Touboul Thesis Completed

290
UNIVERSITE PARIS 7 - Denis DIDEROT UFR des Sciences Physiques de la Terre INSTITUT DE PHYSIQUE DU GLOBE DE PARIS DOCTORAT Géochimie Accrétion et différenciation des planétésimaux et des planètes telluriques: Contraintes apportées par le système 182 Hf- 182 W Accretion and differentiation of planetesimals and terrestrial planets: constraints from the 182 Hf- 182 W system Mathieu Touboul Thèse dirigée par: Bernard Bourdon Soutenance le 19 novembre 2008 en présence du jury composé de MM. : Max Schmidt, professeur (ETH Zürich) Président Francis Albarède, professeur (ENS Lyon) Rapporteur Alex Halliday, professeur (Université d’Oxford) Rapporteur Jean-Louis Birck, professeur (IPGP) Examinateur Bernard Bourdon, professeur (ETH Zürich) Directeur de thèse Thorsten Kleine, directeur de recherche (ETH Zürich) Directeur de thèse

description

UNIVERSITE PARIS 7 - Denis DIDEROT UFR des Sciences Physiques de la Terre INSTITUT DE PHYSIQUE DU GLOBE DE PARISDOCTORAT GéochimieAccrétion et différenciation des planétésimaux et des planètes telluriques: Contraintes apportées par le système 182Hf-182WAccretion and differentiation of planetesimals and terrestrial planets: constraints from the 182Hf-182W systemMathieu TouboulThèse dirigée par: Bernard BourdonSoutenance le 19 novembre 2008 en présence du jury composé de MM. : Max Schmidt

Transcript of Touboul Thesis Completed

Page 1: Touboul Thesis Completed

UNIVERSITE PARIS 7 - Denis DIDEROT UFR des Sciences Physiques de la Terre

INSTITUT DE PHYSIQUE DU GLOBE DE PARIS

DOCTORAT Géochimie

Accrétion et différenciation des planétésimaux et

des planètes telluriques: Contraintes apportées par le système 182Hf-182W

Accretion and differentiation of planetesimals and terrestrial planets:

constraints from the 182Hf-182W system

Mathieu Touboul

Thèse dirigée par:

Bernard Bourdon

Soutenance le 19 novembre 2008 en présence du jury composé de MM. : Max Schmidt, professeur (ETH Zürich) Président Francis Albarède, professeur (ENS Lyon) Rapporteur Alex Halliday, professeur (Université d’Oxford) Rapporteur Jean-Louis Birck, professeur (IPGP) Examinateur Bernard Bourdon, professeur (ETH Zürich) Directeur de thèse Thorsten Kleine, directeur de recherche (ETH Zürich) Directeur de thèse

Page 2: Touboul Thesis Completed

2

Front cover: Artist’s view of the giant impact. The hot debris ejected during this event re-accreted to form the Moon . Credit: James Garry, FastLight.

Page 3: Touboul Thesis Completed

3

Sommaire - List of contents

Résumé-Abstract …………………………………………………………………….. 11

Chapitre 1 Introduction - Introduction …………………………………………..15

1.1 Introduction - Intoduction……………………………………………….. 17

1.2 Le chronomètre 182Hf-182W à courte période 182Hf-182W short-lived chronometer…………….……………………….. 19

1.2.1 Notations - Notations…………………………………………..... 20

1.2.2 Ages Hf-W des météorites - Hf-W ages of meteorites…………... 20

1.2.2.1 Isochrone interne et age relatif Internal isochron and relative age…………………….. 20

1.2.2.2 Calibration sur une échelle de temps absolu Calibration onto an absolute timescale………………...22

1.2.2.3 Température de fermeture du système Closure temperature of the system…………………….. 24

1.2.3 Datation de la differenciation planétaire Hf-W ages of planetary differentiation……………...…………....26

1.2.3.1 Systématiques 182Hf-182W des chondrites et des objets planétaires totaux 182Hf-182W systematics of chondrites and bulk planetary objects ……………………………………... 26

1.2.3.2 Ages Hf-W de formation du noyau Hf-W ages of core formation……………………….…..27

1.2.3.3 Ages Hf-W de differentiation mantellique Hf-W ages of mantle differentiation…………………… 29

1.3 Contenu de la thèse - Contents of the thesis…………………..…………30

Références - Reference………………………….……………………………33

Page 4: Touboul Thesis Completed

4

Chapitre 2 Techniques analytiques – analytical techniques……………….…….39

2.1 Introduction - Introduction……………………………………………...41

2.2 Préparation des échantillons - Sample preparation.…...……………..... 42

2.2.1 Roches totales - Whole rocks…………………………………..... 42

2.2.2 Séparation magnétique - Magnetic separation…………………...42

2.2.3 Séparation des plagioclases d’anorthosites lunaires Plagioclase separation from lunar anorthosites…………….…...43

2.3 Procédures chimiques - Chemical procedures…………………...…….. 43

2.3.1 Procédure sur les métaux -Procedure for metals………...……… 44

2.3.2 Procédure sur des silicates et roches totales Procedure for silicate fractions and whole rocks……………….. 45

2.3.3 Procédure particulière développée pour les plagioclases d’anorthosites lunaires Procedure for plagioclases from lunar anorthosites……………. 47

2.4 Spectrométrie de masse - Mass spectrometry………………………….. 49

2.4.1 Protocole de mesure des compositions isotopiques du W Measurements of W isotopic composition…..…………………… 50

2.4.2 Discrimination de masse instrumentale et interférences Mass bias and interferences……..……...………………………..51

2.4.3 Reproductibilité externe et justesse des mesures External reproductibility and accuracy………....………………. 53

2.5 Mesure de concentrations par dilution isotopique

Concentration measurements by isotope dilution………….………..... 54

2.6 Conclusions - Conclusions……….……………………………………... 56

Références -References…………………………………………………….... 57

Page 5: Touboul Thesis Completed

5

Chapitre 3 Thermochronométrie Hf-W des météorites: Contraintes sur l’accrétion et l’évolution thermique des corps parents

Thermochonometry of meteorites : Constraints on the accretion anf thermal evolution of parent bodies…………................. 59

3.1 Accretion and thermal history of the acapulcoite-lodranite parent body inferred from Hf-W thermochronometry…………………………. 61

3.1.1 Introduction…………………………………………………….... 64

3.1.2 Samples and analytical techniques……………….……………… 65

3.1.2.1 Samples…………………………………………………65

3.1.2.2 Analytical techniques………………………………….. 68

3.1.3 Results…………………….……………………………………... 70

3.1.4 Discussion……………………………………………………….. 73

3.1.4.1 Hf-W isochron ages for acapulcoites and lodranites…...73

3.1.4.2 Closure temperature for the Hf-W system in acapulcoites and lodranites……………………………..75

3.1.4.3 Accretion and cooling history of the acapulcoite- lodranite parent body…………………………………...77

3.1.4.4 Bulk Hf-W systematics of the acapulcoite-lodranite parent body – nebular vs. parent body processes……… 82

3.1.5 Conclusions……………………………………………………… 84

References……………………………………..………………………… 87

3.2 Hf–W thermochronometry: Closure temperature and constraints on the accretion and cooling history of the H chondrite parent body..... 93

3.2.1 Introduction……………………………….……………………... 95

3.2.2 Analytical methods……..………………………………………...96

3.2.3 Results……………………………………..…………………….. 98

3.2.4 Discussion……………………………………………………….. 99

3.2.4.1 Hf-W isochron ages for H chondrites…………………. 99

3.2.4.2 Closure temperature for the Hf-W system in equilibrated H chondrites…………………..………….100

3.2.4.3 Significance of the Hf-W ages…..…………………….102

3.2.4.4 Constraints on the accretion and cooling history……...104

3.2.4.5 Hf-W fractionation among chondrite parent materials in the solar nebula…………………………..105

3.2.5 Conclusions…………….………………………………………. 105

References…………………………..………………………………….. 106

Page 6: Touboul Thesis Completed

6

Chapitre 4 Chronométrie Hf-W de la Lune: Contraintes sur l’impact

géant et la cristallisation de l’océan magmatique lunaire Hf-W chronometry of the Moon: Constraints on the giant impact and the crystallization of the lunar magma ocean................. 109

4.1 Late formation and prolonged differentiation of the Moon inferred from W isotopes in lunar metals……………………………. 111

4.1.1 Main text……………………………………………………….. 112

4.1.2 Methods summary……………………………………………… 121

4.1.3 Acknowledgements…………………………………………...... 122

4.1.4 References……………………………………………………… 122

4.1.5 Supplementary information……………………………………..126

4.1.5.1 Lunar metals…………………………………………..126

4.1.5.2 Analytical methods…………………………………....126

4.1.5.3 Cosmogenic effects on W isotope ratios in lunar samples………………………………………….127

4.1.5.3.1 Correction for W isotope data obtained in this study………………………………….127

4.1.5.3.2 Correction of W isotope data obtained in earlier studies………………………………..129

4.1.5.4 Hf/W fractionation in the crystallizing lunar magma ocean………………………………………….131

4.1.5.5 Supplementary references……………………………..132

4.2 182Hf-182W systematics of ferroan anorthosites and the lifetime of the lunar magma ocean……………………………………………..135

4.2.1 Introduction ................................................................................. 137

4.2.2 Samples and analytical methods.................................................. 139

4.2.3 Results ......................................................................................... 141

4.2.4 Discussion.................................................................................... 144

4.2.4.1 Tungsten isotope homogeneity in the Moon ................ 144

4.2.4.2 Duration of magma ocean solidification ...................... 146

4.2.5 Conclusions…………………………………………………...... 148

References……………………………………………………………… 148

Page 7: Touboul Thesis Completed

7

4.3 Early differentiation of the Earth and the Moon...…………………….151

4.3.1 Introduction ................................................................................. 153

4.3.2 New constraints on the age of the Moon and termination of Earth accretion......................................................................... 154

4.3.2.1 New 182W data and chronological implications............ 154

4.3.2.1.1 New 182W data and lunar mantle evolution ... 157

4.3.2.1.2 Exploring scenarios for Moon-Earth equilibration following the giant impact.. ..... 160

4.3.2.1.3 Isotope equilibration of Earth and Moon after the Giant impact ................................... 163

4.3.2.2 Chronology of terrestrial accretion and astronomical implications…. ........................................................... 166

4.3.3 Constraints on the early differentiation of the Earth …………... 168

4.3.3.1 The Age of Early Earth differentiation......................... 168

4.3.3.2 The Nd isotope composition of terrestrial planets........ 171

4.3.4 Conclusions ................................................................................. 178

References ............................................................................................... 179

Chapitre 5 Conclusions et perspectives - Conclusions and outlooks..………….185

Appendice 1 Systématiques Hf-W des eucrites cumulats et la chronologie du corps parent des eucrites Hf- W systematics of cumulate eucrites and the chronology of the eucrite parent body...…………………………………….……195

A1.1 Introduction .......................................................................................... 197

A1.2 Samples and analytical methods........................................................... 197

A1.3 Results .................................................................................................. 197

A1.4 Discussion............................................................................................. 198

A1.4.1 Evidence of W contamination in eucrites and chronological implications ................................................................................. 198

A1.4.2 Chronology of the eucrite parent body........................................ 198

A1.4.3 Origin of cumulate eucrites ......................................................... 198

A1.5 Conclusions .......................................................................................... 198

References ..................................................................................................... 198

Page 8: Touboul Thesis Completed

8

Appendice 2 Chronométrie Hf-W: Accrétion et évolution précoce des astéroïdes et des planètes telluriques Hf-W chronometry and accretion and evolution of asteroids and terrestrial planets...………………………………........ 199

Abstract ……………………………………………………………………. 201

A2.1 Introduction .......................................................................................... 203

A2.2 The 182Hf-182W chronometer ................................................................ 204

A2.2.1 Hf-W fractionation during planetary differentiation and in meteorites .................................................................................... 204

A2.2.2 Notation and Hf-W isotope systematics ...................................... 206

A2.2.3 Reference parameters for Hf-W chronology ............................... 208

A2.2.3.1 Hf-W systematics of chondrites ................................... 208

A2.2.3.2 Initial 182Hf/180Hf and 182W/184W of CAIs.................... 209

A2.2.3.3 Calibration of the 182Hf-182W chronometer and conversion to absolute ages............................................................. 211

A2.2.4 Closure temperature of the Hf-W system.................................... 212

A2.3 Timescales for the accretion and early evolution of planetesimals ...... 213

A2.3.1 Iron meteorites - remnants of the first planetesimals .................. 213

A2.3.2 Chronology of IAB-IIICD iron meteorites.................................. 216

A2.3.3 Chronology of the eucrite parent body........................................ 218

A2.3.3.1 Accretion and primordial differentiation...................... 218

A2.3.3.2 Magmatism and thermal metamorphism ...................... 219

A2.3.4 Timing of magmatism on the angrite and mesosiderite parent bodies........................................................................................... 221

A2.3.5 Chronology of the H chondrite parent body................................ 222

A2.3.6 Planetesimal accretion and evolution - the dominant role of 26Al heating ........................................................................................ 224

A2.4 Timescales for core formation and early mantle differentiation in Mars

A2.4.1 182W-142Nd systematics of the Martian mantle ............................ 226

A2.4.2 Age of the Martian core............................................................... 228

A2.4.3 Early mantle differentiation in Mars ........................................... 229

A2.5 Hf-W chronometry of the Moon........................................................... 230

A2.5.1 Cosmogenic vs. radiogenic 182W in lunar samples...................... 231

A2.5.2 Lifetime of the lunar magma ocean............................................. 233

Page 9: Touboul Thesis Completed

9

A2.6 Timescales for accretion and differentiation of Earth .......................... 235

A2.6.1 Models of core formation and metal-silicate equilibration ......... 235

A2.6.2 Hf-W systematics in continuous core formation models............. 239

A2.6.3 Tungsten model age for the Moon-forming impact..................... 243

A2.6.4 Early mantle differentiation and implications of a non-chondritic bulk Earth .................................................................................... 244

A2.7 Conclusions .......................................................................................... 246

References ..................................................................................................... 248

Figure captions .............................................................................................. 257

Tables………………………………………………………………………..261

Figures ........................................................................................................... 267

Remerciements - Acknowledgments…………………………………………....… 289

Page 10: Touboul Thesis Completed

10

Page 11: Touboul Thesis Completed

RESUME

Le chronomètre à courte période 182Hf-182W (T1/2 = 8,9 Ma) est un instrument

puissant d’investigation de l’évolution du système solaire primitif. Hf et W sont des

éléments réfractaires et sont supposés être présents en abondances chondritiques dans les

objets du système solaire. Toutefois, le W est un élément modérément sidérophile et très

incompatible tandis que l’Hf est strictement lithophile et également incompatible mais à

un degré moindre puisqu’il est préférentiellement incorporé dans l’ilménite et le

clinopyroxene. Ces différences de propriétés géochimiques permettent une application

variée de ce chronomètre, notamment pour contraindre les échelles de temps de

l’accrétion, du métamorphisme thermique et de la differentiation des planètes et des

planétésimaux.

Les isochrones internes réalisées sur les acapulcoites, les lodranites et les

chondrites H5 et H6 conduisent à des âges Hf-W de 5,1 ± 0,9, 5,6 ± 1,0, 5,9 ± 0,9 Ma et

9,6 ± 1,0 Ma après la formation des CAI respectivement. Des simulations numériques de

la diffusion du W montrent que les températures de fermeture du système Hf-W dans ces

différentes météorites sont proches des températures maximales atteintes lors du pic

thermique du métamorphisme affectant leurs corps parents. La connaissance de

l’évolution thermique à haute température de ces planétésimaux, en conjonction avec un

modèle thermique de refroidissement d’objets chauffés par la désintégration de l’26Al,

permet de définir avec plus de précision leur âge d’accrétion. La séquence chronologique

ainsi obtenue révèle que l’accrétion du corps parent des chondrites H (entre ∼2 et ∼4 Ma

après la formation du système solaire) est plus récente que celle du corps parent des

acapulcoites et des lodranites (entre ∼1,5 et ∼2 Ma après la formation du système solaire),

elle-même postérieure à l’accrétion des planétésimaux différenciés (moins d’∼1 Ma la

formation du système solaire). Elle suggère que le degré de différenciation des

planétésimaux est essentiellement contrôlé par l’abondance de 26Al incorporé lors

l’accrétion donc par leur temps d’accrétion et que le chauffage par impacts n’a joué qu’un

rôle secondaire. Nos résultats montrent également que les corps parents des chondrites H

et des acapulcoites-lodranites ont des rapports Hf/W similaires (0,63 ± 0,20) mais

significativement inférieurs au rapport Hf/W des chondrites carbonées (1,21 ± 0,06).

11

Page 12: Touboul Thesis Completed

Cette différence, inattendue pour des éléments réfractaires, révèle un fractionnement Hf-

W au cours des deux premiers Ma du système solaire qui pourrait être produit lors de

processus au sein de la nébuleuse avant ou durant l’accrétion des planétésimaux.

Contrairement aux résultats des études précédentes, nos nouvelles données

isotopiques du W obtenues pour une série de métaux lunaires et deux anorthosites

lunaires montrent que les produits de la cristallisation de l’océan magmatique lunaire ont

tous une composition isotopique identique, dont la valeur est similaire à celle du manteau

terrestre. Nous avons montré que les anomalies de 182W rapportées précédémment sont

d’origine cosmogénique, et sont générées par la production de 182W à partir de 181Ta par

capture neutronique, liée à l’intense exposition de la surface lunaire aux rayonnements

cosmiques. En revanche, les compositions isotopiques déterminées dans notre étude sont

dépourvues de composant cosmogénique : les fractions de métaux lunaires obtenues sont

suffisamment pures et ne contiennent pas de tantale; les anorthosites sélectionnées ont

des âges d’exposition très jeunes (< 2 Ma) et des fractions pures de plagioclases en ont

été extraites. La mise en évidence de signatures isotopiques identiques dans les produits

de l’océan magmatique lunaire en dépit de leurs rapports Hf/W différents conduit à une

révision importante de l’âge Hf-W de différenciation de la Lune : prédemment estimé à

~30Ma (Kleine et al., 2005), celle-ci est en réalité intervenue plus de 60Ma après la

formation du système solaire. De plus, la similarité des compositions isotopiques des

manteaux terrestre et lunaire indiquent que la formation de la Lune a dû avoir lieu le plus

vraisemblablement à après la formation du système solaire et suggére une

rééquilibration des isotopes du tungstène entre l’océan magmatique terrestre et le disque

proto-lunaire. L’impact géant à son origine étant le dernier événement majeur de

l’accrétion terrestre, cet âge correspond à l’âge de la fin de l’accrétion de la Terre. En

accord avec les prédictions des modèles dynamiques, il constitue, à l’heure actuelle, la

meilleure estimation du temps qui a été nécessaire pour évoluer d’un disque de poussières

et de gaz à un système planétaire.

Ma62 1090

−+

Mots clés : 182Hf, 182W, radioactivité éteinte, chronometrie, système solaire primitif,

planétésimal, planète, Lune, météorite, accrétion, différenciation, impact géant, océan de

magma, équilibration isotopique, âge de la Terre.

12

Page 13: Touboul Thesis Completed

ABSTRACT

The short lived 182Hf-182W chronometer (T1/2 = 8.9 Ma) is a powerful tool for

investigating the evolution of the early solar system. Hf and W are refractory elements

and are supposed to be present in chondritic abundances in the solar system objects.

However, W is a moderately siderophile and very incompatible, while Hf is strictly

lithophile and also incompatible but to a lesser degree than W, because it is preferentially

incorporated in ilmenite and clinopyroxene. These differences in geochemical properties

allow to widely apply Hf-W chronometry, in particular to constrain the timescales of

accretion, thermal metamorphism and differentiation of planets and planetesimals.

The internal isochrons for acapulcoites, lodranites and H5 and H6 chondrites lead

to Hf-W ages of 5.1 ± 0.9, 5.6 ± 1.0, 5.9 ± 0.9 Ma and 9.6 ± 1.0 Ma after the formation of

CAI respectively. Numerical simulations of the diffusion of W show that the closure

temperatures of the Hf-W system in these meteorites are close to the maximum

temperatures reached during the peak of thermal metamorphism, affecting their parent

bodies. The knowledge of thermal evolution at high temperature for those planetesimals,

in conjunction with a thermal model of cooling objects, initially heated by the decay of 26Al, allows a more precise determination the accretion ages of these bodies. The

obtained chronological sequence shows that the accretion of the H chondrite parent body

(between ∼2 and ∼4 Ma after the formation of the solar system) is more recent than that

of the acapulcoite-lodranite parent body (between ∼1.5 and ∼2 Ma after solar system

formation), which occurred after the accretion of differentiated planetesimals (less than

∼1 Ma after solar system formation). This sequence of events suggests that the degree of

differentiation of planetesimals is mainly controlled by the abundance of 26Al

incorporated during accretion and, therefore, by their accretion time and that impact

heating has played a secondary role. Our results also show that the H chondrite and

acapulcoite-lodranite parent bodies have similar Hf/W ratio (0.63 ± 0.20), which is

13

Page 14: Touboul Thesis Completed

significantly lower than the Hf/W ratio of carbonaceous chondrites (1.21 ± 0.06). This

difference, unexpected for refractory elements, reveals an early Hf-W fractionation

during the first two million years of solar system, which could be produced during solar

nebula processes before or during the accretion of planetesimals.

Our new W isotope data obtained for a series of lunar metals and two lunar

anorthosites show that the products of lunar magma ocean crystallization all have an

identical W isotopic composition, which is similar to the composition of the Earth's

mantle, in contrast with what has been reported in earlier studies. We have shown that

previously reported 182W anomalies have a cosmogenic origin and are generated by the

production of 182W from 181Ta via neutron capture due to the intense exposure to cosmic

rays at the lunar surface. In contrast, the W isotope compositions determined in this study

do not contain any cosmogenic component: the lunar metals obtained here are sufficiently

pure and contain no tantalum, the selected anorthosites have very young exposure ages

(<2 Ma) and the extracted plagioclase fractions are pure. The identification of identical

isotopic signatures among the products of the lunar magma ocean in spite of different

Hf/W ratios yields a significant revision of lunar differentiation Hf-W ages from ~30 Ma

to more than ~60Ma after the formation of the solar system. In addition, the similarity of

W isotope compositions of lunar and terrestrial mantles indicates that the formation of the

Moon most likely occured after the formation of the solar system and suggests

an equilibration of W isotopes between the terrestrial magma ocean and the proto-lunar

disk in the aftermath of the giant impact. The giant impact being the last major event of

terrestrial accretion, its age is the age of the termination of the Earth’s accretion.

Consistently with the predictions of dynamic models, this age is, at present, the best

estimate of the time necessary for the solar system to evolve from a disk of dust and gas

to a planetary system.

Ma62 1090

−+

Keywords : 182Hf, 182W, extinct radionuclide, chronometry, early solar system,

planetesimal, planet, Moon, meteorite, accretion, differentiation, giant impact, magma

ocean, isotope equilibration, age of the Earth.

14

Page 15: Touboul Thesis Completed

Chapitre 1

Introduction

15

Page 16: Touboul Thesis Completed

16

Page 17: Touboul Thesis Completed

1.1 Introduction

Les météorites, fragments d’astéroïdes, et les échantillons disponibles de Mars, de

la Lune et de la Terre constituent des témoins uniques de l’évolution du sytème solaire

primitif. Les études isotopiques entreprises sur ces objets permettent de préciser notre

connaissance des processus régissant l’évolution du système solaire primitif, en

conjonction avec les observations astronomiques et les simulations théoriques. Plus

particulièrement, l’étude des isotopes radiogéniques, dont l’abondance relative évolue ou

a évolué avec le temps du fait de leur production par désintégration radioactive,

fournissent les contraintes les plus robustes sur les échelles de temps et la chronologie de

ces processus. Elles ont notamment permis l’identification des objets les plus vieux du

système solaire, les inclusions réfractaires riches en calcium et en aluminium (CAI) parmi

les composants des météorites. Ces condensats de haute température constituent les

premiers solides formés lors du refroidissement du gaz nébulaire après que la contraction

de la nébuleuse pré-solaire, initiallement froide, ait entraîné son échauffement et la

formation du soleil en son centre. Bien que leur âge exact fasse toujours débat, les

datations par les isotopes du Pb conduisent à des âges absolus de 4568,5±0,5 Ma

(Bouvier et al., 2005) et 4567,11±0,16 Ma (Amelin et al., 2002; Amelin et al., 2006) et

constituent à l’heure actuelle la meilleure estimation de l’âge minimum du système

solaire.

Excepté pour les météorites lunaires et martiennes, les météorites sont des

fragments d’astéroïdes éjectés lors de collisions au sein de la ceinture d’astéroïdes située

entre les orbites de Mars et Jupiter. Elles échantillonnent de petits corps parents appelés

planétésimaux qui sont des objets de taille kilométrique, et sont donc des intermédiaires

évolutifs de l’accrétion des planètes. Formée à partir d’aggrégats de poussières milli à

centimétriques, ces planétésimaux se sont collisionnés et accrétés par attraction

gravitationnelle pour donner des embryons planétaires. Ce processus de croissance est

appellé ‘runaway growth’. Selon les simulations dynamiques, ces embryons planétaires et

donc les planétésimaux, doivent apparaître rapidement, entre 0,01 et 1 Ma après la

formation des CAI (Chambers et al., 2004). On distingue deux types principaux de

météorites. Les météorites non différenciées ou chondrites sont des fragments de

planétésimaux qui n’ont jamais atteint la température de fusion. Par opposition, les

17

Page 18: Touboul Thesis Completed

météorites différenciées sont issues de planétésimaux qui ont été fondus et se sont

différenciés en un noyau métallique, un manteau silicaté et parfois une croûte ignée.

Métamorphisées ou altérées à différents degrés, les chondrites se composent de

proportions variables de chondres, de CAI, de grains pré-solaires et d’une matrice. La

présence de ces composants primitifs ont longtemps laissé penser que les corps parents

des chondrites s’étaient formés avant ceux des météorites différenciées. Cependant, les

datations isotopiques des météorites de fer (<1 Ma après la formation des CAI, Halliday,

2003, Markowski et al., 2006), qui sont des fragments de noyau de planétésimal, et des

chondres (~2 Ma après la formation des CAI, Bizzarro et al., 2005), dont la formation

doit précéder l’accrétion des corps parents des chondrites, révèle une séquence inverse,

où l’accrétion des planétésimaux différenciés est antérieure à l’accrétion des

planétésimaux non différenciés. Cette séquence suggère que le degré de différenciation

des planétésimaux est essentiellemnent déterminé par l’abondance d’26Al présent et

décroît donc avec le temps d’accrétion et que, au contraire, les impacts n’ont joué qu’un

rôle marginal dans le chauffage de ces objets. Elle est cependant encore à préciser,

notamment en contraignant au mieux les âges d’accrétion des corps parents des

chondrites, d’autres météorites différenciés, tels que les eucrites, ou d’achondrites

primitives, comme les acapulcoïtes et les lodranites, qui sont des intermédiaires entre

météorites différenciées et non différenciés.

Les embryons planétaires, qui sont des objets pouvant atteindre la taille de Mars,

entrent à leur tour en collision par attraction gravitationnelle pour former les planètes

telluriques (Wetherill et al., 1986 ; Agnor et Asphaug, 2004 ; Asphaug et al., 2006).

Beaucoup plus énergétiques que les collisions de planétésimaux, car elles mettent en jeu

des corps plus massifs, les collisions d’embryons planétaires obéissent à une physique

différente. En effet, l’énergie relâchée lors de ces impacts géants produit la fusion totale

ou partielle des planètes en formation engendrant une différenciation globale et la

formation d’océans de magma. Les simulations dynamiques prédisent que l’accrétion de

planètes à partir des collisions d’embryons planétaires doit prendre fin dans les cent

premiers millions d’années du système solaire (Chambers, 2004). La Lune est le meilleur

témoin de cette étape évolutive. Sa formation, lors d’un impact géant entre la proto-Terre

et un impacteur de la taille de Mars (Canup et Asphaug, 2001) à partir des débris ejectés

18

Page 19: Touboul Thesis Completed

qui ont constitué un disque proto-lunaire s’accrétant rapidement (10 à 100 ans, Thompson

and Stevenson, 1988, Canup et Ward, 2000), correspond au dernier événement majeur de

l’accrétion terrestre. Initiallement totalement ou au moins partiellement fondue (Pritchard

et Stevenson, 2000), l’évolution précoce de la Lune est caractérisée par la cristallisation

d’un océan magmatique (voir Warren, 1985). Si les données isotopiques n’ont pas encore

permis de déterminer l’âge exact de formation de la Lune, les chronométries 182Hf-182W

et 146Sm-142Nd sur les échantillons lunaires permettent la datation de la cristallisation de

l’océan magmatique lunaire qui correspond à l’âge le plus jeune auquel la Lune a pu se

former. Les âges obtenus par ces deux chronomètres, respectivement ~40 Ma (Kleine et

al., 2005) et ~200 Ma (Nyquist et al., 1995 ; Rankenburg et al., 2006 ; Boyet et al., 2007)

sont apparemment en désaccord.

La présente thèse est focalisée sur l’application du chronomètre isotopique Hf-W

dans l’objectif de contraindre plus précisément les échelles de temps de l’accrétion des

planétésimaux et des planètes. Ce système isotopique est basé sur la courte durée de vie

de l’182Hf (9 Ma), aujourd’hui éteint, mais dont la présence dans les premières dizaines de

Ma du système solaire peut être détectée par le biais du produit de sa désintégration

radioactive, le 182W, car elle se traduit par de petites variations des rapports 182W/184W.

Deux axes principaux de recherche ont été choisis: la précision de la séquence d’accrétion

des planétésimaux, grâce à l’étude des acapulcoites, des lodranites et des chondrites H, et

la précision des âges Hf-W de formation et de différenciation de Lune, grâce à l’étude

d’échantillons lunaires.

1.2 Le chronomètre 182Hf-182W

Depuis sa première application par Harper et al. (1991) sur la météorite ferreuse

Toluca, le radiochronomètre 182Hf-182W s’est avéré un instrument puissant d’investigation

des échelles de temps des processus géochimiques dans le système solaire primitif. Ces

quinze dernières années, ce système isotopique a été utilisé extensivement sur des

matériels variés, car il possède plusieurs caractéristiques particulières:

(1) L’182Hf se désintègre en 182W avec une période de demi-vie courte de 8.9 ± 0,09 Ma

(Vockenhuber et al., 2004), ce qui est du même ordre de grandeur que les temps

d’accrétion et de différenciation des planétésimaux et des planètes.

19

Page 20: Touboul Thesis Completed

(2) Hafnium et tungstène sont deux éléments réfractaires et par conséquent sont

supposés avoir une abondance relative identique à celle des chondrites dans les

objets planétaires totaux (Lee and Halliday, 1996).

(3) L’hafnium est strictement lithophile tandis que le tungstène est modérément

sidérophile, si bien que ces éléments fractionnent fortement lors de ségrégation

métal-silicate (Walter and Thibault, 1995, Walker et al., 2000).

(4) L’hafnium et le tungstène sont deux éléments généralement incompatibles mais

certains minéraux tels que l’ilménite et le clinopyroxene incorporent

préférentiellement l’hafnium (Righter and Shearer, 2003). Des fractionnements Hf-

W entre silicates peuvent donc aussi être produits par cristallisation ou fusion

partielle.

Ces différences de caractères géochimiques permettent une application du

système Hf-W d’une part à la datation de météorites individuelles par la méthode

d’isochrone interne et d’autre part à la datation de la ségrégation du noyau et de la

différenciation du manteau silicaté des planétésimaux différenciés et des planètes.

1.2.1 Notations

Les variations des rapports 182W/184W étant très petites, on exprime en général les

données isotopiques du W en unités ε182W. Le ε182W correspond à la déviation du rapport 182W/184W d’un échantillon par rapport à sa valeur pour le standard terrestre en part pour

10000.

4

dardtans184

182néchantillo

184

182

182 101

WW

WW

W ×

⎟⎟⎟⎟⎟

⎜⎜⎜⎜⎜

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

1.2.2 Ages Hf-W des météorites

1.2.2.1 Isochrone interne et âge relatif

Au sein des échantillons météoritiques, on retrouve des phases minérales avec des

rapports Hf/W très contrastés : le métal qui ne contient virtuellement aucun Hf a un

rapport Hf/W de ~0. Au contraire, l’ilménite et le clinopyroxène, enrichis en Hf

20

Page 21: Touboul Thesis Completed

relativement au W, ont des rapports Hf/W élevés. Quant aux autres phases, telles que

l’olivine ou le plagioclase par exemple, elles ne fractionnent pas significativement Hf et

W et ont donc des rapports Hf/W intermédiaires entre les deux pôles précédents. Si le

fractionnement Hf-W entre ces différentes phases a eu lieu alors que le 182Hf était

toujours présent (<60 Ma après la formation des CAI) et que le système est resté fermé

jusqu’à aujourd’hui, on observe alors des variations systématiques du rapport 182W/184W

corrélées avec le rapport élémentaire Hf/W. Ceci est illustré dans la figure 1.1 et peut être

démontré à partir de la loi classique de la désintégration radioactive :

( )t

ttt

eWHf

WW

WW λ−−×⎟⎟

⎞⎜⎜⎝

⎛+⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛1

0184

182

0184

182

184

182

t est le temps (t = 0 au moment de la formation des CAI), t0 est le temps auquel le

système s’est clos, c'est-à-dire le temps au-delà duquel la diffusion du W entre les

différences phases a cessé. λ est la constante de désintégration du 182Hf. Aujourd’hui

(t=4568,5 ± 0,5 Ma, Bouvier et al., 2007), l’ensemble du 182Hf présent initialement s’est

désintégré en 182W et on obtient :

⎟⎠⎞

⎜⎝⎛××⎟⎟

⎞⎜⎜⎝

⎛+⎟⎟

⎞⎜⎜⎝

⎛=

⎟⎟⎠

⎞⎜⎜⎝

⎛×⎟⎟

⎞⎜⎜⎝

⎛+⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛+⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

WHf

HfHf

WW

WHf

HfHf

WW

WHf

WW

WW

tt

ttttmesuré

181 0

180

182

0184

182

184

180

0180

182

0184

182

0184

182

0184

182

184

182

.

Les rapports isotopiques initiaux du W et de l’Hf sont identiques dans les

différentes phases minérales si ces dernières sont bien cogénétiques, et sont restés

imperturbées depuis leur formation et l’équilibration des isotopes du W entre les

différentes phases a bien été atteinte avant le fermeture du système. En revanche, elles

peuvent avoir des rapports Hf/W différents, ce qui produit aujourd’hui des différences de

signature isotopique . La pente de la corrélation obtenue dans le diagramme isochrone

permet ainsi de déterminer le rapport 182Hf/180Hf au temps t0.

21

Page 22: Touboul Thesis Completed

Figure 1.1: Schéma de diagramme isochrone pour trois fractions minérales avec des rapports Hf/W différents.

La comparaison entre les pentes obtenues pour deux objets distincts (A et B)

conduit ensuite au calcul d’un âge relatif ( BAt −Δ ), puisque les processus géochimiques ne

fractionnent pas significativement les isotopes lourds donc que le rapport 182Hf/180Hf

n’évolue qu’avec la désintégration du 182Hf.

)( BA

BA

tt

tt

eHfHf

HfHf

00

00

180

182

180

182−×⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛ λ d’où

⎟⎟⎟⎟⎟⎟

⎜⎜⎜⎜⎜⎜

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

=−=−

B0

A0

t180

182

t180

182

B0

A0BA

HfHf

HfHf

ln1tttλ

Δ

1.2.2.2 Calibration sur une échelle de temps absolu

La conversion des âges Hf-W relatifs en âges absolus précis, afin de faciliter

l’intercomparaison avec les autres chronomètres, requiert un ancrage du système sur une

échelle de temps absolu. Une bonne référence chronométrique doit remplir deux

principaux critères. Le premier est de disposer d’un âge absolu suffisamment précis du

matériel. On se concentrera sur le chronomètre Pb-Pb qui fournit les âges absolus les plus

précis pour des objets vieux de ∼4.5 Ga tels que les météorites, notamment pour celles

22

Page 23: Touboul Thesis Completed

dont les rapports U/Pb sont élevés. Le second est que la météorite datée doit avoir refroidi

suffisament rapidement si bien que la différence de température de fermeture des deux

systèmes isotopiques, à courte et à longue période, ne résulte pas en une différence d’âge

significative. Les CAI et les angrites remplissent toutes deux ce deuxième critère. Au

contraire des CAI (Amelin et al.,, Bouvier et al., 2007), les angrites présentent l’avantage

d’avoir des âges Pb-Pb variés (Lugmair and Galer, 1992; Amelin, 2008) qui s’étalent sur

une période de ∼7 Ma. Ceci permet une évaluation de la cohérence des deux

chronomètres. Le choix des angrites pour l’ancrage permet également de contourner le

débat sur les âges Pb-Pb conflictuels des CAIs (Amelin et al., Bouvier et al., 2007). Dans

un diagramme ln (182Hf/180Hf) en fonction des âges Pb-Pb (Fig. 1.2), les données

obtenues pour une série d’angrites définissent une droite dont la pente, correspondant à la

constante de désintégration du 182Hf, conduit à une valeur de 0,0758 ± 0,0068 Ma-1

(Kleine et al., 2008), en parfait accord avec la valeur expérimentale déterminée par

Vockenhuber et al. (2004). Par conséquent, l’ancrage sur les angrites assure la cohérence

des âges Hf-W et Pb-Pb et peut être utilisé pour convertir les âges Hf-W relatifs en âges

absolus. On choisira préférentiellement les angrites D’Orbigny et Sahara 99555 dont les

âges Hf-W sont les plus précis. Ces deux angrites ont des âges Pb-Pb identiques (Amelin,

2008, Connelly et al., 2008) qui, combinés donnent un âge Pb-Pb de 4564,50 ± 0,23 Ma.

Markowski et al. (2007) ont obtenu des isochrones internes Hf-W identiques pour ces

deux angrites conduisant à un rapport initial 182Hf/180Hf de (7,31 ± 0,16) × 10-5. L’âge Hf-

W absolu de tout échantillon A ancré sur les angrites d’Orbigny et Sahara 99555,

( ) Sah/Orb'DA

0t , peut maintenant être déterminé à partir de son rapport initial 182Hf/180Hf

comme suit :

( )⎟⎟⎟⎟⎟⎟

⎜⎜⎜⎜⎜⎜

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

+−= Sah/Orb'D

t180

182

A

t180

182

D'Orb/SahSah/Orb'DA0

Sah/Orb'D0

A0

HfHf

HfHf

lnλ1PbPbt

23

Page 24: Touboul Thesis Completed

Figure 1.2: ln(182Hf/180Hf) initial en fonction de l’âge Pb-Pb pour les angrites (d’après Kleine et al., 2008). Les âges Hf-W pour les angrites sont tirés de Markowski et al. (2007) et Kleine et al. (2008), les âges Pb-Pb de Amelin and Irving (2007) and Amelin (2008).

1.2.2.3 Température de fermeture du système

Les âges Hf-W dérivés des isochrones internes datent la dernière équilibration du

W entre les phases constitutives d’une roche. Celle-ci intervient lorsque la roche a

refroidi au dessous de la température de fermeture du système, qui correspond à la

température en dessous de laquelle les échanges diffusifs du W cessent entre les

différents minéraux présents (Dodson, 1973; Ganguly and Tirone, 2001). La comparaison

des âges Hf-W aux résultats obtenus par d’autres chronomètres nécessite la connaissance

de cette température de fermeture. Ceci est d’une importance particulière dans le cas des

météorites à refroidissement lent, car les différences de température de fermeture vont s’y

traduire par des différences d’âge significatives, ce qui permet de retracer leur histoire

thermique (voir chapitre 3).

La température de fermeture du système Hf-W reflète le comportement diffusif du

W entre métal, principal hôte du 182W non radiogénique, et silicates/oxydes, en particulier

clinopyroxène et ilménite qui sont les hôtes principaux du 182W radiogénique. La

24

Page 25: Touboul Thesis Completed

diffusion du W est quelques ordres de grandeur plus rapide dans le métal que dans les

silicates et les oxides, si bien que la température de fermeture Hf-W est essentiellement

contrôlée par la vitesse de diffusion dans le clinopyroxène et l’ilménite. Malgré l’absence

de données expérimentales, la température de fermeture du système Hf-W peut être

estimée en simulant numériquement la production et la diffusion simultanées du W

radiogénique entre clinopyroxène et métal (Van Orman et al., 2001, 2008, Kleine et al.,

2008, Touboul et al., soumis). Un exemple des résultats obtenus par ces simulations pour

une température initiale et des tailles de grain appropriées pour des météorites très

métamorphisés est présenté sur la figure 1.3. Pour une large gamme de vitesses de

refroidissement et de taille des clinopyroxènes, la température de fermeture du système

Hf-W est élevée (800-1100°C) et proche de la température initiale choisie. Elle est

supérieure aux températures de fermeture du Pb dans les pyroxènes ou les phosphates

(Cherniak et al., 1991; Cherniak, 1998) et du Mg dans les anorthites (LaTourrette and

Wasserburg, 1998). Par conséquent, dans le cadre d’un refroidissement, le système Hf-W

se ferme avant les autres chronomètres et cela se traduit pour les échantillons à

refroidissement suffisamment lent par des âges Hf-W plus anciens que les âges Pb-Pb et

Al-Mg (Kleine et al, 2008, Touboul et al, soumis). Ainsi, le chronomètre Hf-W date des

processus associés avec l’évolution la plus précoce des corps parents des météorites.

Figure 1.3: Température de fermeture du système Hf-W en fonction du diamètre des clinopyroxènes et de la vitesse de refroidissement. La température initiale a été fixée à 1100-1200°C (d’après Touboul et al., 2008)

25

Page 26: Touboul Thesis Completed

1.2.3 Datation de la differenciation planétaire

1.2.3.1 Systématique 182Hf-182W des chondrites et des objets planétaires totaux

Hafnium et tungstène sont deux éléments réfractaires et sont donc supposés avoir

une abondance chondritique dans la plupart des objets planétaires. La systématique des

chondrites fournit donc une estimation de la composition des objets planétaires totaux et

sert de référence pour le calcul des âges Hf-W de formation du noyau. La détermination

de l’évolution de la composition isotopique du W dans le réservoir chondritique requiert

la connaissance de son rapport 182W/184W actuel et de ses rapports initiaux 182W/184W et 182Hf/180Hf.

Les premières données obtenues sur les chondrites (Lee and Halliday, 1995; Lee

and Halliday, 1996) semblaient montrer qu’elles avaient un rapport 182W/184W identique

au manteau terrestre. Cependant, les travaux ultérieurs (Kleine et al., 2002; Schoenberg et

al., 2002a; Yin et al., 2002) ont mis en évidence que les chondrites carbonées présentent

une anomalie négative en 182W d’environ -2 ε par rapport au manteau terrestre.

Actuellement, la détermination la plus précise du rapport pour les chondrites carbonées

est basée sur 14 échantillons dont la moyenne conduit à un ε182W de -1,9 ± 0,1 (Kleine et

al., 2004).

La meilleure méthode de détermination des rapports initiaux 182Hf/180Hf et 182W/184W du système solaire est l’utilisation d’isochrone interne pour les CAI, car ce

sont les objets les plus vieux du système solaire (Amelin, 2008, Bouvier et al., 2007).

Burkhardt et al. (2008) ont obtenu une isochrone interne unique pour des fractions

minérales provenant de 8 CAI des chondrites CV3 Allende et NWA 2364. La pente de

cette isochrone correspond à un rapport 182Hf/180Hf initial de (9,70 ± 0,40)×10-5 et son

intersection avec l’axe des ordonnées donne un initial ε182W de -3,30 ± 0,12. Le système

Hf-W n’est pas affecté par le métamorphisme thermique de basse température et

l’altération aqueuse du fait de la haute température de fermeture de ce système. Par

conséquent, l’isochrone interne Hf-W pour les CAI conduit bien aux rapports initiaux 182Hf/180Hf and 182W/184W au moment de leur formation des CAI, c'est-à-dire il y a

∼4568,5 Ma.

La composition isotopique du réservoir chondritique a ainsi évolué depuis -3,3 ε

lors de la formation des CAI jusqu’à la valeur actuelle plus radiogénique de -1,9 ε, sous

26

Page 27: Touboul Thesis Completed

l’effet de la désintégration du 182Hf dont l’abondance relative initiale est donnée par le

rapport initial des CAI. Cette évolution correspond à un réservoir ayant un rapport Hf/W

de 1.21, ce qui est cohérent avec les rapports mesurés dans les chondrites carbonées (de

1.1 à 1.6, Kleine et al., 2004, Burkhardt et al., 2008).

Figure 1.5: Evolution du ε182W avec le temps dans les réservoirs résultant de la formation du noyau, supposé instantanée, au sein d’un objet planétaire caractérisé par une évolution chondritique. CHUR = réservoir chondritique uniforme.

1.2.3.2 Ages Hf-W de formation du noyau

Lors de la formation du noyau, l’Hf, qui est strictement lithophile, s’accumule

entièrement dans le manteau tandis que le W, sidérophile, est incorporé

préférentiellement dans le noyau métallique. Dépourvu d’hafnium, le noyau conserve un

rapport 182W/184W correspondant à la composition isotopique du W de la planète totale au

moment de la ségrégation du noyau (Fig.1.5). Connaissant l’évolution de la composition

isotopique du réservoir chondritique, le rapport 182W/184W de tout échantillon dérivé de

noyau planétaire permet donc de déterminer un âge de formation du noyau. Par exemple,

Markowski et al. (2007) ont montré que les météorites de fer ont des ε182W (corrigés des

27

Page 28: Touboul Thesis Completed

effets cosmogéniques) proches du ε182W initial des CAI, indiquant que la ségrégation du

noyau au sein du corps parent est intervenu au cours du premier million d’années du

système solaire.

Au contraire du noyau, le manteau a hérité d’un rapport Hf/W superchondritique.

Le coefficient de partage métal-silicate du W dépend entre autres de la fugacité en

oxygène du manteau, de la pression, de la température et de la composition du liquide

silicaté (Walter and Thibault, 1995; Walter et al., 2000). Ainsi, selon les conditions qui

prévalent à la formation du noyau, l’appauvrissement du manteau en W peut varier d’un

objet planétaire à un autre mais aussi au sein d’un même objet au cours de ce processus.

Le modèle le plus simple permettant de calculer un âge fait l’hypothèse que la formation

du noyau est un événement discret ayant lieu à un instant précis. Depuis une valeur

initialement chondritique à cet instant, le rapport 182W/184W du manteau évolue avec le

temps vers des valeurs d’autant plus radiogéniques que son rapport Hf/W est élevé (Fig.

1.5). La calcul d’un âge modèle à deux stades pour la formation du noyau à partir du

ε182W mantellique requiert donc l’estimation du rapport Hf/W caractéristique du manteau

total. Cependant, ce dernier ne peut être mesuré directement dans la plupart des cas, car

des fractionnements Hf-W significatifs ont pu avoir lieu lors de processus ignées

ultérieurs au sein du manteau (Righter and Shearer, 2003). Les rapports Hf/W

mantelliques doivent donc être estimés en utilisant l’abondance relative d’un élément

strictement réfractaire lithophile dont l’incompatibilité est identique au W. Deux éléments

réfractaires lithophiles, Th and U, ont des comportements incompatibles similaires au W

(Palme and Rammensee, 1981b; Newsom et al., 1996). Par conséquent, les rapports Hf/W

des manteaux planétaires totaux peuvent être déterminés à partir des rapports Th/W et

U/W d’échantillons d’origine mantellique combinés aux rapports Hf/Th et Hf/U des

chondrites.

En utilisant la différence de ~2 ε entre le manteau terrestre et les chondrites en

combinaison avec le rapport Hf/W du manteau terrestre ainsi déterminé (~17), on obtient

un âge modèle à deux stades de formation du noyau terrestre de ~30 Ma. Les meilleures

estimations actuelles de ces paramètres dans le manteau martien donnent quant à eux des

âges de formation du noyau martien compris entre 0 et 8 Ma. Cependant, l’hypothèse de

formation instantanée du noyau n’est pas réaliste, notamment dans le cas d’objets

28

Page 29: Touboul Thesis Completed

planétaires larges comme la Terre et Mars où cette différenciation a probablement été

continue au cours de l’accrétion et largement affectée par les impacts géants marquant la

fin de l’accrétion. Cette hypothèse se traduit par une sous-estimation de l’âge réel de

formation du noyau. Ces âges modèle à deux stades fournissent cependant des contraintes

temporelles importantes puisqu’ils correspondent aux âges les plus vieux auxquels les

noyaux terrestre et martien ont pu être formés (Halliday et al., 1996; Kleine et al., 2004b,

Jacobsen et al., 2005, Nimmo et Kleine, 2007).

1.2.3.3 Ages Hf-W de différenciation mantellique

Des fractionnements additionnels à ceux intervenant lors de la formation du noyau

peuvent avoir lieu au sein de la partie silicatée des objets différenciés. Ils résultent de la

plus grande incompatibilité du W par rapport à l’Hf, qui est préférentiellement incorporé

dans certains minéraux silicatés et oxydes tels que les pyroxènes riches en Ca et les

ilménites. Ces fractionnements, parfois plus importants que ceux produit lors de la

formation du noyau, créent des réservoirs mantelliques et crustaux dont les rapports Hf/W

diffèrent entre eux et bien souvent du rapport Hf/W du manteau total, ce qui complique et

parfois empêche la détermination d’un âge de formation du noyau. Cependant, si ces

fractionnements interviennent alors que le 182Hf n’est pas totalement éteint (dans les 60

premiers Ma du système solaire), les différences de rapport Hf/W vont produire des

différences de ε182W entre les réservoirs mantelliques, ce qui fournit des contraintes

chronologiques importantes sur la différenciation mantellique. Par exemple, la

cristallisation de l’océan magmatique lunaire conduit à des réservoirs mantelliques avec

des rapports Hf/W différents. Les basaltes de mer pauvres en Ti échantillonnent les

premiers cumulats qui cristallisent dans l’océan de magma et qui ont constitué une source

mantellique essentiellement composée d’olivines et d’orthopyroxènes, donc non

fractionnée par rapport à l’océan magmatique total (Hf/W~26). Les basaltes de mer riches

en Ti échantillonnent une source mantellique à rapport Hf/W élevé (~100) du fait de la

présence d’ilménites et de clinopyroxènes. En revanche, les « KREEP » (échantillons

enrichi en Potassium, en Terre-rares et en Phosphore), qui correspondent au liquide

résiduel de la cristallisation de l’océan magmatique lunaire et sont donc hautement

enrichis en éléments incompatibles dont le W, ont hérité d’un rapport Hf/W bas (~20).

29

Page 30: Touboul Thesis Completed

Kleine et al. (2005) ont obtenu des variations systématiques du ε182W en fonction des

rapports Hf/W entre les KREEP et les sources mantelliques des basaltes de mer pauvres

et riches en Ti. Ces variations isotopiques au sein du manteau lunaire indiqueraient que la

cristallisation de l’océan magmatique lunaire s’est achevée entre 30-50 Ma après la

formation des CAI, comme l’illustre la figure 1.6.

Figure 1.6: Diagramme isochrone ε182W vs. (180Hf/184W)source pour les basaltes lunaires de mer pauvres et riches en Ti et pour les KREEP (d’après Kleine et al., 2005)

1.3 Contenu de la thèse

Cette thèse est consacrée à l’étude de l’évolution précoce des planétésimaux et

des planètes grâce au système isotopique 182Hf-182W. Son objectif principal est

l’établissement d’une chronologie précise de l’accrétion, du métamorphisme thermique

ou de la différenciation de ces objets mais les résultats obtenus apportent également des

contraintes importantes sur la nature des processus mis en jeu.

Le Chapitre 2 s’attache à décrire les techniques analytiques employées, en

particulier les procédures chimiques permettant la séparation de l’Hf et du W, adaptée

pour le propos de cette étude à partir de la méthode développée par Kleine et al. (2004).

30

Page 31: Touboul Thesis Completed

Les rendements et les blancs de la procédure ont ainsi été améliorés permettant de

travailler sur de faibles quantités de W. La méthode de mesure par spectrométrie de

masse à source plasma est également décrite. La sensibilité de l’instrument est le facteur

limitant pour les quantités minimum de W requises : au moins 8 ng de W sont nécessaires

pour atteindre une reproductibilité externe supérieure ou égale à 0,4 ε.

Le chapitre 3 illustre l’utilisation du systéme isotopique Hf-W en tant que

thermochronomètre, afin de contraindre l’histoire thermique des corps parents des

météorites. Il comporte deux articles. Le premier, en préparation pour être soumis à

Geochimica et Cosmochimica Acta, concerne les acapulcoïtes et les lodranites qui

appartiennent au groupe des achondrites primitives et dérivent d’un corps parent

commun. En terme de température atteinte lors du pic thermique, les acapulcoïtes et les

lodranites sont des intermédiaires entre météorites différenciées et non différenciées. Les

acapulcoites et les lodranites ont fondu partiellement mais la fusion limitée n’a pas

permis la migration efficace des liquides, si bien qu’elles ne se sont pas différenciées.

Nous présentons des isochrones Hf-W pour une série d’acapulcoïtes et de lodranites,

permettant de dater le refroidissement de ces météorites au dessous de la température de

fermeture du système Hf-W. En collaboration avec James Van Orman, cette température

de fermeture a pu être estimée par modélisation numérique de la diffusion du W entre

métaux et silicates. La contrainte thermochronométrique ainsi fournie par le système Hf-

W est ensuite comparée aux contraintes apportées par les autres chronomètres (ex : Pb-

Pb, Ar-Ar), ce qui révèle l’histoire thermique du corps parent des acapulcoites et des

lodranites et permet de contraindre l’âge d’accrétion de leur corps parent. Le second

article a été écrit par Thorsten Kleine et publié dans Earth and Planetary Science Letters.

Il s’agit d’une étude parallèle à laquelle j’ai collaboré et dans laquelle la

thermochronométrie Hf-W est appliquée aux chondrites ordinaires H selon une approche

similaire.

Le chapitre 4 se concentre sur la détermination des âges de formation et de

differentiation de la Lune et rassemble trois articles. Le premier, publié dans Nature,

présente les nouvelles données Hf-W obtenues sur des métaux lunaires dérivés de

KREEP et de basaltes de mer pauvres et riches en Ti. Dans les roches lunaires, la

production de 182W par capture neutronique par le 181Ta, due à l’intense irradiation de la

31

Page 32: Touboul Thesis Completed

surface lunaire par les rayonnements cosmiques, masque leur composition isotopique

endogène (originelle). Les métaux présentent l’avantage de ne pas contenir de Ta et leur

composition isotopique n’est donc pas affectée par cet effet cosmogénique. Ils ont donc

préservé la signature isotopique originelle des KREEP et des sources mantelliques des

basaltes de mer dont ils sont extraits. Les rapports 182W/184W de ces réservoirs, dont les

rapports Hf/W diffèrent, sont utilisés afin de déterminer un âge de différenciation

mantellique et d’estimer une composition isotopique du manteau lunaire total. En

combinaison avec les rapports Hf/W des manteaux lunaire et terrestre, cette dernière

permet de calculer un âge de formation de la Lune. Dans la continuité de ce premier

article, le second, accepté par Icarus, présente nos données Hf-W obtenues pour deux

anorthosites ferreuses lunaires. Faiblement exposées aux rayonnements cosmiques, leur 182W est également dépourvue de tout composant cosmogenique significatif. La

motivation de cette étude résidait dans la présence d’anomalies du 182W obtenues

précédemment sur ce type d’échantillon lunaire (Lee et., 1997), ce qui est en désaccord

avec une formation tardive de Lune (> 60 Ma après la formation des CAI) déduite des

résultats obtenus sur les métaux lunaires. Enfin le troisième article, écrit par Bernard

Bourdon et publié dans Philosophical Transactions of the Royal Society A, auquel j’ai

collaboré, explore plus en détail le lien entre degré d’équilibration entre impacteur et la

proto-Terre et l’âge de formation de la Lune, à la lumière des nouvelles données Hf-W

présentées dans les deux articles précédents. Il met également en parallèle les données

Sm-Nd lunaires et martiennes afin d’estimer la composition isotopique initiale du Nd.

Enfin, le chapitre 5 est consacré aux conclusions principales ainsi qu’à quelques

perspectives de recherche future qu’ouvrent les nouveaux résultats obtenus au cours de

cette thèse.

L’appendice comprend tout d’abord une courte discussion des données

préliminaires obtenues, pour la première fois, pour une série conséquente d’eucrites

cumulats. Ces données Hf-W sont utilisées pour contraindre la formation du noyau et la

différenciation mantellique du corps parent des eucrites. La seconde partie de l’appendice

est constituée d’un article de Thorsten Kleine auquel j’ai collaboré. Ce dernier est

consacré au passage en revue de l’ensemble des contraintes sur l’évolution précoce des

32

Page 33: Touboul Thesis Completed

planétésimaux et des planètes, qu’a apportées la chronométrie 182Hf-182W jusqu’à

aujourd’hui.

Références

Agnor C. et Asphaug E. , 2004. Accretion efficiency during planetary collision. Astrophy.

J. 613(2), L157-L160.

Amelin Y., Krot A.N., Hutcheon I.D., Ulyanov A.A., 2002. Lead isotopic ages of

chondrules and calcium-aluminium-rich inclusions, Science 297, 1678-1683.

Amelin, Y., Wadhwa, M., and Lugmair, G. W., 2006. Pb-isotopic dating of meteorites

using 202Pb-205Pb double spike: comparison with other high-resolution

chronometers. Lunar Planet. Sci. Conf. XXXVII, 1970

Amelin, 2008, U-Pb ages of angrites. Geochim. Cosmochim. Acta 72, 221-232.

Amelin Y., Irving A. J., Seven Million Years of Evolution on the Angrite Parent Body

from Pb-Isotopic Data. Workshop on the Chronology of Meteorites and the Early

Solar System, LPI Contribution No. 1374, 20-21.

Asphaug E., Agnor C.B., Williams Q., 2006. Hit-and-run planetary collisions. Nature,

439, 155-160.

Bizzarro, M., Baker, J. A., Haack, H., and Lundgaard, K. L., 2005. Rapid timescales for

accretion and melting of differentiated planetesimals inferred from 26Al-26Mg

chronometry. Astrophys. J. 632, L41-L44.

Bouvier A., Blichert-Toft J.; Moynier F., Vervoort J.D., Albarede F., 2007. Pb-Pb dating

constraints on the accretion and cooling history of chondrites, Geochim.

Cosmochim. Acta 71, 1583-1604.

Boyet, M. & Carlson, R. W. 2007 A highly depleted moon or a non-magma ocean origin

for the lunar crust? Earth Planet. Sci. Lett. 262, 505-516.

Burkhardt, C., Kleine, T., Palme, H., Bourdon, B., Zipfel, J., Friedrich, J., and Ebel, D.

Hf-W isochrons for CAIs and timescales for planetesimal accretion Geochim.

Cosmochim. Acta, submitted.

33

Page 34: Touboul Thesis Completed

Canup, R. M. and Asphaug, E., 2001. Origin of the Moon in a giant impact near the end

of the Earth's formation. Nature 412, 708-712.

Canup R.M., Ward W.R., Cameron A.G.W., 2001. A scalling law for satellite-forming

impact. Icarus 150, 288-296.

Chambers, J. E., 2004. Planetary accretion in the inner solar system. Earth Planet. Sci.

Lett. 223, 241-252.

Cherniak, D. J., 1998. Pb diffusion in clinopyroxene. Chem. Geol. 150, 105-117.

Cherniak, D. J., Lanford, W., and Ryerson, F. J., 1991. Lead diffusion in apatite and

zircon using ion implantation and Rutherford backscattering techniques. Geochim.

Cosmochim. Acta 55, 1663-1673.

Connelly, J., Amelin, Y., Bizzarro, M., Thrane, K., and Baker, J. A., 2008. The Pb-Pb age

of angrite Sah99555 revisited. Lunar Planet. Sci. Conf. XXXIX, 2386.pdf.

Dodson, M. H., 1973. Closure temperature in cooling geochronological and petrological

systems. Contrib. Mineral. Petrol. 40, 259-274.

Ganguly, J. and Tirone, M., 2001. Relationship between cooling rate and cooling age of a

mineral: Theory and applications to meteorites. Meteorit. Planet. Sci. 36, 167-175.

Halliday A.N, 2003. The origin and earliest history of the Earth. In Treatrise on

geechemistry. Meteorites, comets and planets. Vol. 1 (éd. Davis A.M., Holland

H.D., Turekian K.K.).

Halliday, A., Rehkämper, M., Lee, D. C., and Yi, W., 1996. Early evolution of the Earth

and Moon: New constraints from Hf-W isotope geochemistry. Earth Planet. Sci.

Lett. 142, 75-89.

Harper, C. L., Völkening, J., Heumann, K. G., Shih, C. Y., and Wiesmann, H., 1991. 182Hf-182W: New cosmochronometric constraints on terrestrial accretion, core

formation, the astrophysical site of the r-process, and the origin of the solar system.

Lunar Planet. Sci. Conf. XXII, 515-516.

Jacobsen, S. B., 2005. The Hf-W isotopic system and the origin of the Earth and Moon.

Ann. Rev. Earth Planet. Sci. 33, 531-570.

Kleine T., Touboul M., Van Orman J. A., Bourdon B., Maden C., Mezger K., and

Halliday A., 2008. Hf-W thermochronometry: closure temperature and constraints

34

Page 35: Touboul Thesis Completed

on the accretion and cooling history of the H chondrite parent body. Earth Planet.

Sci. Lett. 270, 106-118.

Kleine, T., Palme, H., Mezger, K., and Halliday, A. N., 2005. Hf-W chronometry of lunar

metals and the age and early differentiation of the Moon. Science 310, 1671-1674.

Kleine, T., Mezger, K., Münker, C., Palme, H., and Bischoff, A., 2004. 182Hf-182W

isotope systematics of chondrites, eucrites, and Martian meteorites: Chronology of

core formation and mantle differentiation in Vesta and Mars. Geochim. Cosmochim.

Acta 68, 2935-2946.

Kleine, T., Mezger, K., Palme, H., Scherer, E., and Münker, C., 2004. The W isotope

evolution of the bulk silicate Earth: constraints on the timing and mechanisms of

core formation and accretion. Earth Planet. Sci. Lett. 228, 109-123.

Kleine, T., Münker, C., Mezger, K., and Palme, H., 2002. Rapid accretion and early core

formation on asteroids and the terrestrial planets from Hf-W chronometry. Nature

418, 952-955.

Kleine, T., Bourdon, B., Burkhardt, C., and Irving, A. J., 2008. Hf-W chronometry of

angrites: constraints on the absolute age of CAIs and planetesimals accretion

timescales. Lunar Planet. Sci. Conf. XXXIX, 2367.pdf.

LaTourrette, T. and Wasserburg, G. J., 1998. Mg diffusion in anorthite: implications for

the formation of early solar system planetesimals. Earth Planet. Sci. Lett. 158, 91-

108.

Lee, D. C. and Halliday, A. N., 1995. Hafnium-tungsten chronometry and the timing of

terrestrial core formation. Nature 378, 771-774.

Lee, D. C. and Halliday, A. N., 1996. Hf-W isotopic evidence for rapid accretion and

differentiation in the early solar system. Science 274, 1876-1879.

Lee, D. C., Halliday, A. N., Snyder, G. A., and Taylor, L. A., 1997. Age and origin of the

moon. Science 278, 1098-1103.

Lugmair, G. W. and Galer, S. J. G., 1992. Age and isotopic relationships among the

angrites Lewis Cliff 86010 and Angra dos Reis. Geochim. Cosmochim. Acta 56,

1673-1694.

35

Page 36: Touboul Thesis Completed

Markowski, A., Quitté, G., Halliday, A. N., and Kleine, T., 2006. Tungsten isotopic

compositions of iron meteorites: chronological constraints vs. cosmogenic effects.

Earth Planet. Sci. Lett. 242, 1-15.

Markowski, A., Quitté, G., Kleine, T., Halliday, A., Bizzarro, M., and Irving, A. J., 2007.

Hf-W chronometry of angrites and the earliest evolution of planetary bodies. Earth

Planet. Sci. Lett. 262, 214-229.

Newsom, H. E., Sims, K. W. W., Noll, P., Jaeger, W., Maehr, S., and Beserra, T., 1996.

The depletion of W in the bulk silicate Earth: constraints on core formation.

Geochim. Cosmochim. Acta 60, 1155-1169.

Nimmo, F. and Kleine, T., 2007. How rapidly did Mars accrete? Uncertainties in the Hf-

W timing of core formation. Icarus 191, 497-504.

Nyquist, L. E., Wiesmann, H., Bansal, B., Shih, C. Y., Keith, J. E., and Harper, C. L.,

1995. Sm-146-Nd-142 Formation Interval for the Lunar Mantle. Geochim.

Cosmochim. Acta 59, 2817-2837.

Pritchard M.E, Stevenson D.J., 2000. Thermal implications of a lunar origin by giant

impact. In: Origin of the Earth and Moon. University Arizona Press (éd. Canup

R.M., Righter K) , 179-196.

Rankenburg, K., Brandon, A. D., and Neal, C. R., 2006. Neodymium isotope evidence for

a chondritic composition of the Moon. Science 312, 1369-1372.

Righter, K. and Shearer, C. K., 2003. Magmatic fractionation of Hf and W: Constraints

on the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497-2507.

Schoenberg R., Kamber B. S., Collerson K. D. ,Eugster O., 2002. New W-isotope

evidence for rapid terrestrial accretion and very early core formation. Geochim.

Cosmochim. Acta 66, 3151-3160.

Thompson C., Stevenson D.J., 1988. Gravitational instability of two phase disks and the

origin of the Moon. Atrophys. J. 333, 452-481.

Touboul M., Kleine T., Bourdon B., Van Orman J. A., Maden C., Irving A. J., Zipfel J.,

Bunch T.E., Accretion and thermal history of the acapulcoite-lodranite parent body

inferred from Hf-W thermochronometry, submitted

36

Page 37: Touboul Thesis Completed

Van Orman, J. A., Grove, T. L., and Shimizu, N., 2001. Rare earth element diffusion in

diopside: Influence of temperature, pressure and ionic radius, and an elastic model

for diffusion in silicates. Contrib. Mineral. Petrol. 141, 687-703.

Van Orman, J. A., Saal, A. E., Bourdon, B., and Hauri, E. H., 2006. Diffusive

fractionation of U-series radionuclides during mantle melting and shallow level

melt-cumulate interaction. Geochim. Cosmochim. Acta 70, 4797-4812.

Vockenhuber, C., Oberli, F., Bichler, M., Ahmad, I., Quitté, G., Meier, M., Halliday, A.

N., Lee, D. C., Kutschera, W., Steier, P., Gehrke, R. J., and Helmer, R. G., 2004.

New half-life measurement of 182Hf: Improved chronometer for the early solar

system. Phys. Rev. Lett. 93, art. no. 172501.

Walter, M. J. and Thibault, Y., 1995. Partitioning of tungsten and molybdenum between

metallic liquid and silicate melt. Science 270, 1186-1189.

Walter, M. J., Newsom, H. E., Ertel, W., and Holzheid, A., 2000. Siderophile elements in

the Earth and Moon: metal/silicate partitioning and implications for core formation.

In:, Origin of the Earth and Moon (éd.Canup, R. M. and Righter, K.). Lunar and

Planetary Institute, Houston.

Warren P.H., 1985. The magma ocean concept and lunar evolution. Annual Review in

Earth and Planetary Science, 13, 201-240.

Wetherill G.W., 1986. Accumulation of terrestrial planets and implcations concerning the

lunar origin. In: Origin of the Moon (éd. Hartmann W.K., Phillips R.J., Taylor J.G.).

Lunar and planetary institute, 519-551.

Yin, Q. Z., Jacobsen, S. B., Yamashita, K., Blichert-Toft, J., Télouk, P., and Albarède, F.,

2002. A short timescale for terrestrial planet formation from Hf-W chronometry of

meteorites. Nature 418, 949-952.

37

Page 38: Touboul Thesis Completed

38

Page 39: Touboul Thesis Completed

Chapitre 2

Techniques analytiques

Analytical techniques

39

Page 40: Touboul Thesis Completed

40

Page 41: Touboul Thesis Completed

2.1 Introduction

Ces quinze dernières années, durant lesquelles les mesures de la composition

isotopique du W ont suscité un intérêt croissant et se sont popularisées, deux techniques

de mesure ont été communément utilisées. Les premières données (par exemple Harper et

al., 1991, Harper et Jacobsen, 1996) ont été produites par spectrométrie de masse à

thermo-ionisation (N-TIMS) qui permet d’atteindre une précision de l’ordre de 1 ε sur le

rapport 182W/184W. Malgré le potentiel d’ionisation élevé du W, cette technique permet

d’analyser de petites quantités de W (à partir de 400 pg). Depuis l’avènement des

nouvelles générations de spectromètre de masse à source plasma (MC-ICPMS) à partir

des années 2000, la plupart des données isotopiques du W sont produites sur ce type

d’instrument (par exemple, Yin et al., 2002, Kleine et al, 2004, 2005, Markowski et al,,

2006) . Bien que moins sensible que la thermo-ionisation (quelques dizaines de ng de W

sont nécessaires), la précision atteinte est environ 5 fois meilleure (~0.2 ε sur le rapport 182W/184W). Un autre avantage des sources plasma est que la préparation des échantillons

se trouve simplifiée, la phase de dépôt sur filament d’ionisation n’étant pas nécessaire.

Les échantillons sont repris dans une solution d’introduction qui, pompée au moyen d’un

cathéter, est injectée dans la chambre de nébulisation. Le spray obtenu passe dans la

torche du plasma pour y être atomisé et ionisé. Les ions produits pénètrent dans la partie

spectromètre de masse de l’instrument, où, de manière classique, ils sont accélérés dans

un champs électrique puis déviés dans un champs magnétique selon une trajectoire qui

dépend de leur masse. Les isotopes sont ainsi collectés séparément dans les cages de

Faraday. Afin d’éviter les problèmes d’interférences et les effets de matrice pouvant

perturber l’ionisation de l’échantillon et le faisceau d’ions au cours des mesures

isotopiques, il est fondamental de purifier l’élément d’intérêt (W) par séparation

chimique.

Dans ce chapitre, les techniques utilisées pour la préparation des échantillons,

permettant d’obtenir des fractions pures de métaux et de silicates, sont tout d’abord

présentées. Les méthodes de purification du W par chromatographie ionique sont ensuite

décrites. Initialement adaptée à partir de celle développée par Kleine et al.(2004), la

technique de séparation chimique a été progressivement améliorée et modifiée pour le

41

Page 42: Touboul Thesis Completed

propos de cette thèse, afin de permettre de travailler sur des échantillons très pauvres en

W, comme les anorthosites lunaires. Enfin, nous exposons le protocole de mesure des

compositions isotopiques du W sur le MC-ICPMS Nu Plasma.

2.2 Préparation des échantillons

Avant toute opération ultérieure, les surfaces des échantillons sont abrasées avec

du papier de verre puis lavées avec du HNO3 0,05 M, de l’eau déionisée et enfin de

l’éthanol ultrapur distillé dans un bain ultrasonique afin d’éliminer toute contamination

éventuelle, introduite au cours du découpage de l’échantillon massif.

2.2.1 Roches totales

Les poudres de roches totales sont obtenues en général par broyage d’un morceau

de météorite de 0,5 à 1 g dans un mortier en agathe. Dans le cas des eucrites basaltiques

disponibles en large quantité, nous avons broyé ~7 g afin de minimiser les effets liés à

l’hétérogénéité des échantillons. Tout d’abord réduits grossièrement en poudre dans un

mortier en agathe, les échantillons eucritiques sont ensuite transférés dans une meule

automatique permettant l’obtention d’une poudre très fine et homogène.

2.2.2 Séparation magnétique

Cette méthode de séparation à l’aide d’un aimant a été utilisée afin d’obtenir des

fractions de métal de haute pureté à partir des échantillons lunaires (~1 g pour les

KREEP, ~4 g pour les basaltes de mer) et des fractions à rapports Hf/W contrastés

(métaux et silicates) à partir des échantillons d’acapulcoites et de lodranites (~2 à 4 g).

L’échantillon est tout d’abord délicatement broyé dans un mortier en agathe et séparé en

utilisant un tamis en nylon en fractions de moins de 40 μm et de 40 à 150 μm. Dans le cas

des échantillons lunaires où seuls les grains de métaux nous intéressent, la roche est

directement broyée très finement. Les grains de métal sont récupérés en utilisant un

aimant. Les poussières silicatées attachées ainsi que les silicates intimement inclus dans

les grains de métal sont ensuite éliminés en répétant le broyage et la séparation

magnétique dans l’éthanol. Cette étape a été réalisée avec un soin particulier dans le cas

des métaux lunaires, pour lesquels elle est répétée plusieurs fois afin de s’assurer de leur

42

Page 43: Touboul Thesis Completed

pureté. Les fractions de 40 à 150 μm sont toujours un peu magnétiques, probablement du

fait de la présence de petites inclusions métalliques dans les grains de silicates et

d’oxydes. Elles sont séparées en utilisant un aimant en fractions faiblement magnétiques

et non-magnétiques. Elles sont ensuite elles-mêmes séparées en plusieurs fractions

faiblement magnétiques (WM) et non-magnétiques (NM). Inspectées à la loupe

binoculaire, les fractions NM révèlent un enrichissement en ilménites/clinopyroxenes et

les fractions WM et un enrichissement en olivine/orthopyroxenes/plagioclases. Toutes les

fractions (de 150 à 300 mg) sont ensuite pulvérisées dans un mortier en agathe et les

éventuels grains de métal restants sont à nouveau séparés en utilisant l’aimant.

2.2.3 Séparation des plagioclases d’anorthosites lunaires

Les plagioclases des anorthosites ferreuses lunaires sont supposés s’être formés

par cristallisation dans l’océan de magma lunaire et s’être accumulés à la surface par

flottation pour former la croûte primitive de la Lune. Au sein des anorthosites, ces

plagioclases ont été associés à des minéraux basiques tels que des olivines et des

orthopyroxènes (quelques %) et nécessitent donc d’être séparés. L’anorthosite (~5 g) est

tout d’abord broyée délicatement dans un mortier en agathe jusqu’à ce que l’ensemble de

l’échantillon ait une taille de grain inférieure à 500 μm. La fraction inférieure à 40 μm,

trop fine pour les liqueurs denses, est éliminée. Une fraction de plagioclases pure est

obtenue à partir de la fraction de 40 à 500 μm par séparation à l’aide de liqueur dense

(iodure de méthyle) puis purifiée par prélèvement manuel sous une loupe binoculaire et

par élimination des éventuels grains ou d’inclusions de métal à l’aide d’un aimant. La

fraction pure de plagioclases est ensuite lavée à éthanol distillé dans un bain à ultrasons,

séchée puis pulvérisée dans un mortier en agathe.

2.3 Procédures chimiques

Afin d’éviter les problèmes d’interférences au cours des mesures isotopiques aux

spectromètres de masse, il est fondamental de purifier les éléments d’intérêt (W, Hf) par

séparation chimique de la solution de roche obtenue par attaque acide. La procédure

chimique est essentiellement basée sur la capacité du W à être complexé par l’HF et

l’H2O2. Toute la procédure chimique est effectuée en salle blanche sous hôte de classe 10

43

Page 44: Touboul Thesis Completed

en utilisant des réactifs purifiés. On utilise, entre autres, du peroxyde d’hydrogène 31%

Ultrapur®, de l’eau milliQ et des acides nitrique, chlorhydrique et acétique Pro

Analysis®, qui sont distillés à l’aide des boilers en quartz, tout comme l’éthanol. Une

purification plus poussée est nécessaire concernant l’acide fluorhydrique, qui est distillé 2

à 3 fois pour atteindre des blancs de W satisfaisants. Les blancs mesurés pour tous ces

réactifs sont inférieurs à 1 ou 2 pg/ml.

2.3.1 Procédure sur les métaux

Les fractions de métaux sont entièrement dissoutes dans des béchers en téflon

Savillex de 15ml sur plaques chauffantes à 120°C en utilisant une solution de HCl 6M et

HF 0,06M. A ce stade, un aliquot de 1 à 10% est prélevé et spiké avec un traceur mélangé

enrichi en 180Hf et 183W afin de déterminer les concentrations en Hf et en W par dilution

isotopique.

