, the Pt fractions were dried down with a mixture of

15 M HNO3 and 12 M HClO4 to volatilise any remaining Os (as OsO4), before being dissolved in 0.5 M HCl ready for MC-ICP-MS analysis.

4.3.2 Mass spectrometry

4.3.2.1 Instrumentation and data collection protocol

The isotope measurements were performed using a Nu Instruments Nu Plasma HR MC-ICP-MS at Imperial College London whilst samples were delivered via a Nu Instruments DSN-100 desolvating nebuliser at an uptake rate of ~140 µl/min. Typical sensitivity for Pt was 180–210 V/ppm, resulting in a 196Pt ion beam of 6–7 × 10−11 A (6–7 V using 1011 Ohm resistors) for solutions with ~100 ppb Pt. The data were acquired in a single-sequence measurement routine with a simultaneous collection of the 192Pt, 194Pt, 195Pt, 196Pt, 198Pt, as well as 188Os, 200Hg and 191Ir ion beams. Due to the very low abundance of 190Pt (0.01%), adequate measurements of this isotope are unattainable with this set-up. The Faraday cup configuration is indicated in Table 4.2. All analyses utilised 3 blocks with 20 integrations of 5 seconds each. Each block was preceded by

114 Chapter 4 a 30 second on-peak baseline measurement while the ion beam was deflected by the electrostatic analyser.

Table 4.2 Faraday cup configuration for measurement of Pt isotopes

Faraday Cup L5 L4 L3 L2 L1 Ax H1 H2 H3 H4 H5 H6 amu 188 190 191 192 194 195 196 198 200

4.3.2.2 Interference and mass bias corrections

Possible spectral interferences on Pt are shown in Table 4.3. Isobaric interferences from Os were corrected using 188Os as the interference monitor and the Os isotopic abundances of Berglund and Wieser (2011). Interferences from Hg isobars were corrected using 200Hg as interference monitor and the Hg isotopic abundances from Berglund and Wieser (2011).

Table 4.3 Important interferences on platinum isotopes

Interference 192Pt 194Pt 195Pt 196Pt 198Pt

Isobaric 192Os 196Hg 198Hg

Argides 152Sm40Ar 154Sm40Ar 155Gd40Ar 156Gd40Ar 158Gd40Ar 152Gd40Ar 154Gd40Ar 156Dy40Ar 158Dy40Ar

Oxides 176Hf16O 178Hf16O 179Hf16O 180Hf16O 182W16O 176Lu16O 180Ta16O 176Yb16O 180W16O

The instrumental mass bias was corrected by normalisation to 198Pt/195Pt = 0.216277, 196Pt/194Pt = 0.770802, 198Pt/194Pt = 0.222737 and 196Pt/195Pt = 0.748446, using the exponential law (Young et al., 2002; Wombacher and Rehkämper, 2003; Creech et al., 2013). Since these ratios have isobaric interferences (Hg on 196Pt and 198Pt), an iterative procedure was used to subtract these effects from the normalising ratios. Data are reported in εiPt notation (Equation 4.3), calculated relative to the mean of several bracketing runs of an IRMM-010 Pt SRM solution, made up to closely match the Pt concentration of the samples (~100 ppb).

i x Pt Pt εiPt = sample − 1 × 104 Equation 4.3 iPt xPt standard where xPt is the same isotope used in the denominator of the normalising ratio.

Mass-independent platinum isotope anomalies in iron meteorites 115

4.4 Results

4.4.1 Standard solutions and reference materials

The external reproducibility (2σ) and internal precision (2se) of the bracketing runs of the IRMM-010 Pt standard reference material (typically n = 4), for all normalisations, are shown in Table 4.4 and Table 4.5. It is notable that the different normalisations yield similar precisions – this reflects the fact that 194Pt, 195Pt and 196Pt have similar abundances (of 32.9, 33.8 and 25.2%, respectively) and that the different normalisation ratios have only a limited range of mass differences (1–3 amu difference). For all εiPt values, except ε192Pt, the external reproducibility is very similar to the internal precision at about 0.1 to 0.2 ε, demonstrating the excellent precision of the procedures. However, the external reproducibility for ε192Pt is about a factor of 2 worse than the internal precision. A similar observation was made by Kruijer et al. (2013) and attributed to the low abundance of 192Pt (0.78%), which renders measurements more susceptible to matrix effects than the other isotopes, which are far more abundant.

Table 4.4 External reproducibility (2σ) of Pt isotope measurements for n = 4 bracketing IRMM-010 Pt SRM runs

Normalising ratio ε192Pt ε194Pt ε195Pt ε196Pt ε198Pt 198Pt/195Pt ±2.25 ±0.15 - ±0.12 - 196Pt/194Pt ±2.22 - ±0.09 - ±0.28 198Pt/194Pt ±2.22 ±0.11 ±0.14 - 196Pt/195Pt ±2.26 ±0.19 - - ±0.35

Table 4.5 Typical internal precision (2se) of Pt isotope measurements for single IRMM-010 Pt SRM runs

Normalising ratio ε192Pt ε194Pt ε195Pt ε196Pt ε198Pt 198Pt/195Pt ±1.13 ±0.12 - ±0.08 - 196Pt/194Pt ±1.13 - ±0.08 - ±0.19 198Pt/194Pt ±1.12 - ±0.09 ±0.15 196Pt/195Pt ±1.17 ±0.15 - - ±0.25

The robustness and reproducibility of the Pt separation procedure and the MC-ICP-MS Pt isotope measurements were routinely evaluated by repeated analyses of terrestrial standard reference materials. NIST SRM 129c (High-Sulphur Steel) was chosen for this purpose due to a major element composition that is somewhat similar to iron meteorites, with a high S content of about 2%. For analysis, this SRM was doped with IRMM-010 Pt to approximate the Pt concentration in the samples prior to ion-exchange chromatography. A second in-house

116 Chapter 4

Pt standard solution, denoted Pt-α, was analysed versus IRMM-010 Pt to monitor data quality. All analyses of these SRMs routinely yielded terrestrial Pt isotope compositions for all normalisation schemes (Table 4.6).

Table 4.6 Platinum isotope data for terrestrial standard reference materials Normalised to 198Pt/195Pt Reference Material N a ε192Pt ε194Pt ε196Pt NIST SRM 129c b 35 0.48 ± 2.50 0.14 ± 0.26 0.10 ± 0.14 Pt-α 9 –0.29 ± 2.80 –0.04 ± 0.09 0.05 ± 0.19

Normalised to 196Pt/194Pt Reference Material N a ε192Pt ε195Pt ε198Pt NIST SRM 129c b 35 0.11 ± 1.76 –0.13 ± 0.20 –0.03 ± 0.22 Pt-α 9 –0.11 ± 2.56 –0.01 ± 0.10 –0.17 ± 0.40

Normalised to 198Pt/194Pt Reference Material N a ε192Pt ε194Pt ε196Pt NIST SRM 129c b 35 0.06 ± 2.04 –0.10 ± 0.19 0.02 ± 0.11 Pt-α 9 –0.16 ± 3.01 0.03 ± 0.07 0.08 ± 0.20

Normalised to 196Pt/195Pt Reference Material Na ε192Pt ε195Pt ε198Pt NIST SRM 129c b 35 0.79 ± 2.47 0.26 ± 0.40 –0.31 ± 0.42 Pt-α 9 –0.19 ± 2.53 0.03 ± 0.21 –0.15 ± 0.56 a Number of times the SRM was analysed; b SRM doped with IRMM-010 Pt; All uncertainties are 2σ.

The typically encountered Os/Pt and Hg/Pt ratios of less than 1 × 10−5 for purified Pt fractions were associated with maximum corrections of 350 ppm on ε192Pt and 40 ppm on ε198Pt. Molecular interferences (Table 4.3) were also monitored but these were sufficiently small to be completely negligible. Particular attention was paid to monitoring nebuliser uptake rates, since the presence of any organic compounds released from the ion exchange resin in the high molarity HNO3 used for the final Pt elution could cause flow-rate fluctuations. However, no such problems were encountered, and the chemistry, particularly the n-heptane step, was deemed sufficiently effective at removing any such compounds. During the course of this study, the abundance sensitivity of the Nu Plasma HR MC-ICP-MS, measured on 237U versus 238U, was ~1 ppm. For a typical Ir/Pt ratio of ~0.05, this implies that the ε192Pt values require a correction of less than 1 ε. No such correction was applied in this study, as the typically encountered ε192Pt measurement uncertainty was already significantly larger at about ±2 to ±2.5 ε-units.

Mass-independent platinum isotope anomalies in iron meteorites 117

4.4.2 Iron meteorites

The Pt isotope data for iron meteorites, normalised to 198Pt/195Pt, are shown in Table 4.7. A large variation in ε192Pt is observed, with values ranging from +55.15 ± 1.12 for Tlacotopec 1 (IVB) to terrestrial compositions for various meteorites. Similarly, there are clear excesses in ε194Pt and ε196Pt, although these anomalies are generally about an order of magnitude smaller than the variations in ε192Pt. Crucially, there are also significant variations within some iron meteorite groups, which is evident from the individual meteorite results and the very large uncertainties for some groups reported in Table 4.7 (e.g., IVB group mean ε192Pt = +31.52 ± 24.13). As an example, the variations within the IIIAB group of irons are displayed in Figure 4.6. In detail, Bear Creek, Lenarto, Verkhne Udinsk and Williamette have compositions that are within uncertainty of terrestrial Pt, whilst Cape York, Charcas 1 and 2, Henbury 1 and 2, and Santa Apolonia 1 display excesses, to varying degrees, in ε192Pt, ε194Pt and ε196Pt.

Figure 4.6 Platinum isotope data for IIIAB iron meteorites, normalised to 198Pt/195Pt = 0.216277. Measured relative to IRMM-010 Pt SRM, which represents the terrestrial Pt isotope composition, and normalised to 198Pt/195Pt = 0.216277. Uncertainties are 2se = 2σ/√n. Grey-shaded area represents reproducibility (2σ) of the bracketing IRMM-010 Pt SRM runs.

Similar patterns with clearly resolvable variations are observed when the Pt isotope data are normalised to 196Pt/194Pt (Table 4.8), 198Pt/194Pt (Table 4.9) and 196Pt/195Pt (Table 4.10) as the different normalisations just reflect different ways of accounting for instrumental mass bias.

118 Chapter 4

Table 4.7 Platinum isotope compositions of iron meteorites, normalised to 198Pt/195Pt

Group Sample N a Pt (µg/g) b Ir/Pt c ε 192Pt d ε194Pt d ε196Pt d IAB Campo del Cielo 2 8 8.3 0.39 0.74 ± 0.74 0.52 ± 0.30 0.21 ± 0.16 Campo del Cielo 3 4 8.8 0.36 5.45 ± 1.06 0.82 ± 0.10 0.43 ± 0.08 Canyon Diablo 2 3 5.1 0.37 7.01 ± 2.20 0.87 ± 0.07 0.61 ± 0.10 IAB MEAN 5.05 ± 5.54 0.77 ± 0.29 0.46 ± 0.40

IC Arispe 3 4 18.7 0.59 11.80 ± 1.36 0.85 ± 0.13 0.55 ± 0.08 Arispe 4 12 18.9 0.59 12.96 ± 0.45 0.97 ± 0.04 0.61 ± 0.04 Bendego 1 3 11.8 0.02 –0.19 ± 0.88 0.85 ± 0.14 0.61 ± 0.16 Bendego 2 4 11.5 0.02 2.45 ± 2.19 1.20 ± 0.19 0.96 ± 0.02 Santa Rosa 5 5.1 0.01 0.99 ± 0.53 0.00 ± 0.10 0.31 ± 0.08 IC MEAN 4.83 ± 7.54 0.65 ± 0.65 0.56 ± 0.28

IIAB Coahuila 9 31.5 0.46 1.20 ± 0.93 0.00 ± 0.05 –0.04 ± 0.05 Murphy 6 36.6 0.93 –0.39 ± 0.99 –0.21 ± 0.07 –0.19 ± 0.05 North Chile 6 24.3 0.15 1.56 ± 0.75 0.50 ± 0.08 0.31 ± 0.07 Sikhote Alin 1 6 4.3 0.01 –0.73 ± 0.85 –0.37 ± 0.05 –0.14 ± 0.04 Sikhote Alin 2 9 5.2 0.01 1.09 ± 0.88 0.55 ± 0.06 0.35 ± 0.09 IIAB MEAN 0.64 ± 0.90 0.09 ± 0.30 0.05 ± 0.21

IIC Ballinoo 4 14.4 0.62 –2.10 ± 0.98 –0.39 ± 0.08 –0.15 ± 0.09 Kumerina 12 9.3 0.87 –0.17 ± 1.10 0.02 ± 0.03 0.06 ± 0.04 Salt River 4 13.3 0.47 3.36 ± 2.12 1.52 ± 0.05 0.77 ± 0.09 IIC MEAN 0.36 ± 5.54 0.39 ± 2.02 0.22 ± 0.97

IIE Kodaikanal 5 11.1 0.47 1.13 ± 1.99 0.87 ± 0.07 0.39 ± 0.04 Verkhne Dnieprovsk 7 11.8 0.52 –0.40 ± 0.71 –0.11 ± 0.04 –0.06 ± 0.05 Weekeroo Station 7 13.1 0.29 3.96 ± 0.90 0.74 ± 0.09 0.44 ± 0.04 IIE MEAN 1.56 ± 2.55 0.50 ± 0.62 0.26 ± 0.31

IIIAB Bear Creek 4 1.7 0.01 0.42 ± 0.75 –0.32 ± 0.10 –0.17 ± 0.13 Cape York 7 14.3 0.22 0.91 ± 1.16 1.01 ± 0.09 0.47 ± 0.03 Charcas 1 8 12.8 0.15 5.18 ± 0.64 0.49 ± 0.08 0.51 ± 0.06 Charcas 2 7 12.4 0.15 5.92 ± 0.75 0.37 ± 0.04 0.41 ± 0.05 Henbury 1 4 16.1 0.86 14.40 ± 1.80 1.24 ± 0.05 0.62 ± 0.09 Henbury 2 5 16.5 0.83 12.48 ± 1.03 0.96 ± 0.14 0.53 ± 0.05 Lenarto 6 6.5 0.05 0.11 ± 1.25 0.03 ± 0.10 0.05 ± 0.05 Santa Apolonia 1 4 16.5 0.50 4.73 ± 0.72 0.42 ± 0.09 0.30 ± 0.09 Verkhne Udinsk 5 13.1 0.25 0.97 ± 1.71 0.13 ± 0.03 0.01 ± 0.06 Williamette 7 7.1 0.66 –0.12 ± 0.91 –0.24 ± 0.08 –0.17 ± 0.04 IIIAB MEAN 3.25 ± 3.29 0.32 ± 0.37 0.19 ± 0.21

IIICD Nantan 1 4 6.7 0.25 0.54 ± 1.40 0.43 ± 0.12 0.22 ± 0.09

IIIE Staunton 4 6.4 0.02 1.39 ± 0.46 0.77 ± 0.20 0.38 ± 0.03

IIIF Clark County 5 9.7 0.64 3.17 ± 0.61 0.23 ± 0.03 0.08 ± 0.06

IVA Gibeon 2 8 6.0 0.45 –0.94 ± 1.01 –0.27 ± 0.09 0.04 ± 0.03 Obernkirchen 4 6.8 0.47 0.94 ± 0.74 0.36 ± 0.09 0.15 ± 0.08 IVA MEAN 0.00 ± 1.87 0.05 ± 0.63 0.10 ± 0.11

IVB Cape of Good Hope 4 30.1 1.05 23.94 ± 1.49 1.18 ± 0.08 0.56 ± 0.08 Santa Clara 6 23.1 0.78 15.48 ± 0.79 1.26 ± 0.19 0.71 ± 0.07 Tlacotopec 1 4 18.6 1.29 55.15 ± 1.12 2.11 ± 0.11 0.89 ± 0.06 IVB MEAN 31.52 ± 24.13 1.52 ± 0.60 0.72 ± 0.19

a Number of times sample analysed; b Pt concentrations from Chapter 5, determined by isotope dilution with typical uncertainties of ~0.1%; c Ir/Pt ratios calculated using Ir data from Catalogue of Meteorites (2000) and references therein. d 198 195 i i 195 i 195 4 Normalised to Pt/ Pt = 0.216277 using the exponential law, ε Pt = [( Pt/ Pt)smp/( Pt/ Pt)std – 1] × 10 ; Uncertainties are 2se = 2σ/√n, where for samples n is the number of sample analyses, and for group means n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of the specimens is taken to represent that sample when calculating the group mean).

