Paléomagnétisme De La Matière Extraterrestre: Implications Pour La

« Absorbé par les recherches sur l’infiniment grand, puis l’infiniment petit, le scientifique a négligé, de façon coupable, le fantastique potentiel de l’infiniment moyen. »

Grégoire Lacroix

« Alors, tu en es où de tes études ? » « Et bien, j’ai rendu mon manuscrit de thèse au mois d’octobre et je soutiens le 4 décembre 2013 »…. Déjà. Ces trois année s de thèse sont passées à toute allure, entre expériences, rencontres, voyages, rigolades… (moment de nosta lgie… :-/ mais positive hein !! L’heure est donc aux remerciements ;-)).

En premier lieu, je tiens à remercier mes directeurs de thèse, Mr Pierre Rochette et tout particulièrement Mr Jérôme Gattacceca, sans qui toute cette « histoire » n’aurait pas pu être écrite, pour la confiance qu’ils m’ont accordée en me confiant ce trav ail de recherche ainsi que pour leur aide, leurs encouragements et leurs précieux conseils au cours de ces années. Je les remercie aussi pour toutes les opportunités qu’ils m’ont offert es à travers de multiples voyages.

Je remercie également Mr Nicolas Thouveny, directeur du laboratoire du CEREGE pour m’avoir accueillie au sein de cette institution et pour les conseils et les encouragements que j’ai eu l’honneur de recevoir de sa part.

J’ai pu travailler dans un cadre particulièrement agréable, grâce à l’ensemble des membres de l’équipe « géophysique/planétologie » du CEREGE que j’ai intégrée pour ces 3 ans de thèse (et encore pour un an ;)), je vous en remercie.

Ces trois années n’auraient pas été ce qu’elles ont été sans Mr François Demory, ingénieur de recherche, « mon mentor »… fidèle collègue et ami à qui j’aimerais adresser un remerciement particulier pour tout (il y en a tellement à dire mais bon, disons…) ses conseils, sa disponib ilité, son soutien indéfectible et ses tranches de rigolades J autour d’un café. Un grand merci aussi à sa femme, Juliette Lamarche, pour ses tuyaux (tant scientifiques que culinaires ;)) et sa bonne humeur. Merci à vous deux pour tous ces bons moments aux Salles ;) !

Mr Ben P. Weiss et Mr Matthieu Gounelle ont accepté d’être les rapporteurs de cette thèse et je les en remercie. Ils ont également contribué par leurs nombreuses remarques et suggestions à améliorer la qualité de ce mémoire, et je leur en suis très reconnaissante. Merci à Mme France Lagroix et Mr Bertrand Devouard, examinateurs, pour leur participation au jury.

Elles sont arrivées en même temps que moi au CEREGE pour commencer leur thèse, Aurore et Carole, merci. Aurore pour tes qualités d’orga nisatrice et de cuisinière ;), Carole pour ton soutien et ta relecture lors de cette « douloureuse » dernière ligne droite que représente la fin de la thèse.

Je tiens également à remercier Isabelle Hammad, secrétaire de l’Ecole doctorale, pour sa gentilles se et sa patience, notamment dans la galère administrative qu’implique chaque nouvelle inscription. Dans la même optique je remercie également toutes les secrétaires fac et CNRS du CEREGE pour leur aide et leur efficacité lors de chaque mission.

Je souhaite encore remercier toutes ces personnes que j’ai pu croiser ou côtoyer au cours de ces 3 années (en espérant n’oublier personne) : Mme Brigitte Zanda (pour ses conseils et sa disponiblité), Hanane (pour ta bonne humeur et ton amitié), Neija (une vraie maman pour nous dans le désert Chilien), Laurette Scifo, Minoru (Le McGyver du chalet et le photographe du Chili), Millarca, Christine Pailles, Doriane Sabatier, Laëtitia Licari, Anne Alexandre, Yves Gally, Thibault de Garidel-Thoron, Corinne Sonzoni, Matthieu Ghilardi, Clément Suavet, Sonia Tikoo, 4 tous mes consoeurs et confrères thésards du CEREGE, Lauren, Lucie, Hélène, Abir, Marcela, Jade, Adrien, Camille, Sophie, Lise, Anne-Eléonore, Nicolas B., Nicolas B. bis !, Nathan, Elodie, Jérémy, Lucie, Julie Arthur, Fabienne, Alice, Marine, Pierre-Olivier, Laurie. Dédicace spéciale à tous ces « petits » étudiants qui sont passés par « le Chalet » Camille, Estelle, Florent, Frédéric, Hugo, Marine, Thibaut.

Ces trois années dans une nouvelle région m’ont permis de fair e de superbes rencontres Ted (le blond), Lolo (la blonde), Chris (fiston), Aurore (louloute), Céline (la pintche), Jérôme (BG), Brice (Bryce), Minh (le chnaw), Mica, Ludo, François, Lucas, Antoine, Laura, Nico, Julien, Christelle, Florian, Caroline, Hakim, Steve, Marie-Alix, Fabrice, Quentin, Caro, Loïc, Emeline, en espérant vous revoir, je vous remercie tous pour ces soirées et votre bonne humeur J. Je remercie aussi Sabine (fraisou), Tiphaine (pti’écureuil), Sylvain (cartman), Sébastien ( touffu), Yannick (yannou), Jean-Louis (loulou), Thierry (dorifore) et Carole, amis de longue date, qui, malgré la distance, ont toujours été là.

Il a su me donner le sourire même lorsque j’étais épuisée, il m’a soutenue jusqu’à la dernière nuit de rédaction et a participé à la vérification de toute la bibliographie, un immense merci à toi Thibaut (ma branque) ;).

Ces remerciements ne seraient pas complets sans mentionner toutes ces personnes qui font partie de ma famille : mes fréros (Embriqué et Dam), mes grands-parents, mes beaux-parents, mes oncles et tantes, les couz, Juju et Aud (les bofs) et surtout surtout mes parents, sans qui je ne serais pas là aujourd’hui, qui ont suivi mon évolution et qui m’ont toujours soutenue.

Enfin, je remercie mon cher époux, « Doudou », qu i s’est transformé en homme à tout faire pendant tout le temps de ma rédaction, pour sa patience, sa relecture du manuscrit, et son soutien sans faille depuis maintenant plus de 8 ans…

1.1 Généralités 12

1.2 La matière solide du système solaire 13

1.2.1 Formation du Système solaire 13

1.2.2 Les matériaux extraterrestres disponibles à l’étude 14

1.2.3 Classification des météorites 15

1.3 Les grandes questions du magnétisme extraterrestre 18

1.4 Les champs magnétiques dans le système solaire primitif 19

1.4.1 Les champs externes : solaires et nébulaires 19

1.4.2 Les champs de dynamo 21

1.4.3 Les champs crustaux 22

1.4.4 Quels champs enregistrés dans les météorites ? 22

1.5 Structure de la thèse 24

2.1 Introduction 26

2.1.1 Généralités 26

2.1.2 Rappels de magnétisme des roches 27

2.2 Contraintes intrinsèques liées aux échantillons 28

2.3 Propriétés magnétiques des matériaux extraterrestres 29

2.3.1 Minéraux magnétiques dans la matière extraterrestre 29

2.3.2 Des mécanismes d’aimantation particuliers 30

2.3.2.1 Introduction 30

2.3.2.2 Les mécanismes d’aimantation terrestres 31

2.3.2.3 Les mécanismes d’aimantation de la matière extraterrest re 31

2.3.2.4 La contamination terrestre 36

2.3.3 Méthodes d’estimation de la paléointensité 38

2.3.3.1 Méthode Thellier-Thellier 38

2.3.3.2 Les méthodes utilisant les champs alternatifs 38

3.1 Introduction 40

3.2 Etude de gros échantillons Apollo 42

4.1 Les Chondrites 80

4.2 Les Chondrites Carbonées CM 82

4.2.1 Le magnétisme des chondrites carbonées 82

4.2.2 Le magnétisme des chondrites carbonées CM 83

4.3 Les rumurutites 119

136

140

150

Article 2 : Hewins, R.H., Bourot-Denis, M., Zanda, B., Leroux, H., Barrat, J.-A., Humayun, M., Göpel, C., Greenwood, R.C., Franchi, I.A., Pont, S., Lorand, J.-P., Cournède, C., Gattacceca, J., Rochette, P., Kuga, M., Marrocchi, Y., Marty, B., 2014. The Paris meteorite, the less altered CM chondrite so far. M. Geochim. Cosmochim. Acta 124, 190- 222.

Liste des abréviations et symboles:

NRM : natural remanent magnetization , aimantation rémanente naturelle

ChRM : characteristic remanent magnetization , aimantation rémanente caractéristique

ARM : anhysteretic remanent magnetization , aimantation rémanente anhystérétique

CRM : chemical remanent magnetization , aimantation rémanente chimique

IRM : isothermal remanent magnetization , aimantation rémanente isotherme

SIRM : saturated isothermal remanent magnetization , aimantation rémanente isotherme à saturation

GRM : gyroremanent magnetization , aimantation gyro-rémanente

TRM : thermal remanent magnetization , aimantation rémanente thermique pTRM : partial thermal remanent magnetization , aimantation rémanente thermique partielle

VRM : viscous remanent magnétization , aimantation rémanente visqueuse

PRM : piezo-remanent magnetization , aimantation rémanente piézométrique

SRM : shock remanent magnetization , aimantation rémanente de choc

DRM : detrial remanent magnetization , aimantation rémanente détritique dDRM : depositional detrial remanent magnetization , aimantation rémanente détritique dépositionnelle pDRM : post-depositional remanent magnetization , aimantation rémanente post-dépôts

ADRM : accretional detrital remanent magnetization , aimantation rémanente détritique d’accrétion

AF: alternating field , champ alternatif

AARM : anisotropy of anhysteretic remanent magnetization , anisotropie d’ aimantation rémanente anhystérétique

K : magnetic susceptibility , susceptibilité magnétique (SI)

χ : specific magnetic susceptibility , susceptibilité magnétique spécifique (m 3 kg -1)

AMS : anisotropy of magnetic susceptibility , anisotropie de susceptibilité magnétique

P : anisotropy degree for magnetic susceptibility , degré d’anisotropie de la suscepti bilité magnétique

T : shape parameter for magnetic susceptibility , paramètre de forme de la susceptibilité magnétique

Prem : anisotropy degree of remanence , degré d’anisotropie de rémanence

Trem : shape parameter of remanence , paramètre de forme de rémanence

VSM : vibrating sample magnetometer , magnétomèter à échantillon vibrant

Ms : saturation magnetization , aimantation à saturation

Mrs : saturation remanence , aimantation rémanente à saturation

Bcr : coercivity of remanence , champ coercitif de rémanence

Bc : coercivity , champ coercitif

SD : singledomain , monodomaine

PSD : pseudo-single domain , pseudo-monodomaine

MD : multi-domain , poly-domaine

Tc : curie temperature , température de Curie

IDP : interplanetary dust particles , particules de poussières interplanétaires

MDF : median destructive field , champs de destruction médian

MAD : maximum angular deviation , écart angulaire maximum

HC : high corercivity , haute coercivité

MC : medium coercivity , moyenne coercivité

LC : low coercivity , faible coercivité

HT : high temperature , haute température

MT : medium temperature , moyenne température

LT : low temperature , basse température

RT-SIRM : room temperature SIRM , SIRM à temperature ambiante

LT-SIRM : low temperature SIRM , SIRM à basse température

FC : field cooling , refroidissement sous champ

ZFC : zero field cooling , refroidissement en champ nul

CHAPITRE 1 : INTRODUCTION GENERALE

Chapitre 1: Introduction générale 1.1 Généralités

Pourquoi les météorites sont-elles aimantées? Quel est le champ à l’origine de l’aimantation portée par les météorites ? L'aimantation des météorites et des surfaces planétaires peut-elle survivre aux différents évènements qui façonnent leur histoire? (variations thermiques, les chocs, l'exposition à des champs magnétiques, le métamorphisme, l'altération, la déformation...). La Lune et Mars ont-t-elles possédé une dynamo au début de leur histoire? Telles sont les principales questions que nous avons abordées dans le cadre de cette thèse. Depuis plus d’un siècle le p aléomagnétisme et le magnétisme des roches ont contribué à l'avancement des sciences de la Terre, par exemple pour la compréhension de la tectonique des plaques ou pour la datation des sédiments via la magnétostratigraphie. Ces techniques ont été appliquées plus récemment pour l’étude d’un certain nombre de matériaux extraterrestres. L’étude du paléomagnétisme extraterrestre a débuté il y a un peu p lus de soixante ans (Anyzeski, 1949 ; Levinson-Lessing, 1952 ; Fonton, 1954 ). Dès lors s’est ouverte la perspective d’un nouveau domaine d’étude : l’analyse et l’interprétation du magnétisme de la matière extraterrestre. Peu de temps après, à la fin des années 1950, les premières études détaillées ont vu le jour ( Lovering, 1959 ; Stacey and Lovering, 1959 ). Au début des années 1970, ce domaine a connu un nouvel essor avec le retour des échantillons lunaires Apollo et Luna ( Fuller and Cisowski, 1987 pour une synthèse ). S’en sont alors suivies de nombreuses études concernant l’aimantation et les propriétés magnétiques des météorites (Levy and Sonett, 1978 ; Hood and Cisowski, 1983 ; Cisowski, 1987 ; Sugiura and Strangway, 1987 pour une synthèse). Ces travaux qui ont ébauché la connaissance du magnétisme des roches extraterrestres (minéralogie et propriétés magnétiques) ont aussi tenté d’estimer la paléointensité des champs magnétiques du système solaire. Il a alors été émis l’hypothèse que les constituants des chondrites primitives auraient enregistré des champs magnétiques dans le système solaire jeune, lors de leur condensation dans le disque protoplanétaire il y a 4.6 Ga (e.g. Nagata, 1979a ; Sugiura and Strangway, 1981 ). Les études ultérieures proposent une alternative à cette hypothèse, ceci étant plus largement discutée dans la partie 1.5. L’étude du magnétisme extraterrestre connaît un nouvel essor depuis une petite dizaine d'années (e.g., Rochette et al., 2009 ; Weiss et al., 2010 pour une synthèse) provoquée par les progrès des techniques du paléomagnétisme et de l’instrumentation, la collecte d'échantillons de plus en plus nombreux et variés et les progrès sur la théorie de la dynamo. L'enregistrement 12

CHAPITRE 1 : INTRODUCTION GENERALE

paléomagnétique extraterrestre est complémentaire de la géochimie et pétrologie et l’augmentation croissante de données dans ces deux domaines sont autant d’informations cruciales, contextuelles et géochronologiques, pour comprendre la nature et l'origine de l'aimantation rémanente enregistrée dans la matière extraterrestre Dans un contexte plus général, toutes ces nouvelles données laissent entrevoir l’espoir de pouvoir répondre de façon de plus en plus précise aux questions que suscite la diversité de la matière extraterrestre. En contraignant la différenciation précoce et l'histoire thermique des corps du système solaire , ainsi que l’existence de champs magnétiques dans le disque d’accrétion , nous pourrons comprendre comment a pu se former le système solaire tel que nous le connaissons actuellement.

1.2 La matière solide du système solaire

1.2.1 Formation du Système solaire

La formation du système solaire a débuté il y a ~4.57 milliards d'années suite à l'effondrement gravitationnel d'un nuage de gaz et de poussière (Fig 1.1 ). De cet effondrement va naître un disque, appelé disque protoplanétaire. Au centre de ce disque, la plus grande partie de la masse du nuage initial va se contracter sur elle-même pour donner naissance à une étoile, le Soleil (qui rassemble aujourd'hui 99.9% de la matière du système solaire). Le disque protoplanétaire présente dès lors un fractionnement chimique et une variation radiale de température et de pression. Au fur et à mesure que la température va diminuer les poussières solides vont se former par condensation puis s ous l’effet de la gravitation les premiers constituants des futurs corps planétaires vont s’accréter. Les matériaux les plus réfractaires condensent en premier ; on parle d’inclusions alumino- calciques (CAIs). Ces premiers solides datés ~4568 Ma ( Amelin et al., 2002 (4567.2 ± 0.6 Ma) ; Bouvier and Wadhwa, 2010 (4568.2± 0.4 Ma)) sont considérés comme l'instant zéro du système solaire. Condensent ensuite les silicates (olivine et pyroxène essentiellement), les alliages de fer et nickel métalliques puis les sulfures de fer et de nickel. Loin du Soleil, le disque est froid : on trouve alors les glaces d'eau, de méthane, d'ammoniaque et d'oxydes de carbone (monoxyde et dioxyde). Dans une ultime étape, tous les éléments cités ci-dessus vont s’agglomérer pour former des corps de plus en plus massifs, et finalement générer les planètes et leur(s) satellite(s), les astéroïdes et les comètes. Ainsi en quelques millions d’années seulement, le disque initial a donné naissance à tous les corps planétaires et à toute la matière solide que l’ on trouve actuellement dans le Système Solaire.

CHAPITRE 1 : INTRODUCTION GENERALE

Fig. 1.1. Chronologie de la formation et de l’évolution des météorites.

1.2.2 Les matériaux extraterrestres disponibles à l’étude

De nos jours, pour les études en laboratoire nous disposons de quatre grands types de matériaux extraterrestres : les météorites, les micrométéorites, les poussières interplanétaires ( interplanetary dust particles , IDP) et les échantillons récupérés in situ sur des corps extraterrestres (e.g. échantillons lunaires issus des missions Apollo). Dans le système solaire, lors des collisions des fragments d'astéroïdes ou de planètes sont arrachés de leur corps parent. Après un temps de transfert dans l’espace plus ou moins long (de quelques milliers à plusieurs centaines de millions d’années), qui dépend de la région d’où ils proviennent, ces fragments peuvent atteindre la surface terrestre ; il s’agit des météorites . Les météorites peuvent donc provenir des astéroïdes (la grande majorité), de la planète Mars, de la Lune, et certaines météorites ont peut-être une origine cométaire (Gounelle et al., 2006 ). On discerne les « chutes », météorites dont la chute sur Terre a pu être observée et qui ont été retrouvées peu après leur atterrissage, des « trouvailles », météorites découvertes par hasard ou lors de missions de collectes systématiques, essentiellement dans des environnements désertiques (Sahara, Antarctique...) sans que leur chute soit observée. De façon générale, les météorites ont une taille comprise entre quelques millimètres et quelques mètres. Au-delà de cette taille, la météorite est entièrement détruite. Le flux météoritique sur Terre est estimé à 5000 T/an (Halliday et al., 1989 ; Bland et al., 1996 ). On distingue les micrométéorites qui sont, comme leur nom l’ind ique, des météorites de petites tailles, comprise entre 10 μm et 0.1–1 mm (correspondant à la taille minimum des minéraux 14

CHAPITRE 1 : INTRODUCTION GENERALE

constitutifs de la météorite). En deçà de 10 µm on parle de poussière planétaire (IDP) qui est incapable de produire un météore lors de son passage dans l'atmosphère. Les micrométéorites et les IDP constituent l'essentiel du flux de matière extraterrestre vers la Terre (2700 ± 1400 T/an) (Taylor et al., 1998 ). Parmi toutes les missions spatiales, 5 ont permis le retour d’échantillons extra -terrestres. Les missions Apollo (1969-1972) et Luna (1970-1976) ont per mis le retour d’échantillons lunaires (au total ~380 kg et 320 g respectivement). La mission Stardust (lancée en 1999) a ramené sur Terre 100 µg de poussière de la comète Wild 2. Il s’avère que la composition des grains de Stardust ressemble fortement à celle typique des météorites primitives (présence de phases réfractaires dont CAI, compositions isotopiques dans la même gamme). Ceci conforte les modèles de mélange de la matière à très grande échelle dans le système solaire naissant. La sonde Hayabusa (2005) est revenue sur Terre en juin 2010 avec 1534 minuscules échantillons de poussière minérale de l’astéroïde Itokawa . Ces infimes indices ont montré que la composition de la poussière minérale est semblable à celle des météorites que l’on trouve sur Terre, les chondrites (plus exactement les chondrites LL). Enfin, la sonde Genesis (2001-2004) a capturé des particules de vent solaire, le but étant de mieux comprendre la formation et le fonctionnement du Soleil. Le dernier projet en date, le projet Phobos-Grunt (2011) qui avait prévu le retour d’échantillons de la surface de Phobos, un des deux satellites de Mars, au printemps 2013 n’a malheureusement pas aboutit (la sonde n’ayant jamais réussi à quitter l'orbite terrestre).

1.2.3 Classification des météorites

Fig. 1.2. Classification des m étéorites d’après Bischoff, 2001 . Les météorites étudiées dans cette thèse sont figurées par un cercle rouge.

CHAPITRE 1 : INTRODUCTION GENERALE

Rappelons brièvement comment s’organise la classification des météorites. Actuellement on distingue deux grandes catégories de météorites (Fig. 1.2 ) : - Les météorites non-différenciées, appelées aussi chondrites, proviennent d'astéroïdes formés tardivement ou trop petits pour avoir emmagasinés une quantité de chaleur suffisante depuis leur formation pour qu’une différenciation interne ait lieu (Fig 1.3 ). Ces météorites sont caractérisées par la présence de chondres , c’est à dire de petites sphères dont la taille varie entre 20 μm et quelques mm de diamètre, formées à haute température peu de temps (quelque Ma au plus) après la formation des inclusions réfractaires (Fig.1.1 ). Ils sont composés de silicates, de verre, d'alliages fer-nickel sous forme métallique, de sulfures et parfois de CAIs. Les premiers planétésimaux à l’origine des corps parents des chondrites se sont formés par agglomération de chondres, de grains de métal et de sulfure le tout étant cimenté par une matrice finement cristallisée. Les chondrites sont les plus abondantes des chutes observées, et représentent environ ~80 % des chutes sur la période 1800-2000. Parmi les chondrites, on distingue, grossièrement selon la distance croissante entre le lieu de formation et le Soleil : les chondrites à enstatite, les chondrites ordinaires (93 % des chondrites), les chondrites carbonées (4 % des chondrites), les Rumurutites (R) et les Kakangarites (K) ( Weisberg et al., 1996, 2006 ). Ceci est une hypothèse, les variations observées pourraient également être liées à la chronologie de formation de ces différents groupes de météorites. - Les météorites différenciées , sont issues de corps parents plus gros (de diamètres de dix à plusieurs centaines de kilomètres) ce qui a permis d'enclencher une différenciation. Ce phénomène va conduire à une réorganisation de la matière au sein des corps chondritiques initiaux. Grâce à la fusion partielle va se produire une ségrégation gravitaire entre un noyau dense, majoritairement composé des liquides les plus lourds (fer-nickel), et un manteau et une croûte rocheuse moins dense, formés des liquides les plus légers (silicates). Les chondrites différenciées vont donc se diviser en trois grandes classes, chacune représentant l'une des diverses parties d'un corps parent ainsi différencié ; les achondrites (silicatées, ayant pour origine la croûte et/ou le manteau des corps parents), les ferreuses (ayant pour origine les noyaux des corps parents) et les « météorites mixtes » (pallasites et mésosidérites) proviendraient en partie de l’interface entre le noyau métallique et le manteau pierreux , bien que d’autres hypothèse s aient été avancées (Tarduno et al., 2012 ). On notera l’existence d’un « troisième » groupe de météorites, les « achondrites primitives ». Ces dernières possèdent des reliques de chondres mais ont subi une fusion partielle. Elles sont considérées soit comme une classe à part entière au même rang que les chondrites et achondrites (Weisberg et al., 2006 ) soit comme une sous-classe appartenant aux achondrites ( Bischoff et al., 16

CHAPITRE 1 : INTRODUCTION GENERALE

2001 , Fig. 1.2). Cette classe de météorites illustre le continuum entre les météorites différenciées et non différenciées, ce qui laisse entrevoir la possibilité de corps parents à l’histoire plus complexe avec par exemple une différenciation partielle (plus largement discuté dans la partie 1.5). A l’heure actuelle, pour la plupart des météorites, les corps parents ne sont pas identifiés. Les seules météorites rattachées à un corps connu sont les météorites martiennes (encore appelées SNC), les météorites lunaires et les météorites « HED » qui viendraient de l'astéroïde Vesta (e.g., Binzel and Xu, 1993 ).

Fig 1.3 Présentation schématique des grands groupes de météorites. 17

CHAPITRE 1 : INTRODUCTION GENERALE

1.3 Les grandes questions du magnétisme extraterrestre

L’une des grandes questions concernant la formation du système solaire et l’évolution des planètes concerne les processus de formation des corps et leur chronologie. C ’est notamment à travers l’étude du magnétisme de la matière extraterrestre, que l’on peut espérer y répondre. L’étude du paléomagnétisme extraterrestre représente bien plus que la simple détection des champs magnétiques anciens potentiellement enregistrés dans ces objets. En effet, actuellement le champ magnétique interplanétaire est de l’ordre du nT alors qu’il est avéré que certaines météorites portent une aimantation rémanente acquise dans un champ bien plus intense (>µT). Ce simple constat permet d’évoquer l’idée que les météorites ont enregistré des champs primitifs dans la nébuleuse ou des champs de dynamo qui se seraient formés très tôt au sein des corps du système solaire. De ces hypothèses découlent de nombreuses questions, telles que : la formation de dynamo ne concerne-t-elle alors que les corps parents des achondrites ? Les chondrites ont- elles enregistré uniquement des champs nébulaires ? Certains astéroïdes ont-ils pu entretenir une dynamo très peu de temps après leur accrétion comme suggéré par les résultats sur les angrites (Weiss et al., 2008a) ? Une dynamo a-t-elle fonctionnée sur la Lune (e.g., Collinson, 1993 ), sur Mars ? Si oui, pendant combien de temps ? Les mesures paléomagnétiques sur les achondrites vont donc permettre de définir si les corps parents ont subi une différentiation et ont été capable de générer une dynamo interne . La réponse à cette question est d’une grande importance pour comprendre la structure et l’évolution thermique des corps parents ainsi que l’existence de sources d ’énergies appropriées. D’autre part, l’intensité des champs magnétiques dans le disque protoplanétaire (avant l'accrétion des planétésimaux) est mal contrainte. L’estimation de l’intensité de ces champs pourrait fournir une contrain te essentielle en astrophysique et en particulier dans les scénarios d’évolution des disques stellaires. En effet, ces champs magnétiques peuvent fortement influencer le transfert de masse et de moment vers l’intérieur et l’extérieur du disque protoplanétaire et jouent un rôle clé dans le contrôle des conditions dynamiques de l'accrétion de la matière dans le disque (Bouvier et al., 2007 ), en particulier via les instabilités magnéto-rotationnelles ( Balbus, 2003, 2011 ). L'origine des champs est donc d'un intérêt considérabl e, qu’ils soient associés au disque protoplanétaire ou au Soleil dans sa phase T-Tauri, ou bien qu’ils soient générés au sein d'un corps parent planétaire ou d’un astéroïde par un processus de dynamo. L’âge (~4.55 Ma) et la provenance des météorites (~3 UA) sont des atouts pour tenter de répondre à ces questions mais constituent également des inconvénients. En effet, les possibilités de ré-aimantations successives subies au cours de leur histoire (lors d’évènements thermiques,

CHAPITRE 1 : INTRODUCTION GENERALE

chocs…) restreignent la proportion de matériaux capables de conserver leur aimantation primordiale.

1.4 Les champs magnétiques dans le système solaire primitif

Comme nous venons de le voir précédemment, les champs magnétiques présents dans le système solaire primitif et potentiellement à l’origine de l’aimantati on de la matière extraterrestre peuvent-être de différentes natures (Fig. 1.4 ) : les champs magnétiques solaires, les champs magnétiques dans le disque protoplanétaire, les champs magnétiques de dynamo générés dans des noyaux des corps différenciés, ou encore des champs générés localement en surface. Nous détaillons ici brièvement les caractéristiques de ces différents champs magnétiques et les implications en termes de paléomagnétisme extraterrestre.

1.4.1 Les champs externes : solaires et nébulaires

Le Soleil primitif et le disque protoplanétaire sont considérés comme de potentielles sources de génération de champ significatif lors des premiers Ma de l’histoire de la formation du système solaire (Balbus, 2009 ). Au début de son histoire le soleil a peut-être connu une phase d’activité dite, « T-Tauri ». Au cours de cette phase le soleil aurait généré des champs magnétiques intenses et notamment un champ de dynamo stable approximativement dipolaire de longue durée associé à des éruptions de matière transitoire (Vallée, 2011 ). Cette dynamo génère des champs d’environ 10 5 µT à la surface du Soleil (Guenther et al., 1999 ). L’activité paroxysmale de l’étoile crée également des champs (10 4 µT) liés à des flares transitoires (quelques heures) (Vallée, 2003 ). Cependant, dans la région de la ceinture d’astéroïdes (~3 UA), l’intensité des champs de dynamo solaire (T-Tauri) et des flares est beaucoup plus faible, comprise entre 0.001 et 0.01 µT à cause de la décroissance avec l’inverse du cube de la distance au centre stellaire. Le disque protoplanétaire lui-même a pu être à l’origine de champs magnétiques à grande échelle via des phénomènes de dynamo se produisant directement dans le gaz du disque ( Levy, 1978 ). Récemment des simulations magnétohydrodynamiques ont permis d’estimer des intensités de l’o rdre de 100 et 10 µT dans le disque protoplanétaire pour des distances de 1 et 5 UA respectivement. Cependant ces champs ne seraient pas stables dans le temps avec des temps caractéristiques de quelques siècles (Turner and Sano, 2008 ).

CHAPITRE 1 : INTRODUCTION GENERALE

Champs magnétiques d'origine externe

A gauche, image capturée par l’observatoire solaire de la Nasa SDO d’une éruption de catégorie X1,7 – c’est -à-dire spécialement puissante – à la surface du Soleil, le 12 mai 2013. L’astre traverse actuellement un pic d’activité. A droite , image NASA du télescope spatial Hubble, la protoétoile HH30 montre un jet de matière perpendiculaire au disque d'accrétion guidée par des champs magnétiques, échelle

Champs magnétiques d'origine interne

Manifestation des dynamos planétaires : les aurores boréales sur la Terre (à gauche), sur Jupiter (au centre, image NASA télescope spatial Hubble) et sur Saturne (à droite, image NASA, télescope Hubble).

Champs magnétiques de surface

A gauche, illustration des passages rapprochés de la sonde Mars Global Surveyor avec la surface de Martienne révèlent la présence d'anomalies magnétiques. (Crédit photo : MER Science Team). A droite, c artographie des champs magnétiques mesuré s à la surface de Mars (données de Lillis et al., 2008 ).

Fig. 1.4. Manifestations des différents types de champs magnétiques présents dans le système solaire à l’échelle macroscopique.

CHAPITRE 1 : INTRODUCTION GENERALE

Les astronomes ont d'ailleurs détecté des champs magnétiques à géométrie complexe de l’ordre de ~100 μT ( Johansen, 2009 ; Sano et al., 2004 ) dans les zones internes ionisées du disque protoplanétaire. Enfin, des décharges électriques (éclairs) ont pu se produire dans le disque de poussière protoplanétaire avec des champs magnétiques associés localement très forts, jusqu’à plusieurs centaines de mT (Pilipp et al., 1998 ).

1.4.2 Les champs de dynamo

Parmi les événements qui régissent l'histoire planétaire, les plus importants sont peut-être ceux responsables de la différenciation planétaire globale. Dans le système solaire actuel, toutes les géantes gazeuses (Jupiter, Saturne, Uranus, Neptune), mais aussi deux satellites de Jupiter (Ganymède et Io), possèdent un champ magnétique d’origine interne généré par un phénomène de dynamo ( Stevenson, 2003 pour une synthèse). Les planètes telluriques (Mercure, Vénus, Terre, et Mars) et la Lune possèdent toutes une structure radiale différenciée avec un noyau métallique riche en fer dans lequel la génération de dynamo a pu avoir lieu. Mais seules la Terre et Mercure (étudiée par la mission Messenger) possèdent un champ magnétique produit par une dynamo dans leur noyau. Vénus est peut-être actuellement en phase « inter-dynamo » avec des températures intérieures trop faibles pour que la convection thermique se produise et trop importantes pour que la solidification du noyau ait lieu (e.g. Stevenson, 2003 ). Mars a très probablement possédé une dynamo dans le passé (e.g., Lillis et al., 2013 ). Le cas de la Lune est plus largement discuté dans le Chapitre 3 . Dans le système solaire primitif, certains corps ont connu une fusion à grande échelle, conduisant à la formation d'un noyau métallique recouvert par un manteau rocheux silicaté et une croûte. Ces noyaux auraient été initialement fondus et auraient pu avoir des mouvements de convection et générer un champ magnétique de dynamo (Chabot and Haack, 2006 ). Pour les météorites, au-delà de la différenciation avérée des corps parents des achondrites, jusqu’à récemment il était difficile de savoir si les astéroïdes ont été capables de générer des champs magnétiques internes. Les données paléomagnétiques et les nouvelles contraintes issues de la modélisation de la génération des dynamos ont montré que la plupart des planétésimaux primitifs possédant un noyau métallique en convection étaient capables de générer des champs magnétiques très tôt (quelques Ma) après la formation des CAIs et ce pour une durée comprise entre ~10 et 100 Ma ( Elkins-Tanton et al., 2011 ; Fu et al., 2012 ; Sterenborg and Crowley, 2013 ; Tarduno et al., 2012 ; Weiss et al., 2010 ).

CHAPITRE 1 : INTRODUCTION GENERALE

1.4.3 Les champs crustaux

En l ’absence de champ global de dynamo l a Lune et Mars possèdent des champs de surface locaux créés par l’ aimantation rémanente de la croûte. Cette aimantation rémanente peut être le résultat d’ un phénomène de dynamo ayant opéré précédemment sur le corps en question, mais d’autres hypothèses sont envisageables (enregistrement d’un champ externe , impacts, décharges électriques à la surface...). Des éclairs ont par exemple été observés à la surface de Mars ( Ruf et al., 2009 ). Ces aimantations crustales doivent être considérées comme des sources de champ magnétique à part entière. Pour la Lune, les mesures de surface indiquent des champs magnétiques faibles de l’ordre de 100 nT au maximum (Dyal et al., 1970 ) et les mesures satellitaires indiquent une répartition de ces champs magnétiques très hétérogènes ( Richmond and Hood, 2008 ; Mitchell et al., 2008 ). Pour Mars, les champs de surface peuvent atteindre jusqu’à une centaine de µT à certains endroits (surtout au niveau des terrains les plus anciens) (Acuña et al., 1999 ). Dans les deux cas, les champs de surface génèrent des magnétosphères locales qui peuvent protéger la surface des particules solaires ( Blewett et al., 2007 ; Leblanc et al., 2006 ). Dans le même ordre d’idée, l’ab sence de « space weathering » à la surface de Vesta a été interprétée comme la preuve que la croûte de Vesta porte une aimantation rémanente qui protège la surface des particules chargées ( Vernazza et al., 2006 ).

1.4.4 Quels champs enregistrés dans les météorites ?

Les premières études paléomagnétiques sur les chondrites mettaient en évidence des paléochamps de forte intensité (~100 µT) (Banerjee and Hargraves, 1971 ; Brecher and Arrhenius 1974 ; Stacey, 1976 ). Il était également classiquement admis que les petits corps parents, tels que ceux imaginés pour la plupart des chondrites, été probablement trop petits pour se différencier et ainsi permettre le fonctionnement d’ une dynamo interne. Ainsi ces paléochamps étaient attribués à des champs solaires (T-Tauri) ou nébulaires. D’une part des considérations techniques révèlent que ces résultats anciens ne sont pas fiables (Weiss et al., 2010 ). D’autre part d e nouvelles connaissances en astrophysique remettent également en question ces conclusions. En effet la durée de vie des disques d'accrétion est estimée entre 2 et 6 Ma ( Haisch et al., 2001 ), au-delà le soleil entre dans sa séquence principale. Il semblerait que les champs nébulaires et T-Tauri aient disparu lorsque la dissipation du disque d’accrétion a eu lieu ( Evans et al., 2009 ). Ceci implique que la matière solide n’a pu s’ aimanter dans des champs externes que si l’aimantation était acquise avant la fin de la dissipation du disque. Pour un certain nombre de corps parents, le dernier épisode majeur d’altération aqueuse

CHAPITRE 1 : INTRODUCTION GENERALE

et/ou de métamorphisme thermique se produit trop tard (i.e. plus de 6 Ma après la formation des CAIs) pour pouvoir garder la mémoire des éventuels champs externes (Kleine et al., 2009 ). Cette courte durée de vie des champs magnétiques externes associée au fait que dans la région de la ceinture d’astéroïde , l’intensité des champs était probablement trop faible pour créer une aimantation rémanente mesurable dans les astéroïdes, indiquent que contrairement aux idées initiales, la plupart des chondrites ne peuvent pas conserver l’enregi strement de champs magnétiques créés dans le disque protoplanétaire ou par le Soleil primitif. De plus, des études récentes indiquent que certaines chondrites peuvent conserver l’enregistrement de champs d’origine interne. C’est le cas de la météorite Allende (de type CV3) qui serait issue des couches extérieures non-fondues d’un corps partiellement différencié qui possédait un noyau métallique en convection (Carpozen et al., 2011). De nouveaux modèles sur la possibilité de la génération de dynamo sur les petits corps viennent supporter ce nouveau concept (Elkins-Tanton et al., 2011 ). Les modèles d’évolution thermique montrent que des corps parents partiellement différenciés peuvent se former si l’accrétion débute moins de 1.5 Ma après la formation des CAIs, si elle se poursuit durant au moins quelques Ma, et si le rayon final minimum est >~6 km (Weiss and Elkins-Tanton, 2013 ). On notera que la chronométrie Hf/W valide le premier critère pour de nombreux corps parents ( Kleine et al., 2009 ). Un élément supplémentaire en faveur de la formation de ces corps partiellement différenciés est l’existence des achondrites primitives. En résumé, contrairement au paradigme développé aux débuts du paléomagnétisme des météorites, les chondrites pourraient nous renseigner non pas tant sur les champs magnétiques externes primitifs, mais plutôt sur des processus astéroïdaux, en particulier la différenciation. Les achondrites peuvent enregistrer des champs de dynamos liés à la différenciation de leur corps parent. Ceci a été mis en évidence pour certaines angrites ( Weiss et al., 2008a). Les achondrites peuvent également enregistrer des champs crustaux (qui peuvent être interprétés comme la preuve de l’existence d’une dynamo passée) . Une aimantation dans un champ crustal a été mise en évidence dans des eucrites ( Fu et al., 2012 ), et des météorites martiennes (Cisowski, 1986 ; Gattacceca and Rochette 2004 ). Récemment un enregistrement de champ crustal a été proposé pour la shergottite, Tissint ( Gattacceca et al., 2013a). Cet art icle, auquel j’ai contribué, est présenté en Annexe-1. Ces nouvelles avancées, rendues possibles en particulier par les progrès de la thermochronologie et de la modélisation numérique des astéroïdes, fournissent un cadre interprétatif plus robuste et plus varié pour le paléomagnétisme des chondrites.

CHAPITRE 1 : INTRODUCTION GENERALE

1.5 Structure de la thèse

L'essentiel de mon travail de thèse a porté sur le paléomagnétisme extraterrestre. En bref, il s'agit d'appréhender l'évolution de la matière solide du système solaire par le biais des mesures magnétiques. Nous nous sommes donc attachés à étudier l’aimantation rémanente naturelle (NRM 1) enregistrée dans les matériaux extraterrestres ainsi que les propriétés magnétiques indispensables à l’interprétation de l’ NRM qui sont également porteu ses d’information s sur la pétrogenèse des météorites. Nous avons étudié des roches lunaires des missions Apollo, des chondrites carbonées de type CM et des Rumurutites (chondrites R). De par leurs caractéristiques différentes (taille du corps parent, distance estimée de formation par rapport au Soleil, minéralogie magnétique, histoire thermique, chronologie) ces différentes roches permettent d’avoir une vision variée de la nature et de l’évolution des champs magnétiques qui régnaient dans le système solaire. Ces apports permettent de mieux contraindre les processus qui ont conduit à la formation du Système Solaire tel que nous le connaissons actuellement.

Après une brève introduction aux spécificités de l’étude du magnétisme de la matière extraterrestre (Chapitre 2 ), ce travail se divise en trois volets principaux et complémentaires qui font état des différents résultats obtenus pour chaque type d’échantillon . Le Chapitre 3 est consacré au magnétisme lunaire avec des résultats sur la chronologie, l’intensit é et la géométrie du champ de dynamo lunaire. Le Chapitre 4 traite des chondrites et se divise en deux grandes parties. La première partie s’intéresse aux chondrites carbonées CM. En plus d’une étude détaillée des propriétés magnétiques de la météorite Paris (en liaison avec son degré inhabituellement faible d’altération hydrothermale), les résultats de l’étude magnétique et paléomagnétique de sept chondrites de type CM sont présentés. L’essentiel de ces résultats concernent le paléomagnétisme. Nous montrons en particulier que ces chondrites portent le signal paléomagnétique le plus ancien jamais mis en évidence dans la matière solide du système solaire. La seconde partie porte sur les rumurutites où sont présentés les résultats préliminaires obtenus sur ces météorites. Enfin, le Chapitre 5 résume les principales conclusions de ce travail de thèse et discute des questions en suspens ainsi que des perspectives dans le domaine du magnétisme extraterrestre.

1 Tous les termes abrégés seront donnés en anglais afin d’éviter les confusions .

CHAPITRE 1 : INTRODUCTION GENERALE

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

Chapitre 2: Spécificités du magnétisme extraterrestre

Ce chapitre, après quelques rappels de magnétisme des roches, rend compte des spécificités de l’étude du magnétisme ext raterrestre que ce soit les contraintes intrinsèques liées aux échantillons (ab sence d’orientation, rareté) , les minéraux porteurs de l’aimantation ou les mécanismes d’acquisition de l’aimantation .

2.1 Introduction

2.1.1 Généralités

Comme nous l’avons vu dans le chapitre précédent, l'étude de l'aimantation rémanente des matériaux extraterrestres peut apporter des informations sur les champs magnétiques dans le système solaire jeune (phase T-Tauri du soleil, champs magnétiques dans le disque d’accrétion, fonctionnement de dynamo sur les planètes et les astéroïdes). Mais le paléomagnétisme extraterrestre est aussi un traceur de l'histoire de la matière solide du système solaire : évolution thermique, chocs, bréchification, métamorphisme et altération sur le corps parent. Pour cela, avant d’interpréter les mesures d’aimantation, il est indispensable de définir dans un premier temps les propriétés magnétiques intrinsèques des matériaux. Dans le cas de la matière extraterrestre les premières études sur les propriétés magnétiques n’ ont réellement commencé qu ’après le retour des échantillons Apollo (Strangway, 1978 ). Bien que des incertitudes planes encore, l’identification des porteurs magnétiques et la compréhension de leurs mécanismes d’aimantation sont de mieux en mieux contraints. Les études magnétiques vont d’abord s’attacher à définir la minéralogie magnétique des échantillons et ensuite tenter de comprendre comment ces minéraux ont pu s’aimanter, c’est -à- dire définir quel mécanisme est à l’origine de l’aimantation. De façon plus générale, la connaissance détaillée de la minéralogie magnétique des différents types de matériaux extraterrestres constitue également une source d’informations sur les conditions physico-chimiques (en particulier les conditions red-ox, notamment étudiées dans la météorite Tissint ( Gattacceca et al., 2013a-Annexe-1)) qui régnaient lors de la formation et de l'évolution des premiers solides du système solaire.

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

2.1.2 Rappels de magnétisme des roches

Les propriétés magnétiques de la matière tirent leur origine de courants microscopiques présents dans la matière. Ces courants sont le résultat de plusieurs phénomènes ayant lieu à l’échelle atomique; du mouvement des électrons autour des noyaux atomiques et du moment magnétique de spin propre (spin) d'un électron. On parle de moment magnétique (exprimé en Am²). La somme des moments magnétiques des atomes, normalisée au volume (en Am -1) ou à la masse (en Am² kg -1) s’appelle plus communément l’aimantation . Sur une même orbitale atomique, c’est l’appariement des électrons qui va faire que soit, la résultante magnétique sera nulle (spin non appariés, opposés), soit on aura un moment magnétique (spin appariés, parallèles). On notera cependant que certains éléments possèdent un moment magnétique bien que leurs électrons soient non appariés. A l’échelle cristalline, l’organisation de ces ato mes va permettre de distinguer différents types de comportement magnétiques ; le diamagnétisme, le paramagnétisme et le ferromagnétisme (qui se divise à son tour en trois sous-types principaux : ferro-, antiferro- et ferri- magnétisme). La capacité de tou t corps à acquérir une aimantation induite sous l’effet d’un champ inducteur s’appelle la susceptibilité magnétique (lorsque le champ inducteur est coupé, l’aimantation induite s ’annule ). Dans le cadre de l’étude du magnétisme extraterrestre nous ne nous intéresserons qu’au comportement ferromagnétique (au sens large ). Les minéraux ferromagnétiques sont caractérisés par un phénomène de rémanence. Cela signifie que ces minéraux ont la propriété de conserver une aimantation une fois que le champ inducteur est coupé : on parle alors d'aimantation rémanente naturelle (NRM). Ce sont donc ces minéraux qui vont assurer le rôle de «mémoire magnétique» des champs enregistrés il a plusieurs milliards d’années. En plus de la présence d’un champ magnétique et de minéraux ferromagnétiques, il faut également un évènement qui permette l’enregistrement magnétique. Parmi ces évènements on compte par exemple les chauffes, les chocs, la cristallisation de nouveaux minéraux, etc. La NRM des roches nous donnera ainsi des renseignements sur le métamorphisme, l’altération, l’expo sition à des champs magnétique, etc. A noter que l'enregistrement d'un champ magnétique nul (on parle alors de désaimantation) peut aussi être source d'information. Le paléomagnétisme correspond à l'étude de l’enregistrement du champ magnétique (aimantation rémanente) des roches. Le magnétisme des roches proprement dit regroupe grosso modo deux aspects , d’une part l'identification des minéraux magnétiques (minéralogie, taille des grains,

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

orientation des grain s...) et d’autre part la caractérisation de leurs propriétés intrinsèques (capacité à porter une aimantation rémanente, évolution avec la température...).

2.2 Contraintes intrinsèques liées aux échantillons

L’une des principales contraintes lorsque l’on ét udie le paléomagnétisme des météorites est que, contrairement aux roches étudiées dans le cadre du paléomagnétisme terrestre, de nombreux paramètres tels que le corps parent, le contexte géologique et l'orientation originale de presque tous les échantillons sont inconnus. Seuls les corps parents des météorites martiennes, lunaires et des HED (astéroïde Vesta) sont identifiés clairement. La Lune est le seul corps extraterrestre où les sites d’ échantillonnage sont connus (échantillons directement prélevés à la surface de la Lune), l’orientation originelle des échantillons restant cependant inconnue. Cette absence de contraintes contextuelles a conduit les études paléomagnétiques à se focaliser essentiellement sur l’estimation de la paléointensité du champ à l’origine de l’aimantation contenue dans les météorites . Dans le Chapitre 3 nous proposons cependant une méthode utilisant l’anisotr opie de susceptibilité magnétique des roches afin de remédier à ce problème d’orientation . Ceci, combiné à la connaissance des sites d’échantillonnage Apollo , nous permet de contraindre la géométrie du paléochamp lunaire. La rareté de certaines météorites fait que l’étude paléomagnétique se fait fréquemment sur de petites quantités de matériel. Par conséquent, il est parfois difficile de travailler sur des échantillons mutuellement orientés qui sont importants pour déterminer la nature de l’ aimantation (anté- ou post-accrétion par exemple). Les échantillons obtenus sont généralement de petite taille et peuvent donc avoir des moments relativement faibles et une anisotropie magnétique considérable (e.g., Tikoo et al., 2012 ). Les contraintes curatoriales font que souvent seule la méthode de désaimantation par champ alternatif (alternating field , AF) est autorisée. En effet la désaimantation thermique qui est classiquement utilisée pour la détermination des paléochamps des roches terrestres (Thellier and Thellier 1959 ) est à manipuler avec grande précaution (Stacey and Lovering, 1959 ; Stacey et al., 1961 ; Brecher et al., 1973 ; Cisowski et al., 1983 ) car elle peut rapidement modifier la minéralogie magnétique des matériaux extraterrestres qui se sont formés dans des milieux où la fugacité d’oxygène était beaucoup plus faible que sur Terre.

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

2.3 Propriétés magnétiques des matériaux extraterrestres

Comme nous l’avons abordé dans la p artie 2.1, le socle qui permet d’interpréter le signal paléomagnétique est la connaissance de la minéralogie magnétique des échantillons. Les météorites sont caractérisées par la présence de minéraux ferromagnétiques inhabituels et des mécanismes d'acquisition d’aimantation parfois exotiques (Rochette et al., 2009 ). Ceci complique l’interprétati on des données paléomagnétiques par rapport au cas terrestre où les principaux minéraux magnétiques (magnétite, hématite, pyrrhotite) sont bien connus et les mécanismes d’aimantation relativement bien cernés . De plus dans le cas des météorites de nombreux processus secondaires peuvent avoir lieu (chocs, transformations de phase à basse/haute température ainsi que l’a ltération terrestre). Ces processus doivent donc être pris en considération dans l'interprétation des mesures d'aimantation.

2.3.1 Minéraux magnétiques dans la matière extraterrestre

Les porteurs magnétiques extraterrestres diffèrent notablement des porteurs habituels du magnétisme terrestre. En effet, dans les roches terrestres, le fer se trouve sous forme d’oxydes, les minéraux magnétiques les plus abondants sont : la magnétite (Fe 304), la titanomagnétite (Fe 3– xTi xO4) et l’ hématite ( α-Fe 203). Dans la matière extraterrestre, le signal magnétique est dominé par le fer sous forme métallique. Une petite proportion de nickel est toujours associée au fer donnant ainsi une variété d’alliages Fe-Ni. Les matériaux extraterrestres renferment ainsi une diversité de phases ferromagnétiques allant des oxydes terrestres les plus communs aux siliciures et aux alliages de métaux les plus exotiques. Le système Fe-Ni est très complexe en raison de ces nombreuses phases (Cacciamani et al., 2006 ), plus connues sous le nom de kamacite, taénite, tétrataénite et awaruite qui dépendent de la vitesse de refroidissement et la teneur en nickel. Les propriétés magnétiques de ces phases sont peu connues et les processus d’ acquisition de la NRM mal compris (e.g., pour la kamacite (Garrick-Bethell and Weiss, 2010 ) et pour la tétrataénite (Gattacceca et al., 2003 ; Acton et al., 2007 ). Dans les météorites les plus réduites (e.g., aubrites, chondrites à enstatites), on peut trouver des phases de formule générale (Fe,Ni) 3X où X peut être C (cohénite), P (schreibersite) ou Si (suessite). Pour exemple, la cohénite porte la rémanence dans la chondrite à enstatite Abee (Sugiura and Strangway, 1983 ). La rémanence de certaines météorites pourrait être portée par des spinelles de Fe-Cr-Ti ( Yu and Gee, 2005 ). Les materiaux extraterrestres sont aussi communément riche en sulfures. Parmi les sulfures présents dans la matière extraterrestre, la seule phase magnétique est la pyrrhotite (Fe 1-XS) qui joue un rôle majeur dans les propriétés magnétiques des 29

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

météorites martiennes ( Gattacceca et al., 2013a-Annexe-1) ainsi que dans les chondrites carbonées et les rumurutites (Rochette et al., 2005, 2008 ). Les propriétés des minéraux magnétiques terrestres sont de loin les plus étudiées et les mieux connues. Dans les météorites la présence de phases communément rencontrées dans les roches terrestres (tel que la magnétite) rend leur étude paléomagnétique plus « simple ». C’est le cas par exemple des chondrites carbonées (CI, CK, CV), qui contiennent essentiellement de la magnétite ou des angrites et de certaines météorites martiennes dans lesquelles on trouve de la titanomagnétite. Cependant depuis une dizaine d’année une littérature relativement abondante a permis de préciser la nature et la teneur des minéraux ferromagnétiques de la majorité des matériaux extraterrestres (Rochette et al., 2009 pour une synthèse) que ce soit les différents groupes de météorites (Rochette et al., 2003a, 2008, 2009 ), des matériaux lunaires (Rochette et al., 2010 ), des roches martiennes ( Rochette et al., 2005 ), des micrométéorites ( Suavet et al., 2009 ) ou encore des tektites (Pechersky et al., 2012 ). Il est à noter que d ans les météorites il n’est pas rare que plusieurs phases magnétiques coexistent. Par exemple, certaines chondrites carbonées contiennent à la fois du fer métallique, de la magnétite et de la pyrrhotite. Ces mélanges peuvent parfois rendre complexe la détermination du minéral porteur de l’aimantation rémanente.

2.3.2 Des mécanismes d’aimantation particuliers

2.3.2.1 Introduction Une fois que la diversité et la complexité des minéraux magnétiques présents dans la matière extraterrestre ont été intégrées, se pose la question de savoir comment ces minéraux ont pu s'aimanter. Les processus d’aimantation de la matière solide peuvent avoir lieu avant (dans le disque protoplanétaire lorsque la matière est sous forme de poussière), pendant (lors de l’accrétion au corps parent) ou après (sur le corps parent) la formation de la roche. De nombre ux mécanismes peuvent être à l’origine de l’aimantation portée par la matière extraterrestre : la formation de minéraux magnétiques, un événement thermique, un choc. C’est à travers l’analyse des carac téristiques de la NRM que l’on va potentiellement pouvo ir obtenir des informations sur l'événement dans lequel elle a été acquise ainsi que sa chronologie. Ici nous rappelons tout d’abord rapidement quels sont les mécanismes d’aimantation que l’on peut rencontrer sur Terre. Nous verrons le cas de la matière extraterrestre afin de définir quel rôle ces mécanismes peuvent avoir sur l’enregistrement ou la modification du signal magnétique originel.

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

2.3.2.2 Les mécanismes d’aimantation terrestres Avant de détailler le cas de la matière extraterrestre faisons brièvement état des mécanismes d’aimantation connus sur Terre. De nombreuses roches terrestres, en particulier les roches ignées, possèdent une aimantation thermorémanente (thermoremanent magnetization , TRM) acquise par refroidissement dans un champ magnétique H donné, depuis le point de Curie des minéraux ferromagnétiques de la roche jusqu’à la température ambiante. On distingue les aimantations thermorémanentes partielles (partial thermoremanent magnetization , pTRM) acquises dans les mêmes conditions que la TRM mais avec un refroidissement à partir d’une température inférieure à la température de Curie (Thellier and Thellier, 1959 ). Les roches sédimentaires possèdent généralement une aimantation rémanente détritique ( detrital remanent magnetizaiton , DRM) acquise par alignement mécanique des grains ferromagnétiques contenus dans les sédiments dans le champ magnétique ambiant au cours ou peu après le dépôt des sédiments. On distingue l’aimantation acquise lors du dépôt du grain (dDRM aimantation rémanente détritiq ue dépositionnelle) et l’aimantation acquise peu après le dépôt (pDRM aimantation rémanente post - dépôt). La DRM est souvent sur-imprimée par une aimantation rémanente chimique (CRM) acquise plus tard au cours de la diagénèse. La CRM correspond aussi de façon générale à l’enregistrement de la direction du champ magnétique ambiant lors de la cristallisation de grains ferromagnétiques. Les roches peuvent également acquérir une aimantation piezo-rémanente (PRM) définie comme l’aimantation acquise lors de l’application d’une compression uniaxiale statique en présence d’un champ magnétique qui peut atteindre plusieurs dizaines de MPa. Ce phénomène a été décrit pour certaines roches ignées et métamorphiques ( Nagata, 1970 ) mais reste marginal sur Terre . Enfin l’aim antation rémanente viscqueuse ( viscous remanent magnetization , VRM) s’acquiert à température ambiante lorsque le matériau reste en présence d’un champ magnétique ambiant sur de longues périodes. Les grains de faible température de blocage (et en général de faible coercivité) vont alors acquérir spontanément une aimantation qui croît en fonction du logarithmique du temps.

2.3.2.3 Les mécanismes d’aimantation de la matière extraterrestre Dans la matière extraterrestre, en plus de ces mécanismes qui sont bien connus pour les roches terrestres, il existe des mécanismes spécifiques dont les propriétés sont moins bien établies : les chocs, le séjour prolongé dans les champs interplanétaires de quelques nT (équivalent à un stockage en champ nul), la rentrée atmosphérique, ainsi que tous les phénomènes qui peuvent avoir lieu après l’ arrivée sur Terre.

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

Dans la matière extraterrestre, la notion d’aimantation sous pression présente des caractéristiques sensiblement différentes de ce que l’on peut rencontrer sur Terre. L’aimantation sous pression des météorites est liée aux chocs. Bien que la notion ait été évoquée dans le cas des roches lunaires ( Nagata, 1973 ), on ne pa rlera pas de PRM mais plutôt d’aim antation remanente de choc (shock remanent magnetizaiton , SRM) pour les roches extraterrestres. C’ est un phénomène important lorsque l’on aborde le magnétisme extraterrestre. En effet, afin qu’une météorite arrive sur Terre il faut qu’elle soit arrachée à son corps parent , ce qui suppose un impact à grande vélocité (vites ses de l’ordre de 10-30 km.s -1). Toutes les surfaces planétaires anciennes présentent des cratères, témoins d’un intense bombardement. Et si on remonte au moment de la formation des premiers corps de notre sytème solaire, c’est bien par accrétion et donc via des collisions que les corps les plus volumineux se sont formés. Les études sur l’interprétation d es anomalies magnétiques crustales des corps solides du système solaire ont intégré cette contrainte qu’est l’effet des choc s sur le magnétisme : pour Mars (e.g., Hood et al., 2003 ; Artemieva et al., 2005 ), pour la Lune (e.g., Collinson, 1993 ; Halekas et al., 2003 ; Hood and Artemieva, 2008 ), pour les astéroïdes ( Chen et al., 1995 ), ou encore pour les cratères d'impact terrestres ( Pilkington and Grieve, 1992 ; Louzada et al., 2008 ). Dans les chondrites ordinaires par exemple, les effets de choc sont visibles sous forme de déformations de la structure cristalline des silicates, par la présence de veines de chocs et de poches fondues ( Stöeffler et al., 1991 ), ou encore à travers la déformation macroscopique par compaction (Gattacceca et al., 2005 ). Ainsi, la plupart des météorites ont été choquées à des pressions avoisinants 5-10 GPa. Par exemple 90% des chondrites ordinaires ont été choquées au- delà de 5 GPa ( Schulze and Stoffler, 1997 ), tandis que 65% des chondrites carbonées n'ont pas subi de pressions au-delà de 5 GPa ( Scott et al., 1992 ). L ’ensemble d es météorites martiennes ont subi des pressions de choc d’ au moins 10 GPa et fréquemment de 30-40 GPa (Nyquist et al., 2001 ). L’enregistrement magnétique et les propriétés magnétiques intrinsèques des météorites peuvent-être modifiés par les ondes de choc d’au moins trois façons. Premièrement, les chocs peuvent perturber l’aimantation rémanent e et modifier de façon permanente les propriétés magnétiques intrinsèques de roches, comme l’ aimantation rémanente à saturation, la coercivité, la susceptibilité ainsi que l'anisotropie de la susceptibilité (Gattacceca et al., 2005 ; Gattacceca et al., 2008 ; Louzada et al., 2007 ) et de rémanence. Deuxièmement en présence d'un champ ambiant, les chocs peu vent conduire à l'acquisition d’une SRM ( Gattacceca et al., 2008 ). Enfin, lorsque les chocs se produisent en prés ence d’un champ magnétique nul o u faible, ils peuvent effacer

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

partiellement ou complètement l'aimantation rémanente pré-choc ( Gattacceca et al., 2010a). La sensibilité des météorites à la ré-aimantation ou à la désaimantation sous pression dépend de la minéralogie magnétique (e.g., Bezaeva et al., 2010 ; Louzada et al., 2010 ). La pyrrhotite par exemple, qui est un minéral important pour la suite de ce mémoire puisqu’il fait partit des porteurs magnétiques des chondrites CM et des rumurutites, possède une transition à 2.8 GPa (Rochette et al., 2003b ), ce qui implique qu ’un choc au-delà de cette pression effacera une éventuelle aimantation rémanente pré-choc ( Rochette et al., 2001 ; Louzada et al., 2007 ). Globalement, on retiendra des études précédentes que : - la rémanence magnétique des météorites choquées est souvent un mauvais indicateur pour estimer le dernier impact majeur ca r elle dépend de l’intensité de celui-ci. - le champ ambiant au moment de l’ impact peut être transitoire ( Crawford and Schultz, 1999 ; Srnka, 1977 ; Doell et al., 1970 ) ou amplifié (Hide, 1972 ; Hood, 1987 ; Hood and Artemieva, 2008 ) par l’impact lui -même - les indices pétrographiques de choc ne permettent de distinguer que les événements de choc ayant des pressions au-delà de 4-5 GPa (e.g., Stoeffler et al., 1991 ; Scott et al., 1992 ). Les chocs en dessous de 5 GPa ne laissent aucun indice pétrographique mais peuvent quand même affecter sensiblement l’aimantation rémanente . Malgré ces complications, les météorites choquées ne doivent pas être immédiatement rejetées pour des études paléomagnétiques parce que les aimantations de choc et les processus de désaimantation liés aux chocs sont de mieux en mieux compris. Dans nos travaux sur les météorites, nous avons pris cet aspect en compte en réalisant des expériences d’acquisition d’ai mantation de choc en laboratoire, et en comparant les propriétés et les intensités des ces aimantations avec l’aimantation naturelle. Ce travail a par exemple é té réalisé sur toutes les chondrites CM que nous avons étudiées ( Fig. 2.1) (Chapitre 4 ) et fait l’objet d’une publication plus spécifique ( Tikoo et al., en préparation ).

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

Fig 2.1. Evolution de la PRM acquise en fonction de la pression (et normalisée par le champ ambiant) pour la météorite Nogoya (CM2). Cette météorite ne montre pas d’indices pétrographiques de choc ce qui indique qu’elle n’a jamais été choquée au-delà de 5 GPa. L’e xtrapolation des données de PRM jusqu’ à 5 GPa nous donne donc la valeur maximale d’aimantation de choc que peut porter cette météorite . Les points noirs représentent nos données et les points gris les valeurs extrapolées.

Tout comme pour les roches terrestres, la (p)TRM est un type d’aimantation qui peut-être rencontré dans les météorites. Les TRM ou pTRM peuvent s’acquérir à travers de nombreux phénomènes : mise en place d’une roche volcanique (sur la Lune, Mars, ou pour les achondrites), métamorphisme (pour certains groupes de chondrites). Les chocs peuvent aussi conduire à l’acquisition d’une pTRM ou TRM . En effet, les chocs induisent une augmentation de température qui atteint ~100°C à 20 GPa, ~200°C à 30 GPa et ~500°C à 45 GPa (Stöeffler et al., 1991 ; Nyquist et al., 2001 ). L’aimantation de la météorite martienne ALH 84001 a par exemple été attribuée à ce type de phénomène (Weiss et al., 2008b). De la même façon, l’aimantation portée par la météorite Tissint aurait été acquise lors du refroidissement post impact (25-45 GPa) dans un champ magnétique ambiant stable (Gattacceca et al., 2013a-Annexe-1).

Dans le cas des météorites, une CRM due à la cristallisation de nouveaux minéraux magnétiques ou à des changements de phase sous champ peut avoir lieu lors de l ’altération hydrothermale ou le métamorphisme thermique. Cette hypothèse a été évoquée par le passé (Stacey, 1976 ) et reprise récemment pour expliquer l’aimantation rémanente naturelle des chondrites carbonées de type CM ( Chapitre 4 ). Tout comme pour les sédiments terrestres, l’existence d’une DRM dans les météorites pourrait être envisagée. Cependant une nouvelle forme d’ aimantation rémanente propre à la matière extraterrestre a été introduite , il s’agit de l’aimantation rémanente détritique d’accrétion , ADRM (Fu and Weiss, 2012 ). Cette dernière correspondrait à l’alignement des solides

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

ferromagnétiques sur les corps primitifs du système solaire dans les champs magnétiques environnants et localement uniformes de la nébuleuse. L’identification et la caractérisation d’une telle aimantation permettrait de contraindre la force et la géométrie des champs magnétiques dans le système solaire jeune et d’obtenir des informations sur l’accrétion des objets de petite taille (~métrique), sur les régions de formation des corps parents des chondrites et sur l’histoire de l’altération des composants chondritiques.