La méthode de séparation du W par chromatographie échangeuse d’ions utilisée

pour les métaux (Table 2.1) est similaire à celle décrite par Kleine et al. (2004). Après

digestion, les échantillons sont évaporés, redissous en HF 1M-HNO3 0,1M et chargés sur

une colonne échangeuse d’anions (colonne Biorad® de 2 ml , 2 ml de résine AG1X8,

200-400 mesh) prélavée (voir l’étape de prélavage exposé dans la Table 2.1). La matrice

est éliminée de la colonne en utilisant 10 volumes de résine de HF 1M-HNO3 0,1M puis

le W est élué avec du HNO3 6M-HF 0,2M. Après évaporation, les fractions éluées sont à

nouveau dissoutes à l’aide d’un mélange HNO3 1M-HF 0,1M et chargées sur une colonne

échangeuse d’anions prélavée (colonne Biorad® de 2 ml , 1 ml de résine AG1X8, 200-

400 mesh) pour purification. La matrice est rincée avec du HF 1M-HNO3 0,1M à

nouveau puis avec du HCl 6M-HF 0,01M qui permet l’élimination des traces d’Hf. Le W

est ensuite élué en HCl 6M-HF 1M. Le rendement de cet procédure est proche de 100%.

44

Page 45: Touboul Thesis Completed

Table 2.1: Procédure de séparation chromatographique du W destiné aux mesures de composition isotopique dans le cas des métaux.

étapes nb. volumes de résine solution d’élution

Chimie sur colonne anionique (Biorad AG1X8, 200-400 mesh, 2ml) conditionnement 2 x 5 HNO3 1M-HF 0,1M

chargement 5 HNO3 1M-HF 0,1M rinçage de la matrice 2 x 5 HNO3 1M-HF 0,1M

élution du W 5 HNO3 6M-HF 0,2M prélavage 5 HNO3 3 M

5 HNO3 6M-HF 0,2M 2 x 1 HCl 0,5mM-HF 0,5mM 5 HCl 6M-HF 1M 5 HCl 0,5M-HF 0,5M 2 x 1 HCl 0,5mM-HF 0,5mM

Chimie de purification sur colonne anionique (Biorad AG1X8, 1ml) équilibration 2 x 5 HNO3 1M-HF 0,1M chargement 5 HNO3 1M-HF 0,1M

rinçage de la matrice 10 HNO3 1M-HF 0,1M rinçage du HNO3 2 x 1 HCl 0,5mM –HF 0,5mM

élution du W 6 HCl 6M –HF 1M prélavage 2 x 1 HCl 0,5mM-HF 0,5mM

2 x 5 HNO3 3 M 2 x 5 HNO3 6M-HF 0,2M 2 x 1 HCl 0,5mM-HF 0,5mM 2 x 5 HCl 6M-HF 1M 2 x 5 HCl 0,5M-HF 0,5M 2 x 1 HCl 0,5mM-HF 0,5mM

*HFSE= high field strengh elements. 2.3.2 Procédure sur les fractions silicatées et les roches totales

Cette procédure (Table 2.2) a été adaptée à partir de la méthode décrite par Kleine

et al. (2004). Les fractions silicatées et les roches totales sont attaquées dans des béchers

Savillex de 60 ml sur plaque chauffante à 180°C avec un mélange HF-HNO3 concentrés

(7:3). Après digestion les échantillons sont évaporés et dissous en HNO3 14M

accompagné de quelques gouttes de H2O2 (30%) afin de digérer les composés organiques.

Après évaporation, les échantillons sont repris en HCl 6M-HF 0,06M. A ce stade, la

dissolution des échantillons est complète et une aliquote de 1 à 10% est prélevée pour la

dilution isotopique. Après aliquotage, les 90 à 99% de solution restante sont évaporés

puis sont redissous en HF 4M. Les solutions sont ensuite centrifugées et les résidus lavés

plusieurs fois avec du HF 4M afin d’assurer un lessivage quantitatif du W à partir des

précipités fluorés.

45

Page 46: Touboul Thesis Completed

La solution contenant l’échantillon en HF 4M est ensuite chargée sur une colonne

de résine échangeuse d’anions (3,5 ml de résine AG1X8, 100-200 mesh) préalablement

prélavée (Table 2.2). La matrice est éliminée à l’aide de 6 volumes de colonne d’HF 4M

et le W est élué en HNO3 6M-HF 0,2M. Après évaporation, les fractions de W sont

reprises en HCl 1M-HF 0,5M et chargées sur une colonne de résine échangeuse d’anions

prélavée (3 ml de résine AG1X8, 100-200 mesh) pour purification. La matrice est

éliminée à l’aide de 2 volumes de résine de HCl 0,5M-HF 0,5M. Le titane est ensuite élué

grâce à ∼120ml d’un mélange d’acide acétique (HAc), d’HNO3 et d’H2O2. Le Zr, Hf et le

Nb ainsi que les traces résiduelles de Cr sont éliminées par 5 volumes de résine de HCl

6M-HF 0,01M puis le W est élué avec 15 ml d’HCl 6M-HF 1M. Les blancs de cette

procédure varient entre 100 et 300 pg et proviennent pour l’essentiel de mobilisation de

W à partir des béchers lors de la première attaque des échantillons (de l’ordre de

centaines de pg au total) et dans une moindre part de l’acide acétique et fluorhydrique (de

l’ordre de dizaines de pg au total). Ceci est négligeable excepté pour les échantillons les

plus pauvres en W comme certaines fractions NM, où les blancs représentent jusqu’à

10% du W présent. Le rendement de cette procédure, généralement compris entre 70 et

90%, est variable d’un échantillon à un autre. Les pertes de W sont probablement liées à

la présence de W résiduel dans les précipités fluorés.

46

Page 47: Touboul Thesis Completed

Table 2.2 : Procédure de séparation chromatographique du W destiné aux mesures de composition isotopique dans le cas des fractions silicatées et des roches totales.

Étapes Volume de résine Solution d’élution

Chimie sur colonne anionique (Biorad AG1X8, 100-200 mesh, 3,5ml) conditionnement 2x 5 HF 4M

chargement 1 HF 4M rinçage de la matrice 2 HF 4M

élution du W 5 HNO3 6M-HF 0,2M prélavage 2x5 HNO3 3 M

2x5 HNO3 6M-HF 0,2M 2x 1 HCl 0,5mM-HF 0,5mM 2x5 HCl 6M-HF 1M 2x5 HCl 0,5M-HF 0,5M 2x 1 HCl 0,5mM-HF 0,5mM

Chimie de purification sur colonne anionique (Biorad AG1X8, 100-200 mesh, 3ml) équilibration 2x 5 HCl 0,5M-HF 0,5M chargement 1 HCl 1M-HF 0,5M

rinçage de la matrice 2 HCl 0,5M-HF 0,5M rinçage de l’HF 3x 1 HCl 0,5mM-HF 0,5mM

élution du Ti ~40 Hac 3,6M-HNO3 8mM-H2O2 2% rinçage du Ti 3x 1 HAc 9 M

élution des HFSE* dont Hf 4 HCl 6M-HF 0,01M élution du W 5 HCl 6M-HF 1M

prélavage 2x5 HNO3 3 M 2x5 HNO3 6M-HF 0,2M 2x 1 HCl 0,5mM-HF 0,5mM 2x5 HCl 6M-HF 1M 2x5 HCl 0,5M-HF 0,5M 2x 1 HCl 0,5mM-HF 0,5mM

*HFSE= high field strengh elements.

2.3.3 Procédure particulière développée pour les plagioclases d’anorthosites lunaires

Du fait de la très faible concentration en W des anorthosites (~2-4 ppb W), une

grande quantité d’échantillon (~4g) est nécessaire pour obtenir des données isotopiques

suffisamment précises. Les blancs doivent être suffisamment bas (<100 pg), afin que les

corrections de blanc de la composition isotopique mesurée soient négligeables, et les

rendements doivent être améliorés. De plus, la concentration élevée des échantillons en

calcium requiert une élimination efficace de la matrice avant toute purification plus

poussée du W par la technique utilisant des mélanges contenant de l’HF, décrite dans la

section précédente pour les fractions silicatées et les roches totales.

Les 4 g d’échantillon sont attaqués avec 40 ml d’HF 29M et HNO3 14M dans des

béchers Savillex de 60ml sur plaque chauffante pendant trois jours à 120°C. Bien que

47

Page 48: Touboul Thesis Completed

nécessitant une durée plus importante, cette attaque à température relativement basse

permet de limiter le relargage de W par le téflon des béchers. Après évaporation, les

échantillons sont repris plusieurs fois en HNO3-H2O2 pour éliminer les composés

organiques et détruire les précipités fluorés formés lors de la digestion. Les échantillons

sont alors redissous en HCl 6M-HF 0,06M. 50 ml de cette solution sont ajoutés,

l’ensemble est homogénéisé par ultrasonification puis centrifugé. Cette opération est

répétée jusqu’à dissolution complète des résidus, qui est atteinte dans environ 450 ml

d’HCl 6M-HF 0,06M, qui sont transférés dans un bécher Savillex de 500 ml. Une

aliquote de 5% est alors prélevée et spikée pour la dilution isotopique. Le reste est

évaporé puis dissous dans 30ml d’HCl 1M-HF 0,1M. Cette solution est centrifugée,

décantée et le résidu est rincé deux fois avec 10 ml d’HCl 1M-HF 0,1M pour assurer la

mobilisation complète du W contenu dans les précipités fluorés.

La procédure permettant l’élimination de l’essentiel de la matrice (Table 2.3)

s’effectue sur résine échangeuse de cations et est adaptée de la méthode développée par

Tatsumoto et Patchett (1980). Les 50 ml de solution d’échantillon en HCl 1M-HF 0,1M

sont chargés sur 10 colonnes échangeuses de cations prélavées (15 ml de résine BioRad®

AG50WX8, 200-400 mesh). Le W (ainsi que les autres HFSE) est élué directement puis

rincé avec 15 nouveaux ml d’HCl 1M-HF 0,1M, tandis que les autres éléments de la

matrice, notamment l’Al et le Ca, restent adsorbés sur la résine. Les fractions de W sont

ensuite évaporées, redissoutes dans 10 ml de HCl 0,5 M-HF 0,5M puis chargées sur une

colonne échangeuse d’anions prélavée (3 ml de résine Biorad AG1X8, 100-200 mesh)

pour purification. La procédure suit alors le protocole de la chimie de purification décrite

dans la section précédente pour les fractions silicatées et les roches totales (table 2.2) à

cela près qu’elle est répétée une fois afin d’éliminer efficacement l’ensemble du titane.

Le rendement de cette nouvelle procédure est toujours supérieur à 95%, bien

meilleur que celui de la procédure précédente, car la diminution de la quantité de

précipités fluorés avant le chargement sur colonne permet de limiter les pertes de W

résiduel. Les blancs de procédure ont été diminués à 50 pg et sont devenus beaucoup plus

reproductibles (2 SD = ±20 pg). Cette amélioration des blancs a été réalisée grâce à

l’emploi de température plus basse lors de la première attaque acide, de béchers lavés

plusieurs semaines en HF concentré distillé, dont les blancs sont toujours vérifiés avant

48

Page 49: Touboul Thesis Completed

usage, et d’acide acétique distillé deux fois. Une fois développée, cette méthode a donc

été conservée ensuite, notamment pour la préparation des eucrites parmi lesquels les

eucrites cumulats sont également caractérisées par de faible teneur en W.

Table 2.3 : Procédure de séparation chromatographique du W développée pour le cas des anorthosites lunaires.

étapes volume de résine solution d’élution

Chimie sur colonne cationique (Biorad AG50WX8, 200-400 mesh, 15 ml) équilibration 2x 5 HCl 1M-HF 0,1M

Chargement (élution du W) 0,75 HCl 1M-HF 0,1M rinçage du W 1 HCl 1M-HF 0,1M

prélavage 2 HF 4M 2 HCl 6M 2 HF 4M 2 HCl 6M 2x 1 HCl 0,5mM-HF 0,5mM

Chimie de purification sur colonne anionique (Biorad AG1X8, 100-200 mesh, 3ml) équilibration 2x 5 HCl 0,5M-HF 0,5M chargement 1 HCl 1M-HF 0,5M

rinçage de la matrice 2 HCl 0,5M-HF 0,5M rinçage de l’HF 3x 1 HCl 0,5mM-HF 0,5mM

élution du Ti ~40 Hac 3,6M-HNO3 8mM-H2O2 2% rinçage du Ti 3x 1 HAc 9 M

élution des HFSE* dont Hf 4 HCl 6M-HF 0,01M élution du W 5 HCl 6M-HF 1M

prélavage 2x5 HNO3 3 M 2x5 HNO3 6M-HF 0,2M 2x 1 HCl 0,5mM-HF 0,5mM 2x5 HCl 6M-HF 1M 2x5 HCl 0,5M-HF 0,5M 2x 1 HCl 0,5mM-HF 0,5mM

*HFSE= high field strengh elements. 2.4 Spectromètrie de masse

Toutes les mesures isotopiques ont été réalisées sur un instrument Nu Plasma

MC-ICP-MS à l’ETH Zürich, équipé d’un désolvateur Cetac Aridus et d’un nébuliseur en

teflon de type PFA (débit ~ 80 μl.min-1). Avant la mesure, les échantillons sont redissous

à chaud et évaporés plusieurs fois en HNO3-H2O2 afin de détruire les composés

organiques et d’éliminer l’Os qui forme des oxides volatiles. L’échantillon est ensuite

repris en HNO3 0,56M-HF 0,24M (0,2 à 0,5ml) pour la mesure.

49

Page 50: Touboul Thesis Completed

2.4.1 Protocole de mesure des compositions isotopiques du W

Les compositions isotopiques sont typiquement mesurées avec un faisceau d’une

intensité de ~2 V à la masse du 182W, ce qui correspond à une solution de W à ~20 ppb.

La configuration des cages de Faraday employée, pour laquelle le 182W est collecté au

niveau de la cage L1, est décrite dans la table 2.4. Chaque mesure comprend 60 cycles

(temps d’intégration de 5 s) séparés en 3 blocs conduisant à une précision interne de

l’ordre de 0,2 ε. Une ligne de base électronique (bruit de fond∼-5 mV) est mesurée

pendant 30 s par déflection du faisceau et un centrage de pic sur le 182W effectué avant

chaque bloc. Pour les échantillons les plus pauvres en W (certaines des fractions NM des

acapulcoites et des lodranites, les plagioclases d’anorthosites lunaires et certaines eucrites

cumulats), les compositions isotopiques ont été mesurées en 1 ou 2 blocs de 20 cycles

avec une intensité de signal de 0,2 à 1,5 V sur le 182W, résultant en une précision interne

de ~0,4 à ~1,5 ε. La ligne de base n’est alors mesurée qu’une seule fois juste avant

l’introduction de l’échantillon pendant 120 s avec un temps d’intégration de 0,2 s. On

utilise la méthode dite de «bracketing», consistant à intercaler l’échantillon entre deux

mesures isotopiques du standard ALFA AESAR, ce qui permet de calculer les variations

relatives de la composition isotopique du W de l’échantillon par rapport à la composition

terrestre. Les compositions isotopiques des échantillons sont ainsi exprimées en unités

ε18iW qui correspondent à la déviation du rapport 18iW/184W par rapport à la valeur du

standard terrestre en parts par 10.000. Entre chaque mesure, le lavage du système

d’introduction nécessite le passage d’HNO3 0,56M-HF 0,24M pendant ~5 minutes

(juqu’au retour du signal de 182W à moins de 0.1-0.2 mV de la ligne de base

électronique). Les gains des collecteurs de Faraday tendent à se stabiliser après une

journée de fonctionnement (qui n’est donc utilisée qu’à l’optimisation du signal) et ne

nécessitent ensuite d’être étalonnés qu’une fois par jour en début de session. La

reproductibilité externe et la justesse des mesures effectuées peuvent être estimées

notamment par des mesures répétées du standard ALFA AESAR et de solutions

d’échantillon de composition isotopique connue (voir section 2.4.3).

50

Page 51: Touboul Thesis Completed

Table 2.4 : Configuration de cage pour les mesures de composition isotopique du tungstène (W IC) et les dilutions isotopiques pour le W (W ID), pour l’hafnium (Hf ID) et pur le tantale (Ta ID) N° cage H6 H5 H4 H3 H2 H1 Axiale L1 L2 L3 W IC 188Os - 186W - - 184W 183W 182W - -

W ID 188Os - 186W - - 184W 183W 182W - -

Hf ID - - - 182W 181Ta 180Hf 179Hf 178Hf 177Hf 176Hf

Ta ID - - - - - - 182W 181Ta 180Ta 179Hf

2.4.2 Discrimination de masse instrumentale et interférences

En spectrométrie de masse, les rapports isotopiques mesurés sont différents des

rapports isotopiques de l’échantillon, car ils sont affectés par la présence d’interférences

isobares et par la discrimination de masse instrumentale. Ce dernier facteur en

spectrométrie à source plasma est lié à une meilleure extraction des isotopes lourds par

rapport aux plus légers, qui sont préférentiellement déviés de leur trajectoire lors de

collisions entre ions au sein du faisceau. On notera que la discrimination de masse induite

par les spectromètres à source plasma résulte en un effet opposé à celui du

fractionnement de masse s’opérant dans les spectromètres à thermo-ionisation, qui

favorise l’évaporation des isotopes les plus légers. La discrimination de masse théorique

suit une loi exponentielle (Fig. 2.1) : β

⎟⎟⎠

⎞⎜⎜⎝

⎛×⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

j

i

mesj

i

echj

i

MM

WW

WW avec i et j = 182, 183, 184,186, M=masse atomique,

β = facteur de discrimination de masse

Les rapports isotopiques mesurés peuvent être ainsi corrigés de la discrimination

de masse en utilisant un rapport isotopique préalablement connu tel que les rapports

d’isotopes stables. A cet effet, le W présente l’avantage de posséder quatre isotopes

stables (180W, 183W, 184W, 186W). Peu abondant et présentant une inférence importante

avec le 180Hf, le 180W n’est pas utilisé. Le tungstène nous offre donc deux possibilités de

normalisation des rapports 182W/184W : à l’aide des rapports 186W/183W (=1.9859) ou 186W/184W (=0,92767) ainsi qu’une possibilité de vérifier la justesse des rapports corrigés

grâce au troisième rapport d’isotopes stables (183W/184W) :

51

Page 52: Touboul Thesis Completed

⎥⎥⎥⎥⎥

⎢⎢⎢⎢⎢

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎛×⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛mes

ech

WW

WW

MM

mesech MM

WW

WW

184

186

184

186

186

184

184

182184

182

184

182 ln

⎥⎥⎥⎥⎥

⎢⎢⎢⎢⎢

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎛×⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛mes

ech

WW

WW

MM

mesech MM

WW

WW

183

186

183

186

186

183

184

182184

182

184

182 ln

⎥⎥⎥⎥⎥

⎢⎢⎢⎢⎢

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎟⎟⎠

⎞⎜⎜⎝

⎛×⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛mes

ech

WW

WW

MM

mesech MM

WW

WW

183

186

183

186

186

183

184

183184

183

184

183 ln

Un autre biais vient perturber les mesures isotopiques. Bien que la plupart des

éléments autres que le W ait été éliminé par la séparation chimique, certains isobares

résiduels sur les masses des isotopes du W peuvent être à l’origine d’interférences. C’est

le cas notamment de l’Os créant de petites interférences isobares aux masses 184 et 186.

Celles-ci peuvent être corrigées en contrôlant le 188Os mais ces corrections sont toujours

de l’ordre de quelques ppt sur le rapport 182W/184W et sont donc négligeables.

Figure 2.1: Ln (183W/184W) mesuré versus ln(182W/184W) mesuré pour le standard ALFA AESAR au cours de cette étude. Les données sont en parfait accord avec la discrimination de masse théorique. La loi exponentielle permet donc une correction appropriée du fractionnement instrumental.

52

Page 53: Touboul Thesis Completed

2.4.3 Reproductibilité externe et justesse des mesures

La reproductibilité externe peut être estimée par des mesures répétées du standard

de W ALFA AESAR (fig 2.2 et Table 2.5) ou par des mesures répétées de solutions

d’échantillon de métaux ou de silicates (voir tables de résultats des chapitres 3 et 4). Les

deux méthodes conduisent à des résultats similaires, indiquant que les effets de matrice

n’altèrent pas la reproductibilité externe des mesures. Pour une solution de 20 ppb

mesurée pendant 60 cycles, la reproductibilité externe est typiquement de ~30 ppm (2

SD) pour le rapport 182W/184W normalisé au 186W/183W, de ~60 ppm (2 SD) pour le

rapport 182W/184W normalisé au rapport 186W/184W et de ~40 ppm (2 SD) pour le rapport 183W/184W. La reproductibilité externe avec une normalisation 186W/183W étant meilleure

que celle obtenue avec une normalisation au 186W/184W, on utilisera préférentiellement

les compositions isotopiques calculées en utilisant la normalisation au 186W/183W. Pour

des solutions de 2 à 10 ppb mesurée pendant un bloc, elle se détériore pour atteindre 200

et 80 ppm respectivement sur le rapport 182W/184W normalisé au 186W/183W.

Figure 2.2: reproductibilité externe sur le 182W/184W normalisé au 186W/183W pour dix mesures du standard ALFA AESAR en fonction de sa concentration et de la durée d’une mesure. Les courbes pointillées correspondent à une extropolation exponentielle des données pour des mesures isotopiques comportant 20 ou 60 cycles.

53

Page 54: Touboul Thesis Completed

Table 2.5 : Reproductibilité externe sur les rapports 182W/184W normalisé au 186W/183W, 182W/184W normalisé au 186W/184W et 183W/184W normalisé au 186W/183W pour dix mesures (de 60 cycles chacune) du standard ALFA AESAR pour différentes concentrations.

Reproductibilité externe (2SD, n=10, 60 cycles) Concentration en W de la solution

182W/184W norm. au 186W/184W

182W/184W norm. au 186W/183W

183W/184W norm. au 186W/183W

2.5 ppb 164 ppm 96 ppm 94 ppm10 ppb 91 ppm 55 ppm 46 ppm20 ppb 61 ppm 31 ppm 38 ppm40 ppb 50 ppm 21 ppm 30 ppm

La justesse des mesures peut être vérifiée de trois façons différentes.

Premièrement, en vérifiant l’accord des rapports isotopiques obtenus avec les deux

normalisations. Deuxièmement en vérifiant que la valeur obtenue pour le rapport 183W/184W est identique dans la marge d’erreur à la valeur du standard terrestre (voir

tables de résultats des chapitres 3, 4 et 5). Troisièmement, en mesurant un échantillon de

composition isotopique connue. Les mesures répétées au cours de cette étude d’une

poudre de roche d’Allende conduisent à une valeur en accord avec la composition

isotopique des chondrites carbonées rapportée précédemment (ε182W = -1,9 ± 0,1, Kleine

et al., 2004).

2.5 Mesure de concentrations par dilution isotopique

Pour obtenir une isochrone, il est nécessaire de déterminer les rapports Hf/W des

échantillons et donc de mesurer leur concentration en Hf et en W. Celles-ci sont obtenues

par la méthode dite de dilution isotopique, basée sur la détermination de la composition

isotopique d’un mélange d’une quantité connue de ‘spike’ et d’une quantité inconnue

d’un élément de composition isotopique naturelle. Un spike est une solution contenant

une concentration connue d’un élément dont la composition isotopique a été enrichie par

un ou plusieurs de ses isotopes et est connue. Dans cette étude, nous avons employé un

spike enrichi en 180Hf, 183W, 180Ta, 233U et 229Th (Table 2.6). Celui-ci est ajouté aux

aliquotes d’échantillon prélevées lorsque la dissolution est complète (voir sections 2.3.1

et 2.3.2). La matrice des mélanges spike-échantillon (contenant également l’U et le Th)

ainsi que l’Hf, le W et le Ta (du fait des interférences entre 180Hf, 180Ta et 180W) doivent

être séparés avant la mesure des compositions isotopiques, qui s’effectuent donc

54

Page 55: Touboul Thesis Completed

séparément pour chacun des éléments. La procédure de séparation simplifiée sur colonne

échangeuse d’anions est celle décrite par Kleine et al. (2004) et est présentée dans la table

2.7. Les blancs de la procédure pour la dilution isotopique sont généralement compris

entre 10 et 20 pg de W et entre 1 et 2 pg d’Hf et de Ta.

Table 2.6: Compositions isotopiques et concentrations en W, Hf, Mo, Ta, U et Th pour le spike Mo-W I. Les abondances isotopiques naturelles sont également indiquées pour comparaison.

Elements / Isotopes Spike Mo-W I composition naturelle

[W] 97,7

[Hf] 454,8

[Ta] 8,62

[Th] 0,456

Concentration (ppb)

[U] 567,9

Ab 183W 0,9627 0,1431

Ab 184W 0,0158 0,3064

Ab 186W 0,0001 0,2843

Ab 180Hf 0,9827 0,3509

Ab 179Hf 0,0073 0,1362

Ab 177Hf 0,0026 0,1860

Ab 180Ta 0,0546 0,0001

Ab 181Ta 0,9454 0,9999

Ab 229Th 0,9955 0

Ab 232Th 0,0045 ~100

Ab 233U 0,9928 0

Abondance isotopique

Ab 238U 0,0071 0,9927

Afin de déduire les concentrations en W et en Hf de l’échantillon à partir des

compositions isotopiques mesurées, un calcul itératif est nécessaire, car les rapports

devant être utilisés pour déterminer la discrimination de masse (186W/184W et 179Hf/177Hf

55

Page 56: Touboul Thesis Completed

respectivement) n’ont pas leur valeur naturelle dans le mélange spike-échantillon, du fait

de la composition isotopique anormale du spike. Ce calcul est initié en utilisant

directement les rapports 183W/184W et 180Hf/177Hf mesurés (non corrigée de la

discrimination de masse) et finit par converger après quelques itérations. Ce calcul des

concentrations en Hf et en W de l’échantillon nécessite également la connaissance de leur

concentration, de leur composition isotopique dans le spike et des quantités relatives

d’échantillon et de spike dans le mélange.

Table 2.7 : Procédure de séparation chromatographique de l’Hf, du W, du Ta pour les aliquots spikés destinés à la dilution isotopique. L’uranium et le thorium sont récupérés avec la matrice et seront ensuite purifié selon la procédure décrite par Vigier (2000)

étapes volume de résine solution d’élution

Chimie sur colonne anionique (1 ml de résine Biorad AGX8, 200-400 mesh) équilibration 2x 5 HCl 0,5M-HF 0,5M chargement 1,5 HCl 1M-HF 0,5M

rinçage de la matrice (+ U + Th) 1 HCl 0,5M-HF 0,5M 5 HF 1M 3x 1 HCl 0,5mM-HF 0,5mM

élution de l’Hf 3 HCl 6M-HF 0,01M élution du W 4 HCl 6M-HF 1M rinçage HCl 3x 1 HCl 0,5mM-HF 0,5mM rinçage Mo 5 HNO3 6M-HF 0,2M Elution Ta 5 HNO3 6M-HF 0,2M-H2O2 1% prélavage 2x5 HNO3 3M

2x5 HNO3 6M-HF 0,2M 2x 1 HCl 0,5mM-HF 0,5mM 2x5 HCl 6M-HF 1M 2x5 HCl 0,5M-HF 0,5M 2x 1 HCl 0,5mM-HF 0,5mM

2.6 Conclusions

Nous avons donc développé une méthode de séparation magnétique efficace,

capable d’assurer la pureté des fractions de métaux obtenues, ce qui est critique dans le

cas des échantillons lunaires, mais permet également d’obtenir des fractions silicatées à

minéralogie contrastée. Ceci s’avère être un moyen efficace de séparation de fractions à

rapport Hf/W varié, ce qui facilite grandement la réalisation d’isochrones internes. La

nécessité de mesures isotopiques sur de faibles quantités de W (∼6-8 ng) provenant de

quantités importantes d’échantillon (∼4 g), dans le cadre de notre étude des anorthosites

56

Page 57: Touboul Thesis Completed

lunaires, nous a conduit à adapter la procédure initiale de séparation du W. En évitant la

formation de précipités fluorés massifs par l’emploi d’une colonne échangeuse de

cations, qui permet l’élimination de la plus grande partie de la matrice de l’échantillon,

les rendements de procédure sont passés de ∼70% à >90%. De plus, les blancs de W et

leur reproductibilié ont été améliorés significativement (de 250 ± 150 à 50 ± 20 pg),

essentiellement grâce aux lavages intensifs en HF distillé des béchers Savillex et à une

température d’attaque maintenue volontairement relativement basse (120°C contre 180°C

auparavant) mais également, bien que dans une moindre mesure, à une distillation plus

poussée des acides acétique et fluorhydrique.

Références

Markowski, A., Quitté, G., Halliday, A. N., and Kleine, T., 2006. Tungsten isotopic

compositions of iron meteorites: chronological constraints vs. cosmogenic effects.

Earth Planet. Sci. Lett. 242, 1-15.

Harper, C. L., Völkening, J., Heumann, K. G., Shih, C. Y., and Wiesmann, H., 1991. 182Hf-182W: New cosmochronometric constraints on terrestrial accretion, core

formation, the astrophysical site of the r-process, and the origin of the solar system.

Lunar Planet. Sci. Conf. XXII, 515-516.

Harper C.H.J., Jacobsen S.B., 1996. Evidence for 182Hf in the early solar system and

constraints on the timescale of terrestrial accretion and core formation. Geochim.

Cosmochim. Acta 60, 1131-1153.

Kleine T., Mezger K., Münker C., Palme H., and Bischoff A. (2004) 182Hf-182W isotope

systematics of chondrites, eucrites, and Martian meteorites: Chronology of core

formation and mantle differentiation in Vesta and Mars. Geochim. Cosmochim.

Acta 68, 2935-2946.

Kleine, T., Palme, H., Mezger, K., and Halliday, A. N., 2005. Hf-W chronometry of lunar

metals and the age and early differentiation of the Moon. Science 310, 1671-1674.

Patchett P.J., Tatsumoto, M., 1980. A routine high precision method for Lu-Hf

geochemistry and chronology. Contribution to Minerlogy and petrology 75(3), 263-

267.

57

Page 58: Touboul Thesis Completed

Vigier N., 2000. Apport des séries de l’Uranium sur les temps caractéristiques des

processus d’érosion, thèse IPGP.

Yin, Q. Z., Jacobsen, S. B., Yamashita, K., Blichert-Toft, J., Télouk, P., and Albarède, F.,

2002. A short timescale for terrestrial planet formation from Hf-W chronometry of

meteorites. Nature 418, 949-952.

58

Page 59: Touboul Thesis Completed

59

Chapitre 3

Thermochronométrie Hf-W des météorites :

Contraintes sur l’accrétion et l’évolution thermique des corps parents

Hf-W thermochronmetry of meteorites: Constraints on the accretion and thermal evolution of parent bodies

Page 60: Touboul Thesis Completed

60

Page 61: Touboul Thesis Completed

61

Section 3.1

Accretion and thermal history of the acapulcoite-lodranite parent body inferred from Hf-W

thermochronometry*

M. Touboul1, T. Kleine1, B. Bourdon1, J. A. Van Orman2, C. Maden1, A. J. Irving3, J. Zipfel4, T.E. Bunch5

1Institute for Isotope Geochemistry and Mineral Resources, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerland

2Department of Geological Sciences, Case Western Reserve University, Cleveland, OH, USA. 3Forschungsinstitut und Naturmuseum Senckenberg, Frankfurt am Main, Germany. 4Departement of Earth & Space Sciences, University of Washington, Seattle, WA, USA.

5Department of Geology, Northern Arizona University, Flagstaff, AZ, USA

* En préparation pour Geochimica et Cosmochimica Acta

Page 62: Touboul Thesis Completed

62

Page 63: Touboul Thesis Completed

63

Abstract

We obtained Hf-W metal-silicate isochrons for several acapulcoites and lodranites to

constrain the accretion timescale and thermal history of their parent body. The high 180Hf/184W (up to ~60) and radiogenic 182W/184W ratios (up to ~40 ε182W) in some silicate

separates combined with low 180Hf/184W and unradiogenic 182W/184W ratios (ε182W ~-3) in the

metals make it possible to obtain precise Hf-W ages with uncertainties of less than ±1 Ma.

The Hf-W age for the acapulcoites Dhofar 125 and NWA 2775 is ΔtCAI=5.1±0.9 Ma, which

corresponds to an absolute age of 4563.5±0.7 Ma. The Hf-W age for lodranite NWA 2627 is

ΔtCAI=5.6±1.0 Ma, corresponds to an absolute age of 4563.0±0.9 Ma, and is indistinguishable

from the Hf-W age for acapulcoites. Closure temperatures for the Hf-W system in

acapulcoites and lodranites were estimated from numerical simulations of W diffusion in

high-Ca pyroxene, the major host of radiogenic 182W, and are 975±50 °C for acapulcoites and

1025±50 °C for lodranites. These closure temperatures are only ~200 °C lower than the peak

temperatures, such that the Hf-W ages provide information on the earliest high-temperature

evolution of the acapulcoite-lodranite parent body. The peak temperature of acapulcoites and

lodranites in conjunction with their Hf-W ages require that the acapulcoite-lodranite parent

body accreted later than ~1.5 Ma but before ~2 Ma after CAI formation. Thermal modeling

indicates that acapulcoites and lodranites reached their thermal peak ~3 Ma after CAI

formation and that decay of 26Al has been an important heat source in their thermal evolution.

Cooling rates for acapulcoites decreased from ~120°C/Ma just below the thermal peak to

~50°C/Ma at ~600 °C. Over the same temperature interval the cooling rate for lodranites

decreased from ~100°C/Ma to ~40 °C/Ma. These thermal histories may reflect cooling in the

uppermost ∼10-20 km of a parent body with a radius of ∼50-100 km. The acapulcoites and

lodranites analyzed for this study evolved with 180Hf/184W ratios of 0.34 and 1.88, which are

indistinguishable from those of H chondrites but significantly lower than 180Hf/184W ~1.24 for

carbonaceous chondrites. The low 180Hf/184W ratios of acapulcoites and lodranites were

established before ~2 Ma and, hence, prior to partial melting in the parent body at ~3 Ma and

thus must reflect Hf-W fractionation of the precursor material by processes in the solar

nebula.

Combined with Hf-W ages of ΔtCAI<1 Ma for differentiation of the parent bodies of

magmatic iron meteorites and an Hf-W age of ΔtCAI~2.5 Ma for the accretion of the H

chondrite parent body, the Hf-W results for acapulcoites and lodranites reveal an inverse

Page 64: Touboul Thesis Completed

64

correlation between accretion age of asteroids and peak temperature in their interiors. This

indicates that the different thermal histories of meteorite parent bodies primarily reflect

variations in their initial 26Al abundance, which is determined by their accretion time.

3.1.1 Introduction

Determining the accretion timescale and early thermal history of meteorite parent

bodies is key for understanding the formation of planetesimals and for constraining the

parameters that controlled their subsequent evolution. Such information can be obtained by

applying short- and long-lived chronometers to meteorites (or preferably a suite of meteorites

derived from a single parent body) but also requires knowledge of the closure temperature

(Tc) for diffusive exchange of parent and daughter elements among the different minerals in a

meteorite (Dodson, 1973; Ganguly and Tirone, 2001). Most chronometers such as the U-Pb,

Al-Mg or Ar-Ar systems closed at temperatures much lower than the peak temperatures

reached in most meteorite parent bodies and as such provide constraints on the mid- to low-

temperature evolution of meteorite parent bodies (Göpel et al., 1994; LaTourrette and

Wasserburg, 1998; Trieloff et al., 2003). As a consequence, their applicability for determining

the high-temperature thermal history is limited, except to bodies with very fast cooling rates,

but such information is of special interest because it can provide constraints on the accretion

timescale and the heat source(s) responsible for parent body heating. In contrast, the Hf-W

system (t1/2~9 Ma) has a relatively high closure temperature, which for H chondrites is almost

as high as their peak metamorphic temperatures (Kleine et al., 2008b). Thus, the Hf-W system

appears ideally suited for determining the accretion timescale and high-temperature evolution

of meteorite parent bodies.

Recent advances in the chronology of meteorites reveal that the accretion of the parent

bodies of most if not all differentiated meteorites occurred within the first Ma after formation

of Ca-Al-rich inclusions (CAIs) (Bizzarro et al., 2005; Kleine et al., 2005; Markowski et al.,

2006; Schérsten et al., 2006; Burkhardt et al., 2008). In contrast, accretion of chondrite parent

bodies occurred later, more than ~2 Ma after CAI formation (Kita et al., 2000; Kunihiro et al.,

2004). These results indicate that the early evolution of asteroids was determined by their

initial 26Al content, which was sufficiently abundant in the early-formed planetesimals to

cause global melting and differentiation but had decayed to levels insufficient to trigger

melting in the late-formed chondrite parent bodies (Bizzarro et al., 2005; Kleine et al., 2005;

Schérsten et al., 2006). Acapulcoites and lodranites are a group of meteorites with properties

Page 65: Touboul Thesis Completed

65

that place them intermediate between the unmelted chondrites and the differentiated

meteorites (Palme et al., 1981; Zipfel et al., 1995; McCoy et al., 1996; McCoy et al., 1997a)

and as such are of special interest for understanding the differentiation of asteroids as well as

the role of heating by 26Al decay in the thermal history of asteroids. Acapulcoites and

lodranites have mineral and bulk compositions similar to those of ordinary chondrites but they

have non-chondritic textures. These result from recrystallization during high-temperature

metamorphism and from partial melting. The acapulcoites show evidence for partial melting

at the FeS-FeNi cotectic and in some cases silicate melts may have formed (Palme et al.,

1981; Zipfel et al., 1995). The lodranites were heated to higher temperatures than the

acapulcoites and are partial melting residues depleted in troilite and plagioclase (McCoy et

al., 1997a). Several lines of evidence indicate that acapulcoites and lodranites, despite their

different thermal histories, were derived from a common parent body: the occurrence of

meteorites that are transitional between acapulcoites and lodranites (McCoy et al., 1997a);

similar O isotope compositions (Clayton and Mayeda, 1996); similar mineral compositions

(McCoy et al., 1997b); and indistinguishable cosmic ray exposure ages (Eugster and

Lorenzetti, 2005).

To constrain the accretion timescale and high-temperature thermal history of the

acapulcoite-lodranite parent body, we applied the Hf-W chronometer to several acapulcoites

and lodranites. The closure temperatures of the Hf-W system in these meteorites were

estimated numerically using the model of Van Orman et al. (2001, 2006). The Hf-W ages and

Tc estimates are then used in conjunction with previously published ages for acapulcoites and

lodranites and thermal modeling to constrain the accretion timescale and thermal history of

the acapulcoite-lodranite parent body. Finally, these results are compared to the thermal

history of other meteorite parent bodies.

3.1.2 Samples and analytical techniques

3.1.2.1 Samples

Four acapulcoites (NWA 2656, NWA 2775, Dhofar 125) and 2 lodranites (NWA

2627, NWA 4663) were selected for determining internal Hf-W isochrons. In addition, Hf-W

data for a ~700 mg whole-rock of the lodranite Gibson were obtained. In Fig. 3.1.1, back-

scattered electron images for some of these samples are shown. Acapulcoites and lodranites

are predominantly composed of low-Ca pyroxene and olivines and also contain high-Ca

pyroxene, sodic plagioclases, troilite, metal and traces of apatite and chromite. They show

Page 66: Touboul Thesis Completed

66

characteristic recrystallization texture with abundant 120° triple junctions and exhibit only

minor shock effect (S1-S2) (McCoy et al., 1997a, Mittelfehldt et al., 1996, Palme et al.,

1981).

Detailed descriptions of the acapulcoites and lodranites investigated for this study can

be found elsewhere (Greshake et al., 2001, Connelly et al., 2008, Irving et al., 2007). NWA

2775 and 2656 have an average grain size within the typical range of acapulcoites (150-230

μm), whereas Dhofar 125 is somewhat finer grained (∼100μm). The high-Ca pyroxenes in

NWA 2656 have an average grain size of 115±12μm, calculating by averaging long and short

axes, and a modal content of 13 vol.-%. The high-Ca pyroxene in NWA 2775 is larger

(190±12μm) but less abundant (10 vol.-%). The average grain sizes in NWA 2627 and 4663

are larger than in acapulcoites and within the range of lodranites. Note that NWA 2627 has

initially been classified as an acapulcoite but its average grain size clearly indicates that it is a

lodranite. The high-Ca pyroxenes in NWA 2627 and NWA 4663 have average grain sizes of

430±25μm (modal content of 8 vol.-%) and 373±18 μm (modal content of 7 vol.-%). In

addition there are tiny high-Ca pyroxene grains that occur as inclusions in olivine and

probably represent trapped liquid within the crystallizing olivine. All acapulcoites and

lodranites analyzed here show minimal to moderate degree of alteration (W1 and W2, or W3

in the case of NWA 2627), resulting in the some oxidation of metals.

Page 67: Touboul Thesis Completed

metal metal

plag cpx opx + ol

NWA 2775 opx + ol metal

plag cpx

cpx

metal plag

plag cpx opx + ol

metal NWA 2656 metal

opx + ol plag cpx cpx metal cpx

metal metal ol

ph cpx opx

plag opx

opx NWA 4663 ol ol cpx opx metal

ol opx plag Figure 3.1.1: Back-scattered electron (BSE) images for acapulcoites NWA 2775 and NWA 2627, and lodranite NWA 4663. False colors used for NWA 2775 and NWA 2627 are dark grey for FeNi metal and oxide, white for clinopyroxene (cpx), grey for olivine and orthopyroxene (ol+opx) and light grey for plagioclase (plag). False colors used for NWA 4663 are dark grey for FeNi metal and oxide, white for clinopyroxene, grey for orthopyroxene, light grey for plagioclase and phosphate (ph), and very light grey for olivine.

67

Page 68: Touboul Thesis Completed

68

3.1.2.2 Analytical techniques

Pieces of meteorite (with a total weight of ~3-5 g for each sample) were cleaned with

abrasive paper and with 0.05 M HNO3, de-ionized H2O and ethanol in an ultrasonic bath to

remove any contamination introduced during cutting from larger samples. Whole-rock

powders were obtained by crushing a 0.5-1 g piece of each sample in an agate mortar. The

remaining material (~2-4 g) was gently crushed in an agate mortar and separated into <40 μm

and 40-150 μm fractions using nylon sieves. During crushing metal grains were removed

using a hand-magnet and silicate dust attached to or silicate grains intergrown with the metal

grains were removed by repeated crushing of the magnetic fraction under ethanol. Between 15

and 100 mg of pure metal separates were obtained for most samples, except NWA 2656 for

which no pure metal grains could be obtained. Although all visible metal grains were

removed, the 40-150 μm fractions were still slightly magnetic, most likely reflecting the

presence of tiny metal inclusions in the silicate and oxide grains. The 40-150 μm fractions

were further separated using a hand-magnet into a "weakly-magnetic" and "non-magnetic"

fraction. These were then further separated in several weakly- and non-magnetic fractions that

were labeled WM-n and NM-n, n=1, 2, 3…, where n=1 always denotes the least magnetic

fraction among the weakly- and non-magnetic fraction for each sample. All WM and NM

fractions were inspected under the binocular microscope and they consist mainly of olivine

and pyroxene but most fractions also contain some feldspar. All the WM and NM fractions

were cleaned with ethanol in an ultrasonic bath and powdered in an agate mortar. Remaining

metal grains were then removed from these powders using a hand-magnet and the resulting

fractions weighted between 150 and 300 mg.

The metal separates were dissolved in 15 ml Savillex® vials at ~120°C on a hotplate

using 6 M HCl-0.06 M HF. In some cases, a few drops of concentrated HNO3 were added.

The NM fractions were dissolved in 60 ml Savillex® vials at ~180°C on a hotplate using HF-

HNO3 (7:3). After digestion, the samples were dried down and re-dissolved in HNO3-H2O2 to

destroy organic compounds. Then the samples were completely dissolved in 6 M HCl-0.06 M

HF and depending on the W contents a 1-10% aliquot was spiked with a mixed 180Hf-183W

tracer that was calibrated against pure Hf and W metals (Kleine et al., 2004).

The methods for the separation of Hf and W from the sample matrix were similar to

those outlined in Kleine et al. (2008). The metal separates were dried, re-dissolved in 1 M

HF-0.1 M HNO3 and loaded onto pre-cleaned anion exchange columns (2 ml BioRad®

AG1X8, 200-400 mesh). The matrix was washed from the column using 1 M HF-0.1 M

HNO3 and W was eluted in 6 M HNO3-0.2 M HF. After drying down, the W cut was re-

Page 69: Touboul Thesis Completed

69

dissolved in 1 M HF-0.1 M HNO3 and loaded onto a pre-cleaned anion exchange column (1

ml BioRad® AG1X8, 200-400 mesh). Again, the matrix was washed from the column using 1

M HF-0.1 M HNO3, followed by 6 M HCl-0.01 M HF. Tungsten was then eluted in 6 M HCl-

1 M HF.

After aliquoting, the WM and NM fractions were dried and re-dissolved in 4 M HF,

the solution centrifuged and decanted and the residue washed several times with 4 M HF. The

solution was ultrasonicated several times to ensure optimal release of W from the fluoride

residue and was loaded onto pre-cleaned anion exchange columns (3.5 ml BioRad® AG1X8,

100-200 mesh). The matrix was washed from the column using ~6 column volumes of 4 M

HF and W eluted using 6 M HNO3-0.2 M HF. After drying, this cut was re-dissolved in 1 M

HCl-0.5 M HF and loaded onto pre-cleaned anion exchange columns (3 ml BioRad® AG1X8,

100-200 mesh), where W was purified following the procedure of Kleine et al. (2004).

Titanium was washed from the column using a HAc-HNO3-H2O2 mixture, Zr, Hf, and Nb

were rinsed off in 6 M HCl-0.01 M HF and W was finally eluted in 6 M HCl-1 M HF.

Total procedural blanks ranged from ~100 to ~300 pg for the W isotope composition

measurements and ~10 to 30 pg W and ~10 pg Hf for the isotope dilution measurements. For

most samples, blanks are negligible but for some of the most W-depleted fractions blank

corrections on the measured 182W/184W ratios were as high as ∼10% (Table 3.1.1).

All isotope measurements were performed using a Nu Plasma MC-ICP-MS at ETH

Zurich, equipped with a Cetac Aridus desolvating nebuliser. The measurement protocol was

identical to that described in Kleine et al. (2008). Prior to measurement, the samples were re-

dissolved and dried several times in few drops of 14M HNO3 and 30% H2O2 to remove

organic compounds and, in the case of metal-rich samples, volatile Os oxides and then taken

up in a 0.56 M HNO3-0.24 M HF mixture. Tungsten isotope compositions of metals and

whole-rocks were typically measured with a signal intensity of ~2 V on 182W, which was

obtained for a ~20 ppb W solution. For these samples, 60 ratios (3 blocks of 20 ratios) were

measured resulting in within-run statistics of the order of 0.2 ε units (2σ). Owing to the low

W contents in some NM fractions, their W isotope compositions were measured in 1 or 2

blocks of 20 ratios each with signal intensities of ~0.2 to 1 V on 182W. The within-run

statistics of these measurements were typically between 0.5 and 1.5 ε units. Instrumental mass

bias was corrected relative to 186W/183W=1.9859 using the exponential law. Small isobaric

interferences of Os on masses 184 and 186 were corrected by monitoring 188Os and were

negligible.

Page 70: Touboul Thesis Completed

70

The 182W/184W and 183W/184W ratios of all samples were determined relative to two

standard runs bracketing the sample run and are reported in ε18iW units, which is the deviation

of the 18iW/184W ratio from the terrestrial standard value in parts per 10,000. The

reproducibility of the ~20 ppb standard during one measurement day is typically equal to or

better than ~0.3 ε units (2 SD) for the 182W/184W ratio and ~0.2 ε units (2 SD) for the 183W/184W ratio. The external reproducibility of the W isotope measurements typically is 0.3-

0.4 ε units (2 SD) for the 182W/184W ratio and 0.2-0.3 ε units (2 SD) for the 183W/184W ratio

(Kleine et al. 2008). The uncertainties for the W isotope measurements of the W-poor NM-1

and -2 fractions of Dhofar 125 and NWA 2775 were assessed by repeated measurement (1

block of 20 ratios each) of ~2 and ~10 ppb W standard solutions that yielded external

reproducibilities of ~2 and ~0.8 ε182W (2 SD). This is similar to the within-run statistics

obtained for the measurements of the W-poor NM fractions (Table 3.1.1). Note that the major

source of uncertainty in the 182W/184W of the W-poor NM fractions is the blank correction

(3.1. 1).

The accuracy of the measurements was monitored by analyzing several carbonaceous

chondrites, which all yielded the previously determined value of -1.9±0.1 ε182W (Kleine et al.,

2004a). Furthermore, 183W/184W ratios were used as a monitor for accurate measurements and

agree for most samples to within ±0.2 ε units with the terrestrial value (Table 3.1.1). Elevated 183W/184W ratios for some fractions from the NWA 2775 acapulcoite and the NWA 4663 and

NWA 2627 lodranites are attributed to an organic interference on mass 183 that was

successfully removed for all other samples by treatment with HNO3-H2O2. Elevated measured 183W/184W ratios have been observed before during W isotope measurements of some

eucrites, carbonaceous chondrites, CAIs and H chondrites and for these samples the 182W/184W ratio normalized to 186W/184W=0.92767 agrees with 182W/184W ratios for other

samples of these groups (Kleine et al., 2002; Kleine et al., 2004; Burkhardt et al., 2008;

Kleine et al., 2008b). This indicates that only 183W is affected, such that for the fractions with

elevated measured 183W/184W the reported ε182W values were calculated from the 182W/184W

ratio normalized to 186W/184W.

3.1.3 Results

The Hf and W concentrations and the W isotope compositions of metals, whole-rocks

and non-magnetic fractions analyzed for this study are given in Table 3.1.1 and shown in Fig.

3.1.2 to Fig. 3.1.4. The acapulcoite and lodranite metals have W concentrations ranging from

Page 71: Touboul Thesis Completed

71

~850 to 1400 ppb and have indistinguishable ε182W values of ~-3.0. All metal separates have

low Hf/ W ratios of less than ~0.05, indicating that pure metal separates were obtained.

The acapulcoite and lodranite w

180 184

hole-rocks have 180Hf/184W ratios ranging from ~0.3 to ~2 and

ε182W values ranging from ~-2.4 to ~-1.5 but no systematic difference between acapulcoites

and lodranites is apparent. The WM fractions have W contents and ε182W between those of

the whole-rocks and metals in most cases, most likely reflecting a higher abundance of metal

in the WM fractions. The NM fractions have variable Hf and W contents ranging from ~90 to

~660 ppb Hf and from ~4 to ~124 ppb W, respectively. Their 180Hf/184W ratios are between

~3 and ~60, resulting in elevated 182W/184W ratios ranging from ~0 to ~40 ε182W. Owing to

the low W contents and radiogenic 182W/184W of some of the NM fractions, blank corrections

were significant and typically ranged from <1 to ~4 ε182W (Table 3.1.1).

As shown in Figs. 3.1.2 and 3.1.3, the 180Hf/184W ratios and ε182W values correlate for

the acapulcoites Dhofar 125 and NWA 2775, and for the lodranite NWA 2627, such that

precise regression could be obtained for these samples (MSWD<1). The uncertainties on the

slopes of these regressions are better than ~7%, resulting in uncertainties for the ages of better

than ±1 Ma, if these regressions are interpreted as isochrons. In contrast, no regression could

be obtained for NWA 2656 and NWA 4663 and this most likely reflects disturbance of the

Hf-W systematics due to terrestrial weathering. In spite of their different Hf/ W ratios, the

NM fractions of NWA 2656 have a terrestrial W isotope composition, suggesting that this

sample has been contaminated with terrestrial W during weathering in the desert. These

effects are less pronounced in the metal and whole-rock fractions because these have much

higher W contents. Note that for NWA 2656 no metal grain could be separated from the

specimen investigated for this study, indicating substantial oxidation of the metals during

terrestrial weathering. Similarly, in spite of very high Hf/ W ratios the NM fractions of

NWA 4663 have much lower

180 184

180 184

ε182W values compared to fractions from other acapulcoites and

lodranites with similar 180Hf/184W. A plot of ε182W vs. 1/W reveals that the NM fractions of

NWA 4663 plot on a mixing line with terrestrial W, suggesting that their low ε182W values are

due to addition of terrestrial W.

Page 72: Touboul Thesis Completed

Figure 3.1.2: ε182W versus 180Hf/184W for acapulcoites Dhofar 125, NWA 2656, NWA 2775. m = initial 182Hf/180Hf, i = initial ε182W. Regressions are calculated using the model 1 fit of Isoplot (Ludwig, 1991). ΔtCAI is the formation age relative to CAIs; the absolute age t is calculated relative to the angrite D'Orbigny. Linear regressions of the Hf-W data for Dhofar 125 and NWA 2775 yield identical slopes and intercepts and the regression of the combined data for these two acapulcoites provides the best estimate for the initial 182Hf/180Hf and ε182W at the time of the last Hf-W equilibration in acapulcoites. No correlation between ε182W and 180Hf/184W is observed for NWA 2656, indicating that the Hf-W system in this sample is disturbed.

Figure 3.1.3: ε182W versus 180Hf/184W for lodranites NWA 2627 and NWA 4663. m = initial 182Hf/180Hf, i = initial ε182W. Regressions are calculated using the model 1 fit of Isoplot (Ludwig, 1991). ΔtCAI is the formation age relative to CAIs; the absolute age t is calculated relative to the angrite D'Orbigny. Linear regressions of the Hf-W data for NWA 2627 yield a precise isochron but no correlation between ε182W and 180Hf/184W is observed for NWA 4663, indicating that the Hf-W system in this sample is disturbed.

72

Page 73: Touboul Thesis Completed

73

3.1.4 Discussion

3.1.4.1 Hf-W isochron ages for acapulcoites and lodranites

To define an isochron the minerals of a sample must once have been in W isotope

equilibrium, i.e., they must have had the same W isotope composition initially. Given that the

acapulcoite and lodranite fractions were obtained mainly based on their magnetic

susceptibility, the correlation of ε182W with 180Hf/184W could potentially represent a mixing

line between W-rich metal and virtually W-free silicates. Such a mixing line would have no

chronological significance if the two endmembers had different initial 182W/184W ratios.

However, as shown in Fig. 3.1.4, the variations in 180Hf/184W ratios among the fractions of the

analyzed acapulcoites and lodranites require the presence of at least three independent

components for Hf and W among the coexisting phases. The major host of W is metal and the

major host of Hf is high-Ca pyroxene. The third component includes olivine and low-Ca

pyroxene and is characterized by low Hf and W contents. These two minerals are considered

here as one component because no pure olivine and low-Ca pyroxene separates were

obtained. Both olivine and low-Ca pyroxene are not capable of incorporating significant

amounts of either Hf or W (Righter and Shearer, 2003), such that their presence mainly

causes dilution of the high Hf content of high-Ca pyroxene. Therefore, the variability in the

Hf contents of the various NM fractions of one sample most likely results from different

proportions of high-Ca pyroxene.

The presence of at least three independent components with regard to Hf and W

among the coexisting phases of acapulcoites and lodranites reveals that the correlation

between ε182W and 180Hf/184W observed for the fractions of each of the acapulcoites and

lodranites cannot reflect simple binary mixing between W-rich metal and virtually W-free

silicates. This is also apparent from plots of ε182W vs. 1/W, in which binary mixtures should

form straight lines. This is not the case for any of those samples that exhibit linear

correlations in the ε182W vs. 180Hf/184W plots. Each of the fractions, therefore, evolved to

radiogenic ε182W according to their 180Hf/184W. Hence, the Hf-W data for the separates of

acapulcoites and lodranites define isochrons and can be interpreted to have chronological

significance.

Page 74: Touboul Thesis Completed

Figure 3.1.4: Hf versus W contents for the different fractions of the analyzed acapulcoites and lodranites. The Hf and W concentrations in the coexisting phases of these acapulcoites and lodranites are not collinear, indicating the presence of at three independent components for Hf and W among the coexisting phases. These components are high-Ca pyroxene, olivine + low-Ca pyroxene, and metal.

Relative Hf-W ages (or formation intervals), ΔtCAI, are calculated from the initial 182Hf/180Hf ratios obtained from the slopes of the isochrons relative to an initial 182Hf/180Hf =

(9.7±0.4)×10-5 for CAIs (Burkhardt et al., 2008) and refer to the time of Hf-W closure in a

sample elapsed since crystallization of type B CAIs. Acapulcoites Dhofar 125 and NWA 2775

have indistinguishable initial 182Hf/180Hf ratios of (6.7±0.4)×10-5 and (6.5±0.5)×10-5,

respectively (Fig. 3.1.2). All data from these two acapulcoites combined define a single

isochron with an initial 182Hf/180Hf of (6.5±0.3)×10-5, corresponding to an age of ΔtCAI =

5.1±0.9 Ma. The initial 182Hf/180Hf of lodranite NWA 2627 of (6.2±0.4)×10-5 is

indistinguishable from the value for the acapulcoites and corresponds to an age of ΔtCAI =

5.6±1.0 Ma (Fig. 3.1.3).

The comparison of relative Hf-W ages to absolute ages (e.g., Pb-Pb ages) requires

conversion of Hf-W formation intervals to an absolute timescale, which in turn requires

knowledge of the initial 182Hf/180Hf and the absolute age of Hf-W closure in a sample. Due to

differences in closure temperatures of different chronometers, the ideal samples to obtain such

information are angrites because (i) they cooled rapidly, such that differences in closure

74

Page 75: Touboul Thesis Completed

75

temperatures do not result in resolvable age differences, and (ii) they exhibit high U/Pb ratios,

such that precise Pb-Pb ages are available (Lugmair and Galer, 1992; Amelin, 2008; Connelly

et al., 2008). Precise initial 182Hf/180Hf ratios (Kleine et al., 2008a) and Pb-Pb ages are

available for several angrites and all of these can be used to calculate absolute Hf-W ages.

Using the Pb-Pb age of 4564.42±0.12 Ma (Amelin, 2008) and initial 182Hf/180Hf =

(7.2±0.2)×10-5 for the angrite D'Orbigny (Markowski et al., 2007; Kleine et al., 2008a) results

in an absolute Hf-W age of 4563.5±0.7 Ma for the acapulcoites Dhofar 125 and NWA 2775,

and 4563.0±0.9 Ma for lodranite NWA 2627. Identical results are obtained if absolute ages

are calculated relative to the angrites Sahara 99555, NWA 4590, or NWA 4801.

3.1.4.2 Closure temperature for the Hf-W system in acapulcoites and lodranites

To evaluate the significance of the Hf-W ages for constraining the thermal evolution

of the acapulcoite-lodranite parent body, it is necessary to know the closure temperature, Tc,

for W diffusion in the appropriate silicate-metal mixture. Here we use the approach employed

by Kleine et al. (2008b) to calculate Tc as a function of grain size and cooling rate for H

chondrites. This approach is based on the models of Van Orman et al. (2001; 2006) to

estimate the diffusion parameters of W in high-Ca pyroxene, the major host of Hf and hence

radiogenic 182W in acapulcoites and lodranites, and to numerically simulate the diffusion

behavior of W in a high-Ca pyroxene-metal system. Details regarding this model are given in

Kleine et al. (2008b) and Van Orman et al. (2006).

Calculation of Tc as a function of cooling rate requires knowledge of the peak

temperature, the grain size of high-Ca pyroxene, and the high-Ca pyroxene/metal ratio.

Acapulcoites show evidence for melting at the FeNi-FeS eutectic, indicating that peak

temperatures must have been higher than ~1000 °C. There is also evidence for limited melting

of silicates (Zipfel et al., 1995; Mittlefehldt et al., 1996), such that ~1150°C is a reasonable

assumption for the peak temperature of acapulcoites. Lodranites exhibit depletions in troilite

and plagioclase, suggesting mobility of FeNi-FeS and basaltic melts. This requires

temperatures of up to ~1300 °C (Mittlefehldt et al., 1996; McCoy et al., 1997a), which was

chosen here as the peak temperature for lodranites. The grain size of high-Ca pyroxene and

the high-Ca pyroxene/metal ratios in the acapulcoites and lodranites investigated here were

determined using the BSE images shown in Fig. 3.1.1 and results are summarized in the

caption of Fig. 3.1.1.

Page 76: Touboul Thesis Completed

Figure 3.1.5: Closure temperature of the Hf-W system as a function of grain size and cooling rate. The assumed initial temperature is 1150°C for grain sizes of 10-200 μm and 1300°C for larger grain sizes. Note that cooling rates for acapulcoites and lodranites at 900-1100 °C are on the order of ~100 °C/Ma (for details see text).

Fig. 3.1.5 shows closure temperatures calculated as a function of cooling rate. For

grain sizes of high-Ca pyroxenes from ~100 μm to ~200 μm, typical for the acapulcoites

investigated here (Fig. 3.1.5), the assumed initial temperature is 1150 °C. For grain sizes from

~350 μm to ~450 μm, the assumed initial temperature is 1300 °C, as appropriate for

lodranites. Fig. 3.1.5 reveals that Tc for the Hf-W system in acapulcoites ranges from ~900 °C

for slow cooling (dT/dt ~10°C/Ma) to ~1050 °C for fast cooling (dT/dt ~1000°C/Ma). Due to

their larger grain sizes the Hf-W closure temperature for lodranites are slightly higher and

range from ~975 °C for slow cooling (dT/dt ~10°C/Ma) to ~1100 °C for fast cooling (dT/dt

~1000°C/Ma). As we shall see below, the cooling rate of acapulcoites and lodranites at high

temperature is on the order of ~100 °C/Ma, and for this cooling rate Hf-W closure

temperatures for acapulcoites and lodranites are 975±50 °C and 1025±50 °C. Note that these

estimates are not very sensitive to variations in the cooling rate unless cooling is one order of

magnitude slower or faster.

76

Page 77: Touboul Thesis Completed

77

The comparison of the Hf-W ages to other radiometric ages provides a test for the

validity of the Tc estimates for the Hf-W system. Although age information for acapulcoites

and lodranites is sparse and the samples investigated here have not been studied with other

high precision chronometers, the Hf-W ages for the acapulcoites Dhofar 125 and NWA 2775

can be compared with a Pb-Pb age for Acapulco phosphates (Göpel et al., 1992). The Hf-W

age for acapulcoites of 4563.5±0.7 Ma is ~5 Ma older than the 4557±2 Ma Pb-Pb age for

phosphates from Acapulco. The closure temperature of Pb diffusion in apatite (grain sizes

from 100 to 300 µm) range from ~490 °C for slow cooling (dT/dt ~10°C/Ma) to ~655 °C for

fast cooling (dT/dt ~1000°C/Ma) (Cherniak et al., 1991). Hence, Tc for the Pb-Pb system in

Acapulco phosphates is always lower than Tc for high-Ca pyroxene in acapulcoites, consistent

with the observation that the Hf-W age for Dhofar 125 and NWA 2775 is older than the Pb-

Pb age for Acapulco phosphates.

3.1.4.3 Accretion and cooling history of the acapulcoite-lodranite parent body

The Hf-W ages and Tc estimates for acapulcoites and lodranites in conjunction with

thermochronological information from other chronometers and petrological constraints on

peak temperatures can be used to constrain the accretion age and cooling history of the

acapulcoite-lodranite parent body. Temperature profiles for spherical asteroids heated by

energy released from 26Al decay (Carslaw and Jaeger, 1959; Miyamoto et al., 1981) were

calculated and the parameters used are identical to those in Kleine et al. (2008b). We assume

instantaneous accretion and do neither include the insulating effects of a regolith (Akridge et

al., 1998) nor temporal and local variations in physical and thermal parameters (e.g., changes

in thermal conductivity due to a decrease in porosity) (Bennett and McSween, 1996). For

instance, if accretion had taken place over a timescale similar to the 26Al half-life, then a body

would have started retaining the heat produced by 26Al decay before reaching its terminal

mass and peak temperatures would have been reached earlier than estimated when assuming

instantaneous accretion (Merk et al., 2002; Ghosh et al., 2003). A thick insulating regolith

would have resulted in slower cooling rates and hence higher temperatures in the interior,

compared to the model used here (Akridge et al., 1998). Nevertheless, in spite of these

simplifications, the model used here is useful for calculation cooling curves for acapulcoites

and lodranites and for obtaining some information on the accretion time, size and internal

structure of the acapulcoite-lodranite parent body.

In Fig. 3.1.6, calculated temperature profiles for different depths in spherical asteroids

with radii of 30, 50, and 100 km are shown for accretion ages of 1.5, 1.7, and 2 Ma after CAI

Page 78: Touboul Thesis Completed

78

formation. In these models the acapulcoite-lodranite parent body must have had a radius of

more than ~35 km because for smaller radii cooling is faster than indicated by the differences

in Hf-W and Pb-Pb closure temperatures and ages for acapulcoites. In the thermal model used

here, the best fit to the Hf-W and Pb-Pb constraints is obtained for the uppermost 10 to 20 km

of a body with radius of 50 to 100 km. However, if the acapulcoite-lodranite parent body had

a thick insulating regolith, then cooling at a given depth would have been slower, such that

the thermochronological data might also be consistent with a parent body radius of less than

~35 km.

The thermal models also reveal that for accretion at ~2 Ma the peak temperatures of

the lodranites of ~1300 °C are only reached near the center of the parent body but these areas

would have cooled below ~1025 °C (i.e., Tc of the Hf-W system in lodranites) later than ~10

Ma after CAI formation, which is inconsistent with the ~5.6 Ma Hf-W age for lodranites. This

conclusion holds true even if the acapulcoite-lodranite parent body had a thick regolith

because in this case cooling at a given depth would have been even slower (Akridge et al.,

1998). Conversely, for accretion as early as ~1.5 Ma, peak temperatures for lodranites and

acapulcoites are ~1400 °C and ~1200 °C, respectively, which is somewhat too high. These

observations suggest that the acapulcoite-lodranite parent body accreted later than ~1.5 Ma

but before ~2 Ma.

The thermochronometric data are consistent with a parent body of ~100 km radius that

accreted at 1.7 Ma after CAI formation (Fig. 3.1.6). In this model, acapulcoites would have

formed ~7 km below the surface and lodranites originate from a layer that is up to ~1.5 km

deeper. Peak temperatures in the acapulcoite and lodranite regions were reached at ~3.5 Ma

after CAI formation and were ~1150 °C and ~1280°C, respectively, consistent with petrologic

constraints. If this model is correct, then large parts of the acapulcoite-lodranite parent body

(i.e., the innermost ~90 km) must be differentiated into a metal core and silicate mantle

because this area must have maintained temperatures in excess of ~1300 °C for several

millions of years. If any of these materials were delivered to Earth, it has yet not been

identified in the meteorite collections.

Page 79: Touboul Thesis Completed

Figure 3.1.6: Cooling curves for acapulcoites and lodranites. Solid lines indicate calculated temperature profiles for different depths in spherical bodies with radii of 35, 50, and 100 km. Numbers indicate distance in km from the center. Parameters are identical to those used in Kleine et al. (2008b) and are: thermal conductivity K = 1.0 W m-1 K-1; thermal diffusivity κ = 5.0 × 10-7 m2 s-1; density ρ = 3.2 × 103 kg m-3; heat generation A = 11.67 × (26Al/27Al) W m-3; emissivity h = 1.0 m-1. The assumed ambient temperature is T0 = 300 K and the initial 26Al/27Al ratios are 1.38, 1.14, 0.85 and 0.52 × 10-5, corresponding to accretion at 1.5, 1.7, 2 and 2.5 Ma after CAIs. Hf-W ages and closure temperatures are from this study, all other ages are from the literature and are summarized in Table 3.1.2. Ar-Ar ages are shifted by ~30 Ma due to the proposed revision in the 40K decay constant (Renne, 2000; Trieloff et al., 2001). The best fit to the thermochronometric data is obtained for a 100 km body that accreted 1.7 Ma after CAI formation.

79

Page 80: Touboul Thesis Completed

Cooling rates for acapulcoites and lodranites can be obtained from the slope of the

cooling curves and for acapulcoites cooling rates decrease from ~120 °C/Ma at ~1000 °C

(i.e., just below their peak temperature) to ~50 °C/Ma at ~600 °C. Over the same

temperature range the cooling rate of lodranites decreased from ~110 °C/Ma to ~40

°C/Ma (Fig. 3.1.7). This thermal history is different from those of H5 and H6 chondrites,

which cooled at a much slower rate (Fig. 3.1. 7). This could reflect a deeper burial depth

of H5 and H6 chondrites compared to acapulcoites and lodranites. The thermal history at

lower temperatures is less well constrained. Fig. 3.1.6 reveals that the thermal model

presented here is consistent with Ar-Ar ages for acapulcoites and lodranites (McCoy et

al., 1996; Mittlefehldt et al., 1996; McCoy et al., 1997a; Pellas et al., 1997; Renne, 2000)

as well as U-Th-He ages for Acapulco phosphates (Min et al., 2003). However, the

uncertainties on the Ar-Ar and U-Th-He ages are large and would be consistent with

relatively slow cooling of <20 °C/Ma for temperatures below ~400 °C (as would be the

case for the thermal model shown in Fig. 3.1.7) as well as with almost immediate cooling

from ~550 °C to ambient temperature. Evidence for very slow cooling at low

temperatures comes from the determination of 244Pu fission track densities in

orthopyroxene, withlockite and apatite, which indicate a cooling rate of 1.7 °C/Ma in the

temperature range of 280-90 °C (Pellas et al., 1997). This is somewhat lower than the

cooling rates obtained from the model presented here, which decrease from ~7 °C/Ma at

~280 °C to ~1 °C/Ma at ~100 °C. Based on the (U-Th)/He age of 4538±32 Ma, Min et al.

(2003) argued that cooling below ~120 °C must have been rapid, inconsistent with the

slow cooling rates obtained from the 244Pu fission track data. These authors suggest that

track annealing by later thermal disturbance resulted in spuriously young 244Pu fission

track ages. Metallographic cooling rates for acapulcoites and lodranites (McCoy et al.,

1996; McCoy et al., 1997a; McCoy et al., 1997b)and estimates based on Ca zonation in

Acapulco olivines (Zipfel et al., 1995) indeed suggest relatively fast cooling of >1000

°C/Ma for temperatures below ~600 °C. The shift from initially moderate cooling at ~50-

100 °C/Ma (from peak temperatures down to ~600 °C) to rapid cooling at >1000 °C/Ma

(at temperatures below ~600 °C) might be related to excavation of acapulcoites and

lodranites, possibly caused by impacts on their parent body (McCoy et al., 1996; McCoy

80

Page 81: Touboul Thesis Completed

et al., 1997a; McCoy et al., 1997b). Alternatively, it is possible also that one method for

determining cooling rates is not entirely reliable.

Figure 3.1.7: Cooling rates for acapulcoites and lodranites as calculated from the slope of the cooling curves shown in Fig. 3.1.6. Cooling rates for H chondrites are shown for comparison and are from Kleine et al. (2008b).

In the thermal models discussed above it is assumed that 26Al decay is the sole

heat source for melting and metamorphism of acapulcoites and lodranites. However, it

has recently been suggested that impact heating could have been the dominant heat

source (Rubin, 2007). While impact heating of asteroids is difficult to exclude as a heat

source, the data presented here suggest that energy release by 26Al decay must have been

important in the thermal evolution of the acapulcoite-lodranite parent body. This is

because the Hf-W age of 5.1±0.9 Ma for acapulcoites refers to the time of cooling below

~975 °C, such that for reasonable cooling rates, which are on the order of 100 °C/Ma as

constrained by the difference between the Hf-W and Pb-Pb ages, acapulcoites must have

reached their thermal peak of ~1150 °C at ~3-4 Ma after CAI formation. Obviously, their

parent body must have accreted before this time and thermal modeling indicates that in

bodies accreted that early 26Al decay was an important heat source. In other words, 26Al

decay could only not have been an important heat source if accretion, heating, partial

81

Page 82: Touboul Thesis Completed

melting, and cooling below ~975 °C occurred within less than ~1 Ma. This seems highly

unlikely and is also difficult to reconcile with the ~5 Ma difference between the Hf-W

and Pb-Pb ages of acapulcoites that require cooling on the order of ~100 °C/Ma.

3.1.4.4 Bulk Hf-W systematics of the acapulcoite-lodranite parent body – nebular vs.

parent body processes

The 180Hf/184W ratios and ε182W values of the acapulcoite and lodranite whole-

rocks reported in this study show wide variations and are most likely distinct from the

composition of carbonaceous chondrites (Table 3.1.1). This may reflect real differences

in the Hf-W systematics caused by Hf/W fractionation related to melting and segregation

of metals in the acapulcoite-lodranite parent body. However, the variations in the

measured 180Hf/184W ratios among the different whole-rocks may in part also be due to

sample heterogeneities. This seems likely given that in acapulcoites and lodranites almost

all the W is located in the metal, such that small variations in the metal abundance in the

analyzed aliquots will lead to variations in the 180Hf/184W ratio.

The 180Hf/184W ratios of bulk acapulcoites and lodranites are more reliably

determined by their time-integrated W isotope evolution. In a plot of initial ε182W vs.

initial 182Hf/180Hf, the acapulcoites and lodranites analyzed for this study plot below the

evolution line for carbonaceous chondrites, indicating that they evolved with a 180Hf/184W

ratio lower than that of carbonaceous chondrites (Fig. 8). The latest possible time at

which this low 180Hf/184W ratio could have been established is obtained by assuming 180Hf/184W~0 for acapulcoites and lodranites. In this case, acapulcoites and lodranites

departed from the carbonaceous chondrite evolution line at ~2.5 Ma. However, the

acapulcoites-lodranites must have 180Hf/184W>0, indicating that the Hf-W fractionation

event must have taken place well before this time. Given that metal melting in the

acapulcoites and lodranites most likely occurred at ~3 Ma after CAI formation (see above

and Fig. 6), their low 180Hf/184W ratios cannot reflect Hf-W fractionation during metal

segregation in the parent body but rather reflect processes prior to parent body accretion.

Remarkably, acapulcoites and lodranites plot on the same W isotope evolution line as H

chondrites (Kleine et al., 2008b). This is consistent with other compositional similarities

between H chondrites and acapulcoites-lodranites (Palme et al., 1981) and suggests that

82

Page 83: Touboul Thesis Completed

the precursor materials of their parent bodies had similar chemical compositions that

were established by fractionation and/or mixing processes in the solar nebula.

The similarity in time-integrated 180Hf/184W ratios of the acapulcoites and

lodranites implies that the FeS-FeNi melts in lodranite NWA 2627 did not migrate, such

that the chemical composition of this sample remained unaffected by the partial melting

processes that are typical for lodranites (McCoy et al., 1997b). Whether this also applies

to other lodranites remains to be determined by future Hf-W studies on a more

comprehensive set of lodranites.

Figure 3.1.8: W isotope evolution diagram for acapulcoites and lodranites. The evolution line for carbonaceous chondrites is defined by their present-day ε182W of -1.9±0.1 (Kleine et al., 2004) as well as the initial 182Hf/180Hf = (9.7±0.4)×10-5 and initial ε182W=-3.28±0.12 of Allende CAIs (Burkhardt et al., 2008). These values indicate that carbonaceous chondrites evolved with 180Hf/184W=1.23±0.15, consistent with 180Hf/184W=1.25±0.08 measured for carbonaceous chondrites (Kleine et al., 2004; Kleine et al., 2008a). The analyzed acapulcoites and lodranites depart from the evolution line for carbonaceous chondrites but are consistent with the W isotope evolution of H chondrites. The low 180Hf/184W of acapulcoites-lodranites has been established before ~2 Ma after CAI formation and thus must reflect Hf/W fractionation in the solar nebula.

83

Page 84: Touboul Thesis Completed

3.1.5 Conclusions

The Hf-W ages for acapulcoites and lodranites presented here are 5.1±0.9 Ma and

5.6±1.0 Ma after CAI formation and are the most ancient ages yet reported for this group

of meteorites. This confirms the high closure temperature of the Hf-W system in these

meteorites, which for both acapulcoites and lodranites is less than ~200 °C below their

peak temperatures. The Hf-W ages, therefore, provide constraints on the high-

temperature thermal history of the acapulcoite-lodranite parent body, information that is

not obtainable from other chronometers due to their much lower closure temperatures.