Mass-independent platinum isotope anomalies in iron meteorites 119

Table 4.8 Platinum isotope compositions of iron meteorites, normalised to 196Pt/194Pt

Group Sample N a Pt (µg/g) b Ir/Pt c ε 192Pt d ε195Pt d ε198Pt d IAB Campo del Cielo 2 8 8.3 0.39 –0.19 ± 0.67 –0.35 ± 0.21 0.09 ± 0.07 Campo del Cielo 3 4 8.8 0.36 4.22 ± 1.15 –0.64 ± 0.03 –0.10 ± 0.21 Canyon Diablo 2 3 5.1 0.37 5.76 ± 2.12 –0.71 ± 0.08 –0.33 ± 0.16 IAB MEAN 3.89 ± 5.30 –0.60 ± 0.31 –0.17 ± 0.46

IC Arispe 3 4 18.7 0.59 10.63 ± 1.60 –0.72 ± 0.06 –0.22 ± 0.26 Arispe 4 12 18.9 0.59 11.93 ± 0.76 –0.78 ± 0.02 –0.27 ± 0.07 Bendego 1 3 11.8 0.02 –1.49 ± 1.73 –0.73 ± 0.02 –0.17 ± 0.43 Bendego 2 4 11.5 0.02 0.95 ± 2.11 –1.11 ± 0.08 –0.73 ± 0.12 Santa Rosa 5 5.1 0.01 1.32 ± 0.65 –0.12 ± 0.05 –0.58 ± 0.12 IC MEAN 4.11 ± 7.23 –0.60 ± 0.48 –0.42 ± 0.19

IIAB Coahuila 9 31.5 0.46 1.18 ± 1.01 0.00 ± 0.04 –0.04 ± 0.05 Murphy 6 36.6 0.93 –0.05 ± 0.99 0.21 ± 0.04 0.20 ± 0.11 North Chile 6 24.3 0.15 0.75 ± 0.55 –0.39 ± 0.02 –0.10 ± 0.21 Sikhote Alin 1 6 4.3 0.01 –0.16 ± 0.70 0.27 ± 0.02 –0.06 ± 0.07 Sikhote Alin 2 9 5.2 0.01 0.32 ± 0.76 –0.45 ± 0.07 –0.14 ± 0.16 IIAB MEAN 0.49 ± 0.58 –0.07 ± 0.25 –0.01 ± 0.14

IIC Ballinoo 4 14.4 0.62 –1.65 ± 1.09 0.26 ± 0.03 –0.03 ± 0.20 Kumerina 12 9.3 0.87 –0.69 ± 0.90 –0.06 ± 0.04 –0.01 ± 0.09 Salt River 4 13.3 0.47 0.79 ± 2.08 –1.15 ± 0.06 –0.01 ± 0.20 IIC MEAN –0.52 ± 2.47 –0.32 ± 1.48 –0.01 ± 0.03

IIE Kodaikanal 5 11.1 0.47 –0.31 ± 1.79 –0.64 ± 0.06 0.00 ± 0.11 Verkhne Dnieprovsk 7 11.8 0.52 –0.25 ± 0.73 0.09 ± 0.04 0.03 ± 0.09 Weekeroo Station 7 13.1 0.29 2.84 ± 0.87 –0.58 ± 0.03 –0.15 ± 0.14 IIE MEAN 0.76 ± 2.08 –0.38 ± 0.47 –0.04 ± 0.11

IIIAB Bear Creek 4 1.7 0.01 0.72 ± 0.78 0.21 ± 0.07 –0.01 ± 0.41 Cape York 7 14.3 0.22 –0.78 ± 1.00 –0.73 ± 0.02 0.04 ± 0.13 Charcas 1 8 12.8 0.15 4.63 ± 0.75 –0.52 ± 0.01 –0.49 ± 0.15 Charcas 2 7 12.4 0.15 5.61 ± 0.57 –0.39 ± 0.03 –0.48 ± 0.14 Henbury 1 4 16.1 0.86 13.05 ± 1.92 –0.92 ± 0.07 –0.09 ± 0.12 Henbury 2 5 16.5 0.83 11.09 ± 0.92 –0.76 ± 0.08 –0.06 ± 0.16 Lenarto 6 6.5 0.05 0.06 ± 1.16 –0.04 ± 0.04 0.03 ± 0.12 Santa Apolonia 1 4 16.5 0.50 4.56 ± 0.93 –0.33 ± 0.08 –0.18 ± 0.09 Verkhne Udinsk 5 13.1 0.25 0.79 ± 1.67 –0.05 ± 0.04 0.11 ± 0.12 Williamette 7 7.1 0.66 0.20 ± 0.98 0.21 ± 0.04 0.17 ± 0.11 IIIAB MEAN 2.84 ± 3.05 –0.25 ± 0.28 –0.05 ± 0.14

IIICD Nantan 1 4 6.7 0.25 0.10 ± 1.37 –0.33 ± 0.07 0.01 ± 0.29

IIIE Staunton 4 6.4 0.02 0.41 ± 0.79 –0.57 ± 0.08 –0.05 ± 0.20

IIIF Clark County 5 9.7 0.64 2.84 ± 0.52 –0.12 ± 0.02 0.07 ± 0.18

IVA Gibeon 2 8 6.0 0.45 –0.33 ± 1.07 0.13 ± 0.04 –0.40 ± 0.12 Obernkirchen 4 6.8 0.47 0.62 ± 0.86 –0.25 ± 0.05 0.09 ± 0.28 IVA MEAN 0.15 ± 0.95 –0.06 ± 0.38 –0.16 ± 0.49

IVB Cape of Good Hope 4 30.1 1.05 22.29 ± 1.12 –0.83 ± 0.04 0.00 ± 0.14 Santa Clara 6 23.1 0.78 13.84 ± 0.63 –0.99 ± 0.11 –0.11 ± 0.12 Tlacotopec 1 4 18.6 1.29 51.85 ± 0.90 –1.52 ± 0.08 0.24 ± 0.11 IVB MEAN 29.33 ± 23.04 –1.11 ± 0.42 0.04 ± 0.21 a Number of times sample analysed; b Pt concentrations from Chapter 5, determined by isotope dilution with typical uncertainties of ~0.1%; c Ir/Pt ratios calculated using Ir data from Catalogue of Meteorites (2000) and references therein. d 196 194 i i 194 i 194 4 Normalised to Pt/ Pt = 0.770802 using the exponential law, ε Pt = [( Pt/ Pt)smp/( Pt/ Pt)std – 1] × 10 ; Uncertainties are 2se = 2σ/√n, where for samples n is the number of sample analyses, and for group means n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of the specimens is taken to represent that sample when calculating the group mean).

120 Chapter 4

Table 4.9 Platinum isotope compositions of iron meteorites, normalised to 198Pt/194Pt

Group Sample N a Pt (µg/g) b Ir/Pt c ε 192Pt d ε195Pt d ε196Pt d IAB Campo del Cielo 2 8 8.3 0.39 –0.10 ± 0.64 –0. 39 ± 0.22 –0.04 ± 0.03 Campo del Cielo 3 4 8.8 0.36 4.18 ± 1.20 –0.61 ± 0.07 0.05 ± 0.11 Canyon Diablo 2 3 5.1 0.37 5.70 ± 2.09 –0.65 ± 0.05 0.17 ± 0.08 IAB MEAN 3.87 ± 5.17 –0.58 ± 0.22 0.08 ± 0.23

IC Arispe 3 4 18.7 0.59 10.43 ± 1.52 –0.64 ± 0.10 0.11 ± 0.13 Arispe 4 12 18.9 0.59 11.85 ± 0.72 –0.72 ± 0.03 0.13 ± 0.04 Bendego 1 3 11.8 0.02 –1.60 ± 1.40 –0.64 ± 0.10 0.09 ± 0.22 Bendego 2 4 11.5 0.02 0.59 ± 2.23 –0.90 ± 0.14 0.37 ± 0.06 Santa Rosa 5 5.1 0.01 0.95 ± 0.59 0.00 ± 0.08 0.29 ± 0.06 IC MEAN 3.86 ± 7.32 –0.48 ± 0.48 0.21 ± 0.10

IIAB Coahuila 9 31.5 0.46 1.24 ± 0.97 0.00 ± 0.04 0.02 ± 0.03 Murphy 6 36.6 0.93 –0.09 ± 0.99 0.16 ± 0.05 –0.10 ± 0.06 North Chile 6 24.3 0.15 0.67 ± 0.75 –0.40 ± 0.03 0.04 ± 0.09 Sikhote Alin 1 6 4.3 0.01 –0.19 ± 0.78 0.27 ± 0.04 0.03 ± 0.04 Sikhote Alin 2 9 5.2 0.01 0.24 ± 0.75 –0.41 ± 0.05 0.07 ± 0.08 IIAB MEAN 0.46 ± 0.62 –0.08 ± 0.23 0.00 ± 0.07

IIC Ballinoo 4 14.4 0.62 –1.51 ± 1.01 0.29 ± 0.06 0.02 ± 0.10 Kumerina 12 9.3 0.87 –0.70 ± 0.90 –0.02 ± 0.03 0.00 ± 0.04 Salt River 4 13.3 0.47 0.86 ± 2.01 –1.14 ± 0.04 0.00 ± 0.10 IIC MEAN –0.45 ± 2.41 –0.29 ± 1.51 0.01 ± 0.01

IIE Kodaikanal 5 11.1 0.47 –0.31 ± 1.94 –0.65 ± 0.05 0.00 ± 0.06 Verkhne Dnieprovsk 7 11.8 0.52 –0.27 ± 0.72 0.08 ± 0.03 –0.02 ± 0.04 Weekeroo Station 7 13.1 0.29 2.97 ± 0.93 –0.56 ± 0.07 0.07 ± 0.07 IIE MEAN 0.80 ± 2.17 –0.38 ± 0.46 0.02 ± 0.06

IIIAB Bear Creek 4 1.7 0.01 0.83 ± 0.69 0.24 ± 0.08 0.01 ± 0.20 Cape York 7 14.3 0.22 –0.66 ± 1.04 –0.75 ± 0.07 –0.02 ± 0.07 Charcas 1 8 12.8 0.15 4.37 ± 0.72 –0.36 ± 0.06 0.24 ± 0.08 Charcas 2 7 12.4 0.15 5.46 ± 0.61 –0.27 ± 0.03 0.24 ± 0.07 Henbury 1 4 16.1 0.86 12.83 ± 1.85 –0.93 ± 0.04 0.05 ± 0.06 Henbury 2 5 16.5 0.83 11.08 ± 1.06 –0.72 ± 0.09 0.03 ± 0.08 Lenarto 6 6.5 0.05 0.09 ± 1.20 –0.02 ± 0.07 –0.02 ± 0.06 Santa Apolonia 1 4 16.5 0.50 4.40 ± 0.96 –0.31 ± 0.07 0.09 ± 0.05 Verkhne Udinsk 5 13.1 0.25 0.94 ± 1.70 –0.10 ± 0.02 –0.05 ± 0.06 Williamette 7 7.1 0.66 0.30 ± 0.94 0.18 ± 0.06 –0.08 ± 0.05 IIIAB MEAN 2.85 ± 2.97 –0.24 ± 0.28 0.03 ± 0.07

IIICD Nantan 1 4 6.7 0.25 0.00 ± 1.42 –0.32 ± 0.09 0.00 ± 0.15

IIIE Staunton 4 6.4 0.02 0.31 ± 0.51 –0.57 ± 0.15 0.03 ± 0.10

IIIF Clark County 5 9.7 0.64 2.91 ± 0.52 –0.17 ± 0.02 –0.03 ± 0.09

IVA Gibeon 2 8 6.0 0.45 –0.46 ± 1.03 0.20 ± 0.07 0.20 ± 0.06 Obernkirchen 4 6.8 0.47 0.68 ± 0.77 –0.27 ± 0.07 –0.04 ± 0.14 IVA MEAN 0.11 ± 1.14 –0.04 ± 0.47 0.08 ± 0.25

IVB Cape of Good Hope 4 30.1 1.05 22.29 ± 1.04 –0.88 ± 0.06 0.00 ± 0.07 Santa Clara 6 23.1 0.78 13.66 ± 0.60 –0.94 ± 0.14 0.06 ± 0.06 Tlacotopec 1 4 18.6 1.29 51.94 ± 0.87 –1.58 ± 0.08 –0.12 ± 0.05 IVB MEAN 29.30 ± 23.18 –1.14 ± 0.45 –0.02 ± 0.10

a Number of times sample analysed; b Pt concentrations from Chapter 5, determined by isotope dilution with typical uncertainties of ~0.1%; c Ir/Pt ratios calculated using Ir data from Catalogue of Meteorites (2000) and references therein. d 198 194 i i 194 i 194 4 Normalised to Pt/ Pt = 0.222737 using the exponential law, ε Pt = [( Pt/ Pt)smp/( Pt/ Pt)std – 1] × 10 ; Uncertainties are 2se = 2σ/√n, where for samples n is the number of sample analyses, and for group means n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of the specimens is taken to represent that sample when calculating the group mean).

Mass-independent platinum isotope anomalies in iron meteorites 121

Table 4.10 Platinum isotope compositions of iron meteorites, normalised to 196Pt/195Pt

Group Sample N a Pt (µg/g) b Ir/Pt c ε 192Pt d ε194Pt d ε198Pt d IAB Campo del Cielo 2 8 8.3 0.39 1.27 ± 1.14 0.70 ± 0.43 –0.63 ± 0.49 Campo del Cielo 3 4 8.8 0.36 6.56 ± 0.92 1.28 ± 0.07 –1.29 ± 0.23 Canyon Diablo 2 3 5.1 0.37 8.81 ± 2.49 1.43 ± 0.15 –1.81 ± 0.31 IAB MEAN 6.36 ± 6.92 1.21 ± 0.62 –1.39 ± 1.21

IC Arispe 3 4 18.7 0.59 13.49 ± 1.11 1.44 ± 0.11 –1.64 ± 0.25 Arispe 4 12 18.9 0.59 15.15 ± 0.69 1.56 ± 0.05 –1.83 ± 0.11 Bendego 1 3 11.8 0.02 1.57 ± 2.25 1.46 ± 0.05 –1.83 ± 0.48 Bendego 2 4 11.5 0.02 5.28 ± 2.20 2.23 ± 0.17 –2.87 ± 0.07 Santa Rosa 5 5.1 0.01 1.70 ± 0.34 0.25 ± 0.10 –0.91 ± 0.25 IC MEAN 6.48 ± 7.90 1.20 ± 0.97 –1.66 ± 0.83

IIAB Coahuila 9 31.5 0.46 1.22 ± 1.01 0.00 ± 0.08 0.11 ± 0.14 Murphy 6 36.6 0.93 –0.83 ± 0.94 –0.42 ± 0.08 0.57 ± 0.15 North Chile 6 24.3 0.15 1.61 ± 1.45 0.78 ± 0.04 –0.91 ± 0.18 Sikhote Alin 1 6 4.3 0.01 –1.11 ± 0.69 –0.54 ± 0.05 0.42 ± 0.11 Sikhote Alin 2 9 5.2 0.01 2.12 ± 0.86 0.90 ± 0.13 –1.06 ± 0.26 IIAB MEAN 0.63 ± 1.07 0.14 ± 0.50 –0.14 ± 0.63

IIC Ballinoo 4 14.4 0.62 –2.54 ± 1.01 –0.52 ± 0.06 0.46 ± 0.27 Kumerina 12 9.3 0.87 –0.03 ± 1.16 0.11 ± 0.08 –0.17 ± 0.12 Salt River 4 13.3 0.47 5.63 ± 2.03 2.31 ± 0.12 –2.30 ± 0.26 IIC MEAN 1.02 ± 8.36 0.64 ± 2.98 –0.67 ± 2.90

IIE Kodaikanal 5 11.1 0.47 2.31 ± 1.81 1.29 ± 0.11 –1.15 ± 0.13 Verkhne Dnieprovsk 7 11.8 0.52 –0.64 ± 0.52 –0.17 ± 0.08 0.17 ± 0.14 Weekeroo Station 7 13.1 0.29 5.37 ± 0.85 1.17 ± 0.07 –1.31 ± 0.12 IIE MEAN 2.35 ± 3.47 0.76 ± 0.94 –0.76 ± 0.94

IIIAB Bear Creek 4 1.7 0.01 –0.14 ± 0.99 –0.42 ± 0.13 0.51 ± 0.38 Cape York 7 14.3 0.22 2.29 ± 1.07 1.46 ± 0.05 –1.39 ± 0.09 Charcas 1 8 12.8 0.15 6.63 ± 0.65 1.00 ± 0.06 –1.51 ± 0.18 Charcas 2 7 12.4 0.15 7.30 ± 0.60 0.79 ± 0.06 –1.23 ± 0.14 Henbury 1 4 16.1 0.86 16.50 ± 1.82 1.85 ± 0.15 –1.85 ± 0.28 Henbury 2 5 16.5 0.83 13.95 ± 0.98 1.53 ± 0.17 –1.57 ± 0.16 Lenarto 6 6.5 0.05 0.18 ± 1.08 0.08 ± 0.07 –0.15 ± 0.15 Santa Apolonia 1 4 16.5 0.50 5.81 ± 0.73 0.66 ± 0.16 –0.89 ± 0.27 Verkhne Udinsk 5 13.1 0.25 0.82 ± 1.78 0.03 ± 0.04 0.05 ± 0.09 Williamette 7 7.1 0.66 –0.58 ± 0.80 –0.41 ± 0.08 0.51 ± 0.13 IIIAB MEAN 3.82 ± 3.81 0.50 ± 0.57 –0.56 ± 0.63

IIICD Nantan 1 4 6.7 0.25 1.03 ± 1.49 0.66 ± 0.13 –0.67 ± 0.27

IIIE Staunton 4 6.4 0.02 2.61 ± 0.65 1.14 ± 0.17 –1.15 ± 0.08

IIIF Clark County 5 9.7 0.64 3.33 ± 0.55 0.25 ± 0.04 –0.24 ± 0.18

IVA Gibeon 2 8 6.0 0.45 –0.89 ± 1.15 –0.25 ± 0.08 –0.13 ± 0.10 Obernkirchen 4 6.8 0.47 1.30 ± 0.75 0.50 ± 0.11 –0.46 ± 0.23 IVA MEAN 0.20 ± 2.19 0.12 ± 0.76 –0.29 ± 0.33

IVB Cape of Good Hope 4 30.1 1.05 25.62 ± 1.60 1.67 ± 0.08 –1.67 ± 0.25 Santa Clara 6 23.1 0.78 17.58 ± 0.89 1.99 ± 0.22 –2.12 ± 0.20 Tlacotopec 1 4 18.6 1.29 57.75 ± 1.48 3.05 ± 0.15 –2.66 ± 0.17 IVB MEAN 33.65 ± 24.55 2.24 ± 0.83 –2.15 ± 0.58 a Number of times sample analysed; b Pt concentrations from Chapter 5, determined by isotope dilution with typical uncertainties of ~0.1%; c Ir/Pt ratios calculated using Ir data from Catalogue of Meteorites (2000) and references therein. d 196 195 i i 195 i 195 4 Normalised to Pt/ Pt = 0.748446 using the exponential law, ε Pt = [( Pt/ Pt)smp/( Pt/ Pt)std – 1] × 10 ; Uncertainties are 2se = 2σ/√n, where for samples n is the number of sample analyses, and for group means n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of the specimens is taken to represent that sample when calculating the group mean).