Tout comme pour les roches terrestres, l’a imantation rémanente visqueuse (VRM) doit-être prise en compte. Elle peut être acquise, lorsqu’une météorite est exposée de façon prolongée à un champ magnétique faible (par exemple dans le champ terrestre). La VRM peut ainsi se surimposer partiellement à l’aimantation préalablement enregistrée dans la météorite. La croissance de la VRM étant proportionnelle au logarithme du temps, elle affectera plus spécifiquement les trouvailles qui peuvent avoir séjourné dans le champ terrestre pendant plusieurs millier s ou dizaines de milliers d’années. La stabilité d’une VRM peut être évaluée grâce à des diagrammes temps/température établis de manière théorique pour des grains monodomaines : magnétite ( Pullaiha et al., 1975), pyrrhotite (Dunlop et al., 2000 ), kamacite (Garrick-Bethell and Weiss, 2010 ), taénite ( Sato and Nakamura, 2010 ). Cette notion est plus largement discutée pour les chondrites carbonées dans le Chapitre 4 . La même théorie permet d’ estimer la stabilité des aimantations primaires des météorites lorsque celles-ci restent pendant des périodes prolongés dans un champ magnétique nul comme le champ interplanétaire (~ nT). Ainsi, une roche aimantée il y a 4.5 Ga et maintenue en champ quasi nul sur son corps parent à une température de ~-110°C dans la ceinture d’ astéroïdes (e.g., Spencer et al., 1989 ), verra l’aimantation de tous ses grains de pyrrhotite et magnétite avec des températures de déblocage de ~50°C effacée. De plus, le temps de transfert du météoroïde depuis sont astéroïde parent jusqu’à la Terre se passe principalement à proximité de la ceinture d’astéroïde s (Gladman et al., 1997 ). Par conséquent, le temps passé à proximité de l’orbite terrestre (et donc à température plus élevée) est suffisamment court pour ne pas induire de décroissance visqueuse supplémentaire de la NRM originelle. La stabilité des aimantations des météorites sur les longues échelles de temps (plusieurs Ga) semble donc assurée pour les météorites qui n’ont pas subi d’évènements tardifs significatifs (chocs, événement thermique,...) .

Lors des études en laboratoire, pour évaluer la proportion de VRM on réalise des tests d’acquisition de VRM . On considère généralement que le taux d’acquisition et de décroissance de

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

la VRM sont approximativement identiques ( Enkin and Dunlop, 1988 ) mais il est plus sûr de mesurer les deux paramètres. Il est alors possible d’évaluer l’intensité maximale attendu pour une VRM acquise dans le champ terrestre depuis la chute de la météorite. Les VRM acquises pendant des périodes de quelques années à quelques dizaines de milliers d’années peuvent en général être facilement effacées par des désaimantations thermiques à basse température (~80-120°C que l’aimantation soit portée par de la pyrrhotite ( Pullaiha et al., 1975 ) ou de la magnétite (Dunlop et al., 2000 ) ou par des champs alternatifs relativement faibles. L’aimantation gyro-rémanente (GRM) est une aimantation parasite qui peut être acquise lors des désaimantations par champs alternatifs en laboratoire. La rémanence peut augmenter de façon systématique dans les derniers stades de désaimantation (sous champ forts) ( Danker and Zijderveld, 1981 ). Ce phénomène de GRM a fait l’objet d’études qui ont permis de définir des méthodes de correction que nous avons utilisées dans nos travaux (Stephenson, 1993 ; Hu et al., 1998 ) (e.g., SP Fig. 3.B Cournède et al., 2012 , Chapitre 3 ).

2.3.2.4 La contamination terrestre : Pendant et après leur arrivée sur Terre, les météorites peuvent être soumises à plusieurs processus (traversée atmosphérique, altération, le champ magnétique de la Terre, la contamination humaine via l’application d’aimants ) qui peuvent affecter l’aimantation rémanente pré-terrestre qu’elles contiennent . Tous ces éléments, doivent être gardés à l'esprit avant d’aborder l 'interprétation des données paléomagnétiques car ils peuvent compter pour une partie de l'aimantation enregistrée dans les météorites.

-La traversée atmosphérique Il existe des effets liés à l’entrée atmosphérique des météorites. En effet, l ors du passage dans l’atmosphère terrestre, on considère que seul les premiers millimètres des météorites rocheuses/pierreuses sont affectés. Ceci se traduit par la formation d’une mince croûte de fusion en surface (<1mm) et par l’acquisition d’une TRM unidirectionnelle dans le champ terrestre des premiers millimètres de la météorite ( Nagata and Sugiura, 1977 ; Nagata, 1979b ; Weiss et al., 2000, 2002, 2008b ). Plus en profondeur, la météorite n'est quasiment pas chauffée (Sears, 1975 ). On retiendra donc que les effets liés à la traversée atmosphérique sont négligeables pour des échantillons situés à plus de quelques millimètres de la croûte de fusion.

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

-L’altération terrestre : La majeure partie des météorites disponibles pour réaliser des études sont les « trouvailles ». Selon leur zone d’origine, l e temps de résidence terrestre des trouvailles varie entre plusieurs dizaines de milliers d'années (pour les déserts chauds) et plusieurs centaines de milliers d'années (pour l’ Antarctique). L’altération terrestre va conduire à la cristallisation de nouveau x minéraux ferromagnétiques à partir des minéraux ferromagnétiques primaires ce qui peut modifier l’aimantation rémanente pré -terrestre contenue dans les météorites (e.g. Uehara et al., 2012 ). Ce phénomène concerne principalement les phases métalliques et dans une moindre mesure les sulfures. Les minéraux néoformés sont principalement des oxydes et oxyhydroxydes de nature variée (magnétite, maghémite, hématite, akaganéite, lépidocrocite, ferrihydrite, e.g. dans Bland et al., 2006 ; Uehara et al., 2012 ). Ai nsi il est préférable d’éviter les trouvailles pour les études paléomagnétiques , surtout lorsque les porteurs de l’aimantation sont métalliques. Cependant ceci n’est pas toujours possible, certaines classes de météorites n ’étant représentées quasiment que par des trouvailles (comme les rumurutites avec une chute contre une centaine de trouvailles).

-Contamination humaine Que se soit lors des diverses expériences en laboratoire ou lors des campagnes de recherche, la manipulation humaine des échantillons peut définitivement contaminer l’aimantation pré- terrestre contenue dans les météorites. La principale source de contamination (et donc d’inquiétude des paléomagnéticiens !) vient du fait que bien souvent un aimant est appliqué directement sur les échantillons l’objectif étant d’établir la nature extraterrestre e n estimant rapidement la teneur en fer métallique. Typiquement, les aimants classiques (ferrite) produisent des champs de surface de quelques dizaines de mT, tandis que les aimants aux terres rares peuvent produire des champs de surface de plusieurs centaines de mT. Lors de ce genre de test les aimants aux terres rares vont détruire entièrement la NRM originelle au point de contact dès l'application de l'aimant. En plus de détruire le signal magnétique pré-terrestre enregistré dans les météorites depuis des milliards d’années , ce genre de test n'est pas fiable pour affirmer qu'une roche est une météorite. La meilleure méthode permettant de ne pas détruire l’aimantation originelle des échantillons et d’en estimer rapidement le contenu en minéraux magnétiques, est la mesure de la susceptibilité magnétique (Rochette et al., 2008 ; Folco et al., 2006 ).

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

Afin d’éviter au maximum les problèmes liés à l’entrée atmosphérique, l’altération terrestre ou à la contamination, il est préférable de travailler sur des échantillons issus au maximum de l’ intérieur de la météorite (quelques centimètres pour s’affranchir de la contamination liée à un aimant et quelques millimètres lors de la présence d’une croûte de fusion ) lorsque cela est possible. Malgré toutes les difficultés que comporte l’étude magnétique de la matière extraterrestre, les nouvelles techniques et concepts du paléomagnétisme ainsi que les nouveaux instruments (plus sensibles et hautement automatisés) permettent une meilleure mesure et compréhension des propriétés magnétiques des météorites, des effets de chocs, de l’altération et des autres processus secondaires, permettant ainsi aux composantes de l'aimantation primaire et secondaire d’être distinguées et interprétées avec une confiance accrue.

2.3.3 Méthodes d’estimation de la paléointensité

2.3.3.1 Méthode Thellier-Thellier La méthode la plus utilisée en laboratoire pour estimer la paléointensité dans les roches terrestres est la méthode Thellier –Thellier (Thellier and Thellier, 1959 ). Cette méthode repose sur la loi d'additivité qui stipule que la somme des pTRM acquises dans le même champ entre des températures Tl et T2 et entre T2 et T3 est égale à la pTRM acquise entre TI et T3. D'autre part, un réchauffement à la température T suivi d'un refroidissement en champ nul ne détruit que les pTRM acquises à des températures inférieures à T. La méthode consiste à effectuer des chauffes successives en champs nul puis dans un champ connu à des températures de plus en plus élevées. Ceci permet d’établir une courbe évaluant pour chaque étape la perte de NRM comparée à l’acquisition de pTRM, ce qui permet de calculer la paléointensité. Cette méthode a fait ces preuves et est fiable dans le cas où la minéralogie magnétique est stable lors des chauffes. Cependant dans le cas des météorites on ne peut généralement pas l’appliquer, d’une part la minéralogie n’est pas toujours stable et d’autre part les conditions de préservation et l’altération aqueuse font que l’aim antation enregistrée n’est pas toujours une TRM .

2.3.3.2 Les méthodes utilisant les champs alternatifs Etant donné que les méthodes de type Thellier-Thellier ne sont pas souvent applicables aux roches extraterrestres, un certain nombre de méthodes basées sur la désaimantation par champ alternatif ont été développées (Gattacceca and Rochette, 2004 ; Stephenson and Collinson, 1974 ; Kohout et al., 2008 ; Muxworthy and Heslop, 2011 ).

CHAPITRE 2 : SPECIFICITES DU MAGNETISME EXTRATERRESTRE

Nous avons en général utilisé la méthode REM’ ( Gattacceca and Rochette, 2004 ) qui est une méthode de normalisation par l’aimantation rémanente isotherme à saturation (sIRM), aimantation artificielle donnée en l aboratoire dans un champ fort. Il s’agit d’abord de définir l’intervalle sur lequel la NRM a une composante stable (e.g. intervalle A-B mT), de s’assurer que le REM’ est à peu près stable sur cet intervalle et ensuite il suffit de calculer

(NRM B-NRM A)/(IRM B-IRM A)*3000 pour obtenir une paléointensité approximative.

Comparée à la méthode Thellier-Thellier, la méthode du REM’ a l’avantage d’être non - destructive. En effet dans le cas de la méthode Thellier-Thellier, les chauffes peuvent conduire à la formation de nouveaux minéraux magnétiques au sein de la roche modifiant ainsi l’aimantation . Ces deux méthodes ne sont valables que dans le cas où la NRM est une TRM. Dans le cas où la nature de la NRM est différente, ce qui est souvent le cas pour les météorites, ces méthodes ne sont donc pas strictement valables mais permettent cependant d’avoir un ordre de grandeur de l’intensité des cham ps. Ainsi, m ême si l’estimation n’est pas précise cela peut suffire pour interpréter la nature du champ à l’ origine de l’aimantation.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Chapitre 3: Paléomagnétisme Lunaire

3.1 Introduction

Comme nous l’avons vu dans les chapitres précédents les échantillons lunaires ont fait l’objet de nombreuses études magnétiques. Parmi elles, on compte des mesures paléomagnétiques réalisées sur les échantillons Apollo (e.g., Fuller, 1974, 2007 ; Hood and Cisowski, 1983 ; Dunlop and Ozdemir, 1997 ; Garrick-Bethell et al., 2009 ; Shea et al., 2012 ; Suavet et al., 2013 ; Fuller and Cisowski, 1987 , pour une synthèse). Des mesures de champ magnétique ont également été réalisées en surface ( Dyal et al., 1970 ; Dyal and Daily 1978 ), et en périphérie de la Lune (en orbite basse) (e.g., Coleman et al., 1973 ; Richmond and Hood, 2008 ; Mitchell et al., 2008 ). Ces études ont montré que beaucoup de matériaux de la croûte lunaire portent une aimantation rémanente significative. Cependant l'origine de cette aimantation - et en particulier l'origine du champ magnétique qui a permis l'acquisition de cette aimantation – a été une question très débattue. Les réponses à ces questions ont des implications fortes pour la compréhension de l'évolution thermique interne de la Lune et de manière plus générale sur le fonctionnement des dynamos planétaires et astéroïdales. Ainsi, dès les années 1970, deux théories ont été proposées pour expliquer ces aimantations rémanentes (e.g. Collinson, 1993 ) : aimantation par refroidissement de roches magmatiques sous leur température de Curie (TRM) ou aimantation par choc (SRM). Une TRM nécessite un champ magnétique stable pendant la durée de refroidissement (> plusieurs jours), ce qui implique une origine interne, c'est-à-dire une dynamo. Une aimantation de choc (SRM) n'implique pas forcément l'existence d'une dynamo, un champ transitoire lors de la décompression (soit quelques secondes au plus) étant suffisant. Hood et Artemieva (2008) ont ainsi proposé que l'augmentation transitoire du champ magnétique ambiant par un impact couplée à l'occurrence simultanée d'ondes de choc puisse expliquer les anomalies magnétiques antipodales aux grands bassins d'impacts. Récemment des preuves expérimentales ont montré que ce mécanisme d'aimantation par choc pouvait effectivement expliquer l’aimantation portée par certaines roches lunaires ( Gattacceca et al., 2010b). Bien que la Lune ne possède pas de dynamo active actuellement, une variété de données géophysiques et géochimiques indiquent que la Lune est entièrement différenciée et l’existence d’un noyau métallique semble prouvée. En effet, les données sur le moment d'inertie (Konopliv

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

et al., 1998 ) et le moment dipolaire magnétique lunaire induit (Hood et al., 1999 ) soutiennent l'existence d'un noyau de fer de 220-450 km de rayon, alors que les données sismiques Apollo permettent de distinguer l’existence d’ un noyau interne solide de 240 km rayon ( Weber et al., 2011 ), entouré par noyau externe liquide de 330-365 km de rayon (Garcia et al., 2010 ; Weber et al., 2011 ). En outre, des études paléomagnétiques ont montré que la Lune a possédé un champ magnétique de dynamo (Garrick-Bethell et al., 2009 ; Shea et al., 2012 ; Suavet et al., 2013 ) et, par voie de conséquence un noyau métallique liquide animé de mouvements de convection. Ainsi, plutôt que d'autres mécanismes générateurs de champ , c’est cette dernière hypothèse qui est la plus largement admise actuellement pour expliquer l’aimantation rémanente de la plupart des roches lunaires. De nombreuses questions persistent concernant la durée d’activité de cette dynamo. Pour certains, la dynamo se serait arrêtée avant l'éruption des basaltes High-K (avant 3.6 Ga) ( Fuller, 1998 ) tandis que pour d’autres la dynamo se serait arrêtée lentement mais aurait persisté au moins jusqu'à ~3.2 Ga ( Runcorn, 1996 ). Les données paléomagnétiques récentes montre que la dynamo lunaire semble avoir fonctionné au moins entre 4.2 Ga et ~3.6 Ga, avec des champs de surface de ~60 µT ( Garrick-Bethell et al., 2009 ; Shea et al., 2012 ; Suavet et al., 2013 ). Au-delà de 3.6 Ga, l’histoire de la dynamo lunaire reste à écrire. Il devient de plus en plus clair que la durée de vie du champ lunaire est incompatible avec une dynamo entraînée uniquement par convection thermique. Dans les modèles de flux de chaleur, le seuil minimum nécessaire pour maintenir la convection thermique du noyau établi qu’une dynamo lunaire peut avoir persisté au maximum jusqu'à 4.1 Ga ( Stegman et al., 2003 ) soit bien avant l’arrêt estimé de la dynamo. De plus, le petit rayon supposé du noyau lunaire (<460 km, e.g. Wieczorek et al., 2006 ) comparé au rayon lunaire (1738 km) demanderait une dynamo inhabituellement énergétique pour produire un champ magnétique de surface comparable au champ de dynamo terrestre (e.g. Stevenson, 1983 ). Il en ressort qu’ un champ lunaire prolongé et de forte intensité peut exiger des sources d'énergie non traditionnelles (Dwyer et al., 2011 ; Le Bars et al., 2011 ; Wieczorek et al., 2006, 2012 ). Parmi elles, on trouve les effets mécaniques de la précession ( Dwyer et al., 2011 ) ou la formation de bassin d’impacts ( Le Bars et al., 2011 ) pour alimenter la dynamo mécaniquement par un mouvement différentiel entre le noyau liquide et le manteau rocheux. La précession semble être capable d'alimenter une dynamo jusqu'à ~ 1.8-2.7 Ga ( Dwyer et al., 2011 ). Par comparaison, une dynamo causée par un impact pourrait probablement être active uniquement lorsque la formation des bassins impacts a eu lieu, avant ou pendant l’époque Imbrienne précoce ( ≥ ~ 3.72 Ga). L’existence d’u ne dynamo de convection compositionnelle (convection thermochimique) entraînée par la cristallisation du noyau est

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

possible, mais la durée de vie d'une telle dynamo est actuellement incertaine (e.g., Stegman et al., 2003 ). Toutefois, il est à noter que tous les modèles mécaniques actuels de dynamo lunaires permettent de générer des champs de surface allant de 0.2 à 15 µT au maximum (Dwyer et al., 2011 ; Le Bars et al., 2011 ). Les paléointensités élevées définies pour les échantillons lunaires constituent donc un défi à la théorie de la dynamo.

3.2 Etude de gros échantillons Apollo

Notre étude vient compléter et préciser le tableau dressé dans la partie précédente. Ainsi à travers de nouvelles techniques, des instruments plus performants et de nouveaux concepts, nous abordons la question du paléomagnétisme lunaire et du fonctionnement d’une dynamo passée. Pour cette étude nous avons utilisé 17 échantillons Apollo qui se distinguent des études précédentes par leur taille importante (jusqu’à 8 gr). En effet, le caractère non destructif des études magnétiques (tant que l’on ne fait pas de désaimantation thermique) a permis de rassembler ces échantillons qui se trouvaient en prêt dans différents laboratoires français (CRPG, Nancy ; ENS, Lyon ; IPGP, Paris et IAS, Orsay ), dans l’attente d’analyse s destructives de géochimie. La taille importante des échantillons permet de s’affranchir de certains problèmes d’anisotropie qui p euvent être responsables d’un comportement instable lors de la désaimantation (Tikoo et al., 2012 ). Nous avons vu précédemment (Chapitre 2 ) que les échantillons Apollo sont les seuls matériaux extraterrestres pour lesquels les sites d'échantillonnages sont bien connus. Cependant , s’agissant de blocs prélevés en surface, l'orientation originelle des échantillons reste inconnue. Or la connaissance de cette orientation, et en particulier de la paléohorizontale, couplée aux directions paléomagnétiques permettrait de contraindre un aspect de la dynamo lunaire : sa géométrie. Nous avons utilisé l’anisotropie de susceptibilité magnétique (ASM) des roches lunaires, et en particulier leur plan de foliation magnétique, comme indicateur de paléohorizontale. Cette approche est rendue possible par la taille importante des échantillons étudiés. Les directions paléomagnétiques mesurées sur les mêmes échantillons ont permis de déterminer la paléoinclinaison du champ magnétique pour 7 échantillons Apollo. Les mesures des propriétés magnétiques de ces échantillons confirment que les propriétés magnétiques des roches lunaires sont dominées par environ 0.1 % de grains d ’alliage FeNi métallique à l’état multidomaine. L’assemblage des minéraux ferromagnétiques est, au premier ordre le même dans toutes les lithologies lunaires, avec une plus grande concentration dans les basaltes que dans les anorthosites.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Parmi les 17 échantillons étudiés , onze montrent au moins une composante d’aimantation rémanente naturelle stable. Parmi ceux-ci, nous avons déterminé que cinq échantillons avaient acquis leur aimantation sur la Lune et pouvaient, nous renseigner sur le paléochamp lunaire. Pour ces échantillons (ainsi que deux échantillons similaires issus d’une étude précédente), des paléointensités et des paléoinclinaisons ont été estimées. Connaissa nt les sites d’échantillonnage sur la Lune, ces paléoinclinaisons peuvent être expliquées par un champ dipolaire avec un paléopôle magnétique situé à ~ 75° N (Fig. 3.1 ). Ainsi nos résultats suggèrent l’existence sur la Lune au moins entre 3.8 et 3.3 Ga (la gamme d’âge des sept échantillons concernés) , d’un champ dipolaire dont l’axe est proche de l’axe de rotation actuel de la Lune. L’estimation de l’intensité du champ de surface est de plusieurs dizaines de µT, ce qui confirme les résultats précédents qui montrent que la Lune a eu un champ de dynamo fort dans le passé.

Fig. 3.1. Projection stéréographique des possibles paléopôles pour chaque site Apollo étudié. Les sites Apollo sont également représentés. Figure tirée de l’article Cournède et al., 2012 . Les lignes continues et les cercles noirs représentent les petits cercles et les sites dans l’hémisphère Nord et les lignes pointillés et les cercles blancs sont les petits cercles et les sites dans l’hémisphère Sud. Pour les sites Apollo 17, le petit cercle est représenté à partir de la paléolatitude moyenne (± l’écart -type est la zone grise) obtenu pour cinq échantillons. Les zones d’intersections des petits cercles sont représentées en gris foncé. L’étoile noire est la meilleure intersection des petits cercles. Le triangle blanc est la seconde intersection possible des petits cercles. Le triangle noir est la projection de ce dernier dans l’hémisphère Nord. L’étoile grise est le paléopôle défini par Hood (2011).

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Earth and Planetary Science Letters 331-332 (2012) 31-42

Magnetic study of large Apollo samples: Possible evidence for an ancient centered dipolar field on the Moon

Cécile Cournède 1, Jérôme Gattacceca 1, Pierre Rochette 1 1CEREGE, CNRS/ Université d’Aix -Marseille, BP80 13545, Aix en Provence, Cedex 4, France

Abstract

We present new rock magnetic and paleomagnetic results obtained from seventeen Apollo samples. Measurements of the magnetic properties of the samples confirm that the magnetic properties of lunar rocks are dominated by about 0.1 wt.% of multidomain FeNi grains. The ferromagnetic mineral assemblage is to the first order the same in all lunar lithologies, with higher concentration in basalts than in anorthosites (some of which are diamagnetic). Impact processing leads to an increase of the ferromagnetic content. Out of the seventeen samples, eleven show at least one stable component of natural remanent magnetization. In some cases, this magnetization may be a magnetic contamination. However, five basalt samples have a component of magnetization that may have recorded a lunar paleofield acquired on the Moon. For these samples (plus two similar samples from a previous study), using the rock magnetic fabric as a paleohorizontal proxy, paleoinclinations can be estimated from the paleomagnetic data. These paleoinclinations are best explained with a dipolar field and a magnetic paleopole located at 75 N. Therefore our results suggest the existence on the Moon, at least between 3.8 and 3.3 Ga, of a dipolar field whose axis is close to the present-day rotation axis of the Moon. The estimated surface field intensities of several tens of μT support previous result that the Moon had a dynamo field in the past, and the proposed paleofield geometry suggests that this dynamo was centered on the lunar present-day rotation axis.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

1. Introduction

Today the existence of remanent crustal magnetization on the Moon is undisputed (e.g., Halekas et al., 2001; Hood et al., 2001 ). However many questions remain about lunar magnetism: how and when lunar rocks were magnetized, and what is the origin of the magnetizing field? This latter question, with the possibility of an internal field generated by a dynamo in a convective core, is of crucial interest in discussions about the inner structure of the Moon and its thermal history. Indeed, the long debated existence of a lunar liquid core has been recently confirmed (Weber et al., 2011 ). Thanks to Apollo 11 to 17 missions, 370 kg of lunar samples have been returned from the Moon and are a primary source of information about lunar past magnetic fields. These samples have been the focus of numerous paleomagnetic studies (see Fuller and Cisowski (1987) , and Collinson (1993) for a review). It has been shown that many lunar samples carry a significant remanent magnetization, but there is still considerable debate about the interpretation of the results. Many studies proposed the existence of an ancient lunar dynamo, possibly only active in the 3.6 –3.9 Ga interval, with surface fields in the 10 –100 μT range (e.g., Fuller and Cisowski, 1987 ). According to Stephenson et al. (1975) the field intensity decreased from 130 μT to 5 μT between 3.9 and 2.2 Ga suggesting a shutdown of the putative dynamo or solidification of the iron core ( Runcorn et al., 1970 ). However, it has been advocated that most paleointensity determinations obtained through Thellier –Thellier experiments may not be reliable ( Lawrence et al., 2008 ). Only recently robust evidences for a dynamo-generated surface field of at least 1 μT at 4.2 Ga have been given ( Garrick-Bethell et al., 2009 ). This dynamo existed at least until 3.6 Ga (Shea et al., 2012 ). Conversely, Banerjee (1972) proposed that remanent magnetization can originate from thermal cycling at the Moon surface. Cisowski et al. (1974) showed the importance of shock effect on magnetization, which may be at the origin of the measured remanent magnetization. Magnetization in a compressed magnetic field at the antipodes of large impact basins was proposed by Hood and Huang (1991) , modeled by Hood and Artemieva (2008) and experimentally validated by Gattacceca et al. (2010a) . Another theory calls for an intense early solar system field, giving rise to a uniform magnetization of an initially cold moon ( Strangway, 1977 ). Therefore a definitive explanation for the origin of the magnetization and the source of the magnetizing field remains an uncertain but crucial key to our understanding of the evolution of the Moon. Unfortunately, the paleomagnetic study of lunar rocks has many caveats. Most lunar rocks contain very little ferromagnetic grains, the main magnetic carrier is generally multidomain

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

kamacite (a poor magnetic recorder), magnetic contamination during the return of samples to Earth or during storage is a real problem, and most samples are heavily shocked. Moreover sample orientation is unknown which led to focus on paleointensity rather than field geometry. In this study we provide new rock magnetic and paleomagnetic analyses from seventeen Apollo samples (Mare basalts, anorthosites, norite, breccias), among which five have never been studied before. This is a significant increase of the lunar paleomagnetic database, all the more that we used more sensitive apparatus than 40 yr ago when most of the paleomagnetic studies on Apollo rocks were conducted. Measurements of the magnetic properties of the studied samples allowed us to determine the nature and concentration of the magnetic minerals. Measurement of their natural remanent magnetization (NRM) allowed us to investigate the possible existence and origin nature of ancient lunar magnetic fields. Coupling the paleomagnetic and rock magnetic results, we also tentatively determined the paleoinclination of the lunar field, and its geometry.

2. Samples and methods

In order to perform this magnetic and paleomagnetic study, 17 Apollo samples have been temporarily lent from different French laboratories (CRPG (Nancy), IPG (Paris), IAS (Paris) and ENSL (Lyon), detailed in Supplemental Table 3.A), in agreement with NASA. Our aim was to save the magnetic information of these samples before they are analyzed by destructive methods in these laboratories.

2.1. Samples Lunar samples consist of three main groups of material: igneous rocks, breccias and soils. We studied 15 igneous rock and 2 breccia samples (the trocotolitic impact melt breccia 62295, and the anorthositic glassy breccia 68815). Igneous rocks are divided in two sub-groups, volcanic (with 11 mare basalts samples), and plutonic (with 1 norite and 3 anorthosites). Out of these 17 samples (listed in Table 3.1), five have never been studied for magnetism (15475, 71505, 71567, 65315 and 60215). The studied samples have masses ranging from 0.24 to 8.05 g, with a median mass of 1.08 g. Such masses are well above usual ones in Apollo paleomagnetic studies. Sample radiometric ages, determined in previous studies, are between 3.3 and 4.4 Ga ( Supplemental Table 3.A). According to their laboratory of origin, samples were conditioned in different ways. Samples obtained from ENS (Lyon) were provided in non-demagnetized sealed plastic containers. However, we could measure the magnetization of an empty plastic container of the

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

, 3 m 9

− 10

Rhodes et : coercivity cr , j= ref. FeO (%) Laul and Schmitt (1973) 2+ : coercivity, B c Fe (%) , b= χ χ / : anisotropy degree and shape p Rose et al. (1975) (%) χ 4 8 4.8 e rem , i= 2+ 4 59 20 20.2 d , T /Fe 0 1.83 9 6 6.5 k 0.61 32 18 18.0 i rem Fe (%) rem - 0.44 37 19 21.7 a - 0.65 24 23 18.7 e - 4.72 8 6 5.3 l T Willis et al. (1971) 3 -1 0.57 27 29 18.5 f ptibility. P : a= .14 0.33 0.24 45 17 21.8 b .18 -0.07 0.36 50 21 19.2 g rem 1.04 -0.01 0.26 55 22 21.8 b 1.17 0.57 0.50- 31 22 - 19.4 h 0.35 41 16 17.8 j P Rhodes et al. (1976) : saturation magnetization, B s ariation ariation corrected for paramagnetic susceptibility, P, T: ams , h= T f P f: ferromagnetic susceptibility. All susceptibilities are in χ ams P f Δ χ (%) f: susceptibility v χ

Δ Warner et al. (1979) Δ χ (%) f : saturation remanence, M , g= rs log χ log p χ - - 22.8 - - - - 1.09 0.31 ------36.8 - - - - 1.04 0.52 - - - - (1973)

log and 976 Hz logχ δ δ p: paramagnetic susceptibility, . LSPET χ , f= c B (mT) ble are indicated by: -. References for FeO (%) content eptibility, Pf: anisotropy degree of ferromagnetic susce cr ic magnetic properties. M B (mT) ) -1 kg 2 McKinley et al. (1984) s s Duncan et al. (1974) =diamagnetic samples), M m (mA , l= δ ) , e= -1 kg 2 rs M (A m Nava (1974) mass mass (mg) , d= : magnetic susceptibility ( χ 78236 224 1.15E-03 236 25 1.0 3.18 2.11 3.14 0.3 0.3 1.85 1.97 0.38 - - 12002 364 7.13E-04 185 29 1.4 3.05 2.62 2.85 0.1 0.2 1.02 1.03 -0.79 - 15016a 683 4.23E-04 1257001770035 8045 1.73E-03 33 462 1.2 329 2.02E-03 2.94 364 2.6875055 2.59 3.0 7 3.2 374 1.01 1.1 1.03 5.71E-04 -0.71 8 119 3.31 1.0 2.69 3.19 3.38 0.6 2.80 0.7 3.24 1.03 1.04 52 3.1 0.04 1.7 3.2 1.08 - 1.12 2.92 0.11 2.53 1.0 2.69 0.3 0.6 1.02 1.03 -0.01 15016b15475 857 3.93E-0415597 955 88 4.67E-04 214 152 4.00E-0471505 5871567 25 1063 1.3 7.46E-0474275 17 1044 162 9.21E-04 2.92 - 77 591 234 2.57 2.0 1.50E-03 2.66 235 - 2.4 2.86 2.5 71 1.01 2.63 2.52 1.02 2.1 2.47 -0.92 - 11 - 1 1.0 2.96 61 3.4 - 2.65 4.7 3.19 2.6 1.03 2.66 - 1.07 2.67 2.5 -0.29 3.09 3.03 2.6 - - 1.01 2.6 2.59 1.02 2.7 2.92 0.05 1.06 1.09 - - 3.1 1 -0.73 4.7 1.01 - 1.02 0.1 0.14 - - - 0.46 - 15 19.8 c 62295 506 1.86E-03 600 101 1.4 3.20 2.10 3.16 0.1 0.1 1.07 1.08 0.27 - 68815 473 4.69E-03 947 38 1.4 3.66 2.23 3.65 - - 1.10 1.11 0.22 - - 5.59 6002560215 1268 1.29E-05 247 - 1.24E-05 - 32 - 38 - 1.51 - - 3.8 - - - - 1.06 -0.14 - - - - 65315 936 2.70E-06 - 36 - Sample Norite Basalt Breccias Anorthosites Blanchard and McKay (1981) : susceptibility variation between measurement at 15616 Hz , k= χ

Δ , , 1 Waenke et al. (1975) − of of remanence, kg anisotropy degree and shape parameter for magnetic susc parameter remanence.of Parameters that were not measura c= al. (1976) Table 3.1 : Samples studied in this work with their intrins

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

same type, which ranges from 40 μAm² before demagnetization to 20 μAm² after alternating field (AF) demagnetization at 120 mT. Other samples were sealed in oriented plastic containers provided by us. These containers were previously demagnetized by AF up to 150 mT and their residual magnetization was used for background subtraction during NRM measurements of the lunar samples. During all experiments, all samples remained sealed in their plastic containers to avoid contamination for future chemical analyses. This imposed limitations on the type of experiments that we were able to carry out. Moreover, subsampling that could provide information about the spatial homogeneity of the magnetization was not feasible. Also, not all measurements were feasible on each sample because the size of some plastic containers would not fit some of the instruments. As the original orientation of samples on the Moon is unknown, sample orientations were chosen arbitrary. Samples 15016a and 15016b that come from the same Apollo sample are not mutually oriented.

2.2. Magnetic measurements All magnetic measurements were performed at CEREGE (Aix-en- Provence, France). Hysteresis measurements were performed at room temperature with a Princeton Micromag Vibrating Sample Magnetometer (VSM), with a maximum field of 1 T, and a moment Provence, France). Hysteresis measurements were performed at room temperature with a Princeton Micromag Vibrating Sample Magnetometer (VSM), with a maximum field of 1 T, and a moment sensitivity of 10 −9 Am². The analysis of hysteresis loops provided the ratio of saturation remanent magnetization (M rs ) to saturation magnetization (M s) and the coercive force (B c). M rs values were measured with a 2 G cryogenic magnetometer (see below) after saturation in a field of 3 T using a

MMTD pulse magnetizer. M s values are recalculated from the M rs /M s ratio given by the VSM and the M rs value from the 2 G magnetometer. Remanent coercive force (B cr ) was determined by DC back field experiments performed either with the VSM, or with the 2 G magnetometer and the pulse magnetizer for weakly magnetized samples. The low field specific susceptibility (noted χ in m3 kg −1) and its anisotropy were measured using Agico apparatus, either MFK1 or KLY-2 (with sensitivity of 5×10 −13 m3), depending on sample size. KLY-2 operates at a frequency of 920 Hz and at 425 A m −1 peak field, while the MFK1 operates at 200 A m −1 peak field and at a frequency of 976 Hz. Anisotropy of magnetic susceptibility (AMS) was characterized by the shape parameter T ( Jelinek, 1981 ), varying from −1 (prolate) to +1 (oblate), and the anisotropy degree P (ratio of maximum to minimum

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

susceptibility). High field susceptibility ( χhf) was determined by linear fitting of the 0.9 –1 T field interval of the hysteresis loops. To evaluate a possible superparamagnetic contribution, we measured the low field susceptibility at three different frequencies with the MKF1 (976 Hz, 3904 Hz and 15 616 Hz). The anisotropy of anhysteretic remanent magnetization (AARM) was also measured using a three position scheme described in Gattacceca and Rochette (2002) , and was characterized, like the

AMS, by an anisotropy degree P rem and a shape parameter T rem . All remanence measurements were performed with a SQUID cryogenic magnetometer (2G Enterprises, model 755R, with noise level of 10 −11 Am²) with an attached automatic alternating field 3-axis degausser system (maximum peak field 170 mT) placed in a magnetically shielded room. Thermal demagnetization has proved to be problematic with lunar samples because the magnetic phases are irreversibly changed upon heating ( Brecher et al., 1973 ; Cisowski et al., 1983 ; Fuller, 1998 ). In our case, we were restricted to AF demagnetization which has the advantage to be fully non-destructive. The NRM of all samples and its stability against stepwise AF demagnetization up to 160 mT were measured. Before the measurements of NRM, all samples were stored in a shielded room in a double mu-metal box, in a field of ~10 nT for several weeks. At each AF step the sample was demagnetized and measured at least three times in order to reduce spurious anhysteretic effects and measurement noise. The AF demagnetization method is subject to errors caused by either imperfection of the alternating field or by the intrinsic properties of the rock. A gyromagnetic effect produced during static AF demagnetization can create a gyro-remanent magnetization in some anisotropic samples ( Stephenson, 1980 ). In order to avoid this GRM effect, we used the Zijderveld –Dunlop correction method ( Stephenson, 1993 ) on some samples for AF above 10 mT. After the study of NRM all samples were given an ARM (in a DC field of 400 μT, and AC field of 110 mT) that was subsequently stepwise AF demagnetized. Saturation isothermal remanence magnetization (SIRM, acquired in 3 T field with a pulse magnetizer) was also investigated for all samples. The same protocol was followed for the demagnetization of the NRM, the ARM, and the SIRM. For some samples, for instrumental reasons the size and shape of the sealed plastic container prevented AF demagnetization to fields higher than 100 or 120 mT. The rate of acquisition of viscous magnetization of the samples was studied through measurement of the VRM decay rate, since both rates are approximately the same ( Enkin and

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Dunlop, 1988 ). VRM was acquired in a field of 110 μT for about 15 days. The remanent magnetization of the samples was then measured versus time, samples being kept in low (~10 nT) magnetic field.

3. Rock magnetism

Results from rock magnetic measurements are summarized in Table 3.1.

3.1. Hysteresis properties The hysteresis loops of the three anorthositic samples could not be measured because of their weak magnetic signal. All hysteresis loops are very similar, and reveal the presence of metallic iron, evidenced by curvature up to 0.8 T ( Supplemental Fig. 3.A). Some differences between rock types or within a rock type are significant. In basalts, M s varies between 58 and 364 −1 −1 mA m² kg , with standard deviation (s.d.) of 91 mAm² kg (Table 3.2). Breccias with a mean M s of 770± 170 mAm² kg −1 are more magnetic than basalts and the norite that yield a mean of −1 190±87 mAm² kg . M rs follows the same trends as M s. M rs has a mean value of 3.28±1.42 mAm² kg −1 in breccias and a mean value of 0.92±0.54 mAm² kg −1 in basalts and the norite. The mean −3 −3 Mrs /M s ratio is 4.7×10 , and in all samples the ratio is less than 7×10 .

Bcr varies from 7 to 101 mT (with on average higher values in breccias, around 70 mT).

Bcr is extremely large compared with B c. The ratio B cr /B c is usually above 10 except for 70017 and

70035 (values of 6.8 and 7.9, respectively). Breccia 62295 has B cr /B c ratio of 72, which is remarkably higher than other samples. These overall high values of B cr /B c and low values of

Mrs /M s, typical of lunar rocks, indicate the dominance of multidomain grains ( Fig. 3.2 a).

3.2. Magnetic susceptibility For commodity magnetic susceptibility ( χ) is expressed in log units (log χ, with χ in 10 −9 m3 kg −1, Table 3.1). Anorthosites have the lowest values: log χ=1.51 for 60025, and χ=−1.4 and −5.4×10 −9 m3 kg −1 for samples 65315 and 60215 respectively. This latter sample is purely diamagnetic. Breccias and norite have the highest values with a mean susceptibility of 3.35. Basalts are intermediate with a mean of 3.06. In all samples but anorthosites, paramagnetic susceptibility (noted χp) can be approximated by high-field susceptibility ( χhf ) because the diamagnetic contribution is negligible (see similar discussion in Gattacceca et al., 2008 ). The results show log χp ranging from 2.10 to 2.80 (log χp,with χp in 10 −9 m3 kg −1). Paramagnetic minerals contribute significantly (from 24 to 59%) to the total susceptibility in basalts, and for less than 10% in norite and breccias. Ferromagnetic susceptibility ( χf= χ−χhf, expressed as log χf

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

with χf in 10 −9 m3 kg −1) varies from 2.47 to 3.65 with higher mean values in breccias than in igneous rocks (3.40 and 2.86, respectively). Our data show a good linear correlation between magnetic susceptibility and saturation remanence ( Fig. 3.2 b), in agreement with a previous compilation ( Rochette et al., 2010 ). The correlation is better than in the literature data, which may be explained by the fact that in the present study, both measurements were made on the same sample, and the same protocol was used for all samples. The measurements of the frequency dependence of low field susceptibility indicate that superparamagnetic contribution is weak and negligible for most samples. Some basalts exhibit variations above 3% between 976 Hz and 15,616 Hz, which implies a significant but still weak contribution of superparamagnetic grains. Magnetic measurements provide an estimate on the amounts of Fe0 and Fe 2+ (Wasilewski, 1974 ). Indeed, a metallic iron concentration can be computed from the saturation magnetization on the assumption that Ms is entirely due to metallic iron. Similarly χp is proportional to the quantity of ferrous iron (Fe 2+ ) present assuming no other paramagnetic ions contribute significantly, as is the case for most lunar samples. The inferred metallic iron concentration ranges from around 0.09 wt.% for the igneous rocks (without anorthosite) to 0.35 wt.% for the breccias ( Table 3.1). These results agree well with previous data ( Hargraves and Dorety, 1971; Nagata et al., 1970; Pearce et al., 1974 ) and particularly with Gose et al. (1972) who found a native iron concentration typically around 0.1 wt.% for the igneous rocks and 0.5 wt.% for the fragmental rocks and soils.

M M B B ARM NRM Number rs s.d. s s.d. cr s.d. c s.d. Logχ s.d. s.d. s.d. (A m 2 kg -1 ) (A m 2 kg -1 ) (mT) (mT) (A m 2 kg -1 ) (A m 2 kg -1 ) Basalts 11 8.89E-04 5.56E-04 1.86E-01 0.91E-01 36 24.4 1.5 0.52 3.06 0.17 6.22E-05 6.87E-05 2.51E-05 2.88E-05 Breccias 2 3.28E-03 1.42E-03 7.73E-01 1.74E-01 70 31.5 1.4 0 3.43 0.23 5.10E-05 1.26E-05 6.75E-05 5.62E-05 Norite 1 1.15E-03 - 2.36E-01 - 25 - 1.0 - 3.18 - 5.35E-05 - 7.55E-07 - Anorthosites 3 9.33E-06 4.49E-06 - - 37.1 2.47 - - 1.51 - 9.15E-07 4.06E-07 3.81E-07 1.55E-07 Table 3.2 : Mean intrinsic magnetic properties and remanent magnetizations of the four studied lunar lithologies. Standard deviations are also given.

The Fe 2+ concentration derived from paramagnetic susceptibility varies from 15 to 29 wt.% in basalts and is close to 7 wt.% in breccias and the norite. These values are in agreement with the total iron percentage given by chemical analyses ( Table 3.1). The reduction degree (Fe0/Fe 2+ ) is quite variable, varying from less than 0.006 in basalts to above 0.01 in breccias and norite. Breccia 68815 is the most reduced sample which suggests that it has probably been subjected to several heating cycles ( Pearce et al., 1974 ).

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

3.3. Remanent magnetizations ARM measurements give values ranging between 4.5 and 202 μAm 2 kg −1 in basalts, and from 0.41 to 1.4 μAm 2 kg −1 in anorthosites ( Table 3.3). Norite and breccias have intermediate ARM intensity with a mean of 52 μAm 2 kg −1. Median destructive field (MDF: strength of the field necessary to reduce the remanent magnetization to half of its original value) is used to describe the stability of SIRM and ARM against AF demagnetization ( Fig. 3.3 ; Table 3.3). For basalts, MDFs of SIRM range from 3 to 52 mT. One of the breccias (62295) has the highest MDF value of 56 mT. The ARM demagnetization curves likewise exhibits a broad range of stability with MDFs ranging from 2 mT to 63 mT in basalts. In other lithologies MDFs of ARM range from 7 mT (breccia 68815) to 36 mT (anorthosite 60215). VRM acquisition rates (measured for five samples only) allow computation of the maximum VRM (VRM max ) that may have been acquired during the 40 yr of terrestrial residence of the samples ( Table 3.3). This VRM max is computed assuming that samples were kept in a fixed position in the terrestrial magnetic field of about 50 μT, followed by a minimum period of one month in low field (~10 nT) in the shielded room. As in previous studies ( Brecher et al., 1974; Collinson et al., 1973; Gose et al., 1972 ) VRM contributes for less than ~3% of the NRM.

3.4. Magnetic anisotropy Magnetic anisotropy measurements are presented in Table 3.1. Anorthosites are too weakly magnetic to give reliable results. In basalts, anisotropy degrees are weak with P ranging from 1.01 to 1.08 (mean value 1.03) in agreement with previous measurements ( Rochette et al., 2010 ). Breccias are more anisotropic with a mean P of 1.09. The norite sample (78236) is the most anisotropic with P=1.85. Ellipsoid shapes are neutral to prolate. Typical 95% confidence cone on minimum susceptibility axes have a semi-aperture of ~3° for breccias, and ~12° for basalts and norite ( Supplemental Table 3.B). The degree of anisotropy of ferromagnetic susceptibility (Pf), computed assuming an isotropic paramagnetic contribution ( Gattacceca et al., 2008 ) follows the same trends as P.

AARM measurements give a P rem ranging from 1.03 to 1.18 in basalts, and from 1.04 to 1.09 in anorthosites. This is consistent with measurements of anisotropy of IRM by Potter (2011) that give P rem ranging from 1.01 to 1.16 for three basalts samples. These P rem values correspond to

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Sample MDF of ARM MDF of NRM MDF of VRM max 2 -1 2 -1 SIRM (A m kg ) ARM (A m kg ) NRM (A m 2 kg -1 ) (mT) (mT) (mT)

Basalt - 12002 14 1.55E-05 8 6.34E-06 25 - 15016 a 16 2.87E-05 4 1.37E-06 - - 15016 b 16 2.57E-05 5 2.00E-06 - - 15475 10 5.34E-05 8 5.16E-05 14 - 15597 52 4.48E-06 63 5.37E-07 - - 70017 3 2.02E-04 2 6.60E-05 1 5.19E-07 70035 4 1.93E-04 2 8.79E-06 2 - 71505 41 3.15E-05 21 1.95E-05 23 5.8E-07 71567 4 9.93E-05 3 2.86E-05 4 3.3E-07 74275 36 2.02E-05 16 7.79E-06 6 - 75055 29 9.21E-06 8 4.78E-06 3 -

Norite - 78236 12 5.35E-05 10 7.55E-07 - -

Breccias 62295 56 3.83E-05 29 1.11E-05 2 - 68815 17 6.36E-05 7 1.23E-04 3 -

Anorthosites 60025 12 9.28E-07 29 5.40E-07 7 1.61E-08 60215 20 1.40E-06 36 4.31E-07 9 - 65315 18 4.11E-07 32 1.71E-07 6 4.07E-09

Table 3.3 : Remanent magnetizations with the corresponding median destructive field (MDF) and VRM results obtained for studied samples. ARM: anhysteretic remanent magnetization was acquired in a DC field of 400 μT, and AC field of 110 mT. VRM max : maximum viscous remanent magnetization, which may have been acquired during the 40 yr of terrestrial residence of the samples (see text). SIRM intensities are given in Table 3.1 (M rs column).

maximum deviation of the paleomagnetic record of about 4°. For the three samples that have well defined AMS and AARM ellipsoids, the directional agreement between ellipsoid axes is rather poor with an average difference of 33°. This is attributable to the AARM measurement scheme (3-position) that does not allow estimation of the uncertainties that can be fairly large for rocks with rather low magnetic anisotropy (Prem <1.18 here).

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Fig. 3.2. a) M rs /M s versus B cr /B c for various lunar lithologies: basalts (circles), regolith breccias (stars), impact melt breccias (triangles), and norite (boxe). Solid symbols are data from this study, open symbols and stars are data from −3 2 −1 Fuller and Cisowski (1987) . b) Saturation remanence (logM rs with M rs in 10 Am kg ) versus magnetic susceptibility (log χ with χ in 10 −9 m3 kg −1) for various lunar lithologies: basalts (circles), regolith breccias (stars), impact melt breccias (triangles), norite (boxes), and anorthosites (diamonds). Solid symbols are data from this study and data from the two largest samples in Gattacceca et al. (2010a,b) , open symbols and stars are data from Rochette et al. (2010) compilation. A linear fit and the correlation coefficient obtained from our data are indicated.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Fig. 3.3. Normalized intensity of SIRM versus demagnetizing field for a selection of lunar rocks studied in this work. Mare basalts: circles, anorthosites : diamonds, impact melt breccia: triangles.

3.5. Discussion Hysteresis parameters indicate that all samples are strongly dominated by multidomain carriers ( Fig. 3.2 a). ARM and IRM show a broad range of stability against AF demagnetization revealing the presence of fine ferromagnetic grains in some lunar rocks. The frequency dependence of low field susceptibility indicates that some basaltic samples (15475 and 74275) have a non-negligible superparamagnetic contribution.

Mare basalts all have rather similar χ and M s, indicating a mean metallic Fe content of 850 ppm. The three studied anorthosites are the least magnetic among all studied samples. Samples 60215 and 65315 are also, to our knowledge, the only two diamagnetic extraterrestrial rocks ever studied. This confirms that pristine lunar anorthosites are very weakly magnetic rocks, in contrast with impact-processed regolithic samples that are among the most magnetic lunar rocks (Rochette, et al., 2010 ), as evidenced by the two studied breccia samples (62295, and 68815). The norite sample (78236) has magnetic properties close to that of basalts. It is noteworthy that there is a strong linear correlation between χ and M rs (Fig. 3.2 b), which shows that the ferromagnetic mineral assemblage is to the first order identical in all lunar rocks, with only variations in concentration.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Overall there is a good agreement between our rock magnetism results and those from previous studies. The few observed discrepancies can be attributed to the inhomogeneity of some lunar rocks or more modern and accurate measurement techniques used in this work.

4. Paleomagnetism

4.1. Generalities The NRM intensity of basalt and norite samples is between 0.5 and 66 μAm 2 kg −1. Anorthosites have the weakest NRM intensities ranging from 0.2 to 0.5 μAm 2 kg −1. These values are in agreement with values of the literature (e.g., Pearce et al., 1972 ). Breccia 68815 has the highest NRM (123 μAm 2 kg −1). Samples behavior against AF demagnetization varies widely ( Fig. 3.4 ). Correction for GRM was performed on some samples (15016a, 15016b, 60025, 60215, 65315, 71505, 71567, and 70035). This correction improved the results by averaging out measurement noise, although it did not change the directional results significantly. An example of such improvement is given in Supplemental Fig. 3.B. Eleven samples exhibit at least one stable component (12002, 15475, 60025, 60215, 65315, 68815, 70017, 71567, 71505, 74275, and 75055) and are presented in Fig. 3.4 . The intervals of stability are generally restricted to AF below 30 mT, with the exception of samples 70017, 15475 and 71505 which are stable respectively up to 70, 90 and 100 mT ( Table 3.4). For all the stable samples, the softest component (coercivity below 20 mT, e.g., sample 75055 in Fig. 3.4 k) is probably not of lunar origin and may be attributed to exposure to field in the spacecraft or back on Earth. The stable character of the AF demagnetization of the NRM is not related to the radiometric ages of the samples. The other six samples (15016a in Fig. 3.4 b and 15016b, 15597, 62295, 70035, and 78236 in Supplemental Fig. 3.C) have irregular changes in NRM intensity and direction during progressive demagnetization as already observed for many lunar samples (e.g., Banerjee, 1972; Fuller et al., 1974 ).We could not identify a stable component of magnetization in these samples. As far as reproducibility of the AF demagnetization of NRM is concerned, the two samples of basalt 15016 that we have measured show identical erratic behavior against AF demagnetization. Comparison of our NRM demagnetization data with those from previous studies also using AF can be made for five samples: 15016, 62295, 70017 and 70035. The results are comparable for 15016 and 70035 with a similar NRM intensity, and erratic behavior upon AF demagnetization.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Fig. 3.4. Orthogonal projection plots of stepwise alternating field demagnetization data of representative Apollo samples. Open and solid symbols represent projections on two perpendicular planes whose intersection is the horizontal axis. Samples 70215 (m) and 14053 (n) are from Gattacceca et al. (2010a,b) . For 75055 (k), 60025 (d), 70215, and 14053 a closer view of the initial or final demagnetization steps is given. Alternating field values are given in mT. Samples that come from ENSL laboratory (studied in a sample container that was not previously demagnetized, see text) are marked with an asterisk. For samples 12002, 71505, 71567, 70017, 70215, 74275, and 14053 the component of magnetization interpreted as possibly acquired in a lunar paleofield (see text) and computed using the principal component analysis of Kirschvink (1980) is indicated by a thick gray line.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Total AF range Paleofield Sample REM REM' REM 6mT (mT) (μT) Basalt 71505 LC 2.61E-02 2.71E-02 6.6E-02 6-18 198 HC - 3.2E-02 20-100 95 12002 LC 9.60E-03 1.50E-02 1.5E-02 4-14 44 HC - 2.1E-02 16-70 63 71567 3.11E-02 3.11E-02 3.7E-02 4-48 111 70017 3.82E-02 2.21E-02 1.4E-02 7-68 42 15475 1.10E-01 1.34E-01 1.5E-01 4-90 453 70215 8.80E-03 2.93E-03 6.6E-03 7-25 20 74275 5.00E-03 2.67E-03 9.2E-03 4-20 28 14053 5.53E-02 6.26E-02 6.7E-03 7-25 20 75055 6.60E-03 3.19E-03 2.8E-02 0-11 84 70035 4.35E-02 1.14E-02 - - - 15597 4.30E-03 6.37E-03 - - - 15016 3.20E-03 4.30E-03 - - - 15016 5.10E-03 2.45E-03 - - - Norite 78236 1.20E-03 7.79E-03 - - - Breccias 62295 6.20E-03 2.03E-03 - - - 68815 2.62E-02 5.68E-03 2.6E-02 4-14 79 Anorthosites 60025 4.19E-02 3.72E-02 8.1E-02 3-22 243 65315 6.33E-02 3.85E-02 9.6E-02 4-12 286 60215 3.49E-02 2.31E-02 - - -

Table 3.4: Paleofield estimates obtained on studied samples. Total REM=NRM/ SIRM, REM6mT=NRM6mT / IRM6mT, REM ’: REM ’ integrated over the given alternating demagnetization field range (with associated paleofield estimate). LC = low coercivity component, HC=high coercivity component. Samples marked with an asterisk are the seven selected for the discussion.

For sample 62295, previous measurements showed a rather erratic behavior (like in this study), and a large scatter in NRM intensity, with values (in μAm 2 kg −1) of 2 (0.9 g, in Brecher et al., 1973 ), 8 ( Cisowski et al., 1983 ), 64 (sample 62295, 27; 0.25 g in Brecher et al., 1973 ), and 71 (sample 62295, 159; in Fuller and Cisowski, 1987 ). Our measurements, on a 0.51 g sample, give a NRM of 11 μAm 2 kg −1. Conversely, SIRM data on the same samples provide clustered values (all in mA m 2 kg −1): 2.94 ( Brecher et al., 1973 ), 2.5 ( Cisowski et al., 1983 ), 2.0 ( Fuller and Cisowski, 1987 ), 1.86 (this study). Therefore for breccia 62295, the scatter in NRM intensities is not attributable to heterogeneous distribution of ferromagnetic grains at the sample scale, but rather to small-scale heterogeneity of NRM directions (as suggested by decreasing NRM intensities with 58

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

increasing sample mass) or to magnetic contamination of samples with higher NRM. For sample 70017 previous measurements showed NRM and SIRM values in agreement with our results: 20 μAm 2 kg −1 for NRM and 1.18 mAm 2 kg −1 for SIRM, ( Brecher et al., 1974 ) against respectively 66 and 1.73 in this study ( Table 3.1), although NRM is more coercive in Brecher et al. (1974) . To summarize, it appears from the comparison of our data with previous studies, that the NRM stability upon demagnetization is comparable for different subsamples of a given Apollo rock. 2 −1 It is worth mentioning that, although samples with low M rs (<0.5 mA m kg ) generally did not provide reliable paleomagnetic results, there is no clear correlation between the quality of NRM demagnetization and rock magnetic parameters relevant for remanent magnetization (such as B cr ,

Mrs , MDF of ARM or SIRM). It appears that the paleomagnetic behavior of a given lunar rock depends strongly on the history of the rock on the Moon (shock, thermal events …), and can be controlled by a minor fraction of the population of ferromagnetic grains. As such, it cannot be anticipated from its bulk rock magnetic properties only.

4.2. Samples with stable magnetization We identified one or several stable components of magnetization in eleven samples. The directions of these components were computed using the principal component analysis of Kirschvink (1980) , without anchoring to the origin. The inclination, declination, and maximum angular deviation (MAD, as defined by Kirschvink, 1980 ) for each component are listed in Supplemental Table 3.B. For all samples with at least one stable component of NRM, REM ’ values (as defined in Gattacceca and Rochette, 2004 ) versus alternating field are plotted in Fig. 3.5 (for samples for which we suspect a primary NRM, as discussed below) or in Supplemental Fig. 3.D. For each component of magnetization, a paleointensity estimate was computed from the REM ’. For this purpose REM ’ was integrated over the whole AF range on which the component of magnetization of interest is defined ( Gattacceca and Rochette, 2004 ). These AF ranges and paleointensity estimates are given in Table 3.4. It should be noted that the REM ’ method has a factor ~3 uncertainty (e.g., Weiss et al., 2010 ). Samples 75055, 68815, 60215 and 65315, have a single component of magnetization characterized by very low coercivity (<15 mT) and high REM ’ paleointensity (~100 μT). These components may be related to magnetic contamination during sample collection or handling. This is also suggested by the very different behavior of the NRM compared to the ARM against alternating field demagnetization ( Supplemental Fig. 3.D). The seven remaining samples (12002, 15475, 60025, 70017, 71505, 71567, and 74275) have relatively higher coercivity stable components of magnetization. In order to verify that the

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

high coercivity components trend toward the origin, we compared the MAD of such components with the angle (called α in the following, and given in Supplemental Table 3.B) between the direction of the component and the direction of the first (i.e. lowest coercivity) NRM vector of the component. If α is smaller than the MAD, the component trends toward the origin and may be a primary magnetization. In order to evaluate the nature of these high coercivity components, we overplot the AF demagnetization curves of NRM, ARM ( Fig. 3.5 , and Supplemental Fig. 3.D). If the NRM is a TRM (thermal remanent magnetization), it should have a demagnetization behavior similar to ARM (that is a decent analog for TRM, as proposed by Stephenson and Collinson (1974) ). Basalt 12002 shows a soft component isolated between 4 and 14 mT. This component may correspond to magnetic contamination. A higher coercivity component is isolated between 16 and 52 mT. This component does not trend toward the origin. However, it must be noted that even if only 10% of the stable NRM demagnetized between 16 and 52mT can originate from the sample container, the residual magnetization at 52 mT (upper AF limit for the stable component) is due in large part (~50%) to the magnetization of the sample container (see Section 2.1 ). Therefore the high coercivity component of 12002 cannot be ruled out as a primary magnetization. The similar behavior between NRM and ARM ( Fig. 3.5 ) suggests that it may be a primary TRM, as SRM (shock remanent magnetization) is typically much softer ( Gattacceca et al., 2008 ). The REM ’ paleointensity associated with this component is 50 μT, which may correspond to the ambient field at ~3.3 Ga, the age of this basalt (Supplemental Table 3.A). A Thellier – Thellier paleointensity determination was attempted by Helsley (1971) on 12002, but interpretable results were obtained only in the 20 –150 °C temperature range, which is the range of the daily thermal cycling on the Moon. Basalt 15475 has a single component of magnetization trending toward the origin and stable up to 90 mT. Its REM ’ values above 0.1 and the decrease with higher AF indicate likely magnetic contamination by strong field (>50 mT, Verrier and Rochette, 2002 ), which has completely obliterated any possible pre-terrestrial magnetization. Conversely, the NRM and ARM have identical behavior ( Supplemental Fig. 3.D) suggesting that the NRM may be a TRM. However in that case the REM ’ paleointensity would be an unrealistic 450 μT. Therefore we favor the first hypothesis of strong field contamination, which is also consistent with the several year long history of this sample in its laboratory of origin (IAS Orsay). Sample 60025 has a stable but relatively low coercivity component (between 3 and 22 mT, Fig. 3.4 d) trending toward the origin. The paleofield estimate for this component is 240 μT (REM ’=8.1×10 −2). The residual REM calculated at 22 mT (1.4×10 −2) shows that the low

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

coercivity component does not extend to higher coercivity and is probably a secondary magnetization, possibly a magnetic contamination (partial IRM). Indeed VRM is excluded because it cannot account for more than ~3% of the measured NRM ( Table 3.3), and SRM is also unlikely because it would imply a minimum ambient field of 350 μT at the time of impact (as −4 computed from the estimate of maximum SRM from Gattacceca et al., 2010a , SRM max =1.2×10

SIRM * B ambient ). TRM is also excluded by the different behavior of NRM and ARM upon AF demagnetization ( Fig. 3.5 ). Paleofield estimate for sample 70017, computed on its stable component (7 –68 mT) trending toward the origin, is of about 40 μT. This estimate is in broad agreement with previous ones ranging from 30 to 70 μT ( Stephenson et al., 1974 , Thellier –Thellier and ARM techniques), and from 7 to 10 μT ( Brecher et al., 1974 ; single heating experiment). Sample 70017 has a crystallization age of about 3.7 Ga and Nord et al. (1974) determined that 70017 had a single 2 −1 stage cooling history. Moreover VRM effects are negligible (VRM max =0.5 μAm kg which is less than 0.8% of the NRM), and SRM unrealistic in view of the coercivity spectrum of stable NRM. Therefore the observed stable component may be a primary TRM, as suggested by the similar behavior between NRM and ARM upon AF demagnetization ( Fig. 3.5 ). Basalt 71505 has a ratio of NRM to SIRM (total REM in Table 3.4) of 2.6×10 −2 and two well-defined components isolated between 6 –18 mT and 20 –100 mT with an angular distance of 63°. The associated REM ’ paleointensities are 198 μT and 95 μT respectively. The low coercivity component may be a terrestrial contamination but the high coercivity field component, that trends toward the origin, is likely a primary TRM since VRM is negligible, SRM is typically much softer ( Gattacceca et al., 2008 ), and it has a similar behavior as ARM upon AF demagnetization (Fig. 3.5 ). Basalt 71567 has a stable component between 4 and 48 mT, trending toward the origin, with a REM ’ paleointensity of about 110 μT (REM ’=3.68×10 −2). The residual REM computed at 48 mT (2×10 −2) shows that this component of magnetization probably extends to the entire coercivity spectrum, even though the fraction above 48 mT cannot be demagnetized for instrumental reasons (noise, spurious ARM during AF demagnetization …). VRM is negligible and SRM unrealistic (because it would imply a too strong paleofield of 260 μT, see discussion for sample 12002 above) so that the stable component of magnetization, which behaves like ARM upon AF demagnetization ( Fig. 3.5 ) may be a primary TRM. The crystallization age of this basalt is not clearly defined, although Rb/Sr data suggest that this basalt is among the oldest basalt in the Apollo 17 collection ( Nyquist et al., 1976 ).

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Basalt 74275 shows a stable component of magnetization between 4 and 20 mT, with a REM ’ paleointensity of 28 μT (REM ’=9.3×10 −3). This component trends toward the origin. Above 20 mT, demagnetization is chaotic, with a residual REM of about 10 −3 only. This shows that the component isolated between 4 and 20 mT does not extend to higher coercivity, and is probably not a TRM, which is confirmed by the different behavior of NRM and ARM upon AF demagnetization ( Fig. 3.5 ). This component is not likely a magnetic contamination (in view of the low REM ’ values) but rather a SRM (acquired in a minimum field of 40 μT). It is not possible to decide whether the remaining magnetization at 20 mT corresponds to a low field (~ μT) magnetization, or to the magnetization of the sample holder (see Section 2.1 ). Brecher et al. (1974) , with the Thellier –Thellier method, found a paleofield intensity between 3 μT and 13 μT but this may also correspond to a SRM as in our sample.

5. Discussion

5.1. Paleointensities Six out of seventeen samples have no stable component of magnetization (15016a/b, 15597, 62295, 70035, 78236). Their demagnetization plots are given in Supplemental Fig. 3.C. For these samples there are two possibilities to explain this behavior: the rock is intrinsically not capable of being magnetized, or it was “magnetized ” in a too weak magnetic field. This could have been tested by performing paleointensity experiments for laboratory induced magnetization acquired in different paleofields in order to determine the minimum field that can be recovered from the rock ( Lawrence et al., 2008; Tikoo et al., 2010 ), but we did not perform these experiments.