The results presented here indicate that the acapulcoite-lodranite parent body accreted

later than ~1.5 Ma but before ~2 Ma after CAI formation and that 26Al was an important

heat source in its thermal history. Cooling rates for both acapulcoites and lodranites were

~100-120 °C/Ma just below their thermal peak and decreased to ~40-50°C/Ma at 600 °C.

Thus, cooling of acapulcoites and lodranites was roughly one order of magnitude faster

than it was for the most metamorphosed H chondrites (i.e., H6 chondrites most of which

cooled at ~10 °C/Ma in the temperature interval from ~900 to ~200 °C). This difference

results from a deeper burial depth of H6 chondrites compared to acapulcoites-lodranites.

For instance, acapulcoites and lodranites could have been located ~7 and ~9 km below

the surface of a parent body with a radius of ~100 km that accreted at ~1.7 Ma after CAI

formation, whereas H6 chondrites were located ~40-60 km below the surface of an

asteroid with a radius of ~100 km that accreted ~2.5 Ma after CAI formation (Kleine et

al., 2008b).

Constraining the thermal evolution of asteroids does not only provide information

on parent body size and burial depths of a suite of samples but also is key for constraining

the parameters that controlled the evolution of meteorite parent bodies. The Hf-W ages in

conjunction with petrologic constraints on the peak temperatures of acapulcoites and

lodranites indicate that the acapulcoite-lodranite parent body accreted between 1.5 and 2

Ma after CAI formation and, hence, later than the parent bodies of magmatic iron

meteorites (Kleine et al., 2005) but earlier than most (or all) chondrite parent bodies (Kita

et al., 2000; Kunihiro et al., 2004; Rudraswami and Goswami, 2007; Kleine et al.,

2008b). Thus, the accretion age of asteroids is inversely correlated with the peak

temperatures reached in their interiors. Differentiated asteroids (i.e., the parent bodies of

84

Page 85: Touboul Thesis Completed

magmatic iron meteorites) accreted within the first ~1 Ma after CAI formation, partially

differentiated bodies - such as the acapulcoite-lodranite parent body - formed between 1.5

and 2 Ma, and undifferentiated asteroids (i.e., the chondrite parent bodies) accreted later

than ~2 Ma after CAIs. This provides further evidence for earlier conclusions that the

different thermal histories of meteorite parent bodies primarily reflect variations in their

initial 26Al abundance, which, owing to the short 26Al half-life, is determined by the

accretion time (Bizzarro et al., 2005; Kleine et al., 2005; Schérsten et al., 2006).

Table 3.1.1 : Hf-W data for acapulcoites and lodranites

Sample W (ppb)

Hf (ppb)

180Hf/184W ± 2σ ε82W ± 2σ meas.

ε82W ± 2σ corr. ε83W ± 2σ

Dhofar 125 (acapulcoite) WR-1 109.9 190.3 2.042 ± 0.020 -1.79 ± 0.28 -0.01 ± 0.26 WR-2 118.6 189.0 1.881 ± 0.019 -1.54 ± 0.18 0.01 ± 0.15 M 901.1 0.4 0.0005 ± 0.0001 -2.97 ± 0.21 0.11 ± 0.17 -3.33 ± 0.24 -0.20 ± 0.18 mean (2 SE) -3.15 ± 0.36 -0.04 ± 0.31 fines-1 373.9 133.1 0.415 ± 0.004 -2.90 ± 0.21 -0.08 ± 0.11 fines-2 375.1 131.8 0.420 ± 0.004 -2.77 ± 0.17 0.26 ± 0.13 WM-1 91.3 171.8 2.22 ± 0.02 -1.25 ± 0.19 -0.17 ± 0.21 WM-2 105.9 173.1 1.92 ± 0.02 -1.28 ± 0.18 -0.18 ± 0.13 NM-1 14.3 257.9 21.3 ± 0.4 13.2 ± 0.8 13.5 ± 1.2 -0.30 ± 0.56 NM-2 25.9 237.4 10.8 ± 0.2 4.01 ± 0.78 4.11 ± 1.30 -0.11 ± 0.48 NM-3 30.0 270.9 10.6 ± 0.2 4.78 ± 0.52 4.87 ± 1.20 -0.14 ± 0.36 NM-4 29.4 174.2 6.99 ± 0.07 1.92 ± 0.33 2.01 ± 0.50 -0.12 ± 0.27

NWA 2775 (acapulcoite) WR 174.4 179.8 1.21 ± 0.01 -2.30 ± 0.21 -0.47 ± 0.17 -2.22 ± 0.32 -0.44 ± 0.21 mean (2 SE) -2.26 ± 0.09 -0.46 ± 0.04 M 1380.6 6.2 0.0053 ± 0.0001 -3.15 ± 0.37 -0.26 ± 0.18 WM-1 329.0 121.7 0.436 ± 0.004 -2.29 ± 0.23 -0.25 ± 0.13 -2.72 ± 0.14 -0.04 ± 0.12 -2.92 ± 0.18 -0.09 ± 0.11 mean (2 SE) -2.64 ± 0.37 -0.13 ± 0.13 WM-2 416.6 122.8 0.348 ± 0.003 -2.64 ± 0.19 -0.15 ± 0.13 -2.74 ± 0.19 -0.04 ± 0.13 -2.75 ± 0.14 -0.17 ± 0.11 mean (2 SE) -2.71 ± 0.07 -0.12 ± 0.08

85

Page 86: Touboul Thesis Completed

fines 213.4 159.7 0.883 ± 0.009 -2.28 ± 0.18 -0.20 ± 0.15 -2.14 ± 0.18 -0.28 ± 0.14 -1.94 ± 0.26 -0.27 ± 0.19 mean (2 SE) -2.12 ± 0.20 -0.25 ± 0.05 NM-1 3.8 189.9 58.2 ± 5.7 35.4 ± 1.5 39.7 ± 2.9 2.21 ± 0.69 NM-2 5.1 165.4 38.1 ± 2.9 24.2 ± 1.6 26.8 ± 2.3 1.76 ± 1.10 NM-3 27.0 180.2 7.63 ± 0.15 2.41 ± 0.37 2.47 ± 1.20 -0.25 ± 0.55

NWA 2656 (acapulcoite) WR 371.1 106.2 0.337 ± 0.003 -2.41 ± 0.16 -0.06 ± 0.11 -2.29 ± 0.17 -0.09 ± 0.13 mean (2 SE) -2.35 ± 0.11 -0.07 ± 0.03 WM 622.7 26.3 0.0499 ± 0.0005 -2.23 ± 0.18 -0.04 ± 0.13 fines 461.4 285.1 0.729 ± 0.007 -1.06 ± 0.17 -0.08 ± 0.13 -0.93 ± 0.16 -0.10 ± 0.12 mean (2 SE) -1.00 ± 0.13 -0.09 ± 0.01 NM-1 123.9 657.3 6.26 ± 0.06 -0.43 ± 0.44 -0.27 ± 0.26 NM-2 111.3 294.5 3.12 ±0.02 -0.47 ± 0.26 -0.31 ± 0.16 NM-3 58.6 194.3 3.92 ± 0.04 -0.01 ± 1.10 0.63 ± 1.45

NWA 2627 (lodranite) WR 168.6 95.7 0.670 ± 0.007 -2.42 ± 0.16 -0.09 ± 0.23 M-1 1163.4 18.3 0.0185 ± 0.0002 -2.85 ± 0.27 0.22 ± 0.27 M-2 841.7 36.4 0.051 ± 0.001 -2.89 ± 0.30 -0.08 ± 0.27 -3.01 ± 0.22 0.14 ± 0.27 mean (2 SE) -2.95 ± 0.28 0.03 ± 0.28 WM 58.5 66.8 1.346 ± 0.013 -2.15 ± 0.37 -2.24 ± 0.60 1.74 ± 0.23 fines 38.4 125.2 3.85 ± 0.04 -0.42 ± 0.41 -0.38 ± 0.60 -0.30 ± 0.36 NM-1 8.8 126.5 17.0 ± 0.6 7.77 ± 0.39 9.04 ± 0.94 1.74 ± 0.31 NM-2 12.7 92.7 8.59 ± 0.22 3.24 ± 0.35 3.58 ± 0.62 0.31 ± 0.23

NWA 4663 (lodranite)

M 1162.8 7.6 0.0077 ± 0.0001 -3.12± 0.28 0.04 ± 0.15 -2.65± 0.14 0.05 ± 0.09 -2.74± 0.13 0.02 ± 0.12 mean (2 SE) -2.72± 0.16 0.09 ± 0.11 WM-2 224.7 308.9 1.62 ± 0.02 -0.56± 0.21 0.21 ± 0.19 WM-3 664.0 135.7 0.241 ± 0.002 -2.72± 0.14 -0.04 ± 0.12 NM-1 28.5 651.1 26.9 ± 0.4 3.24 ± 0.30 3.39 ± 0.31 0.91 ± 0.53 NM-2 23.7 344.9 17.2 ± 0.4 4.09 ± 0.33 4.38 ± 0.36 0.72 ± 0.15

Gibson (lodranite)

WR 153.5 154.8 1.19 ± 0.01 -1.78 ± 0.24 -0.03 ± 0.15

86

Page 87: Touboul Thesis Completed

M = metal fraction, WM = weackly magnetic fraction, NM = non magnetic fraction, WR = whole rock. The quoted ± 2σ uncertainties for measured ε82W and ε183W are analytical errors. Uncertainties on blank corrected ε82W are calculated by propagating the external reproductibility of isotope measurements and a ∼50% uncertainty on the blank correction. Regression calculation though data assume an external reproductibility of ± 0.3-0.4ε except for the W poor fractions (Dhofar 125 NM-1, 2 and 3, NWA 2656 NM-1 and 2, and NWA 2627 NM-1). Table 3.1.2. Radiochronometric ages and closure temperatures for acapulcoites and lodranites

Samples System Minerals Age Closure

temperature reference Acapulcoites

NWA 2775 – Dhofar 125

182Hf-182W metal-silicate 4563.5 ± 0.7 975 ± 50 This study

Acapulco 206Pb-207Pb phosphates 4557 ± 2 550 ± 100 (1) Acapulco 36Ar-40Ar whole rock,

plagioclases 4514 ± 22 277 ± 50 (2)

Acapulco 244Pu tracks whitlokites 4440 ± 26 117 ± 50 (2) Acapulco U-Th/He apatites 4538 ± 32 120 ± 50 (3)

Lodranites

NWA 2627 182Hf-182W metal-silicate 4563.0 ± 0.9 1025 ± 50 This study Gibson 36Ar-40Ar whole rock,

plagioclases 4520 ± 40 277 ± 50 (4)

(1) Göpel et al. (1992). (2) Pellas et al. (1997). (3) Min et al. (2003). (4) McCoy et al. (1997)

References

Akridge G., Benoit P. H., and Sears D. W. G. (1998) Regolith and megaregolith

formation of H-chondrites: thermal constraints on the parent body. Icarus 132, 185-

195.

Amelin Y. (2008) U-Pb ages of angrites. Geochim. Cosmochim. Acta 72, 221-232.

Bennett M. E. and McSween H. Y. (1996) Revised model calculations for the thermal

histories of ordinary chondrite parent bodies. Meteorit. Planet. Sci. 31, 783-792.

Bizzarro M., Baker J. A., Haack H., and Lundgaard K. L. (2005) Rapid timescales for

accretion and melting of differentiated planetesimals inferred from 26Al-26Mg

chronometry. Astrophys. J. 632, L41-L44.

Burkhardt C., Kleine T., Palme H., Bourdon B., Zipfel J., Friedrich J., and Ebel D. (2008)

Hafnium-tungsten systematics of calcium-aluminum-rich inclusions from

carbonaceous chondrites. Geochim. Cosmochim. Acta, submitted.

87

Page 88: Touboul Thesis Completed

Carslaw H. S. and Jaeger J. C. (1959) Conduction of heat in solids. Oxford University

Press.

Clayton R. N. and Mayeda T. K. (1996) Oxygen isotope studies of achondrites. Geochim.

Cosmochim. Acta 60, 1999-2017.

Connolly, H. C., Smith, C., Benedix, G., Folco, L.,Righter, K., Zipfel, J., Yamaguchi, A.,

Chennaoui Aoudjehane, H., 2007. The Meteoritical Bulletin, No. 93, 2008 March.

Meteorit. Planet. Sci. 43, 3, 571–632.

Connelly J., Bizzarro M., Thrane K., and Baker J. A. (2008) The Pb-Pb age of angrite

Sah99555 revisited. Geochim. Cosmochim. Acta, doi:10.1016/j.gca.2008.06.007.

Dodson M. H. (1973) Closure temperature in cooling geochronological and petrological

systems. Contrib. Mineral. Petrol. 40, 259-274.

Eugster O. and Lorenzetti S. (2005) Cosmic-ray exposure ages for four acapulcoites and

two differentiated achondrites and evidence for a two-layer structure of the

acapulcoite-lodranite parent asteroid. Geochim. Cosmochim. Acta 69, 2675-2685.

Ganguly J. and Tirone M. (2001) Relationship between cooling rate and cooling age of a

mineral: Theory and applications to meteorites. Meteorit. Planet. Sci. 36, 167-175.

Ghosh A., Weidenschilling S. J., and McSween H. Y. (2003) Importance of the accretion

process in asteroid thermal evolution: 6 Hebe as an example. Meteorit. Planet. Sci.

38, 711-724.

Göpel C., Manhès G., and Allègre C. J. (1992) U-Pb Study of the Acapulco Meteorite.

Meteoritics 27, 226-226.

Göpel C., Manhès G., and Allègre C. J. (1994) U-Pb systematics of phosphates from

equilibrated ordinary chondrites. Earth Planet. Sci. Lett. 121, 153-171.

Irving, A. J., Bunch, T. E., Wittke, J. H., Kuehner, S. M. and Rumble, D., 2007.

Assessment of Multi-Component Mixing, Oxidation, Metamorphism and Partial

Melting on the Acapulcoite-Lodranite Parent Bod. Lunar Planet. Sci. Conf.

XXXVIII. LPI Contribution No. 1338, 2254.

Kita N. T., Nagahara H., Togashi S., and Morishita Y. (2000) A short duration of

chondrule formation in the solar nebula: Evidence from 26Al in Semarkona

ferromagnesian chondrules. Geochim. Cosmochim. Acta 64, 3913-3922.

88

Page 89: Touboul Thesis Completed

Kleine T., Münker C., Mezger K., and Palme H. (2002) Rapid accretion and early core

formation on asteroids and the terrestrial planets from Hf-W chronometry. Nature

418, 952-955.

Kleine T., Mezger K., Münker C., Palme H., and Bischoff A. (2004) 182Hf-182W isotope

systematics of chondrites, eucrites, and Martian meteorites: Chronology of core

formation and mantle differentiation in Vesta and Mars. Geochim. Cosmochim.

Acta 68, 2935-2946.

Kleine T., Mezger K., Palme H., Scherer E., and Münker C. (2005) Early core formation

in asteroids and late accretion of chondrite parent bodies: Evidence from 182Hf-182W

in CAIs, metal-rich chondrites and iron meteorites. Geochim. Cosmochim. Acta 69,

5805-5818.

Kleine T., Touboul M., Bourdon B., Nimmo F., Mezger K., Palme H., Yin Q. Z.,

Jacobsen S. B., and Halliday A. N. (2008a) Hf-W chronometry and the accretion

and early evolution of asteroids and terrestrial planets. Geochim. Cosmochim. Acta,

submitted.

Kleine T., Touboul M., Van Orman J. A., Bourdon B., Maden C., Mezger K., and

Halliday A. (2008b) Hf-W thermochronometry: closure temperature and constraints

on the accretion and cooling history of the H chondrite parent body. Earth Planet.

Sci. Lett. 270, 106-118.

Kunihiro T., Rubin A. E., McKeegan K. D., and Wasson J. T. (2004) Initial 26Al/27Al in

carbonaceous-chondrite chondrules: Too little 26Al to melt asteroids. Geochim.

Cosmochim. Acta 68, 2947-2957.

LaTourrette T. and Wasserburg G. J. (1998) Mg diffusion in anorthite: implications for

the formation of early solar system planetesimals. Earth Planet. Sci. Lett. 158, 91-

108.

Ludwig K. (1991) ISOPLOT: a plotting and regression program for radiogenic isotope

data; version 2.53. U.S.G.S. Open File Report 91-0445.

Lugmair G. W. and Galer S. J. G. (1992) Age and isotopic relationships among the

angrites Lewis Cliff 86010 and Angra dos Reis. Geochim. Cosmochim. Acta 56,

1673-1694.

89

Page 90: Touboul Thesis Completed

Markowski A., Quitté G., Halliday A. N., and Kleine T. (2006) Tungsten isotopic

compositions of iron meteorites: chronological constraints vs. cosmogenic effects.

Earth Planet. Sci. Lett. 242, 1-15.

Markowski A., Quitté G., Kleine T., Halliday A., Bizzarro M., and Irving A. J. (2007)

Hf-W chronometry of angrites and the earliest evolution of planetary bodies. Earth

Planet. Sci. Lett. 262, 214-229.

McCoy T. J., Keil K., Clayton R. N., Mayeda T. K., Bogard D., Garrison D., Huss G. R.,

Hutcheon I. D., and Wieler R. (1996) A petrologic, chemical, and isotopic study of

Monument Draw and comparison with other acapulcoites: Evidence for formation

by incipient partial melting. Geochim. Cosmochim. Acta 60, 2681-2708.

McCoy T. J., Keil K., Clayton R. N., Mayeda T. K., Bogard D. D., Garrison D. H., and

Wieler R. (1997a) A petrologic and isotopic study of lodranites: Evidence for early

formation as partial melt residues from heterogeneous precursors. Geochim.

Cosmochim. Acta 61, 623-637.

McCoy T. J., Keil K., Muenow D. W., and Wilson L. (1997b) Partial melting and melt

migration in the acapulcoite-lodranite parent body. Geochim. Cosmochim. Acta 61,

639-650.

Merk R., Breuer D., and Spohn T. (2002) Numerical modeling of 26Al-induced

radioactive melting of asteroids considering accretion. Icarus 159, 183-191.

Min K. W., Farley K. A., Renne P. R., and Marti K. (2003) Single grain (U-Th)/He ages

from phosphates in Acapulco meteorite and implications for thermal history. Earth

Planet. Sci. Lett. 209, 323-336.

Mittlefehldt D. W., Lindstrom M. M., Bogard D. D., Garrison D. H., and Field S. W.

(1996) Acapulco- and Lodran-like achondrites: Petrology, geochemistry,

chronology, and origin. Geochim. Cosmochim. Acta 60, 867-882.

Miyamoto M., Fujii N., and Takeda H. (1981) Ordinary chondrite parent body: An

internal heating model. Proc. Lunar Planet. Sci. Conf. 12B, 1145-1152.

Palme H., Schultz L., Spettel B., Weber H. W., and Wänke H. (1981) The Acapulco

meteorite: chemistry, mineralogy and irradiation effects. Geochim. Cosmochim.

Acta 45, 727-752.

90

Page 91: Touboul Thesis Completed

Pellas P., Fiéni C., Trieloff M., and Jessberger E. K. (1997) The cooling history of the

Acapulco meteorite as recorded by the 244Pu and 40Ar-39Ar chronometers. Geochim.

Cosmochim. Acta 61, 3477-3501.

Renne P. R. (2000) 40Ar/39Ar age of plagioclase from Acapulco meteorite and the

problem of systematic errors in cosmochronology. Earth Planet. Sci. Lett. 175, 13-

26.

Righter K. and Shearer C. K. (2003) Magmatic fractionation of Hf and W: Constraints on

the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497-2507.

Rubin A. E. (2007) Petrogenesis of acapulcoites and lodranites: A shock-melting model.

Geochim. Cosmochim. Acta 71, 2383-2401.

Rudraswami N. G. and Goswami J. N. (2007) Al-26 in chondrules from unequilibrated L

chondrites: Onset and duration of chondrule formation in the early solar system.

Earth Planet. Sci. Lett. 257, 231-244.

Schérsten A., Elliott T., Hawkesworth C., Russell S. S., and Masarik J. (2006) Hf-W

evidence for rapid differentiation of iron meteorite parent bodies. Earth Planet. Sci.

Lett. 241, 530-542.

Trieloff M., Jessberger E. K., and Fieni C. (2001) Comment on "Ar-40/Ar-39 age of

plagioclase from Acapulco meteorite and the problem of systematic errors in

cosmochronology" by Paul R. Renne. Earth Planet. Sci. Lett. 190, 267-269.

Trieloff M., Jessberger E. K., Herrwerth I., Hopp J., Fiéni C., Ghelis M., Bourot-Denise

M., and Pellas P. (2003) Structure and thermal history of the H-chondrite parent

asteroid revealed by thermochronometry. Nature 422, 502-506.

Van Orman J. A., Grove T. L., and Shimizu N. (2001) Rare earth element diffusion in

diopside: Influence of temperature, pressure and ionic radius, and an elastic model

for diffusion in silicates. Contrib. Mineral. Petrol. 141, 687-703.

Van Orman J. A., Saal A. E., Bourdon B., and Hauri E. H. (2006) Diffusive fractionation

of U-series radionuclides during mantle melting and shallow level melt-cumulate

interaction. Geochim. Cosmochim. Acta 70, 4797-4812.

Zipfel J., Palme H., Kennedy A. K., and Hutcheon I. D. (1995) Chemical composition

and origin of the Acapulco meteorite. Geochim. Cosmochim. Acta 59, 3607-3627.

91

Page 92: Touboul Thesis Completed

92

Page 93: Touboul Thesis Completed

Section 3.2

Hf–W thermochronometry: Closure temperature and constraints on the accretion and cooling

history of the H chondrite parent body

Thorsten Kleine 1, Mathieu Touboul 1, James A. Van Orman 2, Bernard Bourdon 1, Colin Maden 1, Klaus Mezger 3, Alex N. Halliday 4

1 Institute for Isotope Geochemistry and Mineral Resources, Department of Earth Sciences, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerland 2 Department of Geological Sciences, Case Western Reserve University, 10900 Euclid Ave, Cleveland, Ohio 44106-7216, United States 3 Institut für Mineralogie, Universität Münster, Corrensstrasse 24, 48149 Münster, Germany 4 Department of Earth Sciences, University of Oxford, Parks Road, OX1 3PR, United Kingdom

Publié dans Earth and Planetary Science Letters, 2008, 270, p.106-118

93

Page 94: Touboul Thesis Completed

94

Page 95: Touboul Thesis Completed

Hf–W thermochronometry: Closure temperature and constraints on the accretionand cooling history of the H chondrite parent body

Thorsten Kleine a,⁎, Mathieu Touboul a, James A. Van Orman b, Bernard Bourdon a, Colin Maden a,Klaus Mezger c, Alex N. Halliday d

a Institute for Isotope Geochemistry and Mineral Resources, Department of Earth Sciences, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerlandb Department of Geological Sciences, Case Western Reserve University, 10900 Euclid Ave, Cleveland, Ohio 44106-7216, United Statesc Institut für Mineralogie, Universität Münster, Corrensstrasse 24, 48149 Münster, Germanyd Department of Earth Sciences, University of Oxford, Parks Road, OX1 3PR, United Kingdom

A B S T R A C TA R T I C L E I N F O

Article history:Received 19 December 2007Received in revised form 3 March 2008Accepted 4 March 2008Available online 15 March 2008

Editor: R.W. Carlson

Keywords:W isotopesclosure temperatureH chondritesthermochronometrychondrules

Editor: R.W. Carlson

We obtained Hf–W metal-silicate isochrons for several H chondrites of petrologic types 4, 5, and 6 toconstrain the accretion and high-temperature thermal history of the H chondrite parent body. The silicatefractions have 180Hf/184W ratios up to ∼51 and 182W/184W ratios up to ∼33 ε units higher than the whole-rock. These high 180Hf/184W and radiogenic W isotope ratios result in highly precise Hf–W ages. The Hf–Wages of the H chondrites become younger with increasing metamorphic grade and range from ΔtCAI=1.7±0.7 Ma for the H4 chondrite Ste. Marguerite to ΔtCAI=9.6±1.0 Ma for the H6 chondrites Kernouvé andEstacado. Closure temperatures for the Hf–W system in H chondrites were estimated from numericalsimulations of W diffusion in high-Ca pyroxene, the major host of radiogenic 182W in H chondrites, and rangefrom 800±50 °C for H4 chondrites to 875±75 °C for H6 chondrites. Owing to these high closure tem-peratures, the Hf–W system closed early and dates processes associated with the earliest evolution of the Hchondrite parent body. Consequently, the high-temperature interval of ∼8 Ma as defined by the Hf–W agesis much shorter than intervals obtained from Rb–Sr and Pb–Pb dating. For H4 chondrites, heating on theparent body probably was insufficient to cause W diffusion in high-Ca pyroxene, such that the Hf–W age ofΔtCAI=1.7±0.7 Ma for Ste. Marguerite was not reset and most likely dates chondrule formation. This isconsistent with Al–Mg ages of ∼2 Ma for L and LL chondrules and indicates that chondrules from all ordinarychondrites formed contemporaneously. The Hf–W ages for H5 and H6 chondrites of ΔtCAI=5.9±0.9 Ma andΔtCAI=9.6±1.0 Ma correspond closely to the time of the thermal peak within the H chondrite parent body.Combined with previously published chronological data the Hf–W ages reveal an inverse correlation ofcooling rate and metamorphic grade: shortly after their thermal peak H6 chondrites cooled at ∼10 °C/Ma, H5chondrites at ∼30 °C/Ma and H4 chondrites at ∼55 °C/Ma. These Hf–W age constraints are most consistentwith an onion-shell structure of the H chondrite parent body that was heated internally by energy releasedfrom 26Al decay. Parent body accretion started after chondrule formation at 1.7±0.7 Ma and probably endedbefore 5.9±0.9 Ma, when parts of the H chondrite parent body already had cooled from their thermal peak.The well-preserved cooling curves for the H chondrites studied here indicate that these samples derive froma part of the H chondrite parent body that remained largely unaffected by impact disruption and reassemblybut such processes might have been important in other areas. The H chondrite parent body has a 180Hf/184Wratio of 0.63±0.20, distinctly lower than the 180Hf/184W=1.21±0.06 of carbonaceous chondrite parent bodies.This difference reflects Hf–W fractionation within the first ∼2 Ma of the solar system, presumably related toprocesses in the solar nebula.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Hafnium–tungsten chronometry has been applied widely todetermine the timescales of differentiation of asteroids and terrestrial

planets (Harper and Jacobsen, 1996; Schoenberg et al., 2002; Yin et al.,2002; Halliday, 2004; Kleine et al., 2004b; Jacobsen, 2005; Nimmo andAgnor, 2006; Nimmo and Kleine, 2007) but its potential for datingchondrites and constraining the thermal evolution of their parentbodies has yet to be explored. To utilize Hf–W chronometry ofmeteorites meaningfully it is necessary to know the closure tem-perature (Tc) for diffusive exchange of parent and daughter elementsamong the different minerals in a rock (Dodson, 1973; Ganguly and

Earth and Planetary Science Letters 270 (2008) 106–118

⁎ Corresponding author.E-mail address: [email protected] (T. Kleine).

0012-821X/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.epsl.2008.03.013

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

95

Page 96: Touboul Thesis Completed

Tirone, 2001). Knowledge of Tc is essential for evaluating whether anage dates the time of mineral growth or some time along the coolingpath. Such information is critical for the interpretation of Hf–Wages incomparison to results from other chronometers and within theframework of models for the thermal evolution of asteroids.

Closure temperatures can be calculated from diffusion rates of theelement of interest in the appropriate minerals. Such data are notavailable for W but here we determine closure temperatures fromnumerical simulations of W diffusion in silicates using the model ofVan Orman et al. (2001, 2006). These results are compared to values ofTc estimated by age comparison. Ideal samples for this (i) should bewell dated with different chronometers (i.e., have a well-definedcooling history), (ii) should exhibit protracted cooling, such thatdifferences in closure temperatures result in well-resolved age dif-ferences, and (iii) should contain components having substantiallydifferent Hf/W ratios, such that precise Hf–W isochrons can bedetermined. These criteria are met by ordinary chondrites. First, thethermal evolution and structure of their parent bodies has alreadybeen studied with several chronometers (Wasserburg et al., 1969;Podosek and Brannon, 1991; Göpel et al., 1994; Trieloff et al., 2003;Amelin et al., 2005; Bouvier et al., 2007). Second, ordinary chondritesexhibit a wide range of metamorphic conditions from type 3 (unequi-librated) to type 6 (highly equilibrated), reflecting widely differentcooling histories (Dodd, 1969). Third, ordinary chondrites containabundant metal, which makes them ideal for Hf–W chronometry.Metals are virtually Hf–free but are enriched in W, resulting in 180Hf/184W∼0 inmetals and elevated 180Hf/184W ratios in the correspondingsilicates. For instance, Kleine et al. (2002) reported 180Hf/184W∼14coupled with radiogenic 182W/184W for a silicate fraction from theH4 chondrite Ste. Marguerite. Such high 180Hf/184W and radiogenic182W/184W ratios make it possible to obtain high-precision Hf–Wages.

We present Hf–W isochrons for several equilibrated H chondrites.Most of the samples investigated here were previously dated withother chronometers, including the 207Pb–206Pb system (Göpel et al.,1994; Amelin et al., 2005; Bouvier et al., 2007). The diffusivity of Pb inthe relevant minerals is relatively well constrained (Cherniak et al.,1991; Cherniak, 1998), facilitating estimates of closure temperaturesby age comparison. These estimates are compared to results fromnumerical simulations of W diffusion in a metal-silicate assemblage,which, in conjunction with the Hf–W ages, are used to assess thesignificance of the Hf–W ages and to constrain the thermal evolutionof the H chondrite parent asteroid.

2. Analytical methods

Pieces of meteorite were cleaned with abrasive paper and with0.05 M HNO3, de-ionized H2O and ethanol in an ultrasonic bath toremove any contamination introduced during cutting from largersamples. Each fragment was crushed in an agatemortar and separatedinto b40 μm and 40–150 μm fractions using nylon sieves. Duringcrushing metal grains were removed using a hand-magnet andseparated into two fractions using a 40 μm nylon sieve.

Where sufficient material was available, the coarser fraction wasfurther separated into several fractions, depending on the size of themetal grains. Silicate dust attached to or intergrown with the metalgrains was removed by repeated crushing of the magnetic fractionunder ethanol. Although all visible metal grains were removed, the40–150 μm fractions were still slightly magnetic, most likely reflectingthe presence of tiny metal inclusions in the silicate and oxide grains.The 40–150 μm fractions were further separated using a hand-magnetto obtain several “non-magnetic” fractions. These were labeled NM-n,n=1, 2, 3…, NM-1 always denoting the least magnetic fraction for eachchondrite. The NM-1 fractions are non- magnetic (i.e., with the hand-magnet used here) and might be entirely metal-free, the NM-2fractions are slightly more magnetic, and the NM-3 fractions again areslightly more magnetic than the NM-2 fractions.

All NM fractions were inspected under the binocular microscope.They consist mainly of olivine and pyroxene but most fractions alsocontain some ilmenite, feldspar and phosphates. All NM fractionswere cleaned with ethanol in an ultrasonic bath and powdered in anagate mortar. Remaining metal grains were removed from thesepowders using a hand-magnet.

The metal separates were dissolved in 15 mL Savillex® vials at∼120 °C on a hotplate using 6 M HCl–0.06 M HF. In some cases, a fewdrops of concentrated HNO3 were added. The NM fractions weredissolved in 60 mL Savillex® vials at ∼180 °C on a hotplate using HF–HNO3–HClO4 (5:4:1). After digestion, the samples were dried and re-dissolved in HNO3–H2O2 to remove organic compounds. Then thesamples were completely dissolved in 6 M HCl–0.06 M HF and a ∼10%aliquot was spiked with a mixed 180Hf–183W tracer that was calibratedagainst pure Hf and W metals (Kleine et al., 2004a).

The methods for the separation of Hf and W from the samplematrix were slightly modified from those outlined in Kleine et al.(2004a). Themetal separateswere dried, re-dissolved in 1MHF–0.1MHNO3 and loaded onto pre-cleaned anion exchange columns (2 mLBioRad® AG1X8, 200–400 mesh). The matrix was washed from thecolumn using ∼5 resin volumes 1 M HF–0.1 M HNO3 and W togetherwith other high field strength elements and Mo was eluted in 6 MHNO3–0.2 M HF (Münker et al., 2001; Weyer et al., 2002; Kleine et al.,2004a). After drying down, the W cut was re-dissolved in 1 M HF–0.1 M HNO3 and loaded onto a pre-cleaned anion exchange column(1 mL BioRad® AG1X8, 200–400 mesh).

Again, the matrix was washed from the column using ∼5 resinvolumes 1 M HF–0.1 M HNO3 but high field strength elements (Hf, Zr,Nb, Ti) were first removed in 6 M HCl–0.01 M HF before W was elutedin 6 M HCl–1 M HF. In this acid mixture, Mo is strongly adsorbed onthe anion resin (Kleine et al., 2004a).

The first part of the ion exchange procedure employed for the NMfractions is similar to thefirst step in theHf chemistry of Salters andHart(Salters and Hart, 1991). After aliquoting, the NM fractions were driedand re-dissolved in 4 M HF. The solution was centrifuged and decantedand the residue washed several times with 4 M HF. The solution wasultrasonicated several times to ensure optimal release of W from thefluoride residue and was loaded onto pre-cleaned anion exchangecolumns (3.5 mL BioRad® AG1X8, 100–200 mesh). The matrix waswashed from the column using ∼6 resin volumes of 4 M HF and Wtogether with Zr, Hf, Ti, Nb, Mo was eluted using 6 M HNO3–0.2 M HF.After drying, this cut was re-dissolved in 1 M HCl–0.5 M HF and loadedonto pre-cleaned anion exchange columns (3 mL BioRad® AG1X8, 100–200 mesh), where W was purified following the procedure of Kleineet al. (2004a). Titaniumwaswashed from the columnusingHAc–HNO3–

H2O2, Zr, Hf, and Nb were rinsed off in 6 M HCl–0.01 M HF and W waseluted in 6 M HCl–1 M HF.

Total procedural blanks ranged from ∼50 to ∼350 pg for the Wisotope composition measurements and ∼12 to 50 pg W and ∼10 pgHf for the isotope dilution measurements. The variable W blanks arecaused by the use of different batches of acetic acid.

All isotope measurements were performed using a Nu Plasma MC-ICP-MS at ETH Zürich, equipped with a Cetac Aridus desolvatingnebuliser. Prior to measurement, the samples were re-dissolved anddried several times in HNO3–H2O2 to remove organic compounds and,in the case of metal-rich samples, volatile Os oxides and then taken upin a 0.56 M HNO3–0.24 M HF mixture. Tungsten isotope compositionsof metals and whole-rocks were typically measured with a signalintensity of ∼2 V on 182W, which was obtained for a ∼20 ppb Wsolution. For these samples, 60 ratios (3 blocks of 20 ratios) weremeasured resulting in within-run statistics of the order of 0.2 ε units(2σ). Owing to the low W contents in the NM fractions, their Wisotope compositions were measured in 1 or 2 blocks of 20 ratios eachwith signal intensities of ∼0.5 to 1 V on 182W. Thewithin-run statisticsof these measurements were typically between 0.5 and 1 ε unit.Instrumental mass bias was corrected relative to 186W/183W=1.9859

107T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

96

Page 97: Touboul Thesis Completed

using the exponential law. Small isobaric interferences of Os onmasses 184 and 186 were corrected by monitoring 188Os and werenegligible. The 182W/184W and 183W/184W ratios of all samples weredetermined relative to two standard runs bracketing the sample runand are reported in ε18iW units, which is the deviation of the 18iW/184W ratio from the terrestrial standard value in parts per 10,000. Thereproducibility of the ∼20 ppb standard during one measurement dayis typically equal to or better than ∼0.3 ε units (2 SD) for the 182W/184W ratio and ∼0.2 ε units (2 SD) for the 183W/184W ratio. The externalreproducibility of the W isotope measurements typically is 0.3–0.4 ε

units (2 SD) for the 182W/184W ratio and 0.2–0.3 ε units (2 SD) for the183W/184W ratio and was estimated by repeated analysis of a whole-rock powder of the Kernouvé H6 chondrite and several H chondritemetals (Table 1). The external reproducibility of the isotope measure-ments of these samples is similar to those obtained for theW standardduring one measurement session, indicating that matrix effects areminor to absent. The uncertainties for theW isotope measurements oftheW-poor NM fractions (i.e., most of the NM-1 and -2 fractions) wereassessed by repeated measurement (1 block of 20 ratios each) of a∼10ppbWstandard solutions that yielded anexternal reproducibilityof

Table 1Hf–W data for metals, whole-rocks and non-magnetic fractions

Sample Hf(ppb)

W(ppb)

180Hf/184W ε182W±2σ meas. ε182W±2σ corr. ε183W±2σ

Ste. Marguerite (H4)Metal 1.81 823.0 0.00251±2 −3.20±0.25 0.05±0.15

−3.19±0.18 −0.21±0.12−3.13±0.19 0.02±0.14

Mean −3.18±0.07 −0.05±0.29NM 223.0 16.31 15.6±1 12.8±0.3 2.9±0.2

Richardton (H5)A-metal (N150 μm) 30.34 660.9 0.0523±4 −3.47±0.19 −0.05±0.11

−3.32±0.19 0.06±0.13−3.27±0.20 −0.03±0.12−3.06±0.18 0.17±0.16

Mean −3.28±0.35 0.04±0.20A-NM-1 212.9 18.31 13.3±2 6.9±0.4 7.4±0.8 0.03±0.19Repl. 212.1 18.23 13.3±1 5.9±0.5 6.9±1.4 0.97±0.52A-NM-2 174.9 18.28 10.9±3 4.1±0.5 4.5±0.9 0.00±0.21A-NM-3 166.9 26.29 7.2±1 2.1±0.3 2.2±0.6 2.63±0.22A-NM-4 162.3 26.31 7.0±3 1.8±0.6 2.0±0.9 0.69±0.29B-metal (N230 μm) 15.2 744.6 0.0233±2 −3.57±0.16 0.04±0.08B-metal (40–230 μm) 8.11 740.9 0.0125±1 −3.18±0.19 −0.12±0.15B-NM 183.0 18.86 11.06±8 3.7±0.7 4.0±1.1 0.55±0.48C-WR 143.6 198.9 0.827±1 −2.79±0.26 −0.14±0.18

ALH84069 (H5)Metal (40–150 μm) 13.9 687.0 0.0231±2 −2.82±0.19 0.10±0.15

−3.20±0.26 0.13±0.18Mean −3.01±0.53 0.12±0.06Metal (N150 μm) 17.4 697.8 0.0284±2 −3.14±0.20 −0.05±0.13

−3.18±0.18 0.07±0.13−3.20±0.17 0.01±0.10−2.85±0.15 0.22±0.13

Mean −3.08±0.39 0.10±0.21NM-1 239.2 5.35 47.4±3 25.3±0.7 32.5±6.1 −0.41±0.55NM-2 151.6 7.48 20.8±2 9.6±0.8 12.6±3.1 −0.16±0.47NM-3 139.3 11.58 12.0±1 4.3±0.5 5.4±1.3 −0.39±0.48

Estacado (H6)Metal (N150 μm) 9.64 826.1 0.01331±9 −2.98±0.21 −0.13±0.12

−2.98±0.16 −0.08±0.12−3.03±0.16 −0.12±0.11−2.82±0.19 −0.02±0.12−2.72±0.15 −0.02±0.10

Mean −2.90±0.26 −0.07±0.10NM-1 148.5 10.5 16.2±1 4.8±1.0 5.1±1.1 −1.15±0.48NM-2 134.3 21.1 7.27±5 1.1±0.5 1.6±0.5 2.44±0.34

Kernouvé (H6)Metal (40–500 μm) 1.94 926 0.00238±2 −2.86±0.19 −0.17±0.16

−3.04±0.24 −0.12±0.14Mean −2.95±0.26 −0.15±0.07Metal (N500 μm) 0.31 857 0.000415±3 −3.04±0.20 −0.13±0.14

−2.72±0.18 −0.02±0.14−2.85±0.21 0.03±0.15

Mean −2.87±0.32 −0.04±0.16NM-1 175.9 3.9 51.0±4 23.1±0.7 25.7±2.3 −0.10±0.30NM-2 125.7 13.3 10.8±1 2.7±0.8 2.9±1.1 −0.84±0.55Whole-rock 141.3 183.4 0.878±7 −2.66±0.22 −0.14±0.14

−2.41±0.20 −0.16±0.17−2.53±0.15 −0.01±0.10

Mean −2.53±0.26 −0.10±0.16

NM=non-magnetic fraction. The quoted 2σ uncertainties for 180Hf/184W refer to the last significant digits. The quoted 2σ uncertainties for themeasured ε182Wand ε183Ware standarderrors of the individual mass spectrometric runs and those for the corrected ε182W were calculated by propagating the external reproducibility of the isotope measurements and a50% uncertainty on the blank correction.The meteorites are from the following collections: Ste. Marguerite (MNHN); Richardton-A (USNM); Richardton-B (MNHN); Richardton-C (Senckenbergmuseum Frankfurt); ALH84069 (NASA); Kernouvé (ETH); Estacado (BM).

108 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

97

Page 98: Touboul Thesis Completed

∼0.8 ε182W (2SD). This is similar to thewithin-run statistics obtained forthemeasurements of theW-poorNMfractions (Table 1),which for thesesamples is used as the uncertainty for the measured 182W/184W. Notethat themajor source of uncertainty in the 182W/184Wof theW-poorNMfractions is the blank correction (see below and Table 1).

The accuracy of the measurements was monitored by analyzingseveral carbonaceous chondrites, which all yielded the previouslydetermined value of −1.9±0.1 ε182W (Kleine et al., 2004a). Furthermore,183W/184W ratios were used as a monitor for accurate measurementsand agree for most samples to within ±0.2 ε units with the terrestrialvalue (Table 1). Elevated 183W/184W ratios for two NM fractions of theRichardton meteorite are attributed to an organic interference on mass183 that was successfully removed for all other samples by treatmentwith HNO3–H2O2. Elevated measured 183W/184W ratios have beenobserved before for W isotope measurements of some eucrites andcarbonaceous chondrites and for these samples the 182W/184W rationormalized to 186W/184W=0.96727 agrees with 182W/184W ratios forother eucrites and carbonaceous chondrites, respectively (Kleine et al.,2004a). This indicates that only 183W is affected, such that for the twoNM fractions from Richardton with elevated measured 183W/184W thereported ε182W values were calculated from the 182W/184W rationormalized to 186W/184W. Note, that the ε182W values thus obtainedare consistent with the Hf–Wdata for the other fractions of Richardtonand also with Hf–W data for the other H5 chondrite ALH84069 (i.e., allfractions plot on one well-defined isochron).

3. Results

The Hf and W concentrations and the W isotope composition ofmetals, whole-rocks and non-magnetic fractions analyzed for thisstudy are given in Table 1. The H chondrite metals have Wconcentrations ranging from ∼660 to ∼926 ppb, consistent withpreviously published results (Rambaldi,1976; Kong and Ebihara,1996;Humayun and Campbell, 2002). These W concentrations are substan-tially higher than those reported for metals from unequilibrated Hchondrites (Rambaldi, 1976; Kong and Ebihara, 1996; Humayun andCampbell, 2002). The 182W/184W ratios of the H chondrite metals areindistinguishable from each other but appear to increase slightly withmetamorphic grade from −3.2 for H4 chondrites to −2.9 ε182W for H6chondrites. Some of the Richardtonmetals do not follow this trend andhave ε182W values as low as −3.57±0.16. Note that given an externalreproducibility of ∼0.3–0.4 ε units for the W isotope measurements,the 182W/184W of this sample is not distinguishable from those of theother H chondrite metals. The various NM fractions have low W

contents between∼4 and∼26 ppb andHf contents ranging from∼125to ∼239 ppb, resulting in high 180Hf/184W ratios from ∼7 to ∼51 andelevated 182W/184W ratios from ∼2 to ∼33 ε182W (Table 1).

Owing to the lowW contents and radiogenic 182W/184W of the NMfractions, the major source of uncertainty in the ages is the blankcorrection, which is significant for some of the NM fractions. Typicallythe blank corrections ranged from b1 to ∼3 ε182W. The NM-1 fractionof ALH 84069 required a larger correction of ∼7 ε182W, reflecting ahigher W blank, which was caused by the use of insufficiently cleanacetic acid (Table 1). In spite of this large correction the Hf–Wdata forthe NM-1 fraction from ALH 84069 are consistent with the Hf–Wdatafor its other fractions that did not require such large corrections.

As shown in Figs. 1–3, the 180Hf/184W ratios and ε182W valuescorrelate for each of the analyzed H chondrites, such that precise

Fig. 1. ε182W versus 180Hf/184W for Ste. Marguerite. Data shown with open symbols arefrom Kleine et al. (2002), those with filled symbols from this study. m=initial 182Hf/180Hf, i= initial ε182W. Regressions are calculated using themodel 1 fit of IsoPlot (Ludwig,1991). Details regarding the calculation of ages are given in the text. ΔtCAI is theformation interval relative to CAIs; the absolute age t is calculated relative to theangrites D'Orbigny and Sahara 99555 (see text).

Fig. 2. ε182W versus 180Hf/184W for H5 chondrites Richardton and ALH84069. m=initial182Hf/180Hf, i= initial ε182W. Regressions are calculated using the model 1 fit of IsoPlot(Ludwig,1991). Details regarding the calculation of ages are given in the text. Data shownas filled grey symbols (Richardtonwhole-rock and coarse-grained metal fractions) werenot included in the regression and are not shown in the isochron plot for the combinedH5 chondrites. ΔtCAI is the formation interval relative to CAIs; the absolute age t iscalculated relative to the angrites D'Orbigny and Sahara 99555 (see text).

109T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

98

Page 99: Touboul Thesis Completed

isochrons could be obtained (MSWDb1 in most cases). The uncer-tainties on the slopes of the isochrons are better than ∼7% in mostcases, resulting in uncertainties for the ages on the order of ∼1 Ma.Only for Estacado the isochron has a higher uncertainty of ∼12%,resulting in an uncertainty of ±1.7 Ma.

For Richardton the scatter on the isochron is slightly largercompared to the other H chondrites, indicating a slight disturbanceof the Hf–W system. A regression including all the Richardtondata yields a precise isochron (MSWD=1.2) corresponding to aninitial 182Hf/180Hf of (6.60±0.35)×10−5 and an initial ε182W of −3.39±0.17. This initial ε182W is indistinguishable from the initial ε182W ofAllende CAIs of −3.30±0.12 (Burkhardt et al., submitted for publica-tion) and corresponds to a W model age of −2±4 Ma (using the Wisotope composition of Kernouvé as a reference), inconsistent withthe age obtained from the initial 182Hf/180Hf of the Richardton

isochron. Compared to the coarse-grained metal, the fine-grainedRichardton metal has a slightly higher ε182W value, which is identicalto the 182W/184W of the ALH84069 metal and also consistent with thetrend of slightly increasing ε182W values defined by the other Hchondrite metals (see Section 4.5). The model age of the fine-grainedRichardton metal is 3±5 Ma, consistent with the isochron age of 5.6±1.1 Ma. This indicates that the 182W/184W ratio of the coarse-grainedRichardton metal has been displaced to too low values and yieldsspurious ages. This has probably also affected the measured 182W/184W of the Richardton whole-rock, which is slightly displaced tolower ε182W values compared to the Kernouvé whole-rock (Table 1)and other H chondrite whole-rocks (Kleine et al., 2007). The origin ofthese low ε182W values remains enigmatic but could be related to theincorporation of irradiated metals with low ε182W values, as has beenobserved for many iron meteorites (Kleine et al., 2005a). However, toour knowledge there is no other evidence for the presence of suchmetal in Richardton. Note that this has no effect on the interpretationof the Hf–W age for Richardton, which is obtained from the slope ofthe isochron. This remains unchanged regardless of whether the twometal fractions with low ε182W and the whole-rock were includedin the isochron regressions. Excluding these fractions from the reg-ression yields a well-defined isochron (MSWD=1.5) with an initial182Hf/180Hf of (6.47±0.44)×10−5 and initial ε182W=−3.23±0.38 (Fig. 2).

4. Discussion

4.1. Hf–W isochron ages for H chondrites

To define an isochron the minerals of an H chondrite must oncehave been in W isotope equilibrium, i.e., they must have had the sameW isotope composition initially. Given that the H chondrite fractionswere obtained mainly based on their magnetic susceptibility, thecorrelation of ε182W with 180Hf/184W could potentially represent amixing line betweenW-rich metal and virtually W-free silicates. Sucha mixing line would have no chronological significance if the two

Fig. 3. ε182W versus 180Hf/184W for H6 chondrites Kernouvé and Estacado. m=initial182Hf/180Hf, i= initial ε182W. Regressions are calculated using the model 1 fit of IsoPlot(Ludwig, 1991). Details regarding the calculation of ages are given in the text.ΔtCAI is theformation interval relative to CAIs; the absolute age t is calculated relative to theangrites D'Orbigny and Sahara 99555 (see text).

Fig. 4. Hf versus W contents for the different fractions of the analyzed H chondrites.Data for some of the NM (non-magnetic) fractions from Ste. Marguerite are from Kleineet al. (2002). The Hf and W concentrations in the coexisting phases of these Hchondrites are not colinear, indicating that presence of at least three independentcomponents for Hf andW among the coexisting phases. These components are high-Capyroxene+ ilmenite, olivine+low-Ca pyroxene, and metal.

110 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

99

Page 100: Touboul Thesis Completed

endmembers had different initial 182W/184W ratios. However, asshown in Fig. 4, the variations in 180Hf/184W ratios among the analyzedH chondrite fractions require the presence of at least threeindependent components for Hf and W among the coexisting phases.The major host of W is metal, which constitutes one component. Themajor hosts of Hf are high-Ca pyroxene and ilmenite (Righter andShearer, 2003) and since no pure high-Ca pyroxene and ilmeniteseparates could be obtained these two phases are considered togetheras one component. The third component encompasses olivine andlow-Ca pyroxene and is characterized by low Hf and W contents.These twominerals are considered as one component because no pureolivine and low-Ca pyroxene separates were obtained. Both olivineand low-Ca pyroxene are not capable of incorporating significantamounts of either Hf or W (Righter and Shearer, 2003), such that theirpresence mainly causes dilution of the high Hf content of high-Capyroxene and ilmenite. Given that high-Ca pyroxene, ilmenite, olivine,and low-Ca pyroxene have similar and low W contents, the slightlyhigher W contents of the NM-3 and -4 fractions compared to the NM-1 and -2 fractions of the same meteorite most likely reflect thepresence of some metal in the NM-3 and -4 fractions.

The presence of at least three independent components withregard to Hf and W among the coexisting phases of H chondritesreveals that the correlation between ε182W and 180Hf/184W observedfor the fractions of each of the H chondrites cannot reflect simplebinary mixing between W-rich metal and virtually W-free silicates.This is also apparent from plots of ε182W vs. 1/W, in which binarymixtures form straight lines. This is not the case for any of themeteorites investigated here, such that the linear correlations inthe ε182W vs. 180Hf/184W plots cannot be mixing lines (for the H6chondrites the data seem to plot on straight lines but in the case ofKernouvé theMSWDof this ε182W vs.1/W line is 4.5 instead of 0.28 forthe isochron). Each of the fractions, therefore, evolved to radiogenicε182W according to their 180Hf/184W. Hence, the Hf–W data for the Hchondrite fractions define isochrons and can be interpreted to havechronological significance.

Relative Hf–W ages (or formation intervals), ΔtCAI, are calculatedfrom the initial 182Hf/180Hf ratios obtained from the slopes of theisochrons relative to an initial 182Hf/180Hf=(1.003±0.045)×10−4 forCAIs and refer to the time of Hf–W closure in a sample elapsed sincecrystallization of type B CAIs (Burkhardt et al., submitted forpublication). With increasing metamorphic grade, the Hf–W agesof the H chondrites become increasingly younger and range fromΔtCAI=1.7±0.7 Ma for the H4 chondrite Ste. Marguerite to ΔtCAI=9.6±1.0 Ma for the H6 chondrites Kernouvé and Estacado. The Hf–Wages for the H5 chondrites Richardton and ALH 84069 are inter-mediate between the ages for the H4 and H6 chondrites and areΔtCAI=5.9±0.9 Ma.

The comparison of relative Hf–W ages and absolute Pb–Pb agesrequires conversion of Hf–W formation intervals to absolute ages,which in turn requires knowledge of the initial 182Hf/180Hf and theabsolute age of Hf–Wclosure in a sample. Due to differences in closuretemperatures of different chronometers, the ideal samples to obtainsuch information are angrites because (i) they cooled rapidly, such thatdifferences in closure temperatures do not result in resolvable agedifferences, and (ii) they exhibit high U/Pb ratios, such that precise Pb–Pb ages are available (Lugmair andGaler,1992; Amelin, 2008). Themostprecise Pb–Pb age for the angrites D'Orbigny is 4564.42±0.12 Ma(Amelin, 2008). For the angrite Sahara 99555, the earlier reported Pb–Pb age of 4566.18±0.14 Ma (Baker et al., 2005) has now been revisedand two identical ages of 4564.58±0.14 and 4564.86±0.38 Ma,obtained using different techniques for the removal of Pb contamina-tion, were reported (Connelly et al., 2008). Mineral separates fromD'Orbigny and Sahara 99555 plot on one well-defined Hf–W isochron(MSWD=1.4) yielding an initial 182Hf/180Hf of (7.31±0.16)×10−5

[recalculated from the Hf–W data reported in Markowski et al. (2007)and using the model 1 fit of IsoPlot], consistent with identical Pb–Pb

ages for these two angrites. Here we calculate absolute Hf–W agesrelative to an initial 182Hf/180Hf=(7.31±0.16)×10−5 at 4564.50±0.23Ma(i.e., the average of the aforementioned Pb–Pb ages for D'Orbigny andthemore precise age for Sahara 99555). Identical results are obtained ifabsolute Hf–W ages were calculated relative to D'Orbigny only orSahara 99555 only.

This approach for calculating absolute Hf–W ages is based on theassumption that the Pb–Pb ages for D'Orbigny and Sahara 99555accurately date the crystallization of these rocks. Several Pb–Pbstudies on angrites, however, reported Pb–Pb ages for the sameangrite that are distinct outside of the reported age uncertainties(Baker et al., 2005; Amelin, 2008; Connelly et al., 2008). Nevertheless,given that two groups obtained identical high-precision Pb–Pb agesfor the angrites D'Orbigny and Sahara 99555 (Amelin, 2008; Connellyet al., 2008), these ages appear reliable and their use as reference agesfor Hf–W chronometry justified. Moreover, Hf–W ages for two otherangrites (Kleine et al., 2008) are consistent with their Pb–Pb ages(Amelin and Irving, 2007), indicating that the intercalibration of Hf–Wand Pb–Pb ages provides reliable results.

4.2. Closure temperature for the Hf–Wsystem in equilibratedH chondrites

To evaluate the significance of the Hf–W ages for constraining thethermal evolution of meteorite parent bodies, it is necessary to knowthe closure temperature for W diffusion in the appropriate silicate-metal mixture. In H chondrites, the major hosts of radiogenic 182Warehigh-Ca pyroxene and ilmenite and each of these minerals might haveits distinct Hf–W closure temperature. In slowly cooled metamorphicrocks such as H chondrites, one of these twominerals may have stayedopen while the other had already closed. This would result in scatteron the isochron but this is not observed for the data presented here.This indicates that there are no significant differences in the Hf–Wclosure temperatures of high-Ca pyroxene and ilmenite in Hchondrites, given that the variable Hf contents in the different NMfractionsmost likely reflect different proportions of ilmenite and high-Ca pyroxene in these fractions.

There are no experimental data available for diffusion ofW in high-Ca pyroxene or ilmenite that would allow calculation of the closuretemperature as a function of effective grain size and cooling rate.Based on the comparison of Hf–Wages for eucrite metals with Pb–Pbages for the host eucrites, Kleine et al. (2005b) estimated the closuretemperature of the Hf–W system in basaltic eucrites to be at least∼600 °C. Here we estimate the closure temperature of the Hf–Wsystem by modeling the diffusion behavior of W in high-Ca pyroxenesand test these results by comparison to Pb–Pb ages for chondrites.

The diffusion behavior of W in high-Ca pyroxene was evaluated by(i) using the model presented by Van Orman et al. (2001) to estimatethe diffusion parameters for W in high-Ca pyroxene and by (ii)modeling the diffusion behavior of W in a high-Ca pyroxene-metalsystem. Tungsten is assumed to have a charge of +4, an ionic radius of0.066 nm (Shannon, 1976) and is assumed to reside on the 6-foldcoordinated M1 site in high-Ca pyroxene, which has an ideal radius of0.072 nm and metal–oxygen bond length of 0.22 nm. Assuming thatthe Van Orman et al. (2001) model applies to cations that occupy theM1 site –which appears reasonable since the model predicts diffusioncoefficients for Fe2+ on the M1 site that are in good agreement withexperimental data (Azough and Freer, 2000) – gives an activationenergy estimate of 453 kJ/mol and a pre-exponential factor of9.53×10−5 m2/s.

Simultaneous production and diffusive exchange of radiogenic Wbetween high-Ca pyroxene and metal was simulated numericallyusing the model of Van Orman et al. (2006). We chose to use thisnumerical model rather than the analytical models for closuretemperature presented by Dodson (1973) and Ganguly and Tirone(2001) because the analytical models make several assumptions thatdo not necessarily apply to the cases considered here. For example, the

111T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

100

Page 101: Touboul Thesis Completed

Dodson (1973) and Ganguly and Tirone (2001) models assume (i) aninfinite sink for radiogenic daughters (which is a valid assumption inthe case of H chondrites); (ii) a decay time that is very long comparedto the cooling time (which might not be valid for short-livedchronometers); and (iii) that heating at peak metamorphic conditionswas sufficient to homogenize the high-Ca pyroxene. The numericalmodel used here does not rely on these assumptions, and is thus amore realistic model for the production and exchange of radiogenicdaughters in short-lived isotope systems. As will be shown below, theassumption that peakmetamorphic conditions were sufficient to resetthe Hf–W system is not valid in the case of the H4 chondrites.Assessing the effects that metamorphism had on the Hf–W systemin H4 chondrites therefore requires a model that can simulate theprograde path, such as the one used here.

In the model, exchange of radiogenic W is controlled by diffusionwithin spherical high-Ca pyroxene grains, which maintain partition-ing equilibrium with metal at their surfaces. Because diffusion in Fe–Ni alloys is many orders of magnitude more rapid than in high-Capyroxene (Watson andWatson, 2003), themetal is assumed to remainhomogeneous throughout the simulation. In most simulations, 182W isassumed to be distributed in chemical and isotopic equilibrium be-tween the high-Ca pyroxene and metal at the peak temperature.However, we also performed simulations, described below as appliedto H4 and H5 chondrites, in which the sample had experienced a coldpre-history, during which radiogenic W was not transferred from thehigh-Ca pyroxene to the metal.

The system is assumed to cool linearly with time from the peaktemperature, and the final age is calculated based on the integratedconcentrations of 182W in the high-Ca pyroxene and the metalresulting from the decay of 182Hf. This age corresponds to a particulartemperature along the cooling path, which is by definition the closuretemperature of the system. The cooling time from peak temperaturesto temperatures at which diffusive exchange becomes negligible is∼10 Ma, which is similar to the 182Hf half-life. In the cases consideredhere, diffusive exchange ceased before 182Hf had completely decayed,such that while the system remains open to exchange, the decay of182Hf is significant.

The closure temperature estimates shown in Fig. 5 assume aninitial temperature of 1000 °C. This temperature provides an upperlimit for the peak temperature of H6 chondrites because at ∼1000 °Cmelting in the FeNi–FeS system begins but the texture of H6chondrites reveals that such melting did not occur. Assuming an

initial temperature of ∼1000 °C appears reasonable because tem-perature estimates for H6 chondrites using the two-pyroxenethermometer (Lindsley, 1983) yield a temperature range of 865–926 °C (Slater-Reynolds and McSween, 2005), only slightly below1000 °C. The peakmetamorphic temperatures for H4–5 chondrites areless well constrained, mainly because in these rocks the pyroxenes arenot completely equilibrated, such that two-pyroxene thermometrycannot be applied. Based on temperature estimates for type 3(obtained from Ni profiles in taenite) and type 6 chondrites, Dodd(1981) estimated the peak temperatures for type 4 and type 5chondrites to 600–700 °C and 700–750 °C, respectively. In morerecent studies olivine–spinel thermometry was used to determinetemperatures for type 4–6 chondrites and the results for H4–6chondrites tightly cluster between 675 and 750 °C (Wlotzka, 2005;Kessel et al., 2007). These temperatures provide a lower limit for thepeak temperatures and the tight cluster of olivine–spinel tempera-tures suggest that peak temperatures for H4–6 chondrites were notvery different (Wlotzka, 2005; Kessel et al., 2007).

Fig. 5 shows closure temperatures calculated as a function ofcooling rate, for an initial temperature of 1000 °C and four differenthigh-Ca pyroxene grain diameters between 1 and 150 μm. The high-Capyroxene/metal ratio in these simulations is 0.5 and the high-Capyroxene/metal partition coefficient for W is 0.01 (Walter andThibault, 1995), but the results are not sensitive to variations inthese parameters unless the high-Ca pyroxene/metal ratio and/orpartition coefficient become much larger.

Fig. 5 reveals that the closure temperature of the Hf–Wsystem in Hchondrites is strongly dependent on the grain size of the high-Capyroxenes, particularly for grain sizes below ∼20 μm. As a conse-quence, Tc increases from H4 to H6 chondrites because the grain sizesof high-Ca pyroxenes increase. In H6 chondrites, the high-Capyroxenes are 5–30 μm in diameter, whereas in H5 chondrites theyare 2–5 μm but can also be larger (Huss et al., 2006). In H4 chondrites,high-Ca pyroxene microcrystallites may have diameters of less than1 μm but high-Ca pyroxenes also occur as euhedral grains of less than∼10 μm in the mesostasis and form rims around olivine and low-Capyroxene that are ∼10–20 μm across (Huss et al., 2006; C. Alexanderand J. Grossmann, pers. comm. 2007). Using these grain sizes thefollowing values for Tc are obtained (Fig. 5): ∼800–875 °C (H6); ∼750–850 °C (H5); ∼725–850 °C (H4). However, these temperatureestimates assume that high-Ca pyroxene and metal are always indirect contact but given that high-Ca pyroxene is only a minorconstituent in H chondrites this will probably not be the case. A morerealistic approach is to assume that the high-Ca pyroxene grains aresurrounded and sometimes even enclosed by large olivine and low-Capyroxene grains (with grain diameters between 20 and 200 μm).These grains would constitute a barrier for W diffusion from the high-Ca pyroxene to the metal and the estimates for Tc presented above arethen lower limits.

A higher limit for Tc is given by the peak temperature of H6chondrites that must have been below ∼1000 °C and above the two-pyroxene temperature of ∼900 °C (see above). It therefore isreasonable to assume that the higher limit of Tc in H6 chondrites is∼950 °C. For H5 chondrites the peak temperature probably wassomewhat lower and here we assume ∼900 °C. Combined with thelower limits for Tc obtained from Fig. 5 and based on the grain sizes ofthe high-Ca pyroxenes the following values for Tc are obtained: 875±75 °C for H6 chondrites and 825±75 °C for H5 chondrites.

The Hf–W closure temperature for H4 chondrites is more difficultto estimate than those for H5 and H6 chondrites because the peaktemperature in H4 chondrites as well as the host phase(s) ofradiogenic 182W and its grain size(s) in H3 chondrites are less wellconstrained. In type 3 chondrites, Hf is enriched in themesostasis (andhigh-Ca pyroxenes therein) of chondrules and in high-Ca rims onolivine and low-Ca pyroxene grains in chondrules (Alexander, 1994).Tungsten most likely is concentrated in metal grains outside

Fig. 5. Closure temperature of the Hf–W system as function of grain diameter andcooling rate. The assumed starting temperature is 1000 °C. Details regarding thecalculations are given in the text. The dashed areas indicate two-pyroxene and olivine–spinel temperatures for H chondrites. The grey boxes indicate typical cooling rates andgrain sizes of high-Ca pyroxene for each of the petrologic types of H chondrites.

112 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

101

Page 102: Touboul Thesis Completed

chondrules and in the matrix (see below). The rate of diffusivetransport of radiogenic 182W from the chondrules into metal outsidethe chondrules is difficult to model and critically depends on the exactlocation of W in the chondrules and, as is evident from Fig. 5, on thehigh-Ca pyroxene grain size. The range in grain sizes of b1 μm to∼10 μm corresponds to a range in Tc of ∼100 °C (Fig. 5). In addition, therate of W diffusion depends on the grain sizes of olivine and low-Capyroxene inside the chondrules, which control how many grainboundary paths there are between the high-Ca pyroxenes and themetals. A lower limit for Tc may be obtained by assuming thatradiogenic 182W is located inside ∼0.1 μm high-Ca pyroxene micro-crystallites, which are in direct contact to metals outside the chon-drules. In this model, a closure temperature of ∼700 °C is calculated.The true value for Tc will be significantly higher because larger high-Ca pyroxenes are present and the high-Ca pyroxenes inside thechondrules andmetal grains outside the chondrules will mostly not bein direct contact. A higher limit for Tc may be obtained by assumingthat all high-Ca pyroxenes have grain sizes of ∼10 μm, in which casethe closure temperature will be ∼850 °C (Fig. 5). Given that many ofthe high-Ca pyroxene grains in H4 chondrites have sizes larger than∼1 μm, Tc will probably be higher than ∼750 °C (Fig. 5) and the bestestimate for Tc in H4 chondrites then is 800±50 °C.

The comparison of the Hf–Wages and other ages provides a test forthe validity of the above closure temperature estimates. For Ste.Marguerite, Kernouvé and Estacado the Hf–W ages are the oldestradioisotopic ages reported for these meteorites so far, for Richardtonthe Hf–W and Pb–Pb chondrule ages are indistinguishable (Table 2).Themost reliable approach for estimating closure temperatures by agecomparison uses slowly cooled samples because then differences inclosure temperatures among different isotope systems and mineralswill result in resolvable age differences. The Hf–Wage for Kernouvé is∼23 Ma older than Pb–Pb whole-rock and pyroxene ages (Göpel et al.,1994; Bouvier et al., 2007) for this meteorite and the Hf–W age forEstacado is ∼30 Ma older than a Pb–Pb chondrule age (Blinova et al.,2007) (Table 2). This indicates that the Hf–W closure temperaturemust be well above the closure temperature for Pb diffusion inpyroxenes (Cherniak, 1998; Amelin et al., 2005), consistent with theobservation that Hf–W ages for eucrite metals are older than Pb–Pbages for their host eucrites (Kleine et al., 2005b). The closure tem-perature for Pb diffusion in pyroxenes in H6 chondrites was estimatedto be 780±100 °C for grain sizes of 20–200 μm (Amelin et al., 2005),which is indistinguishable from but appears to be slightly lower thanour estimate for the Hf–W closure temperature in H6 chondrites of

875±75 °C. This indicates that the closure temperature estimates forthe Hf–W system presented here are reasonable. It is important tonote that the Pb closure temperature for pyroxenes in H chondritesmight be lower than 780±100 °C because in these samples the grainsizes of the high-Ca pyroxene (Huss et al., 2006) are smaller than the20–200 μm range used in the calculation by Amelin et al. (2005). High-Ca pyroxenes are probably an important host of U among the silicateminerals of H chondrites because U fits much better into the M2(i.e., Ca) site of pyroxenes and should therefore be enriched in high-Carelative to low-Ca pyroxene. For grain sizes of 5–30 μm the closuretemperature for Pb diffusion ranges from ∼650 °C to ∼780 °C anddecreases to temperatures as low as ∼550 °C for grain sizes of ∼1 μm.As for the Hf–W system these closure temperature estimates probablyare lower limits because they do not take into account the effects oflarge olivine and low-Ca pyroxenes (e.g., as barriers for Pb diffusion).However, these calculations reveal that the Hf–Wclosure temperaturein H6 chondrites is distinctly higher than the U–Pb closuretemperature in pyroxenes, consistent with the well-resolved differ-ences in Hf–W and Pb–Pb ages for Kernouvé and Estacado.

4.3. Significance of the Hf–W ages

To utilize Hf–W ages for H chondrites for constraining thetimescales of parent body accretion, heating and cooling it is essentialto identify which “events” are being dated. In the case of metamorphicrocks such as H chondrites, these could be (i) cooling from peakmetamorphic temperatures below Tc, (ii) mineral growth duringmetamorphism (this could take place below Tc), or (iii) a pre-metamorphic event (in the case that metamorphism was not capableof resetting the Hf–W system). The interpretation of the Hf–W agescritically depends on (i) whether heating above Tc was achieved, and(ii) how efficiently any initial W isotope heterogeneity, produced bythe decay of 182Hf in phases having different Hf/W, has been erased bythe thermal metamorphism.

4.3.1. Hf–Wage of the H4 chondrite Ste. Marguerite: timing of chondruleformation

The lower limit of the Hf–Wclosure temperature for H4 chondritesoverlaps with the upper limit of ∼750 °C of olivine–spinel tempera-tures determined for this chondrite group (Wlotzka, 2005; Kesselet al., 2007). Peak temperatures of H4 chondrites may be higher thanthe olivine–spinel temperatures but to what extent is unknown. It istherefore difficult to estimate if H4 chondrites were heated above Tcbut it appears that for complete resetting of the Hf–W system heatingto ∼800–850 °C would be required. This is only slightly below thepeak temperatures of H6 chondrites and it seems unlikely that this hasbeen achieved. Moreover, the high olivine–spinel temperature for H4chondrites may in part reflect an earlier high-temperature event (suchas chondrule formation), in which case the peak temperatures of theH4 chondrites could be below 700 °C, consistent with temperatureestimates by Dodd (1981).

To evaluate whether the Hf–W age for Ste. Marguerite reflectsparent body metamorphism, we numerically simulated the diffu-sional exchange of W between clinopyroxene and metal using themodel of Van Orman et al. (see above). In the simulation, we led aclinopyroxene-metal system evolve for 1 Ma at low temperatures,such that no diffusional exchange between clinopyroxene and metalcould occur, then instantaneously heated to 800 °C and led the systemcool at ∼200°/Ma. The simulations indicate that such a scenario is notcapable of erasing the previously accumulated radiogenic W isotopesignature of the clinopyroxene. In this particular scenario, the age ofthe system would only shift by 0.07 Ma, i.e., if 2.7 Ma would be thetime of heating, the Hf–W age would still be 1.77 Ma and, hence,almost entirely reflect the earlier event.

Two lines of evidence further suggest that the Hf–W age for Ste.Marguerite has not been reset by parent body metamorphism. First,

Table 2Compilation of radiometric ages for selected H chondrites (in Ma)

Sample Minerals Age(Ma)±2σ

Tc(°C)

System References

Ste. Marguerite metal-silicate 4566.9±0.5 800±50 182Hf–182W this studywhole-rock,chondrules

4564.4±3.4 650±100 207Pb–206Pb (3,4)

phosphates 4562.7±0.6 477±100 207Pb–206Pb (4)Richardton metal-silicate 4563.0±0.9 825±75 182Hf–182W this study

chondrules 4562.7±1.7 725±100 207Pb–206Pb (1)phosphates 4550.7±2.6 477±100 207Pb–206Pb (1,4)

Estacado metal-silicate 4558.6±1.6 875±75 182Hf–182W this studychondrules 4527.6±6.3 777±100 207Pb–206Pb (2)phosphates 4492±15 477±100 207Pb–206Pb (2)

Kernouvé metal-silicate 4559.2±1.0 875±75 182Hf–182W this studywhole-rock;pyroxene-olivine

4536±7 777±100 207Pb–206Pb (3,4)

phosphates 4522.5±1.5 477±100 207Pb–206Pb (4)

The Pb–Pb ages for whole-rocks and silicates from Ste. Marguerite and Kernouvé areaverages of the ages reported in the literature. The uncertainties are calculated as standarddeviations (2σ) of these ages. For estimates of the closure temperatures see text. (1) Amelinet al. (2005); (2) Blinova et al. (2007); (3) Bouvier et al. (2007); (4) Göpel et al. (1994).

113T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

102

Page 103: Touboul Thesis Completed

the Hf–W age for Ste. Marguerite of 1.7±0.7 Ma is identical to Al–Mgages for chondrules from L and LL chondrites (Russell et al., 1996; Kitaet al., 2000; Rudraswami and Goswami, 2007). Although no ages forchondrules from H chondrites are available it seems likely that Hchondrules formed at the same time as L and LL chondrules, given thatboth L and LL chondrules have identical average Al–Mg ages of ∼2 Ma(Russell et al., 1996; Kita et al., 2000; Rudraswami and Goswami,2007). Hence, H chondrules probably formed at ∼2 Ma, which is thetime given by the Hf–W age for Ste. Marguerite. Second, sincechondrules formed before the assembly of their host parent body,chondrule ages provide the earliest time at which assembly of andhence heating inside the parent body can have started. Thermalmodeling of spherical asteroids heated by 26Al decay indicate that atemperature increase to ∼700 °C is unlikely to have been achievedearlier than ∼1 Ma after accretion. This suggests that the 1.7±0.7 MaHf–Wage for Ste. Marguerite could only date parent body processes ifthe H chondrite parent body formed much earlier than the L and LLchondrite parent bodies. However, this would imply that due to 26Alheating significant parts of the H chondrite parent body would havemelted and differentiated.

However, ordinary chondrite parent bodies appear to be wellsampled by breccias but these do not contain any differentiatedmaterial (Scott, 2006), which would be expected if the H chondriteparent body had been partially differentiated. Based on the similarΔ17O values of H chondrites and silicate inclusions from IIE ironmeteorites it was suggested that IIE metal might represent the core ofthe H chondrite parent body (Clayton and Mayeda, 1996). However,similarity in Δ17O values does not require identical parent bodies, as isevident from meteorites that derive from distinct parent bodies buthave identical Δ17O values (e.g., enstatite chondrites, aubrites,terrestrial and lunar rocks; IVA irons and LL chondrites).

These arguments suggest that in Ste. Marguerite metamorphismon the parent body did not result in significant diffusion of radiogenicW out of high-Ca pyroxene. However, elemental diffusion of W intometals clearly occurred. This is evident from the substantially higherW contents of metals from type 4 chondrites compared tometals fromtype 3 chondrites (Rambaldi, 1976; Kong and Ebihara, 1996; Humayunand Campbell, 2002). Evaluating whether this elemental transfer of Wcould have caused resetting of the Hf–W age critically depends onidentifying the original host of the W that diffused into the metalsduring metamorphism of H4 chondrites. H chondrites contain∼180 ppb W (Table 1) and metals in H3 chondrites have ∼300 ppbW, i.e., only ∼30% of the entire W resides in the metal (assuming ametal mass fraction of 20%). In contrast, metals in H4 chondritesSte. Marguerite have ∼800 ppbW, indicating that almost all W (∼90%)is located in the metals. Consequently, during metamorphism ∼60% ofthe entire W in H chondrites must have diffused into the metals. Noneof the major silicate minerals in H chondrites (i.e., olivine, low-Capyroxene, high-Ca pyroxene) is capable of incorporating such highamounts of W (Righter and Shearer, 2003), as a consequence theseminerals cannot be the original host of the W that has been mobilizedduring metamorphism and incorporated into the metals. Non-magnetic fractions from ordinary chondrites have Ir contents thatare too high to reflect equilibrium distribution between metal andsilicates (Palme et al., 1981). This suggests that ordinary chondritescontain a component with appreciable amounts of siderophileelements (including W) that cannot be separated from silicates witha hand-magnet. This phase could be tiny metal grains (that are toosmall to be separated with a hand-magnet) or small refractory in-clusions that reside in the matrix. It is conceivable that W from thesephase(s) became easily mobilized duringmetamorphism and diffused,probably along grain boundaries, into themetals. ThisW probablywasnot radiogenic because the Hf/W ratios in its host metal grains orrefractory inclusions likely were low. For instance, the 182W/184W ina reservoir with 180Hf/184W∼1.7 (i.e., a value typical for CAIs) onlychanges by ∼0.12 ε/Ma, such that the W that diffused into the metals

of H4 chondrites most likely had aW isotope composition very similarto the 182W/184W of the H3 metal. The elemental diffusion of W fromthe matrix into the metals therefore most likely had no measurableeffect on the 182W/184W of the H4 chondrite metal and hence did notaffect the slope of the metal-high-Ca pyroxene isochron.