122 Chapter 4

4.5 Discussion

Unlike the Mo anomalies observed in Chapter 2, the mass-independent Pt isotope anomalies for meteorites within the same group are typically not identical or even similar (e.g., Figure 4.6). Given that the cores of the magmatic iron meteorite parent bodies were presumably completely homogenised whilst in a liquid state, the observed isotopic anomalies are unlikely to be, at least not entirely, of nucleosynthetic origin. Therefore, another mass-independent isotope effect must be responsible, at least in part, for the observed anomalies. This is not unexpected, given that previous studies demonstrated that the Pt of iron meteorites displays large cosmogenic isotope effects (Kruijer et al., 2013; Hunt et al., 2014).

4.5.1 Cosmogenic Pt isotope anomalies

To further investigate the origin of the observed Pt isotope anomalies, the models quantifying the effects of GCR on Pt isotopes, as outlined in Section 4.2.2, have been applied. This 192 196 modelling indicates that exposure to GCR will create excesses in ε Pt(8/5) and ε Pt(8/5), 192 196 relative to terrestrial Pt (Figure 4.4a). Notably, the data plotted in ε Pt(8/5) vs. ε Pt(8/5) space reveals the great majority of the meteorites do indeed display such excesses (Figure 4.7).

192 196 Figure 4.7 ε Pt(8/5) vs. ε Pt(8/5) of all iron meteorites analysed. All data are sample means with 2se uncertainties. Grey shaded area represents 2σ of bracketing runs of IRMM-010 Pt. Also shown are Ir/Pt(n,γ) slopes for Ir/Pt ratios of 1 (red line), 0.5 (blue line) and 0.25 (green line), calculated by Equation 4.2. The black solid line represents a theoretical s-excess/r-deficit line calculated by mixing terrestrial Pt with s-process material (Bisterzo et al., 2011).

Mass-independent platinum isotope anomalies in iron meteorites 123

The technique introduced and explained in Section 4.2.2 and Figure 4.4 has also been applied to the dataset. Using the Pt concentrations that were determined by isotope dilution as a by-product of the double spike stable isotope analyses (see Chapter 5 for more details) and Ir abundances from literature (e.g., Wasson, 1970), the Ir/Pt ratios of the iron meteorites have 192 been calculated. Based on these, the expected Ir/Pt(n,γ) slopes for the samples in ε Pt(8/5) vs. 196 ε Pt(8/5) space have been determined (Equation 4.2). For instance, the IIIAB irons are shown in more detail in Figure 4.8. For the meteorites that have resolvable mass-independent isotope anomalies (i.e., with εiPt values that are not within uncertainty of zero), the Ir/Pt(n,γ) slopes were plotted to pass through the relevant meteorite compositions. Inspection of the results reveals that the Ir/Pt(n,γ) trends pass through the origin (within uncertainty) for Santa Apolonia, Charcas 1 and Charcas 2, whilst the trend lines of Cape York, Henbury 2 and Henbury 3 do not. Significantly, the Ir/Pt(n,γ) trends of the individual meteorites do not converge at a single point, even when all uncertainties are considered.

192 196 Figure 4.8 ε Pt(8/5) vs. ε Pt(8/5) of the IIIAB iron meteorites. All values are means with uncertainties of 2se = 2σ/√n, where n = number of times a sample was analysed. Grey shaded area represents 2σ of bracketing IRMM-010 Pt SRM runs. The Ir/Pt(n,γ) slopes were calculated from Equation 4.2 and Ir/Pt ratios, and plotted to pass through the relevant meteorite. The black dashed line represents a theoretical s-excess/r-deficit line calculated by mixing terrestrial Pt with varying proportions of s-process material (Bisterzo et al., 2011).

124 Chapter 4

Together, these findings suggest that the irons analysed here have no clearly resolvable nucleosynthetic Pt isotope anomalies (see later discussion). Therefore, the Ir/Pt(n,γ) trends of the meteorites should, in principle, pass through the origin, yet this is not the case in Figure 4.8. Possible causes for this observation include inaccuracies in the measurements of Pt isotope compositions and Pt, Ir concentrations or inadequacies in the modelling of the cosmogenic isotope effects. Given the accuracy and reproducibility that is demonstrated by analyses of various SRMs (see Section 4.4.1 for more detail), the measurements of the Pt isotope compositions are highly unlikely to feature significant inaccuracies. Likewise, the Pt concentrations determined by isotope dilution are very reproducible (uncertainties of ~0.1%) and hence give no cause for doubt (Chapter 5). Whilst the Ir concentrations are taken from literature, Ir is generally fairly homogenously distributed in individual irons, and even a 10 or 20% inaccuracy in the data would have only a minor effect on the Ir/Pt(n,γ) slope. As such, the Ir concentrations used here are deemed sufficiently accurate for the intended purpose. Hence, it can be concluded that the model, which was developed to predict cosmogenic isotope effects on Pt, may yield useful first order estimates but is most likely insufficiently complex to provide accurate predictions. In particular, the model does not provide an accurate prediction of the relationship between Pt isotope effects and Ir/Pt ratio. One issue may be that the calculations of Kruijer et al. (2013) and Leya and Masarik (2013) only account for the shielding depths of meteorites with pre-atmospheric radii of between 10 and 100 cm. However, many meteorites have large masses and may thus feature shielding depths that are not covered by the modelling. It is therefore possible, that the models are able to account for the cosmogenic Pt isotope effects of smaller meteorites but are less accurate at shielding depths that exceed 100 cm, such that the predicted Ir/Pt(n,γ) slopes no longer hold true. 194 194 Conversely, when the data are plotted in ε Pt(8/5) vs. ε Pt(6/5) space, all results are in accord with the cosmogenic isotope trends predicted by the modelling (Figure 4.9). In this case, the cosmogenic Pt isotope trends do not vary significantly as a function of Ir/Pt ratio (see Section 4.2.2 and Figure 4.4b). Instead, the dominant cosmogenic effect is from the reaction 195Pt(n,γ)196Pt. The excellent agreement of the data with the predicted cosmogenic isotope effects reveal the models are able to consistently deliver accurate results when 195Pt(n,γ)196Pt is the dominant reaction.

Mass-independent platinum isotope anomalies in iron meteorites 125

194 194 Figure 4.9 ε Pt(8/5) vs. ε Pt(6/5) of all iron meteorites analysed. All data are means with uncertainties of 2se = 2σ/√n, where n = number of times sample analysed. Grey shaded area represents 2σ of bracketing IRMM-010 Pt SRM runs. Also shown are Ir/Pt(n,γ) slopes for Ir/Pt ratios of 1 (red line), 0.5 (blue line) and 0.25 (green line), using the modelling of Leya and Masarik (2013). The black line represents a theoretical s-excess/r-deficit line calculated by mixing terrestrial Pt with varying proportions of s-process material (Bisterzo et al., 2011).

A crucial observation from Figure 4.9 is that the iron meteorites plot on both sides of the terrestrial composition, which is defined by εiPt = 0. Whilst most samples display positive ε194Pt values, some meteorites have negative ε194Pt (e.g., Sikhote Alin 1). This is an important result, as it suggests that the IRMM-010 Pt SRM features mass-independent Pt isotope effects relative to some meteorites. There are two explanations for this intriguing observation. Firstly, IRMM-010 Pt may have experienced mass-independent isotope fractionation during chemical processing, such as purification. Such effects have been observed for various purified single element standard solutions, including Ni, Cr, Zr and Hf (Vogl and Pritzkow, 2012). However, analyses of a second in-house Pt standard solution (Pt-α), yielded a Pt isotope composition identical to IRMM-010 (Table 4.6). This would require both SRMs were affected in the same manner by chemical processing, which is highly unlikely. Additional investigations of other purified Pt materials may be useful to further explore this possibility, and may help to robustly characterize the Pt isotope composition of the bulk silicate Earth (BSE). Alternatively, the data shown in Figure 4.9 may imply that the composition of IRMM-010 Pt SRM was itself affected by cosmogenic isotope effects. In this case, it is

126 Chapter 4 unlikely that the isotopic signature was acquired on Earth due to the prevalent shielding. This conclusion is reinforced by the identical Pt isotope compositions that were determined for IRMM-010 Pt and the in-house Pt-α SRM. Instead, it is more reasonable that the Pt of the BSE Earth, as represented by the IRMM-010 Pt and Pt-α SRMs, has a cosmogenic isotope signature, which was acquired prior to accretion onto the Earth. In this context, it is, therefore, significant that the present-day BSE inventory of Pt and other highly siderophile elements is generally thought to originate almost exclusively from a late veneer of chondritic material, which was delivered after the main phase of accretion and cessation of core formation (Kimura et al., 1974). Hence, other highly siderophile elements in the BSE may also feature cosmogenic isotope anomalies, and further investigations of such possible effects are thus of particular interest. Our results may also have a significant impact on the use of Pt isotopes as a neutron dosimeter for correcting measurements of W and Pd isotope compositions (Kruijer et al., 2013; Hunt et al., 2015). Such modelling assumes the initial Pt isotope composition of the meteorites, prior to any GCR effects, was identical to the BSE with εiPt = 0. However, the data presented here suggest that the Pt isotope composition of the BSE was also altered by GCR. Crucially, this seriously questions how accurately the W isotope compositions of iron meteorites can be corrected for cosmogenic isotope effects, as the initial Pt compositions of the irons, prior to exposure to GCR, are unknown and cannot be reliably estimated.

4.5.2 Nucleosynthetic Pt isotope anomalies

Significantly, in Figure 4.9 all meteorites plot within the predicted cosmogenic isotope trends, which pass through the terrestrial composition given by εiPt = 0. Furthermore, none of the irons display a deviation toward, or clearly plot on, the trend predicted for mixing with distinct nucleosynthetic material (s-excess mixing line; see Figure 4.5 for explanation). Likewise, in the detailed plot of the IIIAB data (Figure 4.8), the Ir/Pt(n,γ) trends of the individual meteorites do not converge at a single point, even when all uncertainties are considered. This indicates that, at the current analytical precision, the observed mass-independent Pt isotope variations are essentially entirely cosmogenic in origin, with no resolvable nucleosynthetic isotope anomalies, in agreement with the findings of Kruijer et al. (2013) and Wittig et al. (2013). The lack of any nucleosynthetic Pt isotope anomalies is significant and in accord with the thermal processing model presented in Chapter 2. In detail, this model suggests that the

Mass-independent platinum isotope anomalies in iron meteorites 127 nucleosynthetic isotope anomalies of bulk asteroidal bodies are not caused by the ‘moderately vs. highly’ refractory nature of an element but originate from thermal processing of solar system matter which features decoupled ‘light vs. heavy’ r-process components. Crucially, Pt has properties that allow a discrimination between the two models – it has ‘heavy’ r-process components since Z > 56 and a lower 50% condensation temperature than Mo (1403 and 1587 K, respectively; Lodders, 2003). These properties are considered in more detail with respect to the two models in the following. Fundamental to thermal processing models related to ‘moderately vs. highly’ refractory components (e.g., Burkhardt et al., 2012b), is the observation that elements characterised by isotopic homogeneity (Os, Hf, W) are all highly refractory with 50% condensation temperatures of ~1700–1800 K. In contrast, elements that display isotopic heterogeneities in bulk meteorites are less refractory and have lower condensation temperatures (e.g., Mo = 1587 K) (Lodders, 2003). As Pt has a condensation temperature that is even lower than for Mo, one might expect to see evidence of thermal processing with resolvable nucleosynthetic Pt isotope anomalies – yet this is not the case. Instead, the most likely cause is the decoupling of ‘light’ (Z ≤ 56) and ‘heavy’ (Z > 56) r-process nuclides (e.g., Wasserburg et al., 1996; Akram et al., 2013). The heavy r-process isotopes are formed by the main r-process, with the preferred stellar site of neutron star mergers (Tsujimoto and Shigeyama, 2014a). Conversely, the light r-process isotopes are thought to be formed predominantly by the weak r-process and charged particle reactions (CPRs), as well as small contributions from the main r-process, in Type II supernovae and their neutrino driven winds (Wanajo, 2013). This interpretation entails that the host phases of ‘light’ and ‘heavy’ r-process isotopes have different susceptibilities to thermal processing, perhaps due to different grain sizes and/or chemical compositions of the phases. Indeed, all elements with observed s- or r-process heterogeneity in bulk meteorites (Mo, Ba, Ca, Cr, Ni, Ru, Sm, Ti and Zr) have Z ≤ 56, while those homogeneously distributed in the solar nebula have Z > 56 (Pt, Nd, Hf, Os, and W). Therefore, the Pt isotope data acquired in our study provide further support for the thermal processing model presented in Chapter 2, in which the decoupling of ‘light vs. heavy’ r-process components is the controlling factor. Further investigation of elements near the cut-off of Z = 56 and refractory elements with a wide range of condensation temperatures, could further elucidate the validity of this hypothesis.

128 Chapter 4

4.6 Conclusions

Investigations into mass-independent Pt isotope anomalies of iron meteorites reveal large cosmogenic isotope effects. Within the analytical precision, no resolvable nucleosynthetic Pt isotope anomalies could be detected. Based on the thermal processing model that was developed to explain the nucleosynthetic Mo isotope effects of iron meteorites (Chapter 2), nucleosynthetic Pt isotope anomalies were predicted to be absent in such samples. Therefore, the Pt isotope data presented here provide support for the hypothesis that the nucleosynthetic isotope anomalies of bulk asteroidal bodies are controlled by the decoupled sources of ‘light vs. heavy’ r-process nucleosynthetic components. The new data were used to interrogate previously published models, which estimate the cosmogenic isotope effects that are expected for meteorites from exposure to GCR. These models were found to provide good estimates of the neutron capture effects on 194Pt, 195Pt, 196Pt and 198Pt. In contrast, the production of 192Pt due to neutron capture by 191Ir does not appear to be well captured by the models and this discrepancy may be particularly large at greater shielding depths. Finally, it was observed that a few iron meteorites had Pt isotope compositions that were less affected by cosmogenic isotope effects than our terrestrial Pt standard reference materials This may indicate that the late veneer material which provided the present BSE inventory of Pt was affected by interaction with GCR and the associated isotopic signature was imparted to the complete silicate portion of the Earth. Such an outcome is significant because it severely complicates the use of Pt isotopes as a neutron dosimeter for correcting measurements of W and Pd isotope compositions in solar system materials for cosmogenic isotope effects.