However, the ratio of NRM to SIRM (Total REM in Table 3.4, or REM 6mT after AF demagnetization at 6 mT to remove possible magnetic contamination of the low coercivity grains) is a clue to distinguish between the two possibilities. Indeed, if this ratio is in the range 10 −3 to 10 −2, the NRM may have been acquired through a classical low field (1 –10 μT) magnetization mechanism: TRM (e.g., Fuller et al., 1988 ), SRM (e.g., Gattacceca et al., 2010a ), chemical remanent magnetization (CRM) …. Conversely, if this ratio is significantly lower, it means that the sample was magnetized in a very weak (<μT) or null magnetic field. However, in samples with intrinsic low paleomagnetic fidelity, AF demagnetization even at low fields can introduce strong spurious ARM ( Tikoo et al., 2010 ) that invalidates this REM6mT approach. This is probably the case at least for our two samples of 15016.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Fig. 3.5. Normalized intensities of NRM (black line), ARM (dashed line) and of IRM (gray line) versus demagnetizing field (left) and REM ’ values versus alternating field demagnetization (right) for the seven samples selected for the discussion (see text). When two stable components are present (12002 and 71505) the LC component range is indicated by the light gray portion of the REM ’ plot. The horizontal line indicates the integrated REM ’ value and the range of integration.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

−3 For the other four samples (15597, 62295, 70035, 78236) the REM 6mT is in the 1.6×10 to 1.1×10 −2 range, the lowest value being that of the norite sample 78236. Therefore, the unstable behavior of these samples does not prove the absence of magnetic field at the time of their crystallization or last thermal/shock event ( Tikoo et al., 2010 ). Eleven of the seventeen samples have at least one stable magnetization component. Five basalt samples (12002, 70017, 71505, 71567 and 74275) have one stable component trending toward the origin that we interpreted as possibly acquired in a lunar paleofield. We also consider in this discussion two basalt samples (14053, and 70215) that were studied by Gattacceca et al. (2010a) . These two samples have well defined stable components of magnetization trending toward the origin ( Fig. 3.4 , Supplemental Table 3.B). For 70215, a primary TRM origin is ruled out by the markedly softer behavior of NRM upon AF demagnetization with respect to ARM ( Fig. 3.5 ). The magnetization of 70215 was interpreted as a SRM acquired in a paleofield roughly estimated at 95 μT ( Gattacceca et al., 2010a ). 14053 has a complex thermal and crystallization history, and its NRM is carried mostly by cohenite that was formed during a sub-solidus reduction ( Taylor et al., 2004 ). Although a primary TRM is ruled out, the NRM could be a partial TRM acquired during a late heating pulse, or a SRM acquired during a later shock-event. The estimated paleofield intensities are 20 μT and 50 μT for the partial TRM and SRM hypotheses respectively ( Gattacceca et al., 2010a ). It is noteworthy that the comparison of NRM and ARM behaviors upon AF demagnetization is not relevant for 14053 because of its composite magnetic mineralogy (FeNi metal and cohenite) of which only a fraction (cohenite) carries the NRM. Therefore, we consider seven samples (12002, 14053, 70017, 70215, 71505, 71567, and 74275) that may have been magnetized in a lunar paleofield. It is noteworthy that for three of them (14053, 70215, 74275) the magnetization is likely a SRM. SRM has been demonstrated to be a reliable recorder of paleofield intensity and direction ( Gattacceca et al., 2008, 2010b; Pohl et al., 1975 ). Therefore, although the age of the magnetization is not constrained for these samples, and the estimated paleointensities are even more approximate than for samples carrying a TRM, samples carrying a SRM may provide useful insights about lunar paleofield. Contrary to TRM acquisition that requires a stable field over the cooling duration of the rock below its blocking temperatures, SRM acquisition is instantaneous. Therefore a transient field at the time of impact may be sufficient for the rock to acquire a SRM. However, the existence of impact-generated transient fields is based only on small-scale experimental works ( Crawford and Schultz, 1988 ), and their occurrence in the magnetic signature of terrestrial impact craters has never been

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

evidenced (e.g., Carporzen et al., 2012; Louzada et al., 2008 ). For this reason, for samples with a SRM, we favor the hypothesis that the field that has been recorded is a stable lunar paleofield. For these seven samples, the estimated paleointensities are in the range 20 –110 μT (mean 55±34 μT). This suggests that a strong surface magnetic field was present on the Moon when these rocks were magnetized. Radiometric ages are known for five of these samples: 3.3 Ga for basalt 12002 ( Alexander et al., 1972 ), 3.94 for 14053 ( Stettler et al., 1973 ), 3.7 Ga for basalt 70017 (Mattinson et al., 1977 ), around 3.8 Ga for 70215 ( Schaeffer et al., 1977 ), and 3.85 Ga for 74275 (Murthy and Coscio, 1977 ). For 12002 and 70017, that carry a possible primary TRM, this may be the age of magnetization. For 70215 and 74275, that carry a possible SRM, the magnetization is younger (by an unknown duration) than their radiometric age. As discussed in Garrick-Bethell et al. (2009) and Shea et al. (2012) , in this age range (3.85 Ga and younger) the only possible source for a strong (several tens of μT) and stable lunar surface magnetic field is an internally generated field, namely a dynamo field. Such young ages may seem unrealistic for a lunar dynamo, but it has been shown that a dynamo was active on the Moon at 3.6 Ga ( Shea et al., 2012 ).

5.2. Field geometry For these seven basalt samples that have possibly recorded the lunar paleofield (be it through TRM, partial TRM, or SRM), some information may be obtained about the geometry of the paleofield. Indeed, as proposed by Potter (2011) for lunar rocks or by Lovering (1959) for eucrites, the paleohorizontal of Mare basalt samples can be estimated using their magnetic foliation plane. Petrographic observations of all seven samples considered here reveal abundant vugs and vesicles, suggesting that they are indeed effusive rocks and not intrusive. For most terrestrial effusive rocks, the magnetic foliation is indeed close to the horizontal at the time of emplacement (e.g., Canon-Tapia, 2005 ). Although this does not work for a number of terrestrial effusive rocks, we believe that it will more likely be the case on the Moon than on Earth, because the viscosity of mare basalts is much less than that of terrestrial basalts ( Weill et al., 1971 ), allowing better orientation of the grains during flow, as indicated by a higher average anisotropy ratio (usual P values in terrestrial basalts are below 1.02). The measurement of the AMS of mutually oriented sub-samples of lunar basalts, or the comparison of AMS with flow indicators obtained from petrologic studies (unfortunately not possible in this study) still needs to be performed to further support the hypothesis that themagnetic foliation plane corresponds indeed to the paleohorizontal. The seven samples considered here are relatively large (above 1 g except for 12002 and 74275), and have a well-defined anisotropy (mean Pf=1.05). The semi-angles of the 95% confidence cones around the minimum susceptibility axes (given in Supplemental Table

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

3.B) have a median value of 6° for the seven samples considered here. Moreover, multiple AMS measurements on the same sample gave consistent directional results. For all these reasons, the measured magnetic foliations are considered reliable. Because our measurements of AMS and NRM were performed with the same orientation, we can estimate the inclination of the paleofield, computed as the angle between the foliation plane (deduced from AMS measurements as the plane defined by the intermediate and maximum susceptibility axes), and the stable remanent magnetization ( Table 3.5). However in our case, the sign of the inclination is unknown. We made the hypothesis that the lunar field may be dipolar and centered on the present-day rotation axis. From there we attributed positive paleoinclinations to samples from the Northern hemisphere and negative paleoinclinations to samples from the Southern hemisphere.

Sample Latitude Paleoinclination Paleolatitude 95% confidence Departure on the (°) (°) interval (°) to true Moon (°) latitude (°) 12002 3.0 S ±22.8 -11.9 [-38 ,+6] 8.9 14053 3.6 S ±7.6 -3.8 [-6 ,+1] 0.2 70017 20.2 N ±39.8 22.6 [+17 ,+30] 2.4 70215 20.2 N ±32.5 17.7 [+12 ,+25] 2.5 71505 20.2 N ±10.7 5.4 [-5 ,+17] 14.8 71567 20.2 N ±45.6 27.1 [+18 ,+40] 6.9 74275 20.2 N ±5.9 2.9 [-6 ,+12] 17.3

Table 3.5: Paleolatitudes (with their associated 95% confidence intervals) deduced from paleoinclinations obtained for seven basalt samples whose magnetization is interpreted as possibly acquired in a lunar paleofield (see text). The 95% confidence intervals on paleolatitudes are computed according to Demarest (1983) from the maximum angular deviation computed for the stable component of magnetization ( Kirschvink, 1980 ), and the 95% confidence interval on the direction of the minimum susceptibility axis (these latter values are listed in Supplemental Table 3.B). The departure to true latitude is the angular difference between the latitude of the sample on the Moon and the paleolatitude calculated in this study.

Assuming this centered dipolar geometry, paleolatitudes can be estimated from the paleoinclinations (tanI=2tan λ, I being the paleoinclination, and λ the paleolatitude). These paleolatitudes range from −4° to 27°. Their comparison with the latitude of the Apollo sites where the samples were collected shows that these two values are indeed close, with differences ranging from 0° to 17° only (mean value=7.6°), suggesting that the centered dipolar hypothesis may be valid ( Table 3.5). In particular the distribution of these differences is not random at the 75% confidence level ( Supplemental Fig. 3.E). Because the paleomagnetic declination of our samples is unknown, the corresponding paleopoles cannot be determined. However, from the paleolatitude and the site location, we can determine

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

for each sample a small circle that contains all possible paleopoles. These small circles have an intersection ( Fig. 3.6 ), which shows that our data are compatible with the hypothesis of a dipolar field geometry. Moreover the best intersection is located at a latitude of 75°N (and a longitude of ~75°W), suggesting that the dipole was centered close to the present-day rotation axis of the Moon. It must be noted that the small circles have another intersection (although it is not as well defined as the above mentioned one) that defines a paleopole at 45°N latitude and 115°W longitude. None of our two possible paleopoles matches the ones proposed by Potter (2011) , or Runcorn (1983, 1984) , who suggested large polar wonder. Conversely, our paleopole at 75°N is in broad agreement with recent paleopoles determined through the analysis of Nectarian magnetic anomalies ( Hood, 2011 ).

Fig. 3.6. Equal area stereographic projection of possible paleopoles for each studied Apollo site. Apollo sites are also plotted. Solid lines and circles represent small circles and sites in the northern hemisphere, and dashed lines and empty circles are little circles and sites in the southern hemisphere. For Apollo 17 site, the little circle is computed from the mean paleolatitude (±s.d. for the gray area) obtained from five samples. The dark gray areas are defined by circle intersections. The solid star is the best intersection of little circles. The empty triangle is a second possible intersection for the small circles. The solid triangle is its projection in the northern hemisphere. The gray star is the paleopole defined by Hood (2011) .

The paleopole proposed by Potter (2011) is based on only two data (from Apollo 11 and 15, i.e. sites where we do not have paleoinclination data) that lack precise analysis of the NRM stability and origin. It is noteworthy that one of Potter's sample (15015) is a clast from a shock melted breccia, whose shock age is too young (1.2 Ga) to make it a possible recorder of a dynamo field. We thus estimate that the nature of the magnetization measured by Potter (2011) is not well constrained. Concerning Runcorn's paleopoles, we note that they were selected from a pole list

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

produced by Hood (1982) . The original author concluded that his determined poles were randomly oriented and later stressed that the inversion of the complex anomaly studied was highly non unique (Hood, personal communication). Based on this evaluation we estimate that our near axial paleopole is more robust than the mid-latitude paleopoles presented by Potter and Runcorn, although we agree that more data are required to fully secure our interpretation in particular from Apollo 11 and 15 sites.

6. Conclusion We have performed magnetic measurements on seventeen Apollo samples (some of them several grams in weight) of different lithology and from three different Apollo sites. This provides a significant addition to the database of lunar rock magnetism and paleomagnetism. Our rock magnetism results are in good agreement with previous studies with magnetic properties dominated by about 0.1 wt.% of multidomain FeNi grains. We also confirm that impact processing leads to an increase of ferromagnetic content, probably through meteoritic contamination. On the other hand, pristine highland rocks are the least magnetic lithologies. For instance, two out of the three studied anorthositic samples are the first diamagnetic samples ever described among extraterrestrial rocks. Our data show that magnetic susceptibility and saturation remanence are linearly correlated, which indicates that the ferromagnetic mineral assemblage is to the first order identical in all lunar rocks, with variable concentration depending on rock type. Despite many caveats that can affect lunar sample paleomagnetic studies, five samples out of seventeen gave acceptable results, with stable components of magnetization trending toward the origin that have possibly recorded a lunar paleofield. The associated paleointensities suggest that a long-lived magnetic field of several ten of μT was present at the surface of the Moon between at least 3.8 and 3.3 Ga. Using the rock magnetic fabric as a paleohorizontal proxy, we inferred paleoinclinations from our paleomagnetic data. These paleoinclinations are best explained with a dipolar field and a magnetic paleopole located close to the present-day rotation axis of the Moon. Overall, this is in favor of the existence on the Moon, at least in the 3.8 –3.3 Ga age range, of a dynamo generated, centered dipolar field, with surface field intensities of several tens of μT. Additional studies of rock magnetism and paleomagnetism of well dated lunar samples, including measurement of mutually oriented sub-samples and more precise paleointensity determinations, are however necessary to confirm this hypothesis.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Acknowledgments

This manuscript benefited from careful and constructive reviews by B.P. Weiss and an anonymous reviewer. We thank B. Marty (CRPG, Nancy), F. Albarède (ENS, Lyon), J.-L. Birck (IPGP, Paris), and J. Borg (IAS, Orsay) for the temporary loan of the samples. We acknowledge The National Aeronautics Space Administration (NASA) for allowing the loan. F. Demory is acknowledged for his precious assistance in the laboratory.

References

Alexander, E.C., Davis, P.K., Reynolds, J.H., 1972. Rare gas analysis on neutron irradiated Apollo 12 samples. Proc. 3rd Lunar Planet. Sci. Conf., 2, pp. 1787 –1795. Banerjee, S.K., 1972. Iron –Titanium –Chromite, a possible new carrier of remanent magnetization in lunar rocks. Proc. Lunar Planet. Sci. Conf. 3, 2337 –2342. Blanchard, D.P., McKay, G.A., 1981. Remnants from the Ancient Mare Crust III: Norite 78236. 12th Lunar Planet. Sci. Conf. (abstract), pp. 83 –85. Brecher, A., Vaughan, D.J., Burns, R.G., Morash, K.R., 1973. Magnetic and Mössbauer studies of Apollo 16 rock chips 60315 –51 and 62295 –27. Proc. Lunar Planet. Sci. Conf. 3, 2991 –3001. Brecher, A., Menke, W.H., Morash, K.R., 1974. Comparative magnetic studies of some Apollo 17 rocks and soils and their implications. Proc. Lunar Planet. Sci. Conf. 5, 2795–2814. Canon-Tapia, E., 2005. Anisotropy of magnetic susceptibility of lava flows and dykes: a historical account. Magnetic fabric: Methods and applications, Geological society special publication, 238, pp. 205 –225. Carporzen, L.,Weiss, B.P., Gilder, S.A., Pommier, A., Hart, R.J., 2012. Lightning remagnetization of the Vredefort impact crater: no evidence for impact-generated magnetic fields. J. Geophys. Res. 117, 1 –17. Cisowski, C.S., Dunn, J.R., Fuller, M., Rose, M.F., Wasilewski, P.J., 1974. Impact processes and lunar magnetism. Proc. Lunar Planet. Sci. 3, 2841 –2858. Cisowski, S.M., Collinson, D.W., Runcorn, S.K., Stephenson, A., Fuller, M., 1983. A review of lunar paleointensity data and implications for the origin of lunar magnetism. J. Geophys. Res. 88, 691 –704. Collinson, D.W., 1993. Magnetism of the Moon — a lunar core dynamo or impact magnetization? Surv. Geophys. 14, 89 –118. Collinson, D.W., Stephenson, A., Runcorn, S.K., 1973. Magnetic properties of Apollo 15 and 16 rocks. Proc. 4th Lunar Planet. Sci. Conf., 3, pp. 2963 –2976. Crawford, D.A., Schultz, P.H., 1988. Laboratory observations of impact-generated magnetic fields. Nature 336, 50 –52. Demarest, H.H., 1983. Error analysis for the determination of tectonic rotation from paleomagnetic data. J. Geophys. Res. 88, 4321 –4328. Duncan, A.R., Erlank, A.J.,Willis, J.P., Sher, M.K., Ahrens, L.H., 1974. Trace element evidence for a two- stage origin of some titaniferous mare basalts. Proc. 5th Lunar Sci. Conf., 2, pp. 1147 –1157. Enkin, R.J., Dunlop, D.J., 1988. The demagnetization temperature necessary to remove viscous remanent magnetization. Geophys. Res. Lett. 15, 514 –517. Fuller, M., 1998. Lunar magnetism: a retrospective view of the Apollo sample magnetic studies. Phys. Chem. Earth. 23, 725 –735. Fuller,M., Cisowski, S.M., 1987. Lunar paleomagnetism. In: Jacobs, J. (Ed.), Geomagnetism, vol. 2, pp. 307 –455. Fuller, M., Rose, F., Wasilewski, P.J., 1974. Preliminary results of an experimental study of the magnetic effects of shocking lunar soil. Moon 9, 57 –61. Fuller, M., Cisowski, S., Hart, M., Haston, R., Schmidtke, E., 1988. NRM:IRM(S) demagnetization plots; an aid to the interpretation of natural remanent magnetization. Geophys. Res. Lett. 15, 518 –521. Garrick-Bethell, I., Weiss, B.P., Shuster, D.L., Buz, J., 2009. Early lunar magnetism. Science 323, 356 –359.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Gattacceca, J., Rochette, P., 2002. Pseudopaleosecular variation due to remanence anisotropy in a pyroclastic flow succession. Geophys. Res. Lett. 29 (126), 1–4. Gattacceca, J., Rochette, P., 2004. Toward a robust normalized magnetic paleointensity method applied to meteorites. Earth Planet. Sci. Lett. 227, 377 –393. Gattacceca, J., Rochette, P., Gounelle, M., Van Ginneken, M., 2008. Magnetic anisotropy of HED and Martian meteorites and implications for the crust of Vesta and Mars. Earth Planet. Sci. Lett. 270, 280 – 289. Gattacceca, J., Boustie, M., Hood, L., Cuq-Lelandais, J.-P., Fuller, M., Bezaeva, N.S., De Resseguier, T., Berthe, L., 2010a. Can the lunar crust be magnetized by shock: experimental groundtruth. Earth Planet. Sci. Lett. 299, 42 –53. Gattacceca, J., Boustie, M., Lima, E.A., Weiss, B.P., De Resseguier, T., Cuq-Lelandais, J.-P., 2010b. Unraveling the simultaneous shock magnetization and demagnetization of rocks. Phys. Earth Planet. Inter. 182, 42 –49. Gose, W.A., Pearce, G.W., Strangway, D.W., Larson, E.E., 1972. On the applicability of lunar breccias for paleomagnetic interpretations. Moon 5, 106 –120. Halekas, J.S., Mitchell, D.L., Lin, R.P., Frey, S., Acuña, M.H., Binder, A.B., 2001. Mapping of lunar crustal magnetic fields using lunar prospector electron reflectometer data. J. Geophys. Res. 106, 27,841 –27,852. Hargraves, R.B., Dorety, N., 1971. Magnetic properties of some lunar crystalline rocks returned by Apollo 11 and Apollo 12. Proc. 2nd Lunar Sci. Conf., 3, pp. 2477 –2483. Helsley, C.E., 1971. Evidence for an ancient lunar magnetic field. Proc. 2nd Lunar Sci. Conf., 3, pp. 2485 – 2490. Hood, L.L., 1982. Sources of lunar magnetic anomalies and their bulk directions of magnetization. Proc. 12th Lunar Planet. Sci. Conf. , pp. 817 –830. Hood, L.L., 2011. Central magnetic anomalies of Nectarian-aged lunar impact basins: probable evidence for an early core dynamo. Icarus 211, 1109 –1128. Hood, L.L., Artemieva, N.A., 2008. Antipodal effects of lunar basin-forming impacts: initial 3D simulations and comparisons with observations. Icarus 193, 485 –502. Hood, L.L., Huang, Z., 1991. Formation of magnetic anomalies antipodal to lunar impact basins — two- dimensional model calculations. J. Geophys. Res. 96, 9837 –9846. Hood, L.L., Zakharian, A., Halekas, J., Mitchell, D.L., Lin, R.P., Acuña, M.H., Binder, A.B., 2001. Initial mapping and interpretation of lunar crustal magnetic anomalies using lunar prospector magnetometer data. J. Geophys. Res. 106, 27,825 –27,839. Jelinek, V., 1981. Characterization of the magnetic fabric of rocks. Tectonophysics 79, 63 –67. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of palaeomagnetic data. Geophys. J. Int. 62, 699 –718. Laul, J.C., Schmitt, R.A., 1973. Chemical composition of Apollo 15, 16, and 17 samples. Proc. Lunar Planet. Sci. Conf. 2, 1349 –1367. Lawrence, K., Johnson, C., Tauxe, L., Gee, J., 2008. Lunar paleointensity measurements: implications for lunar magnetic evolution. Phys. Earth Planet. Inter. 168, 71 –87. Louzada, K.L., Weiss, B.P., Maloof, A.C., Stewart, S.T., Swanson-Hysell, N.L., Soule, S.A., 2008. Paleomagnetism of lunar impact crater, India. Earth Planet. Sci. Lett. 275, 308 –319. Lovering, J.F., 1959. The magnetic field in a primary meteorite body. Am. J. Sci. 257, 271 –275. Lunar Sample Preliminary Examination Team, 1973. Apollo 17 lunar samples: chemical and petrographic description. Science 182, 659 –672. Mattinson, J.M., Tilton, G.R., Todt, W., Chen, J.H., 1977. Lead isotope studies of mare basalt 70017. Proc. 8th Lunar Sci. Conf. , pp. 1473 –1487. McKinley, J.P., Taylor, G.J., Keil, K., Ma, M.-S., Schmitt, R.A., 1984. Apollo 16: impact sheets, contrasting nature of the Cayley Plains and Descartes Mountains, and geologic history. J. Geophys. Res. Suppl. 89, 513 –524. Murthy, V.R., Coscio, M.R., 1977. Rb –Sr isotopic systematics and initial SR considerations for some lunar samples. Proc. Lunar Planet. Sci. Conf. 8, 706 –708. Nagata, T., Ishikawa, Y., Kinoshita, H., Kono, M., Syono, Y., Fisher, R.M., 1970. Magnetic properties and natural remanent magnetization of lunar materials. Geochim. Cosmochim. Acta Suppl. 1, 2325 –2340. Nava, D.F., 1974. Chemical compositions of some soils and rock types from the Apollo 15, 16, and 17 lunar sites. Proc. Lunar Planet. Sci. Conf. 2, 1087 –1096.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Nord, G.L., Lally, J.S., Heuer, A.H., Christie, J.M., Radcliffe, S.V., Fisher, R.M., Griggs, D.T., 1974. A mineralogical study of rock 70017, an ilmenite-rich basalt, by high voltage electron microscopy. Lunar Planet. Sci. Conf. (abstract) 5, 556 –558. Nyquist, L.E., Bansal, B.M., Wiesmann, H., 1976. Sr isotopic constraints on the petrogenesis of Apollo 17 mare basalts. Proc. Lunar Sci. Conf. 2, 1507 –1528. Pearce, G.W., Strangway, D.W., Gose, W.A., 1972. Remanent magnetization of the lunar surface. Proc. 3rd Lunar Planet. Sci. Conf., 3, 2449 –2464. Pearce, G.W., Strangway, D.W., Gose, W.A., 1974. Magnetic properties of Apollo samples and implications for regolith formation. Proc. 5th Lunar Planet. Sci. Conf., 3, pp. 2815 –2826. Pohl, J., Bleil, U., Hornemann, U., 1975. Shock magnetization and demagnetization of basalt by transient stress up to 10 kbar. J. Geophys. 41, 23 –41. Potter, D.K., 2011. Novel magnetic techniques for rapidly detecting palaeomagnetically important singledomain iron particles and obtaining directional palaeomagnetic data from “unoriented” lunar rock samples. Can. Aeronaut. Space J. 57, 12 –23. Rhodes, J.M., Hubbard, N.J., Wiesmann, H., Rodgers, K.V., Brannon, J.C., Bansal, B.M., 1976. Chemistry, classification, and petrogenesis of Apollo 17 mare basalts. Proc. Lunar Planet. Sci. Conf. 2, 1467 –1489. Rochette, P., Gattacceca, J., Ivanov, A.V., Nazarov, M.A., Bezaeva, N.S., 2010. Magnetic properties of lunar materials: meteorites, Luna and Apollo returned samples. Earth Planet. Sci. Lett. 292, 383 –391. Rose, H.J., Baedecker, P.A., Berman, S., Christian, R.P., Dwornik, E.J., Finkelman, R.B., Schnepfe, M.M., 1975. Chemical composition of rocks and soils returned by the Apollo 15, 16, and 17 missions. Proc. Lunar Planet. Sci. Conf. 2, 1363 –1373. Runcorn, S.K., 1983. Lunar magnetism, polar displacement and primeval satellites in the Earth –Moon system. Nature 304, 589 –596. Runcorn, S.K., 1984. The primeval axis of rotation of the Moon. Philos. Trans. R. Soc. London, Ser. A 313, 77 –83. Runcorn, S.K., Collinson, D.W., O'Reilly, W., Battey, M.H., Stephenson, A., Jones, J.M., Manson, A.J., Readman, P.W., 1970. Magnetic properties of Apollo 11 lunar samples. Geochim. Cosmochim. Acta Suppl. 1, 2369 –2387. Schaeffer, O.A., Muller, H.W., Grove, T.L., 1977. Laser 39Ar –40Ar study of Apollo 17 basalts. Lunar Planet. Sci. Conf. (abstract) 8, 837 –839. Shea, E.K., Weiss, B.P., Cassata, W.R., Shuster, D.L., Tikoo, S.M., Gattacceca, J., Grove, T.L., Fuller, M.E., 2012. A Long-Lived Lunar Core Dynamo. Science 335, 453 –456. Stephenson, A., 1980. A gyroremanent magnetization in anisotropic magnetic material. Nature 284, 49 –51. Stephenson, A., 1993. Three-axis static alternating field demagnetization of rocks and the identification of natural remanent magnetization, gyroremanent magnetization, and anisotropy. J. Geophys. Res. 98, 373 –381. Stephenson, A., Collinson, D.W., 1974. Lunar magnetic field paleointensities determined by an anhysteretic remanent magnetization method. Earth Planet. Sci. Lett. 23, 220 –228. Stephenson, A., Collinson, D.W., Runcorn, S.K., 1974. Lunar magnetic field palaeointensity determinations on Apollo 11, 16, and 17 rocks. Proc. Lunar Sci. Conf. 3, 2859 –2871. Stephenson, A., Runcorn, S.K., Collinson, D.W., 1975. On changes in the intensity of the ancient lunar magnetic field. Proc. 6th Lunar Sci. Conf. , pp. 3049 –3062. Stettler, A., Eberhardt, P., Geiss, J., Grogler, N., Maurer, P., 1973. Ar 39 –Ar 40 ages and Ar 37 –Ar 38 exposure ages of lunar rocks. Proc. 4th Lunar Sci. Conf. , pp. 1865 –1888. Strangway, D.W., 1977. The magnetic fields of the terrestrial planets. Phys. Earth Planet. Inter. 15, 121 – 130. Taylor, L.A., Patchen, A.P., Mayne, R.G., Taylor, D.H., 2004. The most reduced rock from the moon, Apollo 14 basalt 14053: its unique features and their origin. Am. Mineral. 89, 1617 –1624. Tikoo, S.M., Weiss, B.P., Buz, J., Garrick-Bethell, I., Grove, T.L., Gattacceca, J., 2010. Ancient lunar dynamo: absence of evidence is not the evidence of absence. Lunar Planet. Sci. Conf. Abstracts n°1533, p. 2705. Verrier, V., Rochette, P., 2002. Estimating peak currents at ground lightning impacts using remanent magnetization. Geophys. Res. Lett. 29, 1867. Waenke, H., Palme, H., Baddenhausen, H., Dreibus, G., Jagoutz, E., Kruse, H., Palme, C.,

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Spettel, B., Teschke, F., Thacker, R., 1975. New data on the chemistry of lunar samples: primary matter in the lunar highlands and the bulk composition of the moon. Proc. Lunar Sci. Conf. 2, 1313 –1340. Warner, R.D., Taylor, G.J., Conrad, G.H., Northrop, H.R., Barker, S., Keil, K., Ma, M.-S., Schmitt, R., 1979. Apollo 17 high-Ti mare basalts — new bulk compositional data, magma types, and petrogenesis. Proc. Lunar Planet. Sci. Conf. 1, 225 –247. Wasilewski, P., 1974. Magnetochemistry of returned samples from the Apollo landing sites. Meteoritics 9, 418. Weber, R.C., Lin, P.-Y., Garnero, E.J., Williams, Q., Lognonne, P., 2011. Seismic detection of the lunar core. Science 331, 309 –321. Weill, D.F., Grieve, R.A., Mc Callum, S., Bottinga, Y., 1971. Mineralogy-petrology of lunar samples: microprobe studies of samples 12021 and 12022; Viscosity of melts of selected lunar compositions. Proc. Lunar Sci. Conf. 2nd, pp. 413 –430. Weiss, B.P., Gattacceca, J., Stanley, S., Rochette, P., Christensen, U.R., 2010. Paleomagnetic records of meteorites and early planetesimal differenciation. Space Sci. Rev. 152, 341 –390. Willis, J.P., Ahrens, L.H., Danchin, R.V., Erlank, A.J., Gurney, J.J., Hofmeyr, P.K., McCarthy, T.S., Orren, M.J., 1971. Some interelement relationships between lunar rocks and fines, and stony meteorites. Proc. 2nd Lunar Planet. Sci. Conf., 2, pp. 1123 –1138.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

SUPPLEMENTARY FILES

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Laboratory Radiometric Sample of origin ages (Ga) Method ref. Basalts 3.24 ± 0.05 Ar-Ar Turner 1971 12002* ENSL 3.26 ± 0.06 Ar-Ar Alexander et al. 1972 3.36 ± 0.10 Rb-Sr Papanastassiou and Wasserburg 1970 3.96 ± 0.04 Rb-Sr Papanastassiou and Wasserburg 1971 3.94 ± 0.04 Ar plateau Stettler et al. 1973 14053* - 3.94 ± 0.05 Ar plateau Turner et al. 1971 3.92 ± 0.08 Ar plateau Husain et al. 1972 Murthy et al. 1972 15016 a and b CRPG 3.29 ± 0.05 Rb-Sr Evensen et al. 1973 a 3.43 ± 0.15 Rb-Sr 15475 IAS Snyder et al. 1997 3.37 ± 0.05 Sm-Nd 15597 ENSL - -- 3.80 ± 0.03 Ar-Ar Phinney et al. 1975 3.67 ± 0.18 Rb-Sr Nyquist et al. 1975 70017* IPGP 3.7 U-Pb Mattinson et al. 1977 3.67 ± 0.12 Ar-Ar Schaeffer and Schaeffer 1977 a, b 3.82 ± 0.06 Rb-Sr Evensen et al. 1973 a, b 70035 CRPG 3.75 ± 0.07 Ar-Ar Stettler et al. 1973 3.73 ± 0.11 Rb-Sr Nyquist et al. 1974 3.84 ± 0.04 Ar-Ar Kirsten and Horn 1974 70215* - 3.63 -3.85 Ar-Ar Schaeffer et al. 1977 a, b 71505* CRPG - - - 71567* CRPG - - - 3.81 ± 0.32 Rb-Sr Nyquist et al. 1976 a, b 74275* ENSL 3.85 ± 0.08 Rb-Sr Murthy and Coscio 1977 3.83 ± 0.10 Rb-Sr Tatsumoto et al. 1973 3.82 ± 0.05 Ar-Ar Huneke et al. 1973 75055 ENSL 3.76 ± 0.05 Ar-Ar Turner et al. 1973 a, b 3.82 ± 0.05 Ar-Ar Kirsten et al. 1973 a, b 3.77 ± 0.06 Rb-Sr Tera et al. 1974 Norite 4.25 ± 0.09 Pb-Pb Hinthorne et al. 1977 4.38 ± 0.02 Rb-Sr Nyquist et al. 1981 4.43 ± 0.05 Sm-Nd Nyquist et al. 1981 78236 ENSL 4.39 Ar-Ar Nyquist et al. 1981 4.34 ± 0.04 Sm-Nd Carlson and Lugmair 1981 4.11 ± 0.02 Ar-Ar Aeschlimann et al. 1982 4.426 ± 0.065 U-Pb Premo and Tatsumoto 1991 Breccias 3.89 ± 0.05 Ar-Ar Turner et al. 1973 a 62295 ENSL 4.00 ± 0.06 Rb-Sr Mark et al. 1974 3.89 ± 0.012 Ar-Ar Norman et al. 2006 4.12 ± 0.024 68815 ENSL Ar plateau Schaeffer et al. 1976 3.63 ± 0.054 Anorthosites 4.19 ± 0.06 Ar-Ar Schaeffer and Husain 1974 a, b 60025 CRPG 4.44 ± 0.02 Sm-Nd Carlson and Lugmair 1988 65315 CRPG - - - 60215 CRPG --- Supplemental Table 3.A Samples studied in this work with their laboratory of origin and their radiometric ages. Samples marked with an asterisk are the seven that have potentially recorded a lunar paleofield. Radiometric ages not available in the literature are indicated by: -.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

K1 K2 K3 α 95 on AF range Sample α I D I D I D K3 axe (mT) I D MAD Basalt 12002 LC 78 294 3 39 12 129 34.8 4-14 -46.6 335.1 14.3 - 12002 HC* 16-52 -69.5 253.8 25.7 35 14053* 28 77 21 179 54 300 0.2 7-25 -23 334 6.1 5.5 15016 5 189 72 293 18 97 44 - - 32.9 - 15016 6 345 2 255 84 143 40.1 - - 41.6 - 15475 ------4-90 -5 106.8 7 3.9 15597 41 262 13 3 46 107 47.8 - - 33.2 - 70017* 8 233 28 328 60 128 5.8 7-68 -19.8 350.5 9.9 4.0 70035 54 228 7 128 35 33 3.4 - - 42.2 - 70215* 49 354 2.6 7-25 -20 237 12.3 11.3 71505 LC 58 200 29 48 13 311 25.6 6-18 60.8 160 1.7 - 71505 HC* 20-100 7.9 208 3.9 1.9 71567* 48 43 33 178 23 284 6.8 4-48 11.5 238.9 15.1 2.5 74275 12 76 11 343 73 213 15.7 4-20 1.1 98.7 15.6 5.8 75055 41 157 13 56 47 312 20.3 0-11 36.5 146.9 11.7 - Norite 78236 29 92 45 329 32 202 2.3 - - 34.3 - Breccias 62295 57 68 18 188 27 288 3.3 - - 35.8 - 68815 33 64 46 196 26 316 2.3 4-14 -26.1 357.2 10.9 - Anorthosites 60025 54 170 35 10 10 274 38.7 3-22 10.4 169.9 5.3 4.7 65315 34 178 37 58 35 296 27.1 4-12 62.7 334.5 7.4 - 60215 75 305 10 77 11 168 29.1 2-14 -1.3 231.7 8.5 - Supplemental Table 3.B I = inclination and D = declination of K1, K2, K3 that are maximum, intermediate and minimum susceptibility axes respectively. α95 are the semi -angles of the 95% confidence cones around the minimum susceptibility axes, K3. Alternating demagnetization field range of each component (LC = low coercivity component, HC = high coercivity component) and their respective I = inclination, D = declination and MAD = maximum angular deviation are given. For samples with no stable component of magnetization MAD was calculated on a typical range (6 –42 mT) averaged from the others samples. α is the angle between the direction of the component and the direction of the first (i.e. lowest coercivity) NRM vector of the component. Parameters that were not measurable are indicated by: -. Samples marked with an asterisk are the seven that have potentially recorded a lunar paleofield.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Supplemental Figure 3.A Some representative hysteresis loops for a selection of lunar rocks studied in this work.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Supplemental Figure 3.B Example of a) Orthogonal projection plot of stepwise AF demagnetization data for sample 70017 before and after GRM corrections. b) NRM versus demagnetizing field for sample 65315 before and after GRM correction. c) Equal area stereographic projection of stepwise AF demagnetization for sample 65315 before and after GRM correction.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

Supplemental Figure 3.C Orthogonal projection plots of stepwise AF demagnetization data for the six samples that have no stable component of magnetization. Open and solid symbols represent projections on two perpendicular planes whose intersection is the horizontal axis. Samples that come from ENSL laboratory are marked with an asterisk.

CHAPITRE 3 : PALEOMAGNETISME LUNAIRE

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Chapitre 4: Magnétisme des chondrites

4.1 Les Chondrites

Les chondrites sont des météorites non différenciées, elles se divisent en plusieurs groupes : les chondrites ordinaires, les rumurutites, les chondrites carbonées et les chondrites à enstatite. De manière générale, ces météorites peuvent être considérées comme des sédiments composés de divers solides issus du système solaire primitif. Ainsi, les chondrites consistent en un assemblage de chondres, de fragments de chondres et de minéraux, de grains de FeNi métalliques, et de CAIs, englobés dans une fine matrice silicatée riche en phyllosilicates. Les CAIs et les chondres se sont formés tôt dans la nébuleuse solaire. Les CAIs, considérés comme les plus vieux solides du système solaire sont datés à ~4568 Ma ( Amelin et al., 2002 ; Bouvier and Wadhwa, 2010 ). Les chondres sont les objets pré-accrétion les plus abondants en masse. Leur formation aurait commencé quasi simultanément ou peu après celle des CAIs ( Amelin and Krot, 2007 ; Russel et al., 2006 ) e t aurait duré ~3 Ma ( Connelly et al., 2012 ). Peu après, la poussière, les CAIs et les chondres se sont accrétés pour former des petits planétésimaux, qui à leur tour, par accrétion ont donné lieu à la formation des astéroïdes parents des chondrites. Considérées comme des objets primitifs ayant peu évolués depuis leur formation, les chondrites sont donc susceptibles de nous donner des informations sur les champs magnétiques dans le système solaire jeune. Une littérature relativement abondante a abordé le paléomagnétisme des chondrites ordinaires et carbonées ( Weiss et al., 2010 , pour une synthèse). Toutes les chondrites sont aimantées à divers degrés ( Gattacceca et Rochette, 2004 ). Mais l'ensemble des limitations et complications évoquées dans les parties précédentes font que très peu de conclusions claires se sont dégagées, excepté pour une chondrite carbonée (CV) ( Carporzen et al., 2011 ).

Huit classes de météorites carbonées sont recensées : CB, CH, CI, CK, CM, CO, CR, CV. Ces classes se distinguent par la taille des chondres, le % de matrice, le %CAI, le % métal, la composition des olivines (Weisberg et al., 2006 ), ou encore les isotopes de l’ oxygène (Clayton, 1993 ). Une classification pétrographique, basée sur l’intensité du métamorphisme thermique et hydrothermal, permet d’affiner la classification des chondrites (Fig. 4.1 ). Une partie des météorites carbonées a la particularité d'avoir subi un métamorphisme de type hydrothermal. Les chondrites carbonées dont le type pétrographique est noté de 1 à 2 ont subi ce type de

CHAPITRE 4 : MAGNETISME DES CHONDRITES

métamorphisme. Le degré 1 indique le degré le plus hydrothermalisé et correspond à la disparition totale des chondres. Le degré 2 correspond à la présence de phyllosilicates et de carbonates dans la matrice et les chondres sont préservés. Les CM et les CR sont de type pétrographique (1 ou) 2, les CV, les CO et les CH de type 3 (le degré 3 caractérise un hydrothermalisme et un métamorphisme thermique de faible intensité avec des contours de chondres nets, non altérés par l’hydrothermalisme, et des compositions des silicates inhomogènes tout comme pour les types 1 et 2). Toutes les autres classes de météorites peuvent être de type 3 à 6, avec un métamorphisme thermique de plus en plus intense du type 3 à 6 (Huss et al., 2006 ). Ave c l’augmentation de température les chondres sont de moins en moins visibles et la matrice est de plus en plus recristallisée.

TYPE PETROGRAPHIQUE 1 2 3 4 5 6 CI CM CR CO CV CK Chondrites ordinaires Rumurutites GROUPE CHIMIQUE GROUPE Chondrites à enstatite <150°C <200 °C <400 °C <600 °C <700 °C <750 °C <9 50°C

ALTERATION HYDROTHERMALE METAMORPHISME THERMIQUE Fig. 4.1. Classification des chondrites d’après Sephton (2002) . Type pétrographique en fonction du type d’a ltération. Les échantillons étudiés ici sont figurés par un rectangle rouge.

Ces différences de degré de métamorphisme vont donner à chaque classe de chondrites carbonées un intérêt particulier. Par exemple , bien qu’elles aient subit une altération aqueuse importante, les chondrites CI sont les plus proches de la composition du soleil et sont donc les moins fractionnées chimiquement par rapport au système solaire dans son ensemble ( Anders and Grevesse 1989 ). Les chondrites CV sont moins altérées que les chondrites CI et contiennent les CAIs les plus gros et les plus abondants (Brearley and Jones, 1998 ). De nombreuses classes de chondrites carbonées contiennent aussi beaucoup de chondres. Le caractère primitif des composants des chondrites carbonées font donc de ces météorites des enregistreurs potentiels des champs magnétiques précoces (Thomas and Weiss 2006 ). La plupart des rumurutites sont des

CHAPITRE 4 : MAGNETISME DES CHONDRITES

brèches et réunissent souvent plusieurs types pétrographiques (de 3 à 6) dans une même météorite. Cette caractéristique offre la possibilité via l’étude magnétique de retracer différents évènements subis par la météorite.

Dans ce chapitre nous allons nous intéresser à deux groupes de chondrites, les chondrites carbonées CM et les rumurutites.

4.2 Les Chondrites Carbonées CM

4.2.1 Le magnétisme des chondrites carbonées

Parmi les huit groupes de chondrites carbonées, les chondrites CV et CM sont de loin les mieux étudiées, suivie par les chondrites CO et CI. Les chondrites CK et CR n’ont quasiment pas été étudiées excepté un petit nombre d’analyse s réalisées sur Karoonda et Renazzo . A l’exception d’un résumé de congrès (Wasilewski et al., 2000 ) et d’une mesure de NRM ( Guskova, 1970 ), il n’ex iste aucune étude publiée sur les chondrites CB. Aucune étude paléomagnétique n ’a été réalisée pour les chondrites CH Sans rentrer dans le détail (cf Weiss et al., 2010 pour une revue), nous rappelons brièvement ici les principales observations et conclusions obtenues en termes de magnétisme des roches et de paléomagnétisme pour les chondrites carbonées. Les phases ferromagnétiques rencontrées dans ce groupe de météorite sont principalement : pyrrhotite, magnétite, alliages Fe- Ni, cohénite et schreibersite (Hyman and Rowe, 1986 ; Rochette et al., 2008 ). La susceptibilité magnétique qui fournit des informations sur la quantité des phases magnétiques (essentiellement métal et magnétite) varie entre log χ~3.5 (pour les CM les moins magnétiques) et log χ~5.6 (pour les CB et les CH) dans les chondrites carbonées (Rochette et al., 2008 ). Par rapport aux autres groupes, on notera que pour les chondrites CM et CV, la susceptibilité varie sur presque deux ordres de grandeurs, indiquant une grande variabilité dans l’assemblage et /ou de la teneur en minéraux magnétiques (Rochette et al., 2008 ). Parmi toutes les études paléomagnétiques réalisées, très peu ont abouti à des conclusions précises concernant l’origine de l’aimantation contenue dans ces météorites. Les désaimantations par champs alternatifs et thermiques ont révélé la présence d’ aimantations stables pour plusieurs chondrites carbonées : Murchison (CM), Orgueil (CI), Karoonda (CK), Allende (CV)… (Larson et al. 1973 ; Banerjee and Hargraves 1972 ; Guskova, 1976 ; Brecher and Arrhenius, 1974, Carpozen et al., 2011 ). Allende est la chondrite carbonée la plus étudiée en terme de

CHAPITRE 4 : MAGNETISME DES CHONDRITES

paléomagnétisme et probablement celle dont les résultats sont les plus concluants à l’heure actuelle. Les études précédentes ont proposé plusieurs interprétations pour la NRM des chondrites carbonées. Ainsi la NRM a été interprétée comme une aimantation acquise avant l’arrivée sur Terre sous la forme d’une TRM ou une aimantation rémanente thermo-chimique (TCRM), ou une pTRM ou encore une CRM ( Larson et al., 1973 ; Banerjee and Hargraves, 1972 ; Carpozen et al., 2011 ). Bien que la nature de la NRM reste incertaine , de nombreux auteurs ont tenté d’estimer la paléointensité du champ magnétisant. Les valeurs de paléointensité estimées pour les chondrites par la méthode Thellier-Thellier, varient entre 10 et 1600 µT (Stacey et al., 1961 ; Lanoix et al., 1978 ). Avec les méthodes de normalisation basées sur les désaimantations par champ alternatif, ces valeurs ont été revues à la baisse, avec des paléointensités de ~5 µT obtenues pour des CK, CO et CM ( Gattacceca and Rochette, 2004 ; Acton et al., 2007 ). Une compilation des toutes les estimations de paléointensité réalisées pour la météorite Allende donne une paléointensité moyenne de ~20 µT (Carpozen et al., 2011 ). La NRM des chondrites carbonées a été longtemps interprétée comme un enregistrement des champs externes générés dans le sytème solaire primitif soit par le soleil, soit dans le disque lui- même (Levy and Sonett, 1978 ; Sugiura and Strangway, 1988 ; Nagata 1979a ; Acton et al., 2007 ; Cisowski 1987 ; Stacey 1976 ). Cependant de nouvelles contraintes géochimiques et pétrologiques associées à de nouveaux modèles suggèrent la possibilité que les chondrites puissent être issues d’un corps partiellement différencié avec une croûte externe chondritique (d’où est issue la météorite) et à l’intérieur un noyau différencié métallique fondu en convection, tel que proposé pour le corps parent des chondrites CV lors de l’étude de la météorite Allende ( Carpozen et al., 2011 ). La même hypothèse a récemment été faite pour expliquer l’aimantation contenue dans la météorite Kaba (chondrite CV) ( Gattacceca et al., 2013b).

4.2.2 Le magnétisme des chondrites carbonées CM

Pour notre étude nous avons choisi un groupe de chondrites carbonées, les chondrites CM (d’après le nom de la première météorite répertoriée dans cette catégorie, Mighei , tombée en 1889 en Ukraine). Ce groupe des chondrites carbonées a été choisi pour sa diversité texturale (Brearley, 1995 ; Lindgren et al., 2013 ) et son potentiel à avoir conservé des aimantations très anciennes. De plus, les chondrites CM sont reconnues comme étant un des assemblages les plus primitifs de notre système solaire (Anders, 1971 ; Mason, 1962 ; Wood 1963 ), et pourraient s’être formées dans le système solaire externe. Enfin, les CM pourraient représenter la transition entre les astéroïdes et les comètes ( Wasson and Wetherill, 1979 ; Gounelle et al., 2008 ). Ainsi ces

CHAPITRE 4 : MAGNETISME DES CHONDRITES

météorites offrent la possibilité d’estimer l’intensité des champs magnétiques présents dans la partie externe du système solaire (>3.5 UA) ( Clayton et al., 1976, 1977 ). Après accrétion de leur principaux constituants, les chondrites CM ont subi de nombreux processus (altération basse température (Guo and Eiler, 2007 ), métamorphisme thermique basse température (T < 100°C) (Miyamoto, 1991 ; Kitajima et al., 2002 ; Nakamura, 2005 ), chocs <5 GPa (Scott et al., 1992 ) qui, comme nous l’avons vu dans le chapitre 2, peuvent tous avoir contribué à l’acquisition de l’aimantation contenue actuellement dans ces météorites. Les propriétés magnétiques d’une douzaine de chondrites CM2 ont déjà été étudiées précédemment (Weiss et al., 2010 pour une synthèse ; Elmaleh et al., 2012 ). La minéralogie magnétique des chondrites CM est composée, dans des proportions variables de fer métallique (2.0-35.5 wt.%,), de magnétite (<3 wt.%), et de sulfures de Fe-Ni (<3.8 wt%) ( Hyman and Rowe, 1986 ; Rochette et al., 2008 ; Kimura et al., 2011 ). Le paléomagnétisme des chondrites CM a aussi été étudié précédemment (Weiss et al., 2010 pour une synthèse), mais aucune conclusion claire ne se dégage concernant la nature de la NRM et celle du champ magnétisant. L’intensité du champ magnétisant a été estimée dans une large fourchette, entre 0.2 et 20 µT selon la méthode utilisée (Banerjee and Hargraves, 1972 ; Kletetschka et al., 2003 ).

L’article qui suit (Cournède et al. , en préparation) présente l’étude magnétique et paléomagnétique que nous avons réalisée sur sept chondrites CM : Banten, Cold Bokkeveld, Paris, Murchison, Murray, Mighei, Nogoya (toutes sont des chutes à l’ exception de Paris qui est une trouvaille). On notera que la météorite Paris a fait l’objet d’une étude spécifique à cause de ces caractéristiques particulières. En effet, Paris est une chondrite CM qui a en grande partie échappé à l’hydrothermalisme qui caractérise les autres météorites de ce groupe (Zanda et al., 2010 ). Ainsi, l ’étude de la m inéralogie magnétique de la météorite Paris que nous avons réalisée s’intègre dans une étude plus générale (minéralogique, chimique, isotopique et pétrologique) réalisée sur cette météorite sous la direction du Muséum National d’Histoire Naturelle de Paris (MNHN) et qui a donné lieu à la p ublication d’un article (Hewins et al., 2014-Annexe-2). Les CM utilisées pour notre étude ont été choisis afin de couvrir la gamme de degré d’altération hydrothermale la plus large possible, les plus altérées étant Nogoya et Cold Bokkeveled, et les moins altérées Murchison and Paris (Zolensky et al., 1993, 1997 ; Browning et al., 1996 ; Rubin et al., 2007 ; Chizmadia and Brearley, 2008 ; Howard et al. 2011 ; Hewins et al., 2014).

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Par rapport aux études précédentes sur les CM, qui sont restées relativement préliminaires, nous avons étudié des sous-échantillons mutuellement orientés, à la fois via les désaimantations sous champs alternatifs et thermiques. Nous avons également évalué l’aimantation rémanente visqueuse (VRM) et l’aimantation rémanente de choc ( SRM) de chaque météorite. Notre but était de caractériser la minéralogie magnétique des différents échantillons, de déterminer la nature et l’âge de la NRM et enfin d’estimer l’intensité du paléochamp afin de pouvoir discuter son origine. Les mesures des propriétés magnétiques de ces sept chondrites CM confirment que l’assemblage minéralogique est composé d’un mélange de fer métal, de pyrrhotite et de magnétite en proportion variable. Dans la météorite Paris certains échantillons, dans les lithologies décrites comme les moins altérées par l’hydrothermalisme , ne contiennent pas de magnétite mais sont beaucoup plus riches en FeNi métallique comparés aux autres CM.

Les données paléomagnétiques mettent en évidence que toutes ces chondrites CM possèdent au moins une composante d’aimantation stable. Les composantes de basse coercivité sont interprétées comme une aimantation visqueuse ou liée à de la contamination (acquise lors de la conservation, de la manipulation ou de la préparation des échantillons). Paris est ainsi un très bel exemple de contamination par des aimants ! Malgré cela, nous avons réussi à obtenir des résultats relativement concluants sur la gamme de haute coercivité non affectée par cette contamination. Les composantes haute coercivité (HC) et haute température (HT) (isolées respectivement via les désaimantations AF et thermique) des échantillons mutuellement orientés sont homogènes en direction et en intensité (Fig 4.2 ). Ceci indique que l’aimantation a été acquise après l’accrétion du corps pare nt. Dans l’article nous discutons la nature de ces composantes (HC et HT) et nous interprétons finalement cette aimantation comme une CRM. Les températures de déblocage indiquent que la CRM est principalement portée par la pyrrhotite et par la magnétite dans une moindre mesure. Ces deux minéraux se sont formés lors de l’hydrothermalis me sur le corps parent des CM (Choi et al., 1997 ; Hanowski and Brearley, 2001 ; Zolensky and Le, 2003 ; Bullock et al., 2007 ). Cet hydrothermalisme a commencé peu après l’accr étion (~2.5 ± 0.1 Ma, Pravdivtseva et al., 2013 ) et a duré au moins 4 Ma (De Leuw et al., 2009 ; Brearley, 2006 ). Grâce à la méthode de normalisation du REM’ nous estimons une paléointensité minimum de quelques µT (2 ± 1.5 μT).

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig. 4.2. Projection stéréographique des directions des composantes HC et HT obtenues pour les différentes chondrites CM étudiées dans l’article Cournède et al., en préparation pour EPSL. Mighei (gris), Murchison (rouge), Murray (bleu), Nogoya (blanc) et Paris (noir). Le remplissage ou non des symboles correspond respectivement aux projections dans l’hémisphère supérieur ou inférieur.

Au vu de l’intervalle de temps pendant lequel cette aimantation a pu être acquise, ce paléochamp peut être un champ d’origine interne ( champ astéroïdal de dynamo) ou externe (champs de disque, solaires ou nébulaires) compte tenu que les deux types de champ pouvaient coexister à ce moment-là (Kleine et al., 2002 ; Balbus, 2009 ).

En résumé, nous interprétons l’aimantation contenue dans les c hondrites CM comme une CRM pré-terrestre acquise au cours de la cristallisation de la magnétite et de la pyrrhotite lors de l’altération aqueuse dans un champ d’au moins quelque µT (2 ± 1.5 μT). Le champ magnétisant peut-être d’origine interne ou externe. Il est impossible de discriminer ces deux hypothèses. En dépit de cette dernière considération, nous retiendrons que les chondrites CM, reliques du système solaire primitif, portent le plus vieil enregistrement paléomagnétique jamais mis en évidence à ce jour.

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Article en préparation pour EPSL

An Early solar system magnetic field recorded in CM chondrites

Cécile Cournède ¹, Jérôme Gattacceca ¹, Pierre Rochette ¹, Benjamin P. Weiss², Brigitte Zanda 3 ¹ CEREGE, CNRS/Université Aix-Marseille, UM34, Aix-en-Provence, France ² Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge MA 02139, USA 3 LMCM, Museum national d'Histoire naturelle, CP52, 61 rue Buffon, 75005 Paris, France

1. Introduction

The study of the remanent magnetization (paleomagnetism) of extra-terrestrial materials gives clues as to the history of the primitive solar system and its evolution. Indeed, paleomagnetic studies of meteorites provide a unique window into understanding early solar magnetic fields generated externally from planetesimal bodies (e.g., within the proto-planetary nebula), as well as dynamo magnetic fields generated within the planetesimals through convection of a conducting core (e.g., Weiss et al., 2010 ). In this study we focused on CM2 carbonaceous chondrites. This meteorite group is of particular interest because CM2 chondrites are believed to be some of the most primitive material available in our solar system ( Anders, 1971 ; Mason, 1962 ; Wood 1963 ). CM2 may represent the transition between asteroids and comets and may have formed in the outer solar system ( Wasson and Wetherill , 1979 ; Gounelle et al., 2008 ). As such, they offer the possibility to estimate the magnetic fields strength present in the outer early solar system.

1.1 CM chondrites petrography and petrogenesis CM chondrites consist of chondrules, chondrule and mineral fragments, and calcium- aluminium inclusions (CAIs), embedded in a fine-grained phyllosilicate-rich matrix. CAIs and chondrules were formed early in the solar nebula. CAIs, considered as the oldest solids in the solar system, have been dated at 4567.2 ± 0.6 Ma in CV chondrites ( Amelin et al., 2002 ). Chondrule formation appears to have started almost contemporaneously and lasted ~3 Ma (Connelly et al., 2012 ). Shortly after, dust, CAIs and chondrules accreted into small planetesimals that ultimately formed the CM parent asteroid. Following accretion, aqueous alteration took place

CHAPITRE 4 : MAGNETISME DES CHONDRITES

through the action of ice melting. This process generated the main alteration mineral found in CM chondrites such as phyllosilicates, sulfides and iron oxides ( Rosenberg et al., 2001 ). Thermometry of carbonates in Cold Bokkeveld, Murray, and Murchison, demonstrates that they have undergone aqueous alteration at low temperature in the 20 –71 °C range ( Guo and Eiler, 2007 ). Following aqueous alteration, a few CM chondrites have undergone thermal metamorphism ( Miyamoto, 1991 ; Kitajima et al., 2002 ; Nakamura, 2005 ). However, most CM chondrites (and in particular those studied here) escaped significant heating on their parent body (Kimura et al., 2011 ). For instance, Murchison has been heated to a maximum temperature of ~110°C ( Cody et al., 2008 ). Most CM chondrites are brecciated. The combination of brecciation and aqueous alteration has resulted in a remarkable textural diversity and is a complication that makes unraveling the history of these meteorites particularly challenging ( Brearley, 1995; Lindgren et al., 2013 ). A range of degrees of aqueous alteration has been evidenced in CM2 chondrites based on various criteria, e.g.; modal mineralogy, alteration in chondrules, abundance of opaque minerals with the most altered being Nogoya and Cold Bokkeveled, and the least altered being Murchison and Paris (Zolensky et al., 1993, 1997 ; Browning et al., 1996 ; Rubin et al., 2007 ; Chizmadia and Brearley, 2008; Howard et al. 2011; Hewins et al., 2014). A range of aqueous alteration is also sometimes observed within individual meteorites from a single meteorite fall ( Rubin and Wasson, 1986 ; Alexander et al, 2010 ; Zanda et al., 2010a ; Jenniskens et al., 2012 ). The generally accepted view is that aqueous alteration of CM chondrites was a post-acretionnary asteroidal process (e.g. McSween, 1979 ; Tomeoka et al., 1989 ; Hanowski and Brearley, 2000, 2001 ). However aqueous alteration in small precursor planetisimals prior to the formation of the CM parent asteroid is also advocated for some CM chondrites ( Metzler et al., 1992 ; Bischoff, 1998 ; Lauretta et al., 2000 ). The chronology of alteration in CM chondrites is constrained by 53 Mn/ 55 Mn measurement on carbonates, indicating that accretion started contemporaneously with or shortly after CAI formation, and lasted at least 4 Myr ( Lewis and Anders, 1975 ; Niemeyer and Zaikowski, 1980 ; De Leuw et al., 2009 ). Longer alteration duration of ~7.5 ± 2 Myr have been proposed ( Brearley, 2006 ). I-Xe age of 2.4±0.1 Ma after CAI for Murchison magnetite confirms an early onset for aqueous alteration ( Pravdivtseva et al., 2013 ). Ejection from the CM parent body took place ~2 Ma ago (e.g. Eugster et al., 1998 ). All CM chondrites show no petrologic evidence of shock ( Scott et al., 1992 ), indicating maximum shock pressure of 5 GPa, and likely much less in view of their high porosity ( Britt and Consomagno, 2003 ).

1.2 CM magnetic properties: State of the Art

CHAPITRE 4 : MAGNETISME DES CHONDRITES

The magnetic properties of a dozen CM2 chondrites have been studied previously (see review in Weiss et al., 2010 ; Elmaleh et al., 2012 ). Although all CM2 chondrites are chemically similar, considerable differences exist in their mineral constituents, as illustrated by their magnetic phases (Hyman and Rowe, 1983 ). The magnetic mineralogy of CM chondrites is composed, in various proportions, of metallic iron (<1.5 wt.%), magnetite (<3 wt.%), and iron-nickel sulfides (0.7-1.3 wt%) ( Hyman and Rowe, 1986 ; Rochette et al., 2008 ; Howard et al., 2009 ; Burgess et al., 1991 ). Magnetic susceptibility of CM chondrites varies by nearly two orders of magnitude, indicating large variation in the magnetic minerals assemblage and content ( Rochette et al., 2008 ). Magnetite is a secondary mineral probably formed on the parent asteroid by oxidation of metal, as in CV chondrites ( Choi et al., 1997 ). The magnetite content in CM chondrites appears to be positively correlated to the degree of alteration of the meteorite (Petitat and Gounelle, 2010 ). Pyrrhotite origin is still controversial: it could be a secondary mineral formed by replacement of troilite and/or metal ( Hanowski and Brearley, 2001 ; Zolensky and Le, 2003 ; Bullock et al., 2007 ) or could have formed in the solar nebula rather than on the parent body ( Nazarov, 1994, 1996, and 1997 ; Harries and Langenhorst, 2013 ). CM chondrites can also contain metal ( Hewins et al., 2014 ) although it is usually very rare. Rare schreibersite is also observed in some CM chondrites (Nazarov et al., 2009 ). The paleomagnetism of CM2 chondrites has also been previously investigated (see Weiss et al., 2010 for a review). These meteorites possess a measurable natural remanent magnetization (NRM) (Banerjee and Hargraves, 1971 ). Crude partial alternating field (AF) demagnetization of Cold Bokkeveld, Mighei, Murray, and Murchison showed that this NRM is stable ( Larson et al., 1973 ). This NRM was interpreted as a thermal remanent magnetization (TRM), thermo-chemical remanent magnetization (TCRM), or chemical remanent magnetization (CRM) acquired prior to falling on Earth ( Larson et al., 1973; Banerjee and Hargraves, 1972 ). In Murchison, a minimum paleointensity of 20 µT was estimated with the Thellier-Thellier method (Banerjee and Hargraves, 1972 ). Much lower values of 0.2-2 µT were determined by non-heating methods ( Kletetschka et al., 2003 ). From these previous studies, no clear picture regarding the nature of the NRM, and the nature and intensity of the magnetizing field has emerged. Herein, we performed a detailed and comparative magnetic and paleomagnetic study of seven CM carbonaceous chondrites: Banten, Cold Bokkeveld, Paris, Murchison, Murray, Mighei, Nogoya (all falls except Paris that is a very fresh find). Compared to previous studies, we studied mutually-oriented sub-samples, used both AF and thermal demagnetization, and evaluated the viscous remanent magnetization (VRM) and shock remanent magnetization (SRM) of each meteorite. Our aim was to characterize their magnetic mineralogy, to try and correlate this

CHAPITRE 4 : MAGNETISME DES CHONDRITES

mineralogy with alteration degree, determine the nature and age of NRM, and estimate of the paleofield intensity.

2. Samples and methods

2.1. Samples Samples of seven chondrites were supplied by the National Museum of Natural History of Paris (list in Supplemental Table 4.A). All studied meteorites are classified as CM2 carbonaceous chondrite except Paris that is of petrologic type 3 ( Zanda et al., 2010b ). All samples were stored in a magnetically shielded room (field < 400 nT) during at least two weeks after receiving them in order to allow partial decay of the viscous remanent magnetization acquired during exposure of the meteorites to the geomagnetic field since their fall. As the original orientation of samples on the parent body is unknown, sample orientation was chosen arbitrarily. Samples were cut in mutually-oriented sub-samples using a wire saw.

2.2. Magnetic measurements Hysteresis measurements were performed at room temperature with a Princeton Micromag Vibrating Sample Magnetometer (VSM), with a maximum field of 1 T, and a moment sensitivity of 10 −9 Am 2. The analysis of hysteresis loops provided the ratio of saturation remanent magnetization (M rs ) to saturation magnetization (M s) and the coercive force (B c). M rs values were measured with a 2 G cryogenic magnetometer (see below) after saturation in a field of 3 T using a

MMTD pulse magnetizer. Ms values are recalculated from the M rs /M s ratio given by the VSM and the M rs value from the 2 G magnetometer, to account for the sample shape dependence of the calibration of the VSM. Remanent coercive force (Bcr ) was determined by DC back field experiments performed with the VSM. Low temperature (40 K

CHAPITRE 4 : MAGNETISME DES CHONDRITES

susceptibility (AMS) was characterized by the shape parameter T ( Jelinek, 1981 ), varying from −1 (prolate) to +1 (oblate), and the anisotropy degree P (ratio of maximum to minimum susceptibility). High field susceptibility ( χhf ) was determined by linear fitting of the 0.9-1 T field interval of the hysteresis loops. All remanence measurements were performed with a SQUID cryogenic magnetometer (2G Enterprises, model 755R, with noise level of 10 −11 Am 2) with an attached automatic alternating field 3-axis demagnetization system (maximum peak field 170 mT) placed in a magnetically shielded room. Thermal demagnetization was performed using an MMTD furnace, under argon atmosphere above 250 °C. For most samples, we measured the S ratio that is the isothermal remanent magnetization (IRM) obtained after applying a 3 T field and then a back field of − 0.3 T normalized to the IRM acquired in 3 T. IRM thermal demagnetization was done after imparting two orthogonal IRMs in fields of 3 T and 0.3 T after Lowrie (1990) . The NRM of all samples and its stability against stepwise AF demagnetization up to 120 mT were measured. At each AF step the sample was demagnetized and measured at least three times in order to reduce spurious anhysteretic effects and measurement noise. After the study of NRM all samples were given an anhysteretic remanent magnetization (ARM) that was subsequently stepwise AF demagnetized. Piezo-remanent magnetization (PRM), used as an analogue for SRM, was imparted through hydrostatic loading and unloading in the presence of a controlled magnetic field using a non-magnetic pressure cell ( Gattacceca et al., 2010 ). Partial thermoremanent magnetizations (pTRM) were acquired using the MMTD furnace equipped with a coil connected to a stabilized DC power supply. Saturation isothermal remanence magnetization (SIRM, acquired in 3 T field with a pulse magnetizer) was also investigated for all samples. The rates of acquisition and decay of viscous magnetization of the samples were estimated experimentally. The acquisition rate was monitored over a period of one month by periodic measurements of the VRM acquired in a 100 µT field. The decay rate was then measured by periodic measurements with the sample kept in a sub-null (~50 nT) ambient field. Both rates were found to be approximately the same ( Supplemental Table 4.B) as classically observed ( Enkin and Dunlop, 1988 ).