4.3.2. Hf–W ages for H5 and H6 chondrites: timing of the thermal peakThe Hf–W isochrons for the H5 and H6 chondrites investigated

here are shallower than those of Ste. Marguerite, indicating thatdiffusion of radiogenic W from high Hf/W phases occurred in thesesamples. The interpretation of the Hf–W ages for the H5 and H6chondrites critically depends on whether this diffusion completelyerased any preexisting W isotope heterogeneity. To evaluate this weperformed simulations similar to those for H4 chondrites. We let ahigh-Ca pyroxene-metal system evolve for 1 Ma at low temperatures,such that no diffusional exchange between high-Ca pyroxene andmetal could occur, then instantaneously heated to 950 °C and let thesystem cool at 34 °C/Ma (see below). Using a 4 μm effective grain sizefor the high-Ca pyroxenes the system homogenizes almost as soon ascooling begins, suggesting that any initial W isotopic heterogeneity inRichardton has been largely erased by the thermal metamorphism.This is consistent with the observation that the Hf–W age forRichardton is identical to the Pb–Pb age for its chondrules. Owing tothe slower diffusivity of W compared to Pb, incomplete resettingwould be more pronounced for the Hf–W system than for the Pb–Pbsystem, such that in the case of incomplete resetting the apparent Hf–W age would be older than the apparent Pb–Pb age. The differenceshould increase with a decreasing degree of resetting. However, theHf–W and Pb–Pb ages for Richardton of 4563.0±0.9 and 4562.7±1.7 Ma (Amelin et al., 2005), respectively, are identical, such that theeffects of incomplete resetting seem to be minor or absent and bothages should date cooling from peakmetamorphic temperatures. Giventhat H6 chondrites were heated to higher (or at least similar) peaktemperatures and cooled at a slower rate, this implies that the Hf–Wages for H6 chondrites also reflect cooling below the Hf–W closuretemperature.

The difference in Hf–Wages between H4 and H6 chondrites is only∼8 Ma and much shorter than intervals of ∼74 Ma based on theelevated initial 87Sr/86Sr of phosphates from the H6 chondrite Guareña(Wasserburg et al., 1969) and ∼40 Ma based on the differences in Pb–Pb ages of phosphates from Ste. Marguerite (H4) and Kernouvé (H6)(Göpel et al., 1994). Humayun and Campbell (2002) argued that theabundance and isotope composition of W in metals from ordinarychondrites of type 4, 5, and 6 was set at the same time during theprograde path, whereas other ages were interpreted as postmeta-morphic cooling ages. This conclusion was based on the observationthat metals from type 4, 5, and 6 have constant W/Ir ratios that aredistinct from the variable W/Ir ratios observed for metals from type 3ordinary chondrites (Humayun and Campbell, 2002). According toHumayun and Campbell (2002) this reflects termination of Wdiffusion (both elemental and isotopic) from silicates into metal atmetamorphic conditions characteristic for type 4 chondrites. Theseauthors further argue that the transfer of W from silicates to metal isfacilitated by W reduction in the presence of C and that the major C-bearing phases in ordinary chondrites are decomposed in the earlieststages of metamorphism. As a consequence, reduction of W and itstransfer from silicates into metal would no longer be possible in type 5and 6 ordinary chondrites and Hf–W ages for type 4, 5, and 6chondrites should be identical.

However, the Hf–W ages and closure temperature estimatespresented here reveal that this is not the case and that continueddiffusional exchange ofWoccurred in the type 5 and 6 chondrites. Themuch shorter Hf–W interval compared to the Rb–Sr and Pb–Pbintervals rather highlights the fact that the Hf–W system closed earlyand, hence, dates processes associated with the earliest evolution ofthe H chondrite parent body.

114 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

103

Page 104: Touboul Thesis Completed

4.4. Constraints on the accretion and cooling history

The Hf–W ages for the H chondrites can be used to constrain thecooling history and structure of the H chondrite parent body.Temperature profiles for spherical asteroids heated by energy releasedfrom 26Al decay (Carslaw and Jaeger, 1959) were calculated usingparameters similar to those of Miyamoto et al. (1981) and resultssimilar to those of Trieloff et al. (2003) were obtained (details aregiven in the caption of Fig. 6). The model used here is an over-simplification because it assumes instantaneous accretion and doesneither include the insulating effects of a regolith (Akridge et al., 1998)nor temporal and local variations in physical and thermal parameters(e.g., changes in thermal conductivity due to a decrease in porosity)(Bennett and McSween, 1996). If accretion took place over a timescalesimilar to the 26Al half-life, then a body starts retaining the heatproduced by 26Al decay before reaching its terminal mass and peaktemperatures are reached earlier than estimated when assuminginstantaneous accretion (Merk et al., 2002; Ghosh et al., 2003). A thickinsulating regolith results in a more uniform temperature distribu-tion in the interior, such that peak temperatures can be reached atshallower levels compared to the model used here (Akridge et al.,1998).

Nevertheless, to a first order, the thermal model presented here isuseful for calculating cooling curves for individual samples that areconsistent with the chronological data and for constraining thestructure and thermal history of the H chondrite parent body. Thecalculated temperature profiles for different depth in a spherical bodywith a 100 km radius are shown in Fig. 6 and reveal that allchronological data combined are consistent with the simple thermalmodel used here: the high-temperature (N500 °C) cooling historyconstrained by the Hf–Wages is consistent with the low-temperature(b500 °C) cooling history as constrained earlier (Trieloff et al., 2003).For each of the samples well-defined cooling curves are obtained,suggesting that these samples cooled more or less undisturbed fromtheir peak temperatures to less than ∼100 °C. This provides evidence

that at least parts of the H chondrite parent body stayed intact andwere not disrupted and reassembled.

Cooling rates for H chondrites can be obtained from the slope ofthe cooling curves. This reveals a marked decrease in cooling ratesfrom H4 to H6 chondrites (Fig. 7). For instance, at ∼700 °C the H6chondrites Kernouvé and Estacado cooled at ∼10°/Ma, whereas the H5chondrite Richardton cooled at ∼30 °C/Ma; at ∼400 °C the H6chondrites cooled at ∼5 °C/Ma, Richardton at ∼12 °C/Ma and the H4chondrite Ste. Marguerite at ∼50 °C/Ma. This inverse correlation ofcooling rate with petrologic type is most consistent with an onion-shell structure of the H chondrite parent body. However, the lack ofsuch a correlation between metallographic cooling rates and petro-logic type for some other H chondrites suggests that in some parts ofthe H chondrite parent body cooling from ∼500 to ∼300 °C did nottake place in a bodywith onion-shell structure (Taylor et al., 1987). Thelow temperature cooling history of these parts of the H chondriteparent body therefore is most readily explained by impact-relateddisturbance of parts of the parent body.

Although the Hf–W data do not provide direct age constraints onthe timescales of accretion of the H chondrite parent body – becausenone of the Hf–W ages directly relates to the process of accretion –

they can nevertheless be used to obtain minimum andmaximum agesfor accretion. Accretion could not have started before chondruleformation at 1.7±0.7 Ma, as given by the Hf–Wage for Ste. Marguerite.The end of accretion is less well constrained but the Hf–Wage for H5chondrites of ΔtCAI=5.9±0.9 Ma as well as the 4562.7±0.6 Ma Pb–Pbage for phosphates from Ste. Marguerite (H4) indicate that H4 and H5chondrites had already reached their thermal peak and cooled below∼800 °C (H5) and ∼450 °C (H4), respectively, as early as ∼6 Ma afterCAI formation. It therefore is unlikely that significant accretion oc-curred later than ∼6 Ma. Additional constraints on the duration ofaccretion are provided by the distinct sizes and chemical properties ofchondrules from each chondrite group. In a turbulent solar nebula,chondrules would be efficiently mixed on short timescales (Cuzziet al., 2005), such that a characteristic population of chondrules withits distinct size distribution and chemical composition could only bepreserved, if this chondrule population is accreted into larger bodiessoon after chondrule formation. This was quantified by Alexander(2005), who estimated that material in a 1 AU wide area could bemixedwithin less than∼0.5Ma. Since asteroid feeding zones probablywere smaller, these mixing times become even shorter. If theseestimates are correct, chondrite accretion must have occurred almostinstantaneously after chondrule formation (Alexander, 2005). This

Fig. 6. Cooling curves for H chondrites. Solid lines indicate calculated temperatureprofiles for different depth in a spherical body with a 100 km radius. Numbers indicatedistance in km from the center. We used the following parameters adapted fromMiyamoto et al. (Miyamoto et al., 1981): thermal conductivity K=1.0 Wm−1

K−1; thermal

diffusivity κ=5.0×10−7 m2 s−1; density r=3.2×103 kg m− 3; heat generationA=11.67×(26Al/27Al) W m−3; emissivity h=1.0 m−1. The assumed ambient temperatureis T0=300 K and the initial 26Al/27Al is 4.5×10−6, corresponding to accretion at 2.7 Maafter CAIs. Hf–W ages and closure temperatures are from this study, all other ages arefrom the literature (references are given in Table 2). wr=whole-rock, ch=chondrules,ph=phosphates, fsp=feldspar, mrl=merrilite. Ar–Ar and 244Pu ages are shifted by∼30Ma due to the proposed revision in the 40K decay constant (Trieloff et al., 2001; Minet al., 2003; Trieloff et al., 2003).

Fig. 7. Cooling rate versus petrologic type for H chondrites from this study. For both highand low temperatures there is an inverse correlation of cooling rates with petrologictype, consistent with an onion-shell structure of the H chondrite parent body. The lackof such a correlation betweenmetallographic cooling rates and petrologic type for someH chondrites (Taylor et al., 1987) probably reflects impact-related disturbance of parts ofthe H chondrite parent body during the low-temperature interval (b500 °C).

115T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

104

Page 105: Touboul Thesis Completed

does not necessarily imply that the entire H chondrite parent bodywas accreted at 1.7±0.7 Ma (or soon after). It might also be possiblethat several smaller objects formed at that time and were combinedlater to the H chondrite parent body. However, as argued above, thisprocess was largely complete by ∼6 Ma after CAI formation at thelatest.

There are two endmember interpretations for the Hf–Wage of H6chondrites. First, this age could reflect cooling from the thermal peakof ∼950 °C below Tc=875±75 °C for H6 chondrites. In this case, theHf–W age would almost reflect the time when cooling from thethermal peak started. Alternatively, temperatures inside the Hchondrite parent body remained at their peak for longer but Tcincreased due to grain coarsening and the Hf–W system closedalthough cooling had not yet started. In this case, metamorphictemperatures could have remained above Tc for longer than ∼10 Ma.The thermal modeling shown in Fig. 6 reveals that this might be thecase, at least for Estacado. All chronological data for Estacado plot onthe “40 km” cooling curve but at that depth, cooling from peaktemperatures might have started as late as ∼20–30 Ma, significantlylater than the Hf–W age of ∼10 Ma. Therefore, in Estacado the Hf–Wsystem might have closed before cooling started. Note, however, thatboth peak temperature and Hf–W closure temperature are not knownprecisely enough to clearly distinguish between these two end-member interpretations. Owing to the similarity of peakmetamorphicand Hf–W closure temperature, both processes (i.e., cooling andincrease in Tc due to grain coarsening) were probably important for H6chondrites. In any case, the Hf–W age for H6 chondrites correspondsclosely to the time of the thermal peak.

The Hf–W results appear to be most consistent with some of themodels for the H chondrite parent body presented by Ghosh et al.(2003). In their models 3 and 4, accretion of the H chondrite parentbody started at ∼2 Ma and ended ∼2–3.5 Ma later. Peak metamorphictemperatures in the centre are reached at ∼5–6 Ma and althoughGhosh et al. (2003) do not provide information on how long peaktemperatures are maintained in their models 3 and 4, other thermalmodels indicate that in a body with ∼100 km radius temperatures canbe as high as ∼900 °C at ∼10 Ma (Miyamoto et al., 1981; Bouvier et al.,2007) (see also Fig. 6). Amelin et al. (2005) also found that the Pb–Pbdating results for Richardton are most consistent with the model 3 ofGhosh et al. (2003).

4.5. Hf–W fractionation among chondrite parent materials in the solarnebula

The initial ε182W and initial 182Hf/180Hf obtained from the Hchondrite isochrons can be used to constrain the time-integrated Wisotope evolution, and hence 180Hf/184W ratio, of the H chondriteasteroid in comparison to average carbonaceous chondrites. The Wisotope evolution of carbonaceous chondrites is given by the initialε182W and 182Hf/180Hf determined from internal CAI isochrons[ε182W=−3.30±0.12; 182Hf/180Hf=(1.003±0.045)×10−4 (Burkhardtet al., submitted for publication)] and the precisely defined present-day ε182W=−1.9±0.1 (Kleine et al., 2002; Schoenberg et al., 2002; Yinet al., 2002; Kleine et al., 2004a). These data indicate that carbonac-eous chondrites evolvedwith 180Hf/184W=1.21, in excellent agreementwith 180Hf/184W=1.210±0.075 (2σ) measured for 14 carbonaceouschondrites by isotope dilution (Kleine et al., 2004a). In the W isotopeevolution diagram (Fig. 8), the analyzed H chondrites define a trendthat is different from the evolution path of the carbonaceouschondrites, indicating that the H chondrite asteroid has 180Hf/184W=0.63±0.20. With this low 180Hf/184W the radiogenic ingrowth inH chondrite whole-rocks is small (∼0.35 ε182W in the first ∼10 Ma),consistent with the very limited range in ε182W of the H chondritemetals.

Compositional variations among the chondrite groups are com-monly attributed to different proportions of several distinct compo-

nents present in chondritic meteorites (Palme, 2001). Among these,CAIs [180Hf/184W∼1.7, (Burkhardt et al., submitted for publication)]and metallic Fe (180Hf/184W∼0) have Hf/W ratios different from thoseof average carbonaceous chondrites, such that different mixingproportions of CAIs and metal potentially can cause variations in Hf/W ratios among chondrite groups. Fig. 8 shows that the different Hf/Wratios must have been established in the first ∼2 Ma of the solarsystem. Once more high-precision Hf–W isochrons for other chon-drites (in particular LL chondrites with their probably high Hf/Wratios) are available, it should be possible to date the chemical frac-tionation among chondrites more precisely.

5. Conclusions

Key issues regarding the accretion and early evolution of chondriteparent bodies include the timescales of chondrule formation and peakmetamorphism. The relatively high closure temperature of the Hf–Wsystem in high-Ca pyroxene makes internal Hf–W isochrons involvinghigh-Ca pyroxene a powerful tool for obtaining such age constraints.Although it is more or less well established that most chondrulesformed ∼2–3 Ma later than CAIs (Russell et al., 1996; Kita et al., 2000;Kunihiro et al., 2004; Rudraswami and Goswami, 2007), thisinformation is mainly based on Al–Mg ages for chondrules fromonly a few primitive chondrites. Most chondrites, however, are meta-morphosed to some degree and even mild parent body metamorph-ism might reset the Al–Mg system. For instance, Kita et al. (2000)showed that Al–Mg ages for chondrules from chondrites of petrologictypes higher than 3.4 are already partially reset. In contrast, Hf–Wages are more robust and less easily affected by thermal metamorph-ism and even for chondrites of petrologic type 4 appear to datechondrule formation. The capability of Hf–W chronometry for datingchondrule formation, however, will need to be assessed with moreHf–W ages for type 4 chondrites, by investigating the Hf–Wsystematics of type 3 chondrites and those ordinary chondrite groupsfor which Al–Mg chondrule ages are available.

The closure temperature of the Hf–W system in type 5 and 6chondrites is essentially identical to their peak metamorphictemperatures, such that Hf–W ages for type 5 and 6 chondritescorrespond closely to the time of the thermal peak. As we have shownhere for the H chondrite parent body, such information is essential forconstraining models for the thermal evolution of meteorite parentbodies, which is directly related to their accretion rate and terminalsize. Formation intervals determined by Hf–W chronometry are muchshorter than those obtained from Pb–Pb or Rb–Sr chronology becausethe Hf–W system closed early and hence preserved a record of the

Fig. 8. W isotope evolution diagram for H chondrites. The carbonaceous chondriteevolution line is defined by their present-day ε182W and the initial ε182W of AllendeCAIs. The analyzed H chondrites depart from the evolution line of carbonaceouschondrites, indicating evolution in a reservoir with a Hf/W ratio lower than that ofcarbonaceous chondrites.

116 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

105

Page 106: Touboul Thesis Completed

earliest (i.e., high-temperature) evolution of chondrite parent bodies.Therefore, unlike other chronometers that have been used so far toconstrain the (low-temperature) cooling history of chondrite parentbodies – such as the Pb–Pb, Rb–Sr and Ar–Ar systems – the Hf–Wsystem is less easily disturbed by later events and hence appears mostsuitable for determining the initial thermal structure and meta-morphic history of chondrite parent bodies. Hafnium–tungsten agesfor chondrites should therefore permit the determination and directcomparison of accretion rate, terminal size, and thermal structure ofchondrite parent bodies.

Acknowledgements

We thank Rainer Wieler for the comments and discussion. ConelAlexander and Jeff Grossmann are thanked for providing insights intochondrulemineralogy, especially regarding occurrence and grain sizesof high-Ca pyroxenes. Yuri Amelin is thanked for providing preprintsof his unpublished work. We thank the Museum National d' HistorieNaturelle (Paris), The National Museum of Natural History (Washing-ton DC), The Natural History Museum (London), the Senckenbergmu-seum (Frankfurt), and NASA for generously providing the samples forthis study. Comments and suggestion from an anonymous reviewerand Mario Trieloff significantly improved the manuscript. MarioTrieloff pointed out the possibility that the ages for H6 chondrite mayreflect early closure due to grain coarsening.We thank Rick Carlson forhis comments and the efficient editorial handling of our manuscript.This study was supported by a Marie Curie post-doctoral fellowship toThorsten Kleine.

References

Akridge, G., Benoit, P.H., Sears, D.W.G., 1998. Regolith and megaregolith formation of H-chondrites: thermal constraints on the parent body. Icarus 132, 185–195.

Alexander, C.M.D., 1994. Trace element microdistribution within ordinary chondritechondrules: implications for chondrule formation conditions and precursors.Geochim. Cosmochim. Acta 58, 3451–3467.

Alexander, C.M.D., 2005. From supernovae to planets: the view from meteorites andinterplanetary dust particles. In: Krot, A.N., Scott, E.R.D., Reipurth, B. (Eds.), Chon-drites and the Protoplanetary Disk, Astronomical Society of the Pacific ConferenceSeries 341.

Amelin, Y., 2008. U–Pb ages of angrites. Geochim. Cosmochim. Acta 72, 221–232.Amelin, Y., Irving, A.J., 2007. Sevenmillion years of evolution on the angrite parent body

from Pb-isotopic data, Workshop on the chronology of meteorites and the earlysolar system, LPI Contribution No. 1374, pp. 1320–1321.

Amelin, Y., Ghosh, A., Rotenberg, E., 2005. Unraveling the evolution of chondrite parentasteroids by precise U–Pb dating and thermal modeling. Geochim. Cosmochim.Acta 69, 505–518.

Azough, F., Freer, R., 2000. Iron diffusion in single-crystal diopside. Phys. Chem. Miner.27, 732–740.

Baker, J.A., Bizzarro, M., Wittig, N., Connelly, J., Haack, H., 2005. Early planetesimal meltingfrom an age of 4.5662 Gyr for differentiated meteorites. Nature 436, 1127–1131.

Bennett, M.E., McSween, H.Y., 1996. Revisedmodel calculations for the thermal historiesof ordinary chondrite parent bodies. Meteorit. Planet. Sci. 31, 783–792.

Blinova, A., Amelin, Y., Samson, C., 2007. Constraints on the cooling history of the H-chondrite parent body from phosphate and chondrule Pb-isotopic dates fromEstacado. Meteorit. Planet. Sci. 42, 1337–1350.

Bouvier, A., Blichert-Toft, J., Moynier, F., Vervoort, J.D., Albarède, F., 2007. Pb–Pb datingconstraints on the accretion and cooling history of chondrites. Geochim. Cosmochim.Acta 71, 1583–1604.

Burkhardt, C., Kleine, T., Palme, H., Bourdon, B., Zipfel, J., Friedrich, J., Ebel, D., submittedfor publication. Hf–W isochrons for CAIs and timescales for planetesimal accretionGeochim. Cosmochim. Acta.

Carslaw, H.S., Jaeger, J.C., 1959. Conduction of heat in solids. Oxford University Press.Cherniak, D.J., 1998. Pb diffusion in clinopyroxene. Chem. Geol. 150, 105–117.Cherniak, D.J., Lanford, W., Ryerson, F.J., 1991. Lead diffusion in apatite and zircon using

ion implantation and Rutherford backscattering techniques. Geochim. Cosmochim.Acta 55, 1663–1673.

Clayton, R.N., Mayeda, T.K., 1996. Oxygen isotope studies of achondrites. Geochim.Cosmochim. Acta 60, 1999–2017.

Connelly, J., Amelin, Y., Bizzarro, M., Thrane, K., Baker, J.A., 2008. The Pb–Pb age ofangrite Sah99555 revisited. Lunar Planet. Sci. Conf. XXXIX, 2386.pdf.

Cuzzi, J.N., Ciesla, F.J., Petaev, M.I., Krot, A.N., Scott, E.R.D., Weidenschilling, S.J., 2005.Nebular evolution of thermally processed solids: reconciling models and me-teorites. In: Krot, A.N., Scott, E.R.D., Reipurth, B. (Eds.), Chondrites and theProtoplanetary Disk, Astronomical Society of the Pacific Conference Series 341,pp. 732–773.

Dodd, R.T., 1969. Metamorphism of the ordinary chondrites: a review. Geochim.Cosmochim. Acta 33, 161–203.

Dodd, R.T., 1981. Meteorites: A petrologic-chemical synthesis. Cambridge UniversityPress, New York. 368 pp.

Dodson, M.H., 1973. Closure temperature in cooling geochronological and petrologicalsystems. Contrib. Mineral. Petrol. 40, 259–274.

Ganguly, J., Tirone, M., 2001. Relationship between cooling rate and cooling age of amineral: theory and applications to meteorites. Meteorit. Planet. Sci. 36, 167–175.

Ghosh, A., Weidenschilling, S.J., McSween, H.Y., 2003. Importance of the accretionprocess in asteroid thermal evolution: 6 Hebe as an example. Meteorit. Planet. Sci.38, 711–724.

Göpel, C., Manhès, G., Allègre, C.J., 1994. U–Pb systematics of phosphates from equili-brated ordinary chondrites. Earth Planet. Sci. Lett. 121, 153–171.

Halliday, A.N., 2004. Mixing, volatile loss and compositional change during impact-driven accretion of the Earth. Nature 427, 505–509.

Harper, C.L., Jacobsen, S.B., 1996. Evidence for Hf-182 in the early solar system andconstraints on the timescale of terrestrial accretion and core formation. Geochim.Cosmochim. Acta 60, 1131–1153.

Humayun, M., Campbell, A.J., 2002. The duration of ordinary chondrite metamorphisminferred from tungsten microdistribution in metal. Earth Planet. Sci. Lett. 198,225–243.

Huss, G.R., Rubin, A.E., Grossman, J.N., 2006. Thermal metamorphism in chondrites. In:Lauretta, D.S., McSween, H.Y. (Eds.), Meteorites and the Early Solar System II. TheUniversity of Arizona Press, Tucson, pp. 567–586.

Jacobsen, S.B., 2005. The Hf–W isotopic system and the origin of the Earth and Moon.Ann. Rev. Earth Planet. Sci. 33, 531–570.

Kessel, R., Beckett, J.R., Stolper, E.M., 2007. The thermal history of equilibrated ordinarychondrites and the relationship between textural maturity and temperature.Geochim. Cosmochim. Acta 71, 1855–1881.

Kita, N.T., Nagahara, H., Togashi, S., Morishita, Y., 2000. A short duration of chondruleformation in the solar nebula: evidence from 26Al in Semarkona ferromagnesianchondrules. Geochim. Cosmochim. Acta 64, 3913–3922.

Kleine, T., Münker, C., Mezger, K., Palme, H., 2002. Rapid accretion and early coreformation on asteroids and the terrestrial planets from Hf–W chronometry. Nature418, 952–955.

Kleine, T., Mezger, K., Münker, C., Palme, H., Bischoff, A., 2004a. 182Hf–182W isotopesystematics of chondrites, eucrites, and Martian meteorites: chronology of coreformation and mantle differentiation in Vesta and Mars. Geochim. Cosmochim.Acta 68, 2935–2946.

Kleine, T., Mezger, K., Palme, H., Scherer, E., Münker, C., 2004b. The W isotope evolutionof the bulk silicate Earth: constraints on the timing and mechanisms of coreformation and accretion. Earth Planet. Sci. Lett. 228, 109–123.

Kleine, T., Mezger, K., Palme, H., Scherer, E., Münker, C., 2005a. Early core formation inasteroids and late accretion of chondrite parent bodies: evidence from 182Hf–182Win CAIs, metal-rich chondrites and iron meteorites. Geochim. Cosmochim. Acta 69,5805–5818.

Kleine, T., Mezger, K., Palme, H., Scherer, E., Münker, C., 2005b. TheW isotope compositionof eucrites metal: constraints on the timing and cause of the thermal metamor-phism of basaltic eucrites. Earth Planet. Sci. Lett. 231, 41–52.

Kleine, T., Touboul, M., Halliday, A., Zipfel, J., Palme, H., 2007. Cosmochemical frac-tionation of Hf and W in the solar nebula: evidence fromW isotopes in chondrites.Lunar Planet. Sci. XXXVIII.

Kleine, T., Bourdon, B., Burkhardt, C., Irving, A.J., 2008. Hf–W chronometry of angrites:constraints on the absolute age of CAIs and planetesimals accretion timescales.Lunar Planet. Sci. Conf. XXXIX, 2367.pdf.

Kong, P., Ebihara, M., 1996. Distribution of W and Mo in ordinary chondrites andimplications for nebular and parent body thermal processes. Earth Planet. Sci. Lett.137, 83–93.

Kunihiro, T., Rubin, A.E., McKeegan, K.D., Wasson, J.T., 2004. Initial 26Al/27Al incarbonaceous-chondrite chondrules: too little 26Al to melt asteroids. Geochim.Cosmochim. Acta 68, 2947–2957.

Lindsley, D.H., 1983. Pyroxene thermometry. Am. Mineral. 68, 477–493.Ludwig, K., 1991. ISOPLOT: a plotting and regression program for radiogenic isotope

data; version 2.53, U.S.G.S. Open File Report 91-0445.Lugmair, G.W., Galer, S.J.G., 1992. Age and isotopic relationships among the angrites

Lewis Cliff 86010 and Angra dos Reis. Geochim. Cosmochim. Acta 56, 1673–1694.Markowski, A., Quitté, G., Kleine, T., Halliday, A., Bizzarro, M., Irving, A.J., 2007. Hf–W

chronometry of angrites and the earliest evolution of planetary bodies. EarthPlanet. Sci. Lett. 262, 214–229.

Merk, R., Breuer, D., Spohn, T., 2002. Numerical modeling of 26Al-induced radioactivemelting of asteroids considering accretion. Icarus 159, 183–191.

Min, K.W., Farley, K.A., Renne, P.R., Marti, K., 2003. Single grain (U–Th)/He ages fromphosphates in Acapulco meteorite and implications for thermal history. EarthPlanet. Sci. Lett. 209, 323–336.

Miyamoto, M., Fujii, N., Takeda, H., 1981. Ordinary chondrite parent body: an internalheating model. Proc. Lunar Planet. Sci. Conf. 12B, 1145–1152.

Münker, C., Weyer, S., Scherer, E., Mezger, K., 2001. Separation of high field strengthelements (Nb, Ta, Zr, Hf) and Lu from rock samples for MC-ICPMS measurements.Geochem. Geophys. Geosyst. 2.

Nimmo, F., Agnor, C.B., 2006. Isotopic outcomes of N-body accretion simulations:constraints on equilibration processes during large impacts from Hf/W observa-tions. Earth Planet. Sci. Lett. 243, 26–43.

Nimmo, F., Kleine, T., 2007. How rapidly did Mars accrete? Uncertainties in the Hf–Wtiming of core formation. Icarus 191, 497–504.

Palme, H., 2001. Chemical and isotopic heterogeneity in protosolar matter. Philos. Trans.R. Soc. London, A 359, 2061–2075.

117T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

106

Page 107: Touboul Thesis Completed

Palme, H., Schultz, L., Spettel, B., Weber, H.W., Wänke, H., 1981. The Acapulco meteorite:chemistry, mineralogy and irradiation effects. Geochim. Cosmochim. Acta 45,727–752.

Podosek, F.A., Brannon, J.C., 1991. Chondrite chronology by initial 87Sr/86Sr in phosphates?Meteoritics 26, 145–152.

Rambaldi, E.R., 1976. Trace element of metals from L-group chondrites. Earth Planet. Sci.Lett. 31, 224–238.

Righter, K., Shearer, C.K., 2003. Magmatic fractionation of Hf and W: constraints on thetiming of core formation and differentiation in the Moon and Mars. Geochim.Cosmochim. Acta 67, 2497–2507.

Rudraswami, N.G., Goswami, J.N., 2007. Al-26 in chondrules from unequilibratedL chondrites: onset and duration of chondrule formation in the early solar system.Earth Planet. Sci. Lett. 257, 231–244.

Russell, S.S., Srinivasan, G., Huss, G.R., Wasserburg, G.J., Macpherson, G.J., 1996. Evidencefor widespread 26Al in the solar nebula and constraints on nebula time scales.Science 273, 757–762.

Salters, V.J.M., Hart, S.R., 1991. The mantle sources of ocean ridges, islands and arcs – theHf-isotope connection. Earth Planet. Sci. Lett. 104, 364–380.

Schoenberg, R., Kamber, B.S., Collerson, K.D., Eugster, O., 2002. NewW-isotope evidencefor rapid terrestrial accretion and very early core formation. Geochim. Cosmochim.Acta 66, 3151–3160.

Scott, E.R.D., 2006. Meteoritical and dynamical constraints on the growth mechanismsand formation times of asteroids and Jupiter. Icarus 185, 72–82.

Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomicdistances in halides and chalcogenides. Act. Crystall. A32, 751–767.

Slater-Reynolds, V., McSween, H.Y., 2005. Peak metamorphic temperatures in type 6ordinary chondrites: an evaluation of pyroxene and plagioclase geothermometry.Meteorit. Planet. Sci. 40, 745–754.

Taylor, G.J., Maggiore, P., Scott, E.R.D., Rubin, A.E., Keil, K., 1987. Original structures, andfragmentation and reassembly histories of asteroids: evidence from meteorites.Icarus 69, 1–13.

Trieloff, M., Jessberger, E.K., Fieni, C., 2001. Comment on “Ar-40/Ar-39 age of plagioclasefrom Acapulco meteorite and the problem of systematic errors in cosmochronol-ogy” by Paul R. Renne. Earth Planet. Sci. Lett. 190, 267–269.

Trieloff, M., Jessberger, E.K., Herrwerth, I., Hopp, J., Fiéni, C., Ghelis, M., Bourot-Denise,M., Pellas, P., 2003. Structure and thermal history of the H-chondrite parent asteroidrevealed by thermochronometry. Nature 422, 502–506.

Van Orman, J.A., Grove, T.L., Shimizu, N., 2001. Rare earth element diffusion in diopside:influence of temperature, pressure and ionic radius, and an elastic model fordiffusion in silicates. Contrib. Mineral. Petrol. 141, 687–703.

Van Orman, J.A., Saal, A.E., Bourdon, B., Hauri, E.H., 2006. Diffusive fractionation of U-series radionuclides during mantle melting and shallow level melt-cumulateinteraction. Geochim. Cosmochim. Acta 70, 4797–4812.

Walter, M.J., Thibault, Y., 1995. Partitioning of tungsten and molybdenum betweenmetallic liquid and silicate melt. Science 270, 1186–1189.

Wasserburg, G.J., Papanastassiou, D.A., Sanz, H.G., 1969. Initial strontium for a chondriteand the determination of a metamorphism or formation interval. Earth Planet. Sci.Lett. 7, 33–43.

Watson, H.C., Watson, E.B., 2003. Siderophile trace element diffusion in Fe–Ni alloys.Phys. Earth Planet. Inter. 139, 65–75.

Weyer, S., Münker, C., Rehkamper, M., Mezger, K., 2002. Determination of ultra-low Nb,Ta. Zr and Hf concentrations and the chondritic Zr/Hf and Nb/Ta ratios by isotopedilution analyses with multiple collector ICP-MS. Chem. Geol. 187, 295–313.

Wlotzka, F., 2005. Cr spinel and chromite as petrogenetic indicators in ordinarychondrites: equilibration temperatures of petrologic types 3.7 to 6. Met. Planet. Sci.40, 1673–1702.

Yin, Q.Z., Jacobsen, S.B., Yamashita, K., Blichert-Toft, J., Télouk, P., Albarède, F., 2002. Ashort timescale for terrestrial planet formation from Hf–W chronometry ofmeteorites. Nature 418, 949–952.

118 T. Kleine et al. / Earth and Planetary Science Letters 270 (2008) 106–118

107

Page 108: Touboul Thesis Completed

178

Page 109: Touboul Thesis Completed

Chapitre 4

Chronométrie Hf-W de la Lune:

Contraintes sur l’impact géant et la

cristallisation de l’océan magmatique lunaire

Hf-W chronometry of the Moon: constraints on the giant impact and the

crytallization of the lunar magma ocean

109

Page 110: Touboul Thesis Completed

110

Page 111: Touboul Thesis Completed

Section 4.1

Late formation and prolonged differentiation of the Moon

inferred from W isotopes in lunar metals*

M. Touboul1, T. Kleine1, B. Bourdon1, H. Palme2, R. Wieler1

1 Institute for Isotope Geochemistry and Mineral Resources, ETH Zurich, Clausiusstrasse 25, Switzerland.

2 Institut für Geologie und Mineralogie, Universität zu Köln, Zülpicherstrasse 49b, 50674 Köln, Germany.

*Publié dans Nature, 2007, vol. 450, p. 1206-1209

111

Page 112: Touboul Thesis Completed

112

Page 113: Touboul Thesis Completed

Abstract The Moon probably formed from debris ejected by a giant impact with the early

Earth1. Owing to the high energies involved, the Moon melted to form a magma ocean.

The timescales for formation and solidification of the Moon can be quantified using 182Hf-182W and 146Sm-142Nd chronometry2-4 but these methods yielded contradicting

results. In earlier studies3,5-7, 182W anomalies in lunar rocks were attributed to 182Hf-decay

within the lunar mantle and used to infer that the Moon solidified within the first ~60

Myr of the solar system. The dominant 182W component in most lunar rocks, however,

reflects cosmogenic production mainly by neutron capture of 181Ta during cosmic-ray

exposure of the lunar surface3,7, compromising a reliable interpretation in terms of 182Hf-182W chronometry. Here we present W isotope data for lunar metals that do not contain

any measurable Ta-derived 182W. All metals have identical 182W/184W, indicating that the

lunar magma ocean did not crystallize within the first ~60 Myr of the solar system. This

is no longer inconsistent with Sm-Nd chronometry8-11. Our new data reveal that the lunar

and terrestrial mantles have identical 182W/184W. This, in conjunction with 147Sm-143Nd

ages for the oldest lunar rocks8-11, constrains the age of the Moon and Earth to Myr

after solar system formation. The identical

9010-62+

182W/184W of the lunar and terrestrial mantles

require either that the Moon is mainly derived from terrestrial material or that W isotopes

in the Moon and Earth's mantle equilibrated in the aftermath of the giant impact, as has

been proposed to account for identical O isotope compositions of the Earth and Moon12.

4.1.1 Main text

We obtained W isotope data for metals from 2 KREEP-rich samples, 4 low- and 5

high-Ti mare basalts (Table 4.1.1, Fig. 4.1.1). We processed 4-5 times more material

compared to an earlier study3 and monitored the purity of our metal separates by

determining their Hf/W ratios. These indicate that for the analyses reported here any

possible contamination from silicate and oxide grains has no measurable effect on 182W/184W. Most of the samples investigated here have relatively short exposure times

and require corrections of less than 0.1 ε units (ε=0.01%) for burnout of W isotopes13,14

and only for samples 15556 and 70017 (exposure ages ~220 and ~500 Ma) corrections

were larger (~0.4 and ~0.2 ε units). Details regarding the corrections are given in the

113

Page 114: Touboul Thesis Completed

supplement. All samples analyzed here have identical 182W/184W ratios within ±0.32 ε

units (2 standard deviation (SD)) and agree with previously reported data for metals from

KREEP-rich samples3. These data combined average at ε182W=0.09±0.10 (2 standard

error (SE), n=15; ε182W defined in Table 4.1.1).

Table 4.1.1: Hf-W data for lunar metals from KREEP-rich samples, low-Ti and high-Ti mare basalts

Sample W (ppm)

Hf (ppm) Hf/W ε183W ±2σ ε182Wmeas ±2σ ε182Wcorr ±2σ

KREEP-rich samples 68115 23.9 0.407 0.02 0.00±0.10 0.33±0.14 0.33±0.14

-0.01±0.10 0.18±0.12 0.18±0.12 0.01±0.09 0.18±0.12 0.18±0.12 0.00±0.11 0.14±0.15 0.14±0.15

mean (±2σ SD) 0.00±0.02 0.21±0.17 0.21±0.17 68815 27.5 0.725 0.03 0.04±0.11 0.02±0.18 0.02±0.18

-0.02±0.10 0.26±0.18 0.26±0.18 0.03±0.13 0.25±0.15 0.25±0.15 -0.01±0.13 0.22±0.17 0.22±0.17

mean (±2σ SD) 0.01±0.06 0.18±0.23 0.19±0.23 weighted average* (±2σ,n=6) 0.00±0.03 0.11±0.21 0.11±0.21

Low-Ti mare basalts

12004 48.7 0.539 0.11 0.04±0.21 0.00±0.36 0.05±0.36 15058 11.3 0.113 0.01 -0.10±0.15 -0.11±0.19 0.01±0.20 15499 7.64 0.911 0.12 0.03±0.22 0.06±0.31 0.16±0.31 15556 21.6 0.477 0.02 0.02±0.21 -0.14±0.29 0.30±0.36

weighted average (±2σ,n=4) -0.02±0.09 -0.07±0.13 0.09±0.14 High-Ti mare basalts

70017 11.6 2.18 0.19 -0.13±0.14 0.20±0.14 0.38±0.16 70035 10.1 5.09 0.50 0.08±0.13 0.05±0.18 0.14±0.18 74255 1.32 1.30 1.0 -0.11±0.14 0.09±0.17 0.11±0.16 74275 8.92 2.24 0.25 -0.16±0.18 -0.25±0.25 -0.24±0.25 75035 4.65 25.1 5.4 0.07±0.15 0.22±0.18 0.18±0.18

weighted average (±2σ,n=5) -0.04±0.14 0.11±0.19 0.16±0.24

Bulk lunar mantle* (2 SE, n=15) 0.00±0.02 0.02±0.09 0.09±0.10

ε18iW={(18iW/184W)sample/(18iW/184W)standard-1} × 104. Replicates for KREEP-rich samples are repeated measurements of the same solution. Mean values are weighted averages calculated using Isoplot (n=number of samples). * Averages are calculated including data for KREEP-rich samples from ref. 3.

114

Page 115: Touboul Thesis Completed

Figure 4.1.1: ε182W of lunar metals analyzed in this study (closed symbols) compared to previously reported data from Kleine et al.5 (open symbols) and Lee et al.3 (crossed symbols). Some of the previous data (shown with black dots inside the symbols) were corrected for cosmogenic 182W (see text for details). Error bars indicate 2ε external reproducibilities. The hatched area indicates the average ε182W=0.09±0.10 (2 SE, n=15) of lunar metals from this study combined with previously reported data for metals from KREEP-rich samples5 (KREEP stands for enrichment in potassium (K) and Rare Earth Elements). The dashed lines indicate the 2 standard deviation of ±0.32 ε182W of these data.

115

Page 116: Touboul Thesis Completed

In contrast to earlier studies3,5,6, we do not find 182W/184W variations within the

lunar mantle. Lee et al.5 reported ε182W~1.4 for low-Ti mare basalt 15555 but the

exposure age of this sample combined with its Sm isotopic composition and Ta/W ratio

indicates that this anomaly might entirely be due to cosmogenic 182W. Kleine et al.3

reported elevated ε182W~2 for a magnetic separate from high-Ti mare basalt 79155 but

we determined Hf/W=7.5 for an aliquot from the same magnetic separate, most likely

indicating the presence of some ilmenite and hence cosmogenic 182W in this separate. The

calculated cosmogenic 182W component is ~1.7 ε units, hence the elevated 182W/184W

reported for 79155 presumably is mainly due to cosmogenic 182W. Enhanced 182W/184W

in high-Ti mare basalts 72155 (ref. 3) and 75075 (ref. 5) also are not well-resolved from

ε182W=0.09±0.10. Details regarding the quantification of cosmogenic 182W are given in

the supplement. We conclude that in light of the identical 182W/184W we obtained for pure

metal separates from 9 mare basalts (average ε182W=0.12±0.12, 2 SE, n=9), elevated 182W/184W ratios in mare basalts reported earlier likely reflect production of cosmogenic 182W. This production is difficult to quantify, demonstrating that the purely radiogenic 182W/184W is difficult to determine precisely if cosmogenic 182W is present. We conclude

that there are no 182Hf-induced W isotope variations among KREEP and the mare basalt

sources. Given that low- and high-Ti mare basalts sample the majority of cumulates from

the lunar magma ocean (LMO)15 and that KREEP contains a significant part of the

incompatible element inventory of the lunar mantle16, the average ε182W=0.09±0.10 (2

SE) of these lithologies most likely represents the average ε182W of the lunar mantle.

Although some ferroan anorthosites have elevated ε182W up to ~3 (ref. 6), these data are

relatively imprecise and their weighted average ε182W=1.9±1.7 (2σ) is not resolvable

from ε182W=0.09±0.10. Moreover, the analyzed anorthosites might also contain

cosmogenic 182W.

116

Page 117: Touboul Thesis Completed

Figure 4.1.2: ε182W vs. source 180Hf/184W. Assumed Hf/W ratios are 10±10 for KREEP, 26.5±1.1 for the low-Ti mare basalt source, and >40 for the high-Ti mare basalt source (for details see supplement). ε182W values shown for the low-Ti and high-Ti mare basalt sources are weighted averages of the data obtained in this study, the average ε182W of KREEP was calculated including data from Kleine et al.3 Error bars indicate the standard errors (2σ) of these data. Reference isochrons corresponding to 30, 40, 50, 60 and 70 Myr after the start of the solar system are shown. It is assumed that the Hf/W ratio of the bulk lunar mantle is similar to that of the low-Ti mare basalt source. The latter consists mainly of olivine and orthopyroxene, which are not capable of fractionating Hf and W18. Τhe identical ε182W values for KREEP and the low- and high-Ti mare basalt sources require that equilibration of W isotopes within the LMO occurred later than ~60 Myr after the start of the solar system. Note that this conclusion depends on neither the choice of Hf/W ratios of the sources nor on using the Hf/W ratio of the low-Ti mare basalt source as representative of the Hf/W ratio of the bulk lunar mantle.

These new constraints have far reaching implications regarding the lifetime of the

LMO as well as the age and formation of the Moon. The homogeneous 182W/184W ratios

of all lunar samples in spite of strongly fractionated Hf/W ratios in their source areas17,18

indicate that the last equilibration of W isotopes within the LMO occurred later than ~60

Myr (here, Myr refers to time after formation of the first solids in the solar system) (Fig.

4.1.2). Isotopic equilibration among the products of the LMO is possible up to a critical

crystal fraction of ~60% until which convection prevents crystal settling19. Although our

new results only provide the earliest time (>~60 Myr) for ~60% LMO crystallization, this

new age is no longer in conflict with other constraints regarding the lifespan of the LMO.

117

Page 118: Touboul Thesis Completed

The rapid crystallization of the LMO required by the earlier W isotope data3,5,6 implied

that some 147Sm-143Nd ages for ferroan anorthosites8-11 and the 146Sm-142Nd model age2,4

of the lunar mantle reflect post-LMO events3,17. With the revised Hf-W time constraint

this is no longer required and the aforementioned Sm-Nd ages could possibly date

processes associated with the LMO. This suggests that the LMO could have solidified to

~60% as late as ~215 Myr, as given by its 146Sm-142Nd model age2,4.

Our data do not only constrain the lifespan of the LMO but also the timing of the

giant impact. This utilizes the virtually identical 182W/184W of the bulk silicate Moon and

Earth in conjunction with their different Hf/W ratios. The latter can be inferred from U/W

and Th/W ratios because U, Th and W have similar incompatibilities during mantle

melting20,21. The lunar mantle has U/W=1.93±0.08 (2σ)21, distinctly higher than

U/W=1.3±0.4 (2σ) for the bulk silicate Earth [obtained from Th/W=5.5±1.6 (2σ)20 and

Th/U=4.2±0.1 (2σ)22]. If refractory lithophile elements occur in chondritic proportions in

planetary mantles, then Hf/U=13.7 for chondrites23 can be used to calculate

Hf/W=26.5±1.1 (2σ) and Hf/W=18.0±5.2 (2σ) for the bulk silicate Moon and Earth. It is

conceivable however that the Moon and Earth have non-chondritic ratios of refractory

lithophile elements (e.g., Th/U22,23) and in this case the Hf/W ratios of the bulk silicate

Moon and Earth might be different from those calculated above. However, this would

only have a small effect on the calculated ages (Fig. 4.1.3). For instance, using a ~15%

higher Hf/U ratio only changes the ages by ~2 Myr.

A first-order estimate for core formation in the Moon (or the impactor) is obtained

using a two-stage model. This assumes a chondritic initial 182W/184W of the Moon and

results in an age of ~37 Myr. This model implies that the bulk silicate Earth and Moon

started off with slightly different 182W/184W (~0.5 ε units difference at most; see

supplemental information Fig. 4.1.4) and fortuitously evolved to identical present-day 182W/184W. However, as the Moon predominantly consists of mantle material with high

Hf/W and hence most likely radiogenic 182W/184W, this two-stage model is not valid. The

initial 182W/184W of the Moon most likely was higher than chondritic, resulting in an age

younger than ~37 Myr and implying that the bulk silicate Moon and Earth must have had

indistinguishable initial 182W/184W (supplemental information Fig. 4.1.4).

118

Page 119: Touboul Thesis Completed

A more reliable age constraint for core formation is obtained from the identical 182W/184W of the bulk silicate Moon and Earth in conjunction with their distinct Hf/W and

by assuming identical initial 182W/184W for the bulk silicate Moon and Earth. As shown in

Fig. 4.1.3, core formation in the Moon must have occurred later than ~50 Myr, otherwise

the lunar mantle would have a 182W excess relative to the terrestrial mantle. This new

constraint provides the earliest time the giant impact could have occurred because the

segregation of the lunar core probably occurred briefly after formation of the Moon. This

also holds true if the Hf-W systematics of the Moon would largely reflect core formation in

the impactor because this must predate the giant impact.

Owing to the identical 182W/184W of the bulk silicate Moon and Earth, our data

only provide the earliest time the Moon could have formed but the latest time is given by

the age of the oldest lunar rocks. Combined with the 4.456±0.040 Ga 147Sm-143Nd age of

the lunar crust10 we obtain an age of Myr for the formation of the Moon (Fig. 3).

Given that Earth's accretion cannot have terminated before the giant impact, this also

provides an age for the accretion of Earth. This is inconsistent with termination of Earth's

accretion at ~30 Myr (e.g., ref. 24) and also difficult to reconcile with the <30 Myr

9010-62+

146Sm-142Nd model age for differentiation of Earth's mantle25. Our new age constraint,

however, is consistent with most U-Pb model ages for the Earth26.

119

Page 120: Touboul Thesis Completed

Figure 4.1.3: ε182W difference between bulk silicate Moon and Earth as a function of time of core formation. The grey shaded area indicates ε182W=0.09±0.10 (2 SE) for the lunar mantle as determined in this study. Assuming that the lunar and terrestrial mantles had identical initial 182W/184W the time at which these two reservoirs separated can be calculated using

( ) ( ) ( )[ ]{ }

⎪⎪⎪⎪

⎪⎪⎪⎪

⎪⎪⎪⎪

⎪⎪⎪⎪

⎟⎠⎞⎜

⎝⎛

−××⎟⎠⎞⎜

⎝⎛

×

×=

0Hf180Hf182

BSMWUBSEWUCHURUHf1.14

BSEW184W182

4-10W182Δε

lnλ1t , where

CHUR= chondritic uniform reservoir, BSE=bulk silicate Earth, BSM=bulk silicate Moon, and

(182Hf/180Hf)0=1.07×10-4 (ref. 14). ⎭⎬⎫

⎩⎨⎧

−⎟⎠⎞

⎜⎝⎛

⎟⎠⎞

⎜⎝⎛= 1

WHf

WHff

BSEBSMHf/W . Note that the lower uncertainty on

the age is calculated by propagating the uncertainties on Δε182W and 180Hf/184W and hence does not coincide with the intersection of the f=0.1 curve with the grey shaded bar.

The identical 182W/184W ratios of the lunar and terrestrial mantles provide a key

constraint regarding the formation of the Moon. The 182W/184W of any planetary mantle

reflects the timescale of accretion and core formation, the degree of W depletion, and the

extent of re-equilibration of W isotopes during core formation24,27,28. Hence, it is unlikely

that the mantles of the proto-Earth and the impactor evolved to identical 182W/184W ratios.

Successful simulations of the giant impact predict that ~80% of the Moon is derived from

120

Page 121: Touboul Thesis Completed

impactor material1, such that small pre-existing W isotopic differences between proto-

Earth and impactor should be reflected in the composition of the Moon. This is not

observed however in the Hf-W systematics. Similarly, the identical O isotopic

compositions of the Earth and Moon in spite of widespread O isotopic heterogeneity

among inner solar system objects are unexpected. These have been interpreted to reflect

accretion of the Earth and Moon from a similar mix of components formed at the same

heliocentric distance29. As explained above, such a scenario cannot account for the

identical W isotope compositions of the lunar and terrestrial mantles. Hence, either the

Moon almost entirely derived from Earth's mantle (which is contrary to results from

numerical simulations of the giant impact) or lunar and terrestrial materials equilibrated

in the aftermath of the giant impact. Diffusive exchange between the silicate vapour

atmosphere of the proto-Earth and the lunar magma disk might be possible for elements

that became vaporized during formation of the lunar magma disk. It has been shown that

this is possible for O isotopes12 but the efficiency to which W became vaporized and

hence could have been equilibrated isotopically remains to be investigated.

4.1.2 Methods summary

Samples were crushed in an agate mortar and magnetic fractions were obtained

using a hand-magnet. The magnetic separates were purified by repeated grinding,

magnetic separation and ultrasonication in distilled ethanol. After dissolution a ~5% was

spiked with a mixed 180Hf-183W tracer for determination of Hf and W concentrations that

are used to monitor the purity of the metal separates. From the remaining ~95% W was

extracted using anion exchange techniques, slightly modified from Kleine et al.3,30. All

measurements were performed using a Nu Plasma MC-ICPMS at ETH Zürich. Tungsten

isotope measurements were normalized to 186W/183W=1.9859 and 186W/184W=0.92767

using the exponential law. Results obtained using these two normalization procedures

agree for all samples. Tungsten isotope compositions of the samples were determined

relative to the 182W/184W obtained for two bracketing measurements of the ALFA

AESEAR W standard solution. Isobaric Os interferences on masses 186 and 184 were

monitored by measuring 188Os but corrections were insignificant for all samples (< 0.01

p.p.m.). The accuracy of the measurements was monitored using the 183W/184W ratio and

121

Page 122: Touboul Thesis Completed

all samples have 183W/184W identical to the standard value to within ±0.2 ε units. The

external reproducibility of the W isotope measurements was evaluated by repeated

measurements of metals from the two KREEP-rich samples and is ±0.17 (2σ, n=4) and

±0.23 (2σ, n=4) (Table 4.1.1). For the low-Ti mare basalts, the reproducibility is ~0.5 ε

units or better, and for the high-Ti mare basalts, it is 0.3-0.4 ε units.

4.1.3 Acknowledgements

We are grateful to the Curation and Analysis Planning Team for Extraterrestrial

Materials (CAPTEM), NASA curatorial staff, and G. Lofgren for supplying the Apollo

lunar samples. We thank L. Borg and A. Brandon for their thoughtful reviews and F.

Nimmo and J. Van Orman for discussions. This research was supported by a EU Marie

Curie postdoctoral fellowship to T. Kleine.

Supplementary Information accompanies the paper on www.nature.com/nature and after

references in section 4.1.5

4.1.4 References

1. Canup, R. M. & Asphaug, E. Origin of the Moon in a giant impact near the end of

the Earth's formation. Nature 412, 708-712 (2001).

2. Nyquist, L. E. et al. Sm-146-Nd-142 Formation Interval for the Lunar Mantle.

Geochim. Cosmochim. Acta 59, 2817-2837 (1995).

3. Kleine, T., Palme, H., Mezger, K. & Halliday, A. N. Hf-W chronometry of lunar

metals and the age and early differentiation of the Moon. Science 310, 1671-1674

(2005).

4. Rankenburg, K., Brandon, A. D. & Neal, C. R. Neodymium isotope evidence for a

chondritic composition of the Moon. Science 312, 1369-1372 (2006).

122

Page 123: Touboul Thesis Completed

5. Lee, D. C., Halliday, A. N., Leya, I., Wieler, R. & Wiechert, U. Cosmogenic

tungsten and the origin and earliest differentiation of the Moon. Earth Planet. Sci.

Lett. 198, 267-274 (2002).

6. Lee, D. C., Halliday, A. N., Snyder, G. A. & Taylor, L. A. Age and origin of the

moon. Science 278, 1098-1103 (1997).

7. Leya, I., Wieler, R. & Halliday, A. N. Cosmic-ray production of tungsten isotopes

in lunar samples and meteorites and its implications for Hf-W cosmochemistry.

Earth Planet. Sci. Lett. 175, 1-12 (2000).

8. Borg, L. E. et al. Isotopic studies of ferroan anorthosite 62236: A young lunar

crustal rock from a light rare-earth-element-depleted source. Geochim. Cosmochim.

Acta 63, 2679-2691 (1999).

9. Carlson, R. W. & Lugmair, G. W. The age of ferroan anorthosite 60025: oldest

crust on a young Moon? Earth Planet. Sci. Lett. 90, 119-130 (1988).

10. Norman, M. D., Borg, L. E., Nyquist, L. E. & Bogard, D. D. Chronology,

geochemistry, and petrology of a ferroan noritic anorthosite clast from Descartes

breccia 67215: Clues to the age, origin, structure, and impact history of the lunar

crust. Meteorit. Planet. Sci. 38, 645-661 (2003).

11. Nyquist, L. et al. Feldspathic clasts in Yamato-86032: Remnants of the lunar crust

with implications for its formation and impact history. Geochim. Cosmochim. Acta

70, 5990-6015 (2006).

12. Palhevan, K. & Stevenson, D. J. Possible Origin of the geochemical similarity of

the Earth and Moon. Earth Planet. Sci. Lett., in press (2007).

13. Leya, I., Wieler, R. & Halliday, A. N. The influence of cosmic-ray production on

extinct nuclide systems. Geochim. Cosmochim. Acta 67, 529-541 (2003).

14. Kleine, T., Mezger, K., Palme, H., Scherer, E. & Münker, C. Early core formation

in asteroids and late accretion of chondrite parent bodies: Evidence from 182Hf-182W

in CAIs, metal-rich chondrites and iron meteorites. Geochim. Cosmochim. Acta 69,

5805-5818 (2005).

123

Page 124: Touboul Thesis Completed

15. Shearer, C. K. & Papike, J. J. Magmatic evolution of the Moon. Am. Mineral. 84,

1469-1494 (1999).

16. Palme, H. & Wänke, H. A unified trace-element model for the evolution of the

lunar crust and mantle. Proc. Lunar Sci. Conf. 6th, 1179-1202 (1975).

17. Shearer, C. K. & Newsom, H. E. W-Hf isotope abundances and the early origin and

evolution of the Earth-Moon system. Geochim. Cosmochim. Acta 64, 3599-3613

(2000).

18. Righter, K. & Shearer, C. K. Magmatic fractionation of Hf and W: Constraints on

the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497-2507 (2003).

19. Solomatov, V. S. in Origin of the Earth and Moon (eds. Canup, R. M. & Righter,

K.) 323-338 (Lunar and Planetary Institute, Houston, 2000).

20. Newsom, H. E. et al. The depletion of W in the bulk silicate Earth: constraints on

core formation. Geochim. Cosmochim. Acta 60, 1155-1169 (1996).

21. Palme, H. & Rammensee, W. The significance of W in planetary differentiation

processes: Evidence from new data on eucrites. Proc. 12th Lunar Planet. Sci. Conf.,

949-964 (1981).

22. Allegre, C. J., Dupre, B. & Lewin, E. Thorium Uranium Ratio of the Earth. Chem.

Geol. 56, 219-227 (1986).

23. Rocholl, A. & Jochum, K. P. Th, U and other trace elements in carbonaceous

chondrites: Implications for the terrestrial and solar-system Th/U ratios. Earth

Planet. Sci. Lett. 117, 265-278 (1993).

24. Jacobsen, S. B. The Hf-W isotopic system and the origin of the Earth and Moon.

Ann. Rev. Earth Planet. Sci. 33, 531-570 (2005).

25. Boyet, M. & Carlson, R. W. 142Nd evidence for early (>4.53 Ga) global

differentiation of the silicate Earth. Science 309, 576-581 (2005).

26. Allegre, C. J., Manhes, G. & Gopel, C. The age of the Earth. Geochim. Cosmochim.

Acta 59, 1445-1456 (1995).

124

Page 125: Touboul Thesis Completed

27. Kleine, T., Mezger, K., Palme, H., Scherer, E. & Münker, C. The W isotope

evolution of the bulk silicate Earth: constraints on the timing and mechanisms of

core formation and accretion. Earth Planet. Sci. Lett. 228, 109-123 (2004).

28. Nimmo, F. & Agnor, C. B. Isotopic outcomes of N-body accretion simulations:

Constraints on equilibration processes during large impacts from Hf/W

observations. Earth Planet. Sci. Lett. 243, 26-43 (2006).

29. Wiechert, U. et al. Oxygen isotopes and the moon-forming giant impact. Science

294, 345-348 (2001).

30. Kleine, T., Mezger, K., Palme, H., Scherer, E. & Münker, C. The W isotope

composition of eucrites metal: Constraints on the timing and cause of the thermal

metamorphism of basaltic eucrites. Earth Planet. Sci. Lett. 231, 41-52 (2005).

125

Page 126: Touboul Thesis Completed

4.1.5 Supplementary information 4.1.5.1 Lunar metals

Metal grains occur in almost all lunar rocks1. Relatively high Ni and Ir contents in

metals from highland samples indicate that these are meteoritic metals added to the lunar

surface by impacts2. Unlike meteoritic metals, the metals from KREEP–rich highland

rocks have extremely high W contents [24-28 ppm, Table 4.1.1 and ref. 2], reflecting the

strong enrichment of KREEP in W. The high W/Ir ratios in lunar metals confine the

contribution of meteoritic W to less than ~1% (ref. 2,3), precluding any measurable effect

on the 182W/184W of KREEP. Partitioning of W from KREEP into these metals occurred

by impact–induced melting and brecciation during redistribution of KREEP at the lunar

surface2. Metals in mare basalts consist of native Fe and probably formed by

crystallization from a silicate melt under reducing conditions1. The composition of these

metals (namely low Ni contents) distinguishes them from meteoritic metal added to the

lunar surface by impacts1,3.

4.1.5.2 Analytical methods Mare basalts (~4 g) and KREEP-rich samples (~1 g) were crushed in an agate

mortar and separated into several size fractions using nylon sieves. Magnetic fractions

from each of them were separated using a hand-magnet and further purified by repeated

grinding, magnetic separation and ultrasonication in distilled ethanol. The purity of the

metal separates was checked under the binocular and the final separates (1-6 mg) in most

cases did not contain any visible silicate or oxide grains. In some cases ilmenite grains

were still present and could not be removed without loosing too many of the metal grains.

The presence of these impurities in some of the magnetic separates however has no

measurable effect on 182W/184W (see below). The analytical procedure has slightly been

modified from Kleine et al.4-6. The metal separates were dissolved in 6 M HCl-0.06 M

HF in Savillex vials at 120 °C on a hotplate. In this acid mixture most of the potentially

present silicate and oxide impurities should not dissolve. After metal dissolution, a ~5%

aliquot of the sample was spiked with a mixed 180Hf-183W tracer for determination of Hf

and W concentrations. The remaining ~95% solution was dried and redissolved in HNO3-

H2O2, dried and redissolved in 1M HF-0.1 M HNO3. This solution was loaded onto pre-

126

Page 127: Touboul Thesis Completed

cleaned anion exchange columns (BioRad AG1X8, 200-400 mesh) and the sample matrix

eluted using 1 M HF-0.1 M HNO3. Tungsten was then collected using 6 M HCl-1 M HF.

Total procedural blanks were ~20 pg for W isotope composition measurements and

therefore negligible. Total procedural blanks for the concentration measurements are ~2

pg W and ~1 pg Hf, resulting in blank corrections of <4% on the measured Hf

concentrations.

All measurements were performed using a Nu Plasma MC-ICPMS at ETH

Zurich. Tungsten isotope measurements were normalized to 186W/183W=1.9859 and 186W/184W=0.92767 using the exponential law. Results obtained using these two

normalization procedures agree for all samples. Tungsten isotope compositions of the

samples were determined relative to the 182W/184W obtained for two bracketing

measurements of the ALFA AESEAR W standard solution. Isobaric Os interferences on

masses 186 and 184 were monitored by measuring 188Os but corrections were

insignificant for all samples (< 0.01 p.p.m.). The accuracy of the measurements was

monitored using the 183W/184W ratio and all samples have 183W/184W identical to the

standard value to within ±0.2 ε units. The external reproducibility of the W isotope

measurements was evaluated by repeated measurements of metals from the two KREEP-

rich samples and is ±0.17 (2σ, n=4) and ±0.23 (2σ, n=4) (Table 4.1.1). This is similar to

the reproducibility of the W standard measured at the same beam intensity as was

obtained for the KREEP metals. For metals from low- and high-Ti mare basalts less W

was available for analyses and the reproducibility was estimated by repeated

measurement of the W standard using identical beam intensities than those obtained for

the sample runs. For the low-Ti mare basalts, the reproducibility is ~0.5 ε units or better,

and for the high-Ti mare basalts, it is 0.3-0.4 ε units.

4.1.5.3 Cosmogenic effects on W isotope ratios in lunar samples

4.1.5.3.1 Correction for W isotope data obtained in this study

Elevated 182W/184W of lunar whole-rocks largely reflect production of

cosmogenic 182W by neutron capture of 181Ta and subsequent decay of 182Ta to 182W (ref.

6-8). Metals do not contain any Ta and their 182W/184W should not be affected by

cosmogenic 182W production6 but contamination with silicate or oxide grains may result

127

Page 128: Touboul Thesis Completed

in elevated measured 182W/184W. The purity of the metal separates can be monitored by

determining their Hf/W ratios because contamination with silicates or oxides would result

in high concentrations of lithophile elements. The Hf/W ratios of all metal separates from

KREEP-rich samples and low-Ti mare basalts investigated here are below ~0.1. Metal

separates from the high-Ti mare basalts have Hf/W ratios between 0.2 and 5.4. Using

typical Ta/Hf ratios for the whole-rocks (~0.14 for low-Ti mare basalts, ~0.16 for high-Ti

mare basalts, ~0.11 for KREEP9) the Ta/W ratios of the metal separates can be calculated

because the dominant contaminant in the metal separates probably is ilmenite which has

similar Ta/Hf than the whole-rocks (DHf/DTa ~1 in ilmenite, ref. 10). All calculated Ta/W

ratios for metal separates from the KREEP-rich samples and low-Ti mare basalts

investigated here are lower than ~0.02, corresponding to a maximum 181Ta-derived 182W

component of ~0.0003 εW (Supplementary Table 4.1.2). The calculated Ta/W ratios for

metal separates from the high-Ti mare basalts investigated here are higher and range from

0.03 to 0.86 but the 181Ta-derived 182W component in these samples is lower than ~0.02

ε182W for all samples expect for sample 75035, where this component is ~0.1 ε182W.

Hence, contamination of the metal separates investigated here with silicates/oxides has no

measurable effect on the 182W/184W ratios.

The W isotope ratios of metal can also be modified by burnout of W isotopes, as

has been observed for iron meteorites with long exposure times11-13. For this reason, most

of the samples selected for this study have exposure ages younger than ~135 Ma, such

that the downward shift of 182W/184W in the metals of these samples is <0.1 ε units. Only

two of the samples investigated here have older exposure ages and W burnout could have

lowered their 182W/184W by ~0.2 (sample 70017) and ~0.4 (sample 15556) ε units

(Supplementary Table 4.1.2). The combined effects of burnout of W isotopes and 182W

production by neutron capture of 181Ta for all samples analyzed here are summarized in

the Supplementary Table 4.1.2. The corrections for most samples are much smaller than

~0.1 ε182W and hence well below our analytical precision. For the metal separates from

samples 70017 and 15556 (i.e., the two samples with the oldest exposure ages among the

investigated samples) the corrections are similar to or slightly higher than the precision of

our measurements. The uncertainties of these corrections are difficult to assess and a 50%

128

Page 129: Touboul Thesis Completed

uncertainty on the correction is conservatively assumed here. This is a factor of 2 higher

than given in Leya et al.7,14.

Table 4.1.2: Cosmogenic effects on ε182W

Sample Exposure age (Ma) Ta/W Δε182WGCR, burnout Δε182WGCR,capture Δε182WGCR

KREEP 68115 2 0.002 -0.002 0.00001 -0.002±0.001 68815 2 0.003 -0.002 0.00002 -0.002±0.001

Low Ti basalts

12004 60 0.015 -0.05 0.001 -0.05±0.025 15058 135 0.001 -0.12 0.0002 -0.12±0.06 15499 114 0.017 -0.10 0.003 -0.10±0.05 15556 500 0.003 -0.44 0.003 -0.44±0.22

High Ti basalts

70017 220 0.03 -0.19 0.01 -0.18±0.09 70035 120 0.08 -0.11 0.02 -0.1±0.05 74255 25 0.04 -0.02 0.01 -0.01±0.005 74275 32 0.16 -0.03 0.01 -0.02±0.01 75035 72 0.86 -0.06 0.11 0.04±0.02

Δε182WGCR, burnout is the effect on 182W/184W caused by burnout of W isotopes; Δε182WGCR,capture is the cosmogenic production of 182W by neutron-capture of 181Ta; Δε182WGCR is the overall cosmogenic effect, Δε182WGCR=Δε182WGCR, burnout+Δε182WGCR,capture

4.1.5.3.2 Correction of W isotope data obtained in earlier studies

We determined Hf/W~7.5 for an aliquot of the magnetic separate from sample

79155 analyzed earlier by Kleine et al.5 for its W isotope composition. Using Ta/Hf~0.18

determined for the 79155 whole-rock results in Ta/W~1.4. Using this Ta/W ratio, an

exposure age of 575 Ma and correction equations14, a cosmogenic 182W component of

~0.85 ε units is calculated for the magnetic separate from sample 79155. The 79155

whole-rock has ε182W~38 and Ta/W~20 (ref. 6) but using the correction equations from

Leya et al.14 only results in cosmogenic 182W of ~19 ε units. This could indicate that this

correction procedure might underestimate the production of cosmogenic 182W by a factor

of ~2, such that the cosmogenic 182W in the magnetic separate of 79155 might be as high

as ~1.7. This is similar to ε182W=2.3±0.9 measured for the 79155 magnetic separate,

indicating that the elevated ε182W of 79155 most likely reflects contamination with

129

Page 130: Touboul Thesis Completed

silicate and oxide grains that carry cosmogenic 182W. It is plausible that the magnetic

separate from 72155 analyzed earlier by Kleine et al.5 also contains some ilmenite and

hence cosmogenic 182W. However, the situation for this sample remains unclear because

the 182W/184W of its whole-rock is not elevated compared to its magnetic separate, no

Hf/W ratio is available for the latter, and the exposure history of 72155 is not well

constrained. Lee et al.8 reported ε182W=4.7±3.0 for high-Ti mare basalt 75075, based on a

regression of Ta/W and 182W/184W obtained for one pyroxene and one feldspar separate.

The Ta contents and 182W/184W ratios for the feldspar separate however were not

obtained for the same aliquot and the W contents for the two feldspar separates differ by

a factor of almost 4. Hence, the Ta/W ratio of the feldspar separate also analyzed for W

isotopes is not well constrained, rendering it unclear whether this sample has radiogenic

ε182W that is well resolved from ε182W=0.09±0.10.

Lee et al.8 argued that the average ε182W=1.4±0.4 of low-Ti mare basalt 15555

reflects 182Hf decay and that this sample contains no significant cosmogenic 182W because

it exhibits no correlation between the 182W/184W of mineral separates having variable

Ta/W. However, using the 149Sm/150Sm ratio of sample 15555 (ref. 15,16), its Ta/W ratio

of 4.14 (ref. 8) and correction equations14 reveals that this sample may contain a

cosmogenic component of ~0.7 ε182W. Although the presence of cosmogenic 182W should

result in systematic variations of 182W/184W with Ta/W, the entire spread in Ta/W ratios

among the investigated mineral fractions of 15555 is less than ~3, such that the

anticipated cosmogenic 182W/184W variations are similar to the analytical uncertainty of

the W isotope measurements. After correction for cosmogenic 182W, the 15555 whole-

rock (measured εW=1.13±0.36) has ε182W=0.45±0.50 (assuming a 50% uncertainty on the

correction), indistinguishable from the ε182W values determined here for metals from

other low-Ti mare basalts. As outlined above, the correction procedure used here possibly

underestimates the true cosmogenic 182W by a factor of ~2, such that the cosmogenic 182W/184W of the 15555 whole-rock might be as high as ~1.4 ε182W, identical to the ε182W

reported for this sample8.

It is important to note that the possible underestimation of the cosmogenic 182W

production by neutron capture of 181Ta by a factor of ~2 has no significant influence on

the correction of the samples investigated here because the highest correction required

130

Page 131: Touboul Thesis Completed

here is only ~0.1 ε unit (Supplementary Table 4.1.2). In case of the correction for burnout

of W isotopes it is unlikely that the correction procedure underestimates the effects

significantly because the burnout effects observed in iron meteorites are similar to those

that are predicted by the correction method used here11.

4.1.5.4 Hf/W fractionation in the crystallizing lunar magma ocean

The first rocks to crystallize from the LMO consisted mainly of olivine and

pyroxene and were followed by plagioclase-rich cumulates that floated to the surface and

formed the earliest lunar crust14. The late stages of magma ocean crystallization involved

the precipitation of ilmenite and clinopyroxene until solidification of the last few percents

of the magma ocean14. This residual liquid is termed KREEP for its strong enrichment in

incompatible elements including K, Rare Earth Elements and Phosphorus17. Subsequent

melting and mixing among these primary rocks produced the variety observed in the

lunar sample suite. Mare basalts formed by remelting of early mafic cumulates which in

the case of high-Ti mare basalts included assimilation of ilmenite and clinopyroxene.

Impacts on the lunar surface resulted in the redistribution of KREEP, making KREEP the

major carrier of incompatible elements in lunar highland rocks. Hafnium-tungsten

fractionations in the LMO resulted from the compatibility of Hf in clinopyroxene-

ilmenite and the high incompatibility of W18,19. These led to high Hf/W in the high-Ti

mare basalt source and correspondingly low Hf/W in KREEP.

Based on mineral-melt partition coefficients for the relevant minerals in the

crystallizing LMO, Righter and Shearer18 estimated Hf/W=40-80 in the source of high-Ti

mare basalts. Measured Hf/W ratios of high-Ti mare basalts are even higher (up to ~110),

such that it is safe to assume Hf/W>40 for the high-Ti mare basalt source. This is

consistent with estimates for the Hf/W ratio in the low-Ti mare basalt source. This

consists of olivine and orthopyroxene, which are not capable of fractionating Hf and W18,

such that the Hf/W ratio of the low-Ti mare basalt source should be similar to that of the

bulk lunar mantle6. The latter has Hf/W=26.5±1.1, similar to Hf/W ratios measured for

low-Ti mare basalts. This ratio is lower than the Hf/W ratio estimated above for the high-

Ti mare basalt source, consistent with the presence of ilmenite and clinopyroxene.

Estimates for the Hf/W ratio of KREEP range from 12 to 19 (ref. 17,20) and hence are

131

Page 132: Touboul Thesis Completed

lower than the Hf/W ratio of the low-Ti mare basalt source, as expected from the high

incompatibility of W.

Figure 4.1.4: ε182W vs. time for the lunar and terrestrial mantles. The ε182W evolution for chondrites (CHUR) is shown for reference. The hatched area marks the combination of ε182W and time that is inconsistent with the Hf-W data for the lunar and terrestrial mantles.

4.1.5.5 Supplementary references

1. Reid, A. M., Meyer, C., Harmon, R. S. & Brett, R. Metal grains in Apollo 12

igneous rocks. Earth Planet. Sci. Lett. 9, 1-5 (1970).

2. Palme, H., Spettel, B., Wänke, H., Borchardt, D. & Stöffler, D. Can metal

segregation remove siderophiles from lunar impact melts? Lunar and Planetary

Science XIII, 609-610 (1982).

3. Wänke, H. et al. Apollo 12 samples: Chemical composition and its relation to

sample locations and exposure ages, the two component origin of the various soil

samples and studies on lunar metallic particles. Proc. Sec. Lunar Planet. Sci. Conf.

2, 1187-1208 (1971).

132

Page 133: Touboul Thesis Completed

4. Kleine, T., Mezger, K., Münker, C., Palme, H. & Bischoff, A. 182Hf-182W isotope

systematics of chondrites, eucrites, and Martian meteorites: Chronology of core

formation and mantle differentiation in Vesta and Mars. Geochim. Cosmochim.

Acta 68, 2935-2946 (2004).

5. Kleine, T., Mezger, K., Palme, H., Scherer, E. & Münker, C. The W isotope

composition of eucrites metal: Constraints on the timing and cause of the thermal

metamorphism of basaltic eucrites. Earth Planet. Sci. Lett. 231, 41-52 (2005).

6. Kleine, T., Palme, H., Mezger, K. & Halliday, A. N. Hf-W chronometry of lunar

metals and the age and early differentiation of the Moon. Science 310, 1671-1674

(2005).

7. Leya, I., Wieler, R. & Halliday, A. N. Cosmic-ray production of tungsten isotopes

in lunar samples and meteorites and its implications for Hf-W cosmochemistry.

Earth Planet. Sci. Lett. 175, 1-12 (2000).

8. Lee, D. C., Halliday, A. N., Leya, I., Wieler, R. & Wiechert, U. Cosmogenic

tungsten and the origin and earliest differentiation of the Moon. Earth Planet. Sci.

Lett. 198, 267-274 (2002).

9. Munker, C. et al. Evolution of planetary cores and the earth-moon system from

Nb/Ta systematics. Science 301, 84-87 (2003).

10. Klemme, S., Gunther, D., Hametner, K., Prowatke, S. & Zack, T. The partitioning

of trace elements between ilmenite, ulvospinel, annalcolite and silicate melts with

implications for the early differentiation of the moon. Chem. Geol. 234, 251-263

(2006).

11. Kleine, T., Mezger, K., Palme, H., Scherer, E. & Münker, C. Early core formation

in asteroids and late accretion of chondrite parent bodies: Evidence from 182Hf-182W

in CAIs, metal-rich chondrites and iron meteorites. Geochim. Cosmochim. Acta 69,

5805-5818 (2005).