Chapter 5

Platinum stable isotopes in iron meteorites

130 Chapter 5

5.1 Introduction

Measurements of mass-dependent stable isotope fractionations for various elements have long been used as a tool for studying processes in the solar nebula and planetary bodies. For instance, condensation and evaporation were investigated using the stable isotopes of both volatile (Cd; Wombacher et al., 2008) and refractory (Sr; Moynier et al., 2010) elements, whilst planetary differentiation and core formation were studied by Cr, Mo, Si and other isotope systems (Armytage et al., 2011; Moynier et al., 2011; Burkhardt et al., 2014). In contrast, platinum isotopes have yet to be fully explored, with only limited previous investigation (Creech et al., 2014b). As Pt is a highly siderophile element with relatively high concentrations in iron meteorites, Pt stable isotopes offer a potentially useful tool to improve understanding of the processes and conditions relevant for the formation of iron meteorites and their parent bodies. Bulk samples of iron meteorites and, by inference, the respective parent bodies show no evidence of nucleosynthetic Pt isotope anomalies arising from thermal processing of solar nebula material (see Chapter 4). This stands in contrast to Mo isotopes, where evidence of thermal processing is also seen in the Mo stable isotope compositions of the irons (Chapter 2). Therefore, one would not expect to see evidence of such processing in stable Pt isotopes. Instead, stable Pt isotopes may reveal information about other processes operating in early solar system history and give insight into pre-, syn- and post-accretionary processes associated with parent body formation. Here, the results of the first comprehensive investigation of the stable Pt isotope compositions of iron meteorites are presented. A large range of magmatic and non-magmatic irons were analysed and the results investigated to determine the origin of the observed isotopic variations.

5.2 Analytical techniques

Aliquots of the iron meteorite sample solutions that were prepared as described in Chapter 1 were employed for the analyses that are detailed below. All materials and reagents used, and the laboratory conditions, were the same as those outlined in Chapter 1.

Platinum stable isotopes in iron meteorites 131

5.2.1 Platinum double spike preparation and calibration

The preparation and calibration of the Pt double spike was carried out at the Imperial College London MAGIC Laboratories together with Dr Roland Stumpf. As the natural abundances of 190Pt and 192Pt are very low (0.01 and 0.78%, respectively), the four isotopes selected for the double spike data reduction were 194Pt, 195Pt, 196Pt and 198Pt. Enriched single spikes of 198Pt and 196Pt were obtained from Isoflex USA in the form of Pt wire and digested in aqua regia. Based on the double spike toolbox calculations of Rudge et al. (2009), the Pt double spike was gravimetrically prepared by mixing the 198Pt and 196Pt single spike solutions in a ratio of 72.5 to 27.5. The 198Pt-196Pt double spike composition was calibrated in isotopic analyses using external normalisation to admixed Ir for correction of the instrumental mass bias. An Ir-doped solution of IRMM-010 Pt SRM was first analysed by MC-ICP-MS. The Pt isotope ratios of these runs were corrected for instrumental mass bias by both internal and external 196 194 193 191 193 191 normalisation to Pt/ Pt and Ir/ Ir, respectively. The reference ratio ( Ir/ Ir)Ref employed in the latter procedure was hereby empirically optimised to provide the best possible agreement between the Pt isotope ratios normalised to Pt and Ir. This was followed by concentration-matched runs of the Ir-doped Pt double spike solution. In these analyses, the Pt isotope ratios were corrected for instrumental mass bias by external normalisation to 193 191 193 191 Ir/ Ir using the empirically optimised ( Ir/ Ir)Ref ratio from the previous IRMM-010 Pt runs. The resulting Pt isotope ratios were then evaluated to define the isotope compositions of the 198Pt-196Pt double spike (Figure 5.1).

Figure 5.1 Isotope composition of the new 198Pt-196Pt double spike. Shown are the double spike Pt isotope abundances, calibrated as described in text (blue bars), in comparison with terrestrial Pt isotope abundances (red bars) from Creech et al. (2013).

132 Chapter 5

5.2.2 Ion-exchange chromatography

Separation of Pt from the iron meteorite matrix was achieved using the procedure described in Chapter 4, with 20–100 mg of meteorite (~600 ng natural Pt) processed for each sample (Table 5.1). Prior to separation chemistry, the samples were mixed and equilibrated with the Pt double spike by refluxing on a hotplate for at least 24 hours.

Table 5.1 Ion-exchange chemistry for separation of Pt from iron meteorites

Quartz column, 3 ml resin reservoir, 50 ml acid reservoir Resin: Bio-Rad AG1-X8, 200–400 mesh, chloride form (1 ml) Step Resin volumes Acid

a Cleaning 1.5 8 M HNO3 1.5 11 M HCl 1 1 M HCl Pre-condition resin 4 1 M HCl Load Sample 2–4 1 M HCl Elute matrix (inc. Fe, Ni) 40 0.5 M HCl Elute Ir, Os, Re, Pd 50 11 M HCl a

Elute and collect Pt 14 13.5 M HNO3 a Step followed by rinse with 1 ml H2O to prevent aqua regia forming in resin during next step

5.2.3 Mass spectrometry

5.2.3.1 Instrumentation and data collection protocol

The isotope measurements were performed using the Nu Instruments Nu Plasma HR MC-ICP-MS at Imperial College London with sample introduction via a Nu Instruments DSN-100 desolvating nebuliser at an uptake rate of ~140 µl/min. Typical sensitivity for Pt was 180–210 V/ppm for solutions with ~100 ppb Pt. The data were acquired by static multiple collection with the Faraday cups of the instrument, with simultaneous measurement of the 194Pt, 195Pt, 196Pt, 198Pt, 200Hg and 193Ir ion beams (using 1011 Ω amplifiers) in a single cycle. The Faraday cup configuration and positions of the masses are indicated in Table 5.2. The analyses utilised 3 blocks with 20 integrations of 5 seconds each. Each block was preceded by a 30 second on-peak baseline measurement while the ion beam was deflected by the electrostatic analyser.

Platinum stable isotopes in iron meteorites 133

Table 5.2 Faraday cup configuration for measurement of stable Pt isotopes

Faraday Cup L5 L4 L3 L2 L1 Ax H1 H2 H3 H4 H5 H6

amu 193 194 195 196 198 200

Following measurement of the 195Pt/194Pt, 196Pt/194Pt and 198Pt/194Pt ratios all further data reduction took place offline. This employed the iterative approach of Siebert et al. (2001), which was implemented in a Microsoft Excel spreadsheet (Ripperger and Rehkämper, 2007; Xue et al., 2013). The final results are reported in δ198Pt notation (Equation 5.1), calculated relative to the mean of several bracketing runs of solutions of the IRMM-010 Pt reference material to which Pt double spike had been admixed. The mixed double spike – IRMM-010 Pt solutions were thereby made up to closely match both the Pt concentrations (of ~100 ppb) and molar Ptspike/Ptnatural ratios of the samples.

198Pt 194Pt 198 sample 3 δ Pt = 198 194 − 1 × 10 Equation 5.1 Pt Pt standard

5.2.3.2 Interferences

The most important possible spectral interferences on the Pt isotopes are shown in Table 5.3. Isobaric interferences from Hg on 196Pt and 198Pt were corrected using 200Hg as interference monitor and the Hg isotopic abundances of Berglund and Wieser (2011). Spectral interferences from molecular species and doubly charged ions were also monitored but the efficiency of the separation chemistry for Pt was sufficient so that relevant interferences were essentially negligible.

Table 5.3 Important spectral interferences on platinum isotopes

Interference 194Pt 195Pt 196Pt 198Pt

Isobaric 196Hg 198Hg

Argides 154Sm40Ar 155Gd40Ar 156Gd40Ar 158Gd40Ar 154Gd40Ar 156Dy40Ar 158Dy40Ar

Oxides 178Hf16O 179Hf16O 180Hf16O 182W16O 180Ta16O 180W16O

134 Chapter 5

5.2.4 Resolving mass-dependent and mass-independent Pt isotope effects

A graphical representation of the Pt double spike technique is provided in Figure 5.2. Mass-independent isotope effects on δ198Pt in the double spike data reduction were removed based on the approach outlined for Mo in Chapter 3, whereby instead of using the SRM composition for n, the relevant mass-independent isotope anomalies from Chapter 4 are added to the Pt isotope composition of the SRM. In essence, this yields a modified SRM composition, n', that is used in the double spike data reduction to calculate the true isotope composition of the sample, N.

Figure 5.2 Schematic of the Pt double spike technique in four-isotope space. Adapted from Galer (1999). For details of double spike technique see Chapter 1.

The Mo correction hereby relies on Mo isotope data internally normalised to 97Mo/95Mo. The key advantage of this ratio is that it remains essentially invariant even in samples that record clear excesses or deficits of nucleosynthetic components. Unfortunately, no such invariant normalisation ratio is available for Pt isotopes when samples with variable cosmogenic isotope effects are analysed. Therefore, to fully assess the accuracy of the stable Pt isotope corrections, the data internally normalised to 196Pt/194Pt, 198Pt/195Pt and 198Pt/194Pt were used. 198 198 198 The resulting stable Pt isotope compositions, reported as δ Pt(6/4), δ Pt(8/5) and δ Pt(8/4) relative to IRMM-010 Pt, respectively, were then compared. The uncertainties associated with these mass-independent Pt isotope measurements were propagated through all data reduction and correction procedures. To assess the magnitude and importance of the corrections, the

Platinum stable isotopes in iron meteorites 135

δ198Pt data that result if no such corrections are applied are also calculated, and reported as 198 δ Pt(SRM). These results are obtained when the SRM composition is applied as n in the double spike data reduction and no nucleosynthetic isotope anomalies are subtracted from the δ198Pt values generated.

5.3 Results

5.3.1 Standard solutions and reference materials

To ascertain the precision and accuracy of the data, several IRMM-010 Pt SRM solutions with varying proportions of double spike Pt to natural Pt were analysed (Figure 5.3). The various mixtures were analysed relative to a solution with a ratio of spike-derived Pt to natural Pt of S/N = 0.84, which is the optimum S/N as calculated using the double spike toolbox of Rudge et al. (2009). These analyses yielded two main results. (1) The precision of the δ198Pt measurement is a factor of ~1.5 better for solutions with S/N ratios of between 0.7 and 1.1, than for S/N < 0.7 or > 1.1. (2) Ratios of S/N < 0.7 produce positive δ198Pt values, whereas ratios of S/N > 1.1 produce negative δ198Pt. In contrast, S/N values between 0.7 and 1.1 yield δ198Pt = 0.00 ± 0.03‰.

Figure 5.3 δ198Pt values of IRMM-010 Pt standard solutions with variable ratios of spike- derived Pt to natural Pt (S/N). These solutions were analysed relative to a solution with S/N = 0.84. The grey shaded area represents the typical external reproducibility of the bracketing standard solutions. For S/N ratios between 0.7 and 1.1, δ198Pt varies by less than 0.03‰. Crucially, for solutions with S/N between 0.7 and 1.1, the uncertainties in δ198Pt are a factor of ~1.5 better than for solutions with S/N outside this range.

136 Chapter 5

Based on these results, the samples were spiked to obtain S/N ≈ 0.84. In this case, if the actual spiking is slightly off because the Pt concentration of a sample was not characterised well, this will have no significant effect on data quality, as long as the sample solution features an S/N ratio of between about 0.7 and 1.1. In cases where the S/N ratio of a spiked sample was outside this range, the bracketing standard runs utilised a specifically prepared solution of spiked IRMM-010 Pt, which featured the same S/N ratio as the anomalous sample solution. The typical external reproducibility of bracketing runs of the spiked IRMM-010 Pt (S/N ≈ 0.84) was δ198Pt = ±0.060‰ (2σ), with typical internal precision of δ198Pt = ±0.054‰ (2se). The robustness and reproducibility of the Pt separation procedure and MC-ICP-MS Pt isotope measurements were routinely evaluated by repeated analyses of terrestrial standard reference materials. Both NIST SRM 129c (High-Sulphur Steel) and NIST SRM 361 (AISI 4340 Steel) were mixed and equilibrated with solution of double spike Pt and the IRMM-010 Pt SRM (to match the Pt concentration and S/N ratio of samples) prior to ion-exchange chromatography. The analyses of these Pt-doped reference materials consistently yielded terrestrial Pt isotope compositions with δ198Pt results of −0.002 ± 0.064‰ and +0.007 ± 0.073‰ (all uncertainties are 2σ), respectively (Figure 5.4). An in-house Pt standard solution, denoted Pt-α, was also analysed against IRMM-010 Pt at regular intervals during each measurement session to monitor data quality. These analyses yielded a reproducible offset of δ198Pt = −0.072 ± 0.021‰ for Pt-α relative to IRMM-010 Pt (Figure 5.4).

Figure 5.4 δ198Pt of terrestrial standard reference materials. Analysed relative to bracketing IRMM-010 Pt SRM solutions. Error bars denote the 2se in-run precision; uncertainties of reported means are 2σ. Grey-shaded area represents the typical ±0.060‰ reproducibility of the bracketing IRMM-010 Pt runs.

Platinum stable isotopes in iron meteorites 137

5.3.2 Iron meteorites

The Pt stable isotope compositions of the iron meteorites are shown in Table 5.4 and Figure 5.5. Also shown in Table 5.4 are the Pt concentrations determined by isotope dilution using the 198Pt-196Pt double spike.

5.3.2.1 Pt concentrations

The magmatic iron meteorites have Pt concentrations ranging from 1.7 to 36.6 ppm. Non- magmatic irons have a smaller variation of Pt concentrations, from 5.1 to 13.1 ppm. Notably, there are variations within groups. For instance the Pt concentrations of the IIIAB irons vary from 1.7 ppm (Bear Creek) to 16.5 ppm (Henbury 2 and Santa Apolonia 1).

5.3.2.2 δ198Pt values and corrections

The δ198Pt values corrected for mass-independent isotope effects, as well as the δ198Pt data that result if no such corrections are applied, are shown in Table 5.4. The irons are corrected to both higher and lower δ198Pt values, depending on the exact nature of the cosmogenic 198 198 isotope effects. Generally, the corrections for δ Pt(8/5) are larger than those for δ Pt(6/4) and 198 δ Pt(8/4). 198 198 198 198 The largest corrections from δ Pt(SRM) to higher δ Pt(6/4), δ Pt(8/5) and δ Pt(8/4) values are for Tlacotopec 1 (+0.089, +0.326 and +0.114‰, respectively), whilst Gibeon has the most significant corrections to lower δ198Pt values (–0.027, –0.091 and –0.066‰, respectively). Within the iron meteorite groups, the greatest range in δ198Pt values is observed for the IIABs, while the IIIAB and IVA meteorites show the smallest variations, as shown in 198 Figure 5.5 for δ Pt(6/4).