All magnetic measurements were performed at CEREGE (Aix-en-Provence, France), with the exception of MPMS measurements (at IPGP, Paris, France).

3. Rock magnetism

3.1. Hysteresis properties

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Hysteresis properties show large variability among the studied CM chondrites, indicating a

-1 variable magnetic mineral assemblages and contents ( Table 4.1). M s varies between 0.52 Am² kg -1 -2 -1 (Murchison) and 5.69 Am².kg (Murray). M rs varies between 3.58 x 10 (Mighei) and 1.17 x 10 Am² kg -1 (Murray). The hysteresis parameters indicate an overall MD (multi-domain) to PSD - (pseudo-single domain) behavior with a mean M rs /M s of 8.8 x 10 ² ( Fig. 4.3 ). The S ratio varies between -0.83 (Mighei) and -0.93 (Nogoya) ( Table 4.1), indicating the significant presence of a high coercivity mineral. Curvature of hysteresis cycles up to 1T suggests a noticeable contribution of metallic iron (or possibly schreibersite) except in Cold Bokkeveld ( Supplemental Fig. 4.A).

Fig. 4.3. Mrs /M s versus B cr /B c for seven CM chondrites studied in this work. Banten (box), Cold Bokkeveld (triangles), Mighei (grey circle), Murchison (diamond), Murray (star), Nogoya (solid circle), Paris (open circles). The limits for the single domain (SD), pseudo single domain (PSD), and, multi-domain (MD) are indicated.

3.2 Low temperature measurements Low temperature remanence measurements performed on Paris and Murchison meteorites show the presence of a Verwey transition around 120 K ( Supplemental Fig. 4.B) as already observed by Elmaleh et al. (2012) . Monitoring of magnetic susceptibility at low temperature performed on Cold Bokkeveld, Mighei, Nogoya, and Paris confirm the existence of a more or less pronounced Verwey transition in all samples but one (the less aqueously altered lithology of Paris meteorite) ( Supplemental Fig 4.C. a ). This indicates the presence of stoichiometric

CHAPITRE 4 : MAGNETISME DES CHONDRITES

magnetite with variable abundance at the scale of the measured samples (30-150 mg). A faint transition around 40 K, is tentatively identified in Murchison ( Supplemental Fig 4.B. e ) and could correspond to the phase transition of monoclinic pyrrhotite at 34 K (Rochette et al., 1990 ).

S mass M (A m 2 M (A m 2 B B Sample rs s cr c logχ log χ P P T P T ratio (g) -1 -1 (mT) (mT) p asm f asm rem rem kg ) kg ) (%) Banten 0.043 1.02E-01 2.15 89 10 4.03 2.81 1.10 1.11 0.08 1.10 0.15 - Cold Bokkeveld 0.028 6.08E-02 1.53 52 10 3.67 2.91 1.09 1.10 0.26 1.20 0.58 82.7 Mighei 0.249 3.58E-02 0.71 58 19 3.36 2.40 1.05 1.05 0.70 1.13 0.78 90.4 Murchison 0.7658.86E-02 0.52 61 23 3.512.53- - - - - 88.4 Murray 0.273 1.17E-01 5.69 76 6 4.22 3.01 1.08 1.08 -0.16 1.10 0.44 80.6 Nogoya 0.520 8.74E-02 1.05 83 18 3.60 2.45 1.10 1.10 0.41 1.22 0.71 93.0

Paris 1.839 1.24E-01 4.60 66 7 4.282.711.121.130.48 - - ~83 Table 4.1 : Average intrinsic magnetic properties of studied CM chondrites. Results for individual samples are given in details in Supplemental table 4.A. M rs : saturation remanence (acquired in 3T field), M s: saturation magnetization, Bc: coercivity, B cr : coercivity of remanence, χ: low field magnetic susceptibility, χp: paramagnetic susceptibility. All −9 3 −1 susceptibilities are in 10 m kg , P asm , T asm : anisotropy degree and shape parameter for magnetic susceptibility, P f: anisotropy degree of ferromagnetic susceptibility. P rem , T rem: anisotropy degree and shape parameter of anhysteretic remanence, S-ratio (computed as IRM 0.3T /SIRM). Parameters that were not measured or not available are indicated by: -.

3.3. Magnetic susceptibility For commodity magnetic susceptibility (χ) is expressed in log units (log χ, with χ in 10 −9 m3 kg −1 , Table 4.1). Log χ ranges from 3.36 in Mighei to 4.28 in Paris. Our results are in overall agreement with values obtained by Rochette et al. (2008) on larger masses of several grams, except for Murray whi ch give log χ =4.22, much higher than the average obtained for this sample, log χ = 3.82±0.22 measured on 7 samples in Rochette et al. (2008) . Our Murray sample may be an unusual magnetite-rich clast, or may be a mislabeled or misidentified sample although it is clearly a CM chondrite. The relatively large scatter in our susceptibility measurements, indicated by the standard deviations provided in Table 4.1 indicates that magnetic grains are not homogeneously distributed at the scale of our samples (16 mg to 15.5 g).

In all samples paramagnetic susceptibility (noted χ p) can be approximated by high-field susceptibility (χ hf ) because the diamagnetic and antiferromagnetic contributions are negligible (see discussion in Gattacceca et al., 2008 ). The results show logχ p rang ing from 2.40 to 3.01 (log χ p -9 3 -1 with χ p in 10 m kg ). Paramagnetic minerals contribute only to 2 to 14 % to the total susceptibility in all CM studied here. High temperature measurements of magnetic susceptibility under argon atmosphere for four CM chondrites ( Supplemental Fig 4.C. b ) show a Curie temperature at ~580°C, corresponding to magnetite. In all samples a faint downward inflexion around 225°C is observed. It may be attributed to the Curie temperature of schreibersite. Indeed schreibersite has a Curie temperature

CHAPITRE 4 : MAGNETISME DES CHONDRITES

varying linearly with Ni atomic content between 443°C for Fe 3P and -153°C for Fe 0.75 Ni 2.25 P (Gambino et al., 1967 ). For the range of composition observed in CM chondrites, i.e. with Fe/Ni atomic ratio between 0.41 and 4.44 ( Nazarov et al., 2009 ) the expected Curie temperature is in the -110 to 300°C range . However, schreibersite is found in very small amount in the studied CM chondrites, with a volume content in the 0 to 14 ppm range and an average content of 5 ppm (Nazarov et al., 2009 ) .Its contribution should be below the detection limit of the susceptibilitimeter used for this experiment. We have no satisfactory explanation for this inflexion at 225°C. In Paris, two major differences with others CM chondrites can be reported: the small inflexion around 450°C and the residual magnetic susceptibility above 585 °C. The second feature is due to metallic FeNi. The inflexion at 450°C may be attributed to schreibersite with a composition close to Fe 3P.

3.4. Remanent magnetizations We have investigated the ARM, PRM, SIRM and VRM of all studied samples. Their intensities are reported in Table 4.2. PRM intensity increases with pressure up to 2 GPa (the maximum pressure available in our experiments), and the coercivity spectrum is shifted towards higher coercivity values with increasing pressure as classically observed (e.g. Gattacceca et al., 2010 ). More details about these PRM experiments can be found in Tikoo et al., 2013 . Median destructive field (MDF) is used to describe the stability of SIRM and ARM against AF demagnetization ( Table 4.2). MDFs of SIRM range from 37 (Cold Bokkeveld) to 70 mT (Banten) (mean value 50 mT) ( Fig. 4.4 ). The ARM acquired in 100 to 160 mT AF exhibits a broad range of stability with MDFs ranging from 29 (Cold Bokkeveld and Nogoya) to 56 mT (Murchison). PRM acquired in 2 GPa show a more restricted and lower range of coercivity with MDF in the range 2- 15 mT.

MDF of MDF of PRM MDF of M (A m 2 ARM (A m² Sample mass (g) rs SIRM ARM @2GPa PRM -1 -1 -1 kg ) kg µT ) -1 (mT) (mT) (Am² kg ) (mT) Banten 0.043 1.02E-01 70 7.55E-06 * 38 - - Cold Bokkeveld 0.028 6.08E-02 37 4.06E-06 * 29 4.84E-04 11 Nogoya 0.520 8.74E-02 61 3.86E-06 * 29 3.89E-04 14 Mighei 0.249 3.58E-02 40 2.91E-06 * 34 1.37E-04 16 Murray 0.273 1.17E-01 59 9.40E-06 * 31 1.08E-03 3 Murchison 0.765 8.86E-02 38 8.70E-06 + 56 4.18E-04 15

Paris 1.839 1.24E-01 47 1.43E-05 + 45 9.23E-04 11 Table 4.2 : Title: Remanent magnetizations of the studied CM chondrites. ARM: anhysteretic remanent magnetization was acquired in an AF of 100 mT (*), of 160 mT (+). PRM: piezo-remanent magnetization was acquired in 2GPa in a field of 700 µT. SIRM intensities are given in Table 4.1 (M rs ).

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig. 4.4. Normalized intensity of SIRM versus AF for studied CM chondrites.

Thermal demagnetization of the SIRM ( Fig. 4.5 ) shows variable behavior depending on temperature. With the exception of Murchison, all CM chondrites are demagnetized at low temperature with less than 20% of the magnetization remaining at 300 °C. This suggests that sulfides dominate the remanence signal, schreibersite being ruled out in view of its low content (see above). In Mighei and Nogoya thermal demagnetization of IRM reveals two inflections around 270 and 320°C characteristic of hexagonal and monoclinic pyrrhotite respectively. However these Curie points are not well expressed in general and the range of unblocking mainly below 200°C suggest complexity in the magnetic mineralogy, possibly linked with Ni substitution and mineral defects in pyrrhotites. Murchison stands out with a higher resistance to heating. Metallic iron was detected in Murchison and Cold Bokkeveld with a drop at ~ 770 °C during measurements of induced magnetization versus temperature ( Banerjee and Hargraves, 1971, 1972 ). Our experiments show that this metallic iron does not carry any significant remanence compared to pyrrhotite and magnetite. The measurement of VRM acquisition and decay rates allows the computation of the maximum VRM (VRM max ) that may have been acquired during the terrestrial residence of the samples ( Supplemental Table 4.B). This VRM max is computed for the unfavorable case where samples were kept in a fixed position in a terrestrial magnetic field of 50 µT (during two weeks), and taking into account the decay of the VRM in the shielded room. The results show that VRM contributes for less than 10% of the measured NRM in Murray and Cold Bokkeveld, in agreement with previous results (Brecher and Arrhenius, 1974 ). For Mighei and Nogoya VRM max 95

CHAPITRE 4 : MAGNETISME DES CHONDRITES

amounts to 99 and 61 % of the NRM, respectively. In view of the terrestrial residence time of the studied meteorites (50 to ~200 years), the terrestrial VRM is expected to be stable up to 80- 120°C upon laboratory thermal demagnetization whether it is carried by pyrrhotite or magnetite (Pullaiha et al., 1975 ; Dunlop et al., 2000 ).

Fig. 4.5. Normalized intensities of NRM (black line) and SIRM (gray line) versus demagnetization temperature for studied CM chondrites. For Murchison, the sample was lost during the SIRM experiment.

CHAPITRE 4 : MAGNETISME DES CHONDRITES

3.5 Magnetic anisotropy Magnetic anisotropy measurements are presented in Table 4.1. Anisotropy degrees are weak with P ranging from 1.05 (Mighei) to 1.10 (Nogoya and Banten) (mean value 1.08) in agreement with previous measurements that found mean P of ~1.07 by measuring larger samples in the 1- 100 g range (our unpublished data ). Paris meteorite is the most anisotropic, with P=1.14, a difference that may be attributed to the higher content of metallic iron. The shapes of the susceptibility ellipsoids are neutral to oblate. Mutually oriented samples have similar orientation of the anisotropy axes ( Fig. 4.6 ), indicating an homogeneous fabric at the scale of about 1 cm (initial size of the largest studied samples). These results are confirmed by the similar properties of anisotropy of ARM: P rem range from 1.10 to 1.22 with an oblate shape parameter and there is a directional agreement between AMS and AARM ellipsoid axes. Even if the origin of the magnetic fabric in unshocked carbonaceous chondrites is not fully understood (see discussion in Gattacceca et al., 2005 ), the observed homogeneity of the fabric in our CM samples suggests that brecciation took place before crystallization of magnetite and pyrrhotite during aqueous alteration, or that the studied samples were smaller than the clast size. In both cases, the paleomagnetic record of our samples is likely not blurred by brecciation.

Fig. 4.6. Maximum (squares) and minimum (circles) magnetic susceptibility and anisotropy of remanence axes (indicated by “R”) for studied CM chondrites (projection on lower hemisphere). Banten (dotted circles), Cold Bokkeveld (symbols with vertical stripes), Mighei (gray symbols), Murray (symbols with horizontal stripes), Nogoya (solid symbols), Paris (open symbols).

CHAPITRE 4 : MAGNETISME DES CHONDRITES

4. Paleomagnetic results

AF demagnetizations up to 170 mT were performed at least on one sample of each of the six studied CM chondrites. Thermal demagnetizations were also conducted after removal of the low coercivity (LC) component with AF. The directions of the components were computed using the principal component analysis ( Kirschvink, 1980 ), without anchoring to the origin. The results are listed in Table 4.3, and displayed in Fig. 4.7, 4.8 and 4.9 . For Cold Bokkeveld, one sample with fusion crust and four interior mutually-oriented samples were demagnetized with AF. NRM intensity averages 1.2 ± 0.4 x 10 -4 Am².kg -1 in the four interior samples and reaches 2.5 x 10 -3 Am².kg -1 for the sample with fusion crust. The crust sample possesses two stable components of magnetization (Fig. 4.7 ): a low coercivity component (1-6 mT) and a high coercivity component that is interpreted as a TRM acquired in the Earth field during atmospheric entry. In all four interior samples the demagnetization data reveal a single component trending towards the origin and fully demagnetized by 20 mT, with an erratic behavior above this AF level. This magnetization has identical direction in all sub-samples, but different from both directions found in the fusion crust (Fig. 4.7 ). The erratic behavior observed above 20 mT is in agreement with previous measurements showing that the NRM of Cold Bokkeveld is essentially demagnetized by 30 mT ( Brecher and Arrhenius, 1974 ). Thermal demagnetization was performed on one Cold Bokkeveld sample, following AF demagnetization up to 10 mT. A single component of magnetization was isolated below 175 °C. This origin- trending component is similar in direction to components isolated by AF in the other samples (Fig. 4.8 ). For Mighei, two mutually-oriented samples were studied, one with AF demagnetization and one with thermal demagnetization. NRM intensity is homogeneous and averages 1.35 x 10 -5 Am².kg -1. AF demagnetization reveals a LC component (0-8 mT) and a well-defined origin- trending high-coercivity (HC) component (11-105 mT) (Fig. 4.7 ). After removal of the LC component by AF, thermal demagnetization reveals an origin-trending component isolated between 100 and 290°C ( Fig. 4.7 ). This direction is similar to the HC direction ( Fig. 4.8 ). The fairly unstable thermal demagnetization data indicate that mineralogical transformations probably occurred with heating. This is confirmed by the significant changes in the hysteresis properties measured after heating ( Supplemental Table 4.C). For Murchison, two mutually-oriented samples were studied. Both AF and thermal demagnetization evidence two well-defined components of magnetization ( Fig. 4.7 ). The HC and HT components, isolated up to 100 mT and 520 °C respectively, have similar directions ( Fig. 4.8 ). 98

CHAPITRE 4 : MAGNETISME DES CHONDRITES

AF/thermal NRM Paleofield Sample Treatment Total REM range (mT D I MAD n REM' (Am².kg -1 ) (µT) or °C) Banten AF+TH 7.72E-04 LC/LT 1 mT-80°C 267.6 -31.8 12.1 5 HT 120-460°C 194.4 -66.7 4.3 16 Cold Bokkeveld + AF 2.51E-03 5.59E-03 LC 1-7 mT 88.2 -68.9 5.5 10 2.08E-02 62 HC 9-110 mT 18 10.7 0.940 6.59E-03 20 Cold Bokkeveld AF 6.22E-05 1.34E-03 LC 6-15 mT 71.4 0.3 4.0 9 2.59E-03 8 AF 2.16E-04 2.88E-03 LC 6-15 mT 73.4 11.5 2.4 10 6.39E-03 19 AF 7.82E-05 1.87E-03 LC 7-17 mT 77.5 -2.2 7.1 13 3.83E-03 12 AF 1.31E-04 1.75E-03 LC 5-22 mT 69.6 -4.1 6.4 17 2.43E-03 7 AF+TH 1.13E-04 LC/LT 2 mT-175°C 79.1 3.6 2.2 17 Mighei AF 1.55E-05 4.38E-04 LC 0-8 mT 295.8 -41.7 5.4 13 3.67E-03 11 HC 11-105 mT 175.8 26.1 2.0 32 3.11E-04 1 AF+TH 1.14E-05 LC 0-5 mT 310.1 -29.2 3.5 9 HT 100-290 °C 190.7 27.6 10.6 11 Murchison AF 3.44E-05 3.89E-04 LC 1-5 mT 52.3 -30.3 1.9 7 4.87E-03 15 HC 20-100 mT 193.1 -27.4 11.2 29 3.44E-04 1 AF+TH 2.84E-04 LC 1-5 mT 353.7 -20.6 4.1 7 HT 200-520 °C 186.0 -40.3 12.5 8 Murray AF 2.15E-04 1.84E-03 LC 0-4 mT 313.9 -18.4 5.1 9 2.08E-02 63 HC 4-110 mT 262.6 -51.2 2.1 46 7.95E-04 2 AF+TH 1.85E-04 LC 0-4.5 mT 304.7 -12.4 3.8 7 HT 5 mT-270 °C 249.8 -28.8 6.2 13 AF+TH 4.75E-04 LC 1-4 mT 318.6 -34.2 3.3 5 HT 80-250 °C 242.8 -34.4 11.3 6 Nogoya AF 2.50E-05 2.76E-04 LC 0-6 mT 192.3 -23.9 7.5 11 1.84E-03 6 HC 8-78 mT 310.3 -2 2.2 35 1.95E-04 1 AF+TH 1.50E-05 LC 0-3.5 mT 195.0 -9.5 13.3 7 HT 150-290 °C 284.4 -22 10 9 Paris AF 8.80E-04 4.50E-03 LC 2-6 mT 147 28 1.2 7 4.01E-02 120 HC 14-56 mT 167.2 75.5 2.4 11 5.77E-04 2 AF 1.60E-03 1.27E-02 LC 1-4 mT 197.8 0.2 2 7 1.42E-01 426 MC 20-32 mT 51.2 35.6 2.5 7 1.84E-02 55 HC 44-100 mT 89.3 48.1 8.2 10 1.76E-03 5 AF 1.36E-03 1.01E-02 LC 2-10 mT 159.5 25.7 2.2 5 5.38E-02 161 HC 24-90 mT 201.9 77.6 3.7 21 6.15E-04 2 AF 1.55E-03 1.64E-02 LC 0-2 mT 182.8 24.4 1.4 4 3.31E-02 99 MC 11-20 mT 93.7 29.6 2.5 7 5.51E-02 165 HC 26-110 mT 239 83.4 2.1 27 1.32E-03 4 AF 2.29E-04 2.44E-03 LC 2-7 mT 26.4 7 7.6 7 7.34E-03 22 HC 11-58 mT 294.8 75.3 6.4 13 6.29E-04 2 AF+TH 1.25E-04 LC 7mT-150°C 199.7 13.3 14.1 8 HT 175-585°C 32.4 64.4 12.9 11 Table 4.3 : Paleomagnetic interpretation obtained from the studied CM chondrites. Demagnetization treatment used; AF (alternating field) and TH (thermal). Total REM=NRM/SIRM, LC= low coercivity component, HC= high coercivity component, LT= low temperature, HT=high temperature. RE M’: REM’ integrated over the given AF stability range (with associated paleofield estimate). For each component defined inclination, declination, maximum angular deviation (MAD) and the number of points used are given (n).

For Murray three mutually-oriented samples were studied. NRM intensities are homogeneous. AF demagnetization of one sample reveals two components isolated between 0-4 mT and 4-110 mT ( Fig. 4.7 ). AF demagnetization of the two other samples up to ~6 mT reveals a

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig. 4.7. Orthogonal projection AF and AF+thermal demagnetization data for the studied CM chondrites.

100

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig 4.8. Stereographic projection of the stable component of magnetization (and their 95% confidence interval) for the CM chondrites where multiple mutually-oriented samples were studied. Open and solid symbols are projection in the upper and low hemisphere respectively. Boxes: LC components, Circles: HC/HT components (ChRM).

101

CHAPITRE 4 : MAGNETISME DES CHONDRITES

similar LC component. In these samples, stable components are isolated by thermal demagnetization up to 270°C and 250°C respectively, and have the same direction as the HC component (Fig. 4.8 ). For Nogoya, two mutually-oriented samples were studied. After removal of a LC component below ~5 mT, a stable HC origin-trending component was isolated up to 78 mT ( Fig. 4.7 ). Thermal demagnetization reveals a stable origin-trending component between 150 and 290°C (Fig. 4.7 ) with the same direction as the HC components ( Fig. 4.8 ). For Banten, a single sample was investigated through thermal demagnetization. It reveals a LC component (1 mT- 80°C) and an origin trending HC component between 120 and 460°C (Fig. 4.7 ). At higher temperature, demagnetization behaviour becomes unstable probably due to sample alteration that occurred with heating. For Paris six mutually-oriented samples were studied. One sample was thermally demagnetized and five samples were demagnetized with AF. All samples show a LC components (sometimes two) isolated below ~ 12 mT, and a HC component isolated above ~25 mT and up to 85 mT ( Fig. 4.7 ). The LC components lie on a great circle whereas the HC components are clustered ( Fig. 4.8 ). Thermal demagnetization of a sample previously demagnetized by AF up to 10 mT also reveals two components of magnetization: a LC and low temperature (LT) component, isolated up to 150°C, and laying on the same great circle as the LC components, and high temperature (HT) component isolated between 175 and 585°C ( Fig. 4.7 ) that has a similar direction to the HC components ( Fig. 4.8 ). Most demagnetization data show a curved shape and high REM’ above 0.05 up to AF 20 mT ( Table 4.3). This is indicative of contamination by a strong field typical of artificial magnets.

5. Discussion

5.1 Nature of the NRM We have identified two stable components of magnetization in all studied CM chondrites, except in Cold Bokkeveld that shows a single component. These two components were isolated both by AF and thermal demagnetization. The LC component was isolated below ~10 mT, and the HC component between ~20 and ~90 mT. The high temperature component was isolated between ~120 °C and ~300°C (except in Murchison where it is isolated up to 520°C). The directions isolated by AF and thermal demagnetization in mutually-oriented samples are identical, trend towards the origin and are referred to as Characteristic Remanent Magnetization (ChRM) in the following. Overall, our results agree with previous study that found the presence at least of a one

102

CHAPITRE 4 : MAGNETISME DES CHONDRITES

well preserved stable paleoremanence component in Cold Bokkeveld, Mighei, Murray and Murchison meteorites through AF demagnetization up to 30 mT (Larson et al., 1973 ; Brecher and Arrhenius, 1974 ; Kletetschka et al., 2003 ). The LC components of the NRM are likely of viscous origin or result from magnetic contamination during sample curation, handling, and/ or preparation. Accordingly the REM’ method suggest a paleointensity in the 20-60 µT for this component (except in Paris with higher values due to magnet contamination). These terrestrial components are not discussed in further details. The determination of the nature of the ChRM is crucial to their interpretation in terms of paleofield intensity and implication for the evolution of the CM parent body. CM chondrites are brecciated meteorites that have been aqueously altered at low temperatures (< about 70°C, see §1.1), shocked to relatively low pressures (< 5 GPa), and potentially thermally metamorphosed but only to low temperatures (< about 100 °C). As such, a number of mechanisms can be proposed to account for their NRM. The main magnetic minerals in CM chondrites, magnetite and pyrrhotite, are both secondary phases produced by aqueous alteration at low temperature (Zolensky and McSween, 1988 ; Brearley and Jones, 1998 ; Hyman and Rowe, 1983 ). Because the magnetic minerals are secondary and because the characteristic magnetization is homogeneous in direction at least at the cm scale, a pre-accretional magnetization is ruled out. The secondary nature of the magnetic minerals also rules out acquisition of detrital NRM during accretion on the parent body, a possibility proposed by Fu and Weiss (2012) . The following post-accretion mechanisms have to be considered: chemical remanent magnetization (CRM) by crystallization of magnetite and pyrrhotite during aqueous alteration, thermoremanent (TRM) or partial thermoremanent magnetization (pTRM) during metamorphic heating of the parent asteroid, or shock magnetization (SRM) during impacts suffered by the parent asteroid, thermal magnetization acquired in the Earth field during atmospheric entry, viscous magnetization (VRM) in the Earth magnetic field. Comparison of the intensity of the ChRM with the maximum VRM estimates ( Supplemental Table 4.B) show that VRM cannot account for the ChRM except for Mighei. Moreover the ChRM is unblocked up to about 300°C or above, which is notably higher than the expected stability temperature range for a terrestrial VRM (80-120°C, see §3.4). To further assess the nature of the ChRM we compared the AF demagnetization behavior of ChRM, ARM (analogue for TRM, e.g., Stephenson and Collinson, 1974 ), SIRM and PRM (analogue for SRM) on the high coercivity AF interval defined for each CM chondrite ( Fig. 4.9 ).

103

CHAPITRE 4 : MAGNETISME DES CHONDRITES

The behavior of PRM (acquired at 2 GPa) is always different from the one of ChRM. Moreover if the ChRM was a PRM, it would require a minimum ambient field of ~550 µT (range 120-1100 µT) at the time of impact (mean intensity computed for all five chondrites and extrapolated for a maximum SRM acquisition at 5 GPa (the maximum shock pressure potentially suffered by these meteorites). This estimate is too high to be realistic. Thereby ChRM cannot be a SRM. Because the studied CM chondrites have never been subjected to temperatures above ~110 °C ( Cody et al., 2008 ; Kimura et al., 2011 ), the ChRM (unblocked up to ~300 °C or higher) cannot be a TRM or a pTRM. Moreover, if it was a TRM, the ChRM would have a similar demagnetization behavior as ARM, which is not observed except for Murchison ( Fig. 4.9 ). Therefore the only remaining and most robust possibility is that the ChRM of Banten, Mighei, Murchison, Murray, Nogoya, and Paris is a Chemical Remanent Magnetization (CRM), as already suggested (but not established in details) by Banerjee and Hargraves (1971) and Larson et al. (1973) . This conclusion is in good agreement with the similar thermal demagnetization of ChRM and IRM in these meteorites ( Fig. 4.5 ). In all studied samples the unblocking temperatures mostly below 220 °C indicate that the ChRM is mostly carried by sulfide (pyrrhotite), except in Murchison where results show a major inflexion at 570 °C indicating that in Murchison the ChRM is mostly carried by magnetite. The correlation between the ChRM unblocking temperatures and the magnetic mineralogy is a strong indication that the ChRM was not acquired during a thermal event, but is rather the product of crystallization of the magnetic carriers in the presence of a magnetic field. In Cold Bokkeveld, the ChRM is isolated at low temperatures (below 150°C). VRM is excluded because it cannot account for more than 9% of the measured ChRM. SRM would require a paleofield of more than 100 µT. Our pTRM experiments show that the ChRM scales relatively well with a pTRM acquired at about 120 °C in a field of ~50 µT, similar to the Earth field. Therefore the ChRM could have been acquired during moderate heating during atmospheric entry. The ChRM directions of the two interior samples show a 90° difference with the one determined for the fusion crust, but this can be accounted for by the time lag for the diffusion of the thermal wave associated with atmospheric entry heating thermal wave and the fact that the meteorite is rotating in the atmosphere . It is worth mentioning that even though the NRM is entirely demagnetized at ~20 mT, about 80% of the IRM is preserved at this AF level, which implies that a large part of the magnetic grains are not significantly magnetized in the natural state. Residual REM above 22 mT average 5x10 -4, indicating a paleofield of ~1.5 µT. This suggests that Cold Bokkeveld was in a low field environment when its magnetic grains

104

CHAPITRE 4 : MAGNETISME DES CHONDRITES

crystallized. Therefore, regarding its magnetic behaviour, this meteorite deserves further investigations to understand why it is different from all others CM chondrites studied here.

Fig. 4.9. Normalized intensities of ChRM (black line), ARM (black dotted line), IRM (gray line) and PRM (gray dashed line) versus alternating field for the six CM chondrites demagnetized with AF.

5.2 Paleointensities There is no technique to retrieve precise paleointensities from a CRM. However, CRM is usually a magnetization mechanism that is less efficient than TRM ( Mc Clelland, 1996 ) so that a lower limit for the paleointensity can be estimated. For this purpose, we used the REM’ normalization technique that is calibrated for TRM ( Gattacceca and Rochette, 2004 ), except for 105

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Banten that was not studied with AF. The minimum paleointensities derived from the HC component for Paris, Mighei, Nogoya, Murchsion and Murray are in the range 1 to 5 µT, with a mean value of 2.1 ± 1.5 µT ( Table 4.3). For Murchison our minimum paleointensity estimate of ~1 µT is in broad agreement with a previous estimate in the 0.2 to 2 µT range (Kletetschka et al., 2003 ).

5.3 Magnetizing field Considering that the ChRM is a CRM carried by pyrrhotite and magnetite, the question arises about which magnetic paleofield could account for the remanent magnetization recorded in CM chondrites. This paleofield had a minimum paleointensity of ~2 µT. The CRM nature of the magnetization implies that the magnetizing field must have been stable (with respect to the asteroid) for a period of time approaching that of the aqueous alteration of the CM chondrites (up to several Myr, De Leuw et al., 2009 ; Brearley, 2006 ). Two alternatives exist to account for a non-transient magnetic field of ~2 µT in the early solar system: an external magnetic field (in the accretion disk) or an internal field generated within the parent body by some form of dynamo process. The paleointensity and the age of the magnetization may help discriminating between the two possibilities. In view of the nature of the ChRM (CRM) the age of the magnetization is constrained by the age of the aqueous alteration that starts soon after CAI formation and extends for at least 4 Myr (see §1.1). Indeed, it can be estimated from the time-temperature stability diagrams for magnetite and pyrrhotite ( Pullaiha et al., 1975 ; Dunlop et al., 2000 ) that a rock magnetized at 4.5 Gyr and subsequently kept in null magnetic field in the asteroid belt at an average temperature of about -110°C (e.g. Spencer et al., 1989 ) will be viscously erased only to ~20 or 60 °C whether it is carried by pyrrhotite or magnetite. The transfer time from the asteroid belt to the Earth in the form of a meteoroid is spent mostly in the vicinity of the asteroid belt ( Gladman et al., 1997 ). Therefore no significant amount of time is spent in a near-Earth orbit that would result in further viscous decay of the original remanent magnetization. Theoretical analyses indicate that early planetesimals were likely capable of generating core- dynamos soon after the formation of the solar system. For Vesta asteroid, tungsten model age suggests core formation of 3.8 ± 1.3 Myr after the beginning of the Solar System ( Kleine et al., 2002 ). Some doubts arise concerning the minimum asteroid radius for a dynamo and the field intensity that could be generated. In fact, a small radius of the asteroid could not permit the setting up of a dynamo magnetic field. Scaling laws (e.g. MAC balance) allowed to estimate magnetic field strengths generated by small body dynamos; core magnetic field fall in the range

106

CHAPITRE 4 : MAGNETISME DES CHONDRITES

0.1-150 µT ( Weiss et al., 2010 ; Tarduno et al., 2012 ). The field intensity estimated in this study is in the range expected for planetesimal core dynamos. Therefore, it is possible that the CM chondrites parent body had a dynamo field in its early history. Such a dynamo field has been evidenced in the angrite parent body from 4564 to at least 4558 Ma ( Weiss et al., 2008 ), for the CV parent body at ~4559-4560 Ma ( Carporzen et al., 2011 ), for pallasite parent body (Tarduno et al., 2012 ) and for Vesta asteroid ( Fu et al., 2012 ). A field of external origin (or at least some of its components) would have to be stable with respect to the parent body over timescales long enough to allow CRM acquisition. It is generally believed that most carbonaceous chondrites were formed at 3.5 AU or farther from the protosun (Clayton et al., 1976, 1977 ). At this distance, even the most active T-Tauri phase would result in fields of just 0.001-0.01 µT in the asteroid belt region ( Weiss and Elkins-Tanton, 2013 ). Magnetohydrodynamical simulations predict that fields of order 10–100 µT can be generated in the midplane at this location, but these fields periodically reversed every few hundred years (Turner and Sano, 2008 ). A more likely magnetic field source for CM magnetization is the stable, vertical (out-of-the-disk plane) ~ 10 µT field expected to be inherited from the parent molecular cloud during gravitational collapse. Regarding timescale and the estimated duration of aqueous alteration and those of the different magnetic fields it appears clearly that we cannot discriminate between an internal or an external magnetic field origin to account for the magnetization of the CM chondrites.

6. Conclusion

We have performed magnetic measurements on six CM chondrites falls and one fresh find. The magnetic mineralogy of all studied meteorites is composed of variable amounts of multi- and pseudosingle domain pyrrhotite and magnetite formed during aqueous alteration on the parent body. Even though magnetite dominates magnetic susceptibility in all studied CM chondrites, pyrrhotite dominates the remanent properties in all of them with the exception of Murchison (dominated by magnetite). Paris is the sample containing the most significant fraction of metallic FeNi. No clear correlation could be established between the degree of aqueous alteration and the magnetic properties of the meteorites, except for Paris where the metallic FeNi agrees with the unusually low alteration degree (see Hewins et al., 2014). At the centimeter scale, the magnetic fabric of CM chondrites, resulting from preferential alignment of magnetic grains, is homogeneous in direction and intensity. The paleomagnetic data show that six out of seven studied CM chondrites (Cold Bokkeveld being the exception) possess one high coercivity and high temperature stable and origin-trending 107

CHAPITRE 4 : MAGNETISME DES CHONDRITES

component of NRM. This component is homogeneous in direction and intensity at the scale of the meteorite. We interpret this component as a pre-terrestrial CRM acquired during crystallization of magnetite and pyrrhotite during parent body aqueous alteration in a field of at least a few µT (2 ± 1.5 μT). Cold Bokkeveld meteorite requires further investigations. Magnetite and pyrrhotite were formed by aqueous alteration starting immediately after accretion and for a period of at least 4 Ma. During this time interval both internally generated field (from a putative dynamo) and external fields of nebular origin may have existed. It is impossible so far to discriminate between the two hypotheses. Despite this last consideration, it can be retained that CM chondrites, relics from the nebular epoch of the Solar system, bear the oldest paleomagnetic record ever evidenced to date.

References

Alexander, C.M.O.d., Cody, G.D., Bowden, R., Fogel, M. L., Herd, C.D.K., 2010. Parent body modification of the structure, and elemental and isotopic compositions of IOM in Tagish Lake. Meteoritics and Planet. Sci. Suppl. 73, 5367. Amelin, Y., Krot, A.N., Hutcheon, I.D., Ulyanov, A.A., 2002. Lead isotopic ages of chondrules and Calcium-Aluminum-Rich Inclusions. Science 297, 5587, 1678-1683. Anders, E., 1971. Meteorites and the early solar system, Annu Rev. Astron. Astrophys. 9, 1-34. Balbus, S.A., 2009; Magnetorotational instability. Scholarpedia 4, 2409. Banerjee, S.K., Hargraves, R.B., 1971. Natural remanent magnetization of carbonaceous chondrites. Earth Planet. Sci. Lett. 10, 392-396. Banerjee, S.K., Hargraves, R.B., 1972. Natural remanent magnetization of carbonaceous chondrites and the magnetic field in the early solar system. Earth Planet. Sci. Lett. 17, 110-119. Bischoff, A., 1998. Aqueous alteration of carbonaceous chondrites: evidence for pre-accretionary alteration – a review. Meteoritics and Planet. Sci. 33, 1113 –1122. Brearley, A.J., 1995. Aqueous alteration and brecciation in Bells, an unusual, saponite-bearing, CM chondrite. Geochim. Cosmochim. Acta 59, 11, 2291-2317. Brearley, A.J., Jones, R.H., 1998. Chondritic meteorites. Rev. Mineral. 36, 313-398. Brearley, A.J., 2006. The ac tion of water. In ‘ Meteorites and the Early Solar System II ” (Eds Dante Lauretta, H.Y. McSween Jr), Arizona University Press, 587-624. Brecher, A., Arrhenius, G., 1974. The paleomagnetic record in carbonaceous chondrites: natural remanence and magnetic properties. J. Geophys. Res. 79, 14, 2081-2106. Browning, L., McSween, H., Zolensky, M., 1996. Correlated alteration effects in CM carbonaceous chondrites. Geochim. Cosmochim. Acta 60, 2621 –2633. Britt, D.T., Consolmagno, S.J., 2003. Stony meteorite porosities and densities: a review of the data through 2001. Meteor. Planet. Sci. 38, 1161-1180. Bullock, E.S., Grady, M.M., Gounelle, M., Russell, S.S., 2007. Fe-Ni Sulphides as Indicators of Alteration in CM Chondrites. Lunar Planet. Sci. 38, abstract#2057. Burgess, R., Wright, I.P., Pillinger, C.T., 1991. Determination of sulfur-bearing components in C1 and C2 carbonaceous chondrites by stepped combustion. Meteoritics and Planet. Sci. 26, 55 –64. Carporzen, L., Weiss, B.P., Elkins-Taton, L.T., Shuster, D.L., Ebel, D., Gattacceca, J., 2011. Magnetic evidence for a partially differentiated carbonaceous chondrite parent body. Proc. National Acad. Sci. 108, 6386-6389. Chizmadia, L.J., Brearley, A.J., 2008. Mineralogy, aqueous alteration, and primitive textural characteristics of fine-grained rims in the Y-791198 CMS carbonaceous chondrite: TEM observations and comparison to ALHA81002. Geochim. Cosmochim. Acta 72, 602 –625.

108

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Choi, B.G., McKeegan, K.D., Leshin, L.A., Wasson, J.T., 1997. Origin of magnetite in oxidized CV chondrites: in situ measurements of oxygen isotope compositions of Allende magnetite and olivine. Earth Planet. Sci. Lett. 146, 334-349. Clayton, R.N., Onuma, N., Mayeda, T.K., 1976. A classification of meteorites based on oxygen isotopes. Earth Planet. Sci. Lett. 30, 10-18. Clayton, R.N., Onuma, N., Grossman, L., Mayeda, T.K., 1977. Distribution of the pre-solar component in Allende and other carbonaceous chondrites. Earth Planet. Sci. Lett. 34, 209-224. Cody, G.D., Alexander, C.M.O’D., Yabuta, H., Kilcoyne , A.L.D., Araki, T., Ade, H., Dera, P., Fogel, M., Militzer, B., Mysen, B.O., 2008. Organic thermometry for chondritic parent bodies. Earth Planet. Sci. Lett. 272, 446-455. Connelly, J.N., Bizzarro, M., Krot, A.N., Nordlund, A., Wielandt, D., Ivanova, M.A., 2012. The absolute chronology and thermal processing of solids in the solar protoplanetary disk. Science 338, 651-655. De Leuw, S., Rubin, A.E., Schmitt, A.K., Wasson, J.T., 2009. 53 Mn-53 Cr systematics of carbonates in CM chondrites: Implications for the timing and duration of aqueous alteration. Geochim. Cosmochim. Acta 73, 7433-7442. Dunlop D.J., Ozdemir, O., Clark, D.A., Schmidt, P.W., 2000. Time-temperature relations for the remagnetization of pyrrhotite (Fe 7S8) and their use in estimating paleotemperatures. Earth and Planetary Science Letters 176, 107 –116. Enkin, R.J., Dunlop, D.J., 1988. The demagnetization temperature necessary to remove viscous remanent magnetization. Geophys. Res. Lett. 15, 514 –517. Elmaleh, A., Tarantino, S.C., Zema, M., Devouard, B., Fialin, M., 2012. The low-temperature magnetic signature of Fe-rich serpentine in CM2 chondrites: Comparison with terrestrial cronstedtite and evolution with the degree of alteration. Geochem. Geophys. Geosys. 13, 5, Q05Z42. Eugster, O., Eberhardt, P., Thalmann, Ch., Weigel, A., 1998. Neon-E in CM-2 chondrite LEW90500 and collisional history of CM-2 chondrites, Maralinga, and other CK chondrites. Geochim. Cosmochim. Acta 62, 2573-2582. Fu, R.R., Weiss, B.P., 2012. Detrital remanent magnetization in the solar nebula. American Geophysical Union Issue Journal of Geophysical Research: Planets 117, Issue E2. Fu, R.R., Weiss, B.P., Shuster, D.L., Gattacceca, J., Grove, T.L., Suavet, C., Lima, E.A., Li, L., Kuan, A.T., 2012. An Ancient Core Dynamo in Asteroid Vesta. Science, 338, 238-241. Gambino, R.J., McGuire, T.R., Nakamura, Y., 1967. Magnetic properties of the iron-group metal phosphides. J. Appl. Phys. 38, 1253. Gattacceca, J., Rochette, P., 2004. Toward a robust normalized magnetic paleointensity method applied to meteorites. Earth Planet. Sci. Lett. 227, 377-393. Gattacceca, J., Rochette, P., Denise, M., Consolmagno, G., Folco, L., 2005. An impact origin for the foliations of chondrites. Earth and Planetary Science Letters 234, 351 –368. Gattacceca, J., Rochette, P., Gounelle, M., Van Ginneken, M., 2008. Magnetic anisotropy of HED and Martian meteorites and implications for the crust of Vesta and Mars. Earth Planet. Sci. Lett. 270, 280- 289. Gattacceca, J., Boustie, M., Hood, L., Cuq-Lelandais, J.-P., Fuller, M., Bezaeva, N.S., de Resseguier, T., Berthe, L., 2010. Can the lunar crust be magnetized by shock: Experimental groundtruth. Earth Planet. Sci. Lett. 299, 42-53. Gladman, B.J., Migliorini, F., Morbidelli, A., Zappalà, V., Michel, P., Cellino, A., Froeschlé, C., Levison, H.F., Bailey, M., Duncan, M., 1997. Dynamical lifetime of objects injected into asteroid belt resonances. Science 277, 197. Gounelle, M., Morbidelli, A., Bland, P.A., Sephton, M.A., Young, E.D., Spurny, P., 2008. Meteorites from the outer solar system? In The Solar System Beyond Neptune (eds A. Barucci, H. Boehnhardt, D. Cruikshank, A. Morbidelli), 525-541. Arizona University Press: Tucson. Lett. 3, 351-354. Guo, W., Eiler, J.M., 2007. Temperatures of aqueous alteration and evidence for methane generation on the parent bodies of CM chondrites. Geochim. Cosmechim. Acta 71, 5565-5575. Hanowski, N.P., Brearley, A.J., 2000. Iron-rich aureoles in the CM carbonaceous chondrites, Murray, Murchison, and Allan Hills 81002; evidence for in-situ aqueous alteration. Meteoritics and Planet. Sci. 35, 1291-1308. Hanowski, N.P., Brearley, A.J., 2001. Aqueous alteration of chondrules in the CM carbonaceous chondrite, Allan Hills 81002: implications for parent body alteration. Geochim. Cosmochim. Acta 65, 495-518. 109

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Harries, D., Langenhorst, F., 2013. The nanoscale mineralogy of the Fe, Ni sulfides in pristine and metamorphosed CM and CM/CI-like chondrites: Tapping a petrogenetic record. Met. Planet. Sci. 48, 879-903. Hewins, R.H., Bourot-Denis, M., Zanda, B., Leroux, H., Barrat, J.-A., Humayun, M., Göpel, C., Greenwood, R.C., Franchi, I.A., Pont, S., Lorand, J.-P., Cournède, C., Gattacceca, J., Rochette, P., Kuga, M., Marrocchi, Y., Marty, B., 2014. The Paris meteorite, the less altered CM chondrite so far. M. Geochim. Cosmochim. Acta 124, 190-222. Howard, K.T., Benedix, G.K., Bland, P.A., Cressey, G., 2009. Modal mineralogy of CM2 chondrites by X- ray diffraction (PSD-XRD): Part 1. Total phyllosillicate abundance and the degree of aqueous alteration. Geochim. Cosmochim. Acta 73, 4576-4589. Howard, K.T., Benedix, G.K., Bland, P.A., Cressey, G., 2011. Modal mineralogy of CM chondrites by X- ray diffraction (PSD-XRD): Part 2. Degree, nature and settings of aqueous alteration. Geochim. Cosmochim. Acta 75, 2735-2751. Hyman, M., Rowe, M.W., 1983. The origin of magnetite in carbonaceous chondrites. Lunar Planet. Inst. 14, 341-342. Hyman, M., Rowe, M.W., 1986. Saturation magnetization measurements of carbonaceous chondrites. Meteoritics 21, 1, 1-22. Jelinek, V., 1981. Characterization of the magnetic fabric of rocks. Tectonophysics 79, 63 –67. Jenniskens, P., and 69 co-authors, 2012. Radar enabled recovery of the Sutter’s Mill meteorite, a carbonaceous chondrite regolith breccia. Science 338, 1583-1587. Kimura, M., Grossman, J.N., Weisberg, M.K., 2011. Fe-Ni metal and sulfide mineral in CM chondrites: An indicator of thermal history. Meteoritics and Planet. Sci. 46, 3, 431-442. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of palaeomagnetic data. Geophys. J. Int. 62, 699 –718. Kitajima, F., Nakamura, T., Takaoka, N., Murae, T., 2002, Evaluating the thermal metamorphism of CM chondrites by using the pyrolytic behavior of carbonaceous macromolecular matter. Geochim. Cosmochim. Acta 66, 1, 163-172. Kleine, T., Munker, 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. Kletetschka, G., Kohout, T., Wasilewski, P.J., 2003. Magnetic remanence in the Murchison meteorite. Meteoritics and Planet. Sci. 38, 3, 399-405. Larson, E.E., Watson, D.E., Herndon, J.M., Rowe, M.W., 1973. Partial A.F. demagnetization studies of 40 meteorites. J. Geomag. Geoelectr. 25, 331-338. Lauretta, D.S., Hua, X., Buseck, P.R., 2000. Mineralogy of fine-grained rims in the ALH 81002 CM chondrite. Geochim. Cosmochim. Acta 64, 3263 –3273. Lewis, R.S., Anders, E., 1975. Condensation time of the solar nebula from extinct 129 I in primitive meteorites. Proc. Natl. Acad. Sci. 72, 268-273. Lindgren, P., Lee, M.R., Sofe, M.R., Zolensky, M.E., 2013. Clasts in the CM2 carbonaceous chondrite Lonewolf Nunataks 94101: Evidence for acqueous alteration prior to complex mixing. Met. Planet. Sci. 48, 1074-1090. Lowrie, W., 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties. Geophys. Res. Lett. 17, 159-162. Mason, B., 1962. The carbonaceous chondrites. Space Sci. Rev. 1, 621-646. McClelland, E., 1996. Theory of CRM acquired by grain growth, and its implications for TRM discrimination and paleointensity determination in igneous rocks. Geophys. Jour. Int. 126, 271-280. McSween, H.Y., 1979. Alteration in CM carbonaceous chondrites inferred from modal and chemical variation in matrix. Geochim. Cosmochim. Acta 43, 1761 –1770. Metzler, K., Bischoff, A., Stöffler, D., 1992. Accretionary dust mantles in CM chondrites: evidence for solar nebula processes. Geochim. Cosmochim. Acta 56, 2873 –2897. Miyamoto, M., 1991. Thermal metamorphism of CI and CM carbonaceous chondrites: An internal heating model. Meteoritics 26, 111-115. Nakamura, T., 2005. Post-hydration thermal metamorphism of carbonaceous chondrites. Jour. Mine. Petro. Sci. 100, 260-272. Nazarov, M.A., Brandstätter, F., Kurat, G., 1994. P-rich sulfide, barringerite, and other phases in carbonaceous clasts of the Erevan howardite. Lunar Planet. Sci. 25, (abstract) 979 –980.

110

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Nazarov, M.A., Brandstätter, F., Kurat, G., 1996. Phosphides and P-rich sulfides in the Mighei (CM) chondrite. Lunar Planet. Sci. 27, (abstract) 939 –940. Nazarov, M.A., Brandstätter, F., Kurat, G., 1997. Comparative chemistry of P-rich opaque phases in CM chondrites. Lunar Planet. Sci. 28, (abstract) 1003 –1004. Nazarov, M.A., Kurat, G., Brandstätter, F., Ntaflos, T., Chaussidon, M., Hoppe, P., 2009. Phosphorus- bearing sulfides and their associations in CM chondrites. Petrology 17, 2, 101-123. Niemeyer, S., Zaikowski, A., 1980. I-Xe age and trapped Xe components in the murray (C-2) chondrite. Earth Planet. Sci. Lett. 48, 355-347. Petitat, M., Gounelle, M., 2010. Magnetite content and carbonate mineralogy as constraints for parent body hydrothermal alteration. Lunar Planet. Sci. 41, abstract #1673. Pullaiha, G., Irving, E., Buchan, K.L., Dunlop, D.J., 1975. Magnetization changes caused by burial and uplift. Earth Planet . Sci. Lett. 28, 133-143. Pravdivtseva, O., Meshik, A., Hohenberg, C.M., 2013. The I-Xe record: early onset of aqueous alteration in magnetites separated from CM and CV carbonaceous chondrites. Lunar and Planetary Sci. Conf. 44, abstract #3104. Rochette, P., Fillion, G., Matteí, J. L. and Dekkers M. J., 1990. Magnetic transition at 30-34 K in pyrrhotite: Insight into a widespread occurence of this mineral in rocks. Earth Planet. Sci. Lett. 98, 319- 328. Rochette, P., Gattacceca, J., Bonal, L., Bourot-Denise, M., Chevrier, V., Clerc, J.P., Consolmagno, G., Folco, L., Gounelle, M., Kohout, T., Pesonen, L.J., Quirico, E., Sagnotti, L., Skripnik, A., 2008. Magnetic classification of stony meteorites: 2. Non-ordinary chondrites. Meteoritics and Planet. Sci. 43, 959-980. Rosenberg, N.D., Browning, L., Bourcier, W.L., 2001. Modeling aqueous alteration of CM carbonaceous chondrites. Meteoritics and Planet. Sci. 36, 239-244. Rubin, A.E., Wasson, J.T., 1986. Chondrules in the Murray CM2 meteorite and compositional differences between CM-CO and ordinary chondrite chondrules. Geochim. Cosmochim. Acta 50, 307 –315. Rubin, A.E., Trigo-Rodriguez, J.M., Huber, H., Wasson, J.T., 2007. Progressive aqueous alteration of CM carbonaceous chondrites. Geochim. Cosmochim. Acta 71, 2361 –2382. Scott, E.R.D., Keil, K., Stoffler, D., 1992. Shock metamorphism of carbonaceous chondrites. Geoch. Cosmoch. Acta 56, 4281-4293. Stephenson, A., Collinson, D.W., 1974. Lunar magnetic field paleointensities determined by an anhysteretic remanent magnetization method. Earth Planet. Sci. Lett. 23, 220-228. Spencer, J.R., Lebofsky, L.A., Sykes, M.V., 1989. Systematic biases in radiometric diameter determinations. Icarus 78, 337 –354. Tarduno, J.A., Cottrell, R.D., Nimmo, F., Hopkins, J., Voronov, J., et al. 2012. Evidence for a dynamo in the main group pallasite parent body. Science 338, 939-42. Tikoo, S.M., Gattacceca, J., Weiss, B.P., Suavet, C.R., Cournède, C., 2013. Thermal demagnetization characteristics of shock remanent magnetization and implications for interpreting the paleomagnetism of shocked samples. In progress. Tomeoka, K., McSween, H.Y., Buseck P.R., 1989. Mineralogical alteration of CM carbonaceous chondrites: a review. Proc. NIPR Symp. Antarct. Meteorites 2, 221–234. Turner, N.J., Sano, T., 2008. Dead zone accretion flows in protostellar disks. Astrophys. J. 679, L131 –L34. Wasson, J.T., 1976. Relative abundance of CM chondrites in the inner and outer solar system. Meteoritics, 11, 385. Wasson, J.T., Wetherill, G.W.,1979. Dynamical, chemical and isotopic evidence regarding the formation locations of asteroids and meteorites. In “Asteroids”, (edt by Gehrels T.) The University of Arizo na Press, 926 –974. Weiss, B.P., Berdahl, J.S., Elkins-Tanton, L., Stanley, S., Lima, E.A., Carpozen, L., 2008. Magnetism on the Angrite parent body and the early differentiation of planetesimals. Science 322, 713-716. Weiss B.P., Gattacceca J., Stanley S., Rochette P., Christensen U.R., 2010. Paleomagnetic Records of Meteorites and Early Planetesimal Differentiation. Space Science Reviews, 152, 341-390. Weiss, B.P., Elkins-Tanton, L.T., 2013. Differentiation planetesimals and the parent bodies of chondrites. Annu. Rev. Earth and Planet. Sci. 41, 529-560. Wood, J. A., 1963. On the origin of chondrules and chondrites. Icarus 2, 152-180.

111

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Zanda, B., Bourot-Denise, M., Hewins, R.H., Barrat, J.-A., Gattacceca, J., 2010a. The Paris CM chondrite yields new insights on the onset of parent body alteration. 73rd Annual Meteoritical Society Meeting, abstract #5312. Zanda, B., Bourot-Denise, M., Hewins, R.H., Barrat, J.-A., Marrocchi, Y., Pont, S., Gattacceca, J., Greenwood, R.C., Franchi, I.A., 2010b. Paris: Une chondrite CM exceptionnellement peu altérée et peu metamorphique. Meteoritics and Planet. Sci. 45, 222-222. Zolensky, M.E., McSween, H.Y., 1988. Aqueous alteration. Meteorites and the Early Solar System (eds. J.F. Kerridge and M. S. Matthews). University of Arizona Press, 114-143. Zolensky, M., Barrett, R., Browning, L., 1993. Mineralogy and composition of matrix and chondrule rims in carbonaceous chondrites. Geochim. Cosmochim. Acta 57, 3123 –3148. Zolensky, M.E., Mittlefehldt, D.W., Lipschutz, M.E., Wang, M.-S., Clayton, R.N., Mayeda, T.K., Grady, M.M., Pillinger, C., Barber, D., 1997. CM chondrites exhibit the complete petrologic range from type 2 to 1. Geochim. Cosmochim. Acta 61, 5099 –5115. Zolensky, M.E., Le, L., 2003. Iron-nickel sulfide compositional ranges in CM chondrites: No simple plan. Lunar Planet. Sci. 34, abstract #1235.

112

CHAPITRE 4 : MAGNETISME DES CHONDRITES

SUPPLEMENTARY FILES

113

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Supplemental Figure 4.A Hysteresis loop for the seven CM2 chondrites studied in this work. For Paris and Cold Bokkeveld several samples were measured and the two extremes behaviors are shown.

114

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Supplemental Figure 4.B a) Measurement of low temperature remanence acquired after field-cooling (FC) (gray) and zero-field-cooling (ZFC) (black) during warming to 140 K for Paris and Murchison. Measurement of room temperature remanence (RT- SIRM) for c) Murchison and d) Paris during both cooling and warming, as indicated by arrows. FC LT-SIRM for e) Murchison and f) Paris during both cooling and warming, as indicated by arrows.

115

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Supplemental Figure 4.C Normalized susceptibility versus a) low temperature and b) high temperature. Cold Bokkeveld (dotted line), Mighei (black dashed line), Nogoya (black line), Paris (dashed and continuous black line)

116

CHAPITRE 4 : MAGNETISME DES CHONDRITES

NRM Mrs Ms Bcr Bc Sample mass (g) logχ Pasm Tasm (Am² kg -1 ) (A m 2 kg -1 ) (A m 2 kg -1 ) (mT) (mT) Banten 0.050 7.72E-04 4.00 1.08 -0.77 0.043 4.10E-04 4.06 1.10 0.08 1.02E-01 2.15 89 10 Cold Bokkeveld + 0.010 2.51E-03 4.30 1.10 0.67 4.49E-01 3.28 40 10 Cold Bokkeveld 0.028 6.22E-05 3.46 1.06 0.20 4.65E-02 0.31 50 16 0.028 2.16E-04 3.72 1.11 0.33 7.50E-02 2.76 55 4 0.018 7.82E-05 3.43 1.040.12 - - - - 0.086 1.31E-04 3.78 1.110.08 - - - - 0.150 - 3.58 1.08 -0.53 - - - - 0.103 - 3.44 1.060.89 - - - - Mighei 0.0752.93E-053.351.130.23 - - - - 0.249 1.55E-05 3.33 1.05 0.70 3.58E-02 0.71 58 19 0.211 1.14E-05 3.37 1.08 -0.26 - - - - Murchison 0.765 3.44E-05 3.51 - - 8.86E-02 0.52 61 23.2 0.052 1.35E-04 - - - 6.20E-02 0.29 63 24.3 Murray 0.273 2.15E-04 4.40 1.08 -0.16 1.17E-01 5.69 76 5.6 0.522 1.85E-04 4.10 1.06 0.04 - - - - 0.427 4.75E-04 4.15 1.06 -0.03 - - - - Nogoya 0.520 2.50E-05 3.76 1.10 0.41 8.74E-02 1.05 83 18 0.405 1.50E-05 3.54 1.07 0.50 - - - - 0.232 1.19E-05 3.46 1.05 0.64 - - - - Paris 1.500 8.80E-04 4.6 1.14 0.41 1.25E-01 5.68 68 6 0.583 1.60E-03 4.3 1.12 0.04 1.26E-01 5.14 68 7 3.237 1.36E-03 4.4 1.10 0.43 1.38E-01 6.51 66 6 0.928 3.78E-03 4.35 1.16 0.77 1.25E-01 6.56 62 5 1.155 1.55E-03 3.95 1.13 0.61 9.46E-02 1.82 74 12 1.270 5.43E-03 4.44 1.17 0.78 1.46E-01 5.32 61 5 1.100 8.11E-03 4.38 1.14 0.75 1.38E-01 5.05 63 7 0.040 2.29E-04 3.93 - - 9.42E-02 1.77 67 2 0.155 5.08E-03 - - - 1.33E-01 6.16 63 6 0.023 6.09E-03 - - - 1.49E-01 13.13 64 4 0.094 2.98E-03 - - - 1.26E-01 5.77 68 6 17.127 5.59E-04 4.40 1.12 0.59 ---- 0.070 7.49E-04 4.24 1.09 0.35 1.26E-01 3.62 75 9 0.016 3.18E-03 4.49 1.12 -0.26 ---- 0.053 2.55E-03 3.99 1.10 0.77 1.12E-01 1.80 63 13 Supplemental Table 4.A Intrinsic magnetic properties for individual samples. NRM: natural remanent magnetization, χ: magnetic −9 3 −1 susceptibility ( χ in 10 m kg ), P asm , T asm : anisotropy degree and shape parameter for magnetic susceptibility, M rs : saturation remanence, M s: saturation magnetization, B c: coercivity, B cr : coercivity of remanence. Parameters that were not measured are indicated by: -. +: samples with fusion crust (not included in average value given in Table 1 and 2).

117

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Sample Year of fall Sd Sa VRM max VRM/NRM (Am²logs -1 .kg -1 .µT -1 ) (Am²logs -1 .kg -1 .µT -1 ) (Am² kg -1 ) (%) Cold Bokkeveld + 1838 2.10E-07 - 3.08E-05 2.1 Cold Bokkeveld 1838 6.38E-08 - 8.96E-06 10.0 Mighei 1889 6.98E-08 7.28E-08 1.32E-05 99.2 Murchison* 1969 1.00E-07 1.66E-07 2.10E-05 42.0 Murray 1950 2.37E-07 2.40E-07 1.84E-05 1.5 Nogoya 1879 5.42E-08 6.10E-08 1.13E-05 60.8

Paris ? 2.59E-07 --- Supplemental Table 4.B VRM properties for studied CM chondrites. Sd and Sa are respectively the decay and the acquisition rate of the VRM. VRM max is the maximum viscous remanent magnetization, which may have been acquired during the terrestrial residence time of the samples (see text). * =datas from Kletetschka et al., 2003 , assuming VRM was acquired in a 50µT field.

Before heating After heating at 600°C

Mrs /M s Bcr /B c Mrs /M s Bcr /B c Banten 4.76E-02 8.5 7.65E-02 3.8 Mighei 5.03E-02 3.0 1.47E-01 4.2 Murray 2.06E-02 13.4 1.18E-01 3.1

Nogoya 8.30E-02 4.5 6.05E-02 4.4 Supplemental Table 4.C Stability upon heating. Mrs /M s and B cr /B c ratios obtained before and after thermal demagnetization measurements.

118

CHAPITRE 4 : MAGNETISME DES CHONDRITES

4.3 Les rumurutites Nous présentons ici les résultats préliminaires obtenus pour les rumurutites. Afin de respecter l’organisation des parties précédentes ces résultats sont présentés sous forme d’ébauche d’article en anglais. Tout comme pour les chondrites carbonées étudiées dans la partie précédente les rumurutites offrent une nouvelle possibilité d’aborder les questions inhérentes aux champs magnétiques dans le système solaire primitif et à la différenciation des corps parents. Ces météorites présentent l’intérêt de s’être formées à une distance héliocentrique plus grande que les chondrites carbonées et ordinaires ( Khan et al., 2013 ). Les rumurutites pourraient ainsi permettre d’obtenir de nouvelles contraintes sur les champs magnétiques dans de nouvelles « régions » du système solaire primitif. Le groupe des rumurutites n’a été établi comme un nouveau groupe que relativement récemment ( Schulze et al., 1994 ). Les principaux résultats obtenus sur ces météorites sont synthétisés dans Bischoff et al. (2011) . La plupart des rumurutites sont des brèches qui peuvent contenir des lithologies de type pétrographique 3 à 6 au sein d’un e même météorite. Cette caractéristique suggère que certaines parties du corps parent des rumurutites ont été métamorphisées à haute température (> 800°C) et qu’un ou plusieurs impact s ont mélangé des lithologies équilibrées, issues de niveaux plus profonds, avec des lithologies plus proches de la surface (type 3) . Suite à la bréchification ces météorites n’ont pas été réchauffées de façon intense (Bischoff et al., 2011 ), les lithologies de type 3 n’ont pas été équilibrées. Le fer sous forme métallique est rare (~0,1 wt.%, Schulze et al., 1994 ). On trouve le fer principalement sous forme de sulfures (~6 wt.%). Les rumurutites sont caractérisées par un fort état d’oxydation . De la chromite et de la magnétite sont aussi observées dans certaines rumurutites (Kallemeyn et al., 1996 ). Le paléomagnétisme des rumurutites n’a été étudié que sommairement pour cinq d’entre elles (Gattacceca and Rochette, 2004 , repris par Weiss et al., 2010 pour une synthèse). Une aimantat ion stable a été mise en évidence et une paléointensité d’environ 5 µT a été estimée sans que l’origine du champ ou le mécanisme d’aimantation ne soient discutés. Pour notre étude, nous avons sélectionné deux échantillons : PCA91002 et LAP03639 pour lesquels nous avons réalisé une étude magnétique et paléomagnétique détaillée. Le but principal étant d’établir la nature et l’origine du champ magnétique enregistré par ces météorites. Nos résultats préliminaires suggèrent que l’aimantation contenue dans ces ru murutites pourrait être

119

CHAPITRE 4 : MAGNETISME DES CHONDRITES

liée à un champ de surface d’origine interne d’environ 5 µT et donc à une différenciation partielle de leur corps parent.

Paleomagnetic study of Rumuruti chondrites

C. Cournède ⁽¹⁾, J. Gattacceca ⁽¹⁾, P. Rochette ⁽¹⁾ ⁽¹⁾ CEREGE, CNRS/Université Aix-Marseille, Aix-en-Provence, France

1. Introduction

Different types of magnetic fields were at work in the early solar system: fields generated within the protoplanetary nebula, solar fields, dynamo fields generated in the solid bodies. External fields (i.e. nebular or solar) could have played a major role in the accretion process that generated the primary components of our solar system. Internal (i.e. dynamo generated) fields are of substantial interest since they could provide information on parent body evolution, especially on parent body differentiation. Paleomagnetic studies of extraterrestrial materials can help to understand these primordial aspects of our solar system history. In this study we focused on Rumuruti chondrites (R chondrites). This meteorite group is of particular interest because R chondrites parent body is believed to have formed at a heliocentric distance greater than O chondrites and less than C chondrites ( Khan et al., 2013 ). As such, more than a simple new chondrite group, R chondrites offer the possibility to estimate the magnetic fields strength present in a yet unstudied part of the early solar system.