12. Schérsten, A., Elliot, T., Hawkesworth, C., Russell, S. S. & Masarik, J. Hf-W

evidence for rapid differentiation of iron meteorite parent bodies. Earth Planet. Sci.

Lett. 241, 530-542 (2006).

133

Page 134: Touboul Thesis Completed

13. Markowski, A., Quitté, G., Halliday, A. N. & Kleine, T. Tungsten isotopic

compositions of iron meteorites: chronological constraints vs. cosmogenic effects.

Earth Planet. Sci. Lett. 242, 1-15 (2006).

14. Leya, I., Wieler, R. & Halliday, A. N. The influence of cosmic-ray production on

extinct nuclide systems. Geochim. Cosmochim. Acta 67, 529-541 (2003).

15. Nyquist, L. E. et al. Sm-146-Nd-142 Formation Interval for the Lunar Mantle.

Geochim. Cosmochim. Acta 59, 2817-2837 (1995).

16. Rankenburg, K., Brandon, A. D. & Neal, C. R. Neodymium isotope evidence for a

chondritic composition of the Moon. Science 312, 1369-1372 (2006).

17. Warren, P. H. & Wasson, J. T. The origin of KREEP. Rev. Geophys. Space Phys.

17, 73-88 (1979).

18. Righter, K. & Shearer, C. K. Magmatic fractionation of Hf and W: Constraints on

the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497-2507 (2003).

19. Shearer, C. K. & Newsom, H. E. W-Hf isotope abundances and the early origin and

evolution of the Earth-Moon system. Geochim. Cosmochim. Acta 64, 3599-3613

(2000).

20. Palme, H. & Wänke, H. A unified trace-element model for the evolution of the

lunar crust and mantle. Proc. Lunar Sci. Conf. 6th, 1179-1202 (1975).

134

Page 135: Touboul Thesis Completed

Section 4.2

182Hf-182W systematics of ferroan anorthosites and the lifetime of the lunar magma ocean*

M. Touboul1, T. Kleine1, B. Bourdon1, H. Palme2, R. Wieler1

1Institute for Isotope Geochemistry and Mineral Resources, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerland

2Institut für Geologie und Mineralogie, Universität zu Köln, Zülpicherstrasse 49b, 50674 Köln, Germany

*Accepté par Icarus

135

Page 136: Touboul Thesis Completed

136

Page 137: Touboul Thesis Completed

Abstract

New W isotope data for ferroan anorthosites 60025 and 62255 and low-Ti mare

basalt 15555 show that these samples, contrary to previous reports [Lee et al. 1997,

Science 278; Lee et al. 2002, Earth Planet Sci. Lett. 198], have a W isotope composition

that is indistinguishable from KREEP and other mare basalts. This requires crust

extraction on the Moon later than ~60 Ma after CAI formation, consistent with 147Sm-143Nd ages for ferroan anorthosites.

4.2.1 Introduction

The dominant processes in the early evolution of the Moon are the formation and

subsequent solidification of a magma ocean. As this lunar magma ocean (LMO)

crystallized, dense mafic minerals sank to the bottom and lighter plagioclase floated to

the surface, forming an anorthositic crust (Smith et al., 1970; Warren, 1985; Wood et al.,

1970). Most petrologic models for the magma ocean predict that olivine and

orthopyroxene were the first mafic minerals that crystallized, whereas clinopyroxene and

ilmenite formed in the late crystallization stages. The residual liquid of the LMO - termed

KREEP for its high contents of potassium (K), rare earth elements (REE) and phosphorus

(P) - is a distinct chemical component in the Moon and constitutes the major carrier of

incompatible elements in lunar highland rocks (Warren and Wasson, 1979). Later re-

melting of the mafic LMO cumulates gave rise to the formation of mare basalts and their

compositional diversity is thought to reflect melting of different cumulate assemblages:

low-Ti mare basalts were derived from olivine and orthopyroxene cumulates, whereas

high-Ti mare basalts were derived from cumulate layers containing olivine, pyroxenes

and ilmenite (Shearer and Papike, 1999).

The timescale for magma ocean solidification can be quantified using 182Hf-182W

chronometry (half life~9 Ma) because substantial Hf-W fractionations occurred during

crystallization of the LMO (Kleine et al., 2005; Righter and Shearer, 2003; Shearer and

Newsom, 2000). These fractionations stem from the high incompatibility of W and the

compatibility of Hf in clinopyroxene-ilmenite, leading to low Hf/W in KREEP and

complementary high Hf/W in the high-Ti mare basalt sources. Therefore, the presence of

137

Page 138: Touboul Thesis Completed

182W variations in lunar rocks would imply fractionation in the lunar interior within the

first ~60 Ma of the solar system because 182Hf became effectively extinct after that time.

Although 182W variations were initially reported among lunar whole-rock samples (Lee et

al., 1997), Touboul et al. (2007) recently showed that these variations are entirely due to

production of cosmogenic 182W during cosmic-ray exposure of the lunar surface.

Cosmogenic 182W is predominantly produced via the reaction 181Ta(n,γ)182Ta and

subsequent decay to 182W and is significant for lunar whole-rocks that have elevated

Ta/W (Kleine et al., 2005). However, Touboul et al. (2007) investigated metals, which

contain no Ta and hence no cosmogenic 182W. The homogeneous 182W/184W of metals

from KREEP-rich highland rocks, low-Ti and high-Ti mare basalts require that the

separation of KREEP and the mafic cumulates in the LMO occurred later than ~60 Ma

after CAI formation, otherwise high-Ti mare basalts would carry a resolvable positive 182W anomaly relative to KREEP (Touboul et al., 2007).

In contrast to the homogeneous 182W/184W of lunar metals, Lee et al. (1997)

reported positive 182W anomalies for some ferroan anorthosites. These cannot reflect

cosmogenic 182W production at the lunar surface because some of these samples have

young exposure ages of only ~2 Ma (Arvidson et al., 1975). The W isotope data for some

ferroan anorthosites, therefore, seem to imply that anorthosites were extracted from the

mantle within the first ~60 Ma of the solar system. This, however, is inconsistent with 147Sm-143Nd ages for ferroan anorthosites that are thought to reflect crystallization of

these rocks at 4.46±0.04 Ga (Norman et al., 2003). This inconsistency and the

significance of identifying any 182Hf-induced 182W variations among lunar samples for

constraining the timescales of lunar formation and differentiation have inspired the

present study.

We present new W isotope data for pure plagioclase separates from ferroan

anorthosites 60025 and 62255. These data are used to assess whether the anorthositic

lunar crust has a W isotope composition different from the lunar mantle. In addition, we

obtained new W isotope and Ta-Hf-W concentration data for low-Ti mare basalt 15555.

For this sample, Lee et al. (2002) did not find any evidence for cosmogenic 182W and

reported a significantly higher 182W/184W compared to metals from low-Ti mare basalts.

This is inconsistent with the homogeneous 182W/184W of metals from all mare basalt

138

Page 139: Touboul Thesis Completed

investigated so far and requires a reassessment of the W isotope composition of mare

basalt 15555.

4.2.2 Samples and analytical methods

Ferroan anorthosites 60025 and 62255 were selected for this study based on the

following criteria: (i) they have young exposure ages of ~2 Ma, such that the amount of

cosmogenic 182W is negligible; (ii) they have low contents of Ni and Ir, indicating that

they are pristine, i.e., any meteoritic contamination is minor to absent (Warren and

Wasson 1977); (iii) large amounts of sample are available, which is important because

the low W contents of anorthosites require large amounts of sample (~4-5 g) to be

processed.

Ferroan anorthosite 60025 is a coarse-grained cataclastic rock, consisting mainly

of calcic plagioclase. Mafic minerals are mainly olivine and orthopyroxene and are

heterogeneously distributed in 60025. The 60025 sample investigated for this study

contained only ~1 % of mafic material, which was entirely removed during sample

preparation. Ferroan anorthosite 62255 is a breccia of cataclastic anorthosites associated

with finely crystalline impact melt. These melt fragments (~5% of our sample) were first

removed before further processing of the sample. Both anorthosites were gently crushed

in an agate mortar until the entire sample had grain sizes of less than 500 μm. Pure

plagioclase separates were prepared from a 40-500 μm grain size fraction using heavy

liquids. They were further purified by hand picking under a binocular until all mafic

minerals were removed. Any potentially present metal grains were removed using a hand

magnet but the anorthosite samples investigated here did not contain metal. The pure

plagioclase fractions (~4 g for 60025; ~2 g for 62255) were then washed by

ultrasonication in distilled ethanol, dried and powdered in an agate mortar.

Due to the low W contents of anorthosites (~2-4 ppb W), large amounts of sample

needed to be processed to obtain sufficiently precise W isotope data. Moreover, the high

Ca contents require an effective removal of the sample matrix prior to the purification of

W by ion exchange methods that employ HF-containing acid mixtures. Samples were

dissolved in 60 ml Savillex® vials on a hotplate at 120°C for 3 days using 40 ml of 29 M

139

Page 140: Touboul Thesis Completed

HF and 10 ml 14 M HNO3. After digestion the samples were dried and repeatedly re-

dissolved in HNO3-H2O2 to remove organic compounds and dissolve Ca-fluorides that

formed during sample digestion. Then the samples were completely dissolved in 6 M

HCl-0.06 M HF and a ~5% aliquot was spiked with a mixed 180Hf-183W-180Ta tracer for

concentration determination by isotope dilution.

The remaining ~95% were dried and re-dissolved in 30 ml of 1 M HCl-0.1 M

HF, The solution was centrifuged and decanted and the residue washed two times with

~10 ml 1 M HCl-0.1 M HF and ultrasonicated at each step to ensure optimal release of W

from the fluoride residue. The solutions were then loaded onto pre-cleaned cation

exchange columns (15 ml BioRad® AG50WX8, 200-400 mesh) and the ~50 ml solutions

from each sample were split over 10 columns. Tungsten (together with other high field

strength elements) was eluted using 15 ml of a 1 M HCl- 0.1 M HF mixture, whereas

most other matrix elements (most notably Ca and Al) were adsorbed on the resin. This

procedure was adapted from the column A of the Hf chemistry described by Tatsumoto

and Patchett (1981). The W cuts for each sample were combined, dried and re-dissolved

in 10 ml 0.5 M HCl-0.5M HF and loaded onto pre-cleaned anion exchange columns (3 ml

AG1X8, 100-200 mesh), where W is separated from remaining matrix elements and

especially Zr, Nb, Hf, Ti, Mo, and Ta using our previously described methods (Kleine et

al., 2004a; Kleine et al., 2002). Each sample was split over two anion exchange columns

and the W cuts combined. These were re-dissolved in 5 ml 0.5 M HCl-0.5 M HF and W

further purified by an additional pass through the anion exchange columns. With this

procedure we obtained high purity W cuts and >90% of the W originally loaded on the

ion exchange columns.

A non-magnetic fraction from low-Ti mare basalt 15555 was processed using the

same procedure as for the anorthosites, but given its much smaller sample size (~400 mg

compared to ~4 g for the anorthosites), the entire sample could be processed on a single

cation exchange column, followed by purification on anion exchange resins as described

above.

Despite the extensive purification procedure, W blanks for the isotope

measurements were as low as 50±20 pg for the entire procedure. This corresponds to less

140

Page 141: Touboul Thesis Completed

than 1% of the W present in the samples, such that blanks are negligible. The blanks for

the isotope dilution measurements were also insignificant.

All isotope measurements were performed using a Nu Plasma MC-ICPMS at

ETH Zurich, which is equipped with a Cetac Aridus desolvating nebuliser. Details

regarding our measurement protocol are given elsewhere (Kleine et al., 2008). Repeated

measurements of the ALFA AESAR W standard during the course of this study yielded 182W/184W = 0.864843 ± 0.000015 (2σ of the mean) and 183W/184W = 0.467128 ±

0.000006 (2σ of the mean).

4.2.3 Results

Table 4.2.1: Hf-W data for lunar anorthosites and low-Ti mare basalt 15555

Sample W (ppb) Hf (ppb) Ta (ppb) 180Hf/184W ±2σ ε182W ±2σ ε183W ±2σ

Anorthosites

60025 plagioclase 1.8 8.0 0.5 5.2±0.4 -0.03±0.30 0.68±0.26

62255 plagioclase 2.2 17.6 1.8 9.4±0.4 -0.57±0.52 -0.83±0.43

Low-Ti mare basalt

15555 non-magnetic 58.5 2007 287.5 40.5±0.4 0.86±0.10 0.34±0.07

Uncertainties for the ε182W and ε183W values are the 2σ internal errors of the individual mass spectrometric runs.

The W isotope compositions and Hf-W-Ta contents for anorthosites 60025 and

62255 and low-Ti mare basalt 15555 are summarized in Table 4.2.1. The plagioclase

separates from both ferroan anorthosites have Hf and W concentrations that are lower

than the Hf and W concentrations previously reported for whole-rock analyses of these

two anorthosites (Lee et al., 1997). The Hf/W ratios range from ~4 to ~8 and are lower

than those of the mare basalts and KREEP (Kleine et al., 2005). The W isotope

composition of the plagioclase separates from ferroan anorthosites 60025 and 62255 are

indistinguishable from each other and also from the W isotope composition of metals

from KREEP-rich samples and low- and high-Ti mare basalts (Fig. 4.2.1). This contrasts

with results from an earlier study reporting positive ε182W values for some ferroan

anorthosites (Lee et al., 1997), including sample 60025. Note that compared to these

141

Page 142: Touboul Thesis Completed

previously published results for ferroan anorthosites, our new W isotope data are more

precise by a factor of ~5.

Figure 4.2.1: W isotope composition of plagioclase separates from ferroan anorthosites 60025 and 62255. Previously published W isotope data for ferroan anorthosites (Lee et al., 1997) are shown for comparison. The average compositions of KREEP (Kleine et al., 2005, Touboul et al. 2007) and the sources of low- and high-Ti mare basalts (Touboul et al. 2007) are also shown for comparison. ε182W is the deviation of the 182W/184W ratio of a sample relative to the terrestrial standard in parts per 10,000. Error bars shown here correspond to the 2σ external reproducibility (±0.4 ε and ±0.8 ε for 60025 and 62255, respectively), estimated by repeated measurements of ~15 and ~5 ppb ALFA AESER standard solutions.

142

Page 143: Touboul Thesis Completed

The non-magnetic fraction of low-Ti mare basalt 15555 has ε182W=0.9±0.3, ~58.5

ppb W and ~287 ppb Ta, corresponding to Ta/W∼4.9 (Fig. 4.2.2). The ε182W is somewhat

lower than the average ε182W of 1.3±0.4 previously reported for 15555 by (Lee et al.,

2002) and the Ta/W ratio is slightly higher than those reported for several fractions from

sample 15555 (Fig. 4.2.2). The higher Ta/W ratio for the non-magnetic fraction most

likely is due to the removal of metal from the sample analyzed here. Note that the amount

of metal obtained was insufficient for precise W isotope measurements.

Figure 4.2.2: W isotope composition versus Ta/W ratio for mineral separates and whole rocks from low-Ti mare basalt 15555. The predicted cosmogenic 182W production as a function of Ta/W are calculated using equations from Leya et al. (2003) and the Sm isotopic composition of 15555 (solid line, Nyquist et al. 1995) or noble gas exposure ages (dashed lines, York et al., 1972, Podosek et al., 1971; Marti and Lightner, 1971). Dashed curves are labeled with two possible exposure ages for sample 15555. The error bar shown here for the non-magnetic fraction from mare basalt 15555 corresponds to the 2σ external reproducibility of ±0.3 ε, as estimated by repeated measurements of a ~20 ppb ALFA AESAR standard solution.

143

Page 144: Touboul Thesis Completed

4.2.4 Discussion

4.2.4.1 Tungsten isotope homogeneity in the Moon

Our W isotope data for plagioclase separates from ferroan anorthosites 60025 and

62255 contrast with results from Lee et al. (1997), who reported positive ε182W values for

some anorthosites (including 60025). It is unlikely that the higher 182W/184W reported

earlier reflects the presence of cosmogenic 182W because the exposure age of sample

60025 is only ~2 Ma; the calculated cosmogenic 182W production by neutron capture of

181Ta in this case is less than ~0.1 ε units for Ta/W ratios lower than ~15 (note that the

Ta/W ratio of 60025 whole rock is ~2.2 based on data from Lee et al., 1997 and Münker

et al., 2003). The reason for the discrepancy between our and previously reported W

isotope data for ferroan anorthosites is unclear but could be related to the presence of

cosmogenic 182W in the mafic component in the 60025 sample analyzed by Lee et al.

(1997), which is not present in the pure plagioclase separate investigated in the present

study. The presence of a mafic component can also account for the higher W contents

reported by Lee et al. (1997) for their 60025 sample. However, it is difficult to assess

whether this mafic component has cosmogenic 182W, such that the origin of the elevated 182W reported by Lee et al. (1997) remains unclear. Note that the amount of mafic

material separated from the 60025 investigated here is insufficient to obtain precise W

isotope data.

Whatever the reason for the elevated 182W reported for ferroan anorthosites in an

earlier study, several lines of evidence indicate that the 182W/184W of the lunar highlands

is best presented by W isotope data presented here for samples 60025 and 62255. First,

the data presented here are a factor of ~5 and do not show the large range of W isotope

compositions reported earlier. Second, in the present study W isotope data were obtained

on pure plagioclase separates and therefore do not contain components that could have

spuriously elevated 182W/184W. Third, the 182W/184W ratios obtained here for ferroan

anorthosites are consistent with current models for the chronology of the lunar magma

ocean (see below). In contrast, given that anorthosites have low Hf/W ratios, the elevated

and variable 182W/184W reported for anorthosistes earlier would require a complex

petrogenesis involving evolution in a high Hf/W reservoir followed by re-melting of this

reservoir to generate the anorthosites.

144

Page 145: Touboul Thesis Completed

Lee et al. (2002) reported Ta-W data for several mineral separates from low-Ti

mare basalt 15555. In spite of their different Ta/W, the 182W/184W of these mineral

separates are indistinguishable within uncertainty (individual uncertainties range from 0.4

to 1.5 ε182W). Lee et al. (2002) therefore argued that mare basalt 15555 does not contain

any significant cosmogenic 182W and that its elevated ε182W of 1.3±0.4 reflects in-situ

decay of 182Hf. However, noble gas isotope systematics of mare basalt 15555 indicate a

cosmic-ray exposure at the lunar surface over a period of 70 to 90 Ma (York et al., 1972,

Podosek et al., 1971; Marti and Lightner, 1971). Likewise, low 149Sm/150Sm ratios in this

sample reflect burnout of 149Sm by thermal neutrons produced by interaction with the

cosmic rays (Nyquist et al., 1995). The noble gas and Sm isotope data provide two

distinct approaches for calculating the magnitude of cosmogenically produced 182W as a

function of Ta/W ratio. The approach based on Sm isotopes is more reliable because it

provides a direct proxy for the thermal neutron flux, which is the main concern when

determining the amount of cosmogenic produced 182W in lunar samples (Leya et al.,

2003).

Fig. 4.2.2 shows the expected 182W variations due to neutron capture of 181Ta as a

function of their Ta/W ratios, calculated using the 149Sm/150Sm ratio of mare basalt 15555

(Nyquist et al., 1995) and correction equations given in Leya et al. (2003). Fig. 2 also

shows the expected effects using the exposure age of mare basalt 15555 as a proxy for the

thermal neutron flux (see Leya et al., 2003). Both methods yield similar results and reveal

that for a range in Ta/W from ~0 to ~5 (i.e., the range in Ta/W that has been reported for

several fractions of 15555), the 182W/184W can shift by ~1 ε182W. This is significant

because this effect is similar to the reported ε182W anomaly of mare basalt 15555.

Our new data for a non-magnetic fraction of mare basalt 15555, as well as

previously reported data for a 15555 whole-rock (Kleine et al., 2005), are remarkably

consistent with the predicted cosmogenic effect in this sample (Fig. 4.2.2). After

correction for cosmogenic 182W, the non-magnetic fraction and the whole rock of this

sample have ε182W values of 0.1±0.4 and -0.1±0.8, respectively. This provides to

important conclusions. First, after proper correction for cosmogenic effects, sample

15555 has ε182W~0, indistinguishable from the ε182W values of metals from other mare

basalts (Touboul et al., 2007). Second, both the presence of cosmogenic noble gases as

145

Page 146: Touboul Thesis Completed

well as the depletion in 149Sm indicate a significant thermal neutron flux in sample

15555, such that mineral separates having different Ta/W ratios should exhibit different 182W/184W ratios. These variations are small and were not resolvable with the precision

obtained by Lee et al. (2002). However, the data for mineral separates of 15555 reported

by Lee et al. (2002) plot systematically above the line for the predicted cosmogenic 182W

production. The reasons for this are unclear but our new data for sample 15555 in

combination with noble gas and Sm isotope data show that this sample contains

cosmogenic 182W and thus is not suited to determine the indigenous 182W/184W of mare

basalts. Therefore, the indigenous 182W/184W of low-Ti mare basalts is best defined by W

isotope for their metals, all of which have ε182W~0 (Touboul et al., 2007).

4.2.4.2 Duration of magma ocean solidification

The indistinguishable 182W/184W ratios of metals from low- and high-Ti mare

basalts and KREEP (Touboul et al. 2007) combined with the corrected 182W/184W of mare

basalt 15555 indicate that the different Hf/W ratios of the mare basalt sources and

KREEP were established after 182Hf extinction, i.e., later than ~60 Ma after CAI

formation. Likewise, ferroan anorthosites have an indigenous 182W/184W that is

indistinguishable from that of the mare basalts and KREEP. Given that ferroan

anorthosites have Hf/W ratios much lower than KREEP (Hf/W~12-19; Kleine et al.

2005) and the mare basalt sources (Hf/W~26 for low-Ti and Hf/W>40 for high-Ti mare

basalts; Kleine et al. 2005), extraction of ferroan anorthosites from the lunar mantle must

also have occurred later than ~60 Ma after CAI formation. Most importantly, this revised

Hf-W timescale for the formation of the lunar crust is now consistent with 147Sm–143Nd

ages for ferroan anorthosites of 4456±40 Ma (i.e., 112±40 Ma after CAI formation)

(Norman et al. 2003). In most models for the solidification of the lunar magma ocean, the

lunar crust formed by flotation of plagioclase after substantial (~50-70 %) crystallization

of the magma ocean (Snyder et al., 1992). The 182Hf-182W and 147Sm-143Nd systematics

therefore indicate that ~50-70 % crystallization of the magma ocean could have been

achieved at ~60 Ma at the earliest and at ~150 Ma at the latest.

These ages can potentially be used to constrain the duration of magma ocean

solidification but this requires an independent estimate for the formation time of the

146

Page 147: Touboul Thesis Completed

Moon. Based on Hf-W systematics of Earth's mantle and chondritic meteorites it has

been argued that the giant Moon-forming impact might have occurred as early as ~30 Ma

(e.g., Jacobsen, 2005) but the Hf-W systematics are also consistent with a later formation

of the Moon (Halliday, 2004; Kleine et al., 2004b). More recently, Touboul et al. (2007)

showed that the Moon most likely formed later than ~60 Ma after CAI formation,

provided that the Hf/W ratios of the lunar and terrestrial mantles are different, as seems

likely. Therefore, the uncertainties in the age of the Moon and the age of the ferroan

anorthosites currently hamper a precise determination of the duration of magma ocean

solidification and the available age constraints are consistent with both an almost

immediate crystallization and a more protracted timescale of ~100 Ma.

Thermal models of magma oceans indicate that cooling might be very rapid

initially but may become more protracted as soon as an insulating crust forms

(Solomatov, 2000; Tonks and Melosh, 1990). Therefore, if ferroan anorthosites represent

the first crust of the Moon, their crystallization age should closely correspond to the age

of Moon formation. In this case, the indistinguishable 182W/184W ratios of ferroan

anorthosites and other lunar rocks (mare basalts and KREEP) require that the Moon

formed later than ~60 Ma, consistent with the age estimate from Touboul et al. (2007). It

should be noted however that the formation of the ferroan anorthosites might have been

far more complex and might have involved remelting of earlier cumulates (Longhi,

2003). In this case, the crystallization age of ferroan anorthosites might not provide a

close estimate of the time of Moon formation.

Based on 146Sm-142Nd systematics it has been argued that the final solidification

of the lunar magma ocean might have occurred as late as ~200-250 Ma after CAI

formation (Rankenburg et al., 2006). More recently, however, Boyet and Carlson (2007)

argued that the 142,143Nd systematics of lunar samples is most consistent with a

differentiation age of ~4.45 Ga (i.e., ~100 Ma after CAI formation), implying that the 146Sm-142Nd “isochron” of the lunar mantle does not reflect the timing of differentiation.

Likewise, Bourdon et al. (2008) showed that cumulate overturn, magma mixing and

melting following magma ocean solidification at 50-100 Ma could produce a linear array

in the 146Sm-142Nd isochron diagram that yields an apparent age of ~200-250 Ma.

Therefore, the most reliable time constraints on magma ocean solidification are currently

147

Page 148: Touboul Thesis Completed

provided by the Hf-W systematics of lunar rocks and the 147Sm-143Nd ages of ferroan

anorthosites.

4.2.5 Conclusions

We have shown that, contrary to a previous report (Lee et al. 2002), low-Ti mare

basalt 15555 contains cosmogenic 182W and that its indigenous 182W/184W is

indistinguishable from that of metals from all other mare basalts analyzed for W isotopes

so far (Touboul et al., 2007). We have also shown that in contrast to a previous result

(Lee et al., 1997), ferroan anorthosites have an indigenous 182W/184W that is

indistinguishable from that of KREEP and the mare basalts. The homogeneous 182W/184W

ratios of all lunar rocks requires differentiation of the Moon later than ~60 Ma after CAI

formation, consistent with 147Sm-143Nd ages for ferroan anorthosites.

References

Arvidson, R., et al., 1975. Cosmic ray exposure ages of features and events at the Apollo

landing sites. The Moon 13, 259-276.

Bourdon, B., Touboul, M., Caro, G., Kleine T., 2008. Early differentiation of the Earth

and the Moon. Phil. Trans. Royal Soc., in press.

Boyet, M., Carlson, R. W., 2007. A highly depleted moon or a non-magma ocean origin

for the lunar crust? Earth Planet Sci. Lett. 262, 505-516.

Halliday, A. N., 2004. Mixing, volatile loss and compositional change during impact-

driven accretion of the Earth. Nature 427, 505-509.

Jacobsen, S. B., 2005. The Hf-W isotopic system and the origin of the Earth and Moon.

Ann. Rev. Earth Planet. Sci. 33, 531-570.

Kleine, T., et al., 2004a. 182Hf-182W isotope systematics of chondrites, eucrites, and

Martian meteorites: Chronology of core formation and mantle differentiation in

Vesta and Mars. Geochim. Cosmochim. Acta 68, 2935-2946.

Kleine, T., et al., 2004b. The W isotope evolution of the bulk silicate Earth: constraints

on the timing and mechanisms of core formation and accretion. Earth Planet Sci.

Lett. 228, 109-123.

148

Page 149: Touboul Thesis Completed

Kleine, T., et al., 2002. Rapid accretion and early core formation on asteroids and the

terrestrial planets from Hf-W chronometry. Nature 418, 952-955.

Kleine, T., et al., 2005. Hf-W chronometry of lunar metals and the age and early

differentiation of the Moon. Science 310, 1671-1674.

Kleine, T., et al., 2008. Hf-W thermochronometry: closure temperature and constraints on

the accretion and cooling history of the H chondrite parent body. Earth Planet Sci.

Lett. 270, 106-118.

Lee, D. C., et al., 2002. Cosmogenic tungsten and the origin and earliest differentiation of

the Moon. Earth Planet Sci. Lett. 198, 267-274.

Lee, D. C., et al., 1997. Age and origin of the moon. Science 278, 1098-1103.

Leya, I., et al., 2003. The influence of cosmic-ray production on extinct nuclide systems.

Geochim. Cosmochim. Acta 67, 529-541.

Longhi, J., 2003, A new view of lunar ferroan anorthosites: Postmagma ocean

petrogenesis. J. Geophys. Res. 108(E8), 2-1.

Marti, K., Lightner, B. D., 1972. Rare gas record in the largest Apollo 15 rock. Science

175, 421 - 422.

Norman, M. D., et al., 2003. Chronology, geochemistry, and petrology of a ferroan

noritic anorthosite clast from Descartes breccia 67215: Clues to the age, origin,

structure, and impact history of the lunar crust. Met. Planet. Sci. 38, 645-661.

Nyquist, L. E., et al., 1995. 146Sm-142Nd Formation interval for the lunar mantle.

Geochim. Cosmochim. Acta 59, 2817-2837.

Podosek, F. A., Huneke, J. C., Wasserburg, G. J., Gas-Retention and cosmic-ray exposure

ages of lunar rock 15555. Science 175, 423-425.

Rankenburg, K., Brandon, A. D., Neal, C. R., 2006. Neodymium isotope evidence for a

chondritic composition of the Moon. Science 312, 1369-1372.

Righter, K., Shearer, C. K., 2003. Magmatic fractionation of Hf and W: Constraints on

the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497-2507.

Shearer, C. K., Newsom, H. E., 2000. W-Hf isotope abundances and the early origin and

evolution of the Earth-Moon system. Geochimica et Cosmochimica Acta. 64, 3599-

3613.

149

Page 150: Touboul Thesis Completed

Shearer, C. K., Papike, J. J., 1999. Magmatic evolution of the Moon. Am. Min. 84, 1469-

1494.

Smith, J. V., et al., 1970. Petrologic history of the moon inferred from petrography,

mineralogy and petrogenesis of Apollo 11 rocks. Proceedings of the Apollo 11

Lunar Science Conference. 1, 897-925.

Snyder, G. A., Taylor, L. A., Neal, C. R., 1992. A chemical model for generating the

sources of mare basalts - combined equilibrium and fractional crystallization of the

lunar magmasphere. Geochim. Cosmochim. Acta 56, 3809-3823.

Solomatov, V. S., 2000. Fluid dynamics of the terrestrial magma ocean. In: Canup, R. M.

and Righter, K. (eds) Origin of the Earth and Moon. Univ. Arizona Press, pp. 323-

338.

Touboul, M., et al., 2007. Late formation and prolonged differentiation of the Moon

inferred from W isotopes in lunar metals. Nature 450, 1206-1209.

Tonks, W. B. and Melosh, H. J., 1990. The physics of crystal settling and suspension in a

turbulent magma ocean. In: Newsom, H. E. and Jones, J. H. (eds). Origin of the

Earth. Oxford Univ. Press, New York, pp. 151 - 174.

Warren, P. H., 1985. The magma ocean concept and lunar evolution. Ann. Rev. Earth

Planet. Sci. 13, 201-240.

Warren, P. H., Wasson, J. T., 1977. Pristine nonmare rocks and the nature of the lunar

crust. Proceedings of the 8th Lunar Science Conference, 2215-2235.

Warren, P. H., Wasson, J. T., 1979. The origin of KREEP. Rev. Geophys. Space Phys.

17, 73-88.

Wood, J. A., et al., 1970. Lunar anorthosites and a geophysical model of the moon.

Proceedings of the Apollo 11 Lunar Science Conference. 1, 965-988.

York, D., Kenyon, W. J., Doyle, R. J., 1972, 40Ar-39Ar ages of Apollo 14 and 15 samples.

Proceedings of the Lunar Science Conference. 2, 1613-1622.

150

Page 151: Touboul Thesis Completed

Section 4.3

Early differentiation of the Earth and the Moon

Bernard Bourdon1*, Mathieu Touboul1, Guillaume Caro2, Thorsten Kleine1

1 Institute of Isotope Geochemistry and Mineral Resources, ETH Zurich, Clausiusstrasse 25, Zurich, CH-8092 Switzerland 2 Centre de Recherches Pétrographiques et Géochimiques- CNRS, 15 rue Notre Dame des Pauvres, 54501 Vandoeuvre-les-Nancy, France

Publié dans Philosophical Transactions of the Royal Society A doi: 10.1098/rsta.2008.0125

151

Page 152: Touboul Thesis Completed

152

Page 153: Touboul Thesis Completed

Abstract

We examine the implications of new 182W and 142Nd data for Mars and the Moon

for the early evolution of the Earth. The similarity of 182W in the terrestrial and lunar

mantles indicates that the Moon-forming giant impact took place most likely >60 Ma

after CAI formation (4.568 Gyr) because as a result of equilibration between the lunar

magma disk and the Earth's mantle, the lunar and terrestrial mantles must have started off

with identical initial 182W/184W ratios. This is not inconsistent with the apparent U-Pb age

of the Earth. New 142Nd data for martian meteorites show that Mars probably has a super-

chondritic Sm/Nd that could coincide with that of the Earth and Moon. If this is

interpreted by an early mantle differentiation event, this requires a buried enriched

reservoir for the three objects. This is highly unlikely. For the Earth we show, based on

new mass balance calculations for Nd isotopes, that the presence of a hidden reservoir is

difficult to reconcile with the combined 142Nd-143Nd systematics of the Earth's mantle.

We argue that a likely possibility is that the missing component was lost during or prior

to accretion. Furthermore, 142Nd data for the Moon that was used to argue for a ~200 Myr

for the solidification of the magma ocean is reinterpreted. Cumulate overturn, magma

mixing and melting following lunar magma ocean crystallization at 50-100 Myr could

have yielded the 200 Myr model age.

4.3.1 Introduction

Understanding the formation of planets is proven to be a crucial task for

constraining their later evolution. First, this question is inherently linked to the bulk

composition of the planet since designing a scenario for planet accretion and

differentiation is often necessary to derive their composition (Sun and McDonough,

1995, Allègre et al. 1995). Particularly crucial in this respect are the behaviors of

siderophile elements. Namely the incorporation of these elements in the core will be a

function of the accretion scenario and oxidation state of the Earth. This will in turn

determine the chemical composition of the mantle. Second, as early differentiation events

rapidly follow accretion, a description of these processes is essential for determining the

initial conditions of the Earth’s system. The initial energy available upon accretion may

153

Page 154: Touboul Thesis Completed

allow a wholesale differentiation that may be difficult to erase at a later stage. There are

now indications that following magma ocean crystallization, a dense and enriched

reservoir could have formed in the lower mantle (Boyet and Carlson 2005; Labrosse et al.

2007) and fluid dynamical constraints on mantle convection have shown that if an initial

density stratification existed in the Earth then it is likely that it could have persisted over

timescales of several billion years (Davaille et al. 2002). Third, the early evolution of

planets will also determine their rate of cooling and this has implications on stirring rates.

A large planet is likely to cool slower but this can be strongly compounded by the

existence of an insulating lid. Another important parameter is the presence of water. If

water is a component in the accreting material (as is the case on Earth), the lithosphere

becomes more ductile and plate tectonics is more likely to take place. This will ultimately

result in faster cooling rates.

In this contribution, we examine recent data we have obtained for the Moon and

Mars that pertain to the early evolution of the Earth and Moon and we discuss their

implications for the evolution of terrestrial planets. This approach further emphasizes that

the study of nearby planetary objects can provide a wealth of information about the

history of our own planet.

4.3.2 New constraints on the age of the Moon and termination of Earth accretion

4.3.2.1 New 182W data and chronological implications

There are several reasons to argue that the giant impact represents the last

significant stage of terrestrial accretion (Canup and Asphaug, 2000; Morbidelli this

volume). First, significant accretion following Moon formation could only have involved

numerous small impacts because otherwise the angular momentum of the Earth-Moon

system would have been substantially altered and the Moon might have been lost.

Second, if significant amounts of mass had been accreted after the Giant Impact, there

would be a potential problem in accounting for the siderophile elements in both objects.

These considerations limit the amount of mass that can have accreted to the Earth after

the giant impact to less than 0.05 Earth's masses (Canup and Asphaug, 2000). Third, if

significant accretion takes place after the Moon-forming giant impact, then the oxygen

isotopes in the Earth and the Moon could have become different. High precision oxygen

154

Page 155: Touboul Thesis Completed

isotope measurements have shown that within ∼0.02‰ this is not the case (Wiechert et al.

2001). If we assume that the later incoming material had a ∼1‰ deviation in δ18O from

the Earth (equivalent to the difference between Mars and Earth), this limits the mass

fraction of material accreted by a giant impact to less than 2%. A similar calculation

could be made for δ17O. Thus, it would seem that the age of the Moon would provide an

age of termination for the Earth accretion.

Several previous studies have attempted to derive a meaningful age for the

formation of the Moon based on 182Hf-182W chronometry (Lee et al. 1997, Shearer and

Newsom 2000, Righter and Shearer 2003, Lee et al. 2002, Kleine et al. 2005, Touboul et

al. 2007). One major difficulty has been to tackle the issue of cosmogenic production of 182W via the reaction 181Ta(n,γ)182Ta followed by β- decay to 182W. This reaction is

obviously enhanced in silicate rocks for which Ta/W ratios are higher than in metals.

After an initial report of large variations in the 182W/184W of lunar whole-rocks (Lee et al.

1997) it was realized that the enhanced 182W/184W ratios in many samples reflects

cosmogenic 182W production (Leya et al. 2000). Attempts to correct for these effects have

yielded imprecise results because of the large corrections required. The most direct

method to determine the 182W/184W unaffected by cosmic-ray effects is the W isotope

analyses of Fe-metal separated from lunar rocks because these metals do not contain any

Ta and hence cosmogenic 182W (Kleine et al. 2005, Touboul et al. 2007). As shown in

Figure 4.3.1, the new study by Touboul et al. (2007), which used larger sample sizes and

focused on the analysis of high-purity metal separates seems to represent the most self-

consistent data set. Remarkably, the W isotopes are constant among all the investigated

lunar samples and similar to terrestrial samples. In what follows we examine in more

details the implications of these results.

155

Page 156: Touboul Thesis Completed

Figure 4.3.1: Compilation of W isotope data for the Moon. Data source : Touboul et al. (2007); Lee et al. (2002); Kleine et al. (2005). The data clearly shows that the ε182W for the Moon is indistinguishable within error from the terrestrial composition. Circles are for KREEP-rich samples, diamonds are for high Ti-mare basalts and squares for low-Ti mare basalts.

156

Page 157: Touboul Thesis Completed

4.3.2.1.1 New 182W data and lunar mantle evolution

The new 182W isotope data for the Moon has three major important features. First,

the W isotope composition of the Moon is distinct from that of chondrites. Second, the 182W/184W ratio of the lunar mantle is identical to that of the bulk silicate Earth and third,

despite a significant range in Hf/W in their sources, all lunar samples have identical

ε182W (defined as (182W/184Wsample/182W/184Wstd-1)×104). In what follows, we discuss in

more detail, the implications of these features. This discussion represents an extension of

the first report by Touboul et al. (2007).

The formation of the Moon by a giant impact implies that its original temperature

must have been very high. As a consequence, the initial state of the Moon was so hot that

it must have been molten and must have evolved initially as a magma ocean. The magma

ocean concept proposed for the Moon is in part based on the existence of complementary

Eu anomalies in the lunar crust (anorthosite) and lunar basalts (mare basalts) as well as

the existence of the so-called KREEP component assumed to represent the latest stage of

magma ocean crystallization. The data of Touboul et al. (2007) provides an important

outlook on this question because it can be used to estimate a lower limit for the age of the

lunar magma ocean. At face value, the data indicates that the closure of the Hf-W system

took place more than 60 Ma after the beginning of the solar system. In the context of a

lunar magma ocean, one could assume that the closure of the system took place when the

convection in the magma ocean slows down (60% crystal fraction, Solomatov 2000). As 182Hf is short-lived (t1/2=8.9 Ma), the absence of 182W differences in the lunar rocks

implies that the various lunar reservoirs have been remixed and isotopically homogenized

until after 182Hf became effectively extinct. This corresponds roughly to a time of at least

60±10 Myr (Touboul et al. 2007), given the range in Hf/W ratios in the sources of lunar

rocks (Shearer and Newsom 2000; Righter and Shearer 2003; Kleine et al. 2005). This

timescale now seems to be consistent with the 142Nd data available for the Moon,

indicating Sm-Nd closure around 200 Ma, if one takes the data at face value. One

surprising aspect of the Sm-Nd data for the Moon is that it is much younger than the

corresponding Sm-Nd age for Mars and Earth (Caro et al. 2003, Bennett et al. 2007,

Boyet and Carlson 2005). This would imply that the Moon in spite of its smaller size had

a longer cooling time than Mars and Earth. This would seem rather paradoxical and has

157

Page 158: Touboul Thesis Completed

been rationalized by Albarède and Blichert-Toft (2007) who argued that the lunar

anorthosite crust has formed a lid that insulated the Moon and considerably slowed down

the cooling of the Moon (Martin et al. 2006). Such a stable and early lid was not

preserved on Earth and thus the cooling time could have been shorter. It is quite clear that

conductive cooling through a boundary layer is much slower than if the boundary layer is

not stable, which is typically the case when plate tectonics causes the generation of new

seafloor. There is, however, a difficulty in the scenario proposed by Martin et al. (2006)

as the plagioclase forming the anorthosite crust only appears when the lunar magma

ocean has cooled by 500°C and the crystal fraction has reached at least 70% (e.g. Longhi

2003). By that time, the temperature in the lunar mantle would be ∼1250 °C and

convection should be more sluggish. Above a 60% melt fraction, convection of the

crystal-melt assemblage becomes sluggish but the mobility of melts at a planetary scale

could allow some chemical reequilibration. Using the scaling of Solomatov (2007) for the

case of the Moon, a typical differentiation timescale related to melt percolation would be

∼3 Ma. Thus, the chemical and isotope equilibration would effectively stop shortly after

the 60% crystal fraction is reached. This simple calculation does not consider the effect

of compositionally-driven melt convection that could take place in the magma ocean

cumulate pile. Thus, it is not clear how Nd isotopes would effectively equilibrate at a

planetary scale to yield a ~200 Myr isochron. There seems to be a contradiction that calls

upon a reexamination of this question. One possibility to explain the young 146Sm-142Nd

apparent ‘age’ of the Moon is if thermal solid-convection and later differentiation would

have disturbed the original 146Sm-142Nd isochron derived from magma ocean

crystallization. In what follows, we explore this possibility with a simple box model and

its implications for the actual age of lunar differentiation. In this respect, the lunar data

could be compared with terrestrial data. For several long-lived isotope systems, we know

that, in open and continuously differentiating systems, the slope of isochrons does not

directly date a differentiation event (Allègre et al. 1995, Kellogg et al. 2002, Rudge,

2006). Rather, the slope of a mantle isochron is a function of both differentiation and

later mixing. Although the history of the Moon does not involve plate tectonics and

thorough mixing, magmatic activity in the Moon lasted until ~3 Ga ago. Thus, it is worth

158

Page 159: Touboul Thesis Completed

exploring models of continuing differentiation that would yield a 200-250 Myr isochron

in the 146Sm-142Nd system.

To approach this question, we have used a simplified statistical model including

melting of the lunar mantle, radioactive decay and mixing. The model runs over a number

of steps and at each step the melt is assumed to mix with a randomly chosen component

already present in the lunar mantle (or crust). This model attempts to simulate the various

processes affecting the lunar mantle: magma ocean crystallization, cumulate overturn,

melting and contamination. The essential processes involved are thus fractionation due to

melting (or crystallization) and mixing. As shown on Fig. 4.3.2, this statistical model can

produce an array in the 142Nd/144Nd versus 147Sm/144Nd diagram that mimics the

observations in Sm-Nd data. The slope of this array does not represent the age of

fractionation since in this model Sm-Nd fractionation starts at 70 Myr and proceeds until

3.5 Gyr ago. Rather, it is a function of the successive steps of fractionation and mixing.

Arguably, there is more dispersion in the model array in Fig. 4.3.2, which indicates that

the model does not perfectly match observations. On the other hand, the 142Nd data set for

lunar rocks is still limited and more data should enable a better assessment. The basic

conclusion of this simplistic modeling is that the apparent 200 Myr age for the Moon

does not necessarily represent the age of crystallization of the lunar magma ocean.

Rather, it reflects both crystallization of the LMO as well as later processes that both

have induced Sm/Nd fractionation and mixing.

159

Page 160: Touboul Thesis Completed

Figure 4.3.2: 146Sm-142Nd systematics and the age of solidification of the magma ocean. The slope in the 142Nd versus 147Sm/144Nd diagram has been interpreted as reflecting the solidification of the lunar magma ocean. Our simple statistical model of fractionation/mixing shows that this need not be the case. The model operates with five time steps of 70 Ma and at each step, component are fractionated (or not) and the melt can be then re-mixed with a pre-existing component. The model then runs for 350 Ma. One can obtain an array using the following parameters in a melting model: DSm=0.1 and DNd=0.07, melt fraction 0.05, fraction of mixing=0.05 at each time step. We assume that the Moon forms at 70 Ma followed the giant impact.

4.3.2.1.2 Exploring scenarios for Moon-Earth equilibration following the giant

impact

As mentioned above, an important aspect of the 182W isotope in lunar rocks is that

the ε182W in the lunar and terrestrial mantles are identical while their Hf/W ratios are not

identical. Using U/W, Th/W and Th/U of the lunar and terrestrial mantles, Touboul et al.

(2007) estimated the following values: Hf/WMoon=26.4±1.5 Hf/WEarth=18±5. As

explained in Touboul et al. (2007), there are several ways of rationalizing this important

observation. However, in order to derive time constraints on the giant impact, one needs

to define a scenario for the interaction of the impactor with Earth. One reasonable

expectation is that prior to the impact, the impactor was differentiated into a mantle and a

160

Page 161: Touboul Thesis Completed

core. Thus, upon impact it is possible that the impactor’s core merged almost entirely

with the Earth’s core. In this case, the equilibration would have involved equilibration

between the Earth and the impactor’s mantle. Alternatively, the impactor’s core might be

entirely or partially mixed with the Earth’s mantle during the impact, in which case

metal-silicate interaction will be the dominant process (Halliday 2004, Kleine et al. 2004,

Nimmo et al. 2006). Another important aspect is the age of core differentiation for the

impactor and the proto-Earth. As noted by Nimmo et al. (2006), depending on the age of

differentiation and the degree of metal-equilibration, the ε182W could be either highly

radiogenic or simply close to chondritic values. Furthermore, the oxidation state of the

impactor and proto-Earth (Halliday 2004) will strongly influence the degree of Hf/W

fractionation and hence the ε182W of the mantle. A model including all these aspects

would have too many unconstrained parameters (Halliday 2004, Kleine et al. 2004,

Nimmo et al. 2006). To make this problem more tractable while assessing the

implications of the new data of Touboul et al. (2007), we have considered a model for the

Giant Impact using only two parameters. First, we consider the W isotope difference

between the proto-lunar mantle and the proto-Earth and we use this as the first parameter

in the model. The second parameter of this model is the degree of reequilibration (f)

between the proto-Earth’s mantle and the proto-lunar mantle and is defined as:

182

1821 after impact

before impact

Wf

Wεε

Δ= −

Δ

With this definition, f=0 means no equilibration while f=1 means full equilibration. Our

approach was to consider two-stage models for the evolution of the impactor because we

will only focus on the latest event in Earth accretion (i.e., the Giant Impact and Moon

formation). As argued by Morbidelli (this volume), it is very likely that the Moon

represented the latest giant impact while there could have been a number of preceding

‘giant impacts’. It would seem more logical to model the Earth accretion with a

multistage model, but in our case, since we are only dealing with the last event,

considering only a mean composition for Earth and impactor is sufficient. Our modelling

enables to consider only two parameters to represent a large range of possible models

including variable degrees and processes of equilibration between proto-Earth and proto-

Moon, previous history of the impactor and proto-Earth (including variable core sizes,

161

Page 162: Touboul Thesis Completed

162

Hf/W ratios, or age of proto-core formation). It considerably extends the number of

possibilities considered in Touboul et al. (2007).

The results of the modelling are shown in Fig. 4.3.3 and 4.3.4. What this

modeling shows is that there is a priori a large range of possibilities that would satisfy the

observations. As shown in Fig. 4.3.3, an important feature that the theoretical lower age

for the earliest formation of the Moon is ~37 Ma, as obtained from a two-stage model

(Touboul et al. 2007), which assumes a chondritic initial 182W/184W of the Moon.

However, two lines of evidence suggest that this lower limit is highly improbable. First, it

is unlikely that the Moon had a chondritic initial W isotope composition because the

Moon formed mostly from high Hf/W mantle material. Second, the difference in both

Hf/W and initial ε182W between the lunar and terrestrial mantles must have been such that

the ε182W evolved fortuitously to identical present-day values.

Another important result of this modelling is that it shows that some degree of

equilibration between the proto-lunar mantle and the proto-Earth is very likely.

Depending on the Hf/W ratio in the impactor's mantle, the accretion and differentiation

history of the impactor, and the degree to which impactor core material was remixed with

proto-lunar material, the ε182W of the proto-lunar mantle could in principle be either

lower or higher than the ε182W of the proto-Earth. However, consideration of Fig. 4.3.3a

shows that since the present-day Hf/W of the lunar mantle is higher than that of the Earth,

the ε182W in the lunar mantle had to be lower than the ε182W of the Earth, if the giant

impact occurred before 62 Myr (Fig. 4.3.4). This is only possible if the impactor's mantle

was highly oxidized (and hence had a relatively low Hf/W) or if there has been more

silicate-metal equilibration between proto-lunar mantle and impactor core than with the

Earth’s mantle. Given that the numerical simulations of the giant impact suggest that

most of the impactor's core directly merges with Earth's core, this seems implausible.

This scenario would also seem unlikely because the impactor’s mantle should have had a

highly radiogenic ε182W, given that the impactor most likely differentiated early.

Consequently, it is probable that the 182W similarity of the terrestrial and lunar mantles

reflects equilibration in the aftermath of the giant impact, unless the Moon predominantly

consists of terrestrial material.

Page 163: Touboul Thesis Completed

4.3.2.1.3 Isotope equilibration of Earth and Moon after the Giant impact

The identical 182W isotope composition of the lunar and terrestrial mantles

suggests either that the Moon mainly derives from material of the Earth's mantle or that

there has been isotope equilibration between the lunar magma disk and the Earth. The

first of these options appears inconsistent with results from numerical simulations of the

giant impact, all of which indicate that more than ~80% of the Moon is derived from the

impactor. For future simulations, it will be essential to determine whether there are

scenarios in which the Moon could be formed entirely from terrestrial material. In case of

an equilibration between the lunar magma disk and the Earth, one important question is

whether the equilibration process that has been proposed for O isotopes by Pahlevan and

Stevenson (2007) could also be valid for W. The condition for having isotope

equilibration between the Earth and the lunar magma disk is to reach a temperature in the

disk that will be high enough to allow volatilization of W. A key factor in this respect

will be the oxygen fugacity because, as pointed out by Fegley and Palme (1985), WO3 is

far more volatile than the reduced species WO2. If the temperature in the disk reaches

2000 K, then it is quite likely that W will be volatilized and this could ultimately lead to a

loss of W as has been argued for Mo by O’Neill (1991). Although O’Neill (1991) argues

that the case for W loss by volatility is difficult to make based on a thermodynamic

calculation, the temperature of the lunar magma disk might have been substantially

higher than the temperature of 1400 K assumed in the calculation of O'Neill (1991). For a

higher temperature a significant fraction of W might have been lost as WO3.

Volatilization of W would also facilitate isotope equilibration of W isotopes between

Earth and proto-Moon. An additional implication could be that the higher Hf/W of the

Moon (26.4±1.5 as opposed to 18±1.6) is due to volatile loss of W rather than core

formation.

In brief, while one cannot fully ascertain that there has been W isotope

equilibration between the Moon and the Earth at the time of the giant impact, the

mechanism described by Pahlevan and Stevenson (2007) could be plausible. More

detailed modelling will be needed to ascertain this process. It will also be essential to

evaluate in future dynamical models if the Moon could have been formed largely from

terrestrial mantle material.

Page 164: Touboul Thesis Completed

Figure 4.3.3: Degree of reequilibration as a function of the difference in ε182W before impact between the impactor and the proto-Earth for different times of giant impact (black curves). The possibilities of reequilibration scenarios are limited, as indicated by shaded area. First, the Moon obviously must be older than the oldest lunar rocks dated at about 140 Myr after the start of the solar system. Second, the giant impact should have occurred later than 37 Myr after the start of the solar system. Indeed, an earlier formation would required that the Moon initially had a subchondritic ε182W (grey area). Note that it is possible (but not required) to have subchondritic ε182W impactor above the 37 Ma isochron line. Open squares correspond to reequilibration scenarios exemplified by W isotope evolution diagrams shown in Fig. 4. Assuming Hf/W ratios of 26.5±1.1 (2σ) and 18±5.2 (2 σ) for bulk silicate Moon and Earth respectively, their W isotope compositions are strictly identical at 62 Myr, as in the model developed in Touboul et al. (2007). A giant impact at 62 Myr would require that any ε182W heterogeneity between impactor materials and the proto-Earth mantle has been entirely erased by equilibration (f=1). If the giant impact occurred later, the impactor mantle should be more radiogenic than the proto-Earth mantle (Δε182W<0 before equilibration) and the W isotope equilibration is partial. Inversely, a giant impact earlier than 62 Myr would require an impactor less radiogenic than the proto-Earth mantle (Δε182W>0 before equilibration).

164

Figure 4.3.4 (on the right): (A) Schematic diagram showing the various pools of W involved in the make up of the lunar and terrestrial mantle. The vertical double arrows show the show Δε182W before impact (large arrow) and Δε182W after equilibration is shown with the small arrow. (B-E) Time evolution diagrams of ε182W for the bulk silicate Moon and Earth, illustrating reequilibration scenarios as labeled in Fig. 4.3.3. Solid and long-dashed curves show evolutions of lunar and terrestrial mantles respectively, calculated using their present day ε182W combined with their Hf/W ratio. Solid vertical lines indicate reequilibration of W isotope at different times of giant impact. Equation (1) together with input parameters is used to set a Δε182W prior to equilibration. A complete model of impactor and proto-Earth and their equilibration would involve too many unknown parameters. Short-dashed curves give a hypothetical simplistic example of pre-giant impact ε182W evolution for impactor and proto-Earth reservoirs. The chondritic evolution (grey curve) is also shown for reference. (B) Unrealistic scenario with a giant impact at 32 Myr involving essentially materials from the impactor core. (C) Earliest possible scenario (37 Myr), which assumes an undifferentiated impactor (or a reequilibration between its core and its mantle) and no reequilibration between Earth and lunar materials in the aftermath of the giant impact. (D) Scenario with a giant impact at 45 Myr, which requires a terrestrial mantle initially slightly more radiogenic than the lunar mantle. This implies that the impactor mantle had a lower ε182W than the mantle of the proto-Earth prior partial equilibration. (E) Scenario with a giant impact at 62 Myr that requires complete W isotope re-equilibration between terrestrial mantle and proto-lunar materials to yield identical initial ε182W for the bulk silicate Moon and Earth. (F) Scenario with a giant impact at 88 Myr. The terrestrial mantle had to be slightly less radiogenic than the lunar mantle and this scenario therefore requires partial equilibration between the terrestrial mantle and a more radiogenic impactor mantle.

Page 165: Touboul Thesis Completed

165

Page 166: Touboul Thesis Completed

4.3.2.2 Chronology of terrestrial accretion and astronomical implications

An important implication of the new W isotope data for the Moon is that the age

of the Moon and the last episode of terrestrial accretion are younger than previously

thought (∼30 Myr). The new data of Touboul et al. (2007) indicates that the last stage of

Earth accretion occurred later than 50 Ma. As the age of the oldest lunar rocks is

approximately 112±40 Myr (Carlson and Lugmair 1988, Norman et al. 2003), one can

estimate the age of the giant impact and last stage of terrestrial accretion to be between

50 and 150 Myr. This estimate is compatible with earlier estimate for the age of the

Earth based on U-Pb but also on I-Xe systematics (Allègre et al. 1995, Ozima and

Podosek 1999). It has been argued by Harper and Jacobsen (1996) that the elevated

U/Pb ratio of Earth's mantle is due to loss of volatile Pb rather than removal of Pb into

the Earth's core. This was based on the comparison of the Pb abundance of Earth's

mantle to the depletion of lithophile elements with similar volatility. We have

reexamined this question by considering trends in chondrites between Rb and Pb, which

have similar condensation temperature in a gas of solar composition (Allègre et al.

2001). Based on Pb and Rb abundances in chondrites, the estimated 238U/204Pb ratio

ranges between 1 and 1.4 in the bulk Earth (Fig. 4.3.5). If no Pb is segregated into the

Earth’s core, then the 238U/204Pb ratio of the Bulk Silicate Earth (BSE) should be equal

to the 238U/204Pb ratio of the bulk Earth (BE). Yet, the BSE has a 238U/204Pb∼8-9, which

indicates that a significant part of the U-Pb fractionation is not only related to volatile

depletion but must be also to core segregation. As argued by Jacobsen and Harper

(1996), the continuous model age of core formation is almost identical to the two stage

model age for U-Pb, such that, if Pb was indeed partitioned into Earth's core, the U-Pb

model age of ~70-100 Ma should closely reflect the time of terrestrial core formation.

This is consistent with the new constraints on the age of the Giant impact based on W

isotopes. Based on the observation that the Hf-W and U-Pb model ages for formation of

the Earth's core are different, Wood and Halliday (2005) argued that Pb was removed

into Earth's core by late sulfide segregation. The new W isotope data for lunar rocks

however reveal that there must have been ongoing core formation until after 50 Ma,

such that there is not contradiction between the U-Pb and Hf-W systems.

166

Page 167: Touboul Thesis Completed

Figure 4.3.5: Derivation of a bulk Earth 238U/204Pb ratio based on Pb-Rb systematics for meteorite data. Meteorite data is from Wasson and Kayllemen (1988). Data from primitive mantle is from Hofmann (1988). The Pb/Rb versus Rb data for meteorite defines a negative trend reflecting the greater volatility of Rb relative to Pb. The Bulk Earth composition is marked as BE. If one assumes that Pb and Rb are not incorporated in the core, the bulk silicate earth composition should lie on the trend (once Rb is corrected for the mass of the core). The observed value of BSE plots well below this trend and that Pb/Rbpredicted/Pb/Rbobserved defines the degree of depletion of Pb due to incorporation of Pb in the core. One can then infer the degree of depletion of Pb due to core formation.

The W results also have important astronomical implications for the dynamics of

the solar system. As shown by the models of O’Brien et al. (2006), an early termination

of terrestrial accretion (<30 Myr) requires an excentric orbit for Jupiter and Saturn (EJS

model). However, this is not the preferred model of O’Brien et al. (2006) because this

initial condition leads to a rather complete clean up of water-rich planetesimals in the

region that is supposed to deliver water to the Earth at a late stage. Their preferred

model is to have a circular orbit for Jupiter and Saturn (CJS model). Remarkably, this

model leads to a later termination for Earth's accretion (> 70 Myr) and allows the

delivery of water to the Earth. It is also worth pointing out that an initially circular orbit

is easier to justify from a dynamical viewpoint (Morbidelli 2008).

167

Page 168: Touboul Thesis Completed

4.3.3 Constraints on the early differentiation of the Earth

The early differentiation of the Earth’s mantle is currently a highly debated topic

and a recent overview is given in Bourdon and Caro (2007). There is now clear

evidence for an early terrestrial mantle differentiation event (Caro et al., 2003) and the

remaining questions are related to the exact timing of this differentiation as well as the

composition of the bulk Earth. The recent analyses of 142Nd in meteorites (Boyet and

Carlson, 2005) have revealed that (1) the composition of the Earth is distinct from that

of meteorites and (2) that the two main classes of chondrites (ordinary and

carbonaceous chondrites) also have a distinct 142Nd isotope composition despite having

almost similar Sm/Nd ratios. This latter observation can apparently be rationalized by a

difference in the abundance of r- and s-process nuclides (Rasmusen and Sharma 2007,

Carlson et al. 2007). If we assume that the bulk silicate Earth has a chondritic

composition in Nd isotopes, then the positive ε142Nd measured in terrestrial rocks

relative to ordinary chondrites can be explained by early segregation of an enriched

reservoir (with a low Sm/Nd) that would have led to a higher Sm/Nd in the remaining

portion of the mantle. In order to produce a positive 142Nd anomaly, this differentiation

would have had to take place early enough in Earth’s history. In what follows, we

examine the consistency of the W isotope constraints on the timing of the formation of

the Moon inferred by Touboul et al. (2007) and the hypothesis of an early mantle

differentiation of Boyet and Carlson (2005).

4.3.3.1 The Age of Early Earth differentiation

As argued by Boyet and Carlson (2005), if the Earth has a chondritic Nd isotope

composition, then the difference in 142Nd between Earth and chondrites yields an age for

mantle differentiation of less than 30 Myr. The actual age of differentiation given by

this system is likely to be significantly less than 30 Myr. Boyet and Carlson (2005) have

used the ε143Nd in MORBs to estimate the maximum degree of Sm/Nd fractionation in

the early Earth. As shown in their Fig. 4.3.3, if the Early depleted mantle (EDM) had

formed later than 30 Myr, the mean ε143Nd of the observable silicate Earth (=EDM)

should be greater than 10.5. We know that this is not the case and this estimate is

therefore an upper limit, as clearly stated by Boyet and Carlson (2005). An additional

168

Page 169: Touboul Thesis Completed

reason why this is an upper limit include that the effects of later continental crust

extraction are ignored (as is the OIB reservoir that also has a lower Sm/Nd on average

than the MORB reservoir). Here, we attempt to refine this estimate by taking into

account the fact that the continental crust has to be be included for estimating the

composition of the EDM. A conceptual model for this calculation is illustrated in Fig.

4.3.6a. If one takes into account the Nd in the continental crust (CC), one can calculate a

new age for mantle depletion based on the following equations:

( )

( )

147

147

143 143 147

144 144 144

143 143 147

144 144 144

143 143

144 144

1

1

cc

cc

today Tcc todayT

EDM EDM EDMtoday Tcc today

T

CC EDM CC

today

DM

Nd Nd Sm eNd Nd Nd

Nd Nd Sm eNd Nd Nd

Nd NdNd Nd

λ

λ

⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞=⎜ ⎟ ⎜

⎝ ⎠ ⎝( )

( )

147

147 147

147

144

143 143 147

144 144 144

1cc

CC EDM

EDM cc

Tcc todayT

EDM DMT T today

T T

EDM EDM EDM

Sm eNd

Nd Nd Sm e eNd Nd Nd

λ

λ λ

⎛ ⎞+ −⎟ ⎜ ⎟

⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠

( )( )146 0

146

0142 142 144 146

144 144 144 144

0 0142 142 144 146

144 144 144 144 1

EDM

EDM

EDM

today TT T

EDM CHUR EDM CHUR

TT

CHUR CHUR CHUR CHUR

Nd Nd Sm Sm eNd Nd Nd Sm

Nd Nd Sm Sm eNd Nd Nd Sm

λ

λ

− −

⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞= +⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠( )( )

( )

( )

0

147 147 147

144 144 144

143 143 143

144 144 144

1

1

EDMT

DM DMEDM DM CC

DM DMEDM DM CC

Sm Sm SmNd Nd Nd

Nd Nd NdNd Nd Nd

α α

α α

⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞ ⎛ ⎞= + −⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ ⎠ ⎝ ⎠ ⎝ ⎠

The values of input parameters are given in the caption of Fig. 4.3.6. We assume

a chondritic evolution prior to the formation of the EDM reservoir. These eight

equations are non-linear and can be solved iteratively for the eight unknowns assuming

starting input parameters TEDM and αDM, the fraction of Nd in the depleted mantle. This

latter parameter is used to calculate the concentration of Nd in the depleted mantle. The

main results of this calculation is that given that the concentration of Nd in the depleted

mantle should be around 1 ppm at most (Hofmann 1988, Salters and Stracke 2004,

Workman and Hart, 2005), the age of early mantle depletion has to be less than 10 Ma

169

Page 170: Touboul Thesis Completed

after CAI (Fig. 4.3.6b). Essentially, this means that the formation of a hidden reservoir

complementary to the EDM has to be formed before the Earth is completely accreted,

which is unrealistic. This conclusion is very similar to what is discussed below in

section B.2 based on observations of 142Nd in Martian meteorites. A major conclusion is

that it is more plausible to have the loss of enriched material at prior to complete

accretion rather than due to magma ocean differentiation as argued by Boyet and

Carlson (2005). Having an early mantle depletion at 20-30 Myr would require a

prohibitively high Nd concentration in the depleted mantle.

Figure 4.3.6: (a) Conceptual model used to calculate the age of terrestrial differentiation based on coupled 147Sm-144Nd and 146Sm-142Nd systematics of the Earth as inferred. EDM is the early depleted mantle. The original estimate of Boyet and Carlson (2005) assuming a chondritic Earth composition ignored the effect of later crustal extraction and obtained a maximum age of differentiation. The age of the formation of an early depleted mantle is TEDM, while the age of crustal extraction is TCC. (b) Age of early mantle depletion based on mass balance equations and evolution of 142Nd and 143Nd in the bulk silicate Earth, depleted mantle and continental crust. The composition of the EDM was calculated by taking into account the budget of Nd in the continental crust (Nd=16 ppm). The equations are solved iteratively as explained in the main text with the following input parameters: ε142NdEDM=+0.2, ε143NdDM=+11 and ε143NdDM=-20. 147Sm/144Ndchondrite=0.1967, 147Sm/144NdCC=0.115. We made no assumption about the 147Sm/144Nd of the depleted MORB mantle, as the values given by Salters and Stracke (2004) and Workman and Hart (2005) are based on a chondritic model for the Earth. The grey band indicates plausible values for the Nd concentration in the depleted mantle (∼1 ppm). We assumed for this calculation the maximum possible mass for the depleted mantle, i.e. the total mass of the mantle. A lower value will push TEDM to older ages. The arrow indicates the determination of Boyet and Carlson (2005) for the early mantle depletion. This solution, which did not take into account the continental crust, is not viable.

170

Page 171: Touboul Thesis Completed

Even if we ignore this mass balance calculation and take the estimate of Boyet

and Carlson (2005) for the differentiation of the early Earth, it is quite clear that this

early event (<30 Ma) should have taken place prior to the Moon-forming giant impact

(>50 Myr, see above). If that earlier event resulted itself from another giant impact (as is

most likely in order to produce large scale mantle differentiation), then the Earth should

have remained quite hot just prior to the Moon-forming impact (Tonks and Melosh,

1993). In this case, the calculations of Tonks and Melosh (1993) show quite clearly that

the entire Earth must have been molten during the giant impact. It is likely that the giant

impact would have violently disturbed an early segregation. First, much of the

impactor’s core most likely has merged with the terrestrial core either as a single event

or by Rayleigh-Taylor instabilities, and this must have seriously disrupted any dense

layer at the base of the mantle. Second, the heat provided by the segregation of the

impactor’s core is likely to strengthen convection and buoyancy driven instabilities. All

this would appear inconsistent with the idea that the preservation of an earlier enriched

reservoir stored at the bottom of the lower mantle, unless the viscosity and density

contrasts were large enough to avoid remixing of the enriched dense layer. Obviously,

this point will need to be explored more thoroughly with dynamical models. Yet the

new timing for the formation of the Moon does not lend strong support for the very

early storage of an enriched reservoir. In addition, our mass balance calculation on the

Sm-Nd budget casts some serious doubts on the plausibility of a hidden reservoir. In

what follows, we examine this point further and propose an alternative model.

4.3.3.2 The Nd isotope composition of terrestrial planets

In this section, we examine whether one can assume that the bulk composition of

planets in Sm-Nd is chondritic on the basis of new Nd isotope data for Martian

meteorites (Caro et al. 2008). The new data of Caro et al. (2008) shown in Figure 4.3.7

in a ε142Nd versus 147Sm/144Nd diagram clearly shows that the Martian bulk composition

is probably not chondritic. It must be noted that enriched shergottites have an ε142Nd

equal to that of ordinary chondrites but they have lower 147Sm/144Nd ratios and their

ε143Nd and ε176Hf are both negative (-7 and -13 respectively), which is a typical crustal

171

Page 172: Touboul Thesis Completed

signature. On this basis, it would seem that ε142Nd ∼-20 ppm cannot be representative of

a bulk composition for Mars.

Figure 4.3.7: (a) Sm-Nd planetary isochron for Mars, Moon and Earth showing a common composition for all these inner solar system planetary objects. This suggests it could represent the bulk composition of inner solar planets. The figure shows possible scenarios for obtaining planetary isochrons in Mars starting from a chondritic composition. We assume that the bulk Mars composition lies on the planetary isochron formed by shergottites (Caro et al. 2008). If the bulk Sm/Nd of Mars is smaller than the intersect with Earth (line A), then the age of Mars-chondrite fractionation is earlier than the age of the solar system. If the bulk Sm/Nd of Mars is higher than the intersect with Earth then the age of Sm/Nd fractionation is less than 40 Ma. This scenario while being possible would still require a super-chondritic composition for Mars. The common intersect for Mars, Moon and Earth would have to be purely coincidental. (b) Hypothetical scenario for 146Sm-142Nd systematics for the chondritic Moon composition assuming that the 142Nd isotopes equilibrate with the terrestrial composition at the time of the giant impact taken to be 70 Ma. Crystallization of a magma ocean (and later evolution) then produces a fractionation with a model age of 200 Ma. In this scenario that is consistent with W isotopes, the intersection with terrestrial composition should have a much higher Sm/Nd than what is observed. This would also be inconsistent with the data of Rankenburg et al. (2006).

If the intersection of the mantle compositions of Mars, Moon and Earth

represents the mean composition of terrestrial planets, one would infer that the

composition of these planets is superchondritic in Sm/Nd, unless Mars has also

experienced an early episode of mantle differentiation leading to a hidden reservoir with

a composition identical to that of the Earth. Caro et al. (2008) have argued that this is an

unlikely possibility. To make this case more convincing, we explore more

systematically scenarios whereby the common point for terrestrial planets shown on

Figure 4.3.7 can be explained by an early differentiation event as argued for the Earth

by Boyet and Carlson (2005). Unlike Debaille et al. (2007), we assume that that the bulk

172

Page 173: Touboul Thesis Completed

composition of Mars lies on the planetary isochron, implying that there is no hidden

reservoir on Mars. One question is whether the intersect with the Earth is likely to

represent the bulk composition of Mars. If the bulk composition of Mars lies above the

intersection with terrestrial composition (Fig. 4.3.7a), then the chondrite-Mars Sm-Nd

fractionation event would have to post-date the accretion of Mars dated at 0-10 Myr as

shown by Kleine et al. (2002) and Nimmo and Kleine (2007). This putative event would

have to pre-date the large scale mantle differentiation event dated at 40 Myr (Caro et al.

2008), which seems unreasonable. If the bulk composition of Mars is below the

terrestrial composition, then the fractionation in Sm/Nd must be older than the solar

system, which is impossible. Thus, in both cases, it seems likely that the intersect with

the terrestrial composition represents the bulk Martian composition (Fig. 4.3.7a).

Thus, one may conclude that the observations of Caro et al. (2008) imply that

the Sm/Nd fractionation observed in Martian meteorites must have taken place either

before accretion or right after accretion. However, if one calls upon the segregation of a

hidden reservoir, the Sm/Nd fractionation must have been strikingly identical to that

observed in the Earth. Although there has not been a physically plausible scenario

proposed by Boyet and Carlson (2005) to explain their hidden reservoir, it is not clear

why the hidden reservoir on Earth would lead to identical Sm/Nd in the Martian mantle.

A different line of argument has been used by Debaille et al. (2007) who have

argued that the linear array formed by shergottites in an ε142Nd versus ε143Nd diagram

cannot be explained by a super-chondritic mantle identical to that of the Earth. For their

argument they have assumed that the ε143Nd for bulk Mars would be the same as the

depleted terrestrial mantle (i.e. ε143Nd∼10.7, which is the value used by Boyet and

Carlson (2005). However, the choice of this particular composition for Earth (or Mars)

is not justified since we know that both magma ocean crystallization and later crustal

extraction has modified the 143Nd isotope composition while 142Nd is only affected by

the earlier event. As a consequence the argument of Debaille is not conclusive and an

alternative solution is discussed below (see section 4.3.2).

The 146Sm-142Nd systematics for lunar rocks are also difficult to reconcile with a

hidden reservoir model that would be consistent with both Nd and W observations. A

reasonable scenario would be that the Moon formed at 50 Myr (the earliest age possible

based on W isotopes) with a bulk chondritic composition and that the Nd isotopes

173

Page 174: Touboul Thesis Completed

would equilibrate during the proto-Moon stage as has been argued for O, Si and W

isotopes (Pahlevan and Stevenson 2007, Georg et al. 2007; Touboul et al. 2007). As

shown in Figure 4.3.7b, the intersection of the lunar mantle isochron with the terrestrial

Nd isotope composition would fall at a higher Sm/Nd than is observed. Thus, this

scenario cannot explain the observations. The only realistic explanation is then to have

the Moon formed entirely from the Earth’s mantle but this scenario would then be in

conflict with inferences about the iron abundance in the Moon (O’Neill 1991). It would

also contradict dynamical models that predicts that the Moon is formed mostly from the

impactor (Canup and Asphaug 2001), although more recent models have shown that the

mass fraction of proto-Earth in the Moon could be as large as 0.5 (Canup 2004).

One may note that there is a substantial uncertainty in the intersection of Mars,

Moon and Earth in the Sm-Nd diagram. In principle, the 147Sm/144Nd value required to

produce a 20 ppm offset in 142Nd between chondrites and planets is 0.21, which is

slightly higher than the value given by the intersect. This discrepancy could come from

several sources of uncertainties. First, one may note that the 142Nd of carbonaceous

chondrites is obtained based on a correction using 148Nd isotope (Carlson et al. 2007),

which introduces an additional source of uncertainty. If one considers only the

difference between planets and ordinary chondrites then the difference in 142Nd

becomes 17 ppm. Another source of uncertainty is the initial 146Sm/144Sm ratio

(=0.008±1) used to determine the slope of the 4.568 Ga isochron in the 146Sm-142Nd

diagram. This slope is slightly steeper if 146Sm/144Sm is increased to 0.009 and the 147Sm/144Nd needs to be increased to 0.207 to match a difference of ∼17 ppm in between

ordinary chondrites and Earth.

An important implication is that the Sm/Nd fractionation observed in Mars,

Earth and the Moon would predate the accretion and the high Sm/Nd would characterize

the chemical composition of the inner solar system. As there is no direct meteorite

sample from this inner solar system region, it is difficult to test that inference. Because

Sm and Nd are both lithophile and refractory, the expectation is that the Sm/Nd ratio of

the bulk planets should be identical to that of chondrites. There are reports of REE

fractionation under specific nebular conditions. One could consider the following two

hypotheses: (1) given that chondrules have systematically higher Sm/Nd than bulk

chondrites (Amelin 2004; Krestina et al. 1999), a preferential accumulation of

174

Page 175: Touboul Thesis Completed

chondrules in the inner solar system region should in principle lead to a higher Sm/Nd

in terrestrial planets; (2) as the Earth, Mars and possibly Moon-forming impactor

formed from precursory planetesimals, the impact history of these differentiated object

could have led to removal of a crust with a low Sm/Nd.

The observation that each chondrite group has a characteristic size distribution

of chondrules has led to the hypothesis that chondrules have been sorted by a dynamical

process (Kuebler et al. 1999). Furthermore, modelling of turbulence in the solar nebula

has also shown that the abundances of chondrules themselves could be enhanced by the

process of turbulent concentration (Cuzzi et al. 1996, 2001). More importantly, the

prediction is that chondrules should be concentrated in the terrestrial planet regions

(Cuzzi et al. 1998), while porous aggregates would be concentrated in the outer planet

regions. In essence, this is exactly what our inference would be based on the 146Sm-142Nd systematics, suggesting that the offset in Sm/Nd ratio in terrestrial planets could

be due to an accumulation of chondrules with a higher Sm/Nd. In practical terms, the

calculated concentration of chondrules can increase by a factor of ten with a high

probability, Cuzzi et al. 2001). Using the histogram shown on Figure 4.3.8, this is

enough to shift the 147Sm/144Nd from 0.1967 to 0.206.