138 Chapter 5

Table 5.4 Platinum concentrations and stable isotope compositions of the iron meteorites

a b 198 c 198 d 198 e 198 f Group Sample N Pt (ug/g) δ Pt(SRM) δ Pt(6/4) δ Pt(8/5) δ Pt(8/4)

IAB Campo del Cielo 2 5 8.3 −0.047 ± 0.019 −0.012 ± 0.034 0.051 ± 0.096 −0.002 ± 0.034 Campo del Cielo 3 5 8.8 −0.018 ± 0.032 0.033 ± 0.048 0.094 ± 0.060 0.022 ± 0.042 Canyon Diablo 2 5 5.1 −0.075 ± 0.038 −0.037 ± 0.046 0.021 ± 0.044 −0.068 ± 0.044 IAB MEAN −0.053 ± 0.042 −0.013 ± 0.047 0.047 ± 0.051 −0.029 ± 0.078

IC Arispe 3 5 18.7 −0.090 ± 0.012 −0.054 ± 0.039 0.011 ± 0.059 −0.078 ± 0.037 Arispe 4 5 18.9 −0.035 ± 0.023 −0.023 ± 0.054 0.054 ± 0.057 −0.049 ± 0.054 Bendego 1 5 11.8 0.284 ± 0.027 0.331 ± 0.062 0.395 ± 0.108 0.310 ± 0.060 Bendego 2 5 11.5 0.289 ± 0.037 0.362 ± 0.043 0.408 ± 0.080 0.287 ± 0.044 Santa Rosa 5 5.1 0.222 ± 0.035 0.213 ± 0.040 0.165 ± 0.051 0.156 ± 0.040 IC MEAN 0.149 ± 0.215 0.174 ± 0.226 0.200 ± 0.216 0.131 ± 0.211

IIAB Coahuila 5 31.5 −0.022 ± 0.026 −0.023 ± 0.028 −0.029 ± 0.032 −0.027 ± 0.027 Murphy 6 36.6 −0.100 ± 0.020 −0.112 ± 0.026 −0.110 ± 0.038 −0.092 ± 0.025 North Chile 6 24.3 −0.004 ± 0.028 0.000 ± 0.039 0.045 ± 0.061 −0.007 ± 0.036 Sikhote Alin 1 5 4.3 0.316 ± 0.040 0.358 ± 0.041 0.322 ± 0.046 0.351 ± 0.041 Sikhote Alin 2 5 5.2 0.161 ± 0.029 0.185 ± 0.039 0.227 ± 0.034 0.170 ± 0.036 IIAB MEAN 0.028 ± 0.146 0.034 ± 0.165 0.045 ± 0.166 0.034 ± 0.155

IIC Ballinoo 5 14.4 0.004 ± 0.026 −0.015 ± 0.038 −0.065 ± 0.055 −0.020 ± 0.037 Kumerina 8 9.3 0.190 ± 0.021 0.191 ± 0.036 0.194 ± 0.038 0.187 ± 0.035 Salt River 5 13.3 0.007 ± 0.032 0.076 ± 0.044 0.225 ± 0.049 0.075 ± 0.041 IIC MEAN 0.067 ± 0.123 0.084 ± 0.120 0.118 ± 0.183 0.081 ± 0.120

IIE Kodaikanal 5 11.1 −0.066 ± 0.006 −0.019 ± 0.019 0.065 ± 0.032 −0.018 ± 0.017 Verkhne Dneiprovsk 5 11.8 0.001 ± 0.016 −0.002 ± 0.021 −0.007 ± 0.025 0.001 ± 0.020 Weekeroo Station 5 13.1 −0.001 ± 0.030 0.031 ± 0.036 0.099 ± 0.045 0.017 ± 0.036 IIE MEAN −0.022 ± 0.044 0.003 ± 0.029 0.052 ± 0.063 0.000 ± 0.020

IIIAB Bear Creek 3 1.7 0.015 ± 0.026 −0.001 ± 0.031 −0.038 ± 0.032 −0.005 ± 0.030 Cape York 6 14.3 −0.006 ± 0.036 0.009 ± 0.040 0.109 ± 0.054 0.013 ± 0.040 Charcas 1 7 12.8 −0.039 ± 0.013 −0.030 ± 0.028 −0.028 ± 0.043 −0.080 ± 0.028 Charcas 2 16 12.4 −0.012 ± 0.019 −0.009 ± 0.026 −0.011 ± 0.028 −0.056 ± 0.025 Henbury 2 6 16.5 −0.064 ± 0.015 −0.063 ± 0.026 0.034 ± 0.050 −0.070 ± 0.026 Lenarto 8 6.5 0.043 ± 0.017 0.044 ± 0.024 0.053 ± 0.041 0.047 ± 0.024 Santa Apolonia 1 9 16.5 −0.030 ± 0.028 −0.033 ± 0.034 0.003 ± 0.045 −0.051 ± 0.033 Verkhne Udinsk 9 13.1 0.022 ± 0.031 0.027 ± 0.035 0.055 ± 0.036 0.038 ± 0.034 Williamette 8 7.1 0.064 ± 0.016 0.061 ± 0.022 0.044 ± 0.028 0.077 ± 0.022 IIIAB MEAN −0.008 ± 0.034 0.005 ± 0.027 0.033 ± 0.034 −0.001 ± 0.038

IIICD Nantan 1 5 6.7 −0.178 ± 0.027 −0.141 ± 0.052 −0.101 ± 0.074 −0.141 ± 0.063

IIIE Staunton 8 6.4 0.175 ± 0.016 0.168 ± 0.021 0.250 ± 0.034 0.168 ± 0.022

IIIF Clark County 9 9.7 −0.022 ± 0.016 −0.024 ± 0.029 0.012 ± 0.031 −0.019 ± 0.030

IVA Gibeon 2 7 6.0 0.035 ± 0.028 0.009 ± 0.036 −0.056 ± 0.046 −0.031 ± 0.036 Obernkirchen 6 6.8 −0.035 ± 0.014 −0.054 ± 0.023 −0.118 ± 0.037 −0.093 ± 0.022 IVA MEAN 0.000 ± 0.070 −0.023 ± 0.063 −0.087 ± 0.062 −0.062 ± 0.062

IVB Cape of Good Hope 6 30.1 −0.056 ± 0.025 −0.011 ± 0.032 0.109 ± 0.040 −0.008 ± 0.031 Santa Clara 7 23.1 −0.078 ± 0.025 −0.045 ± 0.031 0.070 ± 0.072 −0.057 ± 0.031 Tlacotopec 1 5 18.6 −0.318 ± 0.021 −0.229 ± 0.027 0.007 ± 0.042 −0.205 ± 0.027 IVB MEAN −0.151 ± 0.168 −0.095 ± 0.135 0.062 ± 0.059 −0.090 ± 0.118

a b c ‘ 198 Number of times sample was analysed; Pt concentrations determined by isotope dilution; δ Pt measured’ values obtained if no corrections are applied to the data for mass-independent isotope effects; d ‘δ198Pt corrected’ values are corrected for mass-independent isotope effects, with uncertainties of mass-independent effects propagated; Uncertainties for samples are 2se = 2σ/√n, where n is the number of times the sample was analysed. Uncertainties for group means are 2se = 2σ/√n, where n is the number of ‘unique’ samples analysed for that group (for samples with multiple specimens, the mean of specimens is taken to represent that sample when calculating the group mean).

Platinum stable isotopes in iron meteorites 139

198 Figure 5.5 δ Pt(6/4) of iron meteorites, corrected for mass-independent Pt isotope effects. Uncertainties are propagated 2se = 2σ/√n, where n is the number of sample analyses. Meteorites within the same group are plotted alphabetically.

5.4 Discussion

5.4.1 Magmatic iron meteorite parent bodies

5.4.1.1 Platinum concentrations

To assess the behaviour of Pt during the crystallisation of the magmatic iron meteorite parent bodies, diagrams of Pt concentration against Ir concentration are displayed in Figure 5.6. This study provides the Pt data for all meteorites analysed, while the Ir data data is taken from literature (Catalogue of Meteorites (2000) and references therein). Additionally, for the IIC, IVA and IVB groups, supplementary Pt and Ir concentration data have been employed from literature to complement the limited Pt data from this study (Wasson and Richardson, 2001;

140 Chapter 5

Petaev and Jacobsen, 2004; Walker et al., 2008; Chernonozhkin et al., 2014). The plots of Pt vs. Ir concentrations for the IC, IIAB, IIC, IIIAB and IVA irons reveal resolvable trends, where the meteorites with the highest Pt concentrations for a group also have the highest Ir abundances. The chemical classification of the (magmatic) iron meteorite groups is based on the coherent trends that group members define in siderophile element plots, such as Ge or Ga versus Ni or Ir concentrations. These correlations are thought to reflect the elemental behaviour during progressive fractional crystallisation of metal during cooling and solidification of parent body cores (see Chapter 1). During crystallisation of asteroidal cores, Ir displays compatible behaviour and preferentially partitions into the solid rather than the liquid metal. Therefore, the earliest solid metals have higher Ir concentrations compared to the metals which crystallised later from an increasingly Ir-depleted metallic melt. The correlations that are observed here for the Pt and Ir concentrations of the magmatic irons are hence also best explained as a consequence of fractional crystallisation. In detail, the trends of Figure 5.6 thus imply that Pt displayed compatible behaviour during fractional crystallisation of the IC, IIAB, IIC, IIIAB and IVA parent body cores, with preferential partitioning into the solid metal, and therefore, solid metal – liquid metal partition coefficients, DS/L(Pt), of > 1. This conclusion is supported by previous studies; McCoy et al. (2011), for example, observed compatible behaviour of Pt during fractionation crystallisation of the IVA irons. In contrast, the data for the IVB irons in Figure 5.6 show no clear correlation and hence do not support compatible behaviour of Pt, in accord with the findings of other previous studies. Campbell and Humayun (2005) determined a DS/L(Pt) value of 0.98 ± 0.05 for the IVB parent melt, while Walker et al. (2008) found minimal variations in Pt concentrations during fractional crystallisation of IVBs, in accord with DS/L(Pt) ≈ 1. Interestingly, the Tlacotopec specimen measured here has a Pt concentration of 18.6 ppm, which is significantly lower than previously published Pt concentrations for both Tlacotopec (30.1 ppm) and other IVBs (~30 to 33 ppm) (Walker et al., 2008). Potentially, this implies that the Tlacotopec meteorite is heterogeneous in composition. Another notable anomaly is seen in the IICs. The irons of this group appear to display a resolvable trend in Figure 5.6, in accord with compatible behaviour of Pt during metal crystallisation. However, there is one anomalous specimen (Kumerina), which seems to have a lower Pt and/or higher Ir content than expected. The composition of this sample may represent a trapped melt, whereby the partitioning occurs in a separate closed system, which may feature distinct conditions (e.g., extreme S concentrations).

Platinum stable isotopes in iron meteorites 141

Figure 5.6 Platinum vs. iridium concentration of magmatic iron meteorites. Filled symbols – Pt data from this study, Ir data from Catalogue of Meteorites (2000) and references therein. Open symbols –Pt and Ir data from literature (IIC: Chernonozhkin et al. (2014); IVA: Petaev and Jacobsen (2004) and Wasson and Richardson (2001); IVB: Walker et al. (2008)). Note that fractional crystallisation progresses from high Ir concentrations (right side of diagrams) towards lower Ir contents.

5.4.1.2 Platinum isotope compositions

To robustly assess the δ198Pt values of individual meteorites within a group, one must first consider whether and how the correction for mass-independent isotope effects could produce 198 198 198 any systematic artefacts in the data. In Figure 5.7, the δ Pt(6/4), δ Pt(8/5) and δ Pt(8/4) values are plotted against Pt concentrations. For each group, the same relationships exist, regardless of the δ198Pt value plotted. This demonstrates that δ198Pt is primarily governed by the occurrence or absence of mass-dependent stable isotopes fractionations rather than any cosmogenic isotope effects. Hence, the reported δ198Pt values can be interpreted with confidence as reflecting Pt stable isotope fractionations.

142 Chapter 5

Significant differences between the different δ198Pt values of Fig. 5.7 only arise when an individual sample has a much greater mass-independent isotope effect than the other irons 198 of the group. For instance, in the IIC group, Salt River has a different position in the δ Pt(8/5) plot, due to the larger cosmogenic isotope effects observed for this meteorite in comparison to the other IICs (Chapter 4). However, when considering the IIC group as a whole, the δ198Pt data are still predominantly governed by the mass-dependent stable isotope effects associated with fractional crystallisation of the core. In diagrams of δ198Pt vs. Pt concentration (Figure 5.7), the meteorites of the IIAB and IIC groups display negative correlations, whereby the irons with higher Pt contents have lower δ198Pt values (i.e., lighter Pt isotope compositions). These correlations leads to an intriguing interpretation – during fractional crystallisation of the parent body cores, the lighter Pt isotopes preferentially partition into the solid metal over the liquid metal. Additionally, the IVA and IC groups hint at the same correlation, although insufficient data are available to robustly confirm the trends. Surprising hereby is the slope of the observed correlation because equilibrium isotope fractionation is generally associated with preferential partitioning of the heavier isotopes into the phase with stronger bonding (e.g., Criss, 1999), which is expected to be the solid metal. It is noted, however, that exceptions to this rule of thumb are not uncommon. For example, light Fe isotopes are preferentially enriched in olivine phenocrysts relative to the silicate melt in Hawaiian basalts (Teng et al., 2008). In contrast, the IVB irons display a positive correlation in Fig. 5.7, whereby higher δ198Pt values are associated with higher Pt contents. This may indicate that heavier Pt isotopes preferentially partitioned into the solid metal during fractional crystallisation of the IVB parent body core. In principle, these systematics are more akin with the expected stable isotope fractionation during equilibrium partitioning between solid and liquid metal. The stable Pt isotope compositions of the IIIAB iron meteorites show no clear correlation with Pt concentration, in accord with indifferent partitioning of Pt isotopes during core crystallisation. In this context it is noteworthy that previous investigations concluded the core of the IIIAB parent body crystallised via inward dendritic growth, which led to trapped melts with varying S contents (Haack and Scott, 1992; Wasson, 1999). Such compositional variability may have an effect on the solid metal – liquid metal partitioning and associated isotope fractionation of Pt, as reflected in the lack of a clear trend for the IIIABs in Figure 5.7.

Platinum stable isotopes in iron meteorites 143

198 198 Figure 5.7 δ Pt vs. Pt concentration of magmatic iron meteorites. The panels on the left show δ Pt(6/4) 198 198 198 values, the centre panels δ Pt(8/5) and the right panels δ Pt(8/4). Uncertainties for δ Pt are propagated 2se. Note fractional crystallisation proceeds from right to left in each plot. Differences between plots for the same group arise due to persistent artefacts from cosmogenic isotope effects in the double spike data reduction.

144 Chapter 5

5.4.2 Non-magmatic iron meteorite parent bodies

5.4.2.1 Platinum concentrations

Platinum concentrations from the IAB/IIICD complex display a positive correlation with Ir (Figure 5.8), indicative of the same compatible behaviour of Pt during metal crystallisation as was previously observed for the magmatic irons (Figure 5.6). However, whilst the siderophile element trends exhibited by magmatic irons are in accord with the crystallisation of a common core, the elemental compositions of non-magmatic irons do not fit such a scenario (see Chapter 1 for further detail). Nonetheless, it is conceivable that the correlation of Pt with Ir concentrations for the IAB/IIICD irons may reflect fractional crystallisation but this could have occurred in several distinct pools, either at the surface of the parent body or the interior (e.g., Benedix et al., 2000; Wasson and Kallemeyn, 2002). Alternatively, the IAB/IIICD trend of Figure 5.8 may be a consequence of partial melting during partial differentiation of the parent body. Similarly, the IIE data, in combination with data from previous studies, indicate compatible behaviour of Pt during metal crystallisation (Figure 5.8).

Figure 5.8 Platinum vs. iridium concentration of non-magmatic iron meteorites. Filled symbols – Pt data from this study, Ir data from Catalogue of Meteorites (2000) and references therein. Open symbols – Pt and Ir data from literature (IIICD: Petaev and Jacobsen (2004); IIE: Chernonozhkin et al. (2014)). Note that fractional crystallisation progresses from high Ir concentrations (right side of diagrams) towards lower Ir contents.

5.4.2.2 Platinum isotope compositions

The data for the IAB/IIICD complex irons show no resolvable correlation between δ198Pt and Pt concentrations. This behaviour, which is akin to that observed for the IIIAB meteorites (Figure 5.7), may reflect the distinct compositions and conditions at which metal crystallisation occurred. Notably, it is often argued that the IAB/IIICD parent body was particularly S-rich (e.g., Benedix et al., 2000), and various investigations have considered the

Platinum stable isotopes in iron meteorites 145 effect of higher S and P contents on the crystallising metal systems (e.g., Chabot and Jones, 2003). Furthermore, different pools of segregated melts may have different compositions, particularly with respect to S. As a result, the partitioning and any associated isotope fractionation of Pt may vary slightly from pool to pool, depending on the exact composition of the respective metal melts. Similarly, the irons analysed here from the non-magmatic IIE group reveal no clear trend of Pt isotope fractionation during crystallisation of the parent body (Figure 5.9).

Figure 5.9 δ198Pt vs. Pt concentration of non-magmatic iron meteorites. The uncertainties for δ198Pt are propagated 2se = 2σ/√n.

5.4.3 Initial δ198Pt of solar nebula

As the Pt stable isotope compositions of iron meteorites primarily reflect changes induced by the isotope fractionation that accompanies fractional crystallisation of solid metal, any previous early solar system signatures, for example from nebula processing, are overprinted. However, a tentative estimate of the mean δ198Pt value of the solar system can be obtained based on a plot of δ198Pt vs. Ir concentrations for all iron meteorites analysed here (Figure 5.10). A clear negative correlation, albeit with significant scatter, is observed for this large dataset. The bulk δ198Pt value of the system, taken in the following to approximate the solar system average, can then be estimated from the Ir-rich end of the correlation, because this encompasses the meteorites that first crystallised from the molten metal. Given the observation that the lighter Pt isotopes preferentially partitioned into the solid metal, the initial bulk metal must have had a slightly higher δ198Pt value than the most

146 Chapter 5

Ir-rich magmatic irons. From the data shown in Figure 5.10, it can therefore be cautiously estimated that the bulk metal of the iron meteorites was characterised by δ198Pt ≈ 0‰. As Ir is a highly siderophile element, the cores contain essentially the complete Ir budget of any differentiated asteroid. This, in turn, implies that the average bulk δ198Pt of the iron meteorites can be employed to derive a first-order estimate for the Pt isotope composition of the inner solar system. It will be useful, however, to further investigate the above result based on stable Pt isotope analyses of chondritic meteorites that were not affected by large-scale melting and differentiation processes.

198 Figure 5.10 δ Pt(6/4) vs. Ir concentration of all iron meteorites analysed in this study. The uncertainties for δ198Pt are 2se = 2σ/√n. Ir data from Catalogue of Meteorites (2000) and references therein. An overall trend is observed, whereby more negative δ198Pt values correlate with higher Ir concentrations.