1.1 Rumuruti chondrite petrology and petrogenesis The Rumuruti group was formerly known as the Carlisle Lakes group. Since 1994, R chondrites have been recognized as a new, well-established chondrite group ( Schulze et al., 1994 ). It was officially named for the type specimen Rumuruti that fell in Kenya, Africa, in 1934. The main results of previous studies on R chondrites have been compiled in Bischoff et al., 2011 . It is noteworthy that besides a very preliminary study of five samples in Gattacceca and Rochette (2004) no paleomagnetic study has been undertaken on this meteorite group. Most of the R chondrites are mixtures of unequilibrated lithology (type 3, often > 3.6) ( Bischoff et al., 1998 ; Bischoff, 2000 ) and clasts metamorphosed to various degrees. They are classified as R3-5 or R3-6 breccias (e.g. Bischoff et al., 1994, Schulze et al. 1994, Kallemeyn et al., 1996 ). The

120

CHAPITRE 4 : MAGNETISME DES CHONDRITES

matrix/chondrules modal abundances ratio is high (1:1) ( Bischoff., 2000 ). R chondirites are highly oxidized, as reflected by the rarity or absence of metallic Fe-Ni (e.g., Rubin and Kallemeyn, 1994 ). The iron is either oxidized or found in the form of iron sulfides (mostly pyrrhotite and pentlandite). Mineralogical features reflect variations in local oxidizing conditions on the R-chondrite parent body (e.g. inhomogeneous distribution of water Isa et al., 2011 ). An important aspect in the framework of our paleomagnetic study is to understand when the original metallic FeNi grains did oxidize to account for the FeNi scarcity actually observed in these meteorites. The most accepted view is that oxidation occurred before accretion, in the solar nebula ( Weisberg, 1991 ; Rubin and Kallemeyn, 1994 ; Bischoff, 2000) . However a post accretion oxidation, on the parent body cannot be excluded (Imae and Zolensky, 2003 ). Moreover pyrrhotite mineral is likely formed before parent body accretion ( Jackson and Lauretta, 2010 ). The chronology of the R-chondrite parent body is still unclear. The Rumuruti parent body has suffered several events like thermal metamorphism, brecciation, impacts… All these processes are responsible of the R chondrites group diversity observed actually. Numerous characteristics (the high oxidation state, high matrix/chondrules modal abundance ratio, the relatively low abundance of droplet chondrules, and high ∆O17 (Rubin and Kallemeyn, 1994 and Bischoff et al. 1994 ) suggest that the R chondrites parent body formed at a greater heliocentric distance than ordinary chondrites ( Kallemeyn et al., 1996 ). The Rumuruti parent body formation started with the accretion of its various components in the protoplanetary disk. After the parent body assembly, thermal metamorphism took place, with transformation of the original material into petrographic type 4, 5 and 6 lithologies possibly in an onion shell structure (Lingemann et al., 2000 ). This statement was tested through Ar-Ar (for instance the closure age is >4370 Ma in PCA 91002, Dixon et al., 2003 ) and I-Xe (closure age comprised between 4556 and 4548 Ma, Claydon et al., 2013 ) dating, as one would expect younger ages for higher petrographic types, but no clear conclusion was reached . Mineralogical and petrographic study in petrographic type 4 indicates that metamorphism endding 4 Ma after CAIs formation (Bischoff and Srinivasan, 2003 ). Recently, olivine-chromite geothermometry in ( Wlotzka, 2005 ) indicates that temperatures reached are 550-690°C for R4-5 and 880°C for R6. R chondrites breccias were produced by impacts following metamorphism ( Scott et al., 1985 ). Through these impacts material from the near-surface (type 3 lithology) and fragments from greater depth (equilibrated and recrystallized lithologies of type 5 and 6) were mixed. Moreover heating induced by impacts generated impact melt rocks ( Jäckel et al., 1996 ; Bischoff et al., 2006). These impact-related lithologies cooled rapidly and were incorporated into the surface

121

CHAPITRE 4 : MAGNETISME DES CHONDRITES

breccias. In a last evolutionary phase, during further impact activities the loose regolith was lithified (consolidated). Some insights like the presence of martensite ( Rubin and Kallemeyn. 1989a., 1989b), or the absence of loss of radiogenic He and Ar ( Schultz et al., 2005 ) indicate that the R-chondrite regolith breccias were not significantly reheated (no more than the maximum temperature suffered by the low petrographic type 3, i.e. around 500°C) subsequent to brecciation or during lithification ( Bischoff et al, 2011 ). Using the shock classification scale for ordinary chondrites ( Stöffler et al., 1991 ), it appears that most R chondrites are very weakly (S2) or weakly shocked (S3). It is worth mentioning that in PCA91002, the majority of the rock is of shock stage S3-S4 ( Rubin and Kallemeyn, 1994 ) characterized by numerous sulfide-rich shock veins and melt pockets. For brecciated R chondrites like PCA 91002 however, it is noteworthy that they have not necessarily been shocked to high pressure after brecciation. Indeed, consolidated breccias can be produced even under relatively low shock pressures in the S1-S2 range (Bischoff et al., 1983 ). The presence of hydrous phases ( Jamsja and Ruzicka 2011 ) and oxygen isotopes analyses (Greenwood et al, 2000 ) suggest that subsequent to thermal alteration, an hydrothermal activity may have existed on the Rumuruti parent body.

Only one fall (Rumuruti) is found in the R chondrite group. All other R chondrites originate from hot and cold deserts and show various degrees of weathering. PCA91002 has a weathering index (wi) indicating no or minor alteration (wi=1, based on the modal abundance of crystalline material that is stained brown in thin sections, Rubin and Huber, 2005 ). With increasing degrees of terrestrial weathering of R chondrites, the sulfide modal abundance decreases, and S and Ni become increasingly depleted. Despite the potentially greater heliocentric distance of the R chondrite parent body formation, the transit time of R chondrites and ordinary chondrites to Earth is similar, as suggested by the similar exposure age range of R chondrites to that in ordinary chondrites (Schultz and Weber 2001 ). The shortest exposure age among R chondrites is 0.2 ± 0.1 Ma for NWA 053, whereas a significantly longer exposure age of ~34 Ma was determined for PCA91002 (Schultz and Weber, 2001 ; Schultz et al., 2005 ). A significant number of all R chondrites studied for noble gases may have been ejected from the R chondrite parent body during one large collisional event between 15 and 25 Ma ( Vogel et al., 2011 ).

122

CHAPITRE 4 : MAGNETISME DES CHONDRITES

1.2 Previous magnetic studies The main potentially ferromagnetic opaque mineral in R chondrites is pyrrhotite, with an average abundance of ~6 wt% ( Schulze et al., 1994, Kallemeyn et al., 1996 ). Chromite and magnetite abundance average 0.45 and 0.35 wt% respectively ( Kallemeyn et al., 1996, Rubin and Kallemeyn, 1994 ). Magnetite is sometimes found in larger amounts, e.g. in A-881988 ( Rochette et al., 2008 ). Ilmenite and Fe,Ni-metals are extremely rare (below 0.1 wt%, Schulze et al., 1994 ). The pyrrhotite elemental analyses point toward hexagonal pyrrhotite (Fe 0.93-0.96 S, normally antiferromagnetic) rather than ferromagnetic monoclinic pyrrhotite (Fe 0.87 S). But hexagonal pyrrhotite seems to be present in a metastable ferromagnetic state (Rochette et al. 2008 ), although this requires further clarification. The only paleomagnetic study of R chondrites consists of a rather preliminary study of five meteorites ( Gattacceca and Rochette, 2004 ; and reviewed in Weiss et al. 2010 ). This study, involving only alternating field (AF) demagnetization of one bulk sample for each meteorite (all Antarctic finds including PCA 91002 studied in this work) evidenced that all samples had a stable remanent magnetization. NRM normalization techniques indicated paleointensities around ∼5 μT for each meteorite. In view of the pressure-induced magnetic phase transition of pyrrhotite at 2.8 GPa (e.g., Rochette et al. 2003 ) and the typical peak shock pressures of Rumuruti chondrites (shock stage S2, with peak pressures >5 GPa), it was concluded that the NRM of Rumuruti chondrites is unlikely to predate the last major impact and therefore cannot be a pre-accretionnal magnetization. The paleointensity estimates suggest the existence of a significant (possibly transient) magnetic field at the surface of the Rumuruti parent body at the time of the magnetization.

2. Rationale

In this study we focused on two Rumuruti chondrites samples supplied by NASA : Pecora Escarpment (PCA) 91002 (R3.8-6; S2-4) and La Paz (LAP) 03639 (R4), for which we have performed a detailed magnetic and paleomagnetic study. Both samples are finds recovered in 1991 (PCA 91002) and 2003 (LAP 03639) from Antarctica during the framework of the Antarctic Search for Meteorites Program (ANSMET). Samples were stored in a magnetically shielded room (field < 400 nT) during several months after receiving them in order to allow partial decay of the viscous remanent magnetization acquired during exposure of the meteorites to the geomagnetic field since their fall. As the original orientation of samples on the parent body is unknown, sample orientation was chosen arbitrarily. Samples were cut in mutually-oriented sub-samples with a wire saw. Measurements of the magnetic properties allowed us to characterize their 123

CHAPITRE 4 : MAGNETISME DES CHONDRITES

magnetic mineralogy. Measurement of their natural remanent magnetization (NRM) allowed us to investigate the origin and the nature of the ancient magnetic fields recorded in these objects that prevailed at the beginning of the solar system.

The same methods and techniques than those described for our study on CM chondrites were used here (See Chapter 4, section 4.2).

3. Preliminary results

3.1. Magnetic mineralogy All results are presented and summarized in Table 4.4 .

Hysteresis properties indicate the presence of high coercivity carriers in the two studied samples.

Mrs /M s and B cr /B cr ratio are characteristic of pseudosingle domain carriers. S-ratios are 0.79 and 0.60 in PCA91002 and LAP03639 respectively. These values indicate that the remanence is dominated by a high coercivity mineral. Overall, these are favourable conditions for a paleomagnetic study.

Sample masse Mrs =SIRM Mrs /M s Bcr Bc S-ratio Log χ PT Prem Trem log χ p Log χ f (g) (Am² kg -1 ) (mT) (mT)

PCA91002 91002-57 1610 2.99 1.04 0.77 91002-57D3 134 2.96 1.11 -0.01 91002-57D4 178 2.51E-02 0.25 78.28 32.64 0.79 3.02 1.02 0.54 1.27 0.17 2.66 2.76 91002-57E2 108 2.28E-02 0.78 LAP03639 639-24D3 155 2.97 1.03 -0.28 639-24D4 149 1.62E-02 0.28 116.40 38.90 0.61 2.97 1.02 -0.31 1.19 0.37 2.67 2.67 639-24E2 110 9.45E-03 0.60

Table 4.4 : Samples studied in this work with their intrinsic magnetic properties. Mrs: saturation remanence, Ms: saturation magnetization, Bc: coercivity, Bcr: coercivity of remanence, S-ratio (=(IRM -0.3T )/(IRM 3T )), χ: magnetic susceptibility, χp: paramagnetic susceptibility, χf: ferromagnetic susceptibility. All susceptibilities are in 10 −9 m3 kg −1, χf: susceptibility variation corrected for paramagnetic susceptibility, P, T: anisotropy degree and shape parameter for magnetic susceptibility, Prem , T rem : anisotropy degree and shape parameter of remanence.

Low temperature remanence measurements performed with a VSM at CEREGE are shown in Fig. 4.10 . The most striking features observed are the Verwey transition in PCA91002 observed in the LT-SIRM measurement, characteristic of magnetite, and the overall shape obtained through RT-SIRM measurements for both meteorites, suggestive of monoclinic pyrrhotite (Rochette et al. 1990 ). The increase in RT-SIRM below 60 K can be attributed to the

124

CHAPITRE 4 : MAGNETISME DES CHONDRITES

paramagnetic contribution from the silicates. More robust results, especially regarding the nature of the pyrrhotite, may be obtained by using a MPMS rather than a VSM.

Fig 4.10. Measurements of low temperature (LT) and room temperature (RT) of SIRM for samples studied here. Arrows indicate most striking feature for each measurement. For commodity magnetic susceptibility ( χ) is expressed in log units (log χ, with χ in 10 −9 m3 kg−1 , Table 4.4 ). Magnetic susceptibility was measured on at least three sub-samples in each meteorite which give an average log χ of 2.99±0.03 and 2.97±0.01 for PCA91002 and LAP03639 respectively. The relatively small scatter in our susceptibility measurements, indicated by the standard deviations suggest that magnetic grains are homogeneously distributed at the scale of our samples (~ 0.35 g). Our results are also in agreement with values obtained previously on much larger samples (Rochette et al., 2008), again indicating homogeneous dispersion of

125

CHAPITRE 4 : MAGNETISME DES CHONDRITES

magnetic minerals at the scale of the meteorite (~3.25). It is noteworthy that silicate paramagnetism account for a significant part (50% in both studied samples) of the total susceptibility. We have investigated the properties of the ARM and SIRM of both meteorites. Their intensities are reported in Table 4.5 and Fig. 4.11 . Median destructive field (MDF) is used to describe the stability of SIRM and ARM against AF demagnetization. The two studied meteorites exhibit high MDF values both for ARM and SIRM. The ARM acquired in an alternating field of 110 mT exhibits similar range of stability with MDFs averaging ~48 mT in both rumurutites. It is noteworthy that in view of the dominance of a high coercivity mineral, this ARM does not affect all ferromagnetic grains and this must be kept in mind when interpreting its properties.

Sample IRM MDF of NRM MDF of ARM MDF of (Am² kg -1 ) IRM (Am² kg -1 ) NRM (Am² kg -1 ) ARM

PCA91002 91002-57D3 - 3.46E-05 - 91002-57D4 2.51E-02 62 2.47E-05 120 1.19E-04 46 91002-57E2 2.28E-02 70 2.44E-05 115 LAP03639 639-24D3 - 3.31E-05 - 639-24D4 1.62E-02 120 2.99E-05 >120 7.52E-05 46 639-24E2 9.45E-03 120 1.75E-05 >120

Table 4.5 : Remanent magnetizations with the corresponding median destructive field (MDF) results obtained for studied samples. ARM: anhysteretic remanent magnetization was acquired in a DC field of 100 μT, and AC field of 110 mT. SIRM intensities are given in Table 4.4 (Mrs column).

Fig. 4.11. Normalized intensity of SIRM versus AF for PCA 91002 (red) and LAP 03639 (grey).

126

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Thermal demagnetization of the SIRM ( Fig. 4.12 ) shows that both studied R chondrites are demagnetized at low temperature with less than 10% and 30% of the magnetization remaining at 250 °C in LAP03639 and PCA91002 respectively. Together with the high coercivity and the low temperature behaviour, this indicates that pyrrhotite dominates the remanence signal. It is noteworthy that PCA91002 shows a second inflexion at ~580°C characteristic of magnetite.

Fig 4.12. Normalized intensities of NRM (black line) and SIRM (gray line) versus demagnetization temperature for studied R chondrites

Magnetic anisotropy measurements are presented in Table 4.4. Anisotropy degrees are usually weak (1.02 to 1.04) and fall in the range (1.016-1.065) estimated on larger samples in a previous study (Gattacceca et al., 2005). The anisotropy of ARM (AARM) is moderate, and cannot account for deviations the paleomagnetic directions of more than a few degrees . It is noteworthy that mutually oriented samples have similar orientation of AMS and AARM axes (Fig. 4.13), indicating a homogeneous fabric at the scale of about 1 cm (initial size of the largest studied samples). In particular AARM axes are have similar direction as the AMS axes. The observed homogeneity of the fabric in our rumurutites samples indicates that for the brecciated sample (PCA 91002) the magnetic fabric was acquired after brecciation: either brecciation took place before crystallization of pyrrhotite (which would imply that the oxidation of metal in R chondrites is an asteroidal and not a nebular process), or the fabric was acquired following brecciation for instance during an impact event as already observed in ordinary chondrite breccias (e.g. Gattacceca et al. 2003 ).

127

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig 4.13 Maximum (squares) and minimum (circles) magnetic susceptibility and anisotropy of remanence (indicated by the letter R) axes for the studied R chondrites (projection on lower hemisphere).

3.2. Paleomagnetic results AF demagnetizations were performed at least on two samples for each meteorite. Thermal demagnetizations were also conducted after removal of a low coercivity (LC) component with AF. The directions of the components were computed using the principal component analysis (Kirschvink, 1980 ), without anchoring to the origin. The results are listed in Table 4.6 , and displayed in Fig. 4.14, 4.15.

Sample Treatment AF/thermal I D MAD n REM' Paleofield range (mT (µT) or °C) PCA91002

91002-1 AF LC 3-11 mT 186.5 -10.5 5.4 11 2.07E-03 6.2 HC 12-120 mT 123.2 7.3 1.3 38 7.18E-04 2.2 91002-2 AF LC 4-20 mT 186.1 -5.9 21.6 9 1.33E-03 4.0 HC 22-110 mT 118.7 5.2 3.5 16 1.18E-03 3.5 91002-3 AF+TH LC 2-10 mT 226.5 -24.2 5.4 12 HT 50-520 °C 128.9 7.3 7.1 15 LAP03639 639-1 AF LC 2-11 mT 266 17.9 9.8 13 7.01E-03 21.0 HC 12-120 mT 342.3 -6.2 0.6 39 1.63E-03 4.9 639-2 AF LC 2-10 mT 263.4 21.9 14.9 5 7.26E-03 21.8 HC 16-110 mT 345.4 3.1 1.2 30 1.74E-03 5.2 639-3 AF+TH LC 2 mT-50 °C 261.2 35.1 32.7 12 HT 55-250 °C 348.6 0.6 5.3 13 Table 4.6 : Paleofield estimates obtained on studied samples. REM ’: REM ’ integrated over the given alternating demagnetization field range (with associated paleofield estimate). LC = low coercivity component, HC=high coercivity component HT= high temperature component. 128

CHAPITRE 4 : MAGNETISME DES CHONDRITES

For 91002, three mutually oriented samples (with mass in the 0.11-1.6 g range) were studied two with AF demagnetization and one with thermal demagnetization. NRM intensity is homogeneous and average 2.8 x 10 -5 Am².kg -1. AF demagnetization reveals a LC component isolated below 15 mT, and a well-defined origin-trending high-coercivity (HC) component isolated between 17 and 115 mT ( Fig. 4.14 ). Above 115 mT, the demagnetization path becomes erratic, as is often the case when demagnetizing pyrrhotite with AF. After removal of the LC component by AF, thermal demagnetization reveals a high-temperature (HT) origin-trending component isolated between 50° and 520°C (Fig. 4.14 ). The HT and HC component have the same direction ( Fig. 4.15 ).

For LAP03639, AF demagnetization reveals LC component below 10 mT and a HC component isolated between 15 and 115 mT ( Fig 4.14). Above 115 mT, the demagnetization path becomes erratic. Thermal demagnetization (after removal of the LC component with AF treatment), shows a HT component isolated between 55 and 250°C (Fig 4.14 ) with the same direction as the HC component ( Fig 4.15 ).

Fig. 4.14. Orthogonal projection AF and AF+thermal demagnetization data for the studied R chondrites.

129

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig 4.15. Stereographic projection of the stable component of magnetization (and their 95% confidence interval) for LAP 03639 (black symbols) and PCA91002 (red symbols) where multiple mutually-oriented samples were studied. Open and solid symbols are projection in the upper and low hemisphere respectively. Boxes: LC components, Circles: HC/HT components (ChRM). 3.3. Discussion After removal of a LC component, the magnetization directions isolated by AF and thermal demagnetization in mutually-oriented samples are identical, and trend towards the origin. These magnetizations are referred to as Characteristic Remanent Magnetization (ChRM) in the following). The LC components of the NRM are likely of viscous origin or result from magnetic contamination during sample curation, handling, and/or preparation. These terrestrial components are not discussed in further details. The homogeneity of ChRM directions between mutually oriented sub-samples indicates a post- accretion magnetization for both meteorites and even post-brecciation for PCA91002. It is noteworthy that we have not considered a possible SRM origin for the ChRM. Pyrrhotite has a rather low pressure magnetic transition, at about 2.8 GPa ( Rochette et al., 2003 ), way below the expected peak pressures suffered by these meteorites during the main shocks that led to the shock stages S2 to S4 observed in PCA 91002. Because the ChRM is mostly carried by pyrrhotite, this means that it cannot predate the main shock suffered by these meteorites. However, as discussed above for PCA 91002, the main shock was pre-brecciation, and it is well possible that no shock above 2.8 GPa took place after brecciation. It is also noteworthy that the S4 stage

130

CHAPITRE 4 : MAGNETISME DES CHONDRITES

observed in some PCA 91002 clasts implies shock pressure between 30-35 GPa and an associated temperature increase of ~500°C (Stöffler et al. 1991 ), leasing to post-shock temperatures of about ~350°C (assuming an asteroid temperature of -150°C, typical of temperatures in the asteroid belt), leading to complete remagnetization of pyrrhotite not by shock but by thermal effect. In view of the terrestrial residence time of the studied meteorites (50 to ~100000 years), the terrestrial VRM is expected to be stable up to 80-120°C upon laboratory thermal demagnetization whether it is carried by pyrrhotite (Pullaiha et al., 1975 ) or magnetite ( Dunlop et al., 2000 ). Unblocking temperature are significantly higher (250°C) than this expected temperature stability therefore a VRM origin is ruled out.

In LAP 03639 ChRM is post accretion. The petrographic type 4 of this meteorite implies temperatures in excess of 550°C (Wlotzka, 2005 ), well above the unblocking temperature of ~250 °C defined through NRM thermal demagnetization ( Fig 4.16 ). Therefore the ChRM cannot predate thermal metamorphism. In view of its properties, in particular the similar behaviour of the thermal demagnetization of ChRM and SIRM, it can be interpreted as a CRM or a TRM. The CRM hypothesis would mean that pyrrhotite is formed on the parent body by aqueous alteration following thermal metamorphism, a sequence of events that is not the expected one on asteroidal bodies. The TRM hypothesis would mean that the ChRM is acquired during cooling following thermal metamorphism or during a late thermal event. In that case pyrrhotite formation occurs before the thermal event and is therefore more likely of nebular origin even if an early (pre-thermal event) asteroidal origin cannot be excluded.

In PCA 91002, the ChRM is also post accretion. We have seen that it is also post brecciation and therefore post high-temperature metamorphism. Indeed, unequilibrated petrographic type 3 encountered in this meteorites suggest that post brecciation temperatures were below ~500 °C to preserve this unequilibrated mineralogy. Like in LAP 06639, the NRM could be a CRM or a TRM. In the TRM hypothesis, a post-brecciation temperature increase to ~500°C followed by cooling could account for the ChRM, regardless of the process that generated the temperature increase (vicinity of a melt layer for instance). In the CRM hypothesis CRM, as discussed above for LAP03639, a post brecciation hydrothermal event is required. Because aqueous alteration usually predates thermal metamorphism on asteroidal bodies (e.g. Krot et al., 2006, Huss et al., 2006 ), we favour the TRM hypothesis for both meteorites.

131

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Fig 4.16. Normalized intensities of NRM (black line), ARM (dashed line) and of IRM (gray line) versus demagnetizing field (left) and REM ’ values versus alternating field demagnetization (right) for the two R chondrites demagnetized with AF. The LC (light grey) and HC (black) component range are indicated by the light grey and the black portions on the REM’ plot. The horizontal line indicates the integrated REM’ value and the range of integration.

132

CHAPITRE 4 : MAGNETISME DES CHONDRITES

3.4 Paleointensities and magnetizing field Paleointensities were estimated using the REM’ method ( Gattacceca and Rochette, 2004 ) and are given in Table 4.6 . The paleointensity associated with the ChRM are essentially identical in both studied R chondrites and average 3.9±1.4 µT. These results are in agreement with previous one obtained for R chondrites ( Gattacceca and Rochette, 2004 ). This value must be regarded as a minimum value if ChRM is interpreted as a CRM rather than a TRM. Even though we have not yet conducted SRM acquisition experiments, an even higher field would be required if the ChRM was a SRM, since SRM is a less magnetizing mechanism than TRM (e.g., Gattacceca et al. 2007 ) The final thermal events on the Rumuruti chondrite parent body seem to occur too recently (~10 Ma after the solar system formation considering I-Xe system, Claydon et al., 2013 ), to envisage a nebular origin for the magnetizing field. Therefore, the most likely hypothesis for a “late” stable field of ~4 µT is an internally -generated magnetic field (dynamo field) as proposed recently for angrites (Weiss et al., 2008 ) and CV chondrites ( Carpozen et al ., 2011 ). This would imply that like proposed for the CV chondrites (Carporzen et al., 2011), the R chondrites originate from an undifferentiated layer at the surface of a partially differentiated parent body . However, a better understanding of the thermochronology of R chondrites is necessary to improve to ground this conclusion more firmly. Also in view of the sensitivity of pyrrhotite to pressure remagnetization, the hypothesis that the ChRM is a shock remanence (SRM) should also be considered here. Laboratory SRM acquisition up to 3 GPa are planned in a near future and will help discuss this hypothesis. However, even if the ChRM is a SRM, a late magnetizing field of > 4 µT is still necessary, and likely still require an internally generated dynamo field.

Conclusion

To summarize, the most likely hypothesis is that ChRM measured in both studied R chondrites is a TRM acquired during a late (> ~10 Ma after the formation of the solar system) thermal event in a field of ~5 µT. Considering the primordial magnetic fields timescale and intensities, an internal origin is favoured for the magnetizing field, implying partial differentiation of the R chondrite parent body. Partially differentiated chondritic parent body may have been the rule rather than the exception in the early solar system.

References

Bischoff, A., Rubin, A.E., Keil, K., Stöffler, D., 1983. Lithification of gas-rich chondrite regolith breccias by grain boundary and localized shock melting. Earth Planet. Sci. Lett. 66, 1-10. 133

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Bischoff, A., Geiger, T., Palme, H., Spettel, B., Schultz, L., Scherer, P., Loeken, T., Bland, P., Clayton, R., Mayeda, T.K., Herpers, U, Meltzow, B., Michel, R., Dittrich·Hannen, B., 1994. Acfer 217: A new member of the Rumuruti chondrite group (R). Meteoritics 29, 264-274. Bischoff, A., Weber, D., Bartoschewitz, R., Clayton, R.N., Mayeda, T.K., Spettel, B., Weber, H.W., 1998. Characterization of the Rumuruti chondrite regolith breccia Hughes 030 (R3-6) and implications for the occurrence of unequiliberated lithologies on the R-chondrite parent body. Meteorit. Planet. Sci. 33. A15-A16 (abstract). Bischoff, A., 2000. Mineralogical characterization of primitive, type-3 lithologies in Rumuruti chondrites. Meteoritics and Planetary Science 35, 699-706. Bischoff, A., Srinivasan, G., 2003. 26 Mg excess in hibonites of the Rumuruti chondrite Hugues 030. Meteorit. Planet. Sci. 38, 5-21. Bischoff, A., Scott, E.R.D., Metzler, K., Goodrich, C.A, 2006. Nature and origins of meteoritic breccias. In: Lauretta, D.S., McSween Jr., H.Y. (Eds.), Meteorites and the Early Solar System II.University of Arizona, Tucson, 679-712. Bischoff, A., Vogel, N., Roszjar, J., 2011. The Rumuruti chondrite group. Chemie der Erde 71, 101-133. Carporzen, L., Weiss, B.P., Elkins-Taton, L.T., Shuster, D.L., Ebel, D., Gattacceca, J., 2011. Magnetic evidence for a partially differentiated carbonaceous chondrite parent body. Proc. National Acad. Sci. 108, 6386-6389. Claydon, J.L., Ruzicka, A., Crowther, S.A., Lee, M.Y.P., Bischoff, A., Busemann, H., Gilmour, D., 2013. First I-Xe ages of rumuruti chondrites and the thermal history of their parent body. Lunar and Planet. Sci. Conf. 44, abstract 2211. Dixon, T.E., Bogard, D.D., Garrison, D.H., 2003. 39Ar-40Ar chronology of R chondrites. Meteoritics and Planetary Science 38, 341-355. Dunlop D.J., Ozdemir, O., Clark, D.A., Schmidt, P.W., 2000. Time-temperature relations for the remagnetization of pyrrhotite (Fe7S8) and their use in estimating paleotemperatures. Earth and Planetary Science Letters 176, 107 –116. Gattacceca, J., Rochette, P., Bourot-Denise, M., 2003. Magnetic properties of a freshly fallen LL ordinary chondrite: the Bensour meteorite. Physics of the Earth and Planetary Interiors 140, 343-358. Gattacceca, J., Rochette, P., 2004. Toward a robust normalized magnetic paleointensity method applied to meteorites Earth and Planet. Sci. Lett. 277, 377-393. Gattacceca, J., Rochette, P., Denise, M., Consolmagno, G., Folco, L., 2005. An impact origin for the foliation of chondrites. Earth Planet. Sci. Lett. 234, 351 –368. Gattacceca, J., Berthe, L., Boustie, M., Vadeboin, F., Rochette, P., De Resseguier, T., 2008. On the efficiency of shock magnetization processes. Physics of the Earth and Planetary Interiors 166, 1-10. Greenwood, J.P., Rubin, A.E., Wasson, J.T., 2000. Oxygen isotopes in R-chondrite magnetite and olivine: Links between R chondrites and ordinary chondrites. Geochim. Cosmochim. Acta, 64, 3897 –3911. Huss, G.R., Rubin, A.E., Grossman, J.N., 2006. Thermal Metamorphism in Chondrites. In Lauretta, D.S., Mc Sween, H.Y., (eds) Meteorites and the Early Solar System II. University of Arizona, 567 –586. Imae, N., Zolensky, M.E., 2003. Mineralogy and petrology of a Rumuruti chondrite, including a large unequilibrated clast: PRE95404. Meteorit. Planet. Sci. 38, 5176. Isa, J., Rubin, A.E., 2011. Oxidation in R chondrites. 74th Annual Meteoritical society meeting, Abstract 5200. Jäckel, A., Bischoff, A., Clayton, R.N., Mayeda, T.K., 1996. Dar al Gani 013-A new Saharan Rumuruti- chondrite (R3-6) with highly unequilibrated (type 3) fragments. Lunar Planet. Sci. 27, 595. Jackson, K.M., Lauretta, D.S., 2010. Sulfides in R chondrites: Evidence for sulfidizing conditions in the early solar system. Meteoritics and Planetary Science 45, A94. Jamsja, N., Ruzicka, A., 2011. Presence of hydrous phases in two R chondrites, Northwest Africa 6491 and 6492. 42nd Lunar and Planet. Sci. Conf., abstract 2324. Kallemeyn, G.W., Rubin, A.E., Wasson, J.T., 1996. The compositional classification of chondrites: VII. The R chondrite group. Geochemica and Cosmochimica Acta 60, 2243-2256. Khan, R., Shirai, N., Ebihara, M., 2013. Bulk composition of R chondrites: New data. 44th Lunar and Planet. Sci. Conf., abstract 2059. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of palaeomagnetic data. Geophys. J. Int. 62, 699 –718.

134

CHAPITRE 4 : MAGNETISME DES CHONDRITES

Krot, A.N., Hutcheon, I.D., Brearley, A.J., Pravdivtseva, O.V., Petaev, M.I., Hohenberg, C.M., 2006. Timescales and settings for alteration of chondritic meteorites. In Lauretta, D.S., Mc Sween, H.Y., (eds) Meteorites and the Early Solar System II. University of Arizona, 525-553. Ligemann, C.M., Berlin, J., Stöffler, D., 2000. Rumuruti chondrite: origin and evolution of primitive components. Meteoritics and Planetary Science 35, Abstract 98. Pullaiha, G., Irving, E., Buchan, K.L., Dunlop, D.J., 1975. Magnetization changes caused by burial and uplift. Earth Planet . Sci. Lett. 28, 133-143. Rochette, P., Fillion, G., Mattei, J. L., Dekkers, M., 1990. Magnetic transition at 30 –34 K in pyrrhotite: Insight into a widespread occurrence of pyrrhotite in rocks. Earth Planet . Sci. Lett. 98, 319-328. Rochette, P., Fillion, G., Ballou, R., Brunet, F., Ouladdiaf, B., Hood, L., 2003. High pressure magnetic transition in pyrrhotite and impact demagnetization on Mars. Geophys. Res. Lett. 30, 1683. Rochette, P., Gattacceca, J., Bonal, L., Bourot-Denise, M., Chevrier, V., Clerc, J.-P., Consolmagno, G., Folco, L., Gounelle, M., Kohout, T., Pesonen, L., Quirico, E., Sagnotti, L., Skripnik, A., 2008. Magnetic Classification of Stony Meteorites: 2. Non-Ordinary Chondrites. Meteoritics and Planetary Sciences 43, 959-980. Rubin, A.E., Kallemeyn, G.W.,1989a. A unique chondrite grouplet : Petrology and chemistry of Carlisle Lakes 001 and Allan Hills 85151. Lunar Planetary Science 20, 930-931. Rubin, A.E., Kallemeyn, G.W.,1989b. Carlisle Lakes and Allan Hills 85151 : Members of a new chondrite grouplet. Geochimica and Cosmo Acta 53, 3035-3044. Rubin, A.E., Kallemeyn, G.W., 1993. Carlisle Lakes chondrites : Relationship to other chondrite groups. Meteoritics 28, 255-264. Rubin, A.E., Kallemeyn, G.W., 1994. Pecora Escarpment 91002: A member of the new Rumuruti (R) chondrite group. Meteoritics 29, 255-264. Rubin, A.E., Huber, H., 2005. A weathering index for CK and R chondrites. Meteoritics and Planetary Science 8, 1123-1130. Schultz, L., Weber, H.W., 2001. The irradiation history of Rumurnti-Chondrites. 26th Symposium on Antarctic Meteorites, abstract 128. Schultz, L., Weber, H.W., Franke, L., 2005. Rumuruti chondrites: Noble gases, exposure ages, pairing, and parent body history. Meteoritics and Planetary Science 40, 557-571. Schulze, H., Bischoff, A., Palme, H., Spettel, B., Dreibus, G., Otto, J., 1994. Mineralogy and chemistry of Rumuruti: The first meteorite fall of the new R chondrite group. Meteoritics 29, 275-286. Scott, E.R.D., Lusby, D., Keil, K., 1985. Ubiquitous brecciation after metamorphism in equilibrated ordinary chondrites. Journal of Geophysical Research: Solid Earth 90, 137-148. Stöffler, D., Keil, K., Scott, E.R.D., 1991. Shock metamorphism of ordinary chondrites. Geochimica et Cosmochimica Acta 55, 3845-3867. Vogel, N., Baur, H., Bischoff, A., Wieler, R., 2011. Cosmic ray exposure ages of Rumuruti chondrites from North Africa Chemie der Erde Geochemistry 71, 135-142. Weisberg, M.K., Prinz, M., Kojima, H., Yanai, K., Clayton, R.N., Mayeda, T.K., 1991. The Carlisle Lakes- type chondrites: A new grouplet with high δ17O and evidence for nebula oxidation. Geochim. Cosmochim. Acta 55, 2657 –2669. Weiss, B., 2008. Paleomagnetic records of planetary differentiation and evolution. International Conference on Rock Magnetism, Cargese (France), June 2-7. Weiss, B.P., Gattacceca, J., Stanley, S., Rochette, P., Christensen, U.R., 2010. Paleomagnetic Records of Meteorites and Early Planetesimal Differentiation. Space Science Reviews 152, 341-390. Wlotzka, F., 2005. Cr spinel and chromite as petrogenetic indicators in ordinary chondrites: Equilibration temperatures of petrologic types 3.7 to 6. Meteorit. Planet. Sci. 40, 1673-1702.

135

CHAPITRE 5 : CONCLUSIONS

Chapitre 5: Conclusions

Ce travail sur le magnétisme de la matière extraterrestre nous a permis de faire un retour dans le temps à travers l’investigation des champs e xistants au moment de la formation du système solaire, et les champs anciens générés sur les satellites et les corps parents des météorites. Ainsi, nous avons exploré deux grands aspects : l’étude d’un champ de dynamo sur un corps différencié, la Lune, à travers l’analyse de 17 échantillons issus des missions Apollo ; et les champs magnétiques dans le système solaire primitif à travers l’étude de chondrites (CM et les rumurutites). Les chondrites sont les seules reliques du système solaire primitif disponibles à l’étude. En cela et au-delà des informations sur les champs magnétiques, leur étude magnétique peut nous renseigner sur les conditions physico-chimiques qui régnaient lors de leur formation et sur les processus qu’elles ont pu subir par la suite.

Notre étude sur 17 gros échantillons lunaires a révélé que onze d’entre -eux montrent au moins une composante d’aimantation rémanente naturelle stable. Parmi ceux -ci, nous avons déterminé que cinq échantillons, tous des basaltes, avaient acquis leur aimantation (TRM) sur la Lune et pouvaient nous renseigner sur le paléochamp lunaire. Pour ces cinq échantillons (ainsi que deux échantillons similaires issus d’une étude précédente), des paléointensités et des paléoinclinaisons ont été estimées. La paléointensité moyenne obtenue avec la méthode du REM’ est de 50 µT. Cette forte estimation de l’ intensité du champ de surface confirme les résultats précédents qui montrent que la Lune a eu un champ de dynamo fort dans le passé, mais ceci met en défaut les actuels modèles de génération de dynamo. En utilisant l’anisotropie de susceptibilité magnétique comme indicateur de paléohorizontale, et en c onnaissant les sites d’échantillonnage sur la Lune nous avons déterminé que les paléoinclinaisons peuvent être expliquées par un champ dipolaire avec un paléopôle magnétique situé à ~ 75° N. Ainsi nos résultats suggèrent l’existence sur la Lune au moins entre 3.8 et 3.3 Ga (la gamme d’âge des sept échantillons concernés), d’un champ de dynamo dipolaire dont l’axe était proche de l’axe de rotation actuel de la Lune . L’étude détaillée d’autres échantillons Apollo , en particulier en provenance du site Apollo 15 et Apollo 11, permettrait de valider notre modèle géométrique. De plus, de meilleures contraintes sur l’âge des échantillons permettraient de contraindre la durée effective du phénomène de dynamo et de considére r avec plus d’attention les processus alternatifs nécessaires au maintien d’un champ prolongé de forte intensité sur la Lune.

136

CHAPITRE 5 : CONCLUSIONS

Les résultats obtenus sur les chondrites CM et les Rumurutites montrent que ces météorites portent une aimantation enregistrée dans des champs magnétiques anciens, quelques millions d’années après la formation du système solaire. Six des sept chondrites CM étudiées portent une composante d’ aimantation stable et homogène. Celle-ci est interprétée comme une CRM acquise lors de l’altér ation hydrothermale sur le corps parent à l’ issue de la cristallisation sous champ de la magnétite et de la pyrrhotite. Les datations suggèrent que l’altération hydrothermale a eu lieu, et par conséquent l’aimantation a été acquise, dans les 4 premiers Ma qui ont suivi la formation du système solaire. La méthode du REM’ permet d’estimer une paléointensité moyenne minimum de 2±1.5 µT. Une telle intensité peut potentiellement correspondre à l’intensité du champ présent dans le disque protoplanétaire à cette é poque, ou à l’intensité d’un champ de dynamo. Ainsi, l ’estimation de la paléointensité et les contraintes chronologiques dont nous disposons ne permettent pas de trancher entre un champ d’origine externe (solaire ou nébulaire) ou d’origine interne (dynamo). Cette dernière hypothèse laisse entrevoir la possibilité de la formation de corps partiellement différenciés dès les premiers millions d’année du système solaire. On retiendra cependant de notre étude que l’aimantation rémanente des chondrites CM constit ue probablement le plus ancien enregistrement paléomagnétique jamais mis en évidence .

L’étude préliminaire sur les rumurutites, amène à des conclusions similaires à celles obtenues pour les chondrites CM. En effet, les deux échantillons analysés montrent une composante d’aimantation de haute coercivité stable et homogène dans les deux météorites étudiées. La paléointensité moyenne estimée avec la méthode du REM’ montrent l’existence d’ un champ de surface minium de ~2 µT. L’ordre de grandeur de la paléoint ensité peut là aussi correspondre à l’intensité d’un champ d’origine externe ou d’origine interne. Cependant même si la nature de l’aimantation n’a pas été déterminée de façon certaine, les quelques contraintes temporelles dont nous disposons permettent d’ estimer que l’aimantation est plus récente que dans les chondrites CM et pourrait avoir eu lieu jusqu’ à au moins ~10 Ma après la naissance du système solaire. Dans ce cas, une origine interne est donc favorisée. Au vue de ces résultats, des analyses supplémentaires sont donc prévues afin de préciser par quel mécanisme l’aimantation a été aquise. Et même s’ils demandent encore confirmation et de meilleures contraintes chronologiques, les résultats obtenus ici dans les Rumurutites semblent conforter l’idée que les corps parents peuvent être partiellement différenciés , avec une couche chondritique recouvrant un intérieur différencié et un noyau métallique capable de générer un champ de dynamo .

137

CHAPITRE 5 : CONCLUSIONS

L’étude du paléomagnétisme de la matière extraterrestre est loin d’ avoir livré tous ces secrets et au-delà de l’amélioration des résultats obtenus ici, ce domaine offre encore de nombreuses perspectives. Nous avons vu que le paléomagnétisme se révèle être un outil puissant pour l’étude des conditions internes profondes de s astéroïdes et apporte des preuves paléogéophysiques de différenciation des planétésimaux dans les quelques premiers Ma de la formation du système solaire. Cette différenciation précoce des astéroides pourrait générer des noyaux liquides qui, par convection vigoureuse pendant un cours laps de temps (de l'ordre d'un million d'années), auraient initié des dynamos responsables du paléomagnétisme enregistré dans certaines météorites. Cette possibilité mérite d’être étudiée avec le plus grand intérêt car elle peut nous donner des informations/contraintes sur la limite inférieure de la taille des noyaux nécessaire pour initier une dynamo. De plus, les nouvelles données paléomagnétiques, la diversité croissante des échantillons extraterrestres disponibles à l’étude et les progrès dans leur caractérisation (en particulier la chronologie des événements thermiques, la thermochronométrie) sont autant de facteurs qui contribuent à améliorer notre compréhension de la nature et de la chronologie des processus qui opéraient dans le système solaire primitif.

138

CHAPITRE 5 : CONCLUSIONS

139

BIBLIOGRAPHIE

140

Acton, G., Qing-Zhu, Y., Verosub, K., Jovane, L., Roth, A., Jacobsen, B., Ebel, D.S., 2007. Micromagnetic coercivity distributions and interactions in chondrules with implications for paleointensities of the early solar system. J. of Geophys. Res. 112, B03S90. Acuña, M.H., Connerney, J.E.P., Ness, N.F., Lin, R.P., Mitchell, D., Carlson, C.W., McFadden J., Anderson, K.A, Rème, H., Mazelle, C., Vignes, D., Wasilewski, P., Cloutier P., 1999. Global Distribution of Crustal Magnetization Discovered by the Mars Global Surveyor MAG/ER Experiment. Science 284, 790-793. Amelin, Y., Krot, A.N., Hutcheon, I.D., Ulyanov, A.A., 2002. Lead Isotopic Ages of Chondrules and Calcium-Aluminum-Rich Inclusions. Science 297, 1678-1683. Amelin, Y., Krot, A.N., 2007. Pb isotopic age of the Allende chondrules. Meteorit. Planet. Sci. 42, 1321 – 1335. Anders, E., 1971. Meteorites and the early solar system, Annu Rev. Astron. Astrophys. 9, 1-34. Anders, E., Grevesse, N., 1989. Abundance of the elements- meteoritic and solar. Geochim. Cosmochim. Acta 53, 197 –214. Anyzeski, V., 1949. The possible significance of magnetism in meteorites. Popular Astronomy, 57, 249- 250. Artemieva, N., Hood, L., Ivanov, B.A., 2005. Impact demagnetization of the Martian crust: Primaries versus secondaries. Geophys. Res. Lett. 32, L22204. Balbus, S.A., 2003. Enhanced Angular Momentum Transport in Accretion Disks. Annu. Rev. Astron. Astrophys. 41, 555-597. Balbus, S.A., 2009; Magnetorotational instability. Scholarpedia 4, 2409. Balbus, S.A., 2011. Magnetohydrodynamics of protostellar Disks. Physical Processes in Circumstellar Disks Around Young Stars, ed. by P. Garcia (University of Chicago Press, Chicago, 237) Banerjee, S.K., Hargraves, R.B., 1971. Natural remanent magnetization of carbonaceous chondrites. Earth Planet. Sci. Lett. 10, 392-396. Banerjee, S.K., Hargraves, R.B., 1972. Natural remanent magnetization of carbonaceous chondrites and the magnetic field in the early solar system. Earth Planet. Sci. Lett. 17, 110-119. Bezaeva, N.S., Gattacceca, J., Rochette, P., Sadykov, R.A., Trukhin, V.I., 2010. Demagnetization of terrestrial and extraterrestrial rocks under hydrostatic pressure up to 1.2 GPa. Phys. Earth Planet. Int. 179, 7 –20. Binzel, R., Xu, S., 1993. Chips off of Asteroid 4 Vesta : Evidence for the Parent Body of Basaltic Achondrite Meteorites. Science 260, 186-191. Bischoff, A., 2001. Meteorite classification and the definition of new chondrite classes as a result of successful meteorite search in hot and cold deserts, Planetary and Space Science 49, 769-776. Bischoff, A., Vogel, N., Roszjar, J., 2011. The Rumuruti chondrite group. Chemie der Erde 71, 101-133. Bland, P.A., Smith, T.B., Jull, A.J.T., Berry, F.J., Bevan, A.W.R., Cloudt, S., Pillinger, C.T., 1996. The flux of meteorites to the Earth over the last 50000 years. Mon. Not. R. Astron. Soc. 283, 551-565. Bland, P.A., Zolensky, M.E., Benedix, G.K., Sephton, M.A., 2006. Weathering of chondritic meteorites. In Meteorites and the Early Solar System II, ed. by D.S. Lauretta, H.Y. McSween, (The University of Arizona Press, Tucson), 853-867. Blewett, D.T., Hawke, B.R., Richmond, N.C., Hughes, C.G., 2007. A magnetic anomaly associated with an albedo feature near Airy crater in the lunar nearside highlands. Geophys. Res. Lett. 34, L24206. Bouvier, A., Wadhwa, M., 2010. The age of the Solar System redefined by the oldest Pb –Pb age of a meteoritic inclusion. Nature Geoscience 3, 637-641. Bouvier, J., Alencar, S. H. P., Boutelier, T., Dougados, C., Balog, Z., Grankin, K., Hodgkin, S. T., Ibrahimov, M. A., Kun, M., Yu. Magakian, T., Pinte, C., 2007. Magnetospheric accretion-ejection processes in the classical T Tauri star AA Tauri. Astronomy & Astrophysics 463, 1017-1028. Brearley, A.J., 1995. Aqueous alteration and brecciation in Bells, an unusual, saponite-bearing, CM chondrite. Geochim. Cosmochim. Acta 59, 2291-2317. Brearley, A.J., Jones, R.H., 1998. Chondritic meteorites. Rev. Mineral. 36, 313-398. Brearley, A.J., 2006. The action of water. In ‘Meteorites and the Early Solar System II” (Eds Dante Lauretta, H.Y. McSween Jr), Arizona University Press, 587-624. Brecher, A., Vaughan, D.J., Burns, R.G., Morash, K.R., 1973. Magnetic and Mossbauer studies of Apollo 16 rock chips 60315,51 and 62295,27. Geophy and Cosmo Acta 3, 2991-3001. Brecher, A., Arrhenius, G., 1974. The paleomagnetic record in carbonaceous chondrites: natural remanence and magnetic properties. J. Geophys. Res. 79, 14, 2081-2106.

141

Browning, L., McSween, H., Zolensky, M., 1996. Correlated alteration effects in CM carbonaceous chondrites. Geochim. Cosmochim. Acta 60, 2621 –2633. Bullock, E.S., Grady, M.M., Gounelle, M., Russell, S.S., 2007. Fe-Ni Sulphides as Indicators of Alteration in CM Chondrites. Lunar Planet. Sci. 38, abstract#2057. Cacciamani, G., De Keyzer, J., Ferro, R., Klotz, U.E., Lacaze, J., Wollants, P., 2006. Critical evaluation of the Fe-Ni, Fe-Ti, and Fe-Ni-Ti alloy systems. Intermetallics 14, 1312-1325. Carporzen, L., Weiss, B.P., Elkins-Taton, L.T., Shuster, D.L., Ebel, D., Gattacceca, J., 2011. Magnetic evidence for a partially differentiated carbonaceous chondrite parent body. Proc. National Acad. Sci. 108, 6386-6389. Chabot, N.L., Haack, H. 2006. Evolution of Asteroidal cores. In Lauretta, D.S., Mc Sween, H.Y., (eds) Meteorites and the Early Solar System II. University of Arizona, 747-771. Chen, G., Ahrens, T.J., Hide, R., 1995. Hypervelocity impacts and magnetization of small bodies in the solar system. Icarus 115, 86-96. Chizmadia, L.J., Brearley, A.J., 2008. Mineralogy, aqueous alteration, and primitive textural characteristics of fine-grained rims in the Y-791198 CMS carbonaceous chondrite: TEM observations and comparison to ALHA81002. Geochim. Cosmochim. Acta 72, 602 –625. Choi, B.G., McKeegan, K.D., Leshin, L.A., Wasson, J.T., 1997. Origin of magnetite in oxidized CV chondrites: in situ measurements of oxygen isotope compositions of Allende magnetite and olivine. Earth Planet. Sci. Lett. 146, 334-349. Cisowski, S.M., Collinson, D.W., Runcorn, S.K., Stephenson, A., 1983. A review of lunar paleointensity data and implication for the origin of lunar magnetism. LPSC XIII, Journal of Geophy research 88, 691-704. Cisowski, S.M., 1986. Magnetic studies on Shergotty and other SNC meteorites. Geoch Cosmo. Act 50, 1043 –1048. Cisowski, S. M., 1987. Magnetism of meteorites. Academic, San Diego, Calif., Geomagnetism 2, 525 –560. Clayton, R.N., Onuma, N., Mayeda, T.K., 1976. A classification of meteorites based on oxygen isotopes. Earth Planet. Sci. Lett. 30, 10-18. Clayton, R.N., Onuma, N., Grossman, L., Mayeda, T.K., 1977. Distribution of the pre-solar component in Allende and other carbonaceous chondrites. Earth Planet. Sci. Lett. 34, 209-224. Clayton, R.N., 1993. Oxygen isotopes in meteorites. Ann. Rev. Earth Planet. Sci. 21, 115 –149. Coleman, P.J., Lichtenstein, B.R., Russell, C.T., Schubert, G., Sharp, L., R., 1973. Apollo 16 Prelim. Sci. Rep. NASA SP-315, 23_1-23_13. Collinson, D.W., 1993. Magnetism of the Moon – a Lunar core dynamo or impact magnetization? Survey in Geophysics 14, 89-118. Connelly, J.N., Bizzarro, M., Krot, A.N., Nordlund, A., Wielandt, D., Ivanova, M.A., 2012. The absolute chronology and thermal processing of solids in the solar protoplanetary disk. Science 338, 651- 655. Cournède, C., Gattacceca, J., Rochette, P., 2012. Magnetic study of large Apollo samples: Possible evidence for an ancient centered dipolar field on the Moon. Earth Planet. Sci. Lett. 331-332, 31- 42. Cournède, C., Gattacceca, J., Rochette, P., Weiss, B., Zanda, B., 2013. An Early solar system magnetic field recorded in CM chondrites. En préparation Crawford, D. A., Schultz, P. H., 1999. Electromagnetic properties of impact-generated plasma, vapor and debris. International Journal of Impact Engineering, 23 169-180. Dankers, P.H.M., Zijderveld, J.D.A., 1981. Alternating field demagnetization of rocks, and the problem of gyromagnetic remanence. Earth Planet. Sci. Lett. 53, 89-92. De Leuw, S., Rubin, A.E., Schmitt, A.K., Wasson, J.T., 2009. 53Mn-53Cr systematics of carbonates in CM chondrites: Implications for the timing and duration of aqueous alteration. Geochim. Cosmochim. Acta 73, 7433-7442. Doell, R.R., Gromme, C.S., Thorpe, A.N., Senftle, E.E., 1970. Magnetic studies of lunar samples. Science 167, 695-697. Dunlop, D.J., Özdemir, Ö., 1997. Rock Magnetism: Fundamentals and Frontiers, Cambridge University Press, New York, London and Cambridge, 573. Dunlop, D.J., Ozdemir, O., Clark, D.A., Schmidt, P.W., 2000. Time-temperature relations for the remagnetization of pyrrhotite (Fe7S8) and their use in estimating paleotemperatures. Earth Planet. Sci. Lett. 176, 107-116.

142

Dwyer, C.A., Stevenson, D.J., Nimmo, F., 2011. A long-lived lunar dynamo driven by continuous mechanical stirring. Nature 479, 212 –214. Dyal, P., Parkin, C.W., Sonett, C.P., 1970. Apollo 12 magnetometer: Measurements of a steady magnetic field on the surface of the moon. Science 196, 762-764. Dyal, P., Daily, W.D., 1978. Crustal Magnetic Scale Sizes Measured at Apollo Landing Sites. Lunar and Planetary Science IX, 273-275. Elkins-Tanton, L.T., Weiss, B.P., Zuber, M.T., 2011. Chondrites as samples of differentiated planetesimals. Earth Planet. Sci. Lett. 305, 1-10. Elmaleh, A., Tarantino, S.C., Zema, M., Devouard, B., Fialin, M., 2012. The low-temperature magnetic signature of Fe-rich serpentine in CM2 chondrites: Comparison with terrestrial cronstedtite and evolution with the degree of alteration. Geochem. Geophys. Geosys. 13, 5, Q05Z42. Enkin, R.J., Dunlop, D.J., 1988. The demagnetization temperature necessary to remove viscous remanent magnetization. Geophys. Res. Lett. 15, 514 –517. Evans, N.J., Dunham, M.M., Jorgensen, J.K., Enoch, M.L., Merin, B., Van Dishoeck, E.F., Alcala, J.M., Myers, P.C., Stapelfeldt, K.R., Huard, T.L., Allen, L.E., Harvey, P.M., Van Kempen, T., Blake, G.A., Koerner, D.W., Mundy, L.G., Padgett, D.L., Sargent, A.I., 2009. The spitzer c2d legacy results : star formation rates and efficiences; evolution and lifetimes. Astrophys. J. Suppl. Ser. 181, 321-350. Folco, L., Rochette, P., Gattacceca, J., Perchiazzi, N., 2006. In situ identification, pairing, and classification of meteorites from Antarctica through magnetic susceptibility measurements. Meteori. Planet. Sci. 41, 343-353. Fonton, S.S., 1954. Meteoric dust from sweepings of the fall of the Sikhote-Alin iron meteoritic shower. Meteoritika 11, 121-122. Fu, R.R., Weiss, B.P., 2012. Detrial remanent magnetization in the solar nebula. Journal of Geophy Res. 117, E02003. Fu, R.R., Weiss, B.P., Shuster, D.L., Gattacceca, J., Grove, T.L., Suavet, C., Lima, E.A., Li, L., Kuan, A.T., 2012. An Ancient Core Dynamo in Asteroid Vesta. Science, 338, 238-241. Fuller, M., 1974. Lunar magnetism. Rev. Geophys. Space Physics 12, 23-70. Fuller, M., Cisowski, S.M., 1987. Lunar paleomagnetism. In: Jacobs, J. (Ed.), Geomagnetism 2, 307 –455. Fuller, M., 1998. Lunar magnetism-A retrospective view of the Apollo sample magnetic studies. Phys. Chem. Earth 23, 725-735. Fuller, M.D., 2007. Paleomagnetism, extraterrestrial. Encyclopedia of geomagnetism and paleomagnetism, ed. By D. Gubbins, E. Herrero-Bervera (Springer, Dordrecht, 2007) 788-801. Garcia, R., Gagnepain-Beyneix, J., Chevrot, S., Lognonne, P., 2010. The lunar core revealed by reflected seismic waves : constraints on the deep Moon seismic structure. EOS Transactions, AGU Fall meeting, Abs, #GP43B. Garrick-Bethell, I., Weiss, B.P., Shuster, D.L., Buz, J., 2009. Early lunar magnetism. Science 323, 356 –359. Garrick-Bethell, Y., Weiss, B.P., 2010. Kamacite blocking temperature and application to lunar magnetism. Earth and planetary science Letters 294, 1-7. Gattacceca, J., Rochette, P., Denise, M., 2003. Magnetic properties of a freshly fallen LL ordinary chondrite: the Bensour meteorite, Physics of the Earth and Planetary Interiors 140, 343-358. Gattacceca, J., Rochette, P., 2004. Toward a robust normalized magnetic paleointensity method applied to meteorites Earth and Planet. Sci. Lett. 277, 377-393. Gattacceca, J., Rochette, P., Denise, M., Consolmagno, G., Folco, L., 2005. An impact origin for the foliations of chondrites. Earth Planet. Sci. Lett.234, 351 –368. Gattacceca, J., Berthe, L., Boustie, M., Vadeboin, F., Rochette, P., De Resseguier, T., 2008. On the efficiency of shock magnetization processes. Physics of the Earth and Planetary Interiors 166, 1- 10. Gattacceca, J., Boustie, M., Lima, E., Weiss, B.P., De Resseguier , T., Cuq-Lelandais, J.P., 2010a. Unraveling the simultaneous shock magnetization and demagnetization of rocks. Physics of the Earth and Planetary Interiors 182, 42-49. Gattacceca, J., Boustie, M., Hood, L., Cuq-Lelandais, J.-P., Fuller, M., Bezaeva, N., de Resseguier, T., Berthe, L., 2010b. Can the lunar crust be magnetized by shock: Experimental groundtruth. Earth Planet. Sci. Lett. 299, 42-53. Gattacceca, J., Hewins, R.H., Lorand, J-P., Rochette, P., Lagroix, F., Cournède, C., Uehara, M., Pont, S., Sautter, V., Scorzelli, R.B., Hombourger, C., Munayco, P., Zanda, B., Chennaoui, H., Ferrière, L.,

143

2013a. Opaque minerals, magnetic properties, and paleomagnetism of the Tissint Martian meteorite. Meteoritics and Planet. Sci., 1-18. Gattacceca, J., Weiss, B.P., Gounelle, M., Lima, E.A., Rochette, P., 2013b. More evidence for a partially differentiated CV parent body from the meteorite Kaba. 44 th Lunar and Planet. Sci. Conf., Abs. 1721. Gladman, B.J., Migliorini, F., Morbidelli, A., Zappala, V., Michel, P., Cellino, A., Froeschle, C., Levison, H.F., Bailey, M., Duncan, M., 1997.Dynamical lifetimes of objects injected into asteroid belt resonances. Science 277, 197. Gounelle, M., Spurny, P., Bland, P.A., 2006. The atmospheric trajectory and orbit of the Orgueil meteorite. Meteoritics and Planet. Sci., 41, 135-105. Gounelle, M., Morbidelli, A., Bland, P.A., Sephton, M.A., Young, E.D., Spurny, P., 2008. Meteorites from the outer solar system? In The Solar System Beyond Neptune (eds A. Barucci, H. Boehnhardt, D. Cruikshank, A. Morbidelli), 525-541. Arizona University Press: Tucson. Lett. 3, 351-354. Guenther, E.W., Lehmann, H., Emerson, J.P., Staude, J., 1999. Measurements of magnetic field strength on T Tauri stars. Astron. Astrophys. 341, 768-783. Guo, W., Eiler, J.M., 2007. Temperatures of aqueous alteration and evidence for methane generation on the parent bodies of CM chondrites. Geochim. Cosmechim. Acta 71, 5565-5575. Guskova, Y.G., 1970. Magnetic properties of meteorites in the collection of the Committee on meteorites of the AS USSR. Meteoritika 30, 74-87. Guskova, Y.G.,1976. The magnetic properties of meteorites. NASA, 1-108. Haisch, K.E., Lada, E.A., Lada, C.J., 2001. Disk frequencies and lifetimes in young clusters. The astrophysical journal 553, 153-156. Halekas, J.S., Lin, R.P., Mitchell, D.L., 2003. Magnetic fields of lunar multi-ring impact basins. Meteor. Planet. Sci. 38, 565-578. Halliday, L, Blackwell, A.T., Griffin, A.A., 1989. The flux of meteorites on the Earth's surface. Meteoritics 24, 173-78. Hanowski, N.P., Brearley, A.J., 2001. Aqueous alteration of chondrules in the CM carbonaceous chondrite, Allan Hills 81002: implications for parent body alteration. Geochim. Cosmochim. Acta 65, 495-518. Hewins, R.H., Bourot-Denis, M., Zanda, B., Leroux, H., Barrat, J.-A., Humayun, M., Göpel, C., Greenwood, R.C., Franchi, I.A., Pont, S., Lorand, J.-P., Cournède, C., Gattacceca, J., Rochette, P., Kuga, M., Marrocchi, Y., Marty, B., 2014. The Paris meteorite, the less altered CM chondrite so far. M. Geochim. Cosmochim. Acta 124, 190-222. Hide, R., 1972. Comments on the Moon's magnetism. Moon 4, 39. Hood, L.L., Cisowski, C.S., 1983. Paleomagnetism of the Moon and meteorites. Rev Geophys. Space phys 21, 676-684. Hood, L.L, 1987. Magnetic field and remanent magnetization effects of basin-forming impacts on the Moon. Geophys. Res. Lett. 14, 844-847. Hood, L., Mitchell, D., Lin, R., Acuña, M., Binder, A., 1999. Initial measurements of the lunar induced magnetic dipole moment using Lunar Prospector magnetometer data. Geophys. Res. Lett. 26, 2327 –2330. Hood, L., Richmond, N.C., Pierazzo, E., Rochette, P., 2003. Distribution of crustal magnetic fields on Mars: shock effects of basin-forming impacts. Geophys. Res. Lett. 30(6), 1281. Hood, L., Artemieva, N., 2008. Antipodal effects of lunar basin-forming impacts: Initial 3D simulations and comparisons with observations. Icarus 193, 485-502. Hood, L.L., 2011. Central magnetic anomalies of Nectarian-aged lunar impact basins: probable evidence for an early core dynamo. Icarus 211, 1109 –1128. Howard, K.T., Benedix, G.K., Bland, P.A., Cressey, G., 2011. Modal mineralogy of CM chondrites by X- ray diffraction (PSD-XRD): Part 2. Degree, nature and settings of aqueous alteration. Geochim. Cosmochim. Acta 75, 2735-2751. Hu, S., Appel, E., Hoffmann, V., Schmahl, W.W., Wang, S., 1998. Gyromagnetic remanence acquired by greigite (Fe 3S4) during static three-axis alternating field demagnetization. Geophys. J. Int. 134, 831-842. Huss, G.R., Rubin, A.E., Grossman, J.N., 2006. Thermal Metamorphism in Chondrites. In Lauretta, D.S., Mc Sween, H.Y., (eds) Meteorites and the Early Solar System II. University of Arizona, 567 –586.