One important question that arises is what process would have caused the

Sm/Nd to be greater than chondritic in the first place. This represents a shift of only

4.5% and several processes can a priori be proposed. From a cosmochemical viewpoint,

the relative volatility of rare earths has been explored in variable conditions including

oxidizing and reducing environment (Boynton 1977, Lodders and Fegley 1993, Pack et

al. 2004). In general, there can be an enhancement of Sm or Nd volatility in oxidizing

and reducing conditions but the expectation is that other REE will then be far more

fractionated (i.e. Eu, Ce or Yb) and this extreme fractionation is not observed in

terrestrial mantle peridotites (Jagoutz et al. 1979). Thus, the fractionation due to

volatility, although it has been observed in CAIs and special chondrules (Pack et al.

2004) does not seem to be the process to explain the difference between bulk chondrites

and chondrules.

175

Page 176: Touboul Thesis Completed

Figure 4.3.8: Histogram showing the 147Sm/144Nd for bulk chondrites and chondrules (adapted from Caro et al. 2008). The high Sm/Nd found in chondrules could be a result of REE redistribution during the formation of low Sm/Nd phosphate but the Sm/Nd does not correlate with degree of metamorphism. Further investigations of Sm-Nd systematics in chondrites is required to prove that the concentration of chondrules can effectively explain the high Sm/Nd in inner planets. The mean chondrule 147Sm/144Nd is 0.21, while the bulk chondrite 147Sm/144Nd is 0.1967.

There are only few high precision REE data set on chondrules (Krestina et al.

1999, Amelin 2004) and these studies do not reveal the origin of the fractionation. High

Sm/Nd ratios are observed in ordinary chondrites that also have a low Sm/Nd phosphate

phase. If one assumes that the REE were redistributed during the formation of the

phosphate on the parent body, then the high Sm/Nd ratio in the chondrule would not be

a primary feature and thus would not be relevant to explaining the high Sm/Nd of

planets. It is worth noting, however, that chondrules in carbonaceous chondrites that do

not have a phosphate phase also have a high Sm/Nd ratio. Furthermore, as shown in Fig.

8, the chondrules of one of the least metamorphosed ordinary chondrites (Krymka, LL

3.1) also have a high Sm/Nd ratios, such that this observation is not linked to thermal

processing. It is worth pointing out, however, that recent studies of type 1 chondrules

(Libourel et al. 2007) that include relicts of planetesimal mantles could give clues on the

176

Page 177: Touboul Thesis Completed

observed Sm/Nd fractionation. Although there are very few of these chondrules,

perhaps because many chondrules have been entirely molten, these observations provide

important insights about the origin of chondrules. If the chondrules were part of

differentiated parent bodies, then it is likely that they experienced melt extraction

leading to depletion in Nd relative to Sm. One could speculate that these observations

could explain the higher Sm/Nd in chondrules but further documentation will need to be

given in the future. Clearly, there are potential difficulties in explaining the high Sm/Nd

in Earth, Mars and Moon by accumulation of chondrules and further investigations are

needed to test this hypothesis.

The second hypothesis involving impact erosion of a differentiated crust was

also proposed by Caro et al. (2008) and is discussed at greater length by O’Neill in this

volume. Numerous models (e.g. Benz and Asphaug 1999, Agnor and Asphaug 2004,

Asphaug et al. 2006) have reported that collisions are commonly non-accretionary and

lead to erosion of the outer layers of a growing planet. The inference that one might

make here is that this process should have taken place at a very early stage (within the

first few Ma after the beginning of the solar system) and have led to removal of an early

formed crust on planetesimals that ultimately formed Earth, Mars and Moon. One

additional constraints brought by the 142Nd data is that the age and composition of the

removed crust would have be similar on the precursory material of Earth, Mars and

Moon (Fig. 4.3.9). This requirement would seem easier to fulfill at an early stage (when

bodies have not grown significantly yet) rather than at a late stage once the planets are

formed (see above). Our present state of knowledge does not allow us to conclude on

the actual process that would have led to a chemical depletion in the bulk composition

of the Earth, Mars and Moon.

177

Page 178: Touboul Thesis Completed

Figure 4.3.9: Model for the Nd (top) and Hf (bottom) evolution of Western Greenland rocks (Isua). In order to explain the chondritic Hf isotope in Isua at 3.8 Ga, one needs to consider a three-stage model. First, the Earth’s mantle is depleted which leads to elevated Lu/Hf and Sm/Nd ratios. Then the precursor of Isua rocks is isolated into the crust at an age where 146Sm is almost extinct (e.g. 4.2 Ga), then the Lu/Hf is further fractionated due to intra-crustal refinement at 3.8 Ga. The system would have remained closed until today. This scenario is consistent with a scenario proposed by Kamber et al. (2003) to explain the Pb isotope systematics of Isua rocks.

4.3.4 Conclusions

New 182W data for the Moon provide new constraints on the timing of the Moon-

forming giant impact (later than 50 Myr). Based on this data one can also estimate that

the solidification of the magma ocean took place later than 60 Myr. This solves the

apparent conflict with 142Nd age constraints for differentiation of the lunar magma

ocean. However, because the 146Sm-142Nd is sensitive to fractionation until 4.1-4.2 Gyr,

the relatively young age given by this system (200-250 Myr) for magma ocean

crystallization probably reflects later events including melting, mixing, convection and

contamination. A simple model including these processes shows that this hypothesis is

perfectly plausible.

A wide exploration of models for the giant impact and the subsequent W isotope

evolution suggests that there must have been some degree of W isotope equilibration

178

Page 179: Touboul Thesis Completed

between the Earth and the Moon following the impact. Most likely, part of the core of

the impactor partially reequilibrated with the Earth and Moon mantle.

We have shown based on Sm-Nd mass balance that the hidden reservoir scenario

proposed by Boyet and Carlson (2005) based on the difference in 142Nd abundance

between the Earth and chondrites needs to be questioned. This Sm/Nd fractionation

associated with the formation of a hidden reservoir has to take place prior to the

accretion of the Earth to match the Nd isotope observations. Based on new data for

Martian meteorites, we further hypothesize that the Sm/Nd composition of terrestrial

planets could be higher than that of chondrites. This process could result from

enrichment in a chondrule component by sorting or alternatively by impact erosion the

outer layers of planetesimals forming the Earth and Mars. Deciding between these

alternatives will require deeper investigations of both processes.

Acknowledgements

We would like to thank Stein Jacobsen, Rick Carlson, Hugh O’Neill, Charlie

Langmuir, Alessandro Morbidelli, Alan Brandon, Dave Stevenson, Claude Allègre for

discussions prior to or during the preparation of this manuscript. We also thank one

anonymous reviewer for his constructive and insightful remarks as well as the patient

editorial handling by Andrew Jephcoate.

References

Agnor, C. & Asphaug, E., 2004. Accretion efficiency during planetary collisions.

Astrophys. J. 613, L157–L160.

Albarède, F. & Blichert-Toft, J., 2007. The split fate of the early Earth, Mars, Venus,

and Moon. C. R. Geosci. 339, 917–927.

Allègre, C. J., Manhés, G. & Göpel, C., 1995. The age of the Earth. Geochim.

Cosmochim. Acta 59, 1445–1456.

Allègre, C., Manhès, G. & Lewin, E., 2001. Chemical composition of the Earth and the

volatility control on planetary genetics. Earth Planet. Sci. Lett. 185, 49–69.

Amelin, Y. & Rotenberg, E., 2004. Sm–Nd systematics of chondrites. Earth Planet. Sci.

Lett. 223, 267–282.

179

Page 180: Touboul Thesis Completed

Andreasen, R. & Sharma, M., 2006. Solar nebula heterogeneity in p-process samarium

and neodymium isotopes. Science 314, 806–809. (doi:10.1126/science.1131708)

Asphaug, E., Agnor, C. B. & Williams, Q., 2006. Hit-and-run planetary collisions.

Nature 439, 155–160.

Bennett, V. C., Brandon, A. D. & Nutman, A. P., 2007. Coupled 142Nd–143Nd isotopic

evidence for Hadean mantle dynamics. Science 318, 1907–1910.

Benz, W. & Asphaug, E., 1999. Catastrophic disruptions revisited. Icarus 142, 5–20.

Bourdon, B. & Caro, G., 2007. The early terrestrial crust. C. R. Geosci. 339, 928–936.

Boyet, M. & Carlson, R. W., 2005. 142Nd evidence for early (O4.53 Ga) global

differentiation of the silicate Earth. Science 309, 576–581.

Boyet, M. & Carlson, R. W., 2007. A highly depleted moon or a non-magma ocean

origin for the lunar crust? Earth Planet. Sci. Lett. 262, 505–516.

Boynton, W. V., 1975. Fractionation in the solar nebula: condensation of yttrium and

the rare earth elements. Geochim. Cosmochim. Acta 39, 569–584.

Canup, R. M., 2004. Simulations of a late lunar-forming impact. Icarus 168, 433–456.

Canup, R. M. & Asphaug, E., 2001. Origin of the Moon in a giant impact near the end

of the Earth’s formation. Nature 412, 708–712.

Carlson, R. W. & Lugmair, G. W., 1988. The age of ferroan anorthosite 60025: oldest

crust on a young Moon. Earth Planet. Sci. Lett. 90, 119–130.

Carlson, R. W., Boyet, M. & Horan, M., 2007. Chondrite barium, neodymium, and

samarium isotopic heterogeneity and early Earth differentiation. Science 316,

1175–1178.

Caro, G., Bourdon, B., Birck, J. L. & Moorbath, S., 2003. 146Sm–142Nd evidence from

Isua metamorphosed sediments for early differentiation of the Earth’s mantle.

Nature 423, 428–432.

Caro, G., Bourdon, B., Halliday, A. N. & Quitté, G., 2008. Super-chondritic Sm/Nd in

Mars, Earth, and the Moon. Nature 452, 336–339.

Cuzzi, J. N., Hogan, R. C. & Dobrovolskis, A. R., 1998. Turbulent concentration:

fractal description, scaling laws and generalized applications to planetesimal

accretion. In Proc. 29th Lunar and Planetary Science Conf., Lunar and Planetary

Institute, Houston, TX, p. 1443.

180

Page 181: Touboul Thesis Completed

Cuzzi, J. N., Hogan, R. C., Paque, J. M. & Dobrovolskis, A. R., 2001. Size-selective

concentration of chondrules and other small particles in protoplanetary nebula

turbulence. Astrophys. J. 546, 496–508.

Davaille, A., Girard, F. & Le Bars, M., 2002. How to anchor hotspots in a convecting

mantle? Earth Planet. Sci. Lett. 203, 621–634.

Debaille, V., Brandon, A. D., Yin, Q. Z. & Jacobsen, B., 2007. Coupled 142Nd–143Nd

evidence for a protracted magma ocean in Mars. Nature 450, 525–528.

Fegley, B. & Palme, H., 1985. Evidence for oxidizing conditions in the solar nebula

from Mo and W depletions in refractory inclusions in carbonaceous chondrites.

Earth Planet Sci. Lett. 72, 311–326.

Georg, R. B., Halliday, A. N., Schauble, E. A. & Reynolds, B. C., 2007. Silicon in the

Earth’s core. Nature 447, 1102–1106.

Halliday, A. N., 2004. Mixing, volatile loss and compositional change during impact-

driven accretion of the Earth. Nature 427, 505–509.

Harper Jr, C. L. & Jacobsen, S. B., 1996. Evidence for 182Hf in the early Solar System

and constraints on the timescale of terrestrial accretion and core formation.

Geochim. Cosmochim. Acta 60, 1131–1153.

Hofmann, A. W., 1988. Chemical differentiation of the Earth: the relationship between

mantle, continental crust, and oceanic crust. Earth Planet. Sci. Lett. 90, 297–314.

Jacobsen, S. B. & Harper, C., 1996. Accretion and early differentiation history of the

Earth based on extinct radionuclides. In Earth processes: reading the isotope code

(eds A. R. Basu & S. R. Hart). Geophysical Monograph, no. 95, pp. 47–74.

Washington, DC: American Geophysical Union.

Jagoutz, E., Palme, H., Baddenhausen, H., Blum, K., Cendales, M., Dreibus, G., Spettel,

B., Waenke, H. & Lorenz, V., 1979. The abundances of major, minor and trace

elements in the Earth’s mantle as derived from primitive ultramafic nodules. In

Proc. 10th Lunar and Planetary Science Conf., Lunar and Planetary Institute,

Houston, TX, pp. 2031–2050. New York, NY: Pergamon.

Kellogg, J. B., Jacobsen, S. B. & O’Connell, R. J., 2002. Modeling the distribution of

isotopic ratios in geochemical reservoirs. Earth Planet. Sci. Lett. 204, 183–202.

181

Page 182: Touboul Thesis Completed

Kleine, T., Mu¨nker, C., Mezger, K. & Palme, H., 2002. Rapid accretion and early core

formation on asteroids and the terrestrial planets from Hf–W chronometry. Nature

418, 952–955.

Kleine, T., Mezger, K., Palme, H. & Münker, C., 2004. The W isotope evolution of the

bulk silicate Earth: constraints on the timing and mechanisms of core formation

and accretion. Earth Planet. Sci. Lett. 228, 109–123.

Kleine, T., Palme, H., Mezger, M. & Halliday, A. N., 2005. Hf–W chronometry of

lunar metals and the age and early differentiation of the Moon. Science 310, 1671–

1674.

Krestina, N., Jagoutz, E. & Kurat, G., 1999. The interrelation between core and rim of

individual chondrules from the different meteorites in term of Sm–Nd isotopic

system. In Proc. 30th Lunar and Planetary Science Conf., Lunar and Planetary

Institute, Houston, TX, p. 1918.

Kuebler, K. E., McSween, H. Y., Carlson, W. D. & Hirsch, D., 1999. Sizes and masses

of chondrules and metal–troilite grains in ordinary chondrites: possible

implications for nebular sorting. Icarus 141, 96–106.

Labrosse, S., Hernlund, J. W. & Coltice, N., 2007. A crystallizing dense magma ocean

at the base of the Earth’s mantle. Nature 450, 866–869.

Lee, D. C., Halliday, A. N., Snyder, G. A. & Taylor, L. A., 1997. Age and origin of the

Moon. Science 278, 1098–1103.

Lee, D. C., Halliday, A. N., Leya, I., Wieler, R. & Wiechert, U., 2002. Cosmogenic

tungsten and the origin and earliest differentiation of the Moon. Earth Planet. Sci.

Lett. 198, 267–274.

Leya, I., Wieler, R. & Halliday, A. N., 2000. Cosmic-ray production of tungsten

isotopes in lunar samples and meteorites and its implications for Hf–W

cosmochemistry. Earth Planet. Sci. Lett. 175, 1–12.

Libourel, G. & Krot, A. N., 2007. Evidence for the presence of planetesimal material

among the precursors of magnesian chondrules of nebular origin. Earth Planet Sci.

Lett. 254, 1–8.

Lodders, K. & Fegley Jr, B., 1993. Lanthanide and actinide chemistry at high C/O ratios

in the solar nebula. Earth Planet Sci. Lett. 117, 125–145.

182

Page 183: Touboul Thesis Completed

Longhi, J., 2003. A new view of lunar ferroan anorthosites: postmagma ocean

petrogenesis. J. Geophys. Res. 108, 5083.

Martin, H., Albarède, F., Claeys, P., Gargaud, M., Marty, B., Morbidelli, A. & Pinti, D.

L., 2006. Building of a habitable planet. Earth Moon Planets 98, 97–151.

McDonough, W. F. & Sun, S.-s., 1995. The composition of the Earth. Chem. Geol. 12,

223–253.

Nimmo, F. & Agnor, C. B., 2006. Isotopic outcomes of N-body accretion simulations:

constraints on equilibration processes during large impacts from Hf/W

observations. Earth Planet. Sci. Lett. 243, 26–43.

Nimmo, F. & Kleine, T., 2007. How rapidly did Mars accrete? Uncertainties in the Hf–

W timing of core formation. Icarus 191, 497–504.

Norman, M. D., Borg, L. E., Nyquist, L. E. & Bogard, D. D., 2003. Chronology,

geochemistry, and petrology of a ferroan noritic anorthosite clast from Descartes

breccia 67215: clues to the age, origin, structure, and impact history of the lunar

crust. Meteorit. Planet. Sci. 38, 645–661.

O’Brien, D. P., Morbidelli, A. & Levison, H. F., 2006. Terrestrial planet formation with

strong dynamical friction. Icarus 184, 39–58.

O’Neill, H., 1991. The origin of the Moon and the early history of the Earth - a

chemical model. 1 The Moon. Geochim. Cosmochim. Acta 55, 1135–1157.

O’Neill, H. St. C. & Palme, H., In press. Collisional erosion and the non-chondritic

composition of the terrestrial planets. Phil. Trans. R. Soc. A 366.

Ozima, M. & Podosek, F. A., 1999. Formation age of the Earth from 129I/127I and 244Pu/238U systematics and the missing Xe. J. Geophys. Res. 104, 25 493–25 499.

Pack, A., Shelley, J. M. & Palme, H., 2004. Chondrules with peculiar REE patterns:

implications for solar nebular condensation at high C/O. Science 303, 997–1000.

Pahlevan, K. & Stevenson, D. J., 2007. Equilibration in the aftermath of the lunar-

forming giant impact. Earth Planet. Sci. Lett. 262, 438–444.

Rankenburg, K., Brandon, A. D. & Neal, C. R., 2006. Neodymium isotope evidence for

a chondritic composition of the Moon. Science 312, 1369–1372.

Righter, K. & Shearer, C. K., 2003. Magmatic fractionation of Hf and W: constraints on

the timing of core formation and differentiation in the Moon and Mars. Geochim.

Cosmochim. Acta 67, 2497–2507.

183

Page 184: Touboul Thesis Completed

Rudge, J. F., 2006. Mantle pseudo-isochrons revisited. Earth Planet. Sci. Lett. 249, 494–

513.

Salters, V. J. M. & Stracke, A., 2004. Composition of the depleted mantle. Geochem.

Geophys. Geosyst. 5, Q05B07.

Shearer, C. K. & Newsom, H. E., 2000. W–Hf isotope abundances and the early origin

and evolution of the Earth–Moon system. Geochim. Cosmochim. Acta 64, 3599–

3613.

Solomatov, V. S., 2000. Fluid dynamics of a terrestrial magma ocean. In Origin of the

Earth and Moon (eds R. M. Canup & K. Righter), pp. 323–338. Tucson, AZ:

University of Arizona Press.

Solomatov, V. S., 2007. Magma oceans and primordial mantle differentiation. In

Evolution of the Earth (ed. D. Stevenson). Treatise on geophysics (ed. G.

Schubert), vol. 9, pp. 91–119. Amsterdam, The Netherlands: Elsevier.

Tonks, W. B. & Melosh, H. J., 1993. Magma ocean formation due to giant impacts. J.

Geophys. Res. 98, 5319–5333.

Touboul, M., Kleine, T., Bourdon, B., Palme, H. & Wieler, R., 2007. Late formation

and prolonged differentiation of the Moon inferred from W isotopes in lunar

metals. Nature 450, 1206–1209.

Wasson, J. T. & Kallemeyn, G. W., 1988. Compositions of chondrites. Phil. Trans. R.

Soc. A 325, 535–544.

Wiechert, U., Halliday, A. N., Lee, D. C., Snyder, G. A., Taylor, L. A. & Rumble, D.,

2001. Oxygen isotopes and the moon-forming giant impact. Science 294, 345–

348.

Wood, B. J. & Halliday, A. N., 2005. Cooling of the Earth and core formation after the

giant impact. Nature 437, 1345–1348.

Workman, R. K. & Hart, S. R., 2005. Major and trace element composition of the

depleted MORB mantle (DMM). Earth Planet. Sci. Lett. 231, 53–72.

184

Page 185: Touboul Thesis Completed

Chapitre 5

Conclusions et perspectives

Conclusions and outlooks

185

Page 186: Touboul Thesis Completed

186

Page 187: Touboul Thesis Completed

5.1 Conclusions

Au cours de cette thèse, nous avons appliqué la chronométrie 182Hf-182W à courte période à

des objets variés afin de préciser la chronologie et la nature des processus régissant l’évolution du

système solaire primitif. Dans un premier temps, nous nous sommes intéressés aux âges d’accrétion

et à l’évolution thermique des corps parents de météorites, tels que les chondrites, les acapulcoites

et les lodranites, qui peuvent être déterminés en employant le système 182Hf-182W comme

thermochronomètre. La seconde partie est essentiellement consacrée à l’étude d’échantillons

lunaires, ramenés sur Terre par les misssions Appollo, qui a permis de contraindre les âges de

formation et de différenciation de la Lune. Les mesures isotopiques réalisées sur les anorthosites

lunaires, en particulier, ont nécessité le développement de nouvelles procédures chimiques qui ont

permis une amélioration des rendements et des blancs de W.

Nos études des chondrites ordinaires H (Kleine et al., 2008) et des acapulcoites et lodranites

(Touboul et al., en préparation) ont permis de préciser les échelles de temps d’accrétion et

l’évolution thermique des planétésimaux dont ces météorites sont issues et démontrent les

potentialités du système Hf-W en tant que thermochronomètre. Les isochrones internes conduisent

à des âges Hf-W de 5,1 ± 0,9 et 5,6 ± 1,0 Ma après la formation des CAI, correspondant à un âge

absolu de 4563,5 ± 0,7 et 4563,0 ± 0,9 Ma pour deux acapulcoites et une lodranite respectivement.

Celles réalisées sur des chondrites H5 et H6 donnent des âges Hf-W relatifs de 5,9 ± 0,9 Ma et 9,6

± 1,0 Ma. Les simulations de la diffusion du W entre clinopyroxène et métal mettent en évidence

que les températures de fermeture du système Hf-W dans ces différents objets est proche des

températures maximales atteintes lors de leur métamorphisme. Ces âges Hf-W sont donc proches

de correspondre à l’âge du pic thermique lors du métamorphisme affectant les corps parents des

chondrites H et des acapulcoites-lodranites. Etant donné que la température de fermeture de l’Hf-W

est supérieure aux autres systèmes chronométriques comme le Pb-Pb, l’Al-Mg et l’Ar-Ar, les âges

Hf-W sont systématiquement plus jeunes et fournissent des informations sur l’évolution aux plus

hautes températures et par conséquent la plus précoce de ces planétésimaux. Leur évolution

thermique, ainsi révélée par le biais des données thermochrométriques fournies par les différents

systèmes isotopiques, est compatible avec celle de planétésimaux chauffés de façon interne par la

désintégration radioactive de l’26Al et qui refroidissent par conduction thermique, suggérant que les

impacts n’ont pas joué un rôle majeur en tant que source de chaleur pour les corps parents des

chondrites H et des acapulcoites-lodranites. Bien que le diamètre du corps parent ainsi que la

187

Page 188: Touboul Thesis Completed

localisation en son sein ait une certaine influence, le paramètre, dont dépend de manière critique

son évolution thermique, est l’abondance d’26Al présent. Cette abondance est fixée par le moment

auquel l’accrétion intervient, qui peut donc être contraint relativement précisément. Ainsi, les

données thermochronométriques et les températures des pics thermiques requièrent que les corps

parents des chondrites H et des acapulcoites-lodranites se soient accrétés entre 2 et 4 Ma et entre

1,5 et 2 Ma après la formation des CAI respectivement. Elles suggèrent donc que l’accrétion des

planétésimaux non différenciés (2 à 4 Ma) est postérieure à celle des planétésimaux différenciés

(<1 Ma, Markowski et al., 2007) et que l’accrétion du corps parent des acapulcoites-lodranites (1,5

à 2 Ma), qui sont vus comme des intermédiaires entre météorites non différencié et différencié, est

intervenue dans un intervalle de temps compris entre l’accrétion de leurs corps parents. A la

lumière de cette séquence chronologique, il apparaît donc que le degré de différenciation des

planétésimaux est essentiellement déterminé par leur âge d’accrétion. Une autre implication

importante de nos résultats est que les corps parents des chondrites H et des acapulcoites-lodranites

ont des rapports Hf/W similaires (0,63 ± 0,20) mais significativement inférieurs au rapport Hf/W

des chondrites carbonées (1,21 ± 0,06, Kleine et al., 2004, Burkhardt et al., 2008). Cette différence,

inattendue pour des éléments réfractaires, révèle un fractionnement Hf-W au cours des deux

premiers Ma du système solaire. Ce fractionnement ne peut être lié à des processus au sein des

corps parents mais pourrait être produit lors de processus au sein de la nébuleuse avant ou durant

l’accrétion des planétésimaux.

Nos nouvelles données isotopiques du W obtenues pour les métaux lunaires (Touboul et al.,

2007) et les anorthosites lunaires à faible âge d’exposition (Touboul et al., accepté) conduisent à

une révision importante des âges Hf-W de différenciation de la Lune (Kleine et al., 2005) et

permettent de contraindre son âge de formation. La chronométrie Hf-W est bien adaptée à

l’investigation des échelles de temps de la différenciation précoce de la Lune car des

fractionnements Hf-W importants ont lieu au cours de la cristallisation de l’océan magmatique

lunaire dont les produits ont donc des rapports Hf/W variables (Shearer et Newsom, 2000, Righter

et Shearer, 2003). L’application de la chronométrie Hf-W aux roches lunaires est compliquée du

fait de la production de 182W par capture neutronique par le 181Ta durant l’exposition aux

rayonnements cosmiques à la surface de la Lune, si bien que le composant dominant du 182W dans

la plupart des échantillons lunaires est d’origine cosmogénique et masque leur signature isotopique

originelle (Kleine et al., 2005, Lee et al., 2002, Leya et al., 2003). Afin de contourner cet

188

Page 189: Touboul Thesis Completed

inconvénient, notre étude s’est focalisée sur les métaux extraits de KREEP et de basaltes de mer,

car les métaux ne contiennent pas de Ta, donc de 182W cosmogénique, et sur les anorthosites

faiblement exposées (<2 Ma). Les métaux et les anorthosites lunaires mesurées dans cette étude ont

tous les mêmes rapports 182W/184W, qui sont identiques à la valeur du standard terrestre et nous

montrons également que les anomalies en 182W (Kleine et al., 2005, lee et al, 1997, 2002)

rapportées précédemment reflètent la présence de 182W cosmogénique. Contrairement aux études

précédentes, nous n’avons donc trouvé aucune variation systématique des rapports 182W/184W des

échantillons lunaires en dépit de larges différences des rapports Hf/W de leurs sources. Ceci

indique que la cristallisation de l’océan magmatique s’est prolongée au-delà de l’extinction du 182Hf, c’est-à-dire au-delà des 60 premiers Ma du système solaire. Cette nouvelle contrainte Hf-W

sur la différenciation de la Lune (plus de 60 Ma après la formation du système solaire) n’est plus

incohérente avec les estimations dérivées de la chronométrie 146Sm-142Nd (∼200Ma après la

formation du système solaire ; Nyquist et al., 1995 ; Rankenburg et al., 2006 ; Boyet et al., 2007).

Une autre implication majeure des nouvelles données est qu’elles indiquent que le manteau lunaire

a une composition isotopique du W homogène et apparemment identique à celle du manteau

terrestre. Cette similarité entre les manteaux lunaire et terrestre pourrait indiquer que la Lune est

principalement composée de matériaux dérivés du manteau terrestre, mais les contraintes apportées

par les simulations dynamiques de l’impact géant indiquent, au contraire, que la Lune s’est

constituée essentiellement à partir de matériaux dérivés de l’impacteur (Canup et Asphaug, 2001)

Alternativement, les isotopes du W ont pu se rééquilibrer entre l’océan magmatique terrestre et le

disque proto-lunaire, comme cela a été proposé pour expliquer la similarité des compositions

isotopiques de l’O de la Terre et de la Lune (Pahlevan et Stevenson, 2007). Les compositions

isotopiques similaires du W des manteaux terrestre et lunaire révèlent également que l’impact a dû

arriver suffisamment tard afin que leur différence de rapports Hf/W ne se traduise pas en une

différence de rapports 182W/184W. L’âge maximum de formation de la Lune qui peut ainsi être

déterminé dépend de manière critique du degré de rééquilibration entre l’océan magmatique

terrestre et le disque proto-lunaire. Dans l’article de Bernard Bourdon (2008), nous avons exploré

les différents scénarios d’équilibration possibles et leur influence sur les âges calculés. Dans le cas

le plus probable où cette rééquilibration est totale, l’âge Hf-W maximum dérivé, en conjonction

avec l’âge Sm-Nd des roches lunaires les plus anciennes, permet de contraindre l’âge de formation

de la Lune à après la formation du système solaire. Dans le cas d’une rééquilibration Ma62 1090−+

189

Page 190: Touboul Thesis Completed

partielle, la formation de la Lune peut éventuellement intervenir antérieurement mais en aucune

manière avant 37 Ma après la formation du système solaire. L’impact géant constituant le dernier

événement majeur de l’accrétion terrestre, cet âge constitue actuellement la meilleure estimation de

l’âge maximal de la fin de l’accrétion de la Terre. Ces nouvelles contraintes chronométriques

remettent en cause la datation de la fin de l’accrétion terrestre à 30 Ma après la formation du

système solaire (Jacobsen, 2005) estimée à partir d’un modèle à deux stades des données Hf-W

terrestres et l’âge encore plus ancien obtenu avec un modèle continu. Elles sont également difficiles

à réconcilier avec une différenciation précoce du manteau terrestre à moins de 30 Ma après la

formation du système solaire, suggérée par Boyet and Carlson (2005) pour expliquer la différence

de ε142Nd de 20 ppm entre les chondrites et la Terre. Cette différenciation précoce requiert la

présence d’un réservoir complémentaire caché, isolé à la base du manteau (Boyet and Carlson

2005; Labrosse et al. 2007), à rapport Sm/Nd subchondrititique, qui devrait être préservé lors de

l’impact géant. Ceci est extrêmement improbable et nous suggérons que ce réservoir est

préférentiellement perdu pendant l’accrétion par érosion collisionelle ou que les planètes telluriques

totales aient des rapports Sm/Nd superchondritiques, du fait d’une accumulation préférentielle de

chondres dans la partie interne du système solaire, leur zone nouricière.

5.2 Perspectives

La réalisation d’isochrones internes Hf-W sur d’autres groupes de chondrites ordinaires tels

que les L et LL permettrait de révéler leurs âges et leur composition isotopique initiale. En utilisant

ensuite une approche similaire à celle appliquée ici pour les chondrites H et les achondrites

primitives, il sera possible de déterminer l’âge d’accretion et le rapport Hf/W de leur corps parent.

La combainaison des déterminations de la composition isotopiques du W et de concentrations en U,

en Th, en Hf et en W devrait permettre l’identification parmi les eucrites de celles ayant une

signature ε182W et un rapport Hf/W représentatifs de ces paramètres dans le manteau total de leur

corps parent. Il devrait ainsi être possibe de contraindre l’âge de la formation du noyau et donc de

donner un âge minimum d’accrétion de ce planétésimal. Ces investigations supplémentaires

permettraient donc d’observer si l’ensemble des planétésimaux obéit à la séquence d’accrétion mise

en évidence jusqu’à présent et a donc été chauffé essentiellement par la désintégration de l’26Al ou

si, pour certains d’autres eux, le chauffage par impacts est requis. De plus, la détermination des

190

Page 191: Touboul Thesis Completed

rapports Hf/W d’autres corps parents non différenciés permettrait de préciser l’âge et la nature des

processus géochimiques responsables du fractionnement Hf-W précoce au sein de la nébuleuse.

Comme le montre notre modélisation des nouvelles donneées lunaires, l’âge de formation

de la Lune dépend de manière critique du degré d’équilibration choisi. Des études expérimentales

des échanges isotopiques entre phase gazeuse et phase liquide silicatées à hautes températures sont

requises afin de préciser le comportement du W dans le contexte de l’impact géant et ainsi de

mieux contraindre le degré d’équilibration atteint entre la proto-Terre et le disque proto-lunaire et

donc l’âge de formation de la Lune. D’autre part, la multiplication des déterminations combinées de

concentration d’Hf, de W, d’U et de Th dans les chondrites et les échantillons lunaires et terrestres

fournirait une meilleure estimation des rapports Hf/W des manteaux lunaire et terrestre. Une

meilleure connaissance de ces rapports permettrait d’améliorer la precision des âges déterminés.

Références :

Bourdon, B., Touboul, M., Caro, G., Kleine, T., Early differentiation of the Earth and the Moon,

Phil. Trans. Roy. Soc. A. doi: 10.1098/rsta.2008.0125.

Boyet, M., and Carlson, R. W., 2007 A highly depleted moon or a non-magma ocean origin for the

lunar crust? Earth Planet. Sci. Lett. 262, 505-516.

Boyet, M. and Carlson, R. W., 2005. 142Nd evidence for early (>4.53 Ga) global differentiation of

the silicate Earth. Science 309, 576-581.

Burkhardt, C., Kleine, T., Palme, H., Bourdon, B., Zipfel, J., Friedrich, J., and Ebel, D., 2008. Hf-

W isochrons for CAIs and timescales for planetesimal accretion Geochim. Cosmochim. Acta,

submitted.

Canup, R. M., and Asphaug, E., 2001. Origin of the Moon in a giant impact near the end of the

Earth's formation. Nature 412, 708-712.

Jacobsen, S. B., 2005. The Hf-W isotopic system and the origin of the Earth and Moon. Ann. Rev.

Earth Planet. Sci. 33, 531-570.

Kleine, T., Touboul, M., Van Orman, J. A., Bourdon, B., Maden, C., Mezger, K., and Halliday, A.,

2008. Hf-W thermochronometry: closure temperature and constraints on the accretion and

cooling history of the H chondrite parent body. Earth Planet. Sci. Lett. 270, 106-118.

191

Page 192: Touboul Thesis Completed

Kleine, T., Mezger, K., Münker, C., Palme, H., and Bischoff, A., 2004a. 182Hf-182W isotope

systematics of chondrites, eucrites, and Martian meteorites: Chronology of core formation

and mantle differentiation in Vesta and Mars. Geochim. Cosmochim. Acta 68, 2935-2946.

Kleine, T., Palme, H., Mezger, K., and Halliday, A. N., 2005c. Hf-W chronometry of lunar metals

and the age and early differentiation of the Moon. Science 310, 1671-1674.

Lee, D. C., Halliday, A. N., Snyder, G. A., and Taylor, L. A., 1997. Age and origin of the moon.

Science 278, 1098-1103.

Lee, D. C., Halliday, A. N., Leya, I., Wieler, R., and Wiechert, U., 2002. Cosmogenic tungsten and

the origin and earliest differentiation of the Moon. Earth Planet. Sci. Lett. 198, 267-274.

Leya, I., Wieler, R., and Halliday, A. N., 2003b. The influence of cosmic-ray production on extinct

nuclide systems. Geochim. Cosmochim. Acta 67, 529-541.

Markowski, A., Quitté, G., Halliday, A. N., and Kleine, T., 2006b. Tungsten isotopic compositions

of iron meteorites: chronological constraints vs. cosmogenic effects. Earth Planet. Sci. Lett.

242, 1-15.

Nyquist, L. E., Wiesmann, H., Bansal, B., Shih, C. Y., Keith, J. E., and Harper, C. L., 1995. Sm-

146-Nd-142 Formation Interval for the Lunar Mantle. Geochim. Cosmochim. Acta 59, 2817-

2837.

Pahlevan, K., and Stevenson, D. J., 2007. Possible Origin of the geochemical similarity of the Earth

and Moon. Earth Planet. Sci. Lett. 262, 438-449.

Rankenburg, K., Brandon, A. D., and Neal, C. R., 2006. Neodymium isotope evidence for a

chondritic composition of the Moon. Science 312, 1369-1372.

Righter, K. and Shearer, C. K., 2003. Magmatic fractionation of Hf and W: Constraints on the

timing of core formation and differentiation in the Moon and Mars. Geochim. Cosmochim.

Acta 67, 2497-2507.

Shearer, C. K. and Newsom, H. E., 2000. W-Hf isotope abundances and the early origin and

evolution of the Earth-Moon system. Geochim. Cosmochim. Acta 64, 3599-3613.

Touboul, M., Kleine, T., Bourdon, B., Palme, H., and Wieler, R., 2007. Late formation and

prolonged differentiation of the Moon inferred from W isotopes in lunar metals. Nature 450,

1206-1209.

192

Page 193: Touboul Thesis Completed

Touboul, M., Kleine, T., Bourdon, B., Palme, H., and Wieler, R., Hf–W thermochronometry:

Closure temperature and constraints on the accretion and cooling history of the H chondrite

parent body. Accepted by Icarus.

Touboul M., Kleine T., Bourdon B., Van Orman J. A., Maden C., Irving A. J., Zipfel J., Bunch

T.E., Accretion and thermal history of the acapulcoite-lodranite parent body inferred from Hf-

W thermochronometry. In preparation.

193

Page 194: Touboul Thesis Completed

194

Page 195: Touboul Thesis Completed

Appendice 1

Systématiques Hf-W des eucrites cumulats et la

chronologie du corps parent des eucrites*

Hf-W Systematics of cumulate eucrites and the chronology of the eucrites parent body*

M. Touboul, T. Kleine, B. Bourdon

Institute for Isotope Geochemistry and Mineral Resources, Department of Earth Sciences, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerland

* 39th Lunar and Planetary Science Conference LPI Contribution No 1391, p. 2336

195

Page 196: Touboul Thesis Completed

196

Page 197: Touboul Thesis Completed

Hf-W SYSTEMATICS OF CUMULATE EUCRITES AND THE CHRONOLOGY OF THE EUCRITE PARENT BODY. M. Touboul1, T. Kleine1, B. Bourdon1, 1Institute for Isotope Geology and Mineral Resources, ETH Zürich, 8092 Zürich, Switzer-land ([email protected]).

Introduction: Key issues regarding the early evo-

lution of planetary bodies are the timescales of accre-tion, core formation, and silicate differentiation. The extinct 182Hf-182W decay system has proven particu-larly useful as a chronometer for these early planetary processes because Hf-W fractionations occurred dur-ing both core formation and melting processes in the silicate mantle. The latter is mainly related to the pres-ence of clinopyroxene and ilmenite that have very high Hf/W ratios. For instance, the crystallization of these minerals in the lunar magma ocean produced large Hf/W variations in the lunar mantle, such that Hf-W chronometry could provide a precise age for the crys-tallization of the lunar magma ocean [1].

The combined 147Sm-143Nd and 176Lu-176Hf isotope systematics of cumulate eucrites indicate that ilmenite was present in their source regions [2]. Therefore, cu-mulate eucrites may exhibit highly variable Hf/W ra-tios and the formation of their source region can poten-tially be dated using the Hf-W system. Here we pre-sent new Hf-W data for five cumulate eucrites (Serra de Mage, Moore County, Talampaya, Moama, and Binda) and use these data to constrain the chronology of core formation and mantle differentiation in the eucrite parent body.

Methods: Samples (~1 g) were finely crushed in an agate mortar and ~400 mg of the whole rock pow-der were dissolved in Savillex beakers using HF-HNO3 at 120°C for 72 hours. After drying, the samples are re-dissolved several times in HNO3-H2O2 and fi-nally in 6 M HCl-0.06 M HF. At this stage, complete dissolution is achieved and a ~5% aliquot is spiked with a mixed 180Hf-183W tracer for concentration de-termination by isotope dilution. The remaining ~95% were first loaded onto cation exchange columns to remove most of the sample matrix. Tungsten together with other high field strength elements (HFSEs) was eluted from this column using 1 M HCl-0.1 M HF. Separation of W from the other HFSEs was achieved using our previously established anion exchange tech-niques [3]. The anion exchange chemistry was re-peated once to ensure complete Ti removal. All measurments were performed using a Nu Plasma MC-ICPMS at ETH Zurich. Tungsten isotope compositions of the samples were determined relative to the 182W/184W obtained for two bracketing measurements of the W standard. In each run 40 ratios were meas-ured and, depending on the W content of the samples, the measurements were performed with ~0.6-1.5V for 182W, resulting in within-run statistics of 0.2-0.4 ε

units. The external reproducibility of these measure-ments is ±0.4-0.8 ε (2σ) for the 182W/184W ratio.

Results: In spite of largely different 180Hf/184W ratios, Moama, Serra de Mage and Talampaya have similar ε182W values of 18.0±0.8, 17.2±0.4, 18.2±0.4, respectively (Fig. 1). Moore County has the highest 180Hf/184W of the cumulate eucrites investigated here but has slightly lower ε182W value of 14.9±0.4. In con-trast to the radiogenic ε182W values and low W con-tents (~4-14 ppb) of these four cumulate eucrites, Binda has much lower ε182W=2.6±0.3 and higher W content (∼70 ppb). Compared to basaltic eucrites, the range in 180Hf/184W among the cumulate eucrites is much larger and their ε182W values are lower. Quitté et al. [4] obtained ε182W~0 and ~525 ppb W in their analyses for Serra de Magé, clearly inconsistent with the results presented here.

Figure 1: Hf-W isochron diagram for whole rocks of cumulate and basaltic eucrites. ε182W are the deviation of the 182W/184W ratio of a sample relative to the terrestrial standard in part per 10,000.

Discussion: Chronology of the eucrite parent body. Quitté et al. [4] argue that the high W content they obtained for Serra de Magé reflects sample het-erogeneity, implying the presence of a component with an extreme enrichment of W. Several lines of evidence indicate that this cannot be the case and that the low ε182W and Hf/W obtained for Serra de Mage by Quitte et al. are due to contamination with terrestrial W. First, a linear regression of the Hf-W data presented here and those reported by Quitte et al. [4] yields an initial 182Hf/180Hf that exceeds the solar system initial by a factor of ~2. Second, Quitte et al. reported an ε182W for Serra de Mage that is identical to the terrestrial W isotope composition and a W content that is ~40 times higher than the one obtained here, indicating substan-tial addition of W having terrestrial isotopic composi-

197

Page 198: Touboul Thesis Completed

tion. Third, in a plot of ε182W vs. 1/W, those eucrites that have anomalously high W contents (Serra de Mage and Jonzac [4], Juvinas [5], Binda [this study]) plot on straight lines. This is a characteristic feature of two-component mixtures. Note that the Hf-W data for Binda obtained here plot exactly on the mixing line defined by the Serra de Mage data reported here and by Quitte et al. [4]. Likewise, Jonzac and the Hf-W data point for Juvinas from Yin et al. [5] plot on a straight line defined by the average of basaltic eucrites and terrestrial W.

Therefore, the linear array defined by the basaltic eucrites (+ Serra de Mage as reported by Quitte et al. [4]) in the Hf-W isochron diagram does not represent an isochron but is a mixing line between basaltic eucrites and terrestrial W. If those samples that were clearly contaminated with terrestrial W are excluded from the regression, no statistically meaningful isochron is obtained for the basaltic eucrites. This is because the variation in Hf/W ratios among the re-maining basaltic eucrites is too small.

Figure 2: W isotope composition versus 1/W for cumulate and basal-tic eucrites. Dashed lines show mixing between terrestrial W and W of cumulate and basaltic eucrite endmembers.

The Hf-W data for basaltic eucrite nevertheless provide important age constraints. The combined 147Sm-143Nd and 176Lu-176Hf isotope systematics of basaltic eucrites indicate that they probably formed as large degree melts from a chondritic source [2]. The lack of substantial variations in 176Hf/177Hf ratios among the basaltic eucrites rules out a significant role of ilmenite in their sources. Therefore, only limited fractionation of Hf and W is expected during forma-tion of the basaltic eucrites. This is consistent with the narrow spread in Hf/W ratio and ε182W values among most basaltic eucrites and indicates that the average ε182W and Hf/W of basaltic eucrites may provide a good estimate of these parameters in the bulk mantle of the eucrite parent body. If this is correct, these val-

ues can be used to constrain the timing of core forma-tion.

Using 180Hf/184W= 28.8±1.8 and ε182W = 24.4±3.2 for average basaltic eucrites and assuming a chondritic composition for the bulk eucrite parent body, a two-stage model age for core formation of 2.5±1.2 Ma can be calculated. This corresponds to an absolute age of 4566.1±1.2 Ma (calculated relative to the angrites D’Orbigny and Sahara 99555) and is consistent with the timescales for the differentiation of the eucrite par-ent body deduced from Al-Mg systematics [6], with a Pb-Pb age of 4566.52±0.33 Ma for cumulate eucrite Asuka 881394 [7], and with the 53Mn-53Cr age of 4566.5±0.6 Ma obtained from a eucrite whole-rock isochron [8] (recalculated relative to the angrite D'Or-bigny).

The initial ε182W calculated from the chondrite-mantle isochron of the eucrite parent body is -3.05 ±0.15, slightly higher than ε182W values for magmatic iron meteorites. Therefore, core formation in the eucrite parent body appear to have occurred later as in the parent bodies of magmatic iron meteorites.

Origin of cumulate eucrites. One of the key obser-vations of our study is that cumulate eucrites, in spite of widely different 180Hf/184W ratios, have homogene-ous W isotope composition. The much lower ε182W value of Binda most likely reflects contamination with terrestrial W and the slightly lower ε182W ~15 of Moore County compared to the other cumulate eucrites (ε182W ~18) also is most readily explained by slight terrestrial contamination (Fig. 2). Therefore, cumulate eucrites appear to be characterized by constant ε182W values of ~18, indicating that their widely different 180Hf/184W ratios must have been established after 182Hf became effectively extinct, i.e., later than ~60 Ma after CAI formation. This is consistent with results from 147Sm-143Nd and 176Lu-177Hf whole-rock isochrons corresponding to an age of ~100 Ma after CAI formation [2]. Note, however, that these age con-straints are difficult to reconcile with the ~4566.5 Pb-Pb age for cumulate eucrite Asuka 881394 [7].

Acknowlegment: We thank the teams of the Smithsonian Institution in Washington, the Museum National d’histoire naturelle de Paris and the Museum Victoria in Melbourn for providing the samples.

References: [1] Touboul et al., (2007) Nature 450, 1206-1209. [2] Blichert-Toft J. et al. (2002) EPSL 204, 167-181. [3] Kleine T. et al. (2004) GCA 68, 2935-2946. [4] Quitté G. et al. (2000) EPSL 184, 83-94. [5] Yin Q. et al. (2002) Nature 450, 949-952. [6] Bizzarro M. et al. (2005) ApJ 632, 41-44. [7] Amelin Y. et al. (2006) LPSC XXXVII, #1970. [8] Shukolyukov A. and Lugmair G.W. (1997) GCA 62, 2863–2886 [9] Markowski A. et al. (2006) EPSL 242, 1-15.

198

Page 199: Touboul Thesis Completed

Appendice 2

Chronométrie Hf-W:

Accrétion et évolution précoce des astéroides et des planètes telluriques*

Hf-W chronometry and the accretion and early evolution of

asteroids and terrestrial planets*

T. Kleine1, M. Touboul1, B. Bourdon1, F. Nimmo2, K. Mezger3, H.

Palme4, Q.-Z. Yin5, S.B. Jacobsen6, A.N. Halliday7

1Institute for Isotope Geochemistry and Mineral Resources, ETH Zurich, NW, Clau-siusstrasse 25, 8092 Zurich, Switzerland 2Department of Earth and Planetary Sciences, University of California Santa Cruz, 1156 High St., Santa Cruz CA 95064, USA 3Institut für Mineralogie, Universität Münster, Corrensstrasse 24, 48149 Münster, Ger-many 4Institut für Geologie und Mineralogie, Universität zu Köln, Zülpicherstrasse 49b, 50674 Köln, Germany 5Department of Earth and Planetary Sciences, Harvard University, 20 Oxford Street, Cambridge, MA 02138, USA

6Department of Geology, University of California Davis, One Shields Avenue, Davis, CA 95616, USA 7 Department of Earth Sciences, University of Oxford, Parks Road, OX1 3PR, United Kingdom

* Accepté par Geochimica et Cosmochimica Acta

199

Page 200: Touboul Thesis Completed

200

Page 201: Touboul Thesis Completed

Abstract

We review the chronology for the accretion and earliest evolution of asteroids

and terrestrial planets as obtained by applying Hf-W chronometry to meteoritic and

planetary samples. The Hf-W systematics of meteorites and some of their components

lead to the following succession and duration of events. CAI formation at 4568.5±0.5

Ma was followed by the accretion and differentiation of the parent bodies of magmatic

iron meteorites within less than ~1 Ma. Chondrules from H chondrites formed 1.7±0.7

Ma after CAIs, consistent with 26Al-26Mg ages for L and LL chondrules. Magmatism on

the parent bodies of angrites, eucrites, and mesosiderites started at ~3 Ma after CAI

formation and might have extended until ~10 Ma. A similar timescale is obtained for

the high-temperature metamorphic evolution of the H chondrite parent body. Thermal

modeling combined with these age constraints reveals that the different thermal histo-

ries of meteorite parent bodies primarily reflect their initial abundance of 26Al, which is

determined by their accretion age. Impact-related processes are an important heat source

for the subsequent evolution of asteroids but these do not appear to have induced large-

scale melting. Hafnium-tungsten ages for eucrite metals postdate CAI formation by ~20

Ma and probably reflect impact-triggered thermal metamorphism in the crust of the

eucrite parent body. Hafnium-tungsten data for IAB irons appear most consistent with

mixing of metals and silicates during reassembly of the IAB parent body following its

impact disruption. The timing of this event, however, remains poorly constrained.

The strong fractionation of lithophile Hf from siderophile W during core forma-

tion makes the Hf-W system an ideal chronometer for this major differentiation event.

However, for larger planets such as the terrestrial planets the calculated Hf-W ages are

particularly sensitive to the occurrence of large impacts, the degree to which impactor

cores re-equilibrated with the target mantle during large collisions, and changes in the

metal-silicate partition coefficients of W due to changing fO2 in differentiating plane-

tary bodies. Calculated core formation ages for Mars range from 0-20 Ma after CAI

formation and currently cannot distinguish between scenarios where Mars formed by

runaway growth and where its formation was more protracted. Tungsten model ages for

core formation in Earth range from ~30 Ma to >100 Ma after CAIs and hence do not

provide a unique age for the formation of Earth. However, the identical 182W/184W ratios

of the lunar and terrestrial mantles suggest that the giant Moon-forming impact and the

201

Page 202: Touboul Thesis Completed

termination of Earth's core formation occurred after extinction of 182Hf (i.e., more than

~50 Ma after CAIs). The identical 182W/184W ratios of the lunar and terrestrial mantles

also require that either the Moon consists predominantly of terrestrial material or that W

isotopes in the proto-lunar magma disk had equilibrated with the Earth's mantle.

Hafnium-tungsten chronometry also provides constraints on the duration of

magma ocean solidification in terrestrial planets. Variations in the 182W/184W ratios of

Martian meteorites reflect solidification of a Martian magma ocean during the effective

life-time of 182Hf and are consistent with global differentiation of the Martian mantle at

~40 Ma after CAI formation. In contrast to Mars, no 182W variations exist in the lunar

mantle, indicating magma ocean solidification later than ~60 Ma, consistent with 147Sm-143Nd ages for ferroan anorthosites. Differentiation of the bulk silicate Earth, which

probably was related to the crystallization of a global magma ocean, most likely oc-

curred after the giant Moon-forming impact. This is consistent with a 146Sm-142Nd age

of 50-200 Ma for the differentiation of Earth's mantle but is difficult to reconcile with

the mantle differentiation age of <30 Ma that has been suggested based on different 142Nd abundances in chondrites and Earth's mantle.

202

Page 203: Touboul Thesis Completed

1 INTRODUCTION

The formation of asteroids and terrestrial planets began with the accumulation of dust grains to kilometre-sized planetesimals, which is thought to have occurred in less than ~104 years (Chambers, 2004). Gravity and gas drag caused these planetesimals to collide and form increasingly larger bodies in a period of runaway growth, the products of which include numerous Moon- to Mars-sized planetary embryos that are thought to have formed in ~106 years (Weidenschilling et al., 1997). Collisions among these bodies mark the late stages of accretion, culminating in the formation of a few terrestrial plan-ets that sweep up all the other bodies. Numerical simulations suggest that this stage may have taken 107-108 years (Chambers and Wetherill, 1998; Agnor et al., 1999). The Moon probably formed during this period and involves a 'giant impact' of a Mars-sized body with proto-Earth at the very end of Earth's accretion (Canup and Asphaug, 2001).

Collisions among the planetary embryos and the decay of short-lived radioactive isotopes (especially 26Al) caused the planetary interiors to heat up and eventually melt and differentiate. As a consequence, all major bodies of the inner solar system and also many smaller bodies are chemically differentiated into a metallic core and a silicate mantle (Walter and Tronnes, 2004). However, some objects such as the parent bodies of chondritic meteorites remained undifferentiated.

Determining the timescales for the accretion of asteroids and terrestrial planets is essential for evaluating planetary accretion models and, hence, is key to constraining the nature of the planet formation process. Similarly, determining the timescales for the early, high-temperature evolution of planetary bodies (i.e., chemical differentiation and thermal metamorphism) provides essential information for constraining the initial con-ditions that controlled the subsequent evolution of planets. Short-lived nuclides have proven particularly useful for obtaining such age constraints and, depending on their half-lives, can provide information on different stages of early planetary accretion and evolution. While the 26Al-26Mg chronometer (t1/2 ~0.73 Ma) is suitable for dating proc-esses in the first ~5 Ma of the solar system to high precision, the 182Hf-182W system, owing to its longer half-life of ~8.9 Ma, can potentially be used to date processes in the first ~60 Ma. This timescale is most relevant to the formation and early differentiation of terrestrial planets, processes that are thought to have been largely complete in the first ~100 Ma of the solar system (Chambers, 2004).

In the past ~15 years 182Hf-182W chronometry has successfully been applied to date a variety of processes associated with the formation and earliest evolution of plane-tary bodies. The main interest in the Hf-W system was initially related to its potential

203

Page 204: Touboul Thesis Completed

for dating core formation (Harper et al., 1991; Lee and Halliday, 1995; Harper and Jacobsen, 1996) but Hf-W fractionations also occur during mantle melting processes (Shearer and Newsom, 2000; Righter and Shearer, 2003). This allows the timescales of early mantle differentiation (e.g., magma ocean crystallization) to be determined. There is also an increasing number of applications that use internal Hf-W isochrons for mete-orites to date the timing of Hf-W closure in meteorites and constrain the thermal evolu-tion of their parent bodies (Markowski et al., 2007; Burkhardt et al., 2008; Kleine et al., 2008b). Here we review the chronology of the accretion and earliest evolution of aster-oids and terrestrial planets as obtained by applying Hf-W chronometry to meteoritic and planetary samples.

2 THE 182HF-182W CHRONOMETER

Hafnium-182 is produced by both the r- and s-processes and decays to 182W with a half-life of ~8.9 Ma. Given the range in Hf/W ratios in planetary reservoirs and mete-orite components, 182Hf was sufficiently abundant to create resolvable 182W differences in the first ~60 Ma of the solar system and Hf-W chronometry can be used to date chemical and physical processes that fractionated Hf and W during this period. This time span is most appropriate for studying the formation and early differentiation of planetary objects in the inner solar system, processes that are generally thought to have been complete within the first ~100 Ma of the solar system (Chambers, 2004). Due to the different geo- and cosmochemical properties of Hf and W, these two elements are fractionated by a variety of processes that occurred during accretion, differentiation and early evolution of planetary bodies. This makes the Hf-W system an extremely useful chronometer for the early solar system.

2.1 Hf-W fractionation during planetary differentiation and in meteorites

Hafnium and W are refractory elements and as such should occur in chondritic relative abundances in most bulk planetary objects. The Hf-W systematics of chondrites therefore provide an estimate for the W isotope composition of a bulk planetary object and serve as a reference for calculating Hf-W ages of core formation. During core for-mation lithophile Hf is entirely retained in the mantle, whereas siderophile W is prefer-entially partitioned into the metal core, resulting in Hf/W~0 in the core and correspond-ingly high Hf/W ratios in the silicate mantle. The metal-silicate partition coefficient of W depends on several parameters including fO2, p, T, and silicate melt composition (Walter and Thibault, 1995; Walter et al., 2000). Therefore, depending on the conditions

204

Page 205: Touboul Thesis Completed

of core formation, the depletion of W in the silicate mantle can vary among different planetary objects but also during core formation within one object.

Additional fractionation of Hf from W can occur in the silicate part of differenti-ated planetary objects and in this case is related to the higher incompatibility of W rela-tive to Hf (Righter and Shearer, 2003). Whereas W is one of the most incompatible elements (similar to Th and U), Hf can be incorporated into some minerals such as high-Ca pyroxene and ilmenite. Crystallization of these two minerals - e.g. during solidifica-tion of a magma ocean or during basalt crystallization - can result in the formation of reservoirs with high Hf/W ratios. In some cases, these fractionations can exceed those induced by core formation. For instance, due to the presence of ilmenite and high-Ca pyroxene in their source, lunar high-Ti mare basalts have Hf/W ratios as high as ~100, whereas KREEP-rich samples have much lower Hf/W ratios of ~20 (Kleine et al., 2005c). Such large fractionations greatly enhance the application of the Hf-W system as a chronometer of mantle differentiation.

The Hf/W ratio that is characteristic for the entire mantle of a differentiated plane-tary body cannot be measured directly in most cases because substantial Hf-W frac-tionations occurred during igneous processes within the mantle (Righter and Shearer, 2003). The bulk mantle Hf/W ratio must therefore be inferred by comparing the W con-centrations in mantle-derived samples with a refractory lithophile element that is as in-compatible as W and whose abundance relative to Hf is known. Trace element studies of lunar, terrestrial, and meteoritic basalts indicate that Th, U, and W have similar in-compatibilities (Palme and Rammensee, 1981a; Newsom et al., 1996), such that Th/W and U/W ratios in silicate mantles in conjunction with Hf/Th and Hf/U ratios in chon-drites can be used to determine the Hf/W ratios in bulk planetary mantles.

Table 1 summarizes current best estimates for the Hf/W ratios in the bulk mantles of Mars, the Earth and Moon. Nimmo and Kleine (2007) estimated the Th/W ratio of the Martian mantle to be 0.98±0.13. Using the Hf/Th = 3.1±0.1 of carbonaceous chon-drites (Kleine et al., 2007) this Th/W translates into Hf/W = 3.0±0.4. However, ordinary chondrites have higher Hf/Th ratios of 4.0±0.2 (Kleine et al., 2007), in which case the calculated Hf/W ratio of the Martian mantle would be 3.9±0.6. As shown below, uncer-tainties and variations in the Hf/Th ratio of chondrites have a significant effect on the calculated core formation ages for Mars because the Hf/W ratio of the Martian mantle is only slightly higher than the Hf/W ratio of chondrites (Nimmo and Kleine, 2007).

The Hf/W ratio of the bulk silicate Earth can be inferred from its Th/W ratio of 5.5±1.6 (Newsom et al., 1996), which, using Hf/Th = 3.1±0.1 for carbonaceous chon-drites, corresponds to Hf/W = 17±5. Similarly, the Hf/W ratio of the bulk lunar mantle can be estimated from its U/W ratio of 1.93±0.08 (Palme and Rammensee, 1981b). Us-

205

Page 206: Touboul Thesis Completed

ing Hf/U = 13.7±0.7 for chondrites (Rocholl and Jochum, 1993) this corresponds to an Hf/W ratio of the lunar mantle of 26±2. A slightly lower Hf/W of 25±2 is obtained if Th/U = 4.2±0.2 for the Moon and Hf/Th = 3.1±0.1 for chondrites are used. The differ-ences in these estimates reflect uncertainties and/or variations in the Hf, U and Th con-centrations of chondrites as well as uncertainties in the trace element composition of the lunar and terrestrial mantles. Additional uncertainties may arise if planetary bodies have non-chondritic relative abundances of refractory lithophile elements (in this case Hf, Th and U). For instance, Pb isotope data for terrestrial mantle rocks were used to estimate the terrestrial Th/U to 4.2±0.2 (Allègre et al., 1986), slightly higher than Th/U~3.8 in chondrites (Rocholl and Jochum, 1993). Moreover, Caro et al. (2008) proposed that Earth (+Moon and Mars) have superchondritic Sm/Nd ratios.

The fractionation of Hf from W is not restricted to planetary differentiation but also occurs among the constitutive minerals of many meteorites and some of their com-ponents. Table 1 summarizes measured Hf/W ratios in mineral separates from various meteorites along with the ratios of mineral-melt partition coefficients for Hf and W (Righter and Shearer, 2003). The high Hf/W ratios in high-Ca pyroxene make it possi-ble to obtain precise internal isochron for meteorites because high-Ca pyroxene is an important constituent of numerous meteorites and their components, such as eucrites, angrites and CAIs. Almost all meteorites contain some metal, which has Hf/W~0 and allows direct measurement of the initial 182W/184W of a sample. Precise Hf-W ages can thus be obtained from Hf-W isochrons involving high-Ca pyroxene and other silicates (e.g., olivine, melilite), such as for CAIs (Burkhardt et al., 2008) and angrites (Markowski et al., 2007; Kleine et al., 2008a), or co-genetic metal and silicate fractions, such as for H chondrites (Kleine et al., 2008b). Table 2 summarizes initial 182Hf/180Hf ratios for meteorites that have been obtained from internal isochrons.

2.2 Notation and Hf-W isotope systematics

Tungsten isotope data are commonly expressed in ε182W, which is defined as fol-lows:

( )( )

182 184sample182 4

182 184standard

W Wε W= -1 10

W W

⎧ ⎫⎪ ×⎨⎪ ⎪⎩ ⎭

⎪⎬ (1)

Note that similarly to other short-lived chronometers such as the 53Mn-53Cr and 26Al-26Mg systems, ε182W values are calculated relative to the 182W/184W of the terrestrial standard. However, for modeling purposes some authors (Harper and Jacobsen, 1996; Jacobsen, 2005; Nimmo and Agnor, 2006) also expressed 182W/184W relative to the W isotope composition of the Chondritic Uniform Reservoir (CHUR), as defined by the W

206

Page 207: Touboul Thesis Completed

isotope evolution of carbonaceous chondrites (section 2.3). To distinguish between these notations ( )W tεΔ for a sample or reservoir j is defined as:

( )

( )

182 184

4W 182 184

CHUR

W W ( )ε ( ) 1 10

W W ( )jj

tt

t

⎧ ⎫⎪Δ = − ×⎨⎪ ⎪⎩ ⎭

⎪⎬ (2)

Thus, by definition, of chondrites is always 0 at any time. The present-day W

isotope composition of a sample or reservoir is given by:

( )W tεΔ

182 182 180 182

184 184 184 1800 0

W W Hf HfW W W Hi i

⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛= + ×⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜

⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ f⎞⎟⎠

(3)

where the subscript 0 refers to the present-day and the subscript i to the initial value. Equation (3) reveals that the initial 182Hf/180Hf of a sample or reservoir can be obtained from the slope of an isochron in a plot of 182W/184W vs. 180Hf/184W. Relative ages or formation intervals between any two samples or reservoirs can then be calculated using the following equation:

( )( )

182 1801

182 1802

Hf Hf1 lnHf Hf

⎧ ⎫⎪ ⎪Δ = − × ⎨ ⎬⎪ ⎪⎩ ⎭

(4)

where λ is the decay constant and has a value of 0.078±0.002 Ma-1 (Vockenhuber et al., 2004).

In the simplest model of core formation it is assumed that the core formed instan-taneously from a bulk planet having chondritic Hf/W. Then the two-stage model age of core formation, tcf, can be calculated from Hf-W data for a mantle or core j as follows (e.g., Lee and Halliday, 1995):

0

180 180

184 184182CHUR

180 182 182

184 184CHUR

Hf HfW W1 Hfln

Hf W WW W

jcf

T

j

⎧ ⎫⎛ ⎞ ⎛ ⎞−⎪ ⎪⎜ ⎟ ⎜ ⎟

⎛ ⎞ ⎝ ⎠ ⎝ ⎠⎪ ⎪= × ×⎨⎜ ⎟ ⎬

⎛ ⎞ ⎛ ⎞⎝ ⎠⎪ ⎪−⎜ ⎟ ⎜ ⎟⎪ ⎪⎝ ⎠ ⎝ ⎠⎩ ⎭

(5)

where (0

182 180Hf HfT

) refers to the initial 182Hf/180Hf of the solar system as determined

from internal Hf-W isochrons for CAIs. An alternative equation for calculation two-stage model ages for core formation

was introduced by Harper and Jacobsen (1996) and uses the following notation:

( )

( )

0 0

0

0

182/

W 182 180

182/

182 180

Hfε ( )Hf

HfHf

cf

cf

T T Tj Hf WW j

T

tHf WW j

T

t Q f e e

Q f e

λ λ

λ

− − −

⎛ ⎞ ⎡ ⎤Δ = −⎜ ⎟ ⎢ ⎥⎣ ⎦⎝ ⎠

⎛ ⎞= ⎜ ⎟

⎝ ⎠

(6)

207

Page 208: Touboul Thesis Completed

where ( )4 180 182 4182 CHUR

10 Hf/ W 1.42 10WQ = = × (see below and Table 3). The Hf/W

fractionation in reservoir j relative to CHUR is defined by

( )

( )Hf/W

CHUR

Hf W1

Hf Wj

jf = − (7)

In practice, to determine the age of core formation of a sample or reservoir, we will re-write equation (6) as follows:

0

182/

182 180

W

/

W

HfHf1 ln

ε (0)

1.381 lnε (0)

Hf WW j

Tcf j

Hf Wj

j

Q f

t

f

λ

λ

⎡ ⎤⎛ ⎞⎢ ⎥⎜ ⎟

⎝ ⎠⎢ ⎥= ⎢ ⎥Δ⎢ ⎥

⎢ ⎥⎣ ⎦⎡ ⎤

= ⎢ ⎥Δ⎢ ⎥⎣ ⎦

(8)

Note that the second equality is derived by assuming a carbonaceous chondritic compo-sition, i.e., . ( )4 180 182 4

182 CHUR10 Hf/ W 1.42 10WQ = = ×

Equations (5) and (8) reveal that core formation ages can be calculated from the 182W/184W and 180Hf/184W ratios of a mantle or core once the Hf-W systematics of chondrites and the 182Hf/180Hf at the time of CAI formation are known. We shall see later that the calculation of realistic core formation ages requires more complex models, at least for Earth-sized bodies.

2.3 Reference parameters for Hf-W chronology

2.3.1 Hf-W systematics of chondrites

The first W isotope data for chondrites indicated that Earth's mantle and chon-drites had identical 182W/184W ratios (Lee and Halliday, 1995; Lee and Halliday, 1996). However, in 2002 it was shown that carbonaceous chondrites have 182W/184W ~2 ε units lower than Earth's mantle (Kleine et al., 2002; Schoenberg et al., 2002a; Yin et al., 2002). The currently most precise determination of the 182W/184W ratio of carbonaceous chondrites is based on 14 samples that have an average ε182W of -1.9±0.1 (Kleine et al., 2004a). There are no resolvable differences in the 182W/184W among the carbonaceous chondrite groups (Fig. 1), consistent with their limited spread in 180Hf/184W ratios (180Hf/184W = 1.18 × Hf/W) from ~1.1 to ~1.6 (Kleine et al., 2004a). Ordinary chondrite groups exhibit a larger range in 180Hf/184W ratios and ε182W values (Fig. 1). H chon-drites have lower 180Hf/184W ratios (from ~0.6 to ~1.1) and ε182W values (from ~-2.5 to ~-2.0), whereas LL chondrites have higher 180Hf/184W ratios (from ~1.6 to ~2.1) and

208

Page 209: Touboul Thesis Completed

ε182W values (from ~-1.7 to ~-1.5) (Kleine et al., 2007). Likewise, enstatite chondrites have 180Hf/184W ratios and ε182W values that are similar to those of the H chondrites (Lee and Halliday, 2000).

The different Hf/W ratios among the chondrite groups most likely reflect metal-silicate and refractory element fractionation in the solar nebula. Therefore, bulk plane-tary objects may have different Hf/W ratios, at least to the extent as present in the dif-ferent chondrite groups. In most cases, however, the variations of the Hf/W ratios among bulk chondrites are small compared those produced during planetary differentia-tion (section 2.1).

2.3.2 Initial 182Hf/180Hf and 182W/184W of CAIs

Precise knowledge of the 182Hf/180Hf and 182W/184W at the time of CAI formation is essential for making full use of the time constraints provided by Hf-W chronometry. This is because CAIs are the oldest yet dated material formed in the solar system (Gray et al., 1973) and as such are the most suitable reference point for constraining the dura-tion of processes in the early solar system. Moreover, intervals relative to CAIs deter-mined with the Hf-W chronometer can be directly compared to 26Al-26Mg ages that are also commonly expressed as time intervals calculated based on the initial 26Al/27Al of CAIs.

Several approaches have been used to determine the initial 182Hf/180Hf and 182W/184W ratios of the solar system and for this task Hf-W data for meteorites can be utilized in three different ways (Table 4). The first approach is to assume that the lowest 182W/184W measured in iron meteorites represents the initial W isotope composition of the solar system (Lee and Halliday, 1995; Quitté and Birck, 2004). The difference be-tween this assumed initial value and the present-day 182W/184W of chondrites then represents ingrowth from decay of 182Hf in a reservoir with a chondritic 180Hf/184W ra-tio. However, the 182W/184W of most iron meteorites have been lowered by the interac-tion with cosmic rays (Masarik, 1997; Leya et al., 2003) and the lowest 182W/184W re-ported for iron meteorites does not provide a robust estimate for the initial 182W/184W of the solar system but rather reflects the effects of interaction with thermal neutrons (Masarik, 1997; Leya et al., 2003; Kleine et al., 2005a; Markowski et al., 2006b; Schér-sten et al., 2006). Therefore, using the lowest 182W/184W ratio reported for iron meteor-ites results in estimates for the initial 182Hf/180Hf of the solar system (Quitté and Birck, 2004) that are too high.

The second approach uses internal Hf-W isochrons for meteorites with a well-defined absolute age and relies on two assumptions: (i) the absolute age determined for a sample corresponds to the time of Hf-W closure, and (ii) the absolute age of CAIs is

209

Page 210: Touboul Thesis Completed

known. The first assumption is strictly valid only for rapidly cooled samples such as angrites or CAIs (see below). One problem with this approach is that the absolute age of CAIs is not known with sufficient precision, with Pb-Pb ages ranging from 4567.11±0.16 Ma (Amelin et al., 2002; Amelin et al., 2006) to 4568.5±0.5 Ma (Bouvier et al., 2007). This introduces some uncertainty when the 182Hf/180Hf at any point in time is back-calculated to the time of CAI formation. This approach was employed by Kleine et al. (2002) using Hf-W data for the H4 chondrite Ste. Marguerite and an initial 182Hf/180Hf of (1.09±0.09)×10-4 at the time of CAI formation was calculated (assuming a CAI age of 4566 Ma). The main reason for using this sample was that at that time no precise Hf-W isochrons for rapidly cooled samples were available. Similarly, Yin et al. (2002) used Hf-W data for one CAI and two equilibrated ordinary chondrites to deter-mine an initial 182Hf/180Hf of (1.00±0.08)×10-4 at the time of CAI formation. This ap-proach is based on the assumption that closure of the Hf-W system in those equilibrated ordinary chondrites occurred at the time of CAI formation.

The third and most direct approach to determine the initial 182Hf/180Hf of the solar system is to obtain Hf-W data for the first solids that formed in the solar system. Ireland and Bukovanska (2003) reported ionprobe Hf-W data for zircons from the H5 chondrite Simmern. These zircons have an ultrarefractory-enriched trace element pattern, a feature commonly associated with refractory inclusions, and their initial 182Hf/180Hf of (7.2±4.5)×10-5 might be close the initial value of the solar system. The initial 182Hf/180Hf of the solar system is most reliably and directly determined from internal isochrons for CAIs because these are the oldest known objects that formed in the solar system. Haf-nium-tungsten data are available for 5 bulk CAIs and these exhibit a narrow range in Hf/W and 182W/184W, which are slightly elevated relative to carbonaceous chondrites (Burkhardt et al., 2008). However, in type B CAIs melilite and fassaite have very differ-ent 180Hf/184W ratios, mainly reflecting the compatibility of Hf in fassaites. Their high 180Hf/184W resulted in radiogenic 182W/184W ratios, which makes it possible to obtain precise internal Hf-W isochrons for CAIs (Burkhardt et al., 2008). Whole-rocks and mineral separates from 8 CAIs from the Allende and NWA 2364 CV3 chondrites plot on a single well-defined isochron corresponding to an initial ε182W of -3.28±0.12 and an initial 182Hf/180Hf of (9.72±0.44)×10-5 (Burkhardt et al., 2008). As discussed in detail by Burkhardt et al. (2008), parent body processes such as aqueous alteration or thermal metamorphism had no detectable effect on the Hf-W systematics of these CAIs, such that the CAI isochron provides the initial 182W/184W and 182Hf/180Hf ratios at the time of CAI formation.

210

Page 211: Touboul Thesis Completed

2.3.3 Calibration of the 182Hf-182W chronometer and conversion to absolute ages

The accurate application of Hf-W chronometry to constrain timescales in the early solar system requires that relative Hf-W ages between different early solar system mate-rials are consistent with the differences in their absolute ages. This criterion is strictly applicable only to those samples that cooled rapidly, such that potential differences in closure temperatures did not result in resolvable age differences among various chro-nometers. This requirement is met by angrites. These exhibit high U/Pb ratios, such that precise Pb-Pb ages are available (Lugmair and Galer, 1992; Amelin, 2008) and the available Pb-Pb ages for angrites reveal an age dispersion of ~7 Ma, which makes it possible to intercalibrate the Hf-W and U-Pb system.

In angrites, fassaites have high 180Hf/184W ratios whereas the other constituents - olivine, plagioclase and in one case metal - have low 180Hf/184W ratios (see Table 1), resulting in variations in W isotope compositions of up to ~15 ε182W. This spread makes it possible to obtain precise Hf-W isochrons for angrites (Markowski et al., 2007; Kleine et al., 2008a) and the initial 182Hf/180Hf ratios obtained from the slope of these isochrons are summarized in Table 2. The comparison of Pb-Pb and Hf-W ages can be made for several angrites including D'Orbigny, Sahara 99555 and Northwest Africa 2999, 4590, and 4801. In Fig. 2, the initial 182Hf/180Hf ratios of these angrites are plotted against their Pb-Pb ages. The angrites D'Orbigny and Sahara 99555 as well as North-west Africa 4590 and 4801 plot on a straight line, whose slope corresponds to λ182Hf=0.075±0.007 Ma-1, in excellent agreement with the experimental determination that yielded λ182Hf=0.078±0.002 Ma-1 (Vockenhuber et al., 2004). This provides evi-dence that the absolute Hf-W ages calculated relative to these angrites are robust and accurate. Northwest Africa 2999 plots slightly below, but within uncertainty of this line, most likely indicating a slight disturbance of the Hf-W system in this sample. This is evident from the Hf-W data for one of the whole-rock and the fines fraction of North-west Africa 2999, which plot off the isochron (Markowski et al., 2007).

The Hf-W and Pb-Pb ages of angrites can also be used to convert relative Hf-W ages to absolute ages, which is essential for facilitating a comparison and combination with results from Pb-Pb chronometry. An important feature of Fig. 2 is that the calibra-tion of the Hf-W system onto an absolute timescale yields consistent results regardless of which of the four angrites, D'Orbigny, Sahara 99555, Northwest Africa 4590 or 4801 are used. Here absolute ages are calculated relative to D'Orbigny [initial 182Hf/180Hf = (7.18±0.22) × 10-5 at 4564.42±0.12 Ma (Amelin, 2008)] but identical results are ob-tained if absolute Hf-W ages are calculated relative to any of the other angrites.

Fig. 2 reveals that the absolute Hf-W age for CAIs, as calculated relative to an-grites, is 4568.5±0.5 Ma (Burkhardt et al., 2008) and, hence, ~1.5 Ma older than the

211

Page 212: Touboul Thesis Completed

4567.11±0.16 Ma Pb-Pb age for Efremovka CAIs (Amelin et al., 2002; Amelin et al., 2006). The reasons for this discrepancy are currently unclear but may be related to a partial resetting of the U-Pb system in CAIs. An important observation from Fig. 2 is that if CAIs were used as an age anchor for the Hf-W system, then the Hf-W ages for all angrites would be younger than their Pb-Pb ages, which is highly unlikely.