5.5 Conclusions

The new measurements of Pt concentrations in iron meteorites reveal generally compatible behaviour of Pt during both the fractional crystallisation of the magmatic parent body cores, and the crystallisation of liquid metal in the non-magmatic bodies. Furthermore, stable isotope analyses reveal varying isotope fractionation behaviour. For the cores of the IIAB and IIC (perhaps also the IC and IVA) iron meteorite parent bodies, the lighter isotopes of Pt partition preferentially into the solid metal during fractional crystallisation. This generates trends of increasing δ198Pt values with decreasing Pt concentrations for most groups of iron

Platinum stable isotopes in iron meteorites 147 meteorites. Other groups of irons, however, display trends that are not in accord with these systematics. The IIIAB meteorites show indifferent Pt isotope fractionation between solid and liquid metal, as do the non-magmatic IAB/IIICD complex and IIE group. Conversely, the IVB meteorites indicate preferential partitioning of isotopically heavy Pt into the solid metal, as well as exhibiting indifferent compatible behaviour of Pt during core crystallisation. This, and other distinct behaviour of Pt and Pt isotopes in some groups of iron meteorites may reflect unusual compositions of the liquid metal phase, for example high S contents. Regardless of these effects, the data acquired in this study indicate that the variability of Pt stable isotope compositions in iron meteorites is primarily driven by the isotopic fractionation that accompanies partitioning of Pt between solid and liquid metal. These fractionations are sufficient to overprint any earlier signatures that may have been recorded during elemental processing in the solar nebula.

Chapter 6

Summary

150 Chapter 6

The aim of this study was to use high-precision measurements of mass-independent and mass-dependent isotope effects to investigate conditions in the early solar nebula, including the composition and distribution of material, and the processes that occurred during formation of the earliest planetary bodies. This was achieved during the course of the study, as described in the preceding chapters and summarised in the following.

6.1 Implications for solar nebula evolution

In Chapter 2, analyses of nucleosynthetic Mo isotope anomalies in magmatic iron meteorites afford clear evidence for variable excesses in p- and r-process nuclides, and hence deficits in s-process nuclides, in agreement with several recent high-precision studies. Furthermore, the extensive measurements conducted in this study provide the most precise data for the broadest range of samples analysed to date. This unique dataset allows, for the first time, the resolution of decoupled p-process and r-process isotope effects. Both the magnitude of the excesses and the extent of decoupling are interpreted to vary systematically with heliocentric distance. Comparison of the Mo isotope data with data from studies of other elements (Ba, Cr, Hf, Ni, Os, Ru, Ti, W, Zr) suggest that the most likely cause of the Mo isotope variability is thermal processing and selective destruction/removal of unstable presolar components in the solar nebula before the iron meteorite parent bodies were accreted. The material closest to the Sun was more thermally processed than material further out, and hence experienced more destruction and removal of p- and r-nuclide host phases. Therefore, the magmatic iron meteorites, which formed further out from the Sun than the Earth, were accreted from material that lost less p- and r-process material, and thus they display s-nuclide deficits relative to terrestrial Mo. Further support for this interpretation is provided by the data presented in Chapter 3, which discusses the mass-dependent Mo stable isotope fractionations (δ98Mo) of the iron meteorites. A correlation between δ98Mo and the magnitude of the nucleosynthetic Mo isotope anomalies in magmatic iron meteorites was observed. In detail, with increasing heliocentric distance, the iron meteorites display lighter Mo isotope compositions (with lower δ98Mo) and smaller deficits of r- and p-process nuclides. This is in accord with the proposed thermal processing model, because in the hotter regions close to the Sun, the larger losses of r- and p-nuclide host phases were presumably accompanied by partial evaporation of Mo and concurrent preferential loss of isotopically light Mo.

Summary 151

Based on the thermal processing model, one would not expect to find nucleosynthetic Pt isotope anomalies in iron meteorites. This conclusion follows from the observation that the primary control is a decoupling of ‘light vs. heavy’ r-process nuclides, and thermal susceptibilities therein, rather than due to the ‘moderately vs. highly’ refractory nature of the element as had previously been proposed. Since Pt has ‘heavy’ r-process components, and Mo ‘light’, one would therefore not expect to see evidence of the same thermal processing that affected the Mo isotope compositions, even though Pt has a lower 50% condensation temperature than Mo. Indeed, a detailed investigation of mass-independent Pt isotope effects in iron meteorites (Chapter 4) found no such nucleosynthetic isotope anomalies. However, readily resolvable mass-independent Pt isotope effects were detected in many irons, but these anomalies appear to be entirely of cosmogenic origin, as a result of exposure to galactic cosmic rays. The data was compared to previously published models of neutron capture effects, which were found to provide good estimates for the effects on 194Pt, 195Pt, 196Pt and 198Pt. However, relative to this, the production of 192Pt due to neutron capture by 191Ir does not appear to be well captured by the models and this discrepancy may be related to the effect of different shielding depths.

6.2 Implications for planetary differentiation

Processes pertaining to the evolution and differentiation of the iron meteorite parent bodies are also reflected in the mass-dependent isotope variations investigated here. The iron meteorites formed by segregation of Fe-Ni metal on originally primitive parent bodies and subsequent fractional crystallisation of solid metal during cooling, and such processes are recorded by the stable isotope variations that were observed for Mo and Pt (Chapter 3 and Chapter 5). While magmatic iron meteorites provide fairly consistent δ98Mo values, it was observed that a few individual samples show clear deviations from the otherwise well-defined mean δ98Mo value of a particular group. Such deviations were primarily found for irons with compositions close to the Ni-rich end of a fractional crystallisation sequence. The anomalous δ98Mo values of such meteorites were interpreted to reflect metal–sulphide partitioning of Mo and associated isotope fractionation. Evidence for this comes from two observations. First, troilite (FeS) nodules exhibit a wide range of δ98Mo values, with both heavy and light Mo stable isotope compositions. Second, during the fractional crystallisation of magmatic iron meteorite parent body cores, the S content of the residual liquid metal increases, as S (like Ni)

152 Chapter 6 is incompatible in the solid metal. Segregation of sulphide phases and associated Mo partitioning and isotope fractionation is hence expected to occur predominantly late during fractional crystallisation of metal. Significant variability of δ98Mo was also observed for iron meteorites of the IAB/IIICD complex. For this group, the observed stable isotope effects are also posited to reflect segregation and crystallisation of S-rich melts on the parent body, as such processes are associated with metal–sulphide partitioning of Mo and concomitant Mo isotope fractionation. Measurements of Pt concentrations in iron meteorites revealed generally compatible behaviour of Pt during fractional crystallisation of the parent body cores (Chapter 5). Furthermore, stable isotope analyses showed that the lighter isotopes of Pt typically partition preferentially into the solid metal during fractional crystallisation. Such behaviour is perhaps surprising as equilibrium isotope fractionation is generally associated with preferential partitioning of the heavier isotopes into the phase with stronger bonding, which is expected to be the solid metal. Nevertheless, the data acquired in this study indicate the variability of Pt stable isotope compositions in iron meteorites is primarily driven by the isotopic fractionation that accompanies partitioning of Pt between solid and liquid metal. These fractionations are sufficient to overprint any earlier signatures that may have recorded elemental processing in the solar nebula.

6.3 Outlook and future work

Whilst the investigations that were carried out during the course of this study were able to provide important new constraints on the evolution of the solar nebula and the earliest planetary bodies, there is plenty of scope for further work. The most obvious next objective would be to investigate nucleosynthetic isotope anomalies for elements either side of the cut-off between ‘light vs. heavy’ r-process nuclides at Z = 56 and range of condensation temperatures (e.g., Ba, Cd, Nd, Yb). High-precision isotopic measurements of such elements would test the refined thermal processing model that was proposed here for the origin of nucleosynthetic isotope anomalies in bulk planetary bodies. In detail, one would expect similar s-process deficits to Mo for elements with Z ≤ 56, but no anomalies for elements with Z > 56 (as observed for Pt) regardless of condensation temperature. In addition, stable isotope measurements for elements with Z ≤ 56 are desirable, as such analyses would provide a key test for the proposed thermal processing model. Ideally, the elements investigated will display only negligible stable isotope fractionations associated

Summary 153 with internal parent body processes. In this case, measurements of stable isotope compositions are best placed to reveal the stable isotopic signatures inherited from the solar nebula, with no overprinting from parent body processes, unlike the Pt measurements here, from which no such inherited solar nebula signature could be resolved. This would allow correlations between nucleosynthetic isotope anomalies and stable isotope compositions to be resolved, as observed here for Mo and also expected for other elements.

Bibliography

Akram, W., Schönbächler, M., Bisterzo, S. and Gallino, R., 2015. Zirconium isotope evidence for the heterogeneous distribution of s-process materials in the solar system. Geochimica Et Cosmochimica Acta, 165, 484–500.

Akram, W., Schönbächler, M., Sprung, P. and Vogel, N., 2013. Zirconium-Hafnium Isotope Evidence from Meteorites for the Decoupled Synthesis of Light and Heavy Neutron- Rich Nuclei. Astrophysical Journal, 777, 169–180.

Akram, W. M. and Schönbächler, M., 2013. Zirconium isotope evidence for dust processing in the early solar nebula. 44th Lunar and Planetary Science Conference, The Woodlands, Texas. 2138.

Andreasen, R. and Sharma, M., 2006. Solar nebula heterogeneity in p-process samarium and neodymium isotopes. Science, 314, 806–809.

Andreasen, R. and Sharma, M., 2007. Mixing and homogenization in the early solar system: Clues from Sr, Ba, Nd, and Sm isotopes in meteorites. Astrophysical Journal, 665, 874–883.

Angeli, I., 2004. A consistent set of nuclear rms charge radii: properties of the radius surface R (N, Z). Atomic Data and Nuclear Data Tables, 87, 185–206.

Argast, D., Samland, M., Thielemann, F. K. and Qian, Y. Z., 2004. Neutron star mergers versus core-collapse supernovae as dominant r-process sites in the early Galaxy. Astronomy & Astrophysics, 416, 997–1011.

Arlandini, C., Kappeler, F., Wisshak, K., Gallino, R., Lugaro, M., Busso, M. and Straniero, O., 1999. Neutron capture in low-mass asymptotic giant branch stars: Cross sections and abundance signatures. Astrophysical Journal, 525, 886–900.

Armytage, R. M. G., Georg, R. B., Savage, P. S., Williams, H. M. and Halliday, A. N., 2011. Silicon isotopes in meteorites and planetary core formation. Geochimica Et Cosmochimica Acta, 75, 3662–3676.

Arnould, A. and Goriely, S., 2003. The p-process of stellar nucleosynthesis: astrophysics and nuclear physics status. Physics Reports-Review Section of Physics Letters, 384, 1–84.

156

Baraffe, I., Eleid, M. F. and Prantzos, N., 1992. The s-process in massive stars of variable composition. Astronomy & Astrophysics, 258, 357–367.

Bauswein, A., Goriely, S. and Janka, H. T., 2013. Systematics of Dynamical Mass Ejection, Nucleosynthesis, and Radioactively Powered Electromagnetic Signals from Neutron- Star Mergers. Astrophysical Journal, 773.

Becker, H., Dalpe, C. and Walker, R. J., 2002. High-precision Ru isotopic measurements by multi-collector ICP-MS. Analyst, 127, 775–780.

Becker, H. and Walker, R. J., 2003. Efficient mixing of the solar nebula from uniform Mo isotopic composition of meteorites. Nature, 425, 152–155.

Benedix, G. K., McCoy, T. J., Keil, K. and Love, S. G., 2000. A petrologic study of the IAB iron meteorites: Constraints on the formation of the IAB-Winonaite parent body. Meteoritics & Planetary Science, 35, 1127–1141.

Berglund, M. and Wieser, M. E., 2011. Isotopic compositions of the elements 2009 (IUPAC Technical Report). Pure and Applied Chemistry, 83, 397–410.

Bermingham, K. R., Walker, R. J. and Worsham, E. A., 2015. Refinement of the Mo-Ru isotope cosmic correlation using high precision Mo and Ru isotope data. 46th Lunar and Planetary Science Conference, The Woodlands, Texas. 1588.

Beyersdorf-Kuis, U., Ott, U. and Trieloff, M., 2015. Early cosmic ray irradiation of chondrules and prolonged accretion of primitive meteorites. Earth and Planetary Science Letters, 423, 13–23.

Bigeleisen, J., 1996. Nuclear size and shape effects in chemical reactions. Isotope chemistry of the heavy elements. Journal of the American Chemical Society, 118, 3676–3680.

Bisterzo, S., Gallino, R., Kappeler, F., Wiescher, M., Imbriani, G., Straniero, O., Cristallo, S., Gorres, J. and deBoer, R. J., 2015. The branchings of the main s-process: their sensitivity to alpha-induced reactions on C-13 and Ne-22 and to the uncertainties of the nuclear network. Monthly Notices of the Royal Astronomical Society, 449, 506– 527.

157

Bisterzo, S., Gallino, R., Straniero, O., Cristallo, S. and Kappeler, F., 2011. The s-process in low-metallicity stars - II. Interpretation of high-resolution spectroscopic observations with asymptotic giant branch models. Monthly Notices of the Royal Astronomical Society, 418, 284–319.

Bogard, D. D., Garrison, D. H. and McCoy, T. J., 2000. Chronology and petrology of silicates from IIE iron meteorites: Evidence of a complex parent body evolution. Geochimica Et Cosmochimica Acta, 64, 2133–2154.

Boss, A. P., 2004. Convective cooling of protoplanetary disks and rapid giant planet formation. Astrophysical Journal, 610, 456–463.

Boss, A. P., 2008. Mixing in the solar nebula: Implications for isotopic heterogeneity and large-scale transport of refractory grains. Earth and Planetary Science Letters, 268, 102–109.

Boss, A. P. and Goswami, J. N., 2006. Presolar cloud collapse and the formation and early evolution of the solar nebula. In: Meteorites and the Early Solar System II. Lauretta, D. S. & Mcsween Jr., H. Y. (eds.). Tuscon, Arizona, University of Arizona Press. 171–186.

Brennecka, G. A., Borg, L. E. and Wadhwa, M., 2011. Barium, neodymium & samarium isotopic composition of CAIs: nucleosynthetic anomalies? 42nd Lunar and Planetary Science Conference, The Woodlands, Texas. 9036.

Bridgestock, L. J., Williams, H., Rehkämper, M., Larner, F., Giscard, M. D., Hammond, S., Coles, B. J., Andreasen, R., Wood, B. J., Theis, K. J., Smith, C. L., Benedix, G. K. and Schönbächler, M., 2014. Unlocking the zinc isotope systematics of iron meteorites. Earth and Planetary Science Letters, 400, 153–164.

Buchwald, V. F., 1975. Handbook of Iron Meteorites, Berkeley, University of California Press.

Burbidge, E. M., Burbidge, G. R., Fowler, W. A. and Hoyle, F., 1957. Synthesis of the Elements in Stars. Reviews of Modern Physics, 29, 547–650.

158

Burkhardt, C., Hin, R. C., Kleine, T. and Bourdon, B., 2014. Evidence for Mo isotope fractionation in the solar nebula and during planetary differentiation. Earth and Planetary Science Letters, 391, 201–211.

Burkhardt, C., Kleine, T., Dauphas, N. and Wieler, R., 2012a. Nucleosynthetic tungsten isotope anomalies in acid leachates of the Murchison chondrite: implications for hafnium-tungsten chronometry. Astrophysical Journal Letters, 753, 6.

Burkhardt, C., Kleine, T., Dauphas, N. and Wieler, R., 2012b. Origin of isotopic heterogeneity in the solar nebula by thermal processing and mixing of nebular dust. Earth and Planetary Science Letters, 357, 298–307.

Burkhardt, C., Kleine, T., Oberli, F. and Bourdon, B., 2009. Search for mass-independent Molybdenum isotope anomalies in iron meteorites. 40th Lunar and Planetary Science Conference, The Woodlands, Texas. 2482.

Burkhardt, C., Kleine, T., Oberli, F., Pack, A., Bourdon, B. and Wieler, R., 2011. Molybdenum isotope anomalies in meteorites: Constraints on solar nebula evolution and origin of the Earth. Earth and Planetary Science Letters, 312, 390–400.

Campbell, A. J. and Humayun, M., 2005. Compositions of group IVB iron meteorites and their parent melt. Geochimica Et Cosmochimica Acta, 69, 4733–4744.

Carlson, R. W., Boyet, M. and Horan, M., 2007. Chondrite barium, neodymium, and samarium isotopic heterogeneity and early earth differentiation. Science, 316, 1175– 1178.

Chabot, N. L., 2004. Sulfur contents of the parental metallic cores of magmatic iron meteorites. Geochimica Et Cosmochimica Acta, 68, 3607–3618.

Chabot, N. L. and Drake, M. J., 1999. Crystallization of magmatic iron meteorites: The role of mixing in the molten core. Meteoritics & Planetary Science, 34, 235–246.

Chabot, N. L. and Jones, J. H., 2003. The parameterization of solid metal-liquid metal partitioning of siderophile elements. Meteoritics & Planetary Science, 38, 1425–1436.