144

Hyman, M., Rowe, M.W., 1986. Saturation magnetization measurements of carbonaceous chondrites. 1986. Meteoritics 21, 1. Johansen, A., 2009. The role of magnetic fields for planetary formation. In Cosmic Magnetic Fields: From Planets, to Stars and Galaxies, ed. by K.G. Strassmeier, A.G. Kosovichev, J.E. Beckman, (Proc. IAU Symp. 259), 119-128. Kallemeyn, G.W., Rubin, A.E., Wasson, J.T., 1996. The compositional classification of chondrites: VII. The R chondrite group. Geochemica and Cosmochimica Acta 60, 2243-2256. Khan, R., Shirai, N., Ebihara, M., 2013. Bulk composition of R chondrites: New data. 44th Lunar and Planet. Sci. Conf., abstract 2059. Kimura, M., Grossman, J.N., Weisberg, M.K., 2011. Fe-Ni metal and sulfide mineral in CM chondrites: An indicator of thermal history. Meteoritics and Planet. Sci. 46, 3, 431-442. Kitajima, F., Nakamura, T., Takaoka, N., Murae, T., 2002, Evaluating the thermal metamorphism of CM chondrites by using the pyrolytic behavior of carbonaceous macromolecular matter. Geochim. Cosmochim. Acta 66, 1, 163-172. Kleine, T., Munker, 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., Touboul, M., Bourdon, B., Nimmo, F., Mezger, K., Palme, H., Jacobsen, S.B., Yin, Q.-Z., Halliday, A.N., 2009. Hf –W chronology of the accretion and early evolution of asteroids and terrestrial planets. Geochim. Cosmochim. Acta 73, 5150-5188. Kletetschka, G., Kohout, T., Wasilewski, P.J., 2003. Magnetic remanence in the Murchison meteorite. Meteoritics and Planet. Sci. 38, 3, 399-405. Kohout, T., Kletetschka, G., Donadini, F., Fuller, M., Herrero-Bervera, E., 2008. Analysis of the natural remanent magnetization of rocks by measuring th e efficiency ratio through alternating field demagnetization spectra. Studia Geophysi ca et Geodaetica 52, 225-235. Konopliv, A., Binder, A., Hood, L., Kucinskas, A., Sjogren, W., Williams, J., 1998. Improved gravity field of the Moon from Lunar Prospector. Science 281, 1476 –1480. Lanoix, M., Strangway, D.W., Pearce, G.W., 1978. The primordial magnetic field preserved in chondrules of the Allende meteorite. Geophysical research Letters 5, 73-76. Larson, E.E., Watson, D.E., Herndon, J.M., Rowe, M.W., 1973. Partial A.F. demagnetization studies of 40 meteorites. J. Geomag. Geoelectr. 25, 331-338. Le Bars, M., Wieczorek, M.A., Karatekin, O., Cébron, D., Laneuville, M., 2011. An impact driven dynamo for the early Moon. Nature 479, 215 –218. Leblanc, F., Witasse, O., Winningham, J., Brain, D., Lilensten, J., Blelly, P.-L., Frahm, R. A., Halekas, J. S., Bertaux, J. L., 2006. Origins of the Martian aurora observed by Spectroscopy for Investigation of Characteristics of the Atmosphere of Mars (SPICAM) on board Mars Express. J. Geophys. Res. 111, A09313. Levinson-Lessing, 1952, in selected works, vol 3 (USSR academy of sciences press), 363-366. Levy, E.H., Sonett, C.P., 1978. Meteorite Magnetism and Early Solar-System Magnetic Fields. Protosolar and Planets, T.Gehrels, (University of Arizona press, Tucson), 516-532. Lillis, R.J., Frey, H.V., Manga, M., Mitchell, D.L., Lin, M.H., Acuña, M.H., Bougher, S.W., 2008. An improved crustal magnetic field map of Mars from electron reflectometry : Highland volcano magnetic history and the end of the Martian dynamo.Icarus 194, 575-596. Lillis, R.J., Robbins, S., Manga, M., Halekas, J.S., Frey, H.V., 2013. Time history of the Martian dynamo from crater magnetic field analysis. Journal of geophysical research: planets 118, 1-24. Lindgren, P., Lee, M.R., Sofe, M.R., Zolensky, M.E., 2013. Clasts in the CM2 carbonaceous chondrite Lonewolf Nunataks 94101: Evidence for acqueous alteration prior to complex mixing. Met. Planet. Sci. 48, 1074-1090. Louzada, K.L., Stewart, S.T., Weiss, B.P., 2007. Effect of shock on the magnetic properties of pyrrhotite, the Martian crust, and meteorites. Geophys Res. Lett. 34, L05204. Louzada, K.L., Weiss, B.P., Maloof, A.C., Stewart, S.T., Swanson-Hysell, N.L., Soule, S.A., 2008. Paleomagnetism of Lonar impact crater, India. Earth Planet. Sci. Lett. 275, 308-319. Louzada, K.L., Stewart, S.T., Weiss, B.P., Gattacceca, J., Bezaeva, N.S., 2010. Shock and static pressure demagnetization of pyrrhotite and implications for the Martian crust. Earth Planet. Sci. Lett. 290, 90-101. Lovering, J.F., 1959. The magnetic field in a primary meteorite body. American Journal of Science 257, 271-275.

145

Mason, B., 1962. The carbonaceous chondrites. Space Sci. Rev. 1, 621-646. Mitchell, D.L., Halekas, J., Lin, R.P., Frey, S., Hood, L.L., Acuña, M.H., Binder, A.B., 2008. Global mapping of lunar crustal magnetic fields by Lunar Prospector. Icarus 194, 401-409. Miyamoto, M., 1991. Thermal metamorphism of CI and CM carbonaceous chondrites: An internal heating model. Meteoritics 26, 111-115. Muxworthy, A., Heslop, D., 2011. A Preisach method for estimating absolute paleofield intensity under the constraint of using only isothermal measurements: 1. Theoretical framework, J. Geophys. Res ., 116, B04103. Nagata, T., 1970. Effects of a uniaxial compression on remanent magnetizations of igneous rocks. Pure and Applied Geophysics 78, 100 –109. Nagata, T., 1973. Piezo-remanent magnetization of Lunar rocks. Pure and applied geophysics 110, 2022- 2030. Nagata, T., Sugiura, N., 1977. Paleomagnetic field intensity derived from meteorite magnetization. Phys. Earth Planet. Inter. 13, 373-379. Nagata, T., 1979a. Meteorite magnetism and the early solar system magnetic field, Phys. Earth Planet. Inter. 20, 324-341. Nagata, T., 1979b. Natural remanent magnetization of the fusion crust of meteorites. Mem. Natl. Inst. Polar Res. 15 (Spec. Issue), 253-272. Nakamura, T., 2005. Post-hydration thermal metamorphism of carbonaceous chondrites. Jour. Mine. Petro. Sci. 100, 260-272. Nyquist, L.E., Bogard, D.D., Shih, C.Y., Greshake, A., Stöffler, D., Eugster, O., 2001. Ages and geologic histories of Martian meteorites. Space Sci. Rev. 96, 105 –164. Pechersky, D.M., Markov, G.P., Tsel’movich, V.A., Sharonova, Z.V., 2012. Extraterrestrial magnetic minerals. Physics of the solid Earth 48, 653-669. Pilipp, W., Hartquist, T.W., Morfill, G.E., Levy E.H., 1998. Chondrule formation by lightning in the Protosolar Nebula, Astron. Astrophys., 331, 121-146. Pilkington, M., and Grieve, R.A.F., 1992. The geophysical signature of terrestrial impact craters. Rev. Geophysics 30, 161-181. Pravdivtseva, O., Meshik, A., Hohenberg, C.M., 2013. The I-Xe record: early onset of aqueous alteration in magnetites separated from CM and CV carbonaceous chondrites. Lunar and Planetary Sci. Conf. 44, abstract #3104. Pullaiha, G., Irving, E., Buchan, K.L., Dunlop, D.J., 1975. Magnetization changes caused by burial and uplift. Earth Planet. Sci. Lett. 28, 133-143. Richmond, N.C., Hood, L.L., 2008. A preliminary global map of the vector lunar crustal magnetic field based on Lunar Prospector magnetometer data. J. Geophys. Res. 113, E02010. Rochette, P., Lorand, J.-P., Fillion, G., Sautter, V., 2001. Pyrrhotite and the remanent magnetization of SNC meteorites: a changing perspective on Martian magnetism. Earth Planet. Sci. Lett. 190, 1-12. Rochette, P., Sagnotti, L., Bourot-Denise, M., Consolmagno, G., Folco, L., Gattacceca, J., Osete, M.L., Pesonen, L., 2003a. Magnetic classification of stony meteorites : 1. Ordinary chondrites. Meteor. Planet. Sci. 38, 251-286. Rochette, P., Fillion, G., Ballou, R., Brunet, F., Ouladdiaf, B., Hood, L., 2003b. High pressure magnetic transition in pyrrhotite and impact demagnetization on Mars. Geophys. Res. Lett. 30, 1683. Rochette P., Gattacceca J., Bonal L., Bourot-Denise M., Chevrier V., Clerc J.-P., Consolmagno G., Folco L., Gounelle M., Kohout T., Pesonen L., Quirico E., Sagnotti L., Skripnik A., 2008. Magnetic Classification of Stony Meteorites: 2. Non-Ordinary Chondrites. Meteor. Planet. Sci. 43, 959-980. Rochette P., Gattacceca J., Chevrier V., Lorand J.P., 2005. Matching Martian crustal magnetization and meteorite magnetic properties. Meteor. Planet. Sci. 40, 529-540. Rochette P., Weiss, B.P., Gattacceca J., 2009. Magnetism of extraterrestrial materials. Elements, 5, 229- 234. Rochette, P., Gattacceca, J., Ivanov, A.V., Nazarov, M.A., Bezaeva, N.S., 2010. Magnetic properties of lunar materials: Meteorites, Luna and Apollo returned samples. Earth Planet. Sci. Lett. 292, 383- 391. Rubin, A.E., Trigo-Rodriguez, J.M., Huber, H., Wasson, J.T., 2007. Progressive aqueous alteration of CM carbonaceous chondrites. Geochim. Cosmochim. Acta 71, 2361 –2382.

146

Ruf, C., Renno, N.O., Kok, J.F., Bandelier, E., Sander, M.J., Gross, S., Skjerve, L., Cantor, B., 2009. Emission of non-thermal microwave radiation by a Martian dust storm. Geophys. Research Lett. 36, L13202. Runcorn, S. K., 1996. The formation of the lunar core. Geochim. Cosmochim. Acta 60, 1205-1208. Russell, S.S., Hartmann, L., Cuzzi, J., Krot, A.N., Gounelle, M., Weidenschilling, S., 2006. Timescales of the Solar Protoplanetary Disk. In “Meteorites and the Early Solar System II” (Eds Dante Lauretta, H.Y. McSween Jr), Arizona University Press, 233-251. Sano, T., Inutsuka, S.I., Turner, N.J., Stone, J.M., 2004. Angular momentum transport by magnetohydrodynamic turbulence in accretion Disks: Gas pressures dependence of the Saturation level of the magnetorotational instability. Astrophys. J. 605, 321-339. Sato, Y., Nakamura, N., 2010. Shock melt veins of Tenham chondrite as a possible paleomagnetic recorder: Rock magnetism and high-pressure minerals. Geoch. Geoph. Geosys. 11, Q04Z16. Schulze, H., Bischoff, A., Palme, H., Spettel, B., Dreibus, G., Otto, J., 1994. Mineralogy and chemistry of Rumuruti: The first meteorite fall of the new R chondrite group. Meteoritics 29, 275-286. Schulze, H., Stoffler, D., 1997. Shock Classification of Ordinary Chondrites: Statistics, Effects on Noble Gases, and Implications for Parent Body Processes. Meteoritics 32, A116. Scott, E.R.D., Keil, K., Stoffler, D., 1992. Shock metamorphism of carbonaceous chondrites. Geoch. Cosmoch. Acta 56, 4281-4293. Sears, D.W., 1975. Temperature gradients in meteorites produced by heating during atmospheric passage. Mod. Geol. 5, 155-164. Sephton, M.A., 2002. Organic compounds in carbonaceous meteorites. Natl Prod. Rep. 19, 292-311. Shea, E., Weiss, B., Cassatta, W., Shuster, D., Tikoo, S., Gattacceca, J., Grove, T., Fuller, M., 2012. A long- lived lunar core dynamo. Science 335, 453-456. Spencer, J.R., Lebofsky, L.A., Sykes, M.V., 1989. Systematic biases in radiometric diameter determinations. Icarus 78, 337 –354. Srnka, L.J., 1977. Spontaneous magnetic field generation in hypervelocity impacts. Proc. Lunar Sci. Conf. 8, 785-792. Stacey, F.D., Lovering, J.F., 1959. Natural magnetic moments of two chondritic meteorites. Nature 183, 529-530. Stacey, F.D., Lovering, J.F., Parry, L.G., 1961. Thermomagnetic properties, natural magnetic moments, and magnetic anisotropies of some chondritic meteorites. J. Geophys. Res. 66, 1523-1534. Stacey, F.D., 1976. Paleomagnetism of meteorites. Annu. Rev. Earth Planet. Sci. 4, 147-157. Stegman, D., Jellinek, A., Zatman, S., Baumgardner, J., Richards, M., 2003. An early lunar dynamo driven by thermo chemical mantle convection. Nature 421, 143-146. Stephenson, A., Collinson, D.W., 1974. Lunar magnetic field palaeointensities determined by an anhysteretic remanent magnetization method. Earth Planet. Sci. Let. 23, 220-228. Stephenson, A., 1993. Three-axis static alternating field demagnetization of rocks and identification of natural remanent magnetization, gyroremanent magnetization, and anisotropy. J. Geophys. Res. 98, 373-381. Sterenborg, M.G., Crowley, J.W., 2013. Thermal Evolution of Early Solar System Planetesimals and the Possibility of Sustained Dynamos. Physics of the Earth and Planetary Interiors 214, 53-73. Stevenson, D.J., 1983. Planetary magnetic fields. Rep. Prog. Phys. 46, 555-620. Stevenson, D.J., 2003. Planetary magnetic fields. Earth and Planet. Sci. 208, 1-11. Stöeffler, D., Keil, K., Scott Edward, R.D., 1991. Shock metamorphism of ordinary chondrites. Geoch. Cosmo. Acta 55, 3845 –3867. Strangway, D.W., 1978. Rock magnetism of Extraterrestrial materials. Lunar and Planet. Institute abstract 113. Suavet, C., Gattatcceca, J., Rochette, P., Perchiazzi, N., Folco, L., Duprat, J., Harvey, R.P., 2009. Magnetic properties of micrometeorites. J. of Geophys. Res. 114, B04102. Suavet, C., Weiss, B.P., Cassata, W.S., Shuster, D.L., Gattacceca, J., Chan, L., Garrick-Bethell, I., Head, J.W., Grove, T.L., Fuller, M.D., 2013. Persistence and origin of the lunar core dynamo. PNAS 110, 8453-8458. Sugiura, N., Strangway, D.W., 1981. The magnetic properties of the Albee meteorite: Evidence for a strong magnetic field in the early solar system. Proc Lunar Sci conf 12, 1243-1256. Sugiura, N., Strangway, D.W., 1983. A palaeomagnetic conglomerate test using the Abee E4 meteorite. Earth Planet. Sci. Lett. 62, 169-179.

147

Sugiura, N., Strangway, D.W., 1987. Magnetic studies of meteorites. In, J.F Kerridge and M.S. Matthews (eds) Meteorites and the Early Solar System. University of Arizona, 595-615. Sugiura, N., Strangway, D.W., 1988. In ‘Meteorites and the Early Solar System” (Eds J.F. Kerridge, M.S . Mathews), Arizona University Press, 595 –615. Tarduno, J.A, Cottrell, R.D., Nimmo, F., Hopkins, J., Voronov, J., et al. 2012. Evidence for a dynamo in the main group pallasite parent body. Science 338, 939-942. Taylor, S., Lever, J.H., Harvey, R.P., 1998. Accretion rate of cosmic spherules measured at the South Pole. Nature 392, 899 –903. Thellier, E., Thellier, O., 1959. Sur l’intensité du champ magnétique terrestre dans le passé historique et géologique. Annales Geophysique 15, 285 –376. Thomas, C.A., Weiss, B.P., 2006. Magnetism in the Carbonaceous Chondrite Meteorite Allende. Bulletin of the American Astronomical Society 38, 505. Tikoo, S.M., Weiss, B.P., Buz, J., Lima, E.A., Shea, E.K., Melo, G., Grove, T.L., 2012. Magnetic fidelity of lunar samples and implications for an ancient core dynamo. Earth Planet. Sci. Lett. 337-338, 93- 103. Tikoo, S.M., Gattacceca, J., Weiss, B.P., Suavet, C.R., Cournède, C., 2013 . Thermal demagnetization characteristics of shock remanent magnetization and implications for interpreting the paleomagnetism of shocked samples. En préparation. Turner, N.J., Sano, T., 2008. Dead zone accretion flows in protostellar disks. Astrophys. J. 679, 131-134. Uehara, M., Gattacceca, J., Rochette, P., Demory, F., Valenzuela, E.M., 2012. Magnetic study of meteorites recovered in the Atacama desert (Chile): Implications for meteorite paleomagnetism and the stability of hot desert surfaces. Physics of the earth and Planetary interiors 200, 113-123. Vallée, J.P., 2003. Astral magnetic fields - as observed in starforming nurseries, in stars, and in the Solar system. New astronomy reviews 47, 85-165. Vallée, J.-P., 2011. Magnetic fields in the nearby Universe, as observed in solar and planetary realms, stars, and interstellar starforming nurseries. New Astron. Rev. 55, 23 –90. Vernazza, P., Brunetto, R., Strazzulla, G., Fulchignoni, M., Rochette, P., Meyer-Vernet, N., Zouganelis, I., 2006. Asteroid colors: a novel tool for magnetic field detection? The case of Vesta. Astronomy and Astrophysics 451, 43-46. Wasilewski, P.J., Krot, A.N., Meibom, A., 2000. Magnetic Measurements of the Bencubbin/CH-like Chondrite QUE94411. Lunar Plant. Sci. 31, abstract#1454. Wasson, J.T., Wetherill, G.W.,1979. Dynamical, chemical and isotopic evidence regarding the formation locations of asteroids and meteorites. In “Asteroids”, (edt by Gehrels T.) The University of Arizona Press, 926 –974. Weber, R.C., Lin, P.-Y., Garnero, E.J., Williams, Q., Lognonné, P., 2011. Seismic detection of the lunar core. Science 331, 309-312. Weisberg, M.K., Prinz, M., Clayton, R.N., Mayeda, T.K., Grady, M.M., Franchi, I., Pillinger, C.T., Kallemeyn, G.W., 1996. The K (Kakangari) chondrite grouplet. Geoch Cosmo. Acta 60, 4253- 4263 Weisberg, M.K., McCoy, T.J., Krot, A.N., 2006. Systematics and Evaluation of Meteorite Classification. Systematics and evaluation of meteorite classification. In Lauretta, D.S., Mc Sween, H.Y., (eds) Meteorites and the Early Solar System II. University of Arizona, 19-52. Weiss, B.P., Kirschvink, J.L., Baudenbacher, F.J., Vali, H., Peters, N.T., MacDonald, F.A., Wikswo, J.P., 2000. A low temperature transfer of ALH84001 from Mars to Earth. Science 290, 791-795. Weiss, B.P., Vali, H., Baudenbacher, F.J., Kirschvink, J.L., Stewart, S.T., Shuster, D.L., 2002. Records of an ancient Martian magnetic field in ALH84001. Earth Planet. Sci. Lett. 201, 449-463. Weiss, B.P., Berdahl, S., Elkins-Tanton, L.T., Stanley, S., Lima, E.A., Carpozen, L., 2008a. Magnetism on the angrite parent body and the early differentiation of planetesimals. Science 322, 713-716. Weiss, B.P., Fong, L.E., Vali, H., Lima, E.A., Baudenbacher, F., 2008b. Paleointensity of the ancient Martian magnetic field. Geophysi. Res. Letters 35 (23), L23207. Weiss, B.P., Gattacceca, J., Stanley, S., Rochette, P., Christensen, U.R., 2010. Paleomagnetic records of meteorites and early planetesimal differentiation. Space Sci. rev. 152, 341-390. Weiss, B.P., Elkins-Tanton, L.T., 2013. Differentiated Planetesimals and the Parent Bodies of Chondrites. Annu. Rev. Earth Planet. Sci. 41, 529-560. Wieczorek, M., Jolliff, B., Khan, A., Pritchard, M., Weiss, B., 2006. The constitution and structure of the lunar interior, Rev. Mineral. Geochem. 60, 221-364.

148

Wieczorek, M., Weiss, B., Stewart, S., 2012. An impactor origin for lunar magnetic anomalies. Science 335, 1212-1215. Wood, J. A., 1963. On the origin of chondrules and chondrites. Icarus 2, 152-180. Yu, Y., Gee, J.S., 2005. Spinel in Martian meteorite SaU 008: Implications for Martian magnetism. Earth Planet. Sci. Lett. 232, 287-294. Zanda, B., Bourot-Denise, M., Hewins, R.H., Barrat, J.-A., Marrocchi, Y., Pont, S., Gattacceca, J., Greenwood, R.C., Franchi, I.A., 2010. Paris: Une chondrite CM exceptionnellement peu altérée et peu metamorphique. Meteoritics and Planet. Sci. 45, 222-222. Zolensky, M., Barrett, R., Browning, L., 1993. Mineralogy and composition of matrix and chondrule rims in carbonaceous chondrites. Geochimica et Cosmochimica Acta 57, 3123 –3148. Zolensky, M.E., Mittlefehldt, D.W., Lipschutz, M.E., Wang, M.-S., Clayton, R.N., Mayeda, T.K., Grady, M.M., Pillinger, C., Barber, D., 1997. CM chondrites exhibit the complete petrologic range from type 2 to 1. Geochim. Cosmochim. Acta 61, 5099 –5115. Zolensky, M.E., Le, L., 2003. Iron-nickel sulfide compositional ranges in CM chondrites: No simple plan. Lunar Planet. Sci. 34, abstract #1235.

149

ANNEXE

Article 1 : Gattacceca, J., Hewins, R.H., Lorand, J-P., Rochette, P., Lagroix, F., Cournède, C. , Uehara, M., Pont, S., Sautter, V., Scorzelli, R.B., Hombourger, C., Munayco, P., Zanda, B., Chennaoui, H., Ferrière, L., 2013. Opaque minerals, magnetic properties, and paleomagnetism of the Tissint Martian meteorite. Meteoritics & Planetary Science 1 –18. Article 2 : Hewins, R.H., Bourot-Denis, M., Zanda, B., Leroux, H., Barrat, J.-A., Humayun, M., Göpel, C., Greenwood, R.C., Franchi, I.A., Pont, S., Lorand, J.-P., Cournède, C., Gattacceca, J., Rochette, P., Kuga, M., Marrocchi, Y., Marty, B., 2014. The Paris meteorite, the less altered CM chondrite so far. M. Geochim. Cosmochim. Acta 124, 190- 222.

150

151

Meteoritics & Planetary Science 1–18 (2013) doi: 10.1111/maps.12172

Opaque minerals, magnetic properties, and paleomagnetism of the Tissint Martian meteorite

Jer ome^ GATTACCECA1,2*, Roger H. HEWINS 3, Jean-Pierre LORAND 4, Pierre ROCHETTE 1, France LAGROIX5, C ecile COURNEDE 1, Minoru UEHARA 1, Sylvain PONT 3, Violaine SAUTTER 3, Rosa. B. SCORZELLI 6, Chrystel HOMBOURGER 7, Pablo MUNAYCO 6, Brigitte ZANDA 3,8, Hasnaa CHENNAOUI9, and Ludovic FERRI ERE 10

1CNRS, Aix-Marseille Universite, CEREGE UM34, Aix-en-Provence, France 2Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA 3Laboratoire de Min eralogie et Cosmochimie du Mus eum, MNHN, UMR 7202 Paris, France 4Laboratoire de plan etologie et g eodynamique de Nantes, UMR 6112, Universit e de Nantes, Faculte des Sciences, Nantes, France 5Institut de Physique du Globe de Paris, Paris, France 6Centro Brasileiro de Pesquisas F ısicas, Rio de Janeiro, Brazil 7CAMECA SAS, Gennevilliers, France 8Department of Earth and Planetary Sciences, Rutgers University, Piscataway, New Jersey 08854, USA 9Department of Geochemistry of Earth Sciences, Hassan II University, Casablanca, Morocco 10 Natural History Museum, Vienna, Austria *Corresponding author. E-mail: [email protected]

(Received 19 December 2012; revision accepted 24 June 2013)

Abstract–We present a description of opaque minerals, opaque mineral compositions, magnetic properties, and paleomagnetic record of the Tissint heavily shocked olivine-phyric shergottite that fell to Earth in 2011. The magnetic mineralogy of Tissint consists of about 0.6 wt% of pyrrhotite and 0.1 wt% of low-Ti titanomagnetite (in the range ulv ospinel€ 3–15 magnetite 85 –97). The titanomagnetite formed on Mars by oxidation-exsolution of ulvospinel€ grains during deuteric alteration. Pyrrhotite is unusual, with respect to other shergottites, for its higher Ni content and lower Fe content. Iron deficiency is attributed by an input of regolith-derived sulfur. This pyrrhotite has probably preserved a metastable hexagonal monosulfide solution structure blocked at temperature above 300 °C. The paleomagnetic data indicate that Tissint was magnetized following the major impact suffered by this rock while cooling at the surface of Mars from a post-impact equilibrium temperature of approximately 310 °C in a stable magnetic field of about 2 µT of crustal origin. Tissint is too weakly magnetic to account for the observed magnetic anomalies at the Martian surface.

INTRODUCTION samples from Mars available to date. Although they provide a biased sampling of the Martian crust, the The discovery of the intense crustal magnetic study of their magnetic properties may provide clues to anomalies on Mars was a major result of recent the understanding of Martian crustal magnetism, and Martian exploration (Acu na~ et al. 1999; Langlais et al. hence crustal magnetic mineralogy and dynamics. 2010). However, although there is little doubt that a Opaque minerals in Martian meteorites have been, core dynamo was necessary to magnetize the crust, the in general, studied for information on the evolution of lithology of the source of the large crustal remanent Martian magmas and their oxidation state (e.g., magnetization observed today remains puzzling (e.g., Goodrich et al. 2003; Lorand et al. 2005). On the other Quesnel et al. 2009). Martian meteorites are the only hand, the magnetic properties (e.g., Rochette et al.

2005) and paleomagnetism (e.g., Cisowski 1986) of furnace and a maximum applied field of 1.8 T. Moment nakhlites and shergottites provide the following general sensitivity of the VSM is approximately 5 9 10 À9 Am 2. picture: these meteorites contain pyrrhotite and/or The analysis of hysteresis loops provided the ratio of (titano)magnetite as magnetic minerals, and they most saturation remanent magnetization (M RS ) to saturation likely record Martian surface fields of crustal remanent magnetization (M S) and the coercive force (B C). High- origin (Cisowski 1986; Gattacceca and Rochette 2004). field susceptibility ( vHF ) was determined by a linear fit However, these studies usually originate from different for applied fields >0.9 T of the hysteresis loops. disciplinary groups (petrologists and mineralogists on Remanent coercive force (B CR ) was determined by DC one hand, rock magnetists and paleomagnetists on the back-field experiments performed with the VSM. Other other hand) and have never been really integrated. low-temperature measurements were performed with an In this work, we combine the study of opaque MPMS from Quantum Design â. This instrument has a minerals, as well as magnetic properties and the moment sensitivity of 10 À11 Am 2, and can operate in paleomagnetic record of the Tissint olivine-phyric the 1.9 –400 K temperature range. shergottite, to identify which phases carry the The low-field specific susceptibility (written as v in magnetization and how this magnetization was acquired. m3 kg À1) and its anisotropy were measured using Agico This is the first detailed study envisioning the balance of all MFK1 apparatus with sensitivity of 5 9 10 À13 m3, potential magnetic carriers (i.e., Fe-Ti oxides, chromites, operating at 200 A m À1 peak field and a frequency of and Fe-Ni sulfides) in an olivine-phyric shergottite. 976 Hz. The anisotropy of magnetic susceptibility was The Tissint meteorite, which fell in July 2011 in characterized by the shape parameter T (Jelinek 1981), Morocco, is an olivine-phyric shergottite (as defined by and the anisotropy degree P (ratio of maximum to Goodrich 2002), and only the fifth recovered Martian minimum susceptibility). The anisotropy of isothermal meteorite fall (Chennaoui Aoudjehane et al. 2012). It remanent magnetization (IRM) was measured using a offers the opportunity to study the magnetic mineralogy three-position scheme, and was characterized in the and the paleomagnetic signal of a sample from Mars same way by P IRM and T IRM. that has not suffered terrestrial weathering. This pristine All remanence measurements were performed with a character is crucial for two main reasons: the absence of SQUID cryogenic magnetometer (2G Enterprises, model terrestrial weathering minerals, some of which are 755R, with noise level of 10 À11 Am 2) with an attached ferromagnetic (notably those replacing Fe-Ni sulfides; automatic AF 3-axis degausser system (maximum peak Lorand et al. 2005), and also the better preservation of field 170 mT) placed in a magnetically shielded room. To the paleomagnetic signal of such a meteorite fall. avoid the possible acquisition of gyroremanent magnetization during AF demagnetization, we used the MATERIALS AND METHODS Zijderveld-Dunlop correction method (Stephenson 1993) above 40 mT AF. Thermal demagnetization and We have been able to work on 35 samples from the acquisition of partial or total thermoremanent Tissint meteorite, with mass ranging from 30 mg to magnetizations were performed using an MMTD 29 g. The first samples were available in our laboratory furnace, under argon atmosphere above 250 °C. IRM only 6 months after the fall. We have measured the were imparted using a pulse magnetizer from Magnetic following magnetic properties: natural and artificial Measurements. IRM acquisition curves were obtained remanences and their behavior upon thermal pressure using the VSM. Remagnetization under pressure (PRM: and alternating field (AF) demagnetization, hysteresis piezoremanent magnetization, an analog for the parameters at room and low temperatures, anisotropy magnetization acquired by rocks during a shock event in of magnetic susceptibility. We also used magneto- the presence of an ambient field) was studied using a optical imaging coupled with high spatial resolution nonmagnetic pressure cell and experimental settings electron microprobe analyses, and M ossbauer€ described in Gattacceca et al. (2010a). To determine the spectroscopy. In view of the large numbers of acronyms nature of the natural remanent magnetization (NRM), and the nonintuitive units used in used rock magnetism, we compare its coercivity (AF levels at which it is we provide an overview of the magnetic properties demagnetized) and unblocking temperature (temperature discussed in this work in Table S1. at which is it demagnetized) spectra with different Low-temperature (LT) hysteresis measurements artificial remanences: SIRM (saturation IRM), TRM were performed with a Princeton â Micromag Vibrating (thermoremanent magnetization), and pTRM (partial Sample Magnetometer equipped with a LT cryostat and thermoremanent magnetization). a maximum applied field of 1 T. High-temperature All magnetic measurements were performed at hysteresis measurements were also performed on a CEREGE (Aix-en-Provence, France), with the exception Princeton VSM equipped with a high-temperature of M ossbauer€ spectroscopy (at CBPF, Rio de Janeiro, Opaque minerals and magnetism of Tissint 3

Table 1. Selected analyses of pyrrhotite.

Matrix magmatic Del Norte County troilite Incl in olivine sulﬁdes Shock droplets n = 5 n = 36 a n = 10 b Fe 54.1 54.1 56.0 56.5 57.9 57.85 59.1 61.9 61.8 63.45 (0.05) 63.43 (0.23) 63.30 (0.30) Ni 6.9 6.2 4.8 4.8 4.5 3.0 1.4 1.2 1.27 S 38.9 39.5 39.1 37.6 38.57 38.75 39.5 36.7 36.6 36.53 (0.03) 36.56 (0.20) 36.31 (0.20) SEM EMPA SEM SEM SEM EMPA SEM SEM EMPA SEM EMPA EMPA M/S 0.88 0.875 0.89 0.95 0.91 0.89 0.88 0.99 0.99 0.999 0.998 1.000 Results are in wt%. M/S, metal/sulfur (atomic ratio). Del Norte County is a terrestrial troilite used as external standard. SEM, scanning electron microscope standardless procedure (MNHN); EMPA, electron microprobe analyses (Camparis). aChevrier et al. (2011). bLorand et al. (2012).

Brazil), MPMS (at IPGP, Paris, France), and high- Wavelength-dispersive quantitative analyses of temperature VSM measurements (at IRM, Minneapolis, oxide minerals were made on the Cameca SXFive MN, USA). electron microprobe at the Camparis EMP center Magneto-optical observations were performed using (Universit e Paris VI), using 15 keV and 10 nA, and on a system modified from that described by Uehara et al. the SXFiveFE probe at Cameca SAS (Gennevilliers), (2010). In the previous method, the Faraday rotation using 10 keV and 10 nA. The oxides were imaged on a angle, that is a calibrated function of the surface Tescan VEGA II LSU SEM in conventional mode, and magnetic field, was computed from the brightness of the characterized with an SD3 (Bruker) EDS detector at the magneto-optical image. However, the accuracy of this LMCM in the Mus eum National d’Histoire Naturelle, method strongly depends on the brightness of the light Paris (France). Sulfides were analyzed by WDS at the source and the unevenness of the lightning within the Camparis EMP center (a few grains analyzed in one field of view. A new optical microscope system with a polished section) and by an EDS standardless procedure fixed polarizer and a motor-controlled rotatable (Tescan VEGA II LSU SEM, Mus eum National analyzer has been developed to measure the Faraday d’Histoire Naturelle, Paris) after careful imaging of each rotation angle directly. This system takes several images grain in the BSE mode to avoid phase-mixing at different polarizer-analyzer angles ( h) by rotating the contaminations. SEM was preferred over EMP for its analyzer. At small angles (less than Æ5°), the brightness better spatial resolution (a few hundred nanometers of a given pixel ( I) can be approximated by a parabolic versus a few square micrometers), allowing fine function of h, given by I(h) = ah2 + q, where a and q pentlandite-pyrrhotite intergrowths to be resolved. The are positive constants linked to the illumination and the accuracy and precision of SEM analyses were checked quality of the polarizers. The brightness is minimum by replicate analyses of a terrestrial troilite (Del Norte when the analyzer and the polarizer are crossed ( h = 0). County, California; Chevrier et al. 2011; Lorand et al. When a magneto-optically active film (hereafter called 2012). Both analytical procedures agree within “indicator”) is mounted on the magnetized polished analytical uncertainties (Table 1). sample, the polarization plane is rotated in the indicator by a Faraday rotation angle aF that is a function of the OPAQUE MINERAL PETROGRAPHY AND magnetic field intensity. As a result, the above equation COMPOSITION 2 becomes I(h) = a(h À aF) + q. Thus, aF, that is independent from the illumination (a and q), can be The olivine-phyric shergottites (Goodrich 2002), in determined precisely by polynomial fitting of I(h), and a general, contain pyrrhotite and the following oxides: € map of aF (and therefore of the magnetic field) can be chromite, ulv ospinel, ilmenite (Herd et al. 2002; obtained by conducting such fitting over the entire Goodrich et al. 2003; Gross et al. 2011). Petrographic image (1024 9 1024 pixels). The entire procedure, observations of Tissint showed the presence of including image stacking to reduce the signal –noise chromite, ilmenite, ulv ospinel€ (these oxides amounting ratio, is controlled by LabVIEW software. to less than 1 wt% in total), and pyrrhotite (Chennaoui The 57 Fe M ossbauer€ spectroscopy was performed in Aoudjehane et al. 2012). Among these minerals, transmission geometry at room temperature. pyrrhotite (Rochette et al. 2001) and chromite (Yu et al. Measurements were performed at high velocity 2001) can be ferromagnetic at room temperature, (12 mm s À1), using a 90 mg cm À2 absorber of the bulk depending on their composition. It has also been meteorite sample with recording time of 1 month. Normos proposed that peculiar Fe-Cr-Ti spinels may account for code (Brand 1995) was used for the spectral analysis. the magnetic properties of some shergottites (Yu and 4 J. Gattacceca et al.

Gee 2005). We present here detailed petrographic monoclinic-type (Fe 7S8) and hexagonal-type (Fe 9S10 ) observations and chemical analysis of sulfides and endmembers (Fig. 2). This scatter is reproduced in both oxides. A list of opaque minerals discussed in this work analytical procedures (EMP WDS and SEM EDS), is provided in Table S2. although the SEM analyses tend to cluster around the monoclinic-type composition. However, there is no Mineralogy of Fe-Ni Sulfides reason to assume analytical bias as: (1) the SEM analyses of Del Norte County troilite perfectly match Tissint sulfides show some particularities compared the composition of stoichiometric troilite (FeS) and (2) with shergottites and other SNC meteorites investigated both EMP and SEM procedures produce similar so far for sulfide mineralogy. Two distinct populations variation in the M/S ratios (from 0.86 to 0.92 for of sulfide grains are observed, one predating the major magmatic grains and 0.87–1.03 for shock droplets). The shock suffered by Tissint, and a second one formed scatter in Fig. 2 is probably reflecting actual during this shock event (Fig. S1). The first population compositional variations (note the much higher number consists of polyhedral grains, regularly shaped to of SEM analyses compared with EMP analyses) rather anhedral grains between a few micrometers up to than analytical bias generated by either of the two several tens of micrometers in size, as encountered in analytical methods. the other olivine-phyric shergottites (Goodrich 2002). Ni contents are highly homogeneous on the scale of These grains are all located at multiple junctions of individual grains (relative standard deviation better than major silicates (pyroxene, maskelynite, less commonly 10%), but vary according to location of sulfides neighboring chromite; Fig. 1D). Such grains have been (Fig. 2). The most Ni-rich pyrrhotites are hosted in reported in all SNC meteorites studied so far for sulfide inclusions in olivine. Magmatic pyrrhotite grains minerals, regardless of shock intensities (Lorand et al. associated with pyroxene clusters are poorer in Ni (4 – 2005, 2012; Chevrier et al. 2011). They are considered 5 wt% Ni). Shock pyrrhotite droplets show uniformly crystallization products from magmatic sulfide melts depleted Ni contents around 1 –2 wt%. This Ni trapped in the rock porosity during the igneous variation pattern cannot be ascribed to contamination crystallization history. They will be referred to as of spot analyses by micron-sized pentlandite. “magmatic” sulfides in the following. We include in this Pentlandite is very scarce in Tissint pyrrhotite compared group the rare sulfide grains found in magmatic with similar olivine-phyric shergottites. Pentlandite only inclusions of olivine megacrysts (Figs. 1B and 1C). occurs as a few tiny flame-shaped blebs always Some magmatic sulfides are partially remelted (Fig. 1E) concentrated on discontinuities such as fracture planes or seem to contain bubbles (Figs. 1F and 1G). We or grain edges, leaving large pyrrhotite cores free from interpret this as textural evidence that some shock- pentlandite (Figs. 1B and 1C). If the pyrrhotite analyses related sulfides partly derive from erosion of pre- were polluted by such pentlandite flames, then the Ni existing magmatic sulfides that first become cloudy and content should positively correlate with metal/sulfur, as spongy (e.g., Walton and Spray 2003). The shock- documented in the shergottites previously studied for related droplet-shaped blebs are smaller (10 lm on sulfide mineralogy (Lorand et al. 2005). Clearly, this is average), and are organized as network-forming cells not the case (Fig. 2). inside remelted areas of the rock, which we interpret as Tissint pyrrhotite shows metal/sulfur values recondensation of vaporized sulfide that was separated displaced toward the monoclinic-type endmember with from glassy material (e.g., Sutton et al. 2008). Shock a tail of values plotting inside the two-phase field pyrrhotite droplets in glassy areas define a S-loss trend hexagonal (Hpo) + monoclinic (Mpo) pyrrhotite and toward troilite composition, irrespective of Ni content. intermediate pyrrhotite hexagonal structures (Fe 9S10 to This trend is interpreted as partial devolatilization of S, FeS). This is, at first glance, inconsistent with whole- as recently documented in great detail in NWA 2737, a rock magnetic measurements (see below) that seem to highly shocked chassignite (Lorand et al. 2012). rule out low-T ordered monoclinic pyrrhotite structure Some pyrrhotite analyses are given by Chennaoui as the major magnetic carrier. To reconcile observations Aoudjehane et al. (2012), and we present here new and magnetic measurements, we have to consider the analyses on about 80 individual sulfide grains from 36 following alternatives: (1) two-phase pyrrhotite locations throughout two polished sections (Table 1). intergrowth and (2) high-temperature, metastable Compared with all pyrrhotite analyzed so far in disordered hexagonal pyrrhotite solid solution shergottites and nakhlites, Tissint pyrrhotites are poor (monosulfide solid solution; Mss) showing more in metal, i.e., they have a lower (Fe +Ni)/S ratio, flexibility in metal/sulfur ratios. No intergrowths have regardless of the Fe/Ni ratio. Metal/sulfur ratio of been detected in the BSE mode; hence, the first magmatic grains shows considerable scatter between interpretation lacks evidence. By contrast, the second Opaque minerals and magnetism of Tissint 5

shock sulfides

magmatic sulfides

A B C D chr Po Pn

mk E F G

Fig. 1. Backscattered electron images of Tissint sulfides. Scale bar is 20 lm except for A. A) Illustration of the two sulfide textures described in the text. Field of view = 2 mm. B,C) pyrrhotite inclusion inside magmatic inclusion of olivine megacryst. Enhanced contrast in C reveals small pentlandite grains (Po, pyrrhotite; Pn, pentlandite). D) Unmodified magmatic pyrrhotite (mk, maskelynite; chr, chromite; px, pyroxene). E) Partly remelted magmatic sulfide. F) Magmatic sulfide with bubbling zone inside the pyrrhotite. G) Detail of F. 6 J. Gattacceca et al.

1.1

S volatilization

1 Pn-Po mixing line Metal/sulfur (at. ratio) (at. Metal/sulfur 0.9

0.8 012345678 Ni (wt.%) Fig. 2. Compositional features of Tissint Fe-Ni sulfides. Diamonds and triangle: standardless scanning electron microscope energy dispersive analyses calibrated with Del Norte County troilite (triangles: shock melted sulfides); squares: wavelength- dispersive electron microprobe analyses (Camparis). Shaded area: Shergottite pyrrhotite compositional range (free from terrestrial weathering; Lorand et al. 2005). The pentlandite-pyrrhotite (Pn-Po) mixing arrow assumes hexagonal-type (Fe 9S10 ) pyrrhotite endmember contaminated by pentlandite (Fe/Ni at. = 1). alternative is supported by the high Ni contents that are is related to the Ni-poor compositions of shock melts clearly in solid solution inside the pyrrhotite lattice. Mss and/or quick reheating/cooling after the shock peak. is characterized by complete miscibility between Fe and However, our interpretation leaves open other Ni and metal/sulfur ratios covering the whole range of issues such as the influence of several percent Ni on metal/sulfur ratio of natural pyrrhotite, while possessing pyrrhotite magnetic properties or the contributions of a 1C, disordered hexagonal structure (Fleet [2006] and both populations of sulfides (magmatic, shock-derived) references therein). Tissint pyrrhotite compositions to bulk magnetic properties (it is worth pointing out actually plot inside the Mss field at T = 400 °C, while a that some shock sulfides are close to single-domain few analyses protrude into the S-rich side of the Mss grains in terms of grain size). The question that also stability field at 300 °C in the Fe-Ni-S ternary diagrams arises is why Tissint magmatic pyrrhotite is so metal- (Fig. S1). This indicates that Ni diffusion and structural deficient. Only the olivine-phyric shergottite DAG 476 rearrangement were blocked within the temperature has been reported with similar pyrrhotite composition, range 300–400 °C. Cooling below 300 ° should convert yet this clearly results from terrestrial weathering the Ni into pentlandite while unmixing Hpo and Mpo. (Lorand et al. 2005). A terrestrial origin is very unlikely Incipient pentlandite exsolution occurring only on for Tissint, which is an observed fall in a desert area defects, fracture planes, and outer rims indicates and was collected only weeks after its fall, and is totally heterogeneous nucleation processes that are thought to devoid of sulfide alteration products (i.e., Fe- occur well above 250 °C in Mss. However, the kinetics oxyhydroxides). We may surmise that massive input of of re-equilibration strongly depend on pyrrhotite regolith-derived sulfur raised the fugacity of sulfur composition. For instance, metal-poor (vacancy-rich) (sulfate?) to levels at which metal-deficient pyrrhotite compositions inhibit the exsolution of pentlandite became stable (as documented in nakhlites that bear (Etaschmann et al. 2004). To summarize, Tissint isotopic evidence of atmospheric sulfur recycling; pyrrhotites probably have preserved a metastable Mss Chevrier et al. 2011). Then shock remelted both sulfates structure blocked at temperatures above 300 °C. The and sulfides, promoting additional sulfide melt contrasting Ni partitioning between olivine hosted exsolution concomitant with localized sulfur loss, as grains and matrix pyrrhotites reflects the larger amount indicated by the progressive shift of a few pyrrhotite of Ni available for Ni-Fe exchange inside olivine. compositions toward the FeS-type pyrrhotite Conversely, the Ni-poor composition of shock sulfides endmember in melt pockets. If pervasive devolatilization Opaque minerals and magnetism of Tissint 7

Table 2. Selected analyses of chromite, Cr-Ti spinel, ulv ospinel,€ ilmenite, and magnetite in Tissint.

SiO2 0.14 0.24 0.10 0.11 0.08 0.01 0.05 0.07 0.11 nd 0.00 Cr 2O3 59.29 57.58 52.61 19.68 11.08 6.17 1.49 1.28 1.28 1.21 0.27 a TiO2 0.67 0.83 3.04 20.54 25.19 28.21 33.80 28.23 24.87 3.97 52.92 FeO 25.78 31.47 34.30 51.95 57.64 57.44 58.31 62.35 68.01 81.55b 44.42 MgO 5.77 2.44 1.73 1.13 0.80 1.23 1.06 0.93 0.88 0.19 1.72 Al 2O3 7.46 7.07 5.96 3.76 2.46 2.69 1.71 1.91 1.80 2.00 0.00 V2O3 0.50 0.39 0.55 0.44 0.47 0.38 0.41 0.29 0.36 nd 0.34 MnO 0.78 0.69 0.74 0.82 0.74 0.62 0.68 0.68 0.60 nd 0.73 Total 100.42 100.73 99.06 98.43 98.51 96.75 97.50 95.83 97.91 88.91 100.40 Si 0.005 0.008 0.004 0.004 0.003 0.000 0.003 0.003 0.004 0.000 Cr 1.606 1.596 1.497 0.571 0.324 0.183 0.049 0.039 0.038 0.04 0.005 Ti 0.017 0.022 0.082 0.567 0.700 0.796 0.876 0.806 0.693 0.12 0.987 Fe 3+ 0.035 0.041 0.063 0.111 0.148 0.094 0.100 0.249 0.478 1.63 – Fe 2+ 0.704 0.881 0.969 1.484 1.634 1.708 1.799 1.732 1.630 1.11 0.922 Mg 0.295 0.127 0.093 0.062 0.044 0.069 0.058 0.052 0.048 0.01 0.064 Al 0.301 0.292 0.253 0.163 0.107 0.119 0.080 0.086 0.079 0.09 0.000 Mn 0.023 0.020 0.023 0.025 0.023 0.020 0.022 0.022 0.019 0.015 V 0.014 0.011 0.016 0.013 0.014 0.011 0.013 0.009 0.011 0.010

Oxide Ti-Cr Ti-Cr Ti-Cr name Chromite Chromite Chromite spinel spinel spinel Ulvospinel€ Ulvospinel€ Ulvospinel€ Magnetite Ilmenite Spinel 15.24 14.81 12.78 8.22 5.41 5.99 4.05 4.31 3.97 4.73 Chromite 81.27 80.89 75.71 28.86 16.36 9.20 2.46 1.94 1.89 1.92 Ulvospinel€ 1.74 2.22 8.32 57.31 70.73 80.08 88.42 81.20 69.98 11.95a Magnetite 1.75 2.08 3.19 5.61 7.49 4.73 5.07 12.54 24.15 81.40b Oxides are in wt%. Cation formulae are computed based on ideal structures of 3 cations per 4 oxygens for spinel, and 2 cations per 3 oxygens for ilmenite. aMaximum value. bMinimum value, assuming residual overlap. had occurred, then Tissint would be almost S-free and phyric shergottites, such as EETA79001A and NWA 1110 metal-rich, which is not observed. Crustal sulfur (McSween and Treiman 1998; Goodrich et al. 2003). assimilation and solution into shock melts explain the Chromite cores in the lherzolitic shergottites ALH sulfides that decorate melted areas. This external sulfur 77005 and LEW 88516, and in the basaltic shergottite addition is probably responsible for the metal deficiency EET 79001 (McSween and Treiman 1998) have similar in Tissint pyrrhotite. The fact that Ni/Fe ratios compositions to Tissint chromite grains. The chromite correlate with host mineral compositions (Ni-rich rims in the lherzolitic shergottites have compositions pyrrhotite inside olivine) indicates that the high- falling in the range of Tissint Ti-Cr spinels (McSween temperature partitioning of Fe and Ni between sulfides and Treiman 1998). Tissint ulv ospinel€ compositions are and silicates survived the shock event. also like those for the basaltic shergottites EET 79001 and QUE 94201 (McSween and Treiman 1998), but they Oxides are less Fe-rich than those for Shergotty titanomagnetites (Stolper and McSween 1979). The main oxide phase in Tissint is chromite, which All analyzed oxide grains show bulk compositions shows zoning to intermediate Ti-Cr spinel. Ulv ospinel€ and that correspond to paramagnetic oxides at room ilmenite are also present as separate grains, with rare temperature (Table 2, Fig. 3). However, magneto- micron-sized magnetite grains close to some ulv ospinels.€ optical imaging of magnetically saturated polished Oxide compositions determined by EMP are shown in sections shows that some oxide grains are ferromagnetic Table 2 and Fig. 3. The ulv ospinels€ plot close to the at room temperature. Figure 4A shows an example ulvospinel€ apex of the spinel prism and the ulv ospinel€ end where ferromagnetic sulfides, ferromagnetic oxides, and of the ulv ospinel-magnetite€ join, and there is an almost nonferromagnetic oxides can be seen. The ferromagnetic complete series of intermediate compositions between oxides identified by magneto-optical imaging were chromite and ulv ospinel.€ These spinels are similar in further analyzed by EMP, and all have ulv ospinel€ bulk composition to those of composite spinels in other olivine- compositions. This bulk composition should be 8 J. Gattacceca et al.

Mgt Usp Ti

A B C

Chrom Usp Spin Chrom Fe 2+ Fe 3+

Fig. 3. Compositions of oxide minerals in Tissint. Triangles = chromite; squares = Ti-Cr spinel forming rims on chromite; diamonds = ulvospinel;€ stars = ilmenite; small dots are from Chennaoui Aoudjehane et al. (2012). paramagnetic at room temperature, so one needs to the Fe 2O3 in a coexisting ilmenite phase may be reduced invoke the existence of ferromagnetic Fe-rich domains to magnetite. Righter et al. (2012) proposed a reaction € in the ulv ospinel. Indeed, SEM imaging shows that in which Fe 2O3 + FeS produce magnetite in nakhlites. most ulvospinel€ grains contain fine lamellae, which are The magnetite and ilmenite compositions listed in somewhat deformed and displaced by shock effects Table 2 cannot strictly be used for thermobarometry (Figs. 4B and 4C). The widest lamellae (Fig. 4D) are because of the residual overlap in the electron probe closer to ilmenite than to ulv ospinel€ in probe analyses, data. However, if we assume that the true magnetite and similar in composition to the large grains of composition lies between the value listed and pure ilmenite. The presence of ilmenite lamellae indicates magnetite, we calculate equilibration (Lindsley and oxidation-exsolution in the parent ulv ospinel€ along Spencer 1982; Lepage 2003) at 300 –400 °C and oxygen several crystallographic planes, such that the residual fugacity between FMQ-1 and FMQ-5. This is consistent host spinel phase is magnetite-rich. These Fe-rich with oxygen fugacities in olivine-phyric shergottites and domains are too small for conventional microprobe the observation that oxidation state did not change analysis, so we pursued the problem using the from early- to late-stage crystallization (Goodrich et al. SXFiveFE probe at the Cameca factory. The field 2003). It suggests that the oxy-exsolution reaction emission electron column achieved a very high spatial occurred by cooling in a closed system without the resolution and showed that the small (mostly below introduction of particularly oxidizing fluids, and 1 lm) magnetite islands seen in Figs. 4B, 4C, and 4E is not a product of the hydrothermal alteration contain <4 wt% TiO 2. Ti-free residual spinel would described in Tissint by Chennaoui Aoudjehane et al. require approximately 45% magnetite and (2012). approximately 55% ilmenite, roughly consistent with observed abundances (Fig. 4B). INTRINSIC MAGNETIC PROPERTIES Similar oxidation-exsolution of ulv ospinel€ is seen in QUE 94201 (McSween and Treiman 1998), and suggested The magnetic properties of Tissint are summarized in other shergottites (Cisowski 1986) and common in in Table 3, and a list of acronyms is provided in Table terrestrial rocks (e.g., Buddington and Lindsley 1964). S1. It is noteworthy that hysteresis properties and While FeO in spinel is oxidized to produce magnetite, magnetic susceptibility, measured on bulk samples, are

Fig. 4. A) Reflected light image of a polished section of Tissint previously saturated in a field of 3 T. Two profiles cutting through sulfide and oxide grains are highlighted. The reflected light intensity (relative scale) and the vertical surface magnetic field along these profiles are indicated. Magnetic field profiles were obtained using magneto-optical imaging. The outlined sulfide is ferromagnetic, as well as oxide grains numbered 1 and 2. Oxide grain number 3 is not ferromagnetic. B) Scanning electron microscope (SEM) backscattered electron (BSE) image of a ferromagnetic oxide, with dark (ilmenite) lamellae and bright residual Fe-rich spinel host, and minor pyrrhotite on the edge of the oxide. C) SEM BSE image of a ferromagnetic oxide grain, showing faulting and deformed lamellae. The white box corresponds to the elemental map of Fig. E. D) SEM BSE image of a ferromagnetic oxide, with a wide ilmenite lamella in bright ulv ospinel€ host containing very fine lamellae. E) Combined elemental maps with Ti (red) and Fe (blue) K a X-rays. Ilm, ilmenite (red), Mag, magnetite (blue), Usp, ulvospinel€ (purple or violet). Opaque minerals and magnetism of Tissint 9

12 profile 1 A profile 2 8 oxide 2 oxide 3 4 field (mT) field 0

2 sulfide oxide 1 1

0 sulfide 600 0 200 distance m 400 12 profile 2 8

profile 1 (mT) field 4

oxide 1 2 oxide 2 oxide 3 1

100 m light light 0 0 100 200 300 distance m

B C 5 m oxide 1

5 m

D E Usp

Mag

Usp

Ilm

5 m Ilm 2 m 10 J. Gattacceca et al.

Table 3. Magnetic properties of Tissint meteorite. (Table 3), the median destructive field (MDF) of SIRM Æ Unit Mean Æ SD n of 55 mT, the S À300 ratio of 0.76 0.02 (computed as 2 À1 À2 the ratio of SIRM superimposed with a back-field IRM MRS Am kg 6.23 Æ 0.82 10 16 M Am 2 kg À1 1.72 Æ 0.30 10 À1 10 at 300 mT over SIRM), and the absence of saturation S below 1 T (Fig. 6), all together indicate that the BCR mT 77.5 Æ 4.0 10 magnetic signal at room temperature is dominated by BC mT 46.4 Æ 2.3 10 v m3 kg À1 1.14 Æ 0.23 10 À6 26 high-coercivity minerals. Thermal demagnetization of 3 À1 À7 vf m kg 6.86 Æ 2.05 10 26 SIRM (Fig. 7B) shows two inflexions at about 260 °C, pTRM 80 °C Am 2 kg À1.lT 2.18 10 À7 1 and in the 500 –560 °C range with a maximum À À pTRM 120 °C Am 2 kg 1.lT 6.45 Æ 2.45 10 7 3 unblocking temperature of 585 °C. This suggests that 2 À1 À5 pTRM 340 °C Am kg .lT 1.29 Æ 0.06 10 3 SIRM is carried in equivalent parts by pyrrhotite (Curie TRM 590 °C Am kg À1.lT 2.17 Æ 0.28 10 À 3 À1 À7 temperature TC of 290 and 320 °C for hexagonal and PRM 2 GPa Am kg .lT 2.01 10 1 monoclinic forms, respectively) and a high T phase. PRM 1.2 GPa Am kg À1.lT 1.28 10 À7 1 C À À Comparison of partial TRM acquired at 340 °C and PRM 0.4 GPa Am kg 1.lT 5.10 10 8 1 NRM Am 2 kg À1 2.75 Æ 2.00 10 À5 17 total TRM (Table 3) shows that TRM is also carried in approximately equivalent parts by pyrrhotite and this All abbreviations are defined in the text. high TC phase. The hysteresis parameters B CR /B C = TRM, pTRM, and PRM intensities are normalized to the ambient Æ Æ field. NRM is the average for the 17 samples that were not affected 1.67 0.06 and MRS /M S = 0.38 0.04 indicate an by magnet contamination. overall pseudosingle domain behavior. Overall, the bulk magnetic properties are in the range observed for other homogeneous down to the scale of at least 30 mg, i.e., pyrrhotite-bearing basaltic shergottites (Rochette et al. approximately 10 mm 3. Therefore, in the following, all 2005). values are average values with standard deviations. The The determination of the high TC phase that carries hysteresis loops (Fig. 5), the BCR of 77 Æ 4 mT about half of SIRM and TRM is not straightforward.

30 room temperature

20 350 °C

10 inducedmagnetizaiton (µAm2)

-1.6 -1.2 -0.8 -0.4 0.4 0.8 1.2 1.6 applied field (T)

-10

-20

-30

Fig. 5. Hysteresis loop at room temperature for a 140 mg sample, at room temperature and at 350 °C, after correction for high- ﬁeld susceptibility computed between 1.2 and 1.5 T. Opaque minerals and magnetism of Tissint 11

1 1 NRM TRM 590°C 0.8 350 °C 0.8 pTRM 340 °C room temperature SIRM 0.6 PRM 0.88 GPa 0.6 PRM 2 GPa

IRM (normalized) 0.4

0.4 0.2

0.2 0 magnetizationnormalized at 12 mT 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 A Field (T) 0 Fig. 6. Isothermal remanent magnetization versus magnetizing 0 20 40 60 80 100 120 AF (mT) ﬁeld at room temperature (blue circles) and at 350 °C (red 1 circles).

This high TC phase could be magnetite ( TC of 580 °C), 0.8 a metal phase (as found in two shock-blackened Martian meteorites; Van de Moort ele et al. 2007), or a substituted 0.6 TRM hematite that has been proposed as a candidate for Martian crustal magnetism by McEnroe et al. (2002, 0.4 SIRM 2004). In view of the maximum unblocking temperature pTRM of 585 °C, this titanohematite would be Fe Ti O with 120°C 2- x x 3 0.2 PRM € 2 GPa x ~ 0.1 (Dunlop and Ozdemir 1997). However, such a normalizedremanent magnetization NRM titanohematite should be present in large amount, at least B 2 À1 0 30 wt% based on M ~ 0.5 Am kg for titanohematite 0 100 200 300 400 500 600 S € with this composition (Dunlop and Ozdemir 1997) and a demagnetization temperature (°C) MRS /M S of 0.5) to account for the IRM unblocked Fig. 7. AF (A) and thermal (B) demagnetization of NRM and between 350 and 600 °C. As no titanohematite has been various laboratory remanences (saturation IRM [SIRM], observed by petrographic observation, this hypothesis piezoremanent magnetization [PRM], thermoremanent can be clearly ruled out. magnetization [TRM], and partial thermoremanent Hysteresis measurements at 350 °C, above the T of magnetization [pTRM] at various temperatures) of Tissint. C For AF demagnetization, the intensity is normalized at 12 mT pyrrhotite, allow us to isolate the hysteresis properties to focus on the coercivity range corresponding to the stable of the high TC phase. At this temperature, IRM component of the NRM. For thermal demagnetization, to saturation is reached around 250 mT (Fig. 6), and take into account the two components of magnetizations hysteresis (Fig. 5) indicates pseudosingle domain state (above and below approximately 150 °C), the NRM at a given for this high T phase (B /B = 1.76, M / temperature is the vectorial sum of the pTRM unblocked at C CR C RS all successive demagnetization steps. MS = 0.26). A high-coercivity FeNi phase, such as tetrataenite, is also excluded by the absence of signiﬁcant transformation of the hysteresis properties Overall, these results point toward pseudosingle domain after heating up to 650 °C in Ar or He atmosphere. Ti-poor titanomagnetite for the high TC phase, with a Indeed, tetrataenite would transform into disordered grain size in the 0.5 –1 lm range (Dunlop 2002), in taenite with much lower remanence above 500 °C agreement with the microscopic observations discussed (Wasilewski 1988). Thermomagnetic experiments reveal in the Mineralogy of Fe-Ni Sulﬁdes section. two Curie temperatures of 309 and 563 °C (Fig. 8) that In view of the higher M RS of magnetite with respect correspond to pyrrhotite and Ti-poor titanomagnetite to pyrrhotite (by a factor 6 at most considering the € with as low as 3 mole% of ulv ospinel (TM03), same MRS /M S for both minerals, even though the room respectively. The range of unblocking temperature temperature MRS /M S of magnetite may be slightly lower observed in Fig. 7B suggests either a distribution in than that of pyrrhotite), these results indicate an overall grain size, or, more likely, a range in Ti content of modal dominance of pyrrhotite over magnetite, by a titanomagnetite grains, TM03 being the most Ti-poor factor 6 (mass ratio) at most. Using this ratio, 2 À1 titanomagnetite present. A maximum of about 15 mole% MS = 0.17 Æ 0.03 Am kg would correspond to the ulvospinel€ can be estimated from the unblocking presence of approximately 0.6 wt% pyrrhotite and temperature spectrum in Fig. 8B (Lattard et al. 2006). approximately 0.1 wt% magnetite. 12 J. Gattacceca et al.

1 0.3 A

0.8 /kg) 0.6 2 0.2 FC

magnetite 0.4 pyrrhotite LT-SIRM (Am LT-SIRM 0.2 0.1 Inducedmagnetization (normalized) ZFC 0 100 200 300 400 500 600 Temperature (°C) 0 Fig. 8. Magnetization induced in a ﬁeld of 300 mT as a 0 100 200 300 function of temperature for a 74 mg sample of Tissint. The T (°K) Curie temperatures computed from this curve using the two- tangent method (Gromme et al. 1969) are 309 and 563 °C. 9 B

Low-temperature SIRM experiments conducted /kg) after zero ﬁeld cooling (ZFC) and ﬁeld cooling (FC) 2 cooling

reveal two transitions at 70 K and 120 K (Fig. 9A), in Am

-2 8 close correspondence with the Curie temperature of chromite (Klemme et al. 2000) and the Verwey transition observed for magnetite (Verwey 1939) and Ti- poor titanomagnetite composition with less than 4% Ti RT-SIRM (10 RT-SIRM (Kakol et al. 1994). The M S variation across the 7 warming transition at 70 K is 0.10 Am 2 kg À1 (determined on a 30 mg sample), corresponding to 0.63 wt% of chromite, 2 À1 using a M S of 16 Am kg for pure chromite (Robbins et al. 1971; Gattacceca et al. 2011). The monoclinic pyrrhotite transition at 34 K is not 6 observable in the low-temperature data (Fig. 9B). In 0 100 200 300 T (°K) agreement with unblocking temperatures mostly below 290 °C (Fig. 7B), this suggests that the pyrrhotite phase Fig. 9. Thermal evolution of A) Low-temperature saturation present in Tissint may be mostly the hexagonal form in IRM (SIRM), B) Room-temperature SIRM. its metastable ferromagnetic form, as typically found in other shergottites (Rochette et al. 2005). The Curie listed in Table 4. They allow quantification of the temperature of 309 °C revealed by thermomagnetic proportion of total iron present in titanomagnetite and measurements (Fig. 8) is more indicative of monoclinic pyrrhotite, respectively, 2 and 4 wt%, although with pyrrhotite, but this could well have formed from very large uncertainties. Based on a total FeO content hexagonal pyrrhotite during the heating experiment of 21 wt% (Chennaoui Aoudjehane et al. 2012), this itself. The same explanation accounts for the higher translates into 0.5 and 0.9 wt% for pure magnetite pyrrhotite unblocking temperatures for the laboratory equivalent and pyrrhotite, respectively. This is broadly TRM than for nonheating laboratory remanences like consistent with other magnetic properties and petrologic IRM and PRM (Fig. 7B). The 57 Fe M ossbauer€ spectra investigations. Note that the signal of iron within at room temperature (Fig. S2) exhibit three Fe 2+ paramagnetic oxides (chromite, ilmenite, ulv ospinel)€ is quadrupole doublets, associated with olivine and masked in the paramagnetic silicate doublets. pyroxene (Stevens et al. 1998) superposed on weaker Mass-weighted average low-field magnetic magnetic sextets. Two of the magnetic sextets susceptibility v = 1.14 9 10 À6 m3 kg À1 is in the correspond to titanomagnetite, while the four remaining range observed for other pyrrhotite-bearing basaltic magnetic sextets are consistent with pyrrhotite shergottites (mean 0.94 Æ 0.35 9 10 À6 m3 kg À1, n = 12). (Kondoro 1999). The derived hyperfine parameters are High-field magnetic susceptibility is vHF = 4.56 Æ Opaque minerals and magnetism of Tissint 13

Table 4. M ossbauer€ hyperﬁne parameters for Tissint meteorite. The isomer shifts are reported relative to a-Fe. IS (mm s À1) QS (mm s À1) Bhf (T) FWHM (mm s À1) A (%) D1 1.15 2.92 — 0.28 40.0 Olivine D2 1.16 2.24 — 0.44 29.0 Pyroxene D3 1.13 1.92 — 0.36 25.0 S1 0.79 0.0 47.0 0.44 1.0 Titanomagnetite S2 1.03 0.0 43.5 0.55 1.0 S3 0.64 0.0 28.2 0.45 1.0 Hexagonal pyrrhotite S4 0.69 0.0 25.0 0.45 1.0 S5 0.60 0.0 23.5 0.45 1.0 S6 0.61 0.0 21.5 0.45 1.0 IS = isomet shift (Æ0.005); QS = quadrupole splitting ( Æ0.005); FWHM = full width at half maximum ( Æ0.02); A = relative area (Æ1%).