2.4 Closure temperature of the Hf-W system

The interpretation of Hf-W ages in comparison to results from other chronometers and within the framework of models for the thermal evolution of asteroids requires knowledge of the closure temperature (Tc) for diffusive exchange of Hf and W among the constituent minerals in a rock (Dodson, 1973; Ganguly and Tirone, 2001). Closure temperatures can be calculated from diffusion rates of W in the appropriate minerals but such data are not available yet. Nevertheless, using the model developed by Van Orman et al. (2001), Kleine et al. (2008b) estimated the diffusion parameters of W in high-Ca pyroxene, the major host of radiogenic 182W in most meteorites, and obtained an activa-tion energy of 453 kJ/mol and a pre-exponential factor of 9.53 × 10-5 m2/s. These pa-rameters can be used to calculate Tc for the Hf-W system as a function of cooling rate and effective grain size. Fig. 3 presents Hf-W closure temperatures as a function of cooling rate, calculated for a 10 μm grain size and using the analytical model from Dodson (1973). Also shown are closure temperatures for the U-Pb system in clinopy-roxene (Cherniak, 1998) and phosphates (Cherniak et al., 1991) and for the Al-Mg sys-tem in anorthite (LaTourrette and Wasserburg, 1998). Fig. 3 reveals that for a wide range of cooling rates (and grain sizes), Tc for the Hf-W system is always higher than closure temperatures of the U-Pb system in high-Ca pyroxene and phosphates and the Al-Mg system in anorthite. Consequently, in slowly cooled samples (e.g., in most metamorphosed meteorites) the Hf-W system will have closed at higher temperatures than other chronometers and Hf-W ages are expected to be older than Pb-Pb and 26Al-26Mg ages. Hence, Hf-W chronometry has the potential to date processes associated with the earliest evolution of meteorite parent bodies and is less susceptible to resetting by later thermal events than any other chronometer commonly used in cosmochronol-ogy.

The analytical models for closure temperature presented by Dodson (1973) and Ganguly and Tirone (2001) make several assumptions that do not necessarily apply to short-lived chronometers. For example, the Dodson (1973) and Ganguly and Tirone (2001) models assume (i) an infinite sink for radiogenic daughters; (ii) a decay time that is very long compared to the cooling time (which might not be valid for short-lived chronometers); and (iii) that heating at peak metamorphic conditions was sufficient to

212

Page 213: Touboul Thesis Completed

homogenize any pre-existing isotopic heterogeneity. By contrast, the model of Van Or-man et al. (2006) does not rely on these assumptions, and is thus a more realistic model for the production and exchange of radiogenic daughters in short-lived isotope systems. This model was applied by Kleine et al. (2008b) to numerically simulate the production and diffusive exchange of radiogenic W between high-Ca pyroxene and metal in H chondrites. These simulations show that for highly metamorphosed rocks, such as the H6 and H5 chondrites peak metamorphic temperatures were sufficiently high to ho-mogenize any pre-existing W isotope heterogeneity. In this case, the calculated closure temperatures are identical to those calculated using the Dodson (1973) model. By con-trast, in case of the H4 chondrites, the assumption that peak metamorphic conditions were sufficient to reset the Hf-W system is not valid. Consequently, for such samples the Hf-W closure temperature cannot be calculated using the Dodson (1973) model and assessing the effects of metamorphism on the Hf-W system in H4 chondrites requires a model that can simulate the prograde path, such as the Van Orman et al. (2006).

3 TIMESCALES FOR THE ACCRETION AND EARLY EVOLUTION OF PLANETESIMALS

3.1 Iron meteorites - remnants of the first planetesimals

The first comprehensive investigation of the W isotope composition of iron mete-orites was performed by Horan et al. (1998), who showed that iron meteorites have ε182W values between ~-5 and ~-3. More recently, the W isotope compositions of a large number of iron meteorites was determined to higher precision than was obtainable during the Horan et al. study and these studies confirmed the range in ε182W values ob-tained by Horan et al. (Kleine et al., 2005a; Markowski et al., 2006b; Schérsten et al., 2006; Qin et al., 2008). With the higher precision obtained in these more recent studies variations within groups of iron meteorites became apparent. Moreover, the non-magmatic irons appear to have slightly higher ε182W values than the magmatic irons.

It was observed that the ε182W values of iron meteorites become more negative with increasing exposure ages and values lower than ~-4.0 were only reported for sam-ples having exposure ages as old as ~1 Ga (Fig. 4). This decrease in ε182W reflects burnout of W isotopes by capture of thermal neutrons produced by cosmic rays (Masarik, 1997; Leya et al., 2003). The production rate of thermal neutrons depends on the pre-atmospheric size of and the location within the meteorite, such that cosmogenic effects on the W isotope composition can vary within a single meteorite. This was dem-onstrated for the two iron meteorites Grant and Carbo. In both samples, 182W/184W ra-

213

Page 214: Touboul Thesis Completed

tios were lower close to the pre-atmospheric centre compared to those close to the pre-atmospheric surface (Markowski et al., 2006a).

There is currently no direct proxy available for precisely determining the flux of the thermal neutrons relevant to the effects on W isotopes, such that these effects can currently only be estimated from the concentrations of cosmogenic noble gases in par-ticular samples (Markowski et al., 2006a; Qin et al., 2008). However, since this does not provide a direct measure of the production rate of thermal neutrons, the estimated ther-mal neutron fluxes are model-dependent (Masarik, 1997; Leya et al., 2003). Conse-quently, there is not yet a reliable method to correct for cosmogenic effects on W iso-topes in iron meteorites. Therefore, the currently most reliable age information regard-ing core formation in iron meteorite parent bodies are provided by the W isotope com-positions of magmatic iron meteorites that have not been exposed to thermal neutrons, either because they have young exposure ages or because they were large enough for their interior to have been shielded from thermal neutrons. Among the analyzed iron meteorites, this is only the case for Negrillos (IIAB) and Gibeon (IVA). Negrillos has an exposure age of ~50 Ma (Leya et al., 2000) and several authors have reported W iso-tope data for the IIAB Negrillos averaging at ε182W=-3.42±0.08 (2σ) (Kleine et al., 2005a; Lee, 2005; Markowski et al., 2006b). From the exposure age a downward shift of ε182W of ~0.03 is calculated (Leya et al., 2003), resulting in a corrected ε182W of -3.39±0.08 (assuming a 50% uncertainty on the correction). Note however that the accu-racy of this correction is difficult to assess. The IVA iron meteorite Gibeon has ε182W=-3.38±0.05 (Qin et al., 2007) and the low concentrations of cosmogenic noble gases in this meteorite indicate that it is derived from the inner part of a larger body. Gibeon therefore probably was largely shielded from thermal neutrons produced by the cosmic rays. It should be noted, however, that cosmogenic noble gases are produced by primary protons whereas the effects on W isotopes are caused by secondary neutrons, which may reach deeper inside the meteoroid due to the poor capacity of Fe for slowing down neutrons.

The estimates based on exposure ages indicate that in iron meteorites the 182W/184W ratio might be lowered by ~0.1 ε182W per ~100 Ma of cosmic-ray exposure and for an "average" shielding. This decrease in ε182W is similar to the radiogenic 182W ingrowth that is achieved during the first Ma in a chondritic reservoir. Therefore, even in iron meteorites that were exposed to very low thermal neutron fluxes the effects of 182W burnout can be significant but are essentially undetectable with current analytical techniques. It will thus be important to develop a monitor for the thermal neutron flux in the energy range that is relevant for W isotopes.

214

Page 215: Touboul Thesis Completed

The corrected ε182W=-3.39±0.08 for Negrillos and the measured ε182W=-3.38±0.05 for Gibeon correspond to W model ages for core formation of -1.0±1.3 (2σ) and -0.9±1.2 (2σ) Ma after crystallization of type B CAIs. The uncertainties of these ages were calculated by propagating the uncertainties on the ε182W values of the iron meteorites, the initial ε182W of CAIs, and the present-day ε182W of carbonaceous chon-drites. A more conservative approach is to calculate the range of ages obtained from the minimum and maximum ε182W differences between iron meteorites and CAIs. This ap-proach results in W model ages that range from -2.8 to +0.8 for Negrillos and from -2.5 to +0.6 Ma for Gibeon. Note that the uncertainties of these ages do not include any cosmic-ray effects, which may be present but are essentially undetectable with current analytical techniques. This makes it difficult to exactly determine the time of core for-mation based on W isotope data. Table 5 summarizes W model ages for different group of magmatic iron meteorites that are based on ε182W values that were corrected for cos-mic-ray effects using noble gas isotope systematics (Qin et al., 2008). For the IVB iron meteorites negative ages are obtained, implying that the correction procedure did not fully account for the cosmic-ray effects on their 182W/184W ratios. This highlights the fact that noble gas systematics are not a direct proxy for thermal neutrons. Nevertheless, the range of ages obtained from these corrected ε182W values is similar to the ages ob-tained for Negrillos and Gibeon, which have minor to absent cosmic-ray effects.

In spite of the presence of cosmogenic effects on W isotopes in iron meteorites and difficulties in reliably quantifying them, it can be stated with confidence that core formation in the parent bodies of magmatic iron meteorites predated the formation of chondrules, most of which have Al-Mg and Pb-Pb ages of ~2-3 Ma after CAI formation (Kita et al., 2000; Amelin et al., 2002; Kunihiro et al., 2004; Rudraswami and Goswami, 2007). For instance, the estimated cosmogenic effect on the W isotope composition of Negrillos is ~-0.03, corresponding to an age correction of ~0.3 Ma. Therefore, the cor-rection would need to be ~10 times larger for the core formation age to be younger than the chondrule ages. However, it seems unlikely that the correction equations (Leya et al., 2003) are incorrect by one order of magnitude because they predict the observed effects to within ~50%.

The finding that the parent bodies of magmatic iron meteorites accreted and dif-ferentiated before chondrules formed perhaps is the single most significant result of W isotope investigations in meteorites. This result has several far-reaching implications and it is therefore important to consider cases in which the W isotope composition of iron meteorites would not reflect the timing of core formation: (i) iron meteorite parent bodies did not contain 182Hf, such that their 182W/184W never evolved over time; (ii) temperatures during core formation were too low to allow diffusion of radiogenic W

215

Page 216: Touboul Thesis Completed

from silicates into the metal; (iii) there are nucleosynthetic W isotope anomalies in iron meteorites. As discussed in detail by Burkhardt et al. (2008), these scenarios are incon-sistent with the Hf-W data for meteorites, such that the most straightforward interpreta-tion of the low 182W/184W ratios of iron meteorites is an early metal segregation in their parent bodies.

These results provide two important constraints. First, the assembly of iron mete-orites parent bodies prior to chondrule formation is inconsistent with the standard model for asteroid accretion, in which chondrites represent the precursor material from which asteroids accreted and then differentiated. It is important to note that the early formation of iron meteorite parent bodies is consistent with results from numerical simulations, which predict that accretion of planetesimal occurred in less than 1 Ma (Chambers, 2004). Second, the parent bodies of iron meteorites formed so early that heating by de-cay of 26Al must have been an important heat source. Thermal modeling indicates that for planetesimals that formed within 2 Ma after CAI formation and that are larger than ~20 km in diameter, heating by 26Al decay was sufficient to cause melting and core formation (Hevey and Sanders, 2006). The implications of these results will be dis-cussed in more detail below (section 3.6).

3.2 Chronology of IAB-IIICD iron meteorites

In contrast to the magmatic iron meteorites, the IAB-IIICD irons show little evi-dence for trace element fractionation and probably do not sample a planetary core (Scott and Wasson, 1975; Wasson and Kallemeyn, 2002). Moreover, metal-silicate separation was less efficient in the IAB-IIICD iron meteorites, as is evident from abundant silicate inclusions. Some authors have suggested that the IAB-IIICD iron meteorites formed in localized impact-melt pools in the megaregolith of a chondritic parent asteroid (Choi et al., 1995; Wasson and Kallemeyn, 2002), whereas others proposed formation of the IAB-IIICD irons by impact disruption of a partially differentiated body (Benedix et al., 2000) or incomplete differentiation of a chondritic planetesimal (Takeda et al., 2000).

Horan et al. (1998) were the first to show that most IAB-IIICD iron meteorites have slightly elevated ε182W values compared to magmatic irons. This was confirmed by later W isotope investigations of iron meteorites (Kleine et al., 2005a; Markowski et al., 2006b; Schérsten et al., 2006). Whereas most magmatic iron meteorites have ε182W values below the initial ε182W of CAIs, all the non-magmatic irons have ε182W values similar to or slightly higher than the CAI initial (Fig. 5). Both groups of iron meteorites have a similar range in exposure ages, such that the higher ε182W of the IAB-IIICD irons compared to the magmatic irons cannot be caused by cosmic-ray effects on W isotopes in magmatic irons (see section 3.1). They rather reflect a later event on the

216

Page 217: Touboul Thesis Completed

IAB-IIICD parent asteroid. Due to the presence of cosmic-ray effects on W isotopes the exact difference in the 182W/184W ratios of non-magmatic and magmatic irons is difficult to assess but can be estimated from samples that were not exposed to substantial ther-mal neutron fluxes. This is the case for the non-magmatic iron Caddo County, whose 182W/184W is 0.2±0.1 ε units more radiogenic than values for the magmatic irons Gibeon and Negrillos.

The W isotope composition of the Caddo County metal can be interpreted in dif-ferent ways. It may represent the W isotope composition at the time of metal-silicate separation during partial melting of a chondritic source, in which case its two-stage model age of ΔtCAI = 0.9±2.2 Ma would date this event. An alternative interpretation is that the W isotope composition was partially or completely reset during impact-triggered mixing of metal and silicates. This requires diffusion of radiogenic W from the silicate inclusions into the metal. The metamorphic temperature of IAB silicates is similar to those of H6 chondrites [i.e., ~900 °C (Benedix et al., 2000)], which, depend-ing on the grain sizes of the silicates and the cooling rate might be sufficiently high to cause W diffusion (Kleine et al., 2008b). Schulz et al. (2006) reported W isotope data for several bulk IAB silicates that have ε182W values ranging from ~6 to ~30. Silicate inclusions and host metals do not define isochrons, indicating that silicates and metals are not in W isotope equilibrium. This may reflect incomplete resetting of the Hf-W system during thermal metamorphism of IAB silicates, consistent with 26Al-26Mg (Liu et al., 2002) and 129I-129Xe ages (Bogard et al., 2005) for silicate inclusions from Caddo County of >7 Ma and ~12 Ma, which postdate the W model age for the Caddo County metal.

The W isotope composition of the Caddo County metal nevertheless provides im-portant age information. Partial (or complete) resetting of the Hf-W system during thermal metamorphism results in an increase of the 182W/184W ratio of the metal, such that the two-stage model age is too young. Therefore, metal segregation in the IAB par-ent body must have occurred earlier than 0.9±2.2 Ma. At such an early time, 26Al was sufficiently abundant to cause global melting of asteroids, indicating that the partial melting that resulted in the formation of IAB metals was caused by internal heating from 26Al decay.

Taken together the Hf-W data for IAB iron meteorites appear most consistent with the model of Benedix et al. (2000). This model invokes an early differentiation of the IAB parent body combined with impact disruption followed by mixing of molten metals with a variety of solid silicates during reassembly. Such processes seem capable of causing the partial resetting observed for the Hf-W system because silicates from differ-ent sources and possibly different ages will be mixed together. In such a scenario, sili-

217

Page 218: Touboul Thesis Completed

cates and metals are not co-genetic and hence should not define an isochron, consistent with the observed Hf-W systematics.

3.3 Chronology of the eucrite parent body

3.3.1 Accretion and primordial differentiation

The first W isotope data for eucrites were obtained by Lee and Halliday (1996) and Quitté et al. (2000) presented the first comprehensive investigation of the Hf-W systematics of basaltic eucrites. Subsequently, Yin et al. (2002) reported Hf-W data for one Juvinas whole-rock, Kleine et al. (2004a) reported Hf-W data for several additional basaltic eucrites and Kleine et al. (2005b) determined the W isotope composition of eucrite metals. Most recently, the W isotope composition of zircons in eucrites was de-termined (Srinivasan et al., 2007) and the Hf-W systematics of bulk-rock eucrites re-investigated (Touboul et al., 2008a).

Most basaltic eucrites exhibit high 180Hf/184W and radiogenic 182W/184W ratios (Lee and Halliday, 1996; Quitté et al., 2000; Kleine et al., 2004a). Quitté et al. (2000) reported a large range in 180Hf/184W from ~0.1 to ~46 and ε182W values from ~0 to ~39. In contrast, Kleine et al. (2004a) found a much narrower range in 180Hf/184W from ~12 to ~32 and ε182W from ~14 to ~26. The Hf-W data for eucrites define a linear trend in a plot of 182W/184W vs. 180Hf/184W, which yields a slope of (7.25±0.05)×10-5 and an initial ε182W of -0.5±0.3. Quitté et al. (2000) initially interpreted the Hf-W data for eucrites as indicating differentiation of the eucrite parent body at ~11 Ma after CAI formation. However, this age was obtained using an initial 182Hf/180Hf of ~2.75×10-4 for CAIs but this value was later corrected to ~1×10-4 (Kleine et al., 2002; Yin et al., 2002). Subse-quently, Quitté and Birck (2004) used the Hf-W data for eucrites to infer an age of ~8 Ma for parent body differentiation but this age was obtained using an initial 182Hf/180Hf of ~1.6×10-4 derived from the 182W/184W of the iron meteorite Tlacotepec. However, the 182W/184W of this sample has been lowered by the interaction with cosmic rays, such that the age estimate from Quitté and Birck (2004) is too young.

Recently, Touboul et al. (2008a) showed that the Hf-W data for basaltic eucrites do not define an isochron. The correlation line is mainly defined by three samples [Serra de Magé and Jonzac (Quitté et al., 2000) and Juvinas (Yin et al., 2002)] that have an-omalously high W contents. Using larger samples sizes, Touboul et al. (2008a) were not able to reproduce the Hf-W data for these three samples and obtained much lower W contents and much higher 182W/184W. Moreover, these authors could show that, in a plot of 182W/184W vs. 1/W, the earlier reported W data for these three samples plot on mixing lines between terrestrial W and the W data obtained by Touboul et al. (2008a).

218

Page 219: Touboul Thesis Completed

This relationship sheds doubt on the usefulness of the W data of these three samples to constrain the age of eucrite differentiation.

The Hf-W data for basaltic eucrite may nevertheless be used to extract age infor-mation on the differentiation of the eucrite parent body. The narrow range in Sm/Nd and Lu/Hf ratios as obtained from the combined 147Sm-143Nd and 176Lu-176Hf isotope sys-tematics of basaltic eucrites indicate that they may have formed as large degree melts from a source having chondritic relative abundances of refractory lithophile elements (Blichert-Toft et al., 2002). The lack of substantial variations in 176Hf/177Hf ratios among the basaltic eucrites rules out a significant role of ilmenite in their sources be-cause this would have resulted in substantial Lu-Hf fractionations. Therefore, only lim-ited fractionation of Hf and W is expected during formation of the basaltic eucrites. This is consistent with the narrow spread in Hf/W ratios and ε182W values among basaltic eucrites (Kleine et al., 2004a; Touboul et al., 2008a). If this correct, the 180Hf/184W and 182W/184W ratios of basaltic eucrites may closely reflect that of their sources and the Hf-W data can be used to constrain the timing of the earliest differentiation of the eucrite parent body. However, there is substantial variation in the 180Hf/184W and 182W/184W ratios reported for basaltic eucrites, which probably reflects sample heterogeneities and the small sample sizes used for these analyses. Therefore, the average 180Hf/184W ratio and ε182W value of basaltic eucrites are rather imprecisely defined and are 180Hf/184W=27±8 and ε182W=22±8 [calculated using the data reported in Kleine et al. (2004a; 2005b)]. Using these values a two-stage model age for core formation of ΔtCAI = 3±6 Ma is calculated. The large uncertainty on this age reflects uncertainties in the Hf/W ratio and W isotope composition of the bulk mantle of the eucrite parent body and highlights the need for Hf-W data for representative bulk rock samples of basaltic eucrites.

3.3.2 Magmatism and thermal metamorphism

Basaltic eucrites have magmatic textures similar to those of terrestrial and lunar basalts, indicating formation as lava flows at or near the surface of their parent body, presumably the asteroid Vesta. As a result of rapid cooling near the surface, the pyrox-enes in terrestrial and lunar basalts are chemically zoned. In contrast, pyroxenes in most eucrites show no chemical zoning and contain exsolution lamellae, indicating slow cooling and/or a protracted thermal metamorphism after crystallization (Takeda and Graham, 1991). In addition to this thermal overprint, most eucrites are brecciated as a result of impacts on the surface of Vesta.

Most basaltic eucrites contain some zircons, which owing to their very high Hf/W ratios (up to ~17'000) can be dated with the Hf-W system (Ireland and Bukovanska,

219

Page 220: Touboul Thesis Completed

2003; Srinivasan et al., 2007). Srinivasan et al. (2007) reported ion probe Hf-W data for zircons and pyroxenes from basaltic eucrites Asuka 881388 and Asuka 881467. The zircon-pyroxene isochrons yield 182Hf/180Hf ratios of ~7.5×10-5 and ~6.0×10-5, corre-sponding to Hf-W ages of ΔtCAI~3 and ~6 Ma and absolute ages of ~4565 Ma and ~4562 Ma, respectively.

All eucrites contain at least some metal (typically <0.5%), which formed either during crystallization of the basalts or during later metamorphism (Duke, 1965). The low Ni contents of these metals exclude a meteoritic origin by impacts on the surface of Vesta. The Camel Donga eucrite has the highest metal contents among eucrites and the metal in this meteorite evidently formed by reduction of FeO and FeS during thermal metamorphism (Palme et al., 1988). Metals from five basaltic eucrites (Camel Donga, Juvinas, Bereba, Bouvante, Ibitira) have radiogenic W isotope compositions ranging from ~11 to ~16 ε182W but only for Camel Donga could a metal-silicate isochron be obtained (Kleine et al., 2005b). This isochron yields an initial 182Hf/180Hf of (1.7±0.7)×10-5, corresponding to an age of ΔtCAI = 22±5 Ma and an absolute age of 4546±5 Ma (Kleine et al., 2005b) (Fig. 6). The W model ages for the metals from the other basaltic eucrites, calculated relative to the average 180Hf/184W and 182W/184W of basaltic eucrites, yield similar ages (Kleine et al., 2005b).

The major hosts of radiogenic W in basaltic eucrites are high-Ca pyroxene, ilmen-ite and zircon. Identifying which of these minerals is the source of the radiogenic 182W in the metals requires knowledge of their respective closure temperatures for W diffu-sion as well as of the peak metamorphic temperatures of eucrites. Based on the diffusion profiles in pyroxenes the metamorphic temperatures of eucrites were ~800-900°C (Takeda and Graham, 1991). There are no experimental data available for diffusion of W in high-Ca pyroxene, ilmenite or zircon that could be used to calculate closure tem-peratures. However, the Hf-W closure temperature in a high-Ca pyroxene-metal system has been estimated using the numerical model of Van Orman et al. (2001) [see section 2.4 and (Kleine et al., 2008b)] and, for an assumed starting temperature of 1000 °C, closure temperatures of 750-900 °C were obtained for a wide range of grain sizes and cooling rates (Kleine et al., 2008b).

It therefore seems likely that post-crystallization heating of basaltic eucrites re-sulted in W diffusion from high-Ca pyroxenes into metals. In contrast, W diffusion out of zircons probably did not occur because their Hf-W closure temperature is most likely very high. If W diffusion in zircon is slower than Pb diffusion, as is the case in other silicates such as high-Ca pyroxene (Kleine et al., 2008b), then the Hf-W closure tem-perature in zircons will be higher than the Pb closure temperature of >1000°C (Mezger and Krogstad, 1997). This is higher than the crystallization temperature of zircons in

220

Page 221: Touboul Thesis Completed

basaltic melts, indicating that the zircons crystallized below the Hf-W closure tempera-ture but were not reset during later thermal metamorphism. This is consistent with the higher initial 182Hf/180Hf of the zircons compared to the metal (see below) and indicates that the Hf-W zircon ages most likely dates the crystallization of basaltic eucrites.

Obtaining an age for crystallization from the Hf-W zircon data, however, requires Hf-W data for another phase that is co-genetic with the zircons and remained a closed system for Hf and W after crystallization. Srinivasan et al. (2007) used their Hf-W data for pyroxenes in the isochron calculations but the Hf-W closure temperature for high-Ca pyroxene in conjunction with temperatures estimates for the thermal metamorphism of basaltic eucrites indicate that high-Ca pyroxene did not remain closed for W diffusion during thermal metamorphism (see above). In this case, the slope of the zircon-pyroxene isochron might be too steep and the apparent age too old. More reliable age information may then be obtained from a zircon-whole-rock isochron, provided that the eucrite whole-rocks remained closed systems with regard to Hf and W. Using the average 180Hf/184W and 182W/184W of basaltic eucrites in the regression of the zircon data results in Hf-W ages of ΔtCAI = 6±2 Ma and 9±5 Ma for the Asuka 881467 and Asuka 881388 zircons, corresponding to absolute ages of 4562±2 Ma and 4560±5 Ma (Fig. 6).

An important observation is that eucrite metals plot above the regression line de-fined by the zircon and whole-rock Hf-W data (Fig. 6). This is consistent with a signifi-cant time gap between zircon crystallization and the time of metal closure during ther-mal metamorphism. However, the metal and zircon data were not obtained on the same samples and a comprehensive assessment of the effects that thermal metamorphism had on the Hf-W system in eucrites will require Hf-W data for all major constituents of the same eucrite.

The similarity of Hf-W closure temperature in high-Ca pyroxene and metamor-phic temperature for basaltic eucrites suggests that the Hf-W metal age corresponds closely to the time of the thermal peak and hence provides an age for the thermal meta-morphism. This event occurred ~15 Ma after crystallization of Asuka 881467 zircons, suggesting that the thermal metamorphism is not related to the magmatic activity on the eucrite parent body. Therefore, the thermal metamorphism may be caused either by im-pact heating or by burial of the basaltic eucrites under hot interior material that was ex-cavated by impacts (Kleine et al., 2005b).

3.4 Timing of magmatism on the angrite and mesosiderite parent bodies

The angrites D'Orbigny and Sahara 99555 formed by rapid cooling of basaltic magmas. Therefore, their Hf-W ages of ΔtCAI=3.9±0.7 Ma and ΔtCAI=4.3±0.7 Ma re-flects the crystallization age of these basaltic rocks. The angrites NWA 2999, NWA

221

Page 222: Touboul Thesis Completed

4590, and NWA 4801 have significantly younger Hf-W ages ranging from ~8 to ~10 Ma after CAI formation. This may reflect protracted cooling of more deeply buried ig-neous bodies inside the angrite parent body. Evidence for such a protracted cooling is provided by the annealed textures of NWA 2999 and NWA 4801 and the coarse plu-tonic texture of NWA 4590 and NWA 4801. The absence of chemical zoning in the pyroxenes of NWA 2999 and NWA 4801 provides evidence for annealing at tempera-tures above ~700–800 °C (Takeda and Graham, 1991), consistent with olivine-spinel temperatures of ~870 °C for NWA 2999 (Kuehner et al., 2006).

The chronological data combined with the petrologic evidence therefore suggest that extrusion of basalts occurred at ~4 Ma and that interior parts of the angrite parent body cooled below ~800 °C (i.e., Tc of the Hf-W system) in less than ~10 Ma, as con-strained by the Hf-W ages for NWA 4801 and NWA 4590.

Schönbächler et al. (2000) reported a Hf-W isochron for a basaltic clast from the mesosiderite Vaca Muerta. The plagioclase-pyroxene isochron corresponds to an age of ΔtCAI = 3.0±1.5 Ma and an absolute Hf-W age of 4565.3±1.4 Ma. This age is similar to Hf-W ages for the quenched angrites D'Orbigny and Sahara 99555 as well as to Hf-W zircon ages for basaltic eucrites. These Hf-W ages suggest that extrusion of basaltic melts on the eucrite, angrite and mesosiderite parent bodies occurred in a narrow time interval of ~3-10 Ma after CAI formation. This is an important observation because it is consistent with results from numerical simulations for the thermal evolution of planetesimals heated by decay of 26Al and 60Fe. These simulations show that magma-tism on bodies that accreted within the first 2-3 Ma of CAI formation will initiate within the first ~4 Ma after CAI formation and that the last melting will have occurred before 10 Ma after CAI formation (Sahijpal et al., 2007).

3.5 Chronology of the H chondrite parent body

The early formation of the parent bodies of iron meteorites contrasts with the rela-tively late formation of chondrules. Based on 26Al-26Mg and Pb-Pb ages it is well estab-lished that most chondrules formed ~2-3 Ma after CAIs (Kita et al., 2000; Amelin et al., 2002; Kunihiro et al., 2004; Rudraswami and Goswami, 2007; Connelly et al., 2008a; Kurahashi et al., 2008; Rudraswami et al., 2008). Chondrules must have formed prior to the assembly of chondrite parent bodies, such that the formation ages of chondrules provide the earliest possible time of parent body accretion. Furthermore, the distinct physical and chemical characteristics of chondrules from each chondrite group suggest that parent body accretion occurred briefly after chondrule formation. In a turbulent solar nebula, chondrules would be efficiently mixed on short timescales (Cuzzi et al., 2005), such that a characteristic population of chondrules with its distinct size distribu-

222

Page 223: Touboul Thesis Completed

tion and chemical composition could only be preserved, if this chondrule population is accreted into larger bodies soon after chondrule formation. This was quantified by Alexander (2005), who estimated that material in a 1 AU wide area could be mixed within less than ~0.5 Ma. Since asteroid feeding zones probably were smaller, these mixing times become even shorter. If these estimates are correct, chondrite accretion must have occurred almost instantaneously after chondrule formation at ~2-3 Ma after CAI formation (Alexander, 2005).

This timescale for the accretion of chondrite parent bodies is also consistent with results from thermal modeling, which suggest that the parent bodies of chondrites ac-creted more than ~2 Ma after CAIs because otherwise they would have differentiated due to heating from abundant 26Al. The question then arises as to whether this timescale for accretion is consistent with the long-term thermal evolution of chondrite parent bod-ies. Although the thermal evolution of carbonaceous chondrites is difficult to constrain – because except for CK chondrites there are only few metamorphosed samples – the thermal evolution of ordinary chondrites can be investigated in detail because these samples exhibit a wide range of metamorphic grades (Dodd, 1969). Many ordinary chondrites are brecciated and shocked, such that a signature of their earliest thermal evolution – which is of main interest here – might have partly been erased. However, relatively unshocked H chondrites appear to have preserved their earliest cooling history (Trieloff et al., 2003).

Kleine et al. (2008b) presented Hf-W ages for such H chondrites and the selected samples cover the range of metamorphic conditions characteristic for H4-H6 chondrites. The Hf-W ages of the H chondrites become younger with increasing metamorphic grade and range from ΔtCAI=1.7±0.7 Ma for the H4 chondrite Ste. Marguerite to ΔtCAI=9.6±1.0 Ma for the H6 chondrites Kernouvé and Estacado (Fig. 7a). The closure temperature of the Hf-W system in H4 chondrites is 800±50 °C and thus higher than the peak meta-morphic temperature of H4 chondrites. Therefore, parent body metamorphism did not result in significant diffusion of radiogenic W from silicates into metals, such that the Hf-W age for Ste. Marguerite (H4), which is based on a metal-silicate isochron, was not reset. Consequently, the 1.7±0.7 Ma age for Ste. Marguerite must reflect an earlier high temperature event, and the only viable process in H chondrites is chondrule formation. The 1.7±0.7 Ma Hf-W age for Ste. Marguerite (H4) is consistent with 26Al-26Mg ages for chondrules from L and LL ordinary chondrites (Kita et al., 2000; Rudraswami and Goswami, 2007).

Due to the similarity of peak metamorphic and Hf-W closure temperature in H5 and H6 chondrites, the Hf-W ages for these H chondrites correspond closely to the time of the thermal peak in the respective parts of the H chondrite parent body (Fig. 7b). As a

223

Page 224: Touboul Thesis Completed

consequence, the Hf-W system closed earlier than any other isotope systems and the high temperature interval of ~8 Ma obtained from the Hf-W system is much shorter than intervals obtained from Rb-Sr and Pb-Pb chronometry (Wasserburg et al., 1969; Göpel et al., 1994). Combined with previously published chronological data (Trieloff et al., 2003) the Hf-W ages reveal that shortly after their thermal peak H6 chondrites cooled at ~10°C/Ma, H5 chondrites at ~30 °C/Ma and H4 chondrites at ~55°C/Ma (Kleine et al., 2008b). This inverse correlation of cooling rate and metamorphic grade is most consis-tent with an onion-shell structure (Miyamoto et al., 1981) of an H chondrite parent body that was internally heated internally by 26Al decay (Kleine et al., 2008b). Therefore, heating by 26Al decay does not only account for the peak temperatures reached inside the H chondrite parent body but is also consistent with its long-term thermal evolution. For planetesimals other than the H chondrite parent body, the thermal evolution is less well constrained and it will be important to investigate whether their thermal evolution is also consistent with the dominant role of heating by 26Al decay.

3.6 Planetesimal accretion and evolution - the dominant role of 26Al heating

Hafnium-tungsten chronometry of meteorites as summarized above indicates that melting and differentiation occurred in bodies that accreted within the first ~1 Ma [par-ent bodies of the iron meteorites and possibly eucrites and angrites (Bizzarro et al., 2005)] whereas thermal metamorphism with peak temperatures of 900-1000 °C is char-acteristic for the ordinary chondrite parent bodies that formed shortly after ~2 Ma. The thermal histories of these meteorite parent bodies are consistent with heating by decay of 26Al as the dominant heat source. This is illustrated in Fig. 8, where the peak tem-perature in the centre of an asteroid is plotted against its (instantaneous) accretion age for parent body radii of 10 and 100 km. Also shown are the Hf-W ages for iron meteor-ites, which provide the latest possible time of accretion of their parent bodies, as well as ages for chondrules from various chondrite parent bodies, which provide the earliest possible time of parent body accretion. Fig. 8 predicts that, if 26Al was the dominant heat source in meteorite parent bodies, then the parent bodies of the weakly metamor-phosed carbonaceous chondrites should have accreted more than ~3 Ma after CAI for-mation.

Recent applications of Al-Mg chronometry to chondrules suggest that this indeed is the case. These studies show that chondrules from L and LL chondrites formed in a narrow interval of 1 to 2.5 Ma after CAI formation and that younger ages for L and LL chondrules reflect partial resetting by parent body metamorphism (Rudraswami and Goswami, 2007; Rudraswami et al., 2008). By contrast, chondrules from carbonaceous chondrites show no evidence for resetting by thermal metamorphism but their ages ex-

224

Page 225: Touboul Thesis Completed

tend to more than ~3 Ma after CAI formation. For instance, chondrules from the CO 3.0 chondrite Yamato 81020 have Al-Mg ages ranging from 1.7 to 3.0 Ma (Kunihiro et al., 2004; Kurahashi et al., 2008) and many chondrules from CR chondrites formed more than ~3 Ma after CAIs (Nagashima et al., 2008). If these young chondrule ages do not reflect resetting by low-temperature aqueous alteration on chondrite parent bodies, then the Al-Mg ages for chondrules from carbonaceous chondrites require that their parent bodies accreted more than ~3 Ma after CAI formation and, hence, later than the ordi-nary chondrite parent bodies. In this case, 26Al had decayed to a level too low to cause significant heating, consistent with the scarcity of highly metamorphosed specimens among the carbonaceous chondrites (Fig. 8).

The chronological data summarized above combined with results from thermal modeling (Hevey and Sanders, 2006; Sahijpal et al., 2007) indicate that the abundance of 26Al present at the time of parent body accretion is the essential factor controlling the early evolution of planetesimals: 26Al was sufficiently abundant to melt early-formed planetesimals (<1 Ma), whereas in the late-formed chondrite parent bodies (>2 Ma) too little 26Al remained to raise the temperatures high enough to cause differentiation (Kleine et al., 2005a; Schérsten et al., 2006). Moreover, heating that caused the thermal metamorphism of ordinary chondrites also is consistent with energy release from 26Al decay, as is the scarcity of metamorphosed carbonaceous chondrites, whose parent bod-ies accreted more than ~3 Ma after CAI formation (Fig. 8). Finally, crystallization ages of meteoritic basalts from the eucrite, angrite and mesosiderite parent bodies are re-markably consistent with the timescale for basaltic magmatism predicted by thermal models for asteroids heated by decay of 26Al (Sahijpal et al., 2007). Different thermal evolutions of meteorite parent bodies therefore largely reflect their initial 26Al abun-dance, which, due to the rapid decay of 26Al, is controlled by the time of parent body accretion.

These observations can be rationalized if the variation of accretion time across the asteroid belt is of the order of several 26Al half-lives (Grimm and McSween, 1993). Due to longer orbital periods and a decrease in surface density, the accretion time increases with increasing distance to the Sun (e.g., Weidenschilling, 1977) and this may account for the heliocentric zoning of the asteroid belt in igneous, metamorphic and unaltered asteroids, with igneous asteroids being located closest to the Sun (Gradie and Tedesco, 1982). Grimm and McSween (1993) considered the variation of accretion time as a function of semimajor axis and calculated the subsequent thermal histories for these planetesimals. The results show that the zonation of the asteroid belt and the thermal histories of meteorite parent bodies can be achieved by heating from 26Al decay. The results of these calculations are remarkably consistent with the chronological constraints

225

Page 226: Touboul Thesis Completed

for asteroid accretion and evolution as summarized above. However, Bottke et al. (2006) proposed that the parent bodies initially accreted within the terrestrial planet re-gion, where the fast accretion rates led to early melting even of small planetesimals, and were later scattered in the main asteroid belt. The Hf-W evidence for early core forma-tion in the parent bodies of magmatic iron meteorites is also consistent with this sce-nario. It is remarkable that some of the most primitive material that has been preserved in the meteorite record was accreted and processed last. Although this seems somewhat paradoxical at first glance, it reflects the fact that due to the intense heating by 26Al de-cay only late-formed material could preserve its primitive character.

In addition to 26Al heating, impacts may have been an important heat source in the thermal evolution of asteroids but they probably played only a minor role in the early high-temperature evolution of asteroids. This is because the Hf-W ages for high-temperature events such as core formation, magmatism and parent body-wide thermal metamorphism summarized above are entirely consistent with 26Al as the sole heat source. Nevertheless impact-related processes were an important heat source for the subsequent evolution of asteroids. Hafnium-tungsten ages for eucrite metals postdate CAI formation by ~20 Ma and probably reflect impact-triggered thermal metamorphism in the crust of the eucrite parent body. Hafnium-tungsten data for IAB irons appear most consistent with mixing of metals and silicates during reassembly of the IAB parent body following its impact disruption. The timing of this event, however, remains poorly con-strained.

4 TIMESCALES FOR CORE FORMATION AND EARLY MANTLE DIF-FERENTIATION IN MARS

4.1 182W-142Nd systematics of the Martian mantle

The first W isotope data for Martian meteorites were presented by Lee and Halli-day (1997) and additional analyses were reported by Kleine et al. (2004a) and Foley et al. (2005). The 182W/184W ratios of Martian meteorites mainly fall into two groups (Fig. 9). Basaltic shergottites have ε182W~0.3-0.6 and nakhlites and Chassigny (hereafter NC group) have ε182W~2-3 (Lee and Halliday, 1997; Kleine et al., 2004a; Foley et al., 2005). The only lherzolitic shergottite that has been analyzed for W isotopes so far is ALHA 77005 and its ε182W value of 0.91±0.32 appears to be slightly higher but is not well resolved from the basaltic shergottites (Foley et al., 2005). Orthopyroxenite ALH 84001 has ε182W = 0.49±0.33, similar to the values for basaltic shergottites (Foley et al., 2005). The W isotope data provide two important constraints. First, all Martian meteor-

226

Page 227: Touboul Thesis Completed

ites have elevated 182W/184W ratios relative to chondrites. This is not surprising because Hf and W can be strongly fractionated by large-scale parent body processes such as core formation and mantle melting. Second, there are 182W variations within the Martian mantle, indicating that, in spite of relatively young crystallization ages for most Martian meteorites (Nyquist et al., 2001), a signature of an early mantle differentiation (i.e., <60 Ma after CAI formation) has been preserved.

Application of Hf-W chronometry to date core formation in Mars requires knowl-edge of the 182W/184W ratio of the bulk Martian mantle. Estimating this value requires distinguishing W isotope variations that are due to early mantle differentiation from those that reflect core formation. This in turn requires independent estimates for the timing of mantle differentiation, which can be obtained from 146,147Sm-142,143Nd isotope systematics of Martian meteorites (Harper et al., 1995; Debaille et al., 2007a; Caro et al., 2008). All samples from the NC group studied so far have relatively constant ε142Nd values of ~0.6 (Harper et al., 1995; Caro et al., 2008). In contrast, shergottites show ε142Nd ranging from ~-0.3 to ~+0.7 (Fig. 9) (Debaille et al., 2007a; Caro et al., 2008). Whereas the shergottites define a planetary isochron in the ε143Nd-ε142Nd two-stage evolution diagram, the NC group plots outside the two-stage evolutionary field, indicat-ing that their source(s) had a more complex history (Borg et al., 2003; Caro et al., 2008; Debaille et al., 2008). The NC group has 146Sm-142Nd systematics similar to the de-pleted shergottites, whereas the 147Sm–143Nd systematics reveal differences. This could indicate a disturbance of the Sm-Nd system after extinction of 146Sm (Caro et al., 2008; Debaille et al., 2008). The elevated ε142Nd and ε182W values of the NC group neverthe-less provide evidence for an early differentiation event in their source(s).

Therefore, the timing of mantle differentiation in Mars is currently best dated by the 146,147Sm-142,143Nd systematics of shergottites, which were interpreted to indicate differentiation at 40±18 Ma (Caro et al., 2008) or, alternatively, differentiation in the interval from ~30 to ~100 Ma after CAI formation (Debaille et al., 2007a). This event could potentially have resulted in W isotope variations, such that the effects of silicate mantle differentiation on W isotopes need to be taken into account in any attempt to derive the 182W/184W of the bulk Martian mantle. Kleine et al. (2004a) and Foley et al. (2005) used the 142Nd/144Nd ratios of shergottites for this task, mainly because the Nd isotope data available at that time suggested that at least some of the shergottites have 142Nd/144Nd identical to that of the bulk Martian mantle, assumed to be ε142Nd=0 (i.e., identical to the terrestrial standard). However, two recent high precision 142Nd data sets for meteorites require a reassessment of this approach. Firstly, it was shown that chon-drites, assumed to be representative for the bulk composition of Mars and other terres-trial planets, have ε142Nd values ~20 ppm below those for the terrestrial standard (Boyet

227

Page 228: Touboul Thesis Completed

and Carlson, 2005). Secondly, new highly precise 142Nd data for shergottites show that the 142Nd/144Nd ratios of all shergottites deviate from ε142Nd=0 (Debaille et al., 2007a; Caro et al., 2008). Although some shergottites have ε142Nd values similar to those of ordinary chondrites [ε142Nd ~ -0.2 (Boyet and Carlson, 2005)], their ε143Nd and ε176Hf values are negative, indicating that these are crustal samples (Bourdon et al., 2008; Caro et al., 2008). Therefore, the similar ε142Nd values of these shergottites and chondrites seem fortuitous.

The 142Nd data for chondrites and shergottites show that, as a result of early man-tle differentiation, the 142Nd/144Nd ratios of all shergottites deviate from that of the bulk Martian mantle. It is therefore difficult to assess whether the 182W/184W ratio of any of the shergottites is representative for the bulk Martian mantle. However, in spite of their different ε142Nd values, basaltic shergottites have very similar ε182W values ranging from ~0.3 to ~0.6 (Foley et al., 2005), indicating that the event that caused the 142Nd variations among the shergottites did not result in significant 182W variations. This is consistent with the ~40 Ma 146,147Sm-142,143Nd age, if differentiation in the shergottite source(s) involved only limited Hf/W fractionations. Therefore, it seems reasonable to assume that the bulk Martian mantle has ε182W values between ~0.3 and ~0.6.

4.2 Age of the Martian core

The W model ages for the formation of the Martian core reported in the literature range from ~3 to ~12 Ma after CAI formation (Kleine et al., 2004a; Foley et al., 2005; Jacobsen, 2005). This range in ages primarily reflects the use of different Hf/W ratios for the bulk Martian mantle (Nimmo and Kleine, 2007). Based on Hf, Th and W con-centration data for Martian meteorites and chondrites the currently best estimate for the Hf/W ratio of the bulk Martian mantle is Hf/W~3-4 (see section 2.1). Using this value and ε182W ~0.3-0.6 for the bulk Martian mantle results in two-stage model ages for core formation ranging from ~0 Ma to ~8 Ma after CAI formation (Fig. 10a). As is evident from Fig. 10a, the uncertainty in these model ages is largely due to uncertainty in the Hf/W ratio of the bulk Martian mantle. Fig. 10a also demonstrates that the lowest possi-ble Hf/W ratio of the Martian mantle is ~3 because lower Hf/W ratios would result in negative core formation ages. The Martian mantle likely has a Hf/W ratio that is only slightly higher than this minimum value of ~3, such that the Hf/W ratio must be deter-mined very precisely to yield core formation ages of higher precision.

A different approach for determining the age of the Martian core was employed by Halliday and Kleine (2006), who argued that core formation in Mars might have started within the first Ma of CAI formation. These authors observed that the 182W/184W ratios of SNC meteorites are broadly correlated with their Th/W, which they used as a

228

Page 229: Touboul Thesis Completed

proxy for the Hf/W ratio in the sources. If this is correct, then the W isotope heterogene-ity in the Martian mantle might be due to different degrees of siderophile element deple-tion in the shergottite and NC sources. Later work by Nimmo and Kleine (2007), how-ever, could not identify a systematic difference between the Th/W ratios of shergottites and nakhlites (+chassignites).

Nimmo and Kleine (2007) showed that in the case of Mars two-stage model ages may underestimate the time Mars reached ~90% of its final mass by a factor of more than 3. For instance, for a two-stage model age of ~6 Ma, which is at the upper end of the age range shown in Fig. 10, Mars might have completed 90% of its accretion as late as ~20 Ma. Therefore, Hf-W chronometry can currently not distinguish between scenar-ios in which Mars formed within ~1 Ma during the runaway growth stage of planetary accretion (Halliday and Kleine, 2006) and scenarios in which accretion was more pro-tracted (Nimmo and Kleine, 2007)

4.3 Early mantle differentiation in Mars

The range of ε182W values observed in Martian meteorites combined with the time constraints provided by their 146,147Sm-142,143Nd systematics can be used to infer the Hf/W ratios of the source areas of the shergottites and nakhlites (Foley et al., 2005). Combined with mineral-melt partition coefficients for Hf and W (Righter and Shearer, 2003) this can then be used to constrain the source mineralogy. This is illustrated in Fig. 10b, which shows a possible (but not unique) scenario for the W isotope evolution of the Martian mantle. In this model it is assumed that the bulk Martian mantle has a pre-sent-day ε182W~0.4 and 180Hf/184W = 4. This yields ε182W~0.2 at ~40 Ma, i.e., at the time of mantle differentiation. The corresponding two-stage model age for core forma-tion is ~4 Ma.

The enriched and depleted shergottites might have slightly different W isotope compositions and these could have been produced by Hf-W fractionation during mantle differentiation at ~40 Ma. The inferred 180Hf/184W ratio for the source of the depleted shergottites is ~8, whereas this ratio is ~2 in the source region of the enriched shergot-tites (Fig. 10b). This is consistent with 176Lu-176Hf constraints on the evolution of sher-gottites, which were interpreted to indicate the presence of garnet in the source of the enriched shergottites and derivation of the depleted shergottites from an already de-pleted source (Blichert-Toft et al., 1999). Garnets have DHf/DW~30, such that residual garnet in the source will result in low Hf/W ratios in the melts, as observed for the en-riched shergottites. Likewise, elevated 180Hf/184W in the source of the depleted shergot-tites are consistent with derivation from an already depleted source.

229

Page 230: Touboul Thesis Completed

An important observation is that, if Hf-W fractionation in the NC source occurred at the same time as differentiation in the shergottite source(s), then the NC source must have evolved with very high 180Hf/184W to reach their present-day ε182W of ~3 (Fig. 10b). The inferred 180Hf/184W~56 is consistent with 180Hf/184W~57 as calculated for a majorite-rich garnet-bearing deep mantle based on mineral-melt partition coefficients for Hf and W (Righter and Shearer, 2003). Therefore, if the initial differentiation of the NC source occurred as late as ~40 Ma, then the radiogenic W isotope composition of these samples would require a deep mantle source for the NC group. This is consistent with the 147Sm-143Nd and 176Lu-177Hf systematics of nakhlites (Debaille et al., 2007b).

5 HF-W CHRONOMETRY OF THE MOON

The early evolution of the Moon is characterized by the crystallization of a lunar magma ocean (LMO) in chemically distinct layers (Shearer and Papike, 1999). The first minerals to crystallize from the LMO were mainly olivine and orthopyroxene and were followed by plagioclase that floated to the surface, forming the ferroan anorthosites. Towards the end of the crystallization sequence ilmenite and clinopyroxene precipi-tated, until the residual liquid of the magma ocean, termed KREEP (for enrichments in K, REE, and P), solidified. Subsequent melting and mixing among these primary rocks produced the variety observed in the lunar sample suite. For example, mare basalts formed by remelting of early mafic cumulates, which in the case of high-Ti mare basalts included assimilation of ilmenite and clinopyroxene. The redistribution of KREEP dur-ing impacts on the lunar surface resulted in the contamination of most highland breccias with various amounts of the KREEP-component.

The 182Hf-182W chronometer is potentially well-suited for determining the time-scale of LMO crystallization because substantial Hf-W fractionation occurred during magma ocean solidification. This results from the high incompatibility of W and the compatibility of Hf in clinopyroxene and ilmenite, leading to low Hf/W in KREEP and complementary high Hf/W in the high-Ti mare basalt source (Shearer and Newsom, 2000; Righter and Shearer, 2003; Kleine et al., 2005c). If these Hf-W fractionations occurred during the effective lifetime of 182Hf, variations in the 182W/184W of lunar man-tle reservoirs can be expected.

Lee et al. (1997) reported the first W isotope data for lunar whole-rock samples and found variations in 182W/184W ranging from ~0 to ~7 ε182W. These were interpreted to indicate formation of the Moon ~54 Ma after formation of CAIs (Lee et al., 1997). Later work, however, revealed that elevated 182W/184W ratios in lunar samples largely reflect the presence of cosmogenic 182W. Nevertheless, small 182W variations that were

230

Page 231: Touboul Thesis Completed

attributed to 182Hf decay were thought to be present and implied formation of the Moon within the first ~60 Ma after CAI formation (Leya et al., 2000; Lee et al., 2002; Kleine et al., 2005c). However, Touboul et al. (2007, 2008b) showed that all their measured lunar samples have constant 182W/184W ratios and that elevated 182W/184W ratios in lunar samples entirely reflect cosmogenic 182W production.

5.1 Cosmogenic vs. radiogenic 182W in lunar samples

Lunar samples collected during the Apollo and Luna missions all derive from the surface of the Moon and during their residence at the lunar surface these rocks were exposed to cosmic rays that could have produced 182W via the reaction 181Ta(n,γ)182Ta followed by β- decay to 182W (Leya et al., 2000). Thus, elevated 182W/184W ratios of lunar whole-rock samples can reflect contribution from two sources: decay of 182Hf and cosmogenic production of 182W. Cosmogenic 182W production is significant in lunar samples because of their long exposure times to cosmic rays combined with their high Ta/W ratios, making cosmogenic 182W a dominant component in lunar samples (Fig. 11). Lee et al. (2002) showed that in some high-Ti mare basalts the ε182W of mineral separates is correlated with their Ta/W ratios, providing clear evidence for the presence of cosmogenic Ta-derived 182W in these samples (Fig. 11). Lee et al. (2002) regressed their Ta-W data to determine the purely radiogenic 182W/184W at Ta/W=0. Although this approach yielded rather imprecise corrected 182W/184W ratios with uncertainties of ~2-3 ε units, these corrected 182W/184W ratios revealed variations among the investigated mare basalts that possibly could be as large as ~7 ε units (Lee et al., 2002).

The most direct method to determine the 182W/184W of lunar samples devoid of any cosmogenic 182W is to analyze minerals that do not contain any Ta and hence no Ta-derived cosmogenic 182W. This condition is met by metals, which are found in trace amounts in most lunar samples (Kleine et al., 2005c). Relatively high Ni and Ir contents in metals from highland samples indicate that these are meteoritic metals added to the lunar surface by impacts. Unlike meteoritic metals, however, the metals from KREEP–rich highland rocks have extremely high contents of W [~30 ppm (Kleine et al., 2005c)], reflecting the strong enrichment of W in KREEP. Partitioning of W from KREEP into these metals probably occurred by impact–induced melting and brecciation during redistribution of KREEP at the lunar surface. Metals in mare basalt consist of native Fe and probably formed by crystallization from a silicate melt under reducing conditions. The composition of these metals, particularly their low Ni contents, distin-guishes them from meteoritic metal added to the lunar surface by impacts.

Kleine et al. (2005c) reported the first W isotope data for lunar metals and found small variations in the 182W/184W ratios of metals from high-Ti mare basalts and

231

Page 232: Touboul Thesis Completed

KREEP-rich samples. However, Touboul et al. (2007) recently presented W isotope data for high-purity metal separates from a comprehensive set of lunar mare basalts. These new data reveal that low- and high-Ti mare basalts have identical W isotope compositions that are also identical to the 182W/184W of KREEP. This result is inconsis-tent with the earlier reported W isotope data for magnetic separates for two high-Ti mare basalts (Kleine et al., 2005c) but Touboul et al. (2007) could show that these mag-netic separates probably contained a small fraction of ilmenite with high Ta/W and hence cosmogenic 182W. After correction for cosmogenic 182W, these high-Ti mare ba-salts have 182W/184W indistinguishable from the metals of all other mare basalts (Touboul et al., 2007).

Mineral separates from low-Ti mare basalt 15555, in spite of having variable Ta/W ratios, show no resolvable difference in their ε182W values averaging at 1.4±0.4 (Lee et al., 2002). This elevated 182W/184W appears inconsistent with ε182W~0 for all metals from mare basalts but Touboul et al. (2008b) reported a somewhat lower ε182W of 0.9±0.4 and Ta/W~4.6 for sample 15555 (Fig. 11b). Using correction equations (Leya et al., 2000) and the Sm isotope composition of sample 15555 (Nyquist et al., 1995; Rankenburg et al., 2006), the calculated cosmogenic ε182W at Ta/W~4.6 is ~0.9, remarkably consistent with the measured W isotope composition as reported by Toub-oul et al. (2008). This is also consistent with the ε182W value and Ta/W ratio reported for a 15555 whole-rock analysis by Kleine et al. (2005) and indicates that the W isotope composition of mare basalt 15555 is indistinguishable from ε182W~0 (Fig. 11b). The reason for this systematic offset reported by Lee et al. (2002) for the same sample re-mains unclear.

Similarly, Lee et al. (1997) reported a positive 182W anomaly of 3.1±1.7 ε182W in ferroan anorthosite 60025. This sample has an exposure age of ~2 Ma, such that its ele-vated 182W/184W cannot reflect the presence of cosmogenic 182W. However, more re-cently Touboul et al. (2008b) obtained new W isotope data for pure plagioclase sepa-rates from the two ferroan anorthosites 60025 and 62255 and in contrast to the earlier study by Lee et al. (1997) did not find any resolvable 182W excess in these samples (Fig. 12). The reason for this discrepancy is not known but could potentially reflect the pres-ence of cosmogenic 182W in a mafic component of anorthosite 60025, which is not pre-sent in the pure plagioclase separates analyzed by Touboul et al. (2008b). Therefore, the lunar crust and mantle have indistinguishable W isotope compositions that are also in-distinguishable from that of the terrestrial mantle (Fig. 12).

232

Page 233: Touboul Thesis Completed

5.2 Lifetime of the lunar magma ocean

The homogeneous W isotope composition of the mare basalt sources, KREEP and the lunar crust in spite of their strongly fractionated Hf/W ratios (Fig. 13) indicates that the last equilibration of W isotopes in the LMO occurred after the effective life-time of 182Hf (Touboul et al., 2007; Touboul et al., 2008b). At ~60% crystallization and beyond the viscosity of the crystal-melt mixture in a magma ocean increases strongly, such that mixing (and isotopic equilibration) becomes very slow (Solomatov, 2000) The W iso-tope data could therefore be interpreted to indicate that ~60% crystallization of the LMO was achieved later than ~60 Ma after CAI formation. However, the minerals re-sponsible for Hf-W fractionation in the LMO (i.e., high-Ca pyroxene and ilmenite) form at the very end of the crystallization sequence, most likely after ~80% crystallization (Shearer and Papike, 1999). Therefore, the Hf-W data could also be interpreted to indi-cate that >80% crystallization was achieved later than at ~60 Ma. In any case, crust formation by plagioclase flotation occurs after ~70% crystallization of the LMO, such that the 147Sm-143Nd ages for ferroan anorthosites (Carlson and Lugmair, 1988; Norman et al., 2003; Nyquist et al., 2006) reveal that ~70% LMO crystallization was achieved at 112±40 Ma, entirely consistent with the Hf-W constraints.

Although the revised Hf-W chronology of the LMO (Touboul et al., 2007; Toub-oul et al., 2008a) now appears to be consistent with the ~200 Ma 146Sm-142Nd age of the lunar mantle isochron (Nyquist et al., 1995; Rankenburg et al., 2006), the latter proba-bly does not reflect the timing of magma ocean solidification. Bourdon et al. (2008) recently showed that cumulate overturn, magma mixing and melting following lunar magma ocean crystallization at 50-100 Ma could produce an array in the 142Nd/144Nd-147Sm/144Sm diagram that yields an ~200 Ma model age. Therefore, the crystallization timescale of the LMO appears to be best constrained by the combined Hf-W and 147Sm-143Nd isotope systems that indicate that ~70% crystallization could have been achieved at ~60 Ma at the earliest and at ~150 Ma at the latest (Touboul et al., 2007; Touboul et al., 2008b).

6 TIMESCALES FOR ACCRETION AND DIFFERENTIATION OF EARTH

The application of the 182Hf-182W system as a chronometer for the formation of Earth's core goes back to the work of Harper et al. (1991) and the first comprehensive investigation of W isotope evolution during core formation in Earth was given by Harper and Jacobsen (1996). At that time, however, important parameters for calculat-ing core formation ages had not been determined yet. These include the 182W/184W ratio of chondrites and the initial 182Hf/180Hf at the time of CAI formation (or another well-

233

Page 234: Touboul Thesis Completed

defined point in time). These parameters were first determined by Lee and Halliday (1995, 1996) but their results were revised in 2002 (Kleine et al., 2002; Schoenberg et al., 2002a; Yin et al., 2002). Since then Hf-W chronometry has been utilized in various accretion and core formation models, which are summarized in Table 4.

In the simplest model of core formation it is assumed that the core formed instan-taneously and an age for core formation can then be calculated by assuming a two-stage model. Using the 182W excess of Earth's mantle of Δε182W = +1.9±0.1 relative to chon-drites and a Hf/W ratio of ~17 for the bulk silicate Earth (see section 2.1), results in a two-stage model age for core formation of ~30 Ma (equation 5; section 0). However, at least for bodies as large as the Earth, core formation did not occur as a single event but took place episodically during planetary growth, which probably lasted several tens of millions of years (Halliday et al., 1996; Harper and Jacobsen, 1996; Halliday, 2004; Kleine et al., 2004b; Jacobsen, 2005; Nimmo and Agnor, 2006). Thus, for the Earth the assumption of instantaneous core formation is not justified, and the two-stage model age of ~30 Ma would only date core formation if during this event the entire core was first remixed and homogenized with the entire mantle before segregating again. This seems physically implausible.

The two-stage model age nevertheless provides important age constraints. It as-sumes that Earth's mantle had a chondritic W isotope composition at the time of its for-mation. This corresponds to the minimum 182W/184W Earth's mantle could have had at any time because core formation in Earth cannot result in a mantle with a subchondritic Hf/W ratio. Therefore, the two-stage model age corresponds to the earliest time core formation could have been complete (Halliday et al., 1996; Kleine et al., 2004b), imply-ing that core formation in the Earth could not have ceased before ~30 Ma after CAI formation.

In models of continuous evolution of a reservoir, such as core formation, there is no single "age" for core formation and the reported ages rather correspond to a certain growth stage of the core. In this regard, two different ages have proven to be useful: (i) the mean time of core formation, < tcf >, which corresponds to ~63% of the core (in the case of exponentially decaying rate of accretion) (cf. Harper and Jacobsen, 1996), and (ii) the time of the Moon-forming impact, which most likely determines the termination of the major stage of Earth's accretion and core formation. Obtaining these age con-straints requires an understanding of the mechanisms of core formation in Earth because the magnitude of the W isotope effects in the bulk silicate Earth depend on the type of mechanism by which metal is transported to the core and the degree to which isotopic equilibrium is achieved during metal segregation (Harper and Jacobsen, 1996). This has been addressed in several W isotope evolution models for Earth (Halliday, 2004; Kleine

234

Page 235: Touboul Thesis Completed

et al., 2004b; Jacobsen, 2005; Nimmo and Agnor, 2006) but these models yield different apparent timescales for core formation in Earth, which largely reflects different assump-tions regarding the degree and processes of metal-silicate re-equilibration during core formation.

6.1 Models of core formation and metal-silicate equilibration

6.1.1 Mechanisms of core formation

Two lines of evidence suggest that Earth accreted mostly from planetesimals and planetary embryos already differentiated into mantle and core. First, the Hf-W data for meteorites reveal that differentiation of many planetesimals occurred within the first few Ma of the solar system. Second, current models for the accretion of terrestrial plan-ets suggest that planetary embryos formed within the first Ma of the solar system, in which case they will have differentiated owing to heating from decay of abundant 26Al. Therefore, there are two endmember models for the formation of Earth’s core during growth from pre-differentiated bodies: either Earth's core formed by merging of metal cores from pre-differentiated objects or the cores of these objects dispersed as small metal droplets in the terrestrial magma ocean prior to sinking to join the existing core.

Which of these scenarios is most appropriate for a given collision depends on the relative size of the two colliding bodies. Collisions in which the impactor is much smaller than the target led to vaporization of the impactor, in which case the impactor material can efficiently mix and homogenize within the magma ocean of the target. In this case any information on the differentiation of the impactor is lost, such that the chemical and isotopic consequences for Earth’s mantle are identical to those for addi-tion of an undifferentiated, chondritic body.

What happens in detail during larger collisions, however, is less well understood. Hydrocode simulations of giant impacts (Canup and Asphaug, 2001) show that the cores of target and impactor merge rapidly although more recent simulations indicate that some re-equilibration might occur (Canup, 2004). The problem is that these simula-tions currently provide a resolution on the order of 100 km, whereas the length-scale on which chemical and isotopic re-equilibration occurs is probably on the order of centime-ters (Stevenson, 1990). An alternative approach for constraining the degree of re-equilibration is provided by the abundances of siderophile elements in Earth's mantle. The observed depletions of several siderophile elements in Earth's mantle are consistent with metal-silicate equilibration at high temperatures and pressures in a terrestrial magma ocean (cf. Rubie et al., 2003). This equilibration can be achieved only if the metal cores of the impactor emulsified as small metal droplets in a magma ocean (Rubie

235

Page 236: Touboul Thesis Completed

et al., 2003; Hoink et al., 2006). If this is correct, then efficient W isotope re-equilibration between impactor core and target mantle can be expected. However, Sa-saki and Abe (2007) observed that even in this case re-equilibration might not have been complete because the metal droplets might not have been distributed throughout the entire mantle. Furthermore, full equilibration may have been also limited if the ter-restrial magma ocean was not global.

The 182W/184W ratio of Earth's mantle at time t2 immediately after accretion of a new object at t1 can be expressed as a mixture of three components as follows:

( ) ( ) ( ) ( )BSE BSE 1 I I 1 I I 1BSE 2

BSE I I

mantle mantle core core

mantle core

y R t y R t k y R tR t

y y k y⋅ + ⋅ + ⋅ ⋅

=+ + ⋅

(9)

where BSE denotes bulk silicate Earth, Ri is the 182W/184W ratio, i i Wy m C= is the

amount of W in an object i, and k is the fraction of the impactor's core that first equili-brates with the Earth's mantle before entering Earth's core. Fig. 14 schematically illus-trates the W isotope evolution of Earth’s mantle in the different modes of core forma-tion. In the core-merging model (Fig. 14c), metal-silicate re-equilibration does not occur and the W isotope composition of the target’s mantle immediately after the impact re-sults from addition of impactor mantle material to the target’s mantle. The resulting 182W/184W ratio will always be higher than chondritic because no core material (with subchondritic 182W/184W) is involved in these mixing processes. In the case of core merging, no isotopic record of the collision was generated and the W isotope effect in Earth’s mantle would largely reflect the timing of core formation in the pre-merged ob-jects. By contrast, small metal droplets in a magma ocean could have equilibrated effi-ciently with the surrounding molten silicates (Fig. 14a,b). If re-equilibration was com-plete this is equivalent to adding an undifferentiated object to Earth’s mantle, followed by metal segregation. In this case the resulting W isotope effect reflects the timing of core formation in Earth.

6.1.2 Metal-silicate equilibration in a magma ocean

An additional complication is that the process by which re-equilibration occurs is not well understood and two different models have been used to calculate the effects of full equilibration. In the fully equilibrative plumbing model, Harper and Jacobsen (1996) assume that, during a single impact, each infinitesimal parcel of the impactor fully re-equilibrates with the target's mantle before the next parcel arrives (Fig. 14a). By contrast, other models assume that the entire impactor equilibrates with the entire target mantle at one go (Fig. 14b) (Kleine et al., 2004b; Nimmo and Agnor, 2006; see also equation 8). Thus, these two models essentially assume fractional and batch re-

236

Page 237: Touboul Thesis Completed

equilibration, respectively. Here we summarize the important equations governing these two different equilibration scenarios and then evaluate the reduction in W isotope ef-fects that is achieved in both equilibration models during giant impacts. Fractional equilibration model. The fully equilibrative plumbing model of Harper and Jacobsen (1996) was termed magma ocean differentiation model by Jacobsen (2005). In this model the reservoir content (Nj2) of a stable species j in the magma ocean is given by equation 37 of Jacobsen (2005):

/

2 1 2 22

21 1

m sj j j

jdN C M D M N

dt Mγ

γ γ= −

− −

& & (10)

Here the subscripts 1, 2, and 3 refer to bulk, mantle, and core, Cj1 is the bulk Earth concentration, Dj

m/s is the metal/silicate partition coefficent (considered constant), γ is the core mass fraction (also considered constant), M2 is the mass of the mantle and

2 2M dM dt=& . With these simplifying assumptions, the concentration of j in the mantle

is

( )

2/

1

11 1

jm s

j j

CC D γ

=− +

(11)

This is equivalent to the expression for static models of trace element partitioning (cf. Righter and Drake, 1997) and is also identical to that for a batch equilibration model (see equ. (3) in Nimmo and Agnor, 2006). Jacobsen (2005) showed that the W isotopic evolution of the magma ocean (equ. 72) can, for a very short interval (i.e., no decay) such as directly after an impact, be simplified to

W 2W

2

ε (1 ) εHf Wd Mfdt MΔ

= − + Δ&

(12)

Thus, the difference between the initial (ti) and final states (tf) after an impact is given by

( )( )

( )( )

1W 2

W 2

ε

ε

Hf Wff i

i f

t M tt M t

+⎛ ⎞Δ⎜ ⎟=⎜ ⎟Δ ⎝ ⎠

(13)

Batch equilibration model. Consider a collision between two bodies of mass mA and mB, respectively, where m

B

BB<mA. The core mass fraction in each case is γ and the ratio mB/mB A = g. During the collision, it is assumed that the metal + silicates of body B are com-pletely homogenized with the silicates of body A, and that the metallic fraction from body B then separates from this reservoir. During this separation process elements are partitioned according to the metal/silicate partition coefficient, Dj

m/s = C /C , where C and C are the concentrations in the metal and silicate fractions, respectively. From

met sil

met sil

237

Page 238: Touboul Thesis Completed

mass balance considerations, it can be shown that the final concentration of an element j in the mantle of the resulting object 2

fjC is given by

( )

( )2 1

2 /

1

1 1

ij jf

j m sj

C gCC

g D

γ

γ γ

− +=

1⎡ ⎤− + + −⎣ ⎦ (14)

where is the initial concentration of the element j in the mantle of body A, and

is the bulk concentration of the element in body B (assumed chondritic). 2

ijC

1jC

If Djm/s = 1 then equation (14) shows that the final concentration is just the mass-

weighted average of the initial concentrations, as required. In the limit as the final concentration equals the initial concentration, as required. In the limit as

(i.e., the pre-existing silicate reservoir is vanishingly small), we recover equation (39) of Jacobsen (2005).

0g →g → ∞

After some additional algebra, it may be shown that the final 182W anomaly, ( )W ftεΔ , immediately after the impact, is given by

( ) ( )W W /1ε ε

1 11

f i m sj

t tD

γ

Δ = Δ⎛ ⎞

+ +⎜ ⎟⎜ ⎟−⎝ ⎠

(15)

where ( )W itεΔ is the 182W anomaly of the silicate portion of body 1 immediately prior

to the impact. Recognizing that ( )/ / 1m s

jf D γ γ= − (Jacobsen 2005; equ. 49) we may rewrite

equation (15) as

( ) ( ) ( )W W /

1ε ε1 1

f i Hf Wt t

g fΔ = Δ

+ + (16)

which for yields 1g

( )( )

W /

W

ε1 (1 ) ...

εf Hf W

i

tg f

t

Δ≈ − + +

Δ (17)

Similarly, recognizing that 2 2( ) / ( ) 1 (1 )i fM t M t g= + equation (13) (i.e., the re-

duction of the W isotope effects in the fractional equilibration model) may be written

( )( )

/1W /

W

ε 1 1 (1 ) ...ε 1

Hf Wff Hf W

i

tg f

t g

+Δ ⎛ ⎞= ≈ − +⎜ ⎟Δ +⎝ ⎠

+ (18)

where the approximation is appropriate for 1g . Thus, in the limit of small g the two

approaches give the same answer, as required. Fig. 15 compares the reduction in W isotope effects calculated for different im-

pactor to target ratios and using the fractional and batch equilibration models. As ex-pected, for small impactor to target ratios the reduction is small and the effects are es-

238

Page 239: Touboul Thesis Completed

sentially identical for both the fractional and batch equilibration models. For large im-pactor to target ratios, however, the reduction in the W isotope effects is less pro-nounced for batch equilibration compared to the fractional equilibration. An important observation is that for fractional equilibration and impactor to target ratios of larger than ~0.2, the W isotope effects is reduced to less than ~0.1, such that very large collisions might reduce the ΔεW value of Earth's mantle to almost the chondritic value.

6.2 Hf-W systematics in continuous core formation models

6.2.1 Exponential accretion

One may assume that the accretion of Earth took place at an exponentially decay-ing accretion rate (Jacobsen and Harper, 1996), such that:

( ) 1 t

E

m t eM

α−= − (19)

where m/ME is the cumulative fractional mass of the Earth at time t and α is the time constant of accretion. The mean time of accretion, <t>, is the inverse of the time constant and corresponds to the time taken to achieve ~63% accretion.

Fig. 16a shows the growth curve of Earth, calculated by assuming growth at an exponentially decaying rate and mean times of accretion of 8, 11 and 20 Ma. Using these growth curves and equation (9) the 182W/184W evolution of Earth's mantle can be calculated for each value of α. Thus, in principle, the accretion rate of Earth can be de-termined by finding the α value that yields the present day 182W/184W ratio of Earth's mantle. This is illustrated in Fig. 16b where the W isotope evolution curves for Earth's mantle are shown for different mean times of accretion. In these calculations it is as-sumed that the metal cores of newly accreted planetesimals always entirely re-equilibrated in the terrestrial magma ocean (by the batch equilibration process described above). This is a reasonable assumption if accretion occurred by the incremental growth of small mass fractions but a more realistic view is that Earth accreted by several dis-tinct large collisions that may have occurred at an exponentially decaying rate. In this case, the degree of metal-silicate equilibration is less well constrained.

Fig. 16b reveals that the present-day ΔεW of Earth's mantle of +1.9 can be pro-duced if the mean time of accretion is ~11 Ma. This corresponds to ~63% growth at ~11 Ma and ~90% growth at ~25 Ma. Note that in these models the end of accretion is not well defined because there is a small exponential tail of accretion that is technically still continuing today (Harper and Jacobsen, 1996). Effects of incomplete metal-silicate re-equilibration. Fig. 16b also illustrates the effect of incomplete metal-silicate re-equilibration on the W isotope evolution of Earth's man-

239

Page 240: Touboul Thesis Completed

tle. Metal-silicate re-equilibration during core formation results in a decrease of the 182W excess in Earth's mantle that had previously accumulated due to 182Hf decay. For a given accretion rate, a decreasing degree of metal-silicate re-equilibration (i.e., de-creasing k values) will result in an increasingly radiogenic W isotope composition of Earth's mantle. For instance, if 75 % instead of 100 % of the newly accreted core mate-rial first equilibrated with Earth's mantle before entering Earth's core, the ΔεW value of Earth's mantle would be +2.8 instead of +1.9 (for a given <t> of 11 Ma). Thus, to match the present-day W isotope composition of Earth's mantle, a decreasing degree of metal-silicate re-equilibration must be accompanied by longer accretion timescales. This is illustrated in Fig. 17a, where the calculated age of core formation is plotted as a func-tion of the degree of metal-silicate re-equilibration k. As is evident from this figure, the calculated core formation ages strongly depend on the assumed value for k. Changes in metal-silicate partition coefficient of W. Similar effects are observed if it is assumed that the Hf/W ratios in the mantles of newly accreting planetesimals were dif-ferent from the value observed in Earth’s mantle today. This is likely because the metal-silicate distribution coefficient for W (DW) is sensitive to many parameters such as pres-sure, melt compositions, and particularly oxygen fugacity (see section 2.1). These pa-rameters likely were different in the various objects that accreted to Earth and also probably changed over time, such that considerable variability in the mantle Hf/W ratios is expected (Righter, 2003; Wade and Wood, 2005). Wade and Wood (2005) considered changes in pressure, temperature and oxygen fugacity during accretion. They found the best match to the observed mantle depletions in siderophile elements if the oxygen fu-gacity of Earth's mantle increased during accretion. In their model, DW started off at values >104 and decreased to ~20 at the end of accretion. This would also be consistent with heterogeneous accretion scenarios that predict that Earth was initially highly reduc-ing and became more oxidized by the late addition of volatile-rich material (Wänke and Dreibus, 1988).

However, Halliday (2004) proposed the opposite and suggested that Earth ac-creted from relatively oxidized planetesimals having low Hf/W ratios, similar to the present-day values observed for the Martian mantle. Such a scenario has important con-sequences for interpreting the W isotope record of Earth's mantle because a mantle with a relatively low Hf/W ratio will never develop a large 182W excess. Therefore, upon accretion of such objects to Earth, the W isotope composition of Earth’s mantle can stay at a relatively low value even without significant metal-silicate re-equilibration. Thus, changes in the Hf/W ratios of the mantles of newly accreting objects and different val-ues of k have similar effects on the W isotope evolution of the Earth's mantle. This is illustrated in Fig. 17b, where for a given core formation age (calculated assuming an

240

Page 241: Touboul Thesis Completed

exponentially decaying rate of accretion) the Hf/W ratio in the mantle of the accreting planetesimals is plotted against the degree of metal-silicate re-equilibration k. This plot illustrates that, for a given core formation age (in this case <t>=11 Ma), the present-day 182W/184W of Earth's mantle can be obtained by assuming several distinct pairs of these two parameters. For instance, assuming k=1 and Hf/W=20 in the mantles of the accret-ing planetesimals yields the same result as assuming k=0.7 and Hf/W=4 (Fig. 17b). Therefore, W isotope systematics alone cannot distinguish between scenarios involving a high degree of metal-silicate re-equilibration and those involving low Hf/W ratios in early-formed proto-planets.