Chacko, T., Cole, D. R. and Horita, J., 2001. Equilibrium oxygen, hydrogen and carbon isotope fractionation factors applicable to geologic systems. In: Stable Isotope

159

Geochemistry. Valley, J. W. & Cole, D. R. (eds.). Washington, DC, Mineralogical Society of America. 1–81.

Chambers, J. E., 2004. Planetary accretion in the inner Solar System. Earth and Planetary Science Letters, 223, 241–252.

Chambers, J. E., 2006. Meteoritic Diversity and Planetesimal Formation. In: Meteorites and the Early Solar System II. Lauretta, D. S. & Mcsween Jr., H. Y. (eds.). Tuscon, Arizona, University of Arizona Press. 487–497.

Chen, H. W., Lee, T., Lee, D. C., Shen, J. J. S. and Chen, J. C., 2011. Ca-48 heterogeneity in differentiated meteorites. Astrophysical Journal Letters, 743.

Chen, J. H., Papanastassiou, D. A. and Wasserburg, G. J., 2010. Ruthenium endemic isotope effects in chondrites and differentiated meteorites. Geochimica Et Cosmochimica Acta, 74, 3851–3862.

Chernonozhkin, S. M., Goderis, S., Bauters, S., Vekemans, B., Vincze, L., Claeys, P. and Vanhaecke, F., 2014. Evaluation of pneumatic nebulization and ns-laser ablation ICP- MS for bulk elemental analysis and 2-dimensional element mapping of iron meteorites. Journal of Analytical Atomic Spectrometry, 29, 1001–1016.

Ciesla, F. J., 2009. Two-dimensional transport of solids in viscous protoplanetary disks. Icarus, 200, 655–671.

Clayton, D. D., 1982. Cosmic chemical memory - a new astronomy. Quarterly Journal of the Royal Astronomical Society, 23, 174–212.

Cole, D. R. and Chakraborty, S., 2001. Rates and mechanisms of isotopic exchange. In: Stable Isotope Geochemistry. Valley, J. W. & Cole, D. R. (eds.). Washington, DC, Mineralogical Society of America. 83–223.

Cottrell, E., Walter, M. and Walker, D., 2009. Metal-silicate partitioning of tungsten at high pressure and temperature: Implications for equilibrium core formation in Earth. Earth and Planetary Science Letters, 281, 275–287.

160

Cowan, J. J., Burris, D. L., Sneden, C., McWilliam, A. and Preston, G. W., 1995. Evidence of Heavy Element Nucleosynthesis Early in the History of the Galaxy - the Ultra-Metal- Poor Star Cs-22892-052. Astrophysical Journal, 439, L51–L54.

Creech, J., Baker, J., Handler, M., Schiller, M. and Bizzarro, M., 2013. Platinum stable isotope ratio measurements by double-spike multiple collector ICPMS. Journal of Analytical Atomic Spectrometry, 28, 853–865.

Creech, J. B., Baker, J. A., Handler, M. R. and Bizzarroc, M., 2014a. Platinum stable isotope analysis of geological standard reference materials by double-spike MC-ICPMS. Chemical Geology, 363, 293–300.

Creech, J. B., Handler, M. R., Baker, J. A. and Bizzarro, M., 2014b. Platinum stable isotope tracing of Earth's accretion and differentiation. 77th Annual Meteoritical Society Meeting, Casablanca, Morocco. 5124.

Criss, R. E., 1999. Principles of Stable Isotope Distribution, New York, Oxford University Press.

Dauphas, N., Chen, J. H., Zhang, J., Papanastassiou, D. A., Davis, A. M. and Travaglio, C., 2014. Calcium-48 isotopic anomalies in bulk chondrites and achondrites: Evidence for a uniform isotopic reservoir in the inner protoplanetary disk. Earth and Planetary Science Letters, 407, 96–108.

Dauphas, N., Davis, A. M., Marty, B. and Reisberg, L., 2004. The cosmic molybdenum- ruthenium isotope correlation. Earth and Planetary Science Letters, 226, 465–475.

Dauphas, N., Marty, B. and Reisberg, L., 2002. Molybdenum evidence for inherited planetary scale isotope heterogeneity of the protosolar nebula. Astrophysical Journal, 565, 640– 644.

Dauphas, N., Remusat, L., Chen, J. H., Roskosz, M., Papanastassiou, D. A., Stodolna, J., Guan, Y., Ma, C. and Eiler, J. M., 2010. Neutron-rich chromium isotope anomalies in supernova nanoparticles. Astrophysical Journal, 720, 1577–1591.

Dullemond, C. P., Hollenbach, D., Kamp, I. and D'Alessio, P., 2007. Models of the Structure and Evolution of Protoplanetary Disks. In: Protostars and Planets V. Reipurth, B., Jewitt, D. & Keil, K. (eds.). Tuscon, University of Arizona Press. 555–572.

161

Esbensen, K. H., Buchwald, V. F., Malvin, D. J. and Wasson, J. T., 1982. Systematic compositional variations in the Cape York meteorite. Geochimica Et Cosmochimica Acta, 46, 1913–1920.

Farouqi, K., Kratz, K. L., Pfeiffer, B., Rauscher, T., Thielemann, F. K. and Truran, J. W., 2010. Charged-Particle and Neutron-Capture Processes in the High-Entropy Wind of Core-Collapse Supernovae. Astrophysical Journal, 712, 1359–1377.

Fischer-Gödde, M., Burkhardt, C. and Kleine, T., 2012. Ruthenium isotope anomalies in meteorites and the cosmic Mo-Ru correlation. 43rd Lunar and Planetary Science Conference, The Woodlands, Texas. 2492.

Fischer-Gödde, M., Burkhardt, C., Kruijer, T. S. and Kleine, T., 2015. Ru isotope heterogeneity in the solar protoplanetary disk. Geochimica Et Cosmochimica Acta, 168, 151–171.

Fisker, J. L., Hoffman, R. D. and Pruet, J., 2009. On the origin of the lightest molybdenum isotopes. Astrophysical Journal Letters, 690, L135–L139.

Fujii, T., Moynier, F. and Albarede, F., 2006a. Nuclear field vs. nucleosynthetic effects as cause of isotopic anomalies in the early Solar System. Earth and Planetary Science Letters, 247, 1–9.

Fujii, T., Moynier, F., Telouk, P. and Albarede, F., 2006b. Mass-independent isotope fractionation of molybdenum and ruthenium and the origin of isotopic anomalies in Murchison. Astrophysical Journal, 647, 1506–1516.

Galer, S. J. G., 1999. Optimal double and triple spiking for high precision lead isotopic measurement. Chemical Geology, 157, 255–274.

Gallino, R., Arlandini, C., Busso, M., Lugaro, M., Travaglio, C., Straniero, O., Chieffi, A. and Limongi, M., 1998. Evolution and nucleosynthesis in low-mass asymptotic giant branch stars. II. Neutron capture and the s-process. Astrophysical Journal, 497, 388– 403.

Gault-Ringold, M. and Stirling, C. H., 2012. Anomalous isotopic shifts associated with organic resin residues during cadmium isotopic analysis by double spike MC-ICPMS. Journal of Analytical Atomic Spectrometry, 27, 449–459.

162

Goldberg, T., Gordon, G., Izon, G., Archer, C., Pearce, C. R., McManus, J., Anbar, A. D. and Rehkämper, M., 2013. Resolution of inter-laboratory discrepancies in Mo isotope data: an intercalibration. Journal of Analytical Atomic Spectrometry, 28, 724–735.

Goldstein, J. I., Scott, E. R. D. and Chabot, N. L., 2009. Iron meteorites: Crystallization, thermal history, parent bodies, and origin. Chemie Der Erde-Geochemistry, 69, 293– 325.

Goldstein, J. I., Yang, J. J. and Scott, E. R. D., 2014. Determining cooling rates of iron and stony-iron meteorites from measurements of Ni and Co at kamacite-taenite interfaces. Geochimica Et Cosmochimica Acta, 140, 297–320.

Grady, M. M., 2000. Catalogue of Meteorites, 5th Edition. Cambridge, Cambridge University Press.

Greber, N. D., Puchtel, I. S., Nagler, T. F. and Mezger, K., 2015. Komatiites constrain molybdenum isotope composition of the Earth's mantle. Earth and Planetary Science Letters, 421, 129–138.

Haack, H. and Scott, E. R. D., 1992. Asteroid core crystallization by inward dendritic growth. Journal of Geophysical Research-Planets, 97, 14727–14734.

Halliday, A. and Kleine, T., 2006. Meteorites and the Timing, Mechanisms, and Conditions of Terrestrial Planet Accretion and Early Differentiation. In: Meteorites and the Early Solar System II. Lauretta, D. S. & Mcsween Jr., H. Y. (eds.). Tuscon, Arizona, University of Arizona Press. 775–801.

Hayakawa, T., Iwamoto, N., Kajino, T., Shizuma, T., Umeda, H. and Nomoto, K., 2008. Empirical abundance scaling laws and implications for the gamma process in core- collapse supernovae. Astrophysical Journal, 685, 1089–1102.

Herzog, G. F., 2007. Cosmic-Ray Exposure Ages of Meteorites. In: Treatise on Geochemistry. Holland, H. D. & Turekian, K. K. (eds.). Oxford, Pergamon. 1–36.

Hin, R. C., Burkhardt, C., Schmidt, M. W., Bourdon, B. and Kleine, T., 2013. Experimental evidence for Mo isotope fractionation between metal and silicate liquids. Earth and Planetary Science Letters, 379, 38–48.

163

Hirai, Y., Ishimaru, Y., Saitoh, T. R., Fujii, M. S., Hidaka, J. and Kajino, T., 2015. Enrichment of r-process elements in dwarf spheroidal galaxies in chemo-dynamical evolution model. Astrophysical Journal, 814.

Hunt, A. C., Ek, M. and Schönbächler, M., 2014. The effect of cosmic ray irradiation on platinum isotopes in iron meteorites. 77th Annual Meteoritical Society Meeting, Casablanca, Morocco. 5260.

Hunt, A. C., Ek, M. and Schönbächler, M., 2015. The effect of cosmic ray irradiation on platinum and palladium isotopes in iron meteorites. 46th Lunar and Planetary Science Conference, The Woodlands, Texas. 1835.

Jacobsen, S. B. and Ranen, M. C., 2006. Barium isotopes and incomplete mixing in the solar nebula. Geochimica Et Cosmochimica Acta, 70, A286–A286.

Just, O., Bauswein, A., Pulpillo, R. A., Goriely, S. and Janka, H. T., 2015. Comprehensive nucleosynthesis analysis for ejecta of compact binary mergers. Monthly Notices of the Royal Astronomical Society, 448, 541–567.

Kimura, K., Lewis, R. S. and Anders, E., 1974. Distribution of Gold and Rhenium between Nickel-Iron and Silicate Melts - Implications for Abundance of Siderophile Elements on Earth and Moon. Geochimica Et Cosmochimica Acta, 38, 683–701.

Kracher, A., 1985. The Evolution of Partially Differentiated Planetesimals - Evidence from Iron Meteorite Groups IAB and IIICD. Journal of Geophysical Research, 90, C689– C698.

Kratz, K. L., Farouqi, K., Mashonkina, L. I. and Pfeiffer, B., 2008. Nucleosynthesis modes in the high-entropy-wind of type II supernovae. New Astronomy Reviews, 52, 390–395.

Krot, A. N., Keil, K., Scott, E. R. D., Goodrich, C. A. and Weisberg, M. K., 2007. Classification of Meteorites. In: Treatise on Geochemistry. Holland, H. D. & Turekian, K. K. (eds.). Oxford, Pergamon. 1–52.

Kruijer, T. S., Fischer-Gödde, M., Kleine, T., Sprung, P., Leya, I. and Wieler, R., 2013. Neutron capture on Pt isotopes in iron meteorites and the Hf-W chronology of core formation in planetesimals. Earth and Planetary Science Letters, 361, 162–172.

164

Kruijer, T. S., Touboul, M., Fischer-Gödde, M., Bermingham, K. R., Walker, R. J. and Kleine, T., 2014. Protracted core formation and rapid accretion of protoplanets. Science, 344, 1150–1154.

Lambert, D. L., 1992. The Para-Nuclei – Abundances and Origins. Astronomy and Astrophysics Review, 3, 201–256.

Lee, T., Shu, F. H., Shang, H., Glassgold, A. E. and Rehm, K. E., 1998. Protostellar cosmic rays and extinct radioactivities in meteorites. Astrophysical Journal, 506, 898–912.

Leya, I. and Masarik, J., 2013. Thermal neutron capture effects in radioactive and stable nuclide systems. Meteoritics & Planetary Science, 48, 665–685.

Liang, Y.-H., Siebert, C., Fitton, J. G., Burton, K. W. and Halliday, A. N., 2013. Molybdenum Isotope Fractionation in the Mantle. Mineralogical Magazine, 77, 1607.

Liu, M. H. and Fleet, M. E., 2001. Partitioning of siderophile elements (W, Mo, As, Ag, Ge, Ga, and Sn) and Si in the Fe-S system and their fractionation in iron meteorites. Geochimica Et Cosmochimica Acta, 65, 671–682.

Lodders, K., 2003. Solar system abundances and condensation temperatures of the elements. Astrophysical Journal, 591, 1220–1247.

Lu, Q. and Masuda, A., 1994. The Isotopic Composition and Atomic-Weight of Molybdenum. International Journal of Mass Spectrometry and Ion Processes, 130, 65–72.

Lugaro, M., Heger, A., Osrin, D., Goriely, S., Zuber, K., Karakas, A. I., Gibson, B. K., Doherty, C. L., Lattanzio, J. C. and Ott, U., 2014. Stellar origin of the Hf-182 cosmochronometer and the presolar history of solar system matter. Science, 345, 650– 653.

Lyons, J. R. and Young, E. D., 2005. CO self-shielding as the origin of oxygen isotope anomalies in the early solar nebula. Nature, 435, 317–320.

Markowski, A., Quitte, G., Halliday, A. N. and Kleine, T., 2006. Tungsten isotopic compositions of iron meteorites: Chronological constraints vs. cosmogenic effects. Earth and Planetary Science Letters, 242, 1–15.

165

Masarik, J. and Reedy, R. C., 1994. Rates for Neutron-Capture Reactions on Tungsten Isotopes in Iron-Meteorites. Meteoritics, 29, 497–498.

McCoy, T. J., Walker, R. J., Goldstein, J. I., Yang, J., McDonough, W. F., Rumble, D., Chabot, N. L., Ash, R. D., Corrigan, C. M., Michael, J. R. and Kotula, P. G., 2011. Group IVA irons: New constraints on the crystallization and cooling history of an asteroidal core with a complex history. Geochimica Et Cosmochimica Acta, 75, 6821– 6843.

McDonough, W. F., 2003. Compositional model for the Earth's core. In: The Mantle and Core. Carlson, R. W. (ed.) Oxford, Elsevier-Permagon.

Meyer, B. S. and Clayton, D. D., 2015. Sizes of carbon grains condensing in SNII shells. Meteoritics & Planetary Science, 50.

Moynier, F., Agranier, A., Hezel, D. C. and Bouvier, A., 2010. Sr stable isotope composition of Earth, the Moon, Mars, Vesta and meteorites. Earth and Planetary Science Letters, 300, 359–366.

Moynier, F., Yin, Q.-Z. and Schauble, E., 2011. Isotopic Evidence of Cr Partitioning into Earth’s Core. Science, 331, 1417–1420.

Murphy, K., Rehkämper, M., Kreissig, K., Coles, B. and van de Flierdt, T., 2016. Improvements in Cd stable isotope analysis achieved through use of liquid–liquid extraction to remove organic residues from Cd separates obtained by extraction chromatography. J. Anal. At. Spectrom., 31, 319–327.

Niederer, F. R., Papanastassiou, D. A. and Wasserburg, G. J., 1985. Absolute Isotopic Abundances of Ti in Meteorites. Geochimica Et Cosmochimica Acta, 49, 835–851.

Nishimura, N., Fischer, T., Thielemann, F. K., Frohlich, C., Hempel, M., Kappeli, R., Martinez-Pinedo, G., Rauscher, T., Sagert, I. and Winteler, C., 2012. Nucleosynthesis in Core-Collapse Supernova Explosions Triggered by a Quark-Hadron Phase Transition. Astrophysical Journal, 758.

Norton, O. R., 2002. The Cambridge Encyclopedia of Meteorites, Cambridge, UK, Cambridge University Press.

166

Palme, H. and O'Neill, H. S. C., 2003. Cosmochemical estimates of mantle composition. In: The Mantle and Core. Carlson, R. W. (ed.) Oxford, Elsevier-Permagon.

Pearce, C. R., Cohen, A. S. and Parkinson, I. J., 2009. Quantitative Separation of Molybdenum and Rhenium from Geological Materials for Isotopic Determination by MC-ICP-MS. Geostandards and Geoanalytical Research, 33, 219–229.

Pepin, R. O. and Becker, R. H., 1982. Nitrogen Isotopes in Iron-Meteorites. Meteoritics, 17, 269–269.