0.63 9 10 À7 m3 kg À1, in close agreement with the Four samples were studied through thermal paramagnetic susceptibility of 4.40 9 10 À7 m3 kg À1 that demagnetization up to 340 °C under argon atmosphere. can be computed (see Gattacceca et al. 2008a) from the After the removal of a stable component of bulk composition of the meteorite (Chennaoui magnetization between 25 and approximately 120 °C, Aoudjehane et al. 2012). The anisotropy of magnetic the demagnetization was rather noisy, but a stable susceptibility was measured on two large samples with component was isolated in two samples between 150 masses 7.7 g and 28.5 g. The anisotropy degree is and 300 °C (Fig. 10). In all four samples, the NRM is P = 1.028 (mass-weighted average) and Pf = 1.049 when unblocked mostly above 200 °C, and is essentially corrected for the supposedly isotropic paramagnetic demagnetized at 310 °C. Samples demagnetized by AF contribution (see Gattacceca et al. 2008a). The fabric is show a stable component of magnetization above strongly oblate (mass-weighted average T = 0.73). Two approximately 14 mT and up to at least 100 mT measurements performed on smaller mutually oriented (Fig. 10). samples (with masses 1.32 and 0.35 g, and P = 1.063 and We interpret the stable magnetization isolated 1.072, respectively) give indistinguishable directions below 120 °C as a pTRM acquired in the Earth field for the principal susceptibility axes. This weak, but during heating and cooling of the small meteorite well-defined, oblate, and directionally homogeneous fragments on the desert surface. Indeed, ground fabric may be interpreted as the primary magmatic fabric temperature in July at the fall location can reach about (Gattacceca et al. 2008a) or a shock-induced fabric 80 °C, and can be enhanced inside small crusted (Gattacceca et al. 2007; Nishioka et al. 2007). The meteorites because of the black fusion crust. anisotropy of SIRM, measured on two samples, is also Normalization of the demagnetization of the NRM and weak and oblate with mass-weighted average P SIRM = a pTRM imparted in the laboratory at 120 °C (in a 1.10 and TSIRM = 0.72. field of 49 lT) show that the NRM unblocked below 120 °C can be accounted for by a pTRM acquired in an ambient field of 34 lT, in broad agreement with the PALEOMAGNETISM field intensity at the fall location (41 lT). This low- temperature component of the NRM extends up to Results 120 °C, which is higher than the expected maximum temperature at the surface of the desert. This can be The NRM of 33 samples without fusion crust was attributed to the nonideal behavior of the pseudosingle measured. The intensity of NRM is bimodal, with 17 domain magnetic carriers or more likely thermo-viscous samples with intensity below 10 À4 Am 2 kg À1 (median magnetization acquired during repeated heating and 2.2 9 10 À5 Am 2 kg À1) and 15 samples with intensity cooling cycles at about 80 °C in the Earth’s field. above 5 9 10 À4 Am 2 kg À1 (median 6.9 9 10 À3 Am 2 Indeed, unblocking temperatures of 120 °C during the 1 kg À1). The corresponding median NRM/SIRM ratios hour laboratory heating correspond to a cumulative are 3.5 9 10 À4 and 1.1 9 10 À1. The higher NRM 1 month stay at approximately 65 °C for magnetite population has undoubtedly been contaminated by hand (Pullaiah et al. 1975) and approximately 100 °C for magnet, probably during meteorite hunting, or pyrrhotite (Dunlop et al. 2000), in the range of the subsequent handling; unfortunately, a common practice expected thermal history of Tissint fragments at the among meteorite hunters. However, we were left with desert surface during summer 2011. 17 samples (without fusion crust) that were suitable for The best demagnetization results were obtained on paleomagnetic investigation of extraterrestrial signal. two samples that were first demagnetized at 120 °C, to 14 J. Gattacceca et al.

-8 2 AF, 1.32 g, NRM=2.91 10 Am AF, 36 mg, NRM=1.24 10 -8 Am 2 20 mT 35 48 10 -8 5 mT 42 70

80 10 85 14 -8 2 10 -8 Am 2 10 Am 100 25 40 Thermal, 124 mg, NRM=2.07 10-9 Am 2 Thermal + AF, 160 mg, NRM=1.02 10 -9 Am 2 100 10 -9 225 NRM 4 x 10-10 Am 2 290 44 270 310 2 x 10 -10 80 °C 40 6 mT 20 300 80 °C 10 -9 Am 2 28 NRM 100 120

Fig. 10. Demagnetization data for four samples of Tissint. Open and solid symbols are projections on two perpendicular planes whose intersection is the horizontal axis. The axes represent magnetic moments in Am 2. The demagnetization technique, sample mass, and initial NRM moment are indicated. remove the terrestrial pTRM, and then demagnetized by magnetization originally blocked above 310 °C (i.e., AF (Fig. 10). carried by magnetite) has been erased naturally by There is no relation between NRM intensity and viscous decay. Another possibility would be that the sample mass, which suggests that NRM is homogeneous magnetite formed after cooling below approximately in direction at least down to a scale of 30 mg 310 °C. It is indeed possible that the oxidation (corresponding to 10 mm 3). This homogeneity is happened in stages: the original ulv ospinel€ first formed confirmed by the identical NRM direction observed in ilmenite lamellae and titanium-rich titanomagnetite that two mutually oriented samples. This high-coercivity and was paramagnetic during cooling, and then the high-temperature magnetization cannot be a viscous titanium-poor magnetite would have formed during remanent magnetization (because of the very short further oxidation below approximately 310 °C. residence in the terrestrial field and the removal of a However, if formed at low temperature, magnetite low-temperature component of magnetization) or an should still have recorded a chemical remanent isothermal remanent magnetization acquired by magnetization that is not observed in the NRM thermal contamination with a magnet (because of the low demagnetization (e.g., Haigh 1958). NRM/SIRM value), and is therefore very probably of In view of the shock history of Tissint, shock extraterrestrial origin. remanence needs to be considered as a potential candidate. However, both the coercivity spectrum and Origin of the Natural Remanent Magnetization unblocking temperature spectrum of the laboratory PRM do not fit with those of the NRM (Fig. 7). In The comparison of the coercivity spectrum of the particular, laboratory PRM at 2 GPa unblocks up to NRM with laboratory-induced remanences shows that 550 °C. Even if the maximum pressure used in our the NRM is harder than all other types of remanence, experiments (2 GPa) is much lower than the peak except total TRM acquired at 590 °C (Fig. 7A). pressure suffered by Tissint, there is no reason why a However, it is noteworthy that the NRM is completely natural PRM at higher pressure would not have unblocked at 310 °C, whereas both SIRM and unblocking temperatures above 310 °C. Therefore, the laboratory TRM acquired at 590 °C have unblocking most plausible explanation is that the NRM is a pTRM temperature extending up to 585 °C, with about half of acquired by cooling from about 310 °C in a steady the remanence being unblocked above 310 °C (Fig. 7B). magnetic field. The laboratory pTRM blocked between 590 °C and However, we are left with the issue that the 350 °C is very hard with respect to AF (harder than coercivity spectrum of the stable NRM is significantly total TRM), and there is no possibility that a harder than that of a pTRM acquired at 310 °C Opaque minerals and magnetism of Tissint 15

(Fig. 7B). We attribute the fact that the NRM is harder magnetization acquired at the time of crystallization. than a pTRM at 310 °C to a combination of partial This rock was completely remagnetized following the natural relaxation of the NRM with time and shocks main impact event, by going through pressure-induced that affected this rock after it was magnetized. Indeed, magnetic phase transition of the ferromagnetic minerals. shock remagnetizes preferentially the low-coercivity Indeed, pyrrhotite, at least in its monoclinic form, has a fraction of the ferromagnetic minerals (e.g., Gattacceca phase transition at 2.8 –4.5 GPa (e.g., Rochette et al. et al. 2010b). This is confirmed by our PRM acquisition 2003; Gilder et al. 2011), and magnetite has a transition experiments up to 2 GPa: the coercivity spectrum of at 12 –16 GPa (Ding et al. 2008), much below the peak PRM is significantly shifted toward low AF with respect shock pressure suffered by Tissint. to IRM and NRM (Fig. 7A). Multiple shocks, Our paleomagnetic data show that Tissint was including at least the shock that led to the ejection of magnetized while cooling from 310 °C down to at least the meteorite from Mars about 0.7 Myr ago, would 150 °C in presence of a magnetic field of about 2 lT. result in increasing hardening of the NRM. Indeed, This excludes that the meteorite has been ejected at repeated pressure loadings have been shown to result in temperature above 150 °C because it would have cooled increasing demagnetization (Gilder et al. 2006; Bezaeva in space in a null magnetic field. A similar line of et al. 2007). reasoning was used to show that ALH 84001 had never Scaling the NRM demagnetized between 150 and been heated above 40 °C during its ejection and transfer 310 °C (mean 1.34 9 10 À5 Am 2 kg À1 for the four to Earth (Weiss et al. 2000). samples thermally) with a pTRM acquired at 310 °C The best explanation to account for thermal history and demagnetized at 150 °C, provides a rough estimate of Tissint as revealed by the paleomagnetic data is of 2.3 lT for the paleointensity of the magnetizing field. postshock cooling at the Martian surface. Estimates of For samples demagnetized by AF, NRM/SIRM postshock temperature increase associated with impacts derivative ratios (REM0, see Gattacceca and Rochette for shergottites (e.g., Fritz et al. 2005) combined with a 2004) were also used to estimate the paleofield necessary reasonable initial temperature of 210 K for the upper to account for the high-coercivity magnetization. This surface of Martian subsurface (Mellon et al. 2004) show method, calibrated for both magnetite and pyrrhotite, that a 40 GPa shock will result in a postimpact provides a lower limit for the paleofield because the equilibrium temperature of 310 °C. This peak shock magnetization is a pTRM and not a TRM. The REM 0 value is in good agreement with that evidenced by integrated over the high-coercivity component is petrologic observations (Baziotis et al. 2013; El Goresy 4.99 9 10 À4 (SD = 4.11 9 10 À4, n = 11). This indicates et al. 2013). a lower limit of about 1.5 lT for the paleointensity, in It may be noted that, in this scenario, a shock broad agreement with the paleointensity obtained by remanent magnetization (SRM) acquired during heating experiments. pressure release below about 12 –16 GPa, the transition pressure of magnetite (Ding et al. 2008) should still be DISCUSSION preserved in the rock. Although it is difficult to estimate the unblocking temperatures and the expected intensity Tissint is an olivine-phyric shergottite that has been of such a SRM, it is plausible that it is not visible weathered by fluids at the surface of Mars (Chennaoui during the thermal demagnetization above 310 °C Aoudjehane et al. 2012). Preliminary chronological because the efficiency of SRM acquisition for Tissint analyses suggest a 147Sm-144Nd crystallization age of magnetite that has rather high coercivity (as about 600 Ma (Brennecka et al. 2013). Tissint later demonstrated by the hysteresis cycle at 350 °C, Fig. 5) suffered multiple shock events to levels sufficient to is probably much lower than that of TRM (e.g., form melt veins and pockets, and ubiquitous high- Gattacceca et al. 2008b). pressure minerals (Baziotis et al. 2013), including shock- Our paleomagnetic data also show that this main induced diamonds (El Goresy et al. 2013). Tissint was impact event was distinct from the ejection event whose ejected from Mars approximately 0.7 Myr ago associated maximum postimpact temperature is (Chennaoui Aoudjehane et al. 2012). constrained to below 150 °C, corresponding to a The peak shock pressure in Tissint is estimated to maximum peak pressure of about 35 GPa. In view of approximately 25 GPa with localized excursion up to the young crystallization age of Tissint (approximately 40 GPa (Baziotis et al. 2013), within the peak pressure 600 Ma) compared with the age of the shutdown of the range estimated for other shergottites, is typically Martian dynamo around 4 Ga (e.g., Lillis et al. 2008), 25 –45 GPa (e.g., Nyquist et al. 2001; Fritz et al. 2005). the only likely source for the 2 lT stable surface field In view of this shock history, it is clear that whatever its that magnetized Tissint is the remanent magnetization age, Tissint cannot have retained a putative primary of the Martian crust. Indeed, such crustal fields of 16 J. Gattacceca et al. several lT are commonly expected at the present-day Acknowledgments—The research leading to these results Martian surface (e.g., Langlais et al. 2004), and their has received funding from Agence Nationale de la record has already been evidenced in a number of Recherche (project ANR-09-BLAN-0042), CNRS (UMR nakhlites and shergottites (Cisowski 1986; Gattacceca and 7202), and People Programme (Marie Curie Actions) of Rochette 2004). It is noteworthy that Tissint, if magnetized the European Union’s Seventh Framework Programme in a Martian dynamo magnetic field of 50 lT, would carry (FP7/2077-2013) under REA grant agreement no. a remanent magnetization of only approximately 298355. We thank Francßois Demory (CEREGE) and 3 A m À1. This is way below the 10 –20 A m À1 that is Mike Jackson (IRM) for their precious help in the necessary over a crustal thickness of about 40 km to laboratory, and Luc Labenne from Labenne M et eorites account for the observed magnetic anomalies at the for the loan of part of the samples. Constructive reviews Martian surface, which confirms that shergottite-like rocks by L. Taylor, T. Kohout, and an anonymous reviewer are cannot be the source of these anomalies. greatly acknowledged. S. Gilder also provided useful insights on a previous version of this work. CONCLUSION Editorial Handling—Dr. Gretchen Benedix The scope of this work was to provide a detailed understanding of Tissint opaque minerals, with an REFERENCES emphasis on ferromagnetic phases and their magnetic record. The magnetic mineralogy of Tissint consists of Acuna~ M. H., Connerney J. E. P., Ness N. F., Lin R. P., about 0.6 wt% of pyrrhotite, and 0.1 wt% of low-Ti Mitchell D., Carlson C. W., McFadden J., Anderson K. € A., Reme H., Mazelle C., Vignes D., Wasilewski P., and titanomagnetite (with a low ulv ospinel content in the Cloutier P. 1999. Global distribution of crustal range 3–15 mole%) for those minerals that are magnetization discovered by the Mars Global Surveyor ferromagnetic at temperatures encountered at the MAG/ER experiment. Science 284:790–793. Martian surface. Chromite (with a Curie temperature of Baziotis I. P., Liu Y., DeCarli P. S., Melosh H. J., McSween H. 70 K) is present with an abundance of approximately Y., Bodnar R. J., and Taylor L. A. 2013. The Tissint Martian meteorite as evidence for the largest impact excavation. 0.6 wt%. Magnetite was formed by oxy-exsolution of Nature Communications 4:1404, doi:10.1038/ncomms2414. ulvospinel€ grains during deuteric alteration. Pyrrhotites Bezaeva N. S., Rochette P., Gattacceca J., Sadykov R. A., are the most Fe-poor and Ni-rich ever analyzed in and Trukhin V. I. 2007. Pressure demagnetization of the shergottites and nakhlites thus far. This high Ni content Martian crust: Ground truth from SNC meteorites. reflects the large amount of Ni available inside the Geophysical Research Letters 34:L23202, doi:10.1029/ 2007GL031501. abundant olivine crystals that host pyrrhotite. Iron Brand R. A. 1995. Normos Mossbauer€ fitting program, user’s deficiency may be explained by an input of regolith- guide. Starnberg, Germany: Wissenschaftlich Elektronik derived sulfur in agreement with other evidence of GmbH. Martian weathering detected in this meteorite Brennecka G. A., Borg L. E., Symes S. J. K., and Wadhwa (Chennaoui Aoudjehane et al. 2012). These pyrrhotites M. 2013. The age of Tissint: Sm-Nd & Rb-Sr isotope systematics of a Martian meteorite (abstract #1786). 44th probably have preserved metastable Mss structure Lunar and Planetary Science Conference. CD-ROM. blocked at temperature above 300 °C. Buddington A. F. and Lindsley D. H. 1964. Iron-titanium Tissint was magnetized while cooling from oxide minerals and synthetic equivalents. Journal of approximately 310 °C down to at least 150 °C in the Petrology 5:310–357. presence of a stable magnetic field of about 2 lT that Chennaoui Aoudjehane H., Avice G., Barrat J.-A., Boudouma O., Chen G., Duke M. J. M., Franchi I. A., Gattacceca J., was the surface field generated by the magnetized Grady M. M., Greenwood R. C., Herd C. D. K., Hewins Martian crust. This 310 °C temperature is probably the R., Jambon A., Marty B., Rochette P., Smith C. L., equilibrium temperature that followed the major impact Sautter V., Verchovsky P., Weber P., and Zanda B. 2012. suffered by Tissint, and is compatible with the peak Tissint Martian meteorite: A fresh look at the interior, – shock pressure of 40 GPa inferred from the assemblage surface and atmosphere of Mars. Science 338:785 788. Chevrier V., Lorand J.-P., and Sautter V. 2011. Sulfide petrology of high-pressure minerals. The 150 °C temperature sets of four nakhlites (NWA817, NWA998, Nakhla, Governador an upper limit of Tissint temperature during its ejection Valadares). Meteoritics & Planetary Science 46:769–784. from Mars and transfer to Earth. This emphasizes the Cisowski S. M. 1986. Magnetic studies on Shergotty and other interest of paleomagnetism for assessing the ejection SNC meteorites. Geochimica et Cosmochimica Acta – temperature of Martian meteorites. The relatively weak 50:1043 1048. Ding Y., Haskel D., Ovchinnikov S. G., Tseng Y.-C., Orlov magnetization in Tissint also confirms that shergottite- Y. S., Lang J. C., and Mao H. 2008. Novel pressure- like rocks cannot be the source of the observed Martian induced magnetic transition in magnetite (Fe 3O4). Physical magnetic anomalies. Review Letters 100:045508. Opaque minerals and magnetism of Tissint 17

Dunlop D. J. 2002. Theory and application of the day plot diamonds. Journal of Geophysical Research 116:B10101, (Mrs/Ms versus Hcr/Hc) 1. Theoretical curves and tests doi:10.1029/2011JB008292. using titanomagnetite data. Journal of Geophysical Goodrich C. A. 2002. Olivine-phyric Martian basalts: A new Research 107:B3, doi:10.1029/2001JB000486. type of shergottite. Meteoritics & Planetary Science 37:B31– Dunlop D. and Ozdemir€ O. 1997. Rock magnetism: Fundamentals 34. and frontiers . Cambridge, UK: Cambridge University Press. Goodrich C. A., Herd C. D. K., and Taylor L. A. 2003. 573 p. Spinels and oxygen fugacity in olivine-phyric and Dunlop D. J., Ozdemir€ O.,€ Clark D. A., and Schmidt P. W. lherzolitic shergottites. Meteoritics & Planetary Science 2000. Time-temperature relations for the remagnetization 38:1773–1792. of pyrrhotite (Fe7S8) and their use in estimating Gromme C. S., Wright T. L., and Peck D. L. 1969. Magnetic paleotemperatures. Earth and Planetary Science Letters properties and oxidation of iron-titanium oxide minerals in 176:107–116. Alae and Makaopuhi lava lakes, Hawaii. Journal of El Goresy A., Gillet P., Miyahara M., Ohtani E., Ozawa S., Geophysical Research 74:277–5294. Lin Y., Feng L., and Escerig S. 2013. Multiple shock Gross J., Treiman A. H., Filiberto J., and Herd C. D. K. events and diamond formation on Mars (abstract #1037). 2011. Primitive olivine-phyric shergottite NWA 5789: 44th Lunar and Planetary Science Conference. CD-ROM. Petrography, mineral chemistry, and cooling history imply Etaschmann B., Pring A., Putnis A., Grugic B. A., and Studer a magma similar to Yamato-980459. Meteoritics & A. 2004. A kinetic study of the exsolution of pentlandite Planetary Science 46:116–133. (Ni, Fe)9S8 from the monosulfide solid solution (Fe, Ni)S. Haigh G. 1958. The process of magnetization by chemical American Mineralogist 89:39–50. change. Philosophical Magazine 3:267–286. Fleet M. E. 2006. Phase equilibria at high temperatures. In Herd C. D. K., Borg L. E., Jones J. H., and Papike J. J. 2002. Sulfide mineralogy and geochemistry, edited by Vaughan D. Oxygen fugacity and geochemical variations in the J. Reviews in Mineralogy and Geochemistry, vol. 61. Martian basalts: Implications for Martian basalt pp. 365–419. petrogenesis and the oxidation state of the upper mantle Fritz J., Artemieva N., and Greshake A. 2005. Ejection of Martian of Mars. Geochimica et Cosmochimica Acta 66:2025–2036. meteorites. Meteoritics & Planetary Science 40:1393 –1411. Jelinek V. 1981. Characterization of the magnetic fabric of Gattacceca J. and Rochette P. 2004. Toward a robust normalized rocks. Tectonophysics 79:63–67. magnetic paleointensity method applied to meteorites. Earth Kakol Z., Sabol J., Stickler J., Kozlowski A., and Honig J. M. and Planetary Science Letters 227:377–393. 1994. Influence of titanium doping on the Gattacceca J., Lamali A., Rochette P., Boustie M., and Berthe magnetocrystalline anisotropy of magnetite. Physical L. 2007. The effects of explosive-driven shocks on the Review B 49:12767–12772. natural remanent magnetization and the magnetic Klemme S., O’Neill H. St. C., Schnelle W., and Gmelin E. properties of rocks. Physics of the Earth and Planetary 2000. The heat capacity of MgCr 2O4, FeCr 2O4, and Cr 2O3 Interiors 162:85–98. at low temperatures and derived thermodynamic Gattacceca J., Rochette P., Gounelle M., and van Ginneken properties. American Mineralogist 85:1686–1693. M. 2008a. Magnetic anisotropy of HED and Martian Kondoro J. W. A. 1999. M ossbauer€ study of vacancies in meteorites and implications for the crust of Vesta and natural pyrrhotite. Journal of Alloys Compounds 289:36–41. Mars. Earth and Planetary Science Letters 270:280–289. Langlais B., Purucker M. E., and Mandea M. 2004. The Gattacceca J., Berthe L., Boustie M., Vadeboin F., Rochette crustal magnetic field of Mars. Journal of Geophysical P., and De Resseguier T. 2008b. On the efficiency of shock Research 109, doi:10.1029/2003JE002048. magnetization processes. Physics of the Earth and Langlais B., Lesur V., Purucker M. E., Connerney J. E. P., Planetary Interiors 166:1–10. and Mandea M. 2010. Crustal magnetic fields of terrestrial Gattacceca J., Boustie M., Hood L., Cuq-Lelandais J.-P., planets. Space Science Reviews 152:223–249. Fuller M., Bezaeva N., de Resseguier T., and Berthe L. Lattard D., Engelmann R., Kontny A., and Sauerzapf U. 2010a. Can the lunar crust be magnetized by shock: 2006. Curie temperatures of synthetic titanomagnetites in Experimental groundtruth. Earth and Planetary Science the Fe-Ti-O system: Effects of composition, crystal Letters 299:42–53. chemistry, and thermomagnetic methods. Journal of Gattacceca J., Boustie M., Lima E. A., Weiss B. P., de Resseguier Geophysical Research 111:B12S28, doi:10.1029/2006JB T., and Cuq-Lelandais J.-P. 2010b. Unraveling the 004591. simultaneous shock magnetization and demagnetization of Lepage L. D. 2003. ILMAT: An Excel worksheet for ilmenite – rocks. Physics of the Earth and Planetary Interiors 182:42–49. magnetite geothermometry and geobarometry. Computers Gattacceca J., Rochette P., Lagroix F., Mathe P.-E., and & Geosciences 29:673–678. Zanda B. 2011. Low temperature magnetic transition of Lillis R. J., Frey H. V., Manga M., Mitchell D. L., Lin R. P., chromite in ordinary chondrites. Geophysical Research Acuna~ M. H., and Bougher S. W. 2008. An improved Letters 38:L10203, doi:10.1029/2011GL047173. crustal magnetic field map of Mars from electron Gilder S. A., Le Goff M., and Chevrin J.-C. 2006. Static stress reflectometry: Highland volcano magmatic history and the demagnetization of single and multidomain magnetite with end of the Martian dynamo. Icarus 194:575–596. implications to meteorite impacts. High Pressure Research Lindsley D. H. and Spencer K. J. 1982. Fe –Ti oxide 26:539–547. geothermometry: Reducing analyses of coexisting Ti- Gilder S. A., Egli R., Hochleitner R., Roud S. C., Volk M. magnetite (Mt) and ilmenite (Ilm). Eos Transactions 63:471. W. R., Le Goff M., and de Wit M. 2011. Anatomy of a Lorand J.-P., Chevrier V., and Sautter V. 2005. Sulfide pressure-induced, ferromagnetic-to-paramagnetic transition mineralogy and redox conditions in some shergottites. in pyrrhotite: Implications for the formation pressure of Meteoritics & Planetary Science 40:1257–1272. 18 J. Gattacceca et al.

Lorand J.-P., Barat J.-A., Chevrier V., Sautter V., and Pont S. pyrrhotite and impact demagnetization on Mars. 2012. Metal-saturated sulfide assemblages in chassignite Geophysical Research Letters 30:1683, doi:10.1029/ NWA 2737; Evidence for impact-related sulfur 2003GL017359. devolatilization. Meteoritics & Planetary Science 47:1830– Rochette P., Gattacceca J., Chevrier V., and Lorand J. P. 1841, doi:10.1111/maps.12015. 2005. Matching Martian crustal magnetization and McEnroe S. A., Harrison R. J., Robinson P., and Langhorst meteorite magnetic properties. Meteoritics & Planetary F. 2002. Nanoscale hematite –ilmenite lamellae in massive Science 40:529–540. ilmenite rock: An example of lamellar magnetism with Stephenson A. 1993. Three-axis static alternating field implications for planetary magnetic anomalies. Geophysical demagnetization of rocks and the identification of natural Journal International 151:890–912. remanent magnetization, gyroremanent magnetization, and McEnroe S. A., Brown L. L., and Robinson P. 2004. Earth anisotropy. Journal of Geophysical Research 98:373–381. analog for Martian anomalies: Remanence properties Stevens J. G., Khasanov A. M., Miller J. W., Pollak H., and of helo-ilmenite norites in the Bjerkreim-Sokndal Li Z. 1998. mossbauer€ mineral handbook. Asheville, NC: intrusion, Rogaland, Norway. Journal of Applied Physics Mossbauer€ Effect Data Center, The University of North 56:195–212. Carolina. McSween H. Y. Jr. and Treiman A. H. 1998. Martian Stolper E. and McSween H. Y. Jr. 1979. Petrology and origin meteorites. In Planetary materials, edited by Papike J. J. of shergottite meteorites. Geochimica et Cosmochimica Reviews in Mineralogy, vol. 36. Washington, D.C.: The Acta 43:1475–1498. Mineralogical Society of America. 53 pp. Sutton S. R., Rao M. N., and Nyquist L. E. 2008. Sulfur and iron Mellon M. T., Feldman W. C., and Prettyman T. H. 2004. speciation in gas-rich impact-melt glasses from basaltic The presence and stability of ground ice in the southern shergottites determined by microxanes (abstract #1961). 39th hemisphere of Mars. Icarus 169:324–340. Lunar and Planetary Science Conference. CD-ROM. Nishioka I., Funaki M., and Toshimori S. 2007. Shock- Uehara M., van der Beek C. J., Gattacceca J., Skidanov V., induced anisotropy of magnetic susceptibility: Impact and Quesnel Y. 2010. Advances in magneto-optical experiment on basaltic andesite. Earth, Planets and Space imaging applied to rock magnetism paleomagnetism. 59:e45–e48. Geochemistry Geophysics Geosystems 11:Q05Y09, doi:10. Nyquist L. E., Bogard D. D., Shih C.Y., Greshake D., St offler€ 1029/2009GC002653. D., and Eugster O. 2001. Ages and geologic histories of Van de Moort ele B., Reynard B., Rochette P., Jackson M., Martian meteorites. Space Science Reviews 96:105–164. Beck P., Gillet P., and McMillan P. F. 2007. Shock- Pullaiah G., Irving E., Buchan K. L., and Dunlop D. J. 1975. induced metallic iron nanoparticles in olivine-rich Martian Magnetization changes caused by burial and uplift. Earth meteorites. Earth and Planetary Science Letters 262:37–49. and Planetary Science Letters 28:133–143. Verwey E. J. 1939. Electronic conduction of magnetite (Fe 3O4) Quesnel Y., Sotin C., Langlais B., Costin S., Mandea M., and its transition at low temperature. Nature 144:327–328. Gottschalk M., and Dyment J. 2009. Serpentinization of Walton E. L. and Spray J. G. 2003. Mineralogy, microtexture the Martian crust during Noachian. Earth and Planetary and composition of shock-induced melt pockets in the Los Science Letters 277:184–193. Angeles basaltic shergottite. Meteoritics & Planetary Righter K., Keller L. P., Rahman Z., and Christoffersen R. Science 38:1865–1875. 2012. Exsolution of iron-titanium oxides in magnetite in Wasilewski P. 1988. Magnetic characterisation of the new Miller Range (MIL 03346) nakhlite: Evidence for post magnetic mineral tetrataenite and its contrast with crystallization reduction in the nakhlite cumulate pile isochemical taenite. Physics of the Earth and Planetary (abstract #2417). 43rd Lunar and Planetary Science Interiors 52:150–158. Conference. CD-ROM. Weiss B. P., Kirschvink J. L., Baudenbacher F. J., Vali H., Robbins M., Wertheim G. K., Sherwood R. C., and Buchanan Peters N. T., Macdonald F. A., and Wikswo J. P. 2000. A D. N. E. 1971. Magnetic properties and site distributions in low temperature transfer ALH 84001 from Mars to Earth. 2+ 3+ the system FeCr2O4-Fe3O4(Fe Cr 2-xFe x O4). Journal of Science 290:791–795. Physics and Chemistry of Solids 32:717–729. Yu Y. and Gee J. S. 2005. Spinel in Martian meteorite SaU Rochette P., Lorand J. P., Fillion G., and Sautter V. 2001. 008: Implications for Martian magnetism. Earth and Pyrrhotite and the remanent magnetization of SNC Planetary Science Letters 222:287–294. meteorites: A changing perspective on Martian magnetism. Yu Y. J., Dunlop D. J., Ozdemir€ O., and Ueno H. 2001. Earth and Planetary Science Letters 190:1–12. Magnetic properties of Kurokami pumices from Mt. Rochette P., Fillion G., Ballou R., Brunet F., Oulladiaf B., Sakurajima, Japan. Earth and Planetary Science Letters and Hood L. 2003. High pressure magnetic transition in 192:439–446.

SUPPORTING INFORMATION Fig S2 : M ossbauer€ spectrum of Tissint meteorite at room temperature. D1: olivine, D2 and D3: pyroxene. Additional supporting information may be found in The insert shows the magnetic subcomponents, S1 and the online version of this article: S2: titanomagnetite, S3-S6: hexagonal pyrrhotite. Fig S1 : Plot of Tissint pyrrhotite compositions in a Table S1 : Rock magnetism-related acronyms and simpliﬁed Fe-Ni-S system (phase diagram after Fleet units used in this study. [2006] and references therein). Table S2 : Opaque minerals discussed in this study. Available online at www.sciencedirect.com ScienceDirect

Geochimica et Cosmochimica Acta 124 (2014) 190–222 www.elsevier.com/locate/gca

The Paris meteorite, the least altered CM chondrite so far

Roger H. Hewins a,b,⇑, Miche`le Bourot-Denise a, Brigitte Zanda a, Hugues Leroux c, Jean-Alix Barrat d, Munir Humayun e, Christa Go¨pel f, Richard C. Greenwood g, Ian A. Franchi g, Sylvain Pont a, Jean-Pierre Lorand h, Ce´cile Courne`de i Je´roˆme Gattacceca i,j, Pierre Rochette i, Maı¨a Kuga k, Yves Marrocchi k Bernard Marty k

a Laboratoire de Mine´ralogie et Cosmochimie du Museúm, MNHN and CNRS UMR 7202, 75005 Paris, France b Department of Earth and Planetary Sciences, Rutgers University, Piscataway, NJ 08854, USA c Unite´ Mate´riaux et Transformations, Universite´ Lille 1 and CNRS, UMR 8207, F-59655 Villeneuve d’Ascq, France d Universite´ Europeénne de Bretagne and CNRS UMR 6538, U.B.O-I.U.E.M., 29280 Plouzane´ Cedex, France e Department of Earth, Ocean and Atmospheric Science and National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL 32310, USA f Institut de Physique du Globe de Paris, Sorbonne Paris Cite´, Universite´ Paris Diderot, UMR 7154 CNRS, F-75005 Paris, France g PSS, Open University, Walton Hall, Milton Keynes MK7 6AA, UK h Laboratoire de Plane´tologie et Geódynamique LPG Nantes – UMR CNRS 6112, 44322 Nantes Cedex 3, France i CNRS/Aix-Marseille Universite´, CEREGE UM34, 13545 Aix-en-Provence, France j Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA k Universite´ Lorraine and CNRS, CRPG, UPR 2300, Vandoeuvre les Nancy F-54501, France

Received 13 February 2013; accepted in revised form 14 September 2013; Available online 8 October 2013

Abstract

The Paris chondrite provides an excellent opportunity to study CM chondrules and refractory inclusions in a more pristine state than currently possible from other CMs, and to investigate the earliest stages of aqueous alteration captured within a single CM bulk composition. It was found in the effects of a former colonial mining engineer and may have been an observed fall. The texture, mineralogy, petrography, magnetic properties and chemical and isotopic compositions are consistent with classification as a CM2 chondrite. There are 45 vol.% high-temperature components mainly Type I chondrules (with olivine mostly Fa 0–2, mean Fa 0.9) with granular textures because of low mesostasis abundances. Type II chondrules contain olivine Fa 7 to Fa 76 . These are dominantly of Type IIA, but there are IIAB and IIB chondrules, II(A)B chondrules with minor highly ferroan olivine, and IIA(C) with augite as the only pyroxene. The refractory inclusions in Paris are amoeboid olivine aggregates (AOAs) and fine-grained spinel-rich Ca–Al-rich inclusions (CAIs). The CAI phases formed in the sequence hibonite, perovskite, grossite, spinel, gehlenite, anorthite, diopside/fassaite and forsterite. The most refractory phases are embedded in spinel, which also occurs as massive nodules. Refractory metal nuggets are found in many CAI and refractory platinum group element abundances (PGE) decrease following the observed condensation sequences of their host phases. Mn–Cr isotope measurements of mineral separates from Paris define a regression line with a slope of 53 Mn/55 Mn = (5.76 ± 0.76) Â 10 6. If we interpret Cr isotopic systematics as dating Paris components, particularly the chondrules, the age is 4566.44 ± 0.66 Myr, which is close to the age of CAI and puts new constraints on the early evolution of the solar system. Eleven individual Paris samples define an O isotope mixing line that passes through CM2 and CO3 falls and indicates that Paris is a very fresh sample, with variation explained by local differences in the extent of alteration. The anhydrous precursor to the CM2s was CO3-like, but the two groups differed in that the CMs accreted a higher proportion of water. Paris has little matrix ( 47%, plus 8% fine

⇑ Corresponding author at: Laboratoire de Mine´ralogie et Cosmochimie du Muse´um, MNHN and CNRS UMR 7202, 61 rue Buﬀon, 75005 Paris, France. Tel.: +33 1 4079 3769; fax: +31 4079 5772. E-mail address: [email protected] (R.H. Hewins).

0016-7037/$ - see front matter Ó 2013 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.gca.2013.09.014 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 191 grained rims) and is less altered than other CM chondrites. Chondrule silicates (except mesostasis), CAI phases, submicron forsterite and amorphous silicate in the matrix are all well preserved in the freshest domains, and there is abundant metal preserved (metal alteration stage 1 of Palmer and Lauretta (2011)). Metal and sulfide compositions and textures correspond to the least heated or equilibrated CM chondrites, Category A of Kimura et al. (2011). The composition of tochilinite–cronstedtite intergrowths gives a PCP index of 2.9. Cronstedtite is more abundant in the more altered zones whereas in normal highly altered CM chondrites, with petrologic subtype 2.6–2.0 based on the S/SiO 2 and PFeO/SiO2 ratios in PCP or tochilinite–cronstedtite intergrowths (Rubin et al., 2007), cronstedtite is destroyed by alteration. The matrix in fresh zones has CI chondritic volatile element abundances, but interactions between matrix and chondrules occurred during alteration, modify- ing the volatile element abundances in the altered zones. Paris has higher trapped Ne contents, more primitive organic compounds, and more primitive organic material than other CMs. There are gradational contacts between domains of different degree of alteration, on the scale of 1 cm, but also highly altered clasts, suggesting mainly a water-limited style of alteration, with no significant metamorphic reheating. Ó 2013 Published by Elsevier Ltd.

1. INTRODUCTION who collected artifacts. The stone was later identified as a meteorite by G. Cornen of the Universite´ de Nantes, ac- The recently found Paris meteorite is a 1.3 kg fresh, fu- quired by the MNHN Paris, classified and officially named sion-crusted stone (Fig. 1a). It was purchased by Jacques “Paris”. It is now numbered NMHN 4029. The black shiny Corre´ in 2001 at an auction in the Hoˆtel-Drouot in Paris, fusion crust, which is seen in BSE to contain vesicles and hidden in a box underneath African statuettes. These were magnetite dendrites (Fig. 1b) is so fresh as to suggest the part of the effects of Jean Colonna-Cimera, a senior mining possibility that the meteorite was an observed fall. On the engineer in Africa and SE Asia (between 1940 and 1955) basis of its Na/K ratio (see below), it appears that the sample was collected before being exposed to rain ( Haack et al., 2012). It was probably presented to M. Colonna-Cimera, because of his mining expertise and administrative respon- sibilities, soon after its fall. Paris was classified as a CM chondrites in 2008 by the second author using petrographic techniques and reported by Bourot-Denise et al. (2010) . The classification was confirmed with chemical, oxygen isotopic and chromium isotopic data (Bourot-Denise et al., 2010; Zanda et al., 2010, 2011a,b; Go¨pel et al., 2011 ). Here, we present new bulk analyses of Paris and other CM chondrites, for comparison. Paris is heterogeneous, with domains with hydrated minerals much less developed than in other members of this group, though sufficient for classification as CM2, and anhydrous silicates show no effects related to thermal metamorphism, as in type 3.0 chondrites ( Bourot-Denise et al., 2010; Zanda et al., 2010, 2011a; Hewins et al., 2011 ). Though it has experienced aqueous alteration, petrographic and oxygen isotope evidence place it closer to CO chondrites than most other CM chondrites ( Bourot-Denise et al., 2010; Zanda et al., 2010, 2011a ). There have been several attempts to devise numerical alteration indices for CM chondrites. The alteration index developed by Browning et al. (1996) was based on the Fe 3+ /Si ratio in fine matrix, which reflects the replacement of cronstedtite by chrysotile–greenalite solid solution. Ru- bin et al. (2007) devised a composite index to CM alteration, tracking metal abundance, alteration of chondrule phenocrysts, but in which the composition of PCP (poorly characterized phase) aggregates rather than matrix plays an important role. The petrologic (alteration) subtypes defined range from 2.6 (moderately altered) to 2.0 (totally altered). Rubin et al. (2007) showed that this sequence correlated Fig. 1. (a) The 1.3 kg Paris CM chondrite, covered with fresh with decreasing S/SiO2 and RFeO/SiO2 in PCP, i.e. the ap- fusion crust but with the interior exposed at the top. (b) BSE image proach to a simple serpentine composition. We attempted of the vesicular fusion crust with magnetite dendrites (indicated by to use these composition relations to define the petrologic white arrow in vesicle) on Paris 2010-1. 192 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 subtype of Paris, to see whether the earliest stages of CM the focused ion beam (FIB) technique using an FEI Strata alteration fitted into the sequence. Palmer and Lauretta DB 235 at the University of Lille (see Zega et al., 2007 , for (2011) established an aqueous alteration scale 0–4, based detailed preparation procedure). The FIB sections were ex- on the replacement by S-rich water of metal grains in ma- tracted from the matrix at different distances from a chon- trix or exposed to fluids coming from matrix. drule. They were examined by TEM with an FEI Tecnai Some CM chondrites are poor in hydrous minerals be- G2-20 TWIN (LaB6, 200 kV) equipped with an energy cause of thermal metamorphism ( Nakamura, 2005). We dispersive spectrometer (EDS) at the University of Lille. used several approaches to search for evidence of reheating Mineralogy was studied using bright field TEM imaging – the distribution of Cr in olivine between chromite ( Engi, and electron diffraction. Scanning TEM (STEM) was used 1983; Johnson and Prinz, 1991) and changes in olivine Cr for acquiring EDS data for microanalysis. Quantitative content due to precipitation of chromite ( Grossman and analyses were obtained with k-factors (experimentally Brearley, 2005). We also studied inclusions containing Si, determined) and absorption corrections. Elemental distri- Cr, P or S in metal, which may be formed during slow chon- butions were obtained by EDS X-ray intensity maps, using drule cooling or during metamorphism ( Zanda et al., 1994 ). spectral imaging wherein each pixel of a spectrum image Finally we examined metal and sulfide compositions and contains a full EDS spectrum (see Leroux et al. (2008) for textures, which can indicate the presence and extent of sec- details). ondary heating (Kimura et al., 2011 ). We conclude that Paris has suffered little modification 2.3. Chemical composition and is the least aqueously altered CM currently known. We present here observations on this exceptional sample, A 13 g sample was crushed using a boron carbide mortar which may contain much well preserved pre-accretionary and pestle into a homogenous fine-grained powder in clean material not highly susceptible to aqueous alteration and room conditions at the Institut Universitaire Europeén de allow us to document the earliest stages of alteration on la Mer (IUEM), Plouzane´. A 500 mg aliquot was dissolved the CM parent body. and analyzed for major and trace element concentrations by ICP-AES (inductively coupled plasma – absorption 2. SAMPLES AND METHODS emission spectrometry) using a Horiba Jobin Yvon Ultima 2 spectrometer, and by ICP-SFMS (inductively coupled 2.1. Polished sections plasma – sector field mass spectrometry) using a Thermo Element 2 spectrometer following the procedures described The Paris meteorite NMHN 4029 was first studied in by Barrat et al. (2012) . The concentration reproducibility is thin section 2008 LM, and then in parallel polished sections generally much better than 5% at the chondritic level. 2010-1 through -9, and 2011-1,-2, using optical microscope, Major and trace element concentrations were measured scanning electron microscope (SEM), electron probe micro- by LA-ICP-MS using a New Wave UP193FX ArF excimer analyzer (EPMA), and laser ablation inductively coupled laser system coupled to a Thermo Element XR at the Plas- plasma mass spectrometry (LA-ICP-MS). A sub-sample ma Analytical Facility at the National High Magnetic Field called 4029-2.5 was later sectioned. Wavelength-dispersive Laboratory, Florida State University. This involved in-situ quantitative analyses of minerals were made on the Cameca rastering with LA-ICP-MS of a 0.25 cm 2 square in a com- SX100 electron microprobe at the Universite´ Paris VI, gen- paratively altered region of section 2010-4. In addition, spot erally using 15 kV and 10 nA, except for trace elements in analyses by LA-ICP-MS were used to analyze matrix in the olivine, where we used 300 nA beam current. The sections least altered regions as well as in more altered regions were mapped and textures imaged on a Tescan VEGA II resembling more typical CM chondrites. Major aspects of LSU SEM at the LMCM in the Museúm National d’His- the instrumentation and methodology have been described toire Naturelle (MNHN) in conventional mode, and miner- previously (Humayun et al., 2007, 2010; Gaboardi and als were characterized with an SD 3 (Bruker) energy- Humayun, 2009; Humayun, 2012). disperserive spectrometer (EDS) detector. One section was studied by NanoSIMS at the MNHN by Mostefaoui 2.4. Oxygen isotopes (2011) using a focused Cs + ion beam rastered over 9 Â 9 lm2 areas. Negative secondary ions of the three O- High-precision oxygen isotopic measurements were per- isotopes, 28 Si, and 27 Al 16 O were simultaneously measured formed at the Open University using an infrared laser-as- in multi-collection mode in an unsuccessful search for pre- sisted fluorination system (Miller et al., 1999 ). A total of solar silicates or oxides. A separate section (SPSR-1) was eleven individual analyses of whole rock chips and powders prepared for Remusat et al. (2011) without using epoxy. from Paris were undertaken, with each replicate having a This was also studied by NanoSIMS using rastering of a fo- mass of about 2 mg. After fluorination, the O 2 released was cused Cs + beam to generate secondary ion images of H À, purified by passing it through two cryogenic nitrogen traps À 12 À 16 À 26 À 28 À 32 À D , C , O , CN , Si and S . and over a bed of heated KBr. O 2 was analyzed using a MAT 253 dual inlet mass spectrometer. Analytical precision 2.2. TEM sections (1 r), based on replicate analyses of international (NBS-28 quartz, UWG-2 garnet) and internal standards, is approxi- Four electron-transparent sections were prepared from mately ± 0.04 & for d17 O; ±0.08 & for d18 O; ±0.024 & for SPSR-1 for transmission electron microscopy (TEM) by D17 O ( Miller et al., 1999 ). Oxygen isotopic analyses are R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 193 reported in standard d notation, where d18 O has been calcu- demagnetization was performed using an MMTD furnace, 18 18 16 18 16 lated as: d O = [( O/ Osample/ O/ Oref.)À1] Â 1000(&) under argon atmosphere above 250 °C. Remanence mea- and similarly for d17 O using the 17 O/ 16 O ratio. D17 O, which surements were performed with a SQUID cryogenic magne- represents the deviation from the terrestrial fractionation tometer (2G Enterprises, model 755R, with noise level of line, has been calculated as: D17 O = d17 O–0.52d18 O in order 10 À11 Am 2). to compare our results with those obtained by Clayton and Mayeda (1999). 2.7. Mg and Cr isotopes

2.5. Gas analyses Mineralogical separates were prepared at the IPGP by using the freeze–thaw method and subsequent separation Two fragments of Paris (Paris-A and Paris-B) were ana- by magnetic susceptibility and by density. Handpicking lyzed in order to determine their noble gas abundances and produced pure magnesian and ferroan olivine from Type isotopic compositions. Each sample was weighted and I and II chondrules, a separate of fine-grained material at- placed in a hemispherical cavity in a laser chamber devel- tached to chondrules (presumably fine-grained rims). These oped at the CRPG for the analysis of nitrogen (see Hum- and an aliquot of the bulk rock from the large IUEM sam- bert et al., 2000 for details). The chamber was maintained ple were used for Cr isotope analysis by mass spectrometry for 48 h at 373 K under vacuum to remove atmospheric (Go¨pel et al., 2011 ). Metal and sulfide separates were also contamination. The samples were heated for 1 min with a produced, the latter for S isotope measurements by P. Car- CO 2 laser beam (wavelength of 10.6 lm) to different tem- tigny. Pure hibonite grains handpicked from these fractions perature steps by increasing the laser power, the last step were mounted in epoxy, polished and studied with SEM to corresponding to the fusion of the sample ( Marty et al., identify their mineralogy and morphology. Magnesium iso- 2010). The released gas was purified using two Ti–Zr getters tope measurements were performed on the hibonite with to remove active gases (10 min at 800 °C, 10 min at room CAMECA 1280HR2 at CRPG ( Liu et al., 2012 ). temperature). Analysis of noble gases was performed with a static sector-type mass spectrometer in monocollection 3. CLASSIFICATION, ISOTOPES AND CHEMISTRY mode (Marrocchi et al., 2011 ). Each sample has been brack- eted by two standards for which the amounts of noble gases Paris is classified as a CM2 chondrite. It consists mainly were adjusted to be comparable to the sample abundance. of Type I chondrules in a matrix containing chrysotile– The blanks determined for laser extraction by heating an greenalite–cronstedtite serpentine and PCP (fine-grained empty cavity were less than 0.1% and correspond to (in cronstedtite–tochilinite intergrowths in highly altered CM cm 3) 1.15 Â 10 À10 , 6.51 Â 10 À12 , 2.44 Â 10 À10 and chondrites). It has a grain density of 2.92 ( Zanda et al., 2.78 Â 10 À15 for 4He, 20 Ne, 40 Ar and 130Xe, respectively. 2010), typical of CM chondrites ( Consolmagno et al., 2008). Bulk chemical and O and Cr isotopic data reported 2.6. Magnetic properties below also support a classification for Paris as CM chondrite. We observed no evidence of terrestrial weathering All magnetic measurements were performed at CEREGE in trace element concentrations, or of shock. (Aix-en-Provence, France), with the exception of magnetic Apparent matrix content has been considered as a property measurement system (MPMS) measurements (at measure of the extent of alteration of CM chondrites ( McS- IPGP, Paris, France). Magnetic hysteresis measurements ween, 1979). Preliminary BSE image analyses ( Bourot-De- were performed with a Princeton Micromag Vibrating Sam- nise et al., 2010 ) showed that Paris has less matrix than ple Magnetometer with a maximum applied field of 1 T and Murchison (66% vs. 71%), which McSween (1979) showed a sensitivity of 5 Â 10 À9 Am 2. The analysis of hysteresis to be one of the least matrix-rich, “partially altered” CM loops provided the ratio of saturation remanent magnetiza- falls. We attempted to extend this approach by identifying tion (M RS ) to saturation magnetization (M S) and the coer- chondrule fragments in BSE images down to about cive force (B C). High field susceptibility ( vHF ) was 100 lm and also by counting fine-grained rims on chond- determined by a linear fit for applied fields >0.9 T of the hys- rules separately. Matrix abundance was thus determined teresis loops. Remanent coercive force (B CR ) was determined to be approximately 47%, with 8% fine grained rims by DC back field experiments performed with the VSM. We (FGR) and 45% chondrules. Modal analysis showed 57– measured the S ratio, defined as the isothermal remanent 85% matrix in other CM chondrites ( Grossman and Olsen, magnetization (IRM) obtained after applying a 3 T field 1974; McSween, 1979). and then a back field of À0.3 T normalized to the 3 T IRM. Low-temperature remanence measurements were per- 3.1. Oxygen isotopes Ò formed with an MPMS from Quantum Design . This instrument has a moment sensitivity of 10 À11 Am 2. The low field The results of oxygen isotope measurements of Paris are specific susceptibility (written as v in m 3/kg) and its evolu- given in Table 1 and plotted in Fig. 2a and b. The eleven tion with temperature were measured using an Agico individual analyses of Paris obtained in this study show a MFK1 apparatus with a sensitivity of 5 Â 10 À13 m3, operat- wide range of values with respect to d17 O and d18 O ing at 200 A/m and a frequency of 976 Hz, equipped with a (Fig. 2a), reflecting the isotopically heterogeneous composi- CS3 furnace and a CSL cryostat. IRMs were imparted using tion of its constituent components (chondrules, CAIs, crys- a pulse magnetizer from Magnetic Measurements. Thermal tal fragments, hydrated phases). From an oxygen isotope 194 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Table 1 Oxygen isotope results. Sample COMMENTS d17 O& 1r d18 O& 1r D17 O& 1r Paris (unaltered) Less altered lithology À2.11 2.43 À3.37 Paris (unaltered) Less altered lithology À1.43 4.40 À3.72 Paris (unaltered) Less altered lithology À0.37 6.14 À3.56 Paris (altered) More altered lithology 0.33 6.21 À2.91 Paris (altered) More altered lithology 0.75 6.80 À2.79 Paris (chips) Chips À1.15 4.09 À3.28 Paris (chips) Chips À1.87 3.63 À3.76 Paris (chips) Chips À2.15 3.44 À3.94 Paris Powder 0.62 6.97 À3.00 Paris Powder À0.66 5.11 À3.31 Paris Powder À1.20 4.68 À3.63 MEAN (n = 11) À0.84 1.06 4.90 1.48 À3.39 0.39

slope of 0.69 ( R2 = 0.93) and an intercept of À4.23& (Fig. 2b). In comparison, the combined CM2 fall and find data of Clayton and Mayeda (1999) show significant scatter and define a shallower trend with a slope of 0.57 ( R2 = 0.76) and an intercept of À3.0& (Fig. 2a). The CM2 data of Clayton and Mayeda (1999) included a significant number of analyses from Antarctic finds. If CM falls data alone are considered, a much simpler picture emerges. In Fig. 2a we have plotted just the falls data from Clayton and Mayeda (1999), plus analyses of two other recent CM2 falls: Maribo (Haack et al., 2012 ) and Sayama (Grossman and Zipfel, 2001 ). CM2 fall data define a linear array that is significantly steeper than the combined CM2 fall and find data, having a slope of 0.79 ( R2 = 0.95) and an intercept of À4.9& (Fig. 2a). This suggests that the relatively shallow slope of the combined fall and find CM2 dataset obtained by Clayton and Mayeda (1999) is princi- pally a reflection of the very large number of weathered Antarctic finds that it contains. Compared to meteorite falls, weathered Antarctic finds are generally displaced to less negative D17 O values, as a result of interaction with terrestrial precipitation (Greenwood and Franchi, 2004; Greenwood et al., 2012 ). However, unlike other meteorite Fig. 2. (a) Oxygen isotope composition of Paris meteorite com- groups, it is not possible to use leaching techniques on pared to other CM falls and finds. A1 and A2 are altered lithologies from Paris. L1, L2 and L3 are less altered lithologies from Paris. CM2 samples to mitigate the effects of weathering, as such The mean composition of the Paris meteorite is shown by the open treatment will also remove the indigenous low temperature circle. The dashed regression line is fall and finds data from component. Clayton and Mayeda (1999) only, whereas the solid regression line Clayton and Mayeda (1984, 1999) demonstrated that the if for finds and includes the data of Clayton and Mayeda (1999) anhydrous silicates in CM2s have an oxygen isotopic com- and the recent CM falls Sayama ( Grossman and Zipfel, 2001) and position similar to that of CO3 chondrites and suggested Maribo (Haack et al., 2012 ). (b) Best fit linear regression line that there may be a genetic link between the two groups. through the oxygen isotope data for Paris samples only. The fact The extension of the linear regression line fitted to the Paris that this regression line passes through the relatively narrow field data intersects the CO3 field of Greenwood and Franchi defined by CO3 chondrites provides strong evidence that both the (2004) (Fig. 2a). The slope of the Paris regression line is CM and CO groups are genetically related. CM data: Clayton and Mayeda (1999), Grossman and Zipfel (2001), Haack et al. (2012) . essentially identical in slope to that proposed by Clayton CO3 data: Greenwood and Franchi (2004) . and Mayeda (1999) as being a best fit through their CO and CM whole rocks and CM matrix separates. The fact that the regression line for Paris passes through the rela- perspective, CM2 meteorites are essentially a two compo- tively narrow field defined by CO3 chondrites provides nent mix between 16 O-enriched anhydrous silicates and strong evidence that both the CM and CO groups are 16 O-depleted hydrated phases (Clayton and Mayeda, genetically related. In addition, analyses from Paris signifi- 16 1999). This can clearly be seen in the case of the Paris mete- cantly extend the O-enriched end of the CM2 trend orite, with analyses defining a distinct linear trend with a (Fig. 2a and b); a feature that is consistent with it being less R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 195 hydrated than the least hydrated CM fall and having the 54 Cr in the different materials has especially to be consid- least matrix abundance of any member of the group so ered), the corresponding age for the Paris components, far studied. particularly of Type I and Type II chondrules, is In addition to bulk powders from Paris, chips from two 4566.44 ± 0.66 Myr based on the D’Orbigny anchor ( Glavin distinct lithologies, one altered and one less altered, were et al., 2004; Amelin, 2008 ) recalculated with the U isotopic analyzed during the course of these study. Analyses from composition measured by Brennecka and Wadwha (2012). the less altered material are labeled L1 to L3 on Fig. 2a, The Mn–Cr study suggests early formation ages for the analyses from the more altered lithology are labeled A1 Paris chondrules (<1 Myr after CV CAIs and confirms a and A2 on Fig. 2a. The less altered lithology shows a signif- previous Mn/Cr study on Allende chondrules revealing the icant spread in oxygen isotope compositions and includes antiquity of Allende chondrules ( Yin et al., 2007, 2009 ). the most 16 O-rich material yet analyzed from a CM2 chon- Such an old age is also compatible with a more recent U/ drite (point L1). In comparison, the altered lithology ap- Pb study on refractory inclusions from CV chondrites (Efre- pears to have a more uniform 16 O-poor composition (Fig movka and Allende) as well as on individual chondrules 2a). from Allende and the unequilibrated OC NWA 5697. These U/Pb isotope data define a short formation interval for the 3.2. Chromium isotopes CAIs (4567.30 ± 0.16 Myr), while chondrule ages range from 4567.32 ± 0.42 to 4564.1 ± 0.30 Myr ( Connelly et al., The sequential dissolution pattern, a feature distinguish- 2012). The authors emphasize “these data refute the ing between the different carbonaceous chondrite classes, is long-held view of an age gap between CAIs and chondrules very similar to that of the Murchison CM. The e54 Cr value and indicate that chondrule formation started contempora- of the bulk rock, e54 Cr = 0.925 ± 0.094 ( Go¨pel et al., 2011 ), neously with CAIs”. Such an early chondrule formation is as falls on the correlation line that has been established be- yet difficult to reconcile with the accretion age of the CM tween e54 Cr and D17 O for carbonaceous chondrites ( Trin- parent body (3 Myr after CAIs) and with the Mg/Al quier et al., 2007 ). This is shown by Go¨pel et al. (2012) isotope data of CO chondrules that systematically formed and is consistent with the classification of Paris as a CM 2–2.5 Myr after CAIs (Kita and Ushikubo, 2012; Fujiya chondrite. et al., 2013 ). Forsteritic olivine, fayalitic olivine, respectively from Type I and Type II chondrules, a separate of fine-grained 3.3. Bulk chemistry material attached to chondrules, presumably in fine-grained rims (FGR) and an aliquot of the bulk rock were analyzed Paris, and the CM chondrites Nogoya and Boriskino (2 for Mn/Cr systematics (Go¨pel et al., 2011, 2012 ). The data samples) were analyzed together at UBO. With 32.05 wt.% are presented in Go¨pel et al. (2013) . Fe 2O3 total, 28.72 wt.% SiO2, 28.72 wt.% MgO and LOI All mineral fractions as well as the bulk rock of Paris ex- 12.9 wt.% (Table 2), the aliquot from the 13 g sample of hibit a positive e54 Cr anomaly typical for carbonaceous Paris falls well within the ranges for CM chondrites (27– chondrites. The mineral samples fall on a line with a slope 30, 30–40, and 19–20, respectively; Jarosewich, 1990) and of 53 Mn/55 Mn = (5.76 ± 0.76) Â 10 À6 with an initial major element abundances determined by laser ablation 53 Cr i = À0.132 ± 0.055. are similar. If we interpret these isotopic data of the minerals as pro- The existence of composition clusters in plots of trace viding direct chronological information (the homogeneity of and minor elements of different volatility, corresponding

Table 2 Major element, Cr, Ni and Co abundances in Paris, Nogoya and in Boriskino determined by ICP-AES and compared with selected literature values for Murchison (F: Friedrich et al. (2002) ; KW: Kallemeyn and Wasson (1981); MW: Mittlefehldt and Wetherill (1979); WP: Wolf and Palme (2001)). Oxides and loss on ignition (L.O.I.) in wt.% and Cr, Co and Ni in lg/g. Paris Nogoya Boriskino 1 Boriskino 2 Murchison (literature) WP SiO2 28.72 28.84 WP TiO2 0.10 0.10 0.10 0.11 0.11 WP Al 2O3 2.07 1.96 1.95 2.21 2.19 WP Fe 2O3t 32.05 29.54 30.54 33.02 30.60 MnO 0.22 0.22 0.23 0.24 0.22WP MgO 19.83 18.58 18.43 21.19 20.13WP CaO 1.61 1.84 1.06 1.51 1.79WP KW Na 2O 0.67 0.61 0.57 0.41 0.58 MW K2O 0.04 0.05 WP P2O5 0.28 0.24 0.24 0.25 0.24 L.O.I.** 12.88 Cr 3202 2801 2905 3162 3070WP Co 651 555 597 649 548 F Ni 14170 12660 13570 14360 12600 WP ** Includes 1.93% H, 3.29% C. 196 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 to different degrees of volatile depletion, gives us a useful elements determined for regions analyzed by LA-ICP-MS tool for the taxonomy of carbonaceous chondrites (e.g. (Table 4) are compared with the bulk composition of Paris Kallemeyn et al., 1994 ). We used such plots to compare measured by ICP-AES/SFMS ( Tables 2 and 3), and the both the data by ICP-AES/SFMS ( Table 3) and LA-ICP- average bulk composition of CM chondrites measured by MS ( Table 4) for Paris with other CM, CI and CO chon- radiochemical neutron activation analysis (Wolf et al., drites (Kallemeyn and Wasson, 1981; Friedrich et al., 1980) in Fig. 4. As observed by Wolf et al. (1980) , CMs re- 2002; Barrat et al., 2012 ). Paris falls with the CM chondrites flect a factor of two depletion in the volatile elements rela- for Zn/Mn vs. Sm/Mn or Al/Mn ratios, and is distinct from tive to CI chondrites, consistent with CMs being a 50:50 CI and CO chondrites. mixture of CI-composition matrix and refractory materials The REE pattern in Paris is extremely flat and resembles (chondrules, etc.), in about the same ratio as observed pet- those of most other CM chondrite falls especially Murchi- rographically (Grossman and Olsen, 1974; McSween, son (Fig. 3a). Paris shares the depletion pattern of elements 1979). The bulk composition of Paris determined by LA- more volatile than Mg (which is smooth with the exception ICP-MS (black diamonds) and by ICP-AES/SFMS (gray of high Na relative to K) of other CM chondrites, particu- circles) show the volatile depletion characteristic of CM larly falls (Fig. 3b). Haack et al. (2012) argue that Na is chondrites (blue squares). The altered matrix analysis is depleted in CMs if, as for the Maribo fall, they are exposed the average of two 50 lm spots (M3/4) analyzed by LA- to rain before being collected. ICP-MS (red diamonds) and exhibits a volatile-depleted Differences in bulk compositions of fresh and altered pattern that closely follows the rastered bulk composition matrix are of interest. Abundances of major and volatile with the prominent exception of Li, P, K and Rb. In

Table 3 Trace element abundances (in lg/g) obtained by ICP-SFMS for Paris, Nogoya and Boriskino, and compared with selected literature values for Murchison (F: Friedrich et al. (2002) adjusted to the same standard values as Barrat et al. (2012) ; N: Nakamura (1974), Tatsumoto et al. (1976). Paris Nogoya Boriskino 1 Boriskino 2 Murchison Murchison F N,T Li 1.60 1.51 1.53 1.61 1.7 Be 0.0315 0.0317 0.0303 0.0325 P 1086 1024 997 924 K 380 397 440 237 Sc 8.38 7.69 7.93 8.39 8.39 Ti 628 609 646 629 634 V 72.1 70.2 73.7 73.7 70 Cu 128 125 130 123 116 Zn 180 166 177 168 191 Ga 7.71 7.45 6.85 7.67 7.34 Rb 1.72 1.53 1.81 1.07 1.66 Sr 10.79 10.24 8.01 9.30 10.30 Y 2.11 1.95 2.05 2.10 2.18 Zr 4.83 4.51 4.78 5.09 5.1 Nb 0.384 0.383 0.373 0.410 0.41 Cs 0.131 0.115 0.132 0.114 0.127 Ba 3.24 3.00 3.19 2.98 3.23 3.08 N La 0.329 0.311 0.395 0.329 0.314 0.321 N Ce 0.833 0.776 0.974 0.832 0.815 0.848 N Pr 0.126 0.117 0.137 0.126 0.124 Nd 0.644 0.610 0.675 0.646 0.635 0.641 N Sm 0.211 0.196 0.211 0.212 0.209 0.209 N Eu 0.0797 0.0735 0.0788 0.0801 0.079 0.080 N Gd 0.286 0.273 0.289 0.295 0.282 0.278 N Tb 0.0523 0.0480 0.0525 0.0525 0.052 Dy 0.351 0.324 0.348 0.355 0.351 0.344 N Ho 0.0769 0.0713 0.0739 0.0763 0.079 Er 0.225 0.209 0.217 0.225 0.231 0.222 N Tm 0.0355 0.036 Yb 0.231 0.214 0.229 0.227 0.229 0.226 N Lu 0.0332 0.0305 0.0323 0.0332 0.034 0.036 N Hf 0.144 0.139 0.145 0.154 0.155 Ta 0.0192 0.0206 0.0192 0.0197 W 0.073 0.087 0.087 0.134 0.18 Pb 1.51 1.47 1.75 1.49 1.60 T Th 0.0386 0.0366 0.0404 0.0389 0.039 U 0.0101 0.00940 0.0103 0.00973 0.0098 0.0110 T R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 197

Table 4 LA-ICP-MS analyses of section 2010-4 by in-situ rastering of a 0.25 cm 2 square in a comparatively altered region (SP4). Spot analyses by LA- ICP-MS of matrix: M1/2 less altered zone, M3/4 more altered zone. Atomic mass Concentrations SP4 Average CI chondrites M3/4 average M1/2 average

Li 7 Li7(LR) lg/g 1.539 1.500 3.117 2.216 Be 9 Be9(LR) lg/g 0.021 0.025 0.047 0.039 B 11 B11(LR) lg/g 5.344 0.870 5.761 5.725