6.2.2 Tungsten isotope evolution in N-body simulations of planetary accretion

In the models discussed above, it is assumed that accretion of the Earth occurred at an exponentially decreasing accretion rate. Dynamical modeling, however, suggests that the final stages of terrestrial planet formation is characterized by multiple and sto-chastic impacts that bring in large core masses at once (Agnor et al., 1999; Chambers, 2001). Jacobsen (2005) and Nimmo and Agnor (2006) presented a framework for im-plementing the evolution of W isotopes in N-body simulations of planetary accretion. Nimmo and Agnor (2006) showed that the physical and isotopic characteristics of Earth can be reproduced with the last large impact occurring at ~30 Ma. This result is similar to those of Jacobsen (2005), who showed that the W isotope composition of Earth's mantle is consistent with most of Earth's growth taking place in ~10 Ma and a final gi-ant impact at ~32 Ma. At the time of these studies it was thought that the Moon-forming impact occurred earlier than ~50 Ma (cf. Lee et al., 2002; Kleine et al., 2005) but a more recent study of the Hf-W systematics of the Moon showed that this is not necessarily the case (Touboul et al., 2007).

Therefore, we applied the approach of Nimmo and Agnor (2006) to two accretion simulations from Raymond et al. (2006) that were selected mainly on the basis that they are characterized by a relatively late formation of the Moon. In the first of these simula-tions, a giant impact resembling the Moon-forming collision occurs at ~125 Ma after CAI formation (Fig. 18a), in the second simulation it occurs at ~75 Ma after CAI forma-tion (Fig. 19a). These are near the upper and lower end of the 90

1062+− Ma age for the giant

impact as proposed by Touboul et al. (2007), which is described in more detail below. Another important difference between these two simulations is the history of the impac-tor, whose collision resulted in the formation of the Moon. In the first simulation the silicate part of the impactor had, at the time of its collision with proto-Earth, a W iso-tope composition that was less radiogenic than that of proto-Earth's mantle. This results from a relatively late collision of the impactor with another body at ~40 Ma (Fig. 18b).

241

Page 242: Touboul Thesis Completed

In the second simulation the impactor has highly radiogenic 182W at the time it collided with proto-Earth because this impactor was largely accreted and differentiated within the first few Ma (Fig. 19b).

In the simulations both a high and low degree of metal-silicate equilibration was assumed (Fig. 18b and Fig. 19b) and in the case of full equilibration the effects of both fractional and batch equilibration were calculated (Fig. 18c and Fig. 19c). These simula-tions reveal that, if the degree of metal-silicate equilibration was low throughout accre-tion, Earth’s mantle should exhibit a much more radiogenic 182W anomaly than ob-served. This suggests that substantial metal-silicate re-equilibration occurred but at what stage of accretion this occurred is less well constrained. In both simulations, the present-day ΔεW of Earth's mantle of +1.9 can be obtained for both a high and low degree of metal-silicate re-equilibration prior to the final, Moon-forming impact (Fig. 18b and Fig. 19b). Thus, the conditions of metal-silicate separation during most of Earth’s growth are only poorly constrained by the W isotope data and most of this information is erased by the final giant impact. In the two exemplary accretion scenarios considered here, the degree of metal-silicate re-equilibration must have been high during the final impact because otherwise the 182W anomaly of the bulk silicate Earth would be larger than observed. Whether this is a common feature of all accretion simulations that can reproduce the physical properties of Earth is unclear. In the case of full equilibration, fractional equilibration results in a stronger reduction of the W isotope effects (Fig. 18c and Fig. 19c), as expected from the discussion above.

In summary, the physical and isotopic characteristics of Earth can be reproduced in a variety of accretion and differentiation scenarios. Therefore, the 182W excess of Earth's mantle relative to chondrites cannot be used to firmly establish an age of core formation and models can be constructed that result in core formation ages as early as ~30 Ma or as late as ~125 Ma, for example. Knowing the degree of metal-silicate re-equilibration in particular and the processes by which equilibration occurred (fractional vs. batch) is critical for interpreting the W isotope record of Earth’s mantle and for dis-tinguishing among different scenarios for Earth’s accretion. Nevertheless, the Hf-W systematics provide important constraints on the processes of core formation because the present-day physical and isotopic characteristics of the Earth (and Mars) can only be reproduced if the degree of metal-silicate equilibration during core formation was rela-tively high (Fig. 21). This suggests that core formation largely occurred by the separa-tion and segregation of small metal droplets in a magma ocean.

242

Page 243: Touboul Thesis Completed

6.3 Tungsten model age for the Moon-forming impact

Tungsten isotopes can provide important constraints on the time at which the Moon formed. This is important not only for understanding the formation and early his-tory of the Moon but also provides information on the accretion rate of Earth because most theories for the formation of the Moon suggest that the giant Moon-forming im-pact was the last major event in Earth's accretion. If, as numerical simulations suggest, the Moon mainly consist of impactor mantle material (Canup, 2004), then its W isotope composition should reflect the accretion and differentiation history of the impactor. Therefore, that two such different objects as the proto-Earth and impactor would evolve to identical W isotope compositions in their mantles seems highly unlikely, given the evidence for widely different W isotope signatures in the meteorite record. This is also exemplified in the two accretion simulations shown in Fig. 18, where, regardless of the degree of re-equilibration and accretion history, impactor and Earth have very different 182W anomalies at the time of the Moon-forming impact. Therefore, the identical W isotope compositions of the lunar and terrestrial mantles could indicate that the Moon is largely derived from terrestrial mantle material but this is inconsistent with results from numerical simulations, all of which indicate that the Moon predominantly consists of impactor material. If this is correct, then equilibration of W isotopes between the Earth's mantle and proto-lunar material during or after the giant impact seems to be required. Such equilibration processes could involve metal-silicate re-equilibration between the impactor mantle and core, which would reduce a radiogenic W isotope signature in the proto-lunar material (Bourdon et al., 2008) but it is unlikely that this process fortui-tously resulted in a 182W/184W that is identical to that of the Earth's mantle. Therefore, the equilibration should have involved the Earth's mantle and could potentially have occurred via a shared silicate vapor atmosphere of the lunar magma disk and the terres-trial magma ocean (Pahlevan and Stevenson, 2007). This requires that W became effi-ciently vaporized to enter this atmosphere but the efficiency to which this is possible and hence W isotopes could equilibrate remains to be investigated.

The two-stage Hf-W model time for separation of the Moon from a chondritic reservoir is ~37 Ma (Touboul et al., 2007). This model age could either date formation of a small lunar core or, if the Moon does not possess a metal core, the isolation of the Moon from the proto-lunar circumterrestrial magma disk. In any case, this model age should correspond closely to the time of the giant impact. Note that this age cannot cor-respond to the time of core formation in the impactor - which might be expected if the Moon predominantly consists of impactor material - because the identical 182W/184W of the lunar and terrestrial mantles require either formation of the Moon from terrestrial material or Moon-Earth equilibration. Although the formation of the Moon occurred by

243

Page 244: Touboul Thesis Completed

a single event, such that the timing of this event can in principle be calculated using a two-stage model, the initial W isotope composition of the Moon most likely was higher than chondritic. This is because the Moon is predominantly derived from silicate mantle material having high Hf/W and therefore most likely radiogenic 182W/184W ratios (see for instance Fig. 18). Therefore, the Moon probably formed later than given by its two-stage model age of ~37 Ma.

Fig. 20 shows the difference in 182W/184W between the lunar and terrestrial man-tles that would have developed as a function of age and difference in Hf/W. As is evi-dent from this figure, core formation in Earth could have terminated as early as ~30-40 Ma only if the difference in Hf/W between the bulk silicate Moon and Earth is less than ~10 %. The current best estimates for these Hf/W ratios (Table 1) indicate that this dif-ference might be as high as ~50 % (i.e., fHf/W=0.5 in Fig. 20), in which case the Moon would have formed later than ~50 Ma after CAI formation (Touboul et al., 2007). The Hf/W estimates are based on the U/W and Th/W ratios of the lunar and terrestrial man-tles and assume that these ratios were not fractionated during melting processes within the mantle. It will important in future studies to assess whether this is correct and to determine the U/W and Th/W ratios of the lunar and terrestrial mantles to high preci-sion.

If the Hf/W ratios of the lunar and terrestrial mantles are indeed different, then the Moon formed later than ~50 Ma after CAI formation. Given that the giant impact most likely was the last major event in Earth's accretion (Canup and Asphaug, 2001), this age also provides an age for the termination of Earth's core formation and would be consis-tent with U-Pb age for the formation of Earth (Allègre et al., 1995; Galer and Goldstein, 1996) and with I-Pu-Xe constraints for the formation of Earth's atmosphere (Podosek and Ozima, 2000; Yokochi and Marty, 2005; Pepin and Porcelli, 2006). In this case, the U-Pb, Hf-W, and I-Pu-Xe chronometers converge to provide identical ages for what might be the major early differentiation of the Earth at ~4.50-4.42 Ga (Allègre et al., 2008).

6.4 Early mantle differentiation and implications of a non-chondritic bulk Earth

Hafnium-tungsten chronometry cannot be directly applied to constrain the time-scales of early mantle differentiation in Earth because no 182W variations have yet been detected in terrestrial samples. A possible exception are low 182W/184W ratios in a few Archean metasediments from Greenland but these were attributed to incorporation of meteoritic material into these sediments (Schoenberg et al., 2002b). Traces of an early mantle differentiation have been preserved in some Archean rocks and the slightly ele-vated 142Nd/144Nd ratios of these rocks reflect differentiation at 50-200 Ma after CAI

244

Page 245: Touboul Thesis Completed

formation (Harper and Jacobsen, 1992; Boyet et al., 2003; Caro et al., 2003; Caro et al., 2006). This differentiation probably occurred too late to have resulted in resolvable W isotope variations because these can only be produced in the first ~60 Ma. This is also consistent with the absence of resolvable 182W variations among lunar samples, which indicate differentiation later than ~60 Ma after CAI formation.

However, Boyet and Carlson (2005) observed that the ~20 ppm 142Nd excess of terrestrial samples relative to chondrites could be explained by an early mantle differen-tiation prior to ~30 Ma. This differentiation generated an early depleted reservoir (EDM) and an early enriched reservoir (EER) that since its formation has remained hid-den. In this model the W isotope composition of terrestrial samples would not reflect the composition of the bulk silicate Earth but rather that of the EDR. The Hf/W ratio of the EDR most likely would be elevated relative to that of the bulk silicate Earth because the EDR formed by melt depletion. Thus, in this model the bulk silicate Earth would have a Hf/W ratio and ΔεW value lower than those of the EDR (Hf/W~17, ΔεW ~1.9). However, as discussed in detail above, the earliest time a reservoir having Hf/W~17 and ΔεW ~1.9 could have formed is ~30 Ma after CAI formation. Such a reservoir could only have formed prior to ~30 Ma if its initial ΔεW would be negative, i.e., the 182W/184W would have been lower than chondritic at that time. This however cannot be achieved in Earth's mantle which is characterized by elevated Hf/W relative to chondrites. There-fore, mantle differentiation before 30 Ma, as deduced from 146Sm-142Nd systematics does not seem consistent with the Hf-W constraints.

The 146Sm-142Nd and 182Hf-182W constraints on the chronology of Earth's early differentiation provide a consistent picture if Earth's mantle is characterized by ε142Nd=0, as proposed by Caro et al. (2008), implying that Earth has a non-chondritic Sm/Nd ratio. If this is true, then the Hf/W ratio of bulk Earth most likely also is not chondritic, such that one of the basic assumptions for calculating core formation time-scales may not be valid. The magnitude of this effect is difficult to assess and depends on the process(es) that were responsible for generating non-chondritic ratios of refrac-tory elements in bulk Earth. If this is related to impact erosion of an early-formed crust or explosive volcanism, then W, due to its higher incompatibility relative to Hf, would be preferentially lost. Consequently, bulk Earth would be characterized by elevated Hf/W and 182W/184W ratios relative to chondrites and the two-stage model age of core formation would be >30 Ma, that is, younger than that calculated relative to a chondritic composition of bulk Earth. Therefore, uncertainties in the bulk composition of Earth constitute an additional uncertainty in calculating core formation ages but these effects are difficult to quantify. It is important to note that the age of the Earth and Moon esti-mated using the identical ε182W values of the terrestrial and lunar mantles [Touboul et

245

Page 246: Touboul Thesis Completed

al. (2007) and section 6.3] is insensitive to the actual bulk composition of the Earth and thus provides a more reliable estimate for the termination of Earth’s accretion.

7 CONCLUSIONS

Hafnium-tungsten chronometry of meteorites provides age constraints on several key events in the early evolution of the solar system and its planetary objects, and has led to the establishment of a new model for the accretion and early evolution of aster-oids. Important events that are being dated using the Hf-W chronometer include the formation of CAIs and chondrules as well as core formation, magmatism and thermal metamorphism in meteorite parent bodies (Fig. 22). The chronological data indicate that melting and differentiation was prevalent in planetesimals that accreted within the first ~1 Ma (parent bodies of the iron meteorites and possibly eucrites and angrites), thermal metamorphism is characteristic for bodies formed shortly after ~2 Ma (ordinary chon-drite parent bodies), and aqueous alteration is the dominant parent body process in as-teroids that formed more than ~3 Ma after CAI formation (some carbonaceous chondrite parent bodies). These accretion ages for asteroids are inversely correlated with the peak temperature reached in their interiors, indicating that the different thermal histories of meteorite parent bodies are determined by their accretion age, which determines their initial 26Al abundance. This is most consistent with a model that involves decreasing accretion rates across the asteroid-forming region that is of the order of several 26Al half-lives. Moreover, thermal models for asteroids heated by energy release from 26Al decay not only reproduce the peak temperatures reached inside the meteorite parent bodies but are also consistent with the timescales of thermal metamorphism and cooling of chondrite parent bodies as well as with the time span of basaltic magmatism on aster-oids.

For terrestrial planets, the timescales of accretion and primordial differentiation are much longer, and determined by different parameters. This is consistent with dy-namical models of planetary accretion that indicate that the final stage of terrestrial planet formation is characterized by multiple large and stochastic impacts. Although the Hf-W timescale for the formation of Mars is currently uncertain and would allow core formation ages ranging from 0-20 Ma, Mars clearly formed much earlier than the Earth and Moon. The Hf-W ages for termination of Earth's accretion and core formation range from ~30 Ma to more than ~100 Ma after CAI formation. This range in ages primarily reflects different assumption regarding the mechanisms of accretion and core formation. Unless their Hf/W ratios differ by less than ~10 %, the identical 182W/184W ratios of the lunar and terrestrial mantles provide an independent age estimate for the Moon-forming

246

Page 247: Touboul Thesis Completed

giant impact of Ma. This time constraint is consistent with U-Pb ages for the for-

mation of Earth and I-Pu-Xe ages for the formation of Earth's atmosphere, such that these chronometers combine to provide evidence for the formation and major differen-tiation of the Earth at ~4.50-4.42 Ga.

901062+

The preservation of different 182W abundances in the Martian mantle contrasts with the constant 182W/184W ratios of lunar and terrestrial rocks and indicates that the global differentiation of the Martian mantle occurred when 182Hf was still extant and any later process was not efficient enough to re-homogenize the Martian mantle. This is consistent with 146,147Sm-142,143Nd systematics of Martian rocks. In contrast solidifica-tion of the lunar and terrestrial magma ocean occurred later, after extinction of 182Hf, consistent with the 50-200 Ma 146,147Sm-142,143Nd age for differentiation of Earth's man-tle.

The Hf-W timescale of planetary accretion is remarkably consistent with predic-tions from dynamical simulations. That is, the first planetary objects formed within less than 1 Ma, whereas the final assembly of the terrestrial planets may have taken as long as ~100 Ma. Moreover, implementation of Hf-W systematics in N-body simulations of planetary accretion indicates that the physical and isotopic characteristics of Earth can successfully be reproduced in such simulations.

Acknowledgements - T. Kleine thanks Sasha Krot and Martin Bizzarro for inviting his participation to the Workshop on the Chronology of Meteorites and the Early Solar Sys-tem. We thank Sean Raymond for generously providing the data of his planetary accre-tion simulations, and Rainer Wieler for an informal review of an earlier version of this manuscript. We thank Mary Horan and Maud Boyet for their reviews and Rich Walker for his comments and editorial efforts, which greatly improved our manuscript. We thank the Museum National d' Historie Naturelle (Paris), The National Museum of Natural History (Washington DC), The Natural History Museum (London), the Senck-enbergmuseum (Frankfurt), the Institute for Planetology (Münster), the Naturhis-torisches Museum (Wien), and NASA for generously providing the samples for the studies summarized here. F. Nimmo acknowledges funding by NASA-Origins NNG05G039G.

247

Page 248: Touboul Thesis Completed

REFERENCES

Agnor, C. B., Canup, R. M., and Levison, H. F., 1999. On the character and conse-quences of large impacts in the late stage of terrestrial planet formation. Icarus 142, 219-237.

Alexander, C. M. D., 2005. From supernovae to planets: the view from meteorites and interplanetary dust particles. In: Krot, A. N., Scott, E. R. D., and Reipurth, B. Eds.), Chondrites and the Protoplanetary Disk,

Allègre, C. J., Dupré, B., and Lewin, E., 1986. Thorium Uranium Ratio of the Earth. Chem. Geol. 56, 219-227.

Allègre, C. J., Manhès, G., and Göpel, C., 1995. The age of the Earth. Geochim. Cos-mochim. Acta 59, 1445-1456.

Allègre, C. J., Manhès, G., and Göpel, C., 2008. The major differentiation of the Earth at ~ 4.45 Ga. Earth Planet. Sci. Lett., in press.

Amelin, Y., Krot, A. N., Hutcheon, I. D., and Ulyanov, A. A., 2002. Lead isotopic ages of chondrules and calcium-aluminum-rich inclusions. Science 297, 1678-1683.

Amelin, Y., Wadhwa, M., and Lugmair, G. W., 2006. Pb-isotopic dating of meteorites using 202Pb-205Pb double spike: comparison with other high-resolution chronome-ters. Lunar Planet. Sci. Conf. XXXVII, #1970.

Amelin, Y. and Irving, A. J., 2007. Seven million years of evolution on the angrite par-ent body from Pb-isotopic data. Workshop on the chronology of meteorites and the early solar system, LPI Contribution No. 1374, p.20-21.

Amelin, Y., 2008. U-Pb ages of angrites. Geochim. Cosmochim. Acta 72, 221-232. 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. Met. Planet. Sci. 35, 1127-1141.

Bizzarro, M., Baker, J. A., Haack, H., and Lundgaard, K. L., 2005. Rapid timescales for accretion and melting of differentiated planetesimals inferred from 26Al-26Mg chronometry. Astrophys. J. 632, L41-L44.

Blichert-Toft, J., Gleason, J. D., Télouk, P., and Albaréde, F., 1999. The Lu-Hf isotope geochemistry of shergottites and the evolution of hte Martian mantle-crust system. Earth Planet. Sci. Lett. 173, 25-39.

Blichert-Toft, J., Boyet, M., Télouk, P., and Albarède, F., 2002. 147Sm-143Nd and 176Lu-176Hf in eucrites and the differentiation of the HED parent body. Earth Planet. Sci. Lett. 204, 167-181.

Bogard, D. D., Garrison, D. H., and Takeda, H., 2005. Ar-Ar and I-Xe ages and the thermal history of IAB meteorites. Met. Planet. Sci. 40, 207-224.

Borg, L. E., Nyquist, L. E., Wiesmann, H., Shih, C. Y., and Reese, Y., 2003. The age of Dar al Gani 476 and the differentiation history of the martian meteorites inferred from their radiogenic isotopic systematics. Geochim. Cosmochim. Acta 67, 3519-3536.

Bottke, W. F., Nesvorny, D., Grimm, R. E., Morbidelli, A., and O'Brien, D. P., 2006. Iron meteorites as remnants of planetesimals formed in the terrestrial planet re-gion. Nature 439, 821-824.

Bourdon, B., Touboul, M., Caro, G., and Kleine, T., 2008. Early differentiation of the Earth and the Moon. Phil. Trans. Roy. Soc. London, submitted.

248

Page 249: Touboul Thesis Completed

Bouvier, A., Blichert-Toft, J., Moynier, F., Vervoort, J. D., and Albarède, F., 2007. Pb-Pb dating constraints on the accretion and cooling history of chondrites. Geochim. Cosmochim. Acta 71, 1583-1604.

Boyet, M., Blichert-Toft, J., Rosing, M., Storey, M., Telouk, P., and Albarede, F., 2003. Nd-142 evidence for early Earth differentiation. Earth Planet. Sci. Lett. 214, 427-442.

Boyet, M. and Carlson, R. W., 2005. 142Nd evidence for early (>4.53 Ga) global dif-ferentiation of the silicate Earth. Science 309, 576-581.

Burkhardt, C., Kleine, T., Palme, H., Bourdon, B., Zipfel, J., Friedrich, J., and Ebel, D., 2008. Hafnium-tungsten systematics of calcium-aluminum-rich inclusions from carbonaceous chondrites. Geochim. Cosmochim. Acta, submitted.

Canup, R. M. and Asphaug, E., 2001. Origin of the Moon in a giant impact near the end of the Earth's formation. Nature 412, 708-712.

Canup, R. M., 2004. Dynamics of lunar formation. Annual Review of Astronomy and Astrophysics 42, 441-475.

Carlson, R. W. and Lugmair, G. W., 1988. The age of ferroan anorthosite 60025: oldest crust on a young Moon? Earth Planet. Sci. Lett. 90, 119-130.

Caro, G., Bourdon, B., Birck, J. L., and Moorbath, S., 2003. Sm-146-Nd-142 evidence from Isua metamorphosed sediments for early differentiation of the Earth's man-tle. Nature 423, 428-432.

Caro, G., Bourdon, B., Birck, J. L., and Moorbath, S., 2006. High-precision 142Nd/144Nd measurements in terrestrial rocks: Constraints on the early differentiation of the Earth's mantle. Geochim. Cosmochim. Acta 70.

Caro, G., Bourdon, B., Halliday, A., and Quitté, G., 2008. Superchrondritic Sm/Nd in Mars, Earth and the Moon. Nature 452, 336-339

Chambers, J. E. and Wetherill, G. W., 1998. Making the terrestrial planets: N-body in-tegrations of planetary embryos in three dimensions. Icarus 136, 304-327.

Chambers, J. E., 2001. Making more terrestrial planets. Icarus 152, 205-224. Chambers, J. E., 2004. Planetary accretion in the inner solar system. Earth Planet. Sci.

Lett. 223, 241-252. Cherniak, D. J., Lanford, W., and Ryerson, F. J., 1991. Lead diffusion in apatite and

zircon using ion implantation and Rutherford backscattering techniques. Geochim. Cosmochim. Acta 55, 1663-1673.

Cherniak, D. J., 1998. Pb diffusion in clinopyroxene. Chem. Geol. 150, 105-117. Choi, B. G., Ouyang, X. W., and Wasson, J. T., 1995. Classification and origin of IAB

and IIICD iron meteorites. Geochim. Cosmochim. Acta 59, 593-612. Connelly, J., Amelin, Y., Krot, A. N., and Bizzarro, M., 2008a. Chronology of the solar

system's oldest solids. Astrophys. J. 675, L121-L124. Connelly, J., Bizzarro, M., Thrane, K., and Baker, J. A., 2008b. The Pb-Pb age of an-

grite Sah99555 revisited. Geochim. Cosmochim. Acta, doi:10.1016/j.gca.2008.06.007.

Cuzzi, J. N., Ciesla, F. J., Petaev, M. I., Krot, A. N., Scott, E. R. D., and Weiden-schilling, S. J., 2005. Nebular Evolution of Thermally Processed Solids: Reconcil-ing Models and Meteorites. In: Krot, A. N., Scott, E. R. D., and Reipurth, B. Eds.), Chondrites and the Protoplanetary Disk, 732-773.

Debaille, V., Brandon, A. D., Yin, Q. Z., and Jacobsen, B., 2007a. Coupled 142Nd-143Nd evidence for a protracted magma ocean in Mars. Nature 450, 525-528.

249

Page 250: Touboul Thesis Completed

Debaille, V., Yin, Q. Z., Brandon, A. D., and Jacobsen, B., 2007b. Lu-Hf and Sm-Nd isotopic studies of shergottites and nakhlites: implications for Martian mantle sources. Lunar Planet. Sci. Conf. XXXVIII, 1903.pdf.

Debaille, V., Brandon, A. D., Yin, Q. Z., and Jacobsen, B., 2008. Duration of a Magma Ocean and Subsequent Mantle Overturn in Mars: Evidence from Nakhlites. Lunar Planet. Sci. Conf. XXXIX, 1615.pdf.

Dodd, R. T., 1969. Metamorphism of the ordinary chondrites: A review. Geochim. Cosmochim. Acta 33, 161-203.

Dodson, M. H., 1973. Closure temperature in cooling geochronological and petrological systems. Contrib. Mineral. Petrol. 40, 259-274.

Duke, M. B., 1965. Metallic iron in basaltic achondrites. J. Geophys. Res. 70, 1523-1528.

Foley, C. N., Wadhwa, M., Borg, L. E., Janney, P. E., Hines, R., and Grove, T. L., 2005. The early differentiation history of Mars from 182W-142Nd isotope sys-tematics in the SNC meteorites. Geochim. Cosmochim. Acta 69, 4557-4571.

Galer, S. J. G. and Goldstein, S. L., 1996. Influence of accretion on lead in the Earth, Earth Processes: Reading the isotopic code. American Geophysical Union, Wash-ington (D.C.), 75-98.

Ganguly, J. and Tirone, M., 2001. Relationship between cooling rate and cooling age of a mineral: Theory and applications to meteorites. Meteorit. Planet. Sci. 36, 167-175.

Göpel, C., Manhès, G., and Allègre, C. J., 1994. U-Pb systematics of phosphates from equilibrated ordinary chondrites. Earth Planet. Sci. Lett. 121, 153-171.

Gradie, J. C. and Tedesco, E. F., 1982. Compositional structure of the asteroid belt. Sci-ence 216, 1405-1407.

Gray, C. M., Papanastassiou, D. A., and Wasserburg, G. J., 1973. Identification of early condensates from solar nebula. Icarus 20, 213-239.

Grimm, R. E. and McSween, H. Y., 1993. Heliocentric zoning of the asteroid belt by Aluminium-26 heating. Science 259, 653-655.

Halliday, A., Rehkämper, M., Lee, D. C., and Yi, W., 1996. Early evolution of the Earth and Moon: New constraints from Hf-W isotope geochemistry. Earth Planet. Sci. Lett. 142, 75-89.

Halliday, A. N., 2004. Mixing, volatile loss and compositional change during impact-driven accretion of the Earth. Nature 427, 505-509.

Halliday, A. N. and Kleine, T., 2006. Meteorites and the Timing, Mechanisms, and Conditions of Terrestrial Planet Accretion and Early Differentiation. In: Lauretta, D. S. and McSween, H. Y. Eds.), Meteorites and the Early Solar System II. The University of Arizona Press, Tucson, 775-801.

Harper, C. L., Völkening, J., Heumann, K. G., Shih, C. Y., and Wiesmann, H., 1991. 182Hf-182W: New cosmochronometric constraints on terrestrial accretion, core formation, the astrophysical site of the r-process, and the origin of the solar sys-tem. Lunar Planet. Sci. Conf. XXII, 515-516.

Harper, C. L. and Jacobsen, S. B., 1992. Evidence from Coupled Sm-147 Nd-143 and Sm-146 Nd-142 Systematics for Very Early (4.5-Gyr) Differentiation of the Earths Mantle. Nature 360, 728-732.

Harper, C. L., Nyquist, L. E., Bansal, B., Wiesmann, H., and Shih, C. Y., 1995. Rapid Accretion and Early Differentiation of Mars Indicated by 142Nd/144Nd in SNC Me-teorites. Science 267, 213-217.

250

Page 251: Touboul Thesis Completed

Harper, C. L. and Jacobsen, S. B., 1996. Evidence for Hf-182 in the early solar system and constraints on the timescale of terrestrial accretion and core formation. Geo-chim. Cosmochim. Acta 60, 1131-1153.

Hevey, P. J. and Sanders, I. S., 2006. A model for planetesimal meltdown by 26Al and its implications for meteorite parent bodies. Met. Planet. Sci. 41, 95-106.

Hoink, T., Schmalzl, J., and Hansen, U., 2006. Dynamics of metal-silicate separation in a terrestrial magma ocean. Geochem. Geophys. Geosyst. 7.

Huss, G. R., Rubin, A. E., and Grossman, J. N., 2006. Thermal metamorphism in chon-drites. In: Lauretta, D. S. and McSween, H. Y. Eds.), Meteorites and the Early So-lar System II. The University of Arizona Press, Tucson, 567-586.

Ireland, T. R. and Bukovanska, M., 2003. Initial 182Hf/180Hf in meteoritic zircons. Geochim. Cosmochim. Acta 67, 4849-4856.

Jacobsen, S. B. and Harper, C. L., 1996. Accretion and early differentiation history of the Earth based on extinct radionuclides, Earth Processes: Reading the Isotopic Code, Geophysical monograph. American Geophysical Union, 47-74.

Jacobsen, S. B., 2005. The Hf-W isotopic system and the origin of the Earth and Moon. Ann. Rev. Earth Planet. Sci. 33, 531-570.

Kita, N. T., Nagahara, H., Togashi, S., and Morishita, Y., 2000. A short duration of chondrule formation in the solar nebula: Evidence from 26Al in Semarkona ferro-magnesian chondrules. Geochim. Cosmochim. Acta 64, 3913-3922.

Kleine, T., Münker, C., Mezger, K., and Palme, H., 2002. Rapid accretion and early core formation on asteroids and the terrestrial planets from Hf-W chronometry. Nature 418, 952-955.

Kleine, T., Mezger, K., Münker, C., Palme, H., and Bischoff, A., 2004a. 182Hf-182W isotope systematics of chondrites, eucrites, and Martian meteorites: Chronology of core formation and mantle differentiation in Vesta and Mars. Geochim. Cosmo-chim. Acta 68, 2935-2946.

Kleine, T., Mezger, K., Palme, H., Scherer, E., and Münker, C., 2004b. The W isotope evolution of the bulk silicate Earth: constraints on the timing and mechanisms of core formation and accretion. Earth Planet. Sci. Lett. 228, 109-123.

Kleine, T., Mezger, K., Palme, H., Scherer, E., and Münker, C., 2005a. Early core for-mation in asteroids and late accretion of chondrite parent bodies: Evidence from 182Hf-182W in CAIs, metal-rich chondrites and iron meteorites. Geochim. Cosmo-chim. Acta 69, 5805-5818.

Kleine, T., Mezger, K., Palme, H., Scherer, E., and Münker, C., 2005b. The W isotope composition of eucrites metal: Constraints on the timing and cause of the thermal metamorphism of basaltic eucrites. Earth Planet. Sci. Lett. 231, 41-52.

Kleine, T., Palme, H., Mezger, K., and Halliday, A. N., 2005c. Hf-W chronometry of lunar metals and the age and early differentiation of the Moon. Science 310, 1671-1674.

Kleine, T., Touboul, M., Halliday, A., Zipfel, J., and Palme, H., 2007. Cosmochemical fractionation of Hf and W in the solar nebula: Evidence from W isotopes in chon-drites. Lunar Planet. Sci. Conf. XXXVIII.

Kleine, T., Bourdon, B., Burkhardt, C., and Irving, A. J., 2008a. Hf-W chronometry of angrites: constraints on the absolute age of CAIs and planetesimals accretion timescales. Lunar Planet. Sci. Conf. XXXIX, #2367.

Kleine, T., Touboul, M., Van Orman, J. A., Bourdon, B., Maden, C., Mezger, K., and Halliday, A., 2008b. Hf-W thermochronometry: closure temperature and con-

251

Page 252: Touboul Thesis Completed

straints on the accretion and cooling history of the H chondrite parent body. Earth Planet. Sci. Lett. 270, 106-118.

Krot, A. N., Meibom, A., Weisberg, M. K., and Keil, K., 2002. The CR chondrite clan: Implications for early solar system processes. Meteorit. Planet. Sci. 37, 1451-1490.

Kunihiro, T., Rubin, A. E., McKeegan, K. D., and Wasson, J. T., 2004. Initial 26Al/27Al in carbonaceous-chondrite chondrules: Too little 26Al to melt asteroids. Geochim. Cosmochim. Acta 68, 2947-2957.

Kurahashi, E., Kita, N. T., Nagahara, H., and Morishita, Y., 2008. 26Al-26Mg systemat-ics of chondrules in a primitive CO chondrite. Geochim. Cosmochim. Acta 72, 3865-3882.

LaTourrette, T. and Wasserburg, G. J., 1998. Mg diffusion in anorthite: implications for the formation of early solar system planetesimals. Earth Planet. Sci. Lett. 158, 91-108.

Lee, D. C. and Halliday, A. N., 1995. Hafnium-tungsten chronometry and the timing of terrestrial core formation. Nature 378, 771-774.

Lee, D. C. and Halliday, A. N., 1996. Hf-W isotopic evidence for rapid accretion and differentiation in the early solar system. Science 274, 1876-1879.

Lee, D. C. and Halliday, A. N., 1997. Core formation on Mars and differentiated aster-oids. Nature 388, 854-857.

Lee, D. C., Halliday, A. N., Snyder, G. A., and Taylor, L. A., 1997. Age and origin of the moon. Science 278, 1098-1103.

Lee, D. C. and Halliday, A. N., 2000. Accretion of primitive planetesimals: Hf-W iso-topic evidence from enstatite chondrites. Science 288, 1629-1631.

Lee, D. C., Halliday, A. N., Leya, I., Wieler, R., and Wiechert, U., 2002. Cosmogenic tungsten and the origin and earliest differentiation of the Moon. Earth Planet. Sci. Lett. 198, 267-274.

Lee, D. C., 2005. Protracted core formation in asteroids: Evidence from high precision W isotopic data. Earth Planet. Sci. Lett. 237, 21-32.

Leya, I., Wieler, R., and Halliday, A. N., 2000. Cosmic-ray production of tungsten iso-topes in lunar samples and meteorites and its implications for Hf-W cosmochem-istry. Earth Planet. Sci. Lett. 175, 1-12.

Leya, I., Wieler, R., and Halliday, A. N., 2003. The influence of cosmic-ray production on extinct nuclide systems. Geochim. Cosmochim. Acta 67, 529-541.

Liu, Y. Z., Nyquist, L. E., Wiesmann, H., Shih, C. Y., and Takeda, H., 2002. The age of plagioclase-diopside inclusions in the Caddo County meteorite: Rb-Sr and Al-Mg isochrons. Met. Planet. Sci. 37, 5206.pdf.

Lugmair, G. W. and Galer, S. J. G., 1992. Age and isotopic relationships among the angrites Lewis Cliff 86010 and Angra dos Reis. Geochim. Cosmochim. Acta 56, 1673-1694.

Markowski, A., Leya, I., Quitté, G., Ammon, K., Halliday, A. N., and Wieler, R., 2006a. Correlated helium-3 and tungsten isotopes in iron meteorites: Quantitative cosmogenic corrections and planetesimal formation times. Earth Planet. Sci. Lett. 250, 104-115.

Markowski, A., Quitté, G., Halliday, A. N., and Kleine, T., 2006b. Tungsten isotopic compositions of iron meteorites: chronological constraints vs. cosmogenic effects. Earth Planet. Sci. Lett. 242, 1-15.

252

Page 253: Touboul Thesis Completed

Markowski, A., Quitté, G., Kleine, T., Halliday, A., Bizzarro, M., and Irving, A. J., 2007. Hf-W chronometry of angrites and the earliest evolution of planetary bod-ies. Earth Planet. Sci. Lett. 262, 214-229.

Masarik, J., 1997. Contribution of neutron-capture reactions to observed tungsten iso-topic ratios. Earth Planet. Sci. Lett. 152, 181-185.

Mezger, K. and Krogstad, E. J., 1997. Interpretation of discordant U-Pb zircon ages: An evaluation. Journal of Metamorphic Geology 15, 127-140.

Miyamoto, M., Fujii, N., and Takeda, H., 1981. Ordinary chondrite parent body: An internal heating model. Proc. Lunar Planet. Sci. Conf. 12B, 1145-1152.

Nagashima, K., Huss, G. R., and Krot, A. N., 2008. 26Al in chondrules from CR carbo-naceous chondrites. Lunar Planet. Sci. Conf. XXXIX, #2224.

Newsom, H. E., Sims, K. W. W., Noll, P., Jaeger, W., Maehr, S., and Beserra, T., 1996. The depletion of W in the bulk silicate Earth: constraints on core formation. Geo-chim. Cosmochim. Acta 60, 1155-1169.

Nimmo, F. and Agnor, C. B., 2006. Isotopic outcomes of N-body accretion simulations: Constraints on equilibration processes during large impacts from Hf/W observa-tions. Earth Planet. Sci. Lett. 243, 26-43.

Nimmo, F. and Kleine, T., 2007. How rapidly did Mars accrete? Uncertainties in the Hf-W timing of core formation. Icarus 191, 497-504.

Norman, M. D., Borg, L. E., Nyquist, L. E., and Bogard, D. D., 2003. Chronology, geo-chemistry, and petrology of a ferroan noritic anorthosite clast from Descartes breccia 67215: Clues to the age, origin, structure, and impact history of the lunar crust. Meteorit. Planet. Sci. 38, 645-661.

Nyquist, L., Bogard, D., Yamaguchi, A., Shih, C. Y., Karouji, Y., Ebihara, M., Reese, Y., Garrison, D., McKay, G., and Takeda, H., 2006. Feldspathic clasts in Yamato-86032: Remnants of the lunar crust with implications for its formation and impact history. Geochim. Cosmochim. Acta 70, 5990-6015.

Nyquist, L. E., Wiesmann, H., Bansal, B., Shih, C. Y., Keith, J. E., and Harper, C. L., 1995. Sm-146-Nd-142 Formation Interval for the Lunar Mantle. Geochim. Cos-mochim. Acta 59, 2817-2837.

Nyquist, L. E., Bogard, D. D., Shih, C. Y., Greshake, A., Stoffler, D., and Eugster, O., 2001. Ages and geologic histories of Martian meteorites. Space Sci. Rev. 96, 105-164.

Pahlevan, K. and Stevenson, D. J., 2007. Possible Origin of the geochemical similarity of the Earth and Moon. Earth Planet. Sci. Lett. 262, 438-449.

Palme, H. and Rammensee, W., 1981a. Tungsten and some other siderophile elements in meteoritic and terrestrial basalts. Lunar Planet. Sci. Conf. XII, 796-798.

Palme, H. and Rammensee, W., 1981b. The significance of W in planetary differentia-tion processes: Evidence from new data on eucrites. Proc. 12th Lunar Planet. Sci. Conf., 949-964.

Palme, H., Wlotzka, F., Spettel, B., Dreibus, G., and Weber, H., 1988. Camel Donga: a eucrite with high metal content. Meteoritics 23, 49-57.

Pepin, R. O. and Porcelli, D., 2006. Xenon isotope systematics, giant impacts, and man-tle degassing on the early Earth. Earth Planet. Sci. Lett. 250, 470-485.

Podosek, F. A. and Ozima, M., 2000. The xenon age of the Earth. In: Canup, R. M. and Righter, K. Eds.), Origin of the Earth and Moon. Lunar and Planetary Institute, Houston, 63-72.

253

Page 254: Touboul Thesis Completed

Qin, L. P., Dauphas, N., Janney, P. E., and Wadhwa, M., 2007. Analytical develop-ments for high-precision measurements of W isotopes in iron meteorites. Anal. Chem. 79, 3148-3154.

Qin, L. P., Dauphas, N., Wadhwa, M., Masarik, J., and Janney, P. E., 2008. Rapid ac-cretion and differentiation of iron meteorite parent bodies inferred from 182Hf-182W chronometry and thermal modeling. Earth Planet. Sci. Lett., doi:10.1016/j.epsl.2008.06.018.

Quitté, G., Birck, J. L., and Allegre, C. J., 2000. 182Hf-182W systematics in eucrites: The puzzle of iron segregation in the early solar system. Earth Planet. Sci. Lett. 184, 83-94.

Quitté, G. and Birck, J. L., 2004. Tungsten isotopes in eucrites revisited and the initial 182Hf/180Hf of the solar system based on iron meteorite data. Earth Planet. Sci. Lett. 219, 201-207.

Rankenburg, K., Brandon, A. D., and Neal, C. R., 2006. Neodymium isotope evidence for a chondritic composition of the Moon. Science 312, 1369-1372.

Raymond, S. N., Quinn, T., and Lunine, J. I., 2006. High-resolution simulations of the final assembly of Earth-like planets I. Terrestrial accretion and dynamics. Icarus 183, 265-282.

Righter, K. and Drake, M. J., 1997. Metal-silicate equilibrium in a homogeneously ac-creting earth: New results for Re. Earth Planet. Sci. Lett. 146, 541-553.

Righter, K., 2003. Metal-silicate partitioning of siderophile elements and core formation in the early Earth. Ann. Rev. Earth Planet. Sci. 31, 135-174.

Righter, K. and Shearer, C. K., 2003. Magmatic fractionation of Hf and W: Constraints on the timing of core formation and differentiation in the Moon and Mars. Geo-chim. Cosmochim. Acta 67, 2497-2507.

Rocholl, A. and Jochum, K. P., 1993. Th, U and other trace elements in carbonaceous chondrites: Implications for the terrestrial and solar-system Th/U ratios. Earth Planet. Sci. Lett. 117, 265-278.

Rubie, D. C., Melosh, H. J., Reid, J. E., Liebske, C., and Righter, K., 2003. Mechanisms of metal-silicate equilibration in the terrestrial magma ocean. Earth Planet. Sci. Lett. 205, 239-255.

Rudraswami, N. G. and Goswami, J. N., 2007. 26Al in chondrules from unequilibrated L chondrites: Onset and duration of chondrule formation in the early solar system. Earth Planet. Sci. Lett. 257, 231-244.

Rudraswami, N. G., Goswami, J. N., Chattopadhyay, B., Sengupta, S. K., and Thapli-yal, A. P., 2008. 26Al records in chondrules from unequilibrated ordinary chon-drites: II. Duration of chondrule formation and parent body thermal metamor-phism. Earth Planet. Sci. Lett., doi:10.1016/j.epsl.2008.07.004.

Sahijpal, S., Soni, P., and Gupta, G., 2007. Numerical simulations of the differentiation of accreting planetesimals with Al-26 and Fe-60 as the heat sources. Meteorit. Planet. Sci. 42, 1529-1548.

Sasaki, T. and Abe, Y., 2007. Rayleigh-Taylor instability after giant impacts: Imperfect equilibration of the Hf-W system and its effects on the core formation age. Earths Planets Space 59, 1035-1045.

Schérsten, A., Elliott, T., Hawkesworth, C., Russell, S. S., and Masarik, J., 2006. Hf-W evidence for rapid differentiation of iron meteorite parent bodies. Earth Planet. Sci. Lett. 241, 530-542.

254

Page 255: Touboul Thesis Completed

Schoenberg, R., Kamber, B. S., Collerson, K. D., and Eugster, O., 2002a. New W-isotope evidence for rapid terrestrial accretion and very early core formation. Geochim. Cosmochim. Acta 66, 3151-3160.

Schoenberg, R., Kamber, B. S., Collerson, K. D., and Moorbath, S., 2002b. Tungsten isotope evidence from similar to 3.8-Gyr metamorphosed sediments for early me-teorite bombardment of the Earth. Nature 418, 403-405.

Schulz, T., Münker, C., Mezger, K., and Palme, H., 2006. Age and origin of IAB iron meteorites and their silicate inclusions inferred from Hf/W chronometry. Lunar Planet. Sci. Conf. XXXVII, abstract no. 1401.

Scott, E. R. D. and Wasson, J. T., 1975. Classification and properties of iron meteorites. Rev. Geophys. 13, 527-546.

Shearer, C. K. and Papike, J. J., 1999. Magmatic evolution of the Moon. Am. Mineral. 84, 1469-1494.

Shearer, C. K. and Newsom, H. E., 2000. W-Hf isotope abundances and the early origin and evolution of the Earth-Moon system. Geochim. Cosmochim. Acta 64, 3599-3613.

Slater-Reynolds, V. and McSween, H. Y., 2005. Peak metamorphic temperatures in type 6 ordinary chondrites: An evaluation of pyroxene and plagioclase geother-mometry. Meteorit. Planet. Sci. 40, 745-754.

Solomatov, V. S., 2000. Fluid dynamics of a terrestrial magma ocean. In: Canup, R. M. and Righter, K. Eds.), Origin of the Earth and Moon. University of Arizona Press, Tucson, 323-338.

Srinivasan, G., Whitehouse, M. J., Weber, I., and Yamaguchi, A., 2007. The crystalliza-tion age of eucrite zircon. Science 317, 345-347.

Stevenson, D. J., 1990. Fluid dynamics of core formation. In: Newsom, H. E. and Jones, J. E. Eds.), Origin of the Earth. Oxford University Press, New York, 231-249.

Takeda, H. and Graham, A. L., 1991. Degree of equilibration of eucritic pyroxenes and thermal metamorphism of the earliest planetary crust. Meteoritics 26, 129-134.

Takeda, H., Bogard, D., Mittlefehldt, D., and Garrison, D., 2000. Mineralogy, petrol-ogy, chemistry, and 39Ar-40Ar and exposure ages of the Caddo County IAB iron: evidence for early partial melt segregation of a gabbro area rich in plagioclase-diopside. Geochim. Cosmochim. Acta 64, 1311-1327.

Touboul, M., Kleine, T., Bourdon, B., Palme, H., and Wieler, R., 2007. Late formation and prolonged differentiation of the Moon inferred from W isotopes in lunar met-als. Nature 450, 1206-1209.

Touboul, M., Kleine, T., and Bourdon, B., 2008a. Hf-W systematics of cumulate eucrites and the chronology of the eucrite parent body. Lunar Planet. Sci. Conf. XXXIX.

Touboul, M., Kleine, T., Bourdon, B., Palme, H., and Wieler, R., 2008b. 182Hf-182W systematics of ferroan anorthosites and the lifetime of the lunar magma ocean. Icarus, submitted.

Trieloff, M., Jessberger, E. K., Herrwerth, I., Hopp, J., Fiéni, C., Ghelis, M., Bourot-Denise, M., and Pellas, P., 2003. Structure and thermal history of the H-chondrite parent asteroid revealed by thermochronometry. Nature 422, 502-506.

Van Orman, J. A., Grove, T. L., and Shimizu, N., 2001. Rare earth element diffusion in diopside: Influence of temperature, pressure and ionic radius, and an elastic model for diffusion in silicates. Contrib. Mineral. Petrol. 141, 687-703.

255

Page 256: Touboul Thesis Completed

Van Orman, J. A., Saal, A. E., Bourdon, B., and Hauri, E. H., 2006. Diffusive fractiona-tion of U-series radionuclides during mantle melting and shallow level melt-cumulate interaction. Geochim. Cosmochim. Acta 70, 4797-4812.

Vockenhuber, C., Oberli, F., Bichler, M., Ahmad, I., Quitté, G., Meier, M., Halliday, A. N., Lee, D. C., Kutschera, W., Steier, P., Gehrke, R. J., and Helmer, R. G., 2004. New half-life measurement of 182Hf: Improved chronometer for the early solar system. Phys. Rev. Lett. 93, art. no. 172501.

Wade, J. and Wood, B. J., 2005. Core formation and the oxidation state of the Earth. Earth Planet. Sci. Lett. 236, 78-95.

Walter, M. J. and Thibault, Y., 1995. Partitioning of tungsten and molybdenum between metallic liquid and silicate melt. Science 270, 1186-1189.

Walter, M. J., Newsom, H. E., Ertel, W., and Holzheid, A., 2000. Siderophile elements in the Earth and Moon: metal/silicate partitioning and implications for core forma-tion. In: Canup, R. M. and Righter, K. Eds.), Origin of the Earth and Moon. Lunar and Planetary Institute, Houston, 265-289.

Walter, M. J. and Tronnes, R. G., 2004. Early Earth differentiation. Earth Planet. Sci. Lett. 225, 253-269.

Wänke, H. and Dreibus, G., 1988. Chemical-Composition and Accretion History of Terrestrial Planets. Phil. Trans. Roy. Soc. London 325, 545-557.

Wasserburg, G. J., Papanastassiou, D. A., and Sanz, H. G., 1969. Initial strontium for a chondrite and the determination of a metamorphism or formation interval. Earth Planet. Sci. Lett. 7, 33-43.

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. Geochim. Cosmochim. Acta 66, 2445-2473.

Weidenschilling, S. J., 1977. Distribution of mass in the planetary system and solar nebula. Astrophys. Space Sci. 51, 153-158.

Weidenschilling, S. J., Spaute, D., Davis, D. R., Marzari, F., and Ohtsuki, K., 1997. Accretional evolution of a planetesimal swarm .2. The terrestrial zone. Icarus 128, 429-455.

Yin, Q. Z., Jacobsen, S. B., Yamashita, K., Blichert-Toft, J., Télouk, P., and Albarède, F., 2002. A short timescale for terrestrial planet formation from Hf-W chronome-try of meteorites. Nature 418, 949-952.

Yokochi, R. and Marty, B., 2005. Geochemical constraints on mantle dynamics in the Hadean. Earth Planet. Sci. Lett. 238, 17-30.

256

Page 257: Touboul Thesis Completed

FIGURE CAPTIONS

Fig. 1: W isotope data for chondrites. Data for carbonaceous chondrites are from Kleine et al. (2002, 2004a), Yin et al. (2002) and Schönberg et al. (2002), for ordinary chon-drites from Yin et al. (2002) and Kleine et al. (2007), and for enstatite chondrites from Lee and Halliday (2000). Note that, in contrast to earlier reports (Lee and Halliday, 1995; Lee and Halliday, 1996), all chondrites exhibit a 182W deficit relative to the terres-trial standard.

Fig. 2: Initial 182Hf/180Hf ratios vs. Pb-Pb for angrites. Slightly modified from Burkhardt et al. (2008). References are given in Table 2. Solid lines are decay lines calculated us-ing the 182Hf half-life of 0.078 Ma-1. The dashed line is the best fit line of the data for D’Orbigny, Sahara 99555, NWA 4590 and 4801 and corresponds to λ182Hf = 0.075 ± 0.006 Ma-1. Note that the Hf-W system in NWA 2999 is disturbed and the initial 182Hf/180Hf plotted here is higher than originally reported in Markowski et al. (2007). This higher value is obtained from excluding the Hf-W data for the whole-rock, fines and leached fractions.

Fig. 3: Closure temperature of the Hf-W, U-Pb and Al-Mg systems in various minerals as a function of grain diameter and cooling rate.

Fig. 4: ε182W vs. exposure ages for magmatic iron meteorites. W isotope data are from Kleine et al. (2005a), Lee (2005), Markowski et al. (2006b), Schérsten et al. (2006), and Qin et al. (2008). For exposure ages see these references. The three lines indicate the calculated 182W burnout using different correction procedures (Leya et al., 2003; Qin et al., 2008), assuming that the pre-exposure ε182W of the magmatic irons is identical to the initial ε182W of CAIs (Burkhardt et al., 2008). KASKAD and JEF are two different libraries for cross sections of neutron induced reactions that were used by Qin et al. (2008) and for which reasonable agreement between the predicted effects and those measured in Grant and Carbo (Markowski et al., 2006a) were observed. Fig. 5: ε182W values for IAB-IIICD iron meteorites in comparison to values the two magmatic irons Gibeon and Negrillos and the initial ε182W of CAIs (shown as the hatched area). Data are from Kleine et al. (2005a), Markowski et al. (2006b), and Schérsten et al. (2006).

Fig. 6: Hf-W isochron diagram for basaltic eucrites. The value for average basaltic eucrites is based on the Hf-W data for whole-rock samples (Kleine et al., 2004a; Kleine et al., 2005b). The data for metals and Camel Donga separates are from Kleine et al. (2005b). CC = average carbonaceous chondrites (Kleine et al., 2004a). The model age for Asuka 881467 zircons is obtained from the regression of the zircon data (Srinivasan et al., 2007) and average basaltic eucrites.

257

Page 258: Touboul Thesis Completed

Fig. 7: (a) Hf-W isochrons for H chondrites. Data are from Kleine et al. (2008b). (b) Cooling curves for the high-T evolution (>450 °C) of H chondrites (from Kleine et al. 2008b). Solid lines indicate calculated temperature profiles for different depth in a spherical body with a 100 km radius. Numbers indicate depth in km from the centre. Calculations assume an initial 26Al/27Al of 5.85×10-5 and instantaneous accretion at 2.7 Ma after CAIs.

Fig. 8: Peak temperature vs. accretion age for asteroids (R~10-100 km) that were heated internally by decay from 26Al. Hatched areas indicate the accretion age of and peak temperatures reached in the center of asteroids of different size. For iron meteorite par-ent bodies, the accretion age is derived from the W model ages for iron meteorites, which must postdate accretion. For chondrites, the accretion age is derived from Al-Mg ages for chondrules. Note that parent body accretion must postdate the formation of the youngest chondrule of a given group. Chondrule ages are from the following references: ordinary chondrites (OC) (Kita et al., 2000; Rudraswami and Goswami, 2007; Rudras-wami et al., 2008); CO chondrites (Kunihiro et al., 2004; Kurahashi et al., 2008); CR chondrites (Nagashima et al., 2008). Peak temperature estimates are from the following references: OC (Slater-Reynolds and McSween, 2005); CO (Huss et al., 2006); CR (Krot et al., 2002). To facilitate core formation, temperatures inside the iron meteorite parent bodies must have been above the Fe-FeS eutectic but probably were higher be-cause some melting of silicates may also be required for efficient metal-silicate separa-tion. Solid lines indicate calculated temperature profiles for the centre of spherical bod-ies with radii of 10 and 100 km.

Fig. 9: ε142Nd vs. ε182W for Martian meteorites. Nd isotope data are from Caro et al. (2008) and Debaille et al. (2007a), W isotope data from Lee and Halliday (1997), Kleine et al. (2004a), and Foley et al. (2005). S = Shergotty, LA = Los Angeles, Z = Zagami.

Fig. 10: Hf-W chronology of Mars. (a) Two-stage model age for core formation as a function of Hf/W ratio in the bulk Martian mantle. The grey hatched area indicates the best estimate for the Hf/W ratio of bulk silicate Mars. The solid line represents ages calculated using ε182W=0.4 for the bulk Martian mantle. The dashed lines indicate ages calculated using ε182W values that are 0.2 higher and lower, respectively. (b) Model for the W isotope evolution of Martian mantle reservoirs. In this model it is assumed that bulk silicate Mars has 180Hf/184W~4 (i.e., Hf/W~3.5) and ε182W~0.4. If Martian mantle reservoirs were established by a global event at ~40 Ma, the NC source must have evolved with very high 180Hf/184W, which requires the presence of garnet. In contrast, Hf-W fractionations in the shergottite sources were more modest, consistent with the limited range in ε182W of shergottites. DS = depleted shergottites, ES = enriched sher-gottites.

Fig. 11: ε182W vs. Ta/W for lunar samples. (a) Data for whole-rocks from KREEP-rich samples as well as low- and high-Ti mare basalts (Kleine et al., 2005c; Touboul et al., 2008b). Also shown are data for mineral separates from some high-Ti mare basalts (Lee et al., 2002). Solid lines indicate the variations in ε182W that can be produced by interac-tion with thermal neutrons as a function of exposure age and Ta/W. These lines were

258

Page 259: Touboul Thesis Completed

calculated using equations given in Leya et al. (2000). High-Ti mare basalt 79155 has ε182W~38 and Ta/W~20 (Kleine et al., 2005c) and is not shown here. (b) ε182W vs. Ta/W for low-Ti mare basalt 15555. The predicted cosmogenic 182W is calculated as a function of Ta/W ratio and using the Sm isotope composition of sample 15555 (Nyquist et al., 1995; Rankenburg et al., 2006). Whereas the data reported by Kleine et al. (2005) and Touboul et al. (2008) are consistent with the calculated cosmogenic 182W, the data reported by Lee et al. (2002) appear to be systematically shifted to higher values. After correction for cosmogenic 182W, this sample has ε182W indistinguishable from 0.

Fig. 12: W isotope composition of lunar metals and ferroan anorthosites having young exposure ages (Kleine et al., 2005c; Touboul et al., 2007; Touboul et al., 2008b).

Fig. 13: Hf-W isochron for the lunar magma ocean (Touboul et al., 2007; Touboul et al., 2008b). The estimated 180Hf/184W ratios of the sources are from Kleine et al. (2005c) and Righter and Shearer (2003). For ferroan anorthosites the measured 180Hf/184W are plotted because these were obtained on pure anorthite separates.

Fig. 14: Schematic illustration of different processes of core formation and their effect on the W isotope evolution of the bulk silicate Earth.

Fig. 15: Reduction in W isotope effects during collision between two bodies as a func-tion of impactor to target mass ratio. The reduction is calculated using equations (13) and (16) for fractional and batch equilibration.

Fig. 16: (a) Growth curves for Earth assuming an exponentially decreasing accretion rate and accretion mean times of 8, 11, and 20 Ma. (b) W isotope evolution for the bulk silicate Earth for the growth curves for Earth from (a) and assuming complete metal-silicate re-equilibration. Also shown is the W isotope evolution for an accretion mean time of 11 Ma but for incomplete metal-silicate re-equilibration (i.e., k=0.75). The pre-sent-day ΔεW value of Earth's mantle can be reproduced for mean time of accretion of 11 Ma.

Fig. 17: (a) Effect of incomplete metal-silicate re-equilibration on calculated core for-mation ages. (b) Relationship between the degree of re-equilibration k and Hf/W ratios in the mantles of accreting objects.

Fig. 18: Growth curves and W isotope evolution for a model Earth and Moon-forming impactor from accretion simulations of Raymond et al. (2006). (a) Growth curves are from run 2b of the N-body accretion simulations of Raymond et al. (2006), which track the evolution of 1054 planetesimals as they orbit and collide. The Moon-forming impact occurs at ~125 Ma and the impactor itself underwent a large impact at ~40 Ma. (b) W isotope evolution of these bodies and considering the two extreme cases of core merg-ing and full metal-silicate re-equilibration (batch equilibration). Silicate mass fraction of bodies is variable; W partition coefficient is fixed at DW=29.4 (i.e., Hf/WBSE~17). Other parameter values and methods are as in Nimmo and Agnor (2006). Note that the iso-

259

Page 260: Touboul Thesis Completed

topic evolution of all planetesimals is tracked in these calculations. (c) As for (b) but both fractional and batch equilibration were considered. Fig. 19: (a) As for Fig. 18a except that run 2a of Raymond et al. (2006) is used. The Moon-forming impact occurs at ~75 Ma and the impactor was largely ac-creted in the first few Ma. (b) As for Fig. 18b. (c) As for Fig. 18c.

Fig. 20: Difference in ε182W between bulk silicate Moon (BSM) and Earth (BSE) as a function of time (slightly modified from Touboul et al. 2007). The dotted area indicates the W isotope composition of the lunar mantle. Assuming that BSM and BSE had iden-tical initial 182W/184W ratios, the time of their formation can be calculated from their present-day 182W/184W and Hf/W ratios. The current best estimates for the Hf/W ratios in BSM and BSE indicate a difference of ~50 %. If this is correct, the Moon must have formed more than ~50 Ma after CAI formation.

Fig. 21: a) Mass and semi-major axis of planetary bodies after 150 Ma, from the N-body simulations of Raymond et al. (2006). Solid symbols plot planet in this solar sys-tem, open symbols are model results. b) Present-day 182W anomaly (relative to chon-drites) for all bodies assuming core merging throughout, and using the methods given in Nimmo and Agnor (2006). c) As for b) but assuming complete metal-silicate (batch) equilibration throughout. Note that in case of core merging the present-day characteris-tics of Mars and the Earth cannot be reproduced. By contrast, assuming complete metal-silicate equilibration results in the observed physical and isotopic properties of the Earth and Mars.

Fig. 22: Timing of events in the first ~100 Ma of the solar system as deter-mined mostly by Hf-W chronometry.

260

Page 261: Touboul Thesis Completed

261

Table 1. Hf/W ratios in planetary reservoirs and minerals Planetary reservoirs Hf/W

Chondrites ~0.6-1.8 Metal core ~0 Basaltic eucrites 27±8 Bulk silicate Mars 3-4 Bulk silicate Moon 26±2 Bulk silicate Earth 17±5

Minerals High-Ca pyroxene 8-18 (DHf/DW ~ 5) Fassaites 6-14 Olivine ~0.5 (DHf/DW ~ 1) Plagioclase ~0.6 Melilites ~0.8 Garnet DHf/DW ~ 30

Page 262: Touboul Thesis Completed

Table 2. Summary of initial 182Hf/180Hf and calculated Hf-W ages as obtained from internal isochronsa

Sample (182Hf/180Hf)i × 105 ΔtCAI (Ma) t (Ma) 207Pb-206Pb age (Ma) Reference CAIs 9.72±0.44 =0 4568.5±0.5 4567.11±0.16 Burkhardt et al. (2008) Ste. Marguerite (H4) 8.50±0.23 1.7±0.7 4566.6±0.5 - Kleine et al. (2008b) Richardton (H5) 6.23±0.30 5.7±0.9 4562.6±0.8 - Kleine et al. (2008b) ALH841069 (H5) 6.12±0.62 5.9±1.4 4562.4±1.4 - Kleine et al. (2008b) Kernouvé (H6) 4.66±0.34 9.4±1.1 4558.9±1.0 - Kleine et al. (2008b) Estacado (H6) 4.43±0.53 10.1±1.7 4558.2±1.6 - Kleine et al. (2008b) Vaca Muerta clast 7.69±0.82 3.0±1.5 4565.3±1.4 - Schönbächler et al. (2001)D'Orbigny 7.18±0.22 3.9±0.7 =4564.42±0.12 4564.42±0.12 Markowski et al. (2007) Sahara 99555 6.95±0.22 4.3±0.7 4564.0±0.6 4564.58±0.14 Markowski et al. (2007) NWA 2999 (px-metal) 5.38±0.48 7.6±1.3 4560.7±1.2 4561.79±0.42 Markowski et al. (2007) NWA 4590 4.65±0.17 9.5±0.8 4558.9±0.6 4558.86±0.30 Kleine et al. (2008a) NWA 4801 4.34±0.23 10.3±0.9 4558.0±0.8 4558.06±0.15 Kleine et al. (2008a) Camel Donga 1.7±0.7 22±5 4546±5 - Kleine et al. (2005b)

aall initial 182Hf/180Hf ratios were re-calculated from Hf-W data given in the original references using the model 1 fit of IsoPlot and 180HF/184W = 1.18×Hf/W. The Pb-Pb ages are from the following references: CAIs (Amelin et al., 2002; Amelin et al., 2006); D’Orbigny (Amelin, 2008); NWA 2999, 4590, 4801 (Amelin and Irving, 2007); Sahara 99555 (Connelly et al., 2008b).

262

Page 263: Touboul Thesis Completed

Table 3. Reference parameters for Hf-W chronology Parameter Value Remarks

( )0

182 180Hf HfT

(9.72±0.44)×10-5 Internal isochron for CAIs

( )182 1804564.42±0.12Ma

Hf Hf (7.18±0.22)×10-5 Internal isochron for D’Orbigny

( )0

182 184W WT

0.864579±0.000014 ( ε182W = -3.28±0.12) Initial 182W/184W of CAI isochron

( )BSE182 1840

W W 0.864863±0.000018 ( ε182W ª 0) 182W/184W of terrestrial standard as measured at ETH Zurich

( )CHUR182 1840

W W 0.864699±0.000012 ( ε182W = -1.9±0.1) Average 182W/184W of carbonaceous chondrites

( )180 184CHUR

Hf W = 1.23±0.08

( )4 180 182182 =10 Hf WWQ = 1.42 ×104

263

Page 264: Touboul Thesis Completed

Table 4. Summary of estimates for the initial 182Hf/180Hf of the solar system Initial 182Hf/180Hf Reference Method ~2×10-5 Harper and Jacobsen 1996 Models for 182Hf nucleosynthesis >2.6×10-4 Lee and Halliday 1995 Hf-W data for chondrites, iron meteorites (2.75±0.24)×10-4 Lee and Halliday 2000 Isochrons for Ste. Marguerite, Forest Vale (1.00±0.08)×10-4 Yin et al. 2002 Isochron for Dalgety Downs (L4), Dhurmsala (LL6) (1.09±0.09)×10-4 Kleine et al. 2002 Isochron for Ste. Marguerite (H4) (7.4±4.5)×10-4 Ireland and Bukovanska 2003 Hf-W data for zircons from Simmern (H5) (1.60±0.25)×10-4 Quitté and Birck 2004 Iron meteorites (1.07±0.10)×10-4 Kleine et al. 2005a Isochron for CAIs (9.72±0.44)×10-5 Burkhardt et al. 2008 Mineral isochrons for CAIs

264

Page 265: Touboul Thesis Completed

Table 5. Summary of ε182W values and two-stage model ages for magmatic iron meteoritesa

Sample ε182W ΔtCAI (Ma) Gibeon (IVA) -3.38±0.05 -2.5 to 0.6 Negrillos (IIAB) -3.39±0.08 -2.8 to 0.8 IC -3.63 to -3.35 -4.4 to 0.4 IIAB -3.43 to -3.23 -2.7 to 1.4 IID -3.95 to -3.24 -6.7 to 1.4 IIIAB -3.40 to -3.28 -2.4 to 1.0 IIIE -3.41 to -3.33 -2.5 to 0.6 IIIF -3.35 to -3.00 -1.9 to 3.7 IVA -3.43 to -3.28 -2.7 to 1.0 IVB -3.50 to -3.46 -3.3 to -0.5 aRange of ε182W values for different groups of magmatic iron meteorites are from Qin et al. (2008) and are corrected for cosmic-ray effects using noble gas systematics. Note for IVB irons the calculated W model ages are negative, sug-gesting that this correction procedure did not en-tirely correct the cosmic-ray effects. W model ages, ΔtCAI, are calculated relative to an initial ε182W=-3.28±0.12 for CAIs and include uncer-tainties on the ε182W values of the iron meteor-ites, CAIs and average carbonaceous chondrites (ε182W=-1.9±0.1).

265

Page 266: Touboul Thesis Completed

Table 6. Summary of proposed Hf-W ages for the formation of Earth's corea

Age Based on Accretion/core formation model Metal-silicate re-equilibration Reference t100% >62 Ma Δε182WBSE = 0 Two-stage n.a. Lee and Halliday (1995) t100% ~30 Ma Δε182WBSE = +1.9 Two-stage n.a. Kleine et al. (2002); Schön-

berg et al. (2002); Yin et al. (2002)

t63% = 11 Ma Δε182WBSE = +1.9 Continuous core formation during exponentially decreasing accretion

100 % Yin et al. (2002)

t99% ~50-100 Ma Δε182WBSE = +1.9 Continuous core formation during exponentially decreasing accretion

60-100 % Kleine et al. (2004b)

tGI ~55 Ma Δε182WBSE = +1.9 Hf/WBSE = 15 Hf/WBSI = 15

Multiple giant impacts that occur at an overall exponentially decreasing rate

100 % before the Moon-forming impact; 26 % during the Moon-forming impact; batch equ.

Halliday (2004)

tGI ~50 Ma Δε182WBSE = +1.9 Hf/WBSE = 15 Hf/WBSI = Hf/WBSE = 5 before the giant impact

Multiple giant impacts that occur at an overall exponentially decreasing rate

100 % before the Moon-forming impact; 4 % during the Moon-forming impact; batch equ.

Halliday (2004)

t63% ~11 Ma tGI ~32 Ma

Δε182WBSE = +1.9

Exponentially decreasing accretion, which terminates by the Moon-forming impact

100 %; fractional equ. Jacobsen (2005)

tGI = Ma 901062+

− Δε182WBSE = Δε182WBSM Hf/WBSE < Hf/WBSM

Independent on the accretion/core formation model

n.a. Touboul et al. (2007)

tGI ~125 Ma Δε182WBSE = +1.9 Multiple stochastic giant impacts ~100 % for the final Moon-forming impact

This study

tGI ~75 Ma Δε182WBSE = +1.9 Multiple stochastic giant impacts ~100 % for the final Moon-forming impact

This study

aBSE = bulk silicate Earth; BSM = bulk silicate Moon; BSI = bulk silicate impactor; GI = giant impact (i.e., the Moon-forming impact)

266

Page 267: Touboul Thesis Completed

Fig. 1

267

Page 268: Touboul Thesis Completed

Fig. 2

268

Page 269: Touboul Thesis Completed

Fig. 3

269

Page 270: Touboul Thesis Completed

Fig. 4

270

Page 271: Touboul Thesis Completed

Fig. 5

271

Page 272: Touboul Thesis Completed

Fig. 6

272

Page 273: Touboul Thesis Completed

Fig. 7

273

Page 274: Touboul Thesis Completed

Fig. 8

274

Page 275: Touboul Thesis Completed

Fig. 9

275

Page 276: Touboul Thesis Completed

Fig. 10

276

Page 277: Touboul Thesis Completed

Fig. 11

277

Page 278: Touboul Thesis Completed

Fig. 12

278

Page 279: Touboul Thesis Completed

Fig. 13

279

Page 280: Touboul Thesis Completed

Fig. 14

280

Page 281: Touboul Thesis Completed

Fig. 15

281

Page 282: Touboul Thesis Completed

Fig. 16

282

Page 283: Touboul Thesis Completed

Fig. 17

283

Page 284: Touboul Thesis Completed

Fig. 18

284

Page 285: Touboul Thesis Completed

Fig. 19

285

Page 286: Touboul Thesis Completed

Fig. 20

286

Page 287: Touboul Thesis Completed

Fig. 21

287

Page 288: Touboul Thesis Completed

Fig. 22

288

Page 289: Touboul Thesis Completed

Remerciements - Acknowledgments

Pour conclure je voudrais remercier tous ceux qui ont contribué à la réalisation de cette thèse et qui m’ont soutenu au cours de ces années passées à l’IPGP et à l’ETH.

Je suis tout particulièrement reconnaissant envers Bernard Bourdon et Thorsten Kleine, mes deux superviseurs. Leur encadrement et nos nombreuses et fructueuses discussions m’ont permis d’apprendre beaucoup à leur contact et de m’épanouir dans mon travail. Bernard m’a tout d’abord donné l’opportunité de découvrir la recherche scientifique lors de mon stage de master sur les séries de l’uranium dans les laves de la Guadeloupe. Il m’a ensuite renouvelé sa confiance en me prenant en thèse à l’Institut de Physique du Globe de Paris. Je lui suis sincèrement reconnaissant de m’avoir proposé et conseillé de le suivre à Zürich lorsqu’il a été nommé pour prendre la succession d’Alex Halliday à l’ETH. Cela m’a permis de réaliser mon rêve, travailler sur la formation du système solaire, et de m’ouvrir au monde au contact de l’environnement international de l’ensemble du département des Sciences de la Terre. Ce fut une grande chance pour moi de faire la connaissance de Thorsten Kleine et de pouvoir travailler à Zürich sous sa direction. Il m’a tout appris sur la chronométrie Hf-W et a toujours été disponible. Merci pour les nombreuses heures passées ensemble en chimie ou devant le spectromètre de masse et pour son aide et ses corrections pour les différents articles. Je tenais également à le remercier pour m’avoir soutenu lors de mon séjour à l’hôpital.

Je voudrais remercier Alex Halliday, Francis Albarède et Jean-Louis Birck d’avoir accepté d’être examinateur de cette thèse et de s’être déplacé à Zürich pour sa soutenance.

Je tenais à exprimer toute ma gratitude envers Rainer Wieler et Herbert Palme, pour leur collaboration lors de l’étude des échantillons lunaires, pour les nombreuses discussions scientifiques et pour leur aide dans l’interprétation des données.

Je tenais à remercier James Van Orman et Colin Maden pour leur aide sur les modèles de diffusion du W et de refroidissement des planétésimaux, indispensables à l’interprétation des données.

Je remercie Tony Irving, Juta Zipfel et Ted Bunch pour nous avoir fourni certains des échantillons météoritiques ainsi que leur caractérisation minéralogique.

Je suis sincèrement reconnaissant envers Amélie Hubert et Monique Pierre pour m’avoir initié aux manipulations en salle blanche et aux mesures isotopiques sur Neptune avec l’aide de Pascal Louvat. Un grand merci également à Guillaume Caro pour m’avoir instruit la méthode de séparation du Nd et la mesure de sa composition isotopique sur Triton.

Je remercie Agnès Markowski et Ghylaine Quitté pour leur soutien et pour leur aide à mieux cerner la chronométrie Hf-W lors de mon arrivée à Zürich. Merci à Irene Ivanov pour m’avoir appris la séparation de minéraux par liqueurs denses. Merci également à Andreas Stracke, Ben Reynolds et Felix Oberli pour leurs conseils et pour avoir assurer la maintenance des machines et des salles blanches, permettant de travailler dans des conditions optimales.

289

Page 290: Touboul Thesis Completed

Je tenais à remercier l’ensemble des chercheurs, des étudiants et des techniciens de l’IPGP pour leurs conseils et leurs aides au cours de mon année passée au sein de L’IPGP, et plus particulièrement Laure Meynadier, Jean Louis Birck, Gérard Manhès, Christa Göpel, Manuel Moreira, Gérôme Gaillardet, Claude Allègre, Benjamin Chetelat, Damien Calmels, Anne-Lise Salomé, Joel Dyon, Lydia Zerbib, Sylvie Cangemi.

Merci également à Audrey Bouvier, Vinciane Debaille, Anne Trinquier, Morten

Anderson, Oliver Nebel, Craig O’Neil, Ben Jacobsen, Anders Meibom, Marcus Gutjahr, Ansgar Grimberg, Bastian Georg, Sune Nielsen, Simone de Leuw, Matthias Van Ginneken et Dim Coumou pour les bons moments passés lors des différentes conférences et pour les échanges d’idées scientifiques.

Durant ces quatre années, j’ai tout particulièrement apprécié l’environnement de

L’ETH et les nombreuses relations que j’y ai nouées avec mes collègues. Ils ont énormément contribué à la réussite de ce projet en me soutenant et en me permettant de passer des moments agréables à l’institut et à l’extérieur. Pour cela, je voudrais remercier :

- mes collègues de bureau, Ulrik Hans, Gregory de Souza et Janne Koornneef (une vrai dream team), pour l’ambiance dans le bureau et au labo ainsi que pour les nombreuses soirées passées ensemble.

- l’équipe de la ‘Friday beer’, Jessica Langlade, Marion Louvel, Pierre Bouilhol et Maarten Aerts, pour l’entretien de cette tradition qui permet de renforcer les liens entre les membres du département.

- l’ensemble des personnes non encore citées avec qui j’ai partagé multes rires, bières et autres: Caroline Fitoussi, Ed Tipper, Sarah Aciego, Manuel Petitat, Antoine Roth, Benjamin Le Bayon, Christoph Burkhardt, Olivier Reubi, Jörg Rickli, Ruth Hindshaw , Paula Ardia, Lucas Carrichi, Estelle Auzanneau, Carmen Sanchez, Etienne Médard, Sarah Bureau, Matthias Meier, Veronika Heber, Veronika Klemm, Claudia Pudack, Nadia Vögel, Adélie Delacour.

Je voudrais remercier pour leur indéfectible amitié depuis vingt ans les Malgaches

(Benoît, Lolo, Paulo, Gégé, Papy, la ‘Miche’, Alex, Vincent, le ‘Dude’, Micool, Toto, Guigui, Pi-oui, David, Tata, Yanno) et les Malgachettes (la ‘Marie’, Nono, Valou, Hélène, Caro, Cécé, Dodo). Longue vie à l’association ‘Tripcycle’!

Enfin pour terminer, je voudrais exprimer toute ma gratitude aux membres de ma

famille pour leur soutien, leur encouragement et leur patience. Rien n’aurait été possible sans eux: en tout premier lieu, mes parents Anne Clément et Jean Touboul qui ont su éveiller ma curiosité scientifique et m’apprendre à raisonner ; merci à mes frères et sœurs (Vincent, Cécile, Justine et Marion); merci à Thierry et Isabel ; merci à mes regrettés grands parents (Yvonne, Albert, Monette et Jacques).

Je suis sincèrement reconnaissant à l’Etat Français de m’avoir permis de parvenir

ici grâce à l’enseignement fourni au cours de ma scolarité et de mes études universitaires et au ministère français de la recherche pour m’avoir offert une bourse de recherche pour cette étude.

290