Petaev, M. I. and Jacobsen, S. B., 2004. Differentiation of metal-rich meteoritic parent bodies: I. Measurements of PGEs, Re, Mo, W, and Au in meteoritic Fe-Ni metal. Meteoritics & Planetary Science, 39, 1685–1697.

Pruet, J., Hoffman, R. D., Woosley, S. E., Janka, H. T. and Buras, R., 2006. Nucleosynthesis in early supernova winds. II. The role of neutrinos. Astrophysical Journal, 644, 1028– 1039.

Qian, Y. Z. and Wasserburg, G. J., 2007. Where, oh where has the r-process gone? Physics Reports-Review Section of Physics Letters, 442, 237–268.

Qin, L. P., Carlson, R. W. and Alexander, C. M. O., 2011. Correlated nucleosynthetic isotopic variability in Cr, Sr, Ba, Sm, Nd and Hf in Murchison and QUE 97008. Geochimica Et Cosmochimica Acta, 75, 7806–7828.

Ranen, M. C. and Jacobsen, S. B., 2006. Barium isotopes in chondritic meteorites: Implications for planetary reservoir models. Science, 314, 809–812.

Rauscher, T., Origin of the p-Nuclei in Explosive Nucleosynthesis. 11th Symposium on Nuclei in the Cosmos, 2010 Heidelberg, Germany.

Regelous, M., Elliott, T. and Coath, C. D., 2008. Nickel isotope heterogeneity in the early Solar System. Earth and Planetary Science Letters, 272, 330–338.

Rehkämper, M. and Halliday, A. N., 1997. Development and application of new ion-exchange techniques for the separation of the platinum group and other siderophile elements from geological samples. Talanta, 44, 663–672.

167

Ripperger, S. and Rehkämper, M., 2007. Precise determination of cadmium isotope fractionation in seawater by double spike MC-ICPMS. Geochimica Et Cosmochimica Acta, 71, 631–642.

Rudge, J. F., Reynolds, B. C. and Bourdon, B., 2009. The double spike toolbox. Chemical Geology, 265, 420–431.

Schatz, H., Aprahamian, A., Barnard, V., Bildsten, L., Cumming, A., Ouellette, M., Rauscher, T., Thielemann, F. K. and Wiescher, M., 2001. End point of the rp process on accreting neutron stars. Physical Review Letters, 86, 3471–3474.

Schatz, H., Aprahamian, A., Gorres, J., Wiescher, M., Rauscher, T., Rembges, J. F., Thielemann, F. K., Pfeiffer, B., Moller, P., Kratz, K. L., Herndl, H., Brown, B. A. and Rebel, H., 1998. rp-process nucleosynthesis at extreme temperature and density conditions. Physics Reports-Review Section of Physics Letters, 294, 167–263.

Schiller, M., Paton, C. and Bizzarro, M., 2015. Evidence for nucleosynthetic enrichment of the protosolar molecular cloud core by multiple supernova events. Geochimica Et Cosmochimica Acta, 149, 88–102.

Schönbächler, M., Akram, W. M., Williams, N. H. and Leya, I., 2011. Nucleosynthetic heterogeneities of neutron–rich isotopes in calcium aluminium–rich inclusions and bulk solar system materials. Formation of the First Solids in the Solar System, Kauai, Hawai'i.

Schönbächler, M., Lee, D. C., Rehkämper, M., Halliday, A. N., Fehr, M. A., Hattendorf, B. and Gunther, D., 2003. Zirconium isotope evidence for incomplete admixing of r- process components in the solar nebula. Earth and Planetary Science Letters, 216, 467–481.

Schönbächler, M., Rehkämper, M., Halliday, A. N., Lee, D. C., Bourot-Denise, M., Zanda, B., Hattendorf, B. and Gunther, D., 2002. Niobium-zirconium chronometry and early solar system development. Science, 295, 1705–1708.

Schönbächler, M., Rehkämper, M., Lee, D. C. and Halliday, A. N., 2004. Ion exchange chromatography and high precision isotopic measurements of zirconium by MC-ICP- MS. Analyst, 129, 32–37.

168

Scott, E. R. D., Wasson, J. T. and Buchwald, V. F., 1973. Chemical classification of iron- meteorites. VII. Reinvestigation of irons with Ge concentrations between 25 and 80 ppm. Geochimica Et Cosmochimica Acta, 37, 1957–1976.

Siebert, C., Nagler, T. F. and Kramers, J. D., 2001. Determination of molybdenum isotope fractionation by double-spike multicollector inductively coupled plasma mass spectrometry. Geochemistry Geophysics Geosystems, 2, art. no.–2000GC000124.

Siebert, C., Nagler, T. F., von Blanckenburg, F. and Kramers, J. D., 2003. Molybdenum isotope records as a potential new proxy for paleoceanography. Earth and Planetary Science Letters, 211, 159–171.

Simon, J. I., DePaolo, D. J. and Moynier, F., 2009. Calcium Isotope Composition of Meteorites, Earth, and Mars. Astrophysical Journal, 702, 707–715.

Sprung, P., Scherer, E. E., Upadhyay, D., Leya, I. and Mezger, K., 2010. Non-nucleosynthetic heterogeneity in non-radiogenic stable Hf isotopes: Implications for early solar system chronology. Earth and Planetary Science Letters, 295, 1–11.

Steele, R. C. J., Coath, C. D., Regelous, M., Russell, S. and Elliott, T., 2012. Neutron-Poor Nickel Isotope Anomalies in Meteorites. Astrophysical Journal, 758.

Tanvir, N. R., Levan, A. J., Fruchter, A. S., Hjorth, J., Hounsell, R. A., Wiersema, K. and Tunnicliffe, R. L., 2013. A 'kilonova' associated with the short-duration gamma-ray burst GRB130603B. Nature, 500, 547–549.

Teng, F. Z., Dauphas, N. and Helz, R. T., 2008. Iron isotope fractionation during magmatic differentiation in Kilauea Iki Lava Lake. Science, 320, 1620–1622.

Travaglio, C., Gallino, R., Arnone, E., Cowan, J., Jordan, F. and Sneden, C., 2004. Galactic evolution of Sr, Y, and Zr: A multiplicity of nucleosynthetic processes. Astrophysical Journal, 601, 864–884.

Travaglio, C., Gallino, R., Rauscher, T., Roepke, F. K. and Hillebrandt, W., 2015. Testing the role of SNe Ia for galactic chemical evolution of p-nuclei with two-dimensional models and with s-process seeds at different metallicities. Astrophysical Journal, 799.

169

Travaglio, C., Ropke, F., Gallino, R. and Hillebrandt, W., 2011. Type Ia supernovae as sites of the p-process: two-dimensional models coupled to nucleosynthesis. Astrophysical Journal, 739.

Trinquier, A., Birck, J. L. and Allegre, C. J., 2007. Widespread Cr-54 heterogeneity in the inner solar system. Astrophysical Journal, 655, 1179–1185.

Trinquier, A., Elliott, T., Ulfbeck, D., Coath, C., Krot, A. N. and Bizzarro, M., 2009. Origin of Nucleosynthetic Isotope Heterogeneity in the Solar Protoplanetary Disk. Science, 324, 374–376.

Truran Jr, J. W. and Heger, A., 2003. Origin of the Elements. In: Treatise on Geochemistry. Holland, H. D. & Turekian, K. K. (eds.). Oxford, Pergamon. 1–15.

Tsujimoto, T. and Shigeyama, T., 2014a. Enrichment history of r-process elements shaped by a merger of neutron star pairs. Astronomy & Astrophysics, 565.

Tsujimoto, T. and Shigeyama, T., 2014b. The Origins of Light and Heavy R-Process Elements Identified by Chemical Tagging of Metal-Poor Stars. Astrophysical Journal Letters, 795.

Vogl, J. and Pritzkow, W., 2012. Reference Materials in Isotopic Analysis. In: Isotopic Analysis: Fundamentals and Applications Using ICP-MS. Vanhaecke, F. & Degryse, P. (eds.). Weinham, Wiley-VCH Verlag GmbH & Co. KGaA. 139–163.

Wade, J. and Wood, B., 2005. Core formation and the oxidation state of the Earth. Earth and Planetary Science Letters, 236, 78–95.

Walker, R. J., 2012. Evidence for homogeneous distribution of osmium in the protosolar nebula. Earth and Planetary Science Letters, 351, 36–44.

Walker, R. J., McDonough, W. F., Honesto, J., Chabot, N. L., Mccoy, T. J., Ash, R. D. and Bellucci, J. J., 2008. Modeling fractional crystallization of group IVB iron meteorites. Geochimica Et Cosmochimica Acta, 72, 2198–2216.

Wanajo, S., 2006. The rp-process in neutrino-driven winds. Astrophysical Journal, 647, 1323–1340.

170

Wanajo, S., 2013. The r-process in proto-neutron-star wind revisited. Astrophysical Journal Letters, 770.

Wanajo, S., Janka, H. T. and Kubono, S., 2011. Uncertainties in the vp-process: supernova dynamics versus nuclear physics. Astrophysical Journal, 729.

Wanajo, S., Sekiguchi, Y., Nishimura, N., Kiuchi, K., Kyutoku, K. and Shibata, M., 2014. Production of all the r-process nuclides in the dynamical ejecta of neutron star mergers. Astrophysical Journal Letters, 789.

Wang, M., Audi, G., Wapstra, A. H., Kondev, F. G., MacCormick, M., Xu, X. and Pfeiffer, B., 2012. The AME2012 atomic mass evaluation (II). Tables, graphs and references. Chinese Physics C, 36, 1603–2014.

Wasserburg, G. J., Busso, M. and Gallino, R., 1996. Abundances of actinides and short-lived nonactinides in the interstellar medium: Diverse supernova sources for the r-processes. Astrophysical Journal, 466, L109–L113.

Wasserburg, G. J. and Qian, Y. Z., 2009. Chemical Evolution of the Juvenile Universe. Publications of the Astronomical Society of Australia, 26, 184–193.

Wasson, J. T., 1969. The chemical classification of iron meteorites: III. Hexahedrites and other irons with germanium concentrations between 80 and 120 ppm. Geochimica Et Cosmochimica Acta, 22, 4799–4807.

Wasson, J. T., 1970. Chemical classification of iron meteorites. IV. Irons with Ge concentrations greater than 190 ppm and other meteorites associated with Group I. Icarus, 12, 407–423.

Wasson, J. T., 1974. Meteorites: Classification and Properties, New York, Springer.

Wasson, J. T., 1999. Trapped melt in IIIAB irons; solid/liquid elemental partitioning during the fractionation of the IIIAB magma. Geochimica Et Cosmochimica Acta, 63, 2875– 2889.

Wasson, J. T. and Kallemeyn, G. W., 2002. The IAB iron-meteorite complex: A group, five subgroups, numerous grouplets, closely related, mainly formed by crystal segregation in rapidly cooling melts. Geochimica Et Cosmochimica Acta, 66, 2445–2473.

171

Wasson, J. T. and Richardson, J. W., 2001. Fractionation trends among IVA iron meteorites: Contrasts with IIIAB trends. Geochimica Et Cosmochimica Acta, 65, 951–970.

Wasson, J. T. and Wang, J. M., 1986. A Nonmagmatic Origin of Group IIE Iron Meteorites. Geochimica Et Cosmochimica Acta, 50, 725–732.

White, W. M., 2013. Geochemistry, Oxford, Wiley-Blackwell.

Wieser, M. E. and De Laeter, J. R., 2009. Molybdenum isotope mass fractionation in iron meteorites. International Journal of Mass Spectrometry, 286, 98–103.

Williams, H. M. and Archer, C., 2011. Copper stable isotopes as tracers of metal-sulphide segregation and fractional crystallisation processes on iron meteorite parent bodies. Geochimica Et Cosmochimica Acta, 75, 3166–3178.

Williams, J. P., 2010. The astrophysical environment of the solar birthplace. Contemporary Physics, 51, 381–396.

Wittig, N., Humayun, M., Brandon, A. D., Huang, S. and Leya, I., 2013. Coupled W-Os-Pt isotope systematics in IVB iron meteorites: In situ neutron dosimetry for W isotope chronology. Earth and Planetary Science Letters, 361, 152–161.

Wombacher, F. and Rehkämper, M., 2003. Investigation of the mass discrimination of multiple collector ICP-MS using neodymium isotopes and the generalised power law. Journal of Analytical Atomic Spectrometry, 18, 1371–1375.

Wombacher, F., Rehkämper, M., Mezger, K., Bischoff, A. and Munker, C., 2008. Cadmium stable isotope cosmochemistry. Geochimica Et Cosmochimica Acta, 72, 646–667.

Worsham, E. A., Bermingham, K. R. and Walker, R., 2015. Cosmic ray exposure modification of the molybdenum isotopic composition of iron meteorites. AGU Fall Meeting 2015, San Fransico, California. 65206.

Worsham, E. A. and Walker, R. J., 2015. Molybdenum isotope evidence for diverse genetics among IAB iron meteorite complex subgroups. 46th Lunar and Planetary Science Conference, The Woodlands, Texas. 2524.

172

Xue, Z., Rehkämper, M., Horner, T. J., Abouchami, W., Middag, R., van de Flierdt, T. and de Baar, H. J. W., 2013. Cadmium isotope variations in the Southern Ocean. Earth and Planetary Science Letters, 382, 161–172.

Yang, J., Siebert, C., Barling, J., Savage, P., Liang, Y. H. and Halliday, A. N., 2015. Absence of molybdenum isotope fractionation during magmatic differentiation at Hekla volcano, Iceland. Geochimica Et Cosmochimica Acta, 162, 126–136.

Yin, Q. Z., Jacobsen, S. B. and Yamashita, K., 2002. Diverse supernova sources of pre-solar material inferred from molybdenum isotopes in meteorites. Nature, 415, 881–883.

Yokoyama, T., Alexander, C. M. O. and Walker, R. J., 2010. Osmium isotope anomalies in chondrites: Results for acid residues and related leachates. Earth and Planetary Science Letters, 291, 48–59.

Yokoyama, T., Alexander, C. M. O. and Walker, R. J., 2011. Assessment of nebular versus parent body processes on presolar components present in chondrites: Evidence from osmium isotopes. Earth and Planetary Science Letters, 305, 115–123.

Yokoyama, T., Rai, V. K., Alexander, M. O., Lewis, R. S., Carlson, R. W., Shirey, S. B., Thiernens, M. H. and Walker, R. J., 2007. Osmium isotope evidence for uniform distribution of s- and r-process components in the early solar system. Earth and Planetary Science Letters, 259, 567–580.

Yokoyama, T. and Walker, R. J., 2016. Nucleosynthetic Isotope Variations of Siderophile and Chalcophile Elements in the Solar System. Reviews in Mineralogy and Geochemistry, 81, 107–160.

Yorke, H. W., Bodenheimer, P. and Laughlin, G., 1993. The Formation of Protostellar Disks . I. 1 M☉. Astrophysical Journal, 411, 274–284.

Young, E. D., Galy, A. and Nagahara, H., 2002. Kinetic and equilibrium mass-dependent isotope fractionation laws in nature and their geochemical and cosmochemical significance. Geochimica Et Cosmochimica Acta, 66, 1095–1104.

Yu, T. H., Meyer, B. S. and Clayton, D. D., 2013. Dust Growth in Core-Collapse and Thermo-Nuclear Supernovae. Meteoritics & Planetary Science, 48, A386–A386.

173

Zhang, J., Davis, A. M. and Dauphas, N., 2012. Titanium Isotopic Compositions in Calcium- Alumium-Rich Inclusions. Meteoritics & Planetary Science, 47, A429–A429.

Zinner, E., 2007. Presolar Grains. In: Treatise on Geochemistry. Holland, H. D. & Turekian, K. K. (eds.). Oxford, Pergamon. 1–33.

175

Acknowledgements

First and foremost, I must express my deepest gratitude to my supervisor Professor Mark Rehkämper for his support and guidance throughout the project, and without whom the scientific development achieved by this research would not have been possible.

I would also especially like to thank Barry Coles for his extensive help with the dark art of mass spectrometry and its application to obtain high-precision isotope measurements. I am indebted to Dr Tatiana Goldberg and Dr Roland Stumpf for their crucial help in the development of the Mo and Pt methods respectively. Thanks must also go to Dr Katharina Kreissig and Dr Luke Bridgestock for laboratory training and preparation, and Katy Murphy for help with the processing of data reduction scripts. In addition, I would like to thank Rebekah Moore and David Poole for proofreading this thesis.

The Natural History Museum, London, is acknowledged for providing the majority of meteorite specimens used here. Particular thanks go to the meteorite curator Dr Caroline Smith and assistant meteorite curators Deborah Cassey and Natasha Almeida for help with selecting appropriate specimens.

Finally, I would like to thank the Science and Technology Facilities Council (STFC) for sponsoring the Departmental Scholarship through which my PhD studies were funded.