Na 2O 23 Na23(LR) %m/m 0.783 0.674 0.850 1.326 MgO 25 Mg25(LR) %m/m 21.158 16.400 16.980 17.630

Al 2O3 27 Al27(LR) %m/m 2.939 1.640 2.745 3.436 SiO2 29 Si29(LR) %m/m 33.787 22.765 31.199 33.851 P2O5 31 P31(LR) %m/m 0.336 0.279 0.465 0.359 S 34 S34(LR) %m/m 4.796 6.250 5.398 7.004 Cl 35 Cl35(LR) lg/g 829.924 704.000 700.382 1149.652

K2O 39 K39(LR) %m/m 0.092 0.067 0.129 0.151 CaO 44 Ca44(LR) %m/m 2.289 1.298 3.490 2.266 Sc 45 Sc45(LR) lg/g 11.218 5.820 7.771 9.041

TiO2 47 Ti47(LR) %m/m 0.144 0.073 0.105 0.118 V 51 V51(LR) lg/g 85.579 56.500 82.895 68.637 Cr 53 Cr53(LR) lg/g 3812.490 2660.000 3484.190 3369.411 MnO 55 Mn55(LR) %m/m 0.309 0.257 0.302 0.378 FeO(t) 57 Fe57(LR) %m/m 33.367 24.495 38.335 33.480 Co 59 Co59(LR) lg/g 665.524 502.000 740.556 611.397 Ni 60 Ni60(LR) lg/g 23110.148 11000.000 26083.183 22418.334 Cu 63 Cu63(LR) lg/g 167.660 126.000 174.263 222.204 Zn 66 Zn66(LR) lg/g 283.063 312.000 289.094 608.848 Ga 71 Ga71(LR) lg/g 11.635 10.000 12.040 15.321 Ge 74 Ge74(LR) lg/g 27.561 32.700 30.424 40.293 As 75 As75(LR) lg/g 1.706 1.860 1.600 2.217 Se 82 Se82(LR) lg/g 15.629 18.600 15.868 26.763 Br 79 Br79(LR) lg/g 18.813 3.570 49.471 67.868 Kr Kr83(LR) Rb 85 Rb85(LR) lg/g 2.863 2.300 3.513 3.875 Sr 88 Sr88(LR) lg/g 15.214 7.800 17.611 17.514 Y 89 Y89(LR) lg/g 2.664 1.560 2.015 2.184 Zr 90 Zr90(LR) lg/g 6.973 3.940 5.196 5.396 Nb 93 Nb93(LR) lg/g 0.580 0.246 0.388 0.525 Mo 97 Mo97(LR) lg/g 1.653 0.928 1.756 1.392 Ru Ru102(LR) 1.093 0.712 0.988 0.893 Rh 103 Rh103(LR) lg/g 0.187 0.134 0.157 0.144 Pd 106 Pd106(LR) lg/g 0.666 0.560 0.598 0.564 Ag 107 Ag107(LR) lg/g 0.312 0.199 0.240 0.377 Cd 111 Cd111(LR) lg/g 0.504 0.686 0.545 0.997 In 115 In115(LR) lg/g 0.070 0.080 0.078 0.127 Sn 120 Sn120(LR) lg/g 1.057 1.720 1.148 1.632 Sb 121 Sb121(LR) lg/g 0.148 0.142 0.150 0.246 Te Te125(LR) 0.762 2.320 0.883 1.648 I I127(LR) 0.856 0.433 1.410 2.217 Cs 133 Cs133(LR) lg/g 0.794 0.187 0.706 0.369 Ba 138 Ba138(LR) lg/g 4.847 2.340 4.689 5.130 La 139 La139(LR) lg/g 0.452 0.235 0.330 0.410 Ce 140 Ce140(LR) lg/g 1.174 0.603 0.957 1.053 Pr 141 Pr141(LR) lg/g 0.170 0.089 0.125 0.146 Nd 145 Nd145(LR) lg/g 0.885 0.452 0.622 0.748 Sm 147 Sm147(LR) lg/g 0.295 0.147 0.207 0.251 Eu 153 Eu153(LR) lg/g 0.117 0.056 0.097 0.127 Gd 158 Gd158(LR) lg/g 0.381 0.197 0.269 0.303 Tb 159 Tb159(LR) lg/g 0.067 0.036 0.047 0.054 Dy 164 Dy164(LR) lg/g 0.476 0.243 0.339 0.365 Ho 165 Ho165(LR) lg/g 0.101 0.056 0.074 0.085 Er 166 Er166(LR) lg/g 0.311 0.159 0.224 0.252 Tm 169 Tm169(LR) lg/g 0.049 0.024 0.032 0.042 Yb 174 Yb174(LR) lg/g 0.336 0.163 0.213 0.278 198 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Table 4 ( continued)

Atomic mass Concentrations SP4 Average CI chondrites M3/4 average M1/2 average

Lu 175 Lu175(LR) lg/g 0.048 0.024 0.034 0.036 Hf 180 Hf180(LR) lg/g 0.202 0.104 0.138 0.144 Ta 181 Ta181(LR) lg/g 0.029 0.014 0.021 0.025 W 182 W182(LR) lg/g 0.183 0.093 0.193 0.151 Re 185 Re185(LR) lg/g 0.068 0.037 0.046 0.052 Os 192 Os190(LR) lg/g 0.801 0.486 0.757 0.633 Ir 193 Ir193(LR) lg/g 0.711 0.481 0.623 0.522 Pt 195 Pt195(LR) lg/g 1.384 0.990 1.197 1.228 Au 197 Au197(LR) lg/g 0.385 0.140 0.347 0.384 Hg 202 Hg202(LR) lg/g 2.939 0.258 2.533 7.034 Tl 205 Tl205(LR) lg/g 0.162 0.142 0.150 0.219 Pb 208 Pb208(LR) lg/g 2.994 2.470 2.984 4.902 Bi 209 Bi209(LR) lg/g 0.131 0.114 0.112 0.192 Th 232 Th232(LR) lg/g 0.061 0.029 0.035 0.044 U 238 U238(LR) lg/g 0.017 0.008 0.015 0.017

3 21 38 comparison, the average of two 50 lm spots taken on ma- concentration of He cos, Ne cos and Ar cos from the heat- trix in unaltered regions (green diamonds) has undepleted ing extraction data using the protocol proposed by Eugster abundances like CI chondrite ( Zanda et al., 2011a,b ). et al. (2007) . Due to the high abundance of 4He of Paris (>2 Â 10 À5 cm 3 gÀ1; Table 5), we took into account the 3.4. Noble gases contribution of radiogenic 4He from U and Th decay on 3 the calculated abundance of He cos (Eugster et al., 2007 ). All the noble gas characteristics reveal that Paris is As the amounts of U and Th of Paris have not been deter- clearly linked to the solar gas-rich meteorites. It contains mined, we used the average concentration in CM chondrites large amounts of light noble gases with 4He and 20 Ne con- to determine the abundance of radiogenic 4He ( Wasson and À4 3 À1 À7 3 centrations higher than 1 Â 10 cm g and 1 Â 10 - Kallemeyn, 1988). The He c ages are 2.81 ± 0.41 and cm 3 gÀ1, respectively ( Table 5). These concentrations are 2.75 ± 0.18 Ma for samples A and B, respectively ( Table 7), 21 characteristic of gas-rich meteorites ( Eugster et al., 2007 ) in good agreement with the Ne c ages of 3.65 ± 0.10 and 38 and are best explained by the occurrence of a solar wind 3.55 ± 0.08 Ma. The Ar c ages were not determined be- component. This is conﬁrmed by the 4He/20 Ne ratio that cause the argon content of Paris is dominated by the Q falls within the range determined for the Solar gas-rich component, precluding a good estimation of the abundance meteorites (Eugster et al., 2007 ). In addition, the low of cosmogenic 38 Ar. The CRE ages determined for Paris fall 3He/4He of 3.4–4.2 Â 10 À4 (Table 5) are near the range of within the range of those reported for other CM carbona- solar values (4.64 ± 0.09 Â 10 À4, Heber et al., 2009 ) but dif- ceous chondrites that are characterized by short CRE ages fer from other He components trapped in gas-poor meteor- (Eugster et al., 2006 ). The reasons why CMs have such ites. Likewise, the 20 Ne/22 Ne ratios of 9.61–11.42 indicate short exposure ages are not quite clear but might be linked the presence of solar noble gases (13.78 ± 0.03; Heber to their fragile natures that reduce their ability to survive et al., 2009 ). under space conditions (Scherer and Schultz, 2000). In The 20 Ne/36 Ar ratio and the argon and xenon isotopic any case, the noble gas signatures reported in this study an- compositions (Fig. 5, Tables 5 and 6 ) point out that the chor Paris into the gas-rich meteorite clan and present char- so-called Q component (a ubiquitous noble gas component acteristics close to the CM carbonaceous chondrites. trapped in chondrites; Busemann et al., 2000 ) is also present in Paris. This component, trapped in the insoluble organic 4. PETROGRAPHY, MATRIX AND ALTERATION matter of primitive meteorites ( Marrocchi et al., 2005 ), dominates the heavy noble gas signature of Paris in good 4.1. Petrographic overview agreement with the noble gas characteristics of other CMs (Busemann et al., 2000 ). Paris contains more abundant, well preserved metal in The cosmic-ray exposure age (hereafter CRE age) repre- Type I chondrules and in matrix than other CMs. Type sents the time of exposure to the galactic cosmic rays of a Im (metal-bearing) granular/porphyritic chondrules ( McS- meter-sized meteoroid until its capture by the Earth. This ween, 1977) are especially prominent. We show a BSE im- 3 duration can be calculated from the excess of He cos, age of Paris 2010-04 ( Fig. 6), with contrast such that 21 38 Ne cos and Ar cos relative to solar isotopic composition. forsterite is dark grey, ferroan olivine and PCP are light 3 21 38 Cosmogenic production rates of He cos, Ne cos and Ar cos grey and metal is white. The chondrite is less altered at were computed from the chemical composition of Paris the bottom left and more altered at the top right: the con- determined in this study ( Table 2) and following the proce- trast in metal abundance between the fresher and more al- dure of Eugster and Michel (1995). We also determined the tered domains is striking. There is approximately 3% metal R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 199

Fig. 4. CI- and Si-normalized abundances of major and volatile elements determined by LA-ICP-MS from petrographically chosen contexts in diﬀerent Paris regions (bulk, more altered and less altered matrix) as a function of condensation temperature. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)

Fig. 5. Bulk 136Xe/130Xe and 134Xe/130Xe ratios of Paris (ﬁlled

square) determined with a CO 2 laser. The results show that the typical Q component dominates the heavy noble gas signature of Paris in accordance with the noble gas characteristics of other CMs (Busemann et al., 2000 ). Air composition data are from Ozima and Podosek (2002). Q-Xe composition data were obtained by closed- Fig. 3. Analyses of (a) REE and (b) other elements, in Paris and system sample etching (CSSE) on 8 different meteorites ( Busemann other CM chondrites by SFMS. Paris shows the volatile element et al., 2000 ). Solar wind data (SW) are from Wieler and Baur depletion typical of CM chondrites, and is almost identical to (1994). Murchison. in the freshest areas, mainly within Type I chondrules, but The mineralogical composition of altered matrix varies only traces of metal in the most altered areas. The least widely, as seen in BSE as regions of different average grey altered CM chondrite studied by Rubin et al. (2007) is tones. The widespread CM material “PCP” has been shown QUE 97990, which contains 1% metal. Type II chond- by a large body of work ( Tomeoka and Buseck, 1985; rules (<2%), refractory inclusions (<1%), and chondrule Browning et al., 1996; Rubin et al., 2007; Palmer, 2009 ) debris are also present in Paris. Chondrule silicate minerals to be intergrowths of cronstedtite and tochilinite. In Paris, are fresh and unserpentinized. Chondrule glass is preserved however, intergrowths are very fine grained, may have in inclusions in olivine, especially in Type II chondrules, but compositions poorer in both Si and S than expected, and is replaced in mesostasis by aluminous serpentine ( Table 8). contain other phases including nano-magnetite. Their The preservation of chondrules in Paris leaves little doubt composition and abundance are variable. We therefore pre- that the loose sugary aggregates of olivine and olivine crys- fer to retain the term PCP. There are darker domains in tal clasts in highly altered CM chondrites are derived from BSE images with chondrules containing fresh metal in chondrules. Although there are several lithic clasts in Paris, gradational contact with zones with altered metal and we do not recognize clastic matrix, suggesting it is mainly prominent bright grey PCP grains, e.g. in 2010-4 ( Fig. 7a, “primary rock” (Metzler et al., 1992 ). a close-up of the center of Fig. 6). There are also sharp Table 5 190–222 (2014) 124 Acta Cosmochimica et Geochimica / al. et Hewins R.H. 200 À Helium, neon and argon concentration (cm 3 g 1) and isotopic ratios for Solar gas-rich meteorites, Q-gases and two fragments of Paris. The listed uncertainties on abundances and isotopic ratios (1 r) include blank, standard and sample uncertainties Sample mass laser 3He 3He/ 4He 21 Ne 20 Ne/ 22 Ne 21 Ne/ 22 Ne 38 Ar 38 Ar/ 36 Ar 40 Ar/ 36 Ar 4He/ 20 Ne 20 Ne/ 36 Ar À À À À À À À (mg) steps (10 8 cm 3 g 1) (Â10 4) (10 9 cm 3 g 1) (10 7 cm 3 g 1) Paris-A 0.106 2 23.68 (1.81) 3.40 (0.31) 15.93 (0.51) 11.42 0.079 3.53 (0.03) 0.188 9.2 (0.2) 278.36 1.42 (0.01) (0.24) (0.004) (0.002) (0.02) Paris-B 0.295 4 13.73 (0.80) 4.21 (0.30) 12.88 (0.33) 9.61 0.110 3.04 (0.02) 0.189 9.7 (0.2) 267.73 0.82 (0.01) (0.18) (0.005) (0.001) (0.02) Solar gas-rich 3.02 (0.04) 12.24 0.032 0.179 120–600 47 (3) meteorite a (0.30) (0.016) (0.005) Qb 1.23 (0.02) 10.67 0.029 0.187 112 (10) 0.044 (0.02) (0.001) (0.003) (0.006) a Gas-rich meteorites composition determined by Eugster et al. (2007) b Q-gas composition data obtained by closed-system sample etching (CSSE) on 8 different meteorites ( Busemann et al., 2000 )

Table 6 À Xe concentration (cm 3 g 1) and isotopic ratio for Air, Solar wind, Q-gases and one fragment of Paris. Isotopic ratios Â 100. The listed uncertainties on abundances and isotopic ratios (1 r) include blank, standard and sample uncertainties. À À Sample laser steps mass (mg) 130 Xe (10 10 cm 3Ág 1) 124 Xe/ 130 Xe 126 Xe/ 130 Xe 128 Xe/ 130 Xe 129 Xe/ 130 Xe 131 Xe/ 130 Xe 132 Xe/ 130 Xe 134 Xe/ 130 Xe 136 Xe/ 130 Xe Paris-B 4 0.295 4.57 (0.03) 2.911 (0.14) 4.357 (0.17) 50.20 (0.70) 624.0 (6.8) 487.0 (5.2) 603.0 (6.0) 228.1 (2.9) 193.8 (2.3) Air * 2.337 (0.008) 2.180 (0.011) 47.15 (0.07) 649.6 (0.6) 521.3 (0.6) 660.7 (0.5) 256.3 (0.4) 217.6 (0.2) Q** 2.810 (0.004) 2.505 (0.018) 50.79 (0.02) 645.9 (0.2) 506.4 (0.3) 617.5 (0.2) 233.4 (0.2) 195.4 (0.2) SW *** 2.948 (0.017) 2.549 (0.082) 51.02 (0.54) 627.3 (0.5) 498.0 (0.2) 602.0 (0.3) 220.7 (0.9) 179.7 (0.6) * Air composition data are from Ozima and Podosek (2002) . ** Q-gas composition data obtained by closed-system sample etching (CSSE) on 8 diﬀerent meteorites ( Busemann et al., 2000 ). *** Lunar soil 75101 by the CSSE method ( Wieler and Baur, 1994 ). R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 201

Table 7 Production rates and CRE ages determined for Paris from the abundances of cosmogenic light noble gases. Production rate (10 À8 cm 3 STP/g, Ma CRE ages

P3 P21 P38 T3 T21 T38 Paris-A 1.584 0.294 0.0539 2.81 (0.41) 3.65 (0.10) – Paris-B 1.584 0.294 0.0539 2.75 (0.18) 3.55 (0.08) –

Fig. 6. BSE image of Paris section 2010-04 with metal (white), ferroan olivine (bright grey), matrix phases (light-dark grey) and magnesian olivine plus pyroxene (very dark gray). Broad gradational contact from top left to bottom right. Metal is abundant in Type I chondrules in the fresh domain near bottom left and but is replaced in the moderately altered domain near top right. The white rectangle shows the location of Fig. 7a.

Table 8 Averages (wt.%) and standard deviations for EMP analyses of alteration material.

SiO2 Al2O3 TiO2 “Cr2O3” “ FeO” MnO MgO CaO Na2O K2O “P2O5” S Ni Total Serpentine matrix 27.17 2.86 0.07 0.41 33.72 0.27 14.34 0.49 0.52 0.07 0.16 2.96 1.92 84.95 Serpentine chondrule 25.49 7.46 0.48 0.07 33.57 0.33 10.01 0.93 1.00 0.09 0.77 0.50 0.07 80.77 Cronstedtite 19.34 2.85 0.04 0.11 48.85 0.19 6.94 0.20 0.42 0.04 0.10 2.57 0.94 82.60 PCP I 10.81 3.21 0.03 0.18 53.18 0.14 5.54 0.14 0.50 0.05 0.08 9.03 1.60 84.50 PCP II, n = 5 5.33 5.05 0.02 0.29 53.27 0.14 4.32 2.14 0.82 0.04 1.19 5.43 1.11 79.17 PCP II, n = 4 2.90 1.93 0.01 0.03 67.25 0.12 3.59 0.06 0.40 0.02 0.03 3.95 0.57 80.87 T-I tochilinite 6.53 0.25 0.04 3.12 53.23 0.31 0.84 0.65 0.28 0.04 3.81 7.61 9.20 85.91 T-II tochilinite, n = 2 3.22 2.17 0.03 0.07 63.92 0.07 3.50 0.01 0.14 0.00 0.01 18.38 1.71 93.23 T-II tochilinite, n = 7 2.21 1.78 0.01 0.06 62.68 0.13 3.86 0.09 0.31 0.02 0.01 13.51 0.56 85.23 T-II tochilinite, n = 2 1.08 1.95 0.01 0.03 69.53 0.16 2.93 0.07 0.38 0.01 0.00 4.53 0.43 81.10 Magnetite 0.08 0.02 0.01 0.05 88.33 0.03 0.01 0.05 0.04 0.00 0.03 0.01 0.04 88.71 Serp matrix, n = 20 2.82 0.69 0.04 0.23 5.59 0.08 2.56 0.34 0.18 0.04 0.05 0.97 0.65 2.43 Serp chondrule, n = 6 4.11 2.04 0.27 0.22 9.65 0.14 2.47 1.46 0.26 0.06 0.74 0.33 0.04 5.96 Cron, n = 9 2.28 1.35 0.02 0.08 6.71 0.04 1.89 0.15 0.26 0.04 0.16 1.75 0.82 1.51 PCP I, n = 29 5.13 1.96 0.02 0.58 5.93 0.07 1.74 0.15 0.28 0.03 0.19 4.60 2.05 4.09 PCP II, n = 5 2.10 2.56 0.02 0.53 3.60 0.07 1.49 4.60 0.21 0.02 2.56 1.74 1.20 6.20 PCP, II, n = 4 2.16 0.67 0.01 0.02 2.86 0.05 0.82 0.03 0.09 0.02 0.03 0.90 0.26 1.29 T-I, n = 6 0.77 0.15 0.08 0.24 1.85 0.06 0.27 0.09 0.43 0.01 0.34 0.92 1.93 1.81 T II, n = 2 0.76 0.77 0.02 0.01 1.58 0.02 1.87 0.02 0.19 0.00 0.01 3.24 1.82 4.94 T-II, n = 7 1.05 0.41 0.01 0.05 2.87 0.06 1.25 0.06 0.06 0.01 0.01 1.68 0.40 3.46 T-II, n = 2 0.23 0.76 0.02 0.02 1.70 0.01 0.48 0.02 0.05 0.02 0.00 0.99 0.13 1.99 Mag, n = 4 0.04 0.02 0.01 0.03 0.30 0.04 0.01 0.04 0.04 0.00 0.03 0.00 0.04 0.33 202 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

2010-7 and -8). These grains consist of pyrrhotite rimmed with small pentlandite flames or granules in a more or less continuous peripheral corona (Fig. 8a and b). Semi-quantitative EDS analyses indicate metal/sulfur ratio (at.) close to 0.9 in hexagonal pyrrhotite (Hpo) and Fe/Ni at. = 1.2–1.4 in Fe-rich pentlandite (Pn, richer in Ni than matrix pentlandite). Some enclose spherical forsterite droplets which show Fe-enriched rims in contact with the pyrrhotite. Pt– Ir–(Os) alloys (2–3 lm) have been detected in two grains (former refractory metal nuggets (RMN) completely re- equilibrated with pyrrhotite for the most chalcophile PGEs Ru, Rh and Pd). Type II chondrules contain an Fe sulfide with 1% Ni and the same stoichiometry as Canyon Diablo troilite, and a similar sulfide is found in matrix; its survival suggests limited alteration. Metal in matrix is derived from Type I chondrules and may contain inclusions, e.g. of silica (Fig. 8d). It shows thin rims of tochilinite in the freshest zones (Fig. 8c), but is extensively replaced in the more altered regions. TEM images of the matrix reveal a complex fine-grained assemblage dominated by an amorphous phase, as in Y-791198 (Chizmadia and Brearley, 2008), and phyllosilicates (Fig. 9). Anhydrous silicates, sulfides and carbonaceous matter are also present. The most abundant phase displays a mottled contrast and a sponge-like structure with numerous nano-sized voids (Fig. 9a). Electron diffraction patterns show a diffuse ring indicative of amorphous or poorly crystalline material. Locally a fibrous morphology is clearly distinguishable from the matrix amorphous material, with a gradation in size suggesting a continuum from the sponge-like structure to a coarse-grained fibrous morphology. Only the coarse-grained areas (Fig. 9b) display a crystalline signal in electron diffraction, compatible with a serpentine structure. Fig. 7. (a) Gradational contact, vertical in the center of the figure, The FIB samples also contain relatively large and elon- in Paris section 2010-4 between fresh CM material on left (part of gated grains suggesting plate morphology. Their widths are region in Fig. 6), with Type I chondrules containing small fresh typically a hundred to several hundreds of nanometers and metal globules, and altered CM material on right, containing they can reach several micrometers in length. The electron abundant PCP (bright grey) and chondrules with altered metal. M, diffraction patterns (Fig. 10 a) and compositions indicate O and P are adjacent to metal, olivine and PCP, respectively. (b) the solid solution cronstedtite–thu¨ringite–amesite, where Sharp contact in Paris between moderately altered CM material 2+ 3+ and bright PCP-rich clast at top in 2010-1 (image width 2.6 mm). thuringite is (Fe ,Fe ,Mg,Al) 6(Si,Al)4O10 (O,OH) 8 and amesite is (Mg,Al) 3(Si,Al) 2O5(OH)4. They are iron-rich and are termed cronstedtite in the fol- contacts, e.g. in 2010-1 ( Fig. 7b), between moderately lowing. Discrete cronstedtite crystals are rarer and tiny altered matrix and angular bodies of BSE-bright material (100 nm) in chondrule rims but larger and polycrystalline rich in PCP. in adjacent matrix, as in Y-791198 ( Chizmadia and Brear- ley, 2008). Fig 10 a shows large laths of cronstedtite with 4.2. Matrix mineralogy fine interstratified tochilinite (blue arrows) and associated fibrous regions (red arrows). Cronstedtite and fibrous mate- 4.2.1. Matrix petrography rial are coarser in the matrix of more altered zones. Tochil- The matrix in Paris appears in BSE to consist mainly of inite seen in TEM is mainly a sulfur-rich fibrous phase, fine-grained mixtures of phyllosilicates (cronstedtite–chrys- varying in morphology and crystallinity in different regions. otile–greenalite solid solutions), tochilinite-like material The fibrous regions replacing metal in a chondrule are lo- and PCP (mixtures or intergrowths containing the above cally surrounded by an iron oxide phase, probably magne- phases). The size and abundance of PCP (bright grey in tite, with many crystal defects (Fig. 10 b). Fig. 7) varies with extent of alteration. Pyrrhotite, pent- Anhydrous silicates seen in TEM images consist of for- landite and Fe–Ni–P–S sulfides are observed in phyllosili- sterite and enstatite. Most of them have a rounded mor- cate-rich matrix. There are a few large irregularly shaped phology (Fig. 11 a) and are micron to sub-micron in size. isolated polycrystalline Fe–Ni sulfide grains, up to 80 lm Some of the enstatite grains are elongated along the [100] in maximum dimensions (3 grains observed in thin sections direction. Enstatite contains numerous planar defects R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 203

(ortho-clino inversion) on (100). In rare cases part of an the beam, and ATEM (Tables 8 and 10 ). Note that the enstatite grain is replaced by fibrous Fe-bearing serpentine. characterization of phases is not parallel at the different Other accessory minerals are Ca-phosphate, Cr–Al spinels scales of observation, as discussed in detail below. As the and hollow carbon globules (Fig. 11 b). Carbonates were amorphous phase (Fig. 8) does not have a distinctive stoi- not detected in these TEM-FIB sections but the more al- chiometry, we could not identify it by EMPA. Thus some, tered matrix domains contain 1–3% calcite, 50 lm in perhaps most, of our ‘serpentine’ is probably a mixture of diameter, particularly associated with altered CAI. Dolo- the amorphous phase with cronstedtite and/or other mite and more complex carbonates were not recognized. phases. The EMP data for serpentine and cronstedtite, plus Magnetite spherules, partly clustered into framboids intergrowths (PCP), tochilinite-like material and (rare) (Fig. 12 a), are present in very restricted domains, in which magnetite in matrix in Paris section 2010-04 and 2008- metal in contact with matrix is quite altered. Although LM are summarized in Table 8 and shown in Fig. 13 . We magnetite may be observed in significant amount in some found only slightly different compositions of phases in re- FIB sections (Fig. 10 b), BSE imaging suggests magnetite gions of different degrees of alteration (Section 6.1). The is a trace mineral. However, a precise estimate of a maxi- phyllosilicate endmember compositions are also plotted in mum magnetite content of 6.4 wt.% was made from mag- the Mg–Si–Fe quadrilateral (Fig. 13 a) and it is seen that netic measurements in §4.2. The magnetite in Fig. 12 a phyllosilicates run from about Chrysotile60 –Greenalite40 occurs in a clast 1.8 mm across sharply distinguished from to about Chrysotile 20 –Cronstedtite80 . surrounding matrix much more enriched in PCP-related TEM work performed on FIB sections of fine-grained material (Fig. 12 b). The magnetite is concentrated in clast rims on chondrules and chondrite matrix in moderately al- matrix between the fine-grained rims on Type I chondrules. tered regions of section SPSR-1 permitted precise identification of phases ( Table 9 and Fig. 14 a and b). Compositions 4.2.2. Matrix mineral compositions of the abundant amorphous phase and the finest fibers are The matrix was analyzed using EPMA, using Si, Fe, Mg comparable to the serpentine in Fig. 13 , but more Si-rich. and S concentrations to identify the main minerals under They cross the chrysotile–greenalite join and extend to the

Fig. 8. (a and b) Grains of pyrrhotite rimmed with small pentlandite granules in a more or less continuous peripheral corona in Paris matrix. (c) Metal grain (300 lm) in a more altered part of 2010-04, outside the right edge of Fig 7a, with a tochilinite rim 2–10 lm thick (d) Euhedral precipitates of silica exsolved in kamacite formed in Type I chondrules indicate mild reheating or slow cooling. 204 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 9. Bright field TEM images. (a) Amorphous silicate with a mottled contrast and a sponge-like structure containing numerous Fig. 10. Bright field TEM images in altered region. (a) Large laths nano-sized voids. The inset shows its electron diffraction pattern. of cronstedtite with interstratified tochilinite (blue arrows) and Wispy fibrous phyllosilicate and dark sulfide inclusions are also fibrous serpentine (red arrows). The inset shows an electron present. (b) Well developed fibrous serpentine region. diffraction pattern of the cronstedtite. (b) PCP on chondrule rim. Fibrous tochilinite at the bottom, presumably in replacement of an iron-sulfide grain. The S-rich fibrous region is surrounded by a En–Fs join (Fig. 14 a), suggesting a variable amount of Si thick iron-oxide layer (middle/top) at the interface with a silicate. Small grains at top right are colloidal silica polishing material. (For and a low fraction of crystalline phyllosilicate mixed in with interpretation of the references to color in this figure legend, the the amorphous material. They are similar to the average reader is referred to the web version of this article.) compositions of amorphous material in other primitive material (GEMS in interplanetary dust particles, Acfer 094, ALH 77307, MET 00246, QUE 99177 and Yamato to that of the most Si-poor tochilinite of Tomeoka and 791198: Brearley, 1993; Chizmadia and Brearley, 2008; Buseck (1985) and Palmer (2009). Many grains are richer in Abreu and Brearley, 2010; Keller and Messenger, 2012). Si and poorer in S than this, in part because of intergrowth Coarser-grained fibers are richer in Fe than the sponge- with cronstedtite. However there is a variety (tochilinite I in like areas and the ratio Si/(Mg + Fe) decreases with Table 8), which is S-poor but Cr–P–Ni-rich and Mg-poor, increasing fiber width displacing the compositions towards with affinities to the metal-derived Type I PCP of Tomeoka the chrysotile–cronstedtite join (Fig 14 a). They are similar and Buseck (1985). This composition may indicate the pres- in composition to the most Si-rich PCP-I of Fig. 13 . The ence of schreibersite or P-rich sulfides ( Palmer, 2009; Naza- gradation in size and composition suggests a close relation- rov et al., 2009 ). Tochilinite II (poor in Cr–P–Ni) varies ship between the amorphous (sponge-like) areas and the from close to the Fe tochilinite endmember to poor in both coarse grained fibrous areas. Ni-rich and Ni-poor sulfides, Si and S ( Fig. 13 b). Its composition approaches that of al- pyrrhotite and pentlandite, are widespread in the amor- tered metal and it may contain an Fe-rich phase. ATEM anal- phous material, with an average size of 50 nm ( Fig. 8a). Sul- yses of tochilinite are shown in Table 9 and Fig. 14 . Where it is fide grains are not present in the fibrous areas but S is still very crystalline, it contains no Si ( Fig. 14 a). Where it is finely detectable in abundance, suggesting that it is located in the fibrous, there is Si and also abundant Ni. The high propor- interstratified layers of phyllosilicates or expressed as nano- tion of Si likely indicates a mixture with phyllosilicates. meter-scale tochilinite grains. Thus the coarse fibrous mate- Most PCP intergrowth compositions fall between cron- rial might be considered a nano-PCP. Platy/elongated stedtite and tochilinite (Fig. 13 a and b). We distinguish crystals (cronstedtite) are more Fe-rich and Al-rich (Table 9) them in Table 8 as PCP I and they fit the PCP SiO 2 criterion than the amorphous material. They generally have less S of Rubin et al. (2007) . Blanchard et al. (2011) reported sim- than the cronstedtite of Fig. 13 , but some regions are int- ilar compositions of PCP in Paris sections 2010-5 and -7. erstratified with thin layers of tochilinite. However we also observe PCP-like material, PCP II, which We analyzed several varieties of tochilinite-like material is depleted in both S and Si, plotting on the Fe-rich side of by EMP ( Table 8), including some with compositions close the tochilinite–cronstedtite tie line. This material R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 205

Fig. 11. Bright field TEM images. (a) Large grain of forsterite in fine grained matrix. (b) Hollow carbon nanoglobules (arrows) in matrix. approaches altered metal in composition ( Fig. 13 b) and is mixtures of tochilinite, phyllosilicate and some Fe-rich phase related to alteration of metal, probably magnetite, as seen in association with tochilinite fibers in Fig. 10 b.

4.3. Magnetic properties and mineralogy

Elmaleh et al. (2011, 2012) compared the low-temperate (2–300 K) magnetic properties of single crystals of cronstedtite, and of CM2 chondrites including Paris. The temperature of the magnetic susceptibility peak (Tp, 7 K) of cronstedtite falls with increasing Mg and increases in inter- acting fine grained particles. Elmaleh et al. (2012) observe a Fig. 12. (a) Detail of magnetite framboids (b) Fragment 1.8 mm decrease of Tp among the CM chondrites with increased across containing magnetite spherules, and with a metal-rich alteration based on the mineralogical alteration index chondrule near its center, is in sharp contact with normal altered matrix with PCP at top and right. (Browning et al., 1996 ) and the FeO/SiO 2 content of PCPs (Rubin et al., 2007 ). Metal-rich and metal-poor areas in the Paris chondrite show contrasting low-T magnetic signa- between 1 and 2 g give an average log v = 4.33 ± 0.23, and tures, with a shift from high Tp in the less altered material the five measurements on masses below 1 g give an average to a lower one like Murray’s. The evolution of the low-T logv = 4.47 ± 0.40. magnetic signature is interpreted as due to changes in cron- Monitoring of magnetic susceptibility at low tempera- stedtite’s chemical composition (Elmaleh et al., 2012 ). ture performed on six samples with mass in the 20–100 mg We measured the magnetic susceptibility of Paris on two range shows the existence of a weak Verwey transition in large samples, both with masses of 17.1 g, as log v = 4.310 all samples but one, indicating the presence of stoichiome- and 4.50 (v2 in 10 À10 m3 kg À1). The mean log v = 4.45 is in tric magnetite with variable abundance at this scale the very upper range for CM chondrites, that have mean (Fig. 15 ). This is confirmed by cooling of room temperature susceptibility logv = 3.100 ± 0.43, computed from 52 CM saturation isothermal remanent magnetization (SIRM), as chondrites (Rochette et al., 2008 ). As shown by measure- well as heating of low temperature SIRM that also shows ments taken on 12 samples with mass ranging from 23 mg the Verwey transition (Fig. 16 ). It is noteworthy that the to 17.1 g, magnetic susceptibility is homogeneous down to only sample that does not display a Verwey transition is the scale of 2 g. Below 2 g, there is an increasing scatter the sample that was specifically sampled in the less aqueous- between measurements: the four measurements on masses ly altered lithology of the meteorite, whereas the sample that 206 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 13. EMPA data (at.%) of 300 nm squares of Paris matrix Fig. 14. ATEM data (at.%) for Paris matrix alteration materials, alteration materials, named for the dominant mineral present. The corrected for the presence of sulfide inclusions. Amorphous refers Fe concentration is total Fe in the spot analyzed (silicate ± sulfide). to mottled and porous material, and finely fibrous material. End- Asterisks are end-member chrysotile, greenalite and cronstedtite member chrysotile, greenalite and cronstedtite compositions are compositions; stars are the most Si-poor Type I and II tochilinite of also shown, plus the most Si-poor Type I and II tochilinite of Tomeoka and Buseck (1985). Tomeoka and Buseck (1985). shows the strongest Verwey transition is the one that was specifically sampled in the more aqueously altered lithology. metal, Paris contains a mineral with high coercivity, High temperature measurements of magnetic suscepti- suggesting the presence of sulfides. bility under argon atmosphere (Fig. 17 a) show Curie tem- Thermal demagnetization of SIRM shows major peratures at 581 °C, corresponding to magnetite. The unblocking temperatures in the 100–200 °C range, and an- residual magnetic susceptibility above 585 °C is due to the other minor peak in unblocking around 550–580 °C presence of metallic FeNi. However, it is not possible to (Fig. 17 b). The latter peak corresponds to magnetite. The determine the Ni content of this phase because it is being low temperature peak can be attributed to sulfides. It is destroyed during the heating experiment, as indicated by noteworthy that the FeNi phase that is visible in experi- the irreversible curve upon cooling, with formation of ments involving magnetic susceptibility is not detectable new magnetite (Fig. 17 a). The S ratio of 0.83 ± 0.05 in experiments using magnetic remanence because FeNi (n = 4) indicates that in addition to magnetite and FeNi grains are typically a very poor carrier of remanence but have large magnetic susceptibility. The exact opposite is true for sulfides. The nature of these sulfides cannot be Table 9 determined from the magnetic experiments. However, it is Averages (at.%) and standard deviations for TEM analyses of noteworthy that after heating to 400 °C, SIRM is basically matrix phases. Data are normalized to a 100% anhydrous basis. unchanged, but has a rather different unblocking tempera- Si Al Fe Mg S Ni ture spectrum, with a shift of the low temperature peak Amorphous / 16.2 0.99 12.7 9.3 1.73 0.54 from 100–200 °C to 300–350 °C. This suggests growth of Cronstedtite 12.5 2.77 20.7 6.54 0.64 0.27 pyrrhotite, possibly from S in PCP, or possibly due to Coarse fibers 10.6 1.01 22.0 7.08 4.29 0.40 shrinkage of the pentlandite stability field. Tochilinite 1.65 0.68 30.23 1.16 19.99 2.07 Determining the modal abundances of each of these Amorph, n = 21 1.7 0.16 2.4 1.3 0.56 0.18 phases is not straightforward as their magnetic properties Cron, n = 31 0.9 0.60 2.8 2.31 0.32 0.13 strongly depend on grain size. However, the average satura- Coarse, n = 17 1.3 0.13 1.9 1.15 1.85 0.18 2 À1 tion magnetization (M S) of 5.10 Am kg for Paris, com- Toch, n = 13 0.99 0.21 1.78 0.53 2.15 0.43 pared to the Ms for metal (218 Am 2 kg À1) and magnetite R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 207

Table 10 Compositions of refractory metal nuggets (wt.%, limit of detection 0.5%). Re Os W Ir Mo Ru Pt Rh 4 (2) 8.39 3.4 7.03 42.16 26.88 12.17 7 (1) 18.96 4.54 5.92 38.83 20.5 8.43 1.67 10.9 12.07 22.56 16.21 28.9 4.34 7 (1) (7-III-2) 10.25 49.64 7 (IV I) (7-5) 11.29 14.71 74 7 (IV-2) 13 15.6 71.4 8-I-1 (8 (1)) 18.09 1.42 16.2 13.5 19.41 11.86 3.03 8(II-1) largest RMN 1.52 21.93 1.95 21.55 4.43 21.79 0.11 2.09 (8 (3)) 26.35 3.84 26.89 2.89 20.8 1.47 8-II-2 (8-4) 0.74 23.02 1.86 20.76 9.88 13.58 1.28 1.35 8-VI-1 (8-4’) 8.79 1.14 9.43 15.29 14.05 25.73 1.19 8-XI-1 (8-2’) 0.93 26.29 3.53 24.53 11.82 11.84 6.5 0.84 6-I-1 21.46 0.76 25.35 18.59 29.74 1.19 25.35 27.46 23.48 20.31 29.46 27.11 13.52 21.6 1.52 6-I-2 1.47 26.43 2.28 25.4 17.42 24.33 1.74 0.92 0.53 24.44 2.76 25.34 17.06 23.9 4.58 1.42 0.57 15.24 3.07 20.82 25.92 22.95 8.27 3.13 23.66 1.7 24.12 16.83 27.3 4.3 1.95 6-I-3 0.87 27.34 2.66 27.81 15.25 23.3 0.53 2.22 23.38 27.9 17.71 30.54 19.27 3.38 23.01 17.71 26.75 4.3 11.03 20.43 30.05 25.87 6.87 3.46 6-VII-1 4.54 0.79 6.43 13.31 15.9 32.88 2.54 2.5.2 2.96 47.36 6.5 37.22 3.32 0.54 0.001 0.001 1.42 34.73 3.7 32.16 8.44 16.55 0.001 0.81 41.84 4.45 37.17 1.28 2.16 0.001 2.5-4 2.24 40.27 3.96 41.15 8.6 2.31 0.001 0.001

(102 Am 2 kg À1) indicates a maximum magnetite content of 6.4 wt.%, and a maximum metal content of 2.7 wt.%. The amount of pyrrhotite cannot be assessed in the same way because it has much lower Ms and susceptibility. But in view of the fraction of the SIRM that is demagnetized below 350 °C (Curie temperature of pyrrhotite), and using a hypothetical MRS /M S = 0.3 for pyrrhotite, a maximum content of about 2 wt.% pyrrhotite can be estimated. The hysteresis parameters, with BCR /B C = 11.6 and M RS / À2 MS = 2.73 Â 10 , indicate an overall multidomain behav-

Fig. 15. Magnetic susceptibility as a function of temperature for three representative low temperature experiments: regular sample (solid line), aqueously altered sample (thick solid line), and non Fig. 16. (a) Low temperature cycling of room temperature SIRM aqueously altered sample (dotted line). The Verwey transition for a 54 mg sample of Paris. (b) Thermal evolution of low temperature at 120 K is indicated by a vertical line. temperature SIRM for a 155 mg sample of Paris. 208 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 18. Hysteresis properties of thirteen Paris samples (crosses), and thirteen samples from eight other meteorite falls (circles). The limits of the pseudo-single domain (PSD) and multidomain (MD) behaviors are indicated.

an intrinsic Ms that is about only 2.4 times higher than magnetite), but also indicates a larger modal abundance of metal + magnetite in Paris than in other CM chondrites. Compared to most other CM falls, Paris has hysteresis properties that are in the pseudo-single domain area (Fig. 18 ). Rather than a different grain size for magnetic minerals, this is indicative of a larger proportion in Paris of metallic FeNi vs. magnetite and sulfides. It is noteworthy Fig. 17. (a) Magnetic susceptibility as a function of temperature that despite having a larger M S, the saturation remanent 2 for a 11 mg sample of Paris. (b) Thermal demagnetization of SIRM magnetization of Paris (M RS = 0.13 ± 0.02 Am /kg, mea- of two Paris samples with masses 94 mg (circles) and 40 mg (open sured on 13 samples), that is dominated by sulfides and boxes). The smaller sample was resaturated after the 400 °C heating magnetite to a lesser extent, is equal to the mean value 2 step, and thermally demagnetized again (solid boxes). for other CM falls (M RS = 0.12 ± 0.07 Am /kg, measured on 13 meteorites from 8 different falls, our unpublished data). This suggests that the modal abundance of sulfides ior. This multidomain behavior is typical of multidomain in Paris is about the same as in other CM falls. FeNi grains (i.e. with grain size >100 nm). The grain size To summarize, Paris has a magnetic mineralogy that is of the other magnetic phases (magnetite and pyrrhotite) is homogeneous down to the scale of about 2 g. It contains difficult to assess precisely because of mixture of three fer- magnetite, FeNi metal, as well as poorly crystalline sulfides. romagnetic minerals with contrasted magnetic properties. The less aqueously altered lithology of Paris does not con- But by analogy with other CM chondrites it is likely in tain magnetite. The most striking difference with other the pseudo-single domain area, indicating grain size in the CM chondrites is the significantly higher amount of FeNi 20–280 nm range for magnetite ( Dunlop and O¨ zdemir, metal in Paris, estimated at 2.7 wt.% in the least aqueously 1997), and below 1.5 lm for pyrrhotite ( Menyeh and altered lithology. This higher amount cannot be entirely ex- O’Reilly, 1991), though larger grains could also be present. plained by a lower degree of aqueous alteration of metal in Magnetite of this size was not observed in our BSE images, Paris, and also implies a higher total modal abundance of but was found with tochilinite in some FIB sections of al- metal + magnetite in Paris than in other CM chondrites, tered metal. probably related to the higher ratio of chondrules to matrix. The hysteresis properties of Paris are significantly different from other CM. Saturation magnetization of Paris 4.4. Carbon (5.85 Am 2/kg) is the highest among all CM falls, and is more than five times higher than the mean MS for falls In our TEM work, we observed that organic globules 2 (M S = 0.107 ± 0.46 Am /kg measured on six CM falls are common in the Paris matrix ( Fig. 10 b). Other authors (our unpublished data). This cannot be entirely attributed have reported on organic material in Paris. A separate to a larger proportion of metal vs. magnetite (metal has section (SPSR-1), prepared for Remusat et al. (2011) with- R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 209 out using epoxy, was studied by NanoSIMS using rastering other Paris CAI. The vuggy inclusion 285 (thin section of a focused Cs + beam to generate secondary ion images 2008 LM) consists almost entirely of hibonite laths of H À, D À, 12 CÀ, 16 OÀ, 26 CN À, 28 Si À and 32 SÀ in micron (Fig. 19 c), but hibonite is coated with spinel, which is fol- sized organic particles. C/H ratios in matrix IOM appear lowed in turn by fassaite, in interior cavities, ( Fig. 19 d). significantly lower than in other CM chondrites ( Remusat Kamacite (3–4% Ni) is present in a few inclusions, mantled et al., 2011 ). The D/H ratio in more and less altered regions like spinel by diopside or occurring between fassaite and the is indistinguishable. A few D-rich hot spots are observed, outer mantle of forsterite. with a maximum at dD = 7950 ± 330 &, suggest that aque- Other Paris sections contain similar RI but we observed ous alteration on the CM parent body does not induce a one distinctive and very large (about 800 lm) zoned inclu- significant modification of the composition of organic par- sion in section 2010-05. Perovskite appears to have been ticles (Remusat et al., 2011 ). the first phase to form in this CAI, which is spinel-rich Crushed fragments weighing a few mg were used by throughout. It has a central region where perovskite and Merouane et al. (2012) for mid and far IR and Raman mi- lathy or ragged spinel are encased in massive gehlenite cro-spectroscopy. Some IR spectra of aromatic-rich mi- (Fig. 19 e). This region is surrounded by a highly porous re- cron-sized fragments of Paris resemble those of other CM gion lacking gehlenite but with the same lathy spinel chondrites, but others match spectra in the 3.4 and 6 lm re- (Fig. 19 f), which grades outwards into a region where the gions for organics from the diffuse interstellar medium matrix is a silicate rich in O, Al and Na, i.e. with a zeo- (Merouane et al., 2012 ). They find a CH 2/CH 3 ratio of lite-like composition, and with minor S and Cl. This CAI 2.2 ± 0.2 in agreement with the value for ISM objects. thus appears to record the partial dissolution and replace- The good match for the spectra suggests that Paris may ment of melilite. In the outer part of the altered mantle, have preserved some organic matter of interstellar origin. the spinel is more equigranular and is partly mantled by 1 lm of a Ca–Ti–Al–Mg-rich silicate, probably fassaitic 5. PETROLOGY OF HIGH TEMPERATURE pyroxene. The whole is encased in a rim of diopside/fassaite COMPONENTS 10–25 lm thick, locally forming extensions containing spinel-rich nodules, in turn surrounded by a serpentine-rich 5.1. Refractory Inclusions mantle. The simplest CAI contain spinel mantled by diopside or Ninety five refractory inclusions (RI), of size 100– fassaite, with an outer rim of forsterite in many cases 200 lm, were identified using BSE in the thin section 2008 (Fig. 20 a) and many of them show no alteration. The main LM, making up 0.4% by area of the 4.5 cm 2 section. alteration seen in CAI is formation of cronstedtite between Phases in CAI were identified from EDS spectra and spinel and pyroxene (Fig. 20 b). We note that this horizon semi-quantitative analyses on the Tescan VEGA SEM suf- corresponds to gehlenite, where present, and which is re- ficient to define stoichiometry. The CAI are mainly fine- placed by void space and a zeolite-like Na–Al silicate in grained spinel-rich inclusions. The occurrence of forsterite the large CAI 2010-05-01. Calcite and an Mg-rich, Al-bear- rims on some CAI and of small nodules of CAI inside some ing phase with elevated oxygen, probably a hydroxide, are AOA demonstrates their kinship. The AOA are porous and found in heavily altered inclusions. frequently fractured, with most grains 1–5 lm long Magnetic and density separates were prepared after (Fig. 19 a). In addition to forsterite, they may contain kama- using the freeze–thaw method on Paris. Hibonite-rich cite and/or orthopyroxene. Their CAI minerals include spi- objects were handpicked at IPGP from these fractions for nel and fassaite, while alteration is mainly to cronstedtite SIMS analysis. Mg isotopes of 14 spinel-hibonite spherules with some tochilinite. (SHIBs) and one platy hibonite inclusion (PLACs), 30– Spinel is a major mineral in CAI and it occurs as nod- 40 lm in size in most cases, were analyzed by Liu et al. ules, either massive or containing more refractory phases. (2012), along with samples from the Murchison CM chon- The sequence hibonite, perovskite and grossite is observed drite. Mg isotopic data for SHIBs show heterogeneity and encased in spinel in inclusion 268 (thin section 2008 LM), suggest formation with gradually increasing 26 Al/27 Al to and spinel is mantled by gehlenite, followed by anorthite, roughly the canonical level in the solar nebula. The maxi- followed by diopside ( Fig. 19 b). Hibonite is zoned to Ti– mum value for 26 Al/27 Al, 6 Â 10 À5 suggests that SHIBs pre- Mg-rich rims. The sequence spinel followed by gehlenite dated CAIs (Liu et al., 2012 ). Platy hibonite has no clear is common in CM chondrites, though not predicted by excess in 26 Mg and may have formed before the injection the equilibrium condensation sequence (Simon et al., of 26 Al into the disk. Liu et al. (2012) concluded that 2006). Rarely there are small inclusions of gehlenite in spi- hibonite formed in the first few thousand of years of Solar nel in inclusion 268. These are possibly protrusions of the System history. mantle into the plane of the section, but they might also Refractory Metal Nuggets (RMN) have been identified be evidence that some gehlenite managed to nucleate on in the BSE mode (Tescan SEM, MNHN) during systematic and replace spinel, as in rare cases reported by Simon traversing of polished thin sections at magnification 500 Â. et al. (2006) . RMN occur exclusively inside CAI or attached to disrupted Other CAI contain fewer phases, but with the same for- CAI minerals. RMN-bearing CAI (13 in all) were investi- mation sequence. Perovskite or hibonite may be the only gated with the SEM in five polished sections and were phase earlier than spinel, or there may be no phases earlier found in all of them. Their abundance, size and composi- than spinel. Gehlenite and anorthite are not common in tions vary as a function of host CAI minerals. RMN in 210 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 19. BSE images of refractory inclusions in Paris 2008 LM (a–d), and Paris 2010-05 (e–f). (a) amoeboid olivine aggregate containing forsterite and kamacite. (b) CAI 268 contains nodules of spinel (darkest grey, except for black pores); the sequence hibonite (medium grey needles), perovskite (white) and grossite (medium grey) is mantled by the spinel; the sequence gehlenite (light grey), followed by anorthite (medium grey), followed by diopside/fassaite (medium grey) mantles the spinel nodules; the large kamacite grain (white) is also embedded in the diopside. (c) CAI 285 consists almost entirely of hibonite laths, with spinel on the exterior. (d) CAI 285 is vuggy (detail of (c)), with hibonite laths coated by spinel (medium dark grey), followed in turn by fassaite (very light grey), in interior cavities. (e) CAI 2010-05-01 is a zoned spinel-rich inclusion about 800 lm long, in which perovskite was the ﬁrst phase to form. In the center perovskite and lathy or ragged spinel are encased in massive gehlenite. The surrounding region is porous with no gehlenite and grades outwards into a zeolite-like region encased in a rim of diopside/fassaite 10–25 lm thick, surrounded by a serpentine-rich mantle. (f) CAI 2010-05-01: close-up of transition from gehlenite-rich to porous to zeolite-like, with spinel in all three. Symbols: Ol forsterite, Px diopside or fassaite, K kamacite, Hi hibonite, Pv perovskite, Gr grossite, Sp spinel, Ge gehlenite, An anorthite, Ze zeolite-like phase, Se serpentine. R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 211

the two most refractory CAIs (containing hibonite, Paris sub-sample 2.5, and Y-rich perovskite + Al spinel, Paris 2010-6) are very small-size inclusions (a few tens of nanometers to less than 500 nm). They form clusters of several inclusions inside hibonite. The largest RMN-bearing CAI yet discovered in Paris meteorite appears to be dotted with tens of evenly distributed, small (100 nm to less than 1 lm in maximum dimension) individual RMN which preferentially occur in the Y-rich perovskite phase ( Fig. 20 c). RMN in spinel-perovskite (Y-free) CAI are isolated grains (1–3 per CAI) preferentially included in spinel. These are larger (ca. 500 nm to 1 lm) RMN cuboids, with cubo-octa- hedral or hexagonal cross-section. Isolated RMN grains have also been identified in Al-spinel + Ca–Al pyroxene CAIs and Al-spinel coexisting with forsterite. The largest RMN (2.5 lm) is included in forsterite. Semiquantitative chemical compositions were determined with the SEM on the largest RMN particles using a standardless EDX procedure ( Fig. 21 , Table 10 ). The most refractory alloys occur in the hibonite grain (Paris 4039 2.5), with compositions dominated by Os (47.4) and Ir (37.2), coupled with the highest Re (3.0) and W contents (6.5 wt.%). The RMN analyzed in perovskite + Al-spinel CAI are Os–Ir–Mo–Ru alloys containing detectable amounts of Re (up to 1.5 wt.%) and W (up to 3.3 wt.%), coupled with low contents of the less refractory PGE (<8 wt.% Pt). Those hosted in spinel (i.e. remote from perovskite) are enriched in Pt (up to 29 wt.%) and Rh. RMN enclosed in forsterite are Pt–Fe-rich alloys. Refractory metal nuggets akin to those reported from the Paris meteorite were extensively documented in CM2 (Murchison; Berg et al., 2009; Schwander et al., 2011, 2012; Harries et al., 2012; Croat et al., 2013 ), CV3 (Allende, Palme and Wlotzka, 1976; Blander et al., 1980; Schwander et al., 2013 ; NWA 1934, Leoville; Schwander et al., 2013 ), and very recently, in a quite different petrological setting however (low-density fraction in presolar graphite grains, CI (Orgueil-like) chondrites, Croat et al., 2013 ). Enriched in refractory siderophile elements by 5–7 orders of magnitude over the CI abundances, RMN match compositions of early nebular condensates ( Palme and Wlotzka, 1976); their compositions are reproduced by chemical equilibrium calculations of high temperature condensation into hcp single-phase alloys (Harries et al., 2012 ) from a cooling gas of solar composition. In Murchison, RMN were thought to have been once residing in CAIs, but relationships with CAI’s mineral assemblages were lost because all studies were performed on acid-resistant concentrates ( Berg et al., 2009; Harries et al., 2012; Schwander et al., 2012 ). Murch- ison and Paris show strikingly similar features, regarding size (20 nm–1.3 lm), the regular shape of RMN particles and their compositional range (7–57 wt.% Os vs. 8– Fig. 20. BSE images of CAI. (a) CAI 257 is a typical fine-grained 47 wt.%), in spite of the higher number of analysed grains spinel-rich CAI. There are diopside rims around spinel, and an incomplete fine rim of forsterite (Paris 2008 LM). (b) Altered fine- in Murchison (>100 vs. 25 for in Paris). The polyhedral sur- grained spinel-rich CAI 258. Perovskite inclusions in spinel, rim of face faces indicate growth in a gas phase. Empirical calcu- cronstedtite (bright grey) and incomplete outer rim of fassaite (Paris lation from condensation models indicates a wide range 2008 LM). (c) The largest RMN-bearing CAI with tens of evenly of condensation temperatures for such alloy compositions, distributed submicron RMN which occur preferentially in Y-rich from ca. 1600 K to less than 1450 K, similar in both mete- perovskite phase (Paris 2010-6). Symbols: Ol forsterite, Px diopside orites. Berg et al. (2009) for Murchison assumed the highest or fassaite, K kamacite, Pv perovskite, Sp spinel, Cr cronstedtite. condensation T to be similar to that of perovskite; in Paris 212 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 21. Chondrite-normalized compositions of RMN in Paris CAI. Refractory siderophile elements decrease following the observed condensation sequence of the host oxide and silicate phases.

Fig. 22. BSE images of metal-rich Type I chondrules in Paris 2010-04, -06 and -07. (a) Chondrule 2010-07-6 is a Type IABm with kamacite

(5.4% Ni, 0.5% P, 0.8% Cr), olivine Fa 0.6–0.9 and pyroxene Fs 0.9–1.3Wo 1.0–2.3; the mesostasis contains diopside. (b) 2010-07-11 is a Type IABm chondrule with Fo 1.0–2.0. In the central porphyritic zone, metal replaced mainly by tochilinite is in contact with mesostasis altered mainly to cronstedtite. Metal inside olivine is fresh. Pyroxene in the poikilitic mantle is etched. (c) 2010-07-2 is a Type IBm chondrule with kamacite

(5.4% Ni; 0.5% P, 0.6% Cr) and pyroxene En 1.0–6.7Wo 0.7–1.0 plus minor diopside. (d) 2010-04-6, containing metal droplets up to 35 lm in diameter with 2.95 wt.% Si. The growths on the surface of the chondrule contain cronstedtite and tochilinite. Symbols: OL forsterite, PX diopside, K kamacite, TO tochilinite, CR cronstedtite. R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 213

(Fig. 21 ), the most refractory RMN are hosted in hibonite. Moreover, Fig. 21 shows that abundances of the most refractory PGE decrease following the observed condensation of the host oxide and silicate phases of RMN. This relationship adds further evidence for a nebular condensate origin; RMN served as nucleation centres for host silicates and oxides. RMN in the Paris chondrite escaped significant alteration prior to incorporation into their parent body. There is no sign of significant W and Mo depletion in Paris and Murchison RMNs, in contrast to Allende CAIs that show strong depletion in those easily oxidized and sulfurized elements (e.g. Palme et al., 1994; Schwander et al., 2013 ). Alteration by parent body processes is limited to a few Pt–Ir-rich alloys redistributed in Fe–Ni sulfides; similar assemblages were documented in R chondrites ( Schulze, 2007). However, Paris is devoid of the As-, Sn-bearing noble metal minerals that testify to extensive parent body alteration in R chondrites.

5.2. Type I chondrules

Type I chondrules in Paris are approximately 48% Type IA (PO), 39% Type IAB (POP, mostly metal-rich) and 13% type IB (PP). Metal-rich Type IAB ( Fig. 22 a) and metal- rich Type IB ( Fig. 22 c) are conspicuous in the freshest zones, but can also be recognized easily when the metal is pseudomorphed by alteration minerals ( Fig. 22 b). Mesosta- sis is not very abundant so that the textures are more granular than strictly porphyritic, except in chondrule centers where mesostasis is concentrated. Dusty (metal-bearing) relict grains were not observed. Glass is replaced by phyllosilicates, whereas olivine, pyroxene and metal are generally unaltered. In the more altered regions, metal in Type I chondrules may be replaced by tochilinite and re- Fig. 23. Histograms of olivine and pyroxene compositions in Type lated minerals, especially if in contact with mesostasis I chondrules. (Fig. 25 c). Histograms of olivine and pyroxene compositions in 54 Type I chondrules are given in Fig. 23 . Mean compositions of olivine and pyroxene are Fa 0.9, s.d. 0.7 and Fs 2.3Wo 1.4, s.d. 1.7 and 1.3, respectively. Diopside or fassaite occurs in the mesostases of some of these chondrules. Diopside in some cases contains exsolved pigeonite, with exsolved lamellae on (001) 10 nm wide. Ca decreases and Cr increases sharply in olivine going from Fa 0 to Fa 1 (Fig. 24 ). Metal in Type I chondrules and isolated globules in matrix is on average chondritic in Ni/Co and in Ni/P, with a small range of Ni, mostly 4 to 8 wt.% but some up to 12 wt.%, and scatter of Co (0–0.7 wt.%) and P ( Fig. 25 a and b) related to exsolution. Subsets of the metal analyses have very high and very low P concentrations, consistent Fig. 24. Variation of Ca and Cr concentrations of forsteritic with the observed exsolution of P-rich phases. olivine in Type I chondrules. Metal grains in matrix and Type I chondrules may contain inclusions of other phases. Those with low concentrations of dissolved Si contain small inclusions ( Fig. 8d), and silica glass inclusions were identified in metal by Caillet mainly tabular 100 nm–1 lm inclusions, of an SiO 2 poly- Komorowski et al. (2011) . The crystallographic alignment morph with a common alignment ( Bourot-Denise et al., (Fig. 19 d) indicates an origin by exsolution as negative crys- 2010). These fine inclusions tend to be near the edge of tals. Inclusions of daubreélite, Cr phosphide, chromite, the grains and not in their centers, possibly due to zoning schreibersite and troilite were also observed in metal in Si or to limited diffusion of O into the metal. Cristobalite (Bourot-Denise et al., 2010 ). Assemblages of several of 214 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Fig. 27. Histogram of compositions of olivine in Type II chondrules.

these phases were found near the exteriors of the same metal grains. However, unlike the other elements in metal, Cr concentrations are mostly highly depleted relative to Ni ( Fig. 25 c), Co and P. A correlation of Fa vs. “Cr 2O3” in the most forsteritic olivine (Fa0.2 to 1) and a negative correlation between Cr in metal and Cr 2O3 in olivine indicates that most of this variation is due to oxidation: metal and silicate equilibrated during crystallization of chondrules ( Zanda et al., 1994 ). However the lowest Cr concentrations (<1% Cr) are associated with very low P concentrations, and are due to exclusion of Cr phosphide precipitates from the analyzed regions. The calculated fO 2 for the OIP buffer assemblage and for Cr partitioning between metal and olivine or pyroxene ( Kring, 1986; Zanda et al., 1994 ) is 10 À11 at 1600 °C (IW-3). Si concentrations, with exceptions, are close to EMP detection limits (Fig. 26 ). We found two chondrules in Fig. 25. Metal compositions: Ni vs. (a) Co (b) P (c) Cr. Metal in which metal contains 2.95% Si, s.d. 0.19, and 0.64% Si, Type I chondrules and isolated globules in matrix is generally s.d. 0.11. Some isolated metal droplets in matrix, like grain chondritic but Cr varies from chondritic to highly depleted due to 3 in Fig. 26 , also have significant Si. In Fig. 26 , we use dif- oxidation in the chondrules. ferent data symbols for individual metal grains in the two chondrules. Different metal grains within a given chondrule may have similar yet significantly different Si concentrations (Fig. 26 ). The moderately Si-rich metal in chondrule 2010-04-09 occurs as 50–100 lm droplets. The exception- ally Si-rich metal in chondrule 2010-04-06 occurs as numerous small spheres, from 35 to a few microns in diameter (Fig 22 d), with irregular boundaries as if accreting nanospherules, unlike the irregular blebs in most Type I chondrules (Fig. 22 ). The chondrule resembles a quenched spherule with an emulsion of two liquids. The interior metal is fresh, except very close to the edge, though there are three tochilinite–cronstedtite growths with some minute P/Cr inclusions, and also small magnetite growths on the Fig. 26. Metal in Type I chondrules and in matrix in section 2010- chondrule surface. The Si-rich metal is associated with very 04 is generally poor in dissolved Si, but is Si-rich in chondrules-9 Cr-poor silicates, yielding an fO 2 estimate from Cr parti- À12 and -6, and in matrix grain 3. Different symbols are used to tioning of 10 at 1600 °C (IW-4). These silicates are distinguish different metal grains in the same chondrule. Chondrule not more magnesian than usual, however, indicating some 2010-04-6 is shown in Fig. 21 . redox disequilibrium. R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 215

Fig. 29. FeO–MnO relations in olivine in chondrules in Paris, other CM chondrites and Semarkona. Paris Type II chondrules are typical of those in carbonaceous chondrites and unlike those in UOC. Relict grains overlap the ﬁelds between Type I and Type II chondrule olivine compositions.

Fig. 30. Histogram of wt.% Cr 2O3 in olivine of Type II chondrules and fragments has the form of those in 3.0 chondrites ( Grossman and Brearley, 2005).

5.3. Type II chondrules

Type II chondrules and fragments make up 1–2% of Paris, and there are many tiny ferroan olivine crystal fragments. They are dominantly Type IIA (PO) chondrules, but several kinds of pyroxene-bearing Type II chondrules are present. The range of melt-grown olivine compositions is Fa 7 to Fa 76 (Fig. 27 ). We have included a chondrule with Type IIA PO texture and Fa 7–8 in this histogram, though its FeO/MnO ratio shows it is intermediate between Type I and Type II olivine ( Hewins and Zanda, 2012). Porphy- ritic olivine texture is illustrated in Fig. 28 a, and 100 lm relict forsterite grains are common in such IIA chondrules Fig. 28. BSE images of Type II chondrules in section 2010-04. (a) (Fig. 28 b). Preliminary accounts of these chondrules have Type IIA PO chondrule-17 with olivine Fa 26–33. (b) Type IIA been given in Hewins et al. (2011) and Hewins and Zanda chondrule-9 with olivine Fa 29–40, numerous relict olivine grains Fa 1–11 and one relict spinel grain. (c) Type II(A)B chondrule-16 (2011). Sulﬁde in Type II chondrules has the stoichiometry with pyroxene Fs 40–63Wo 1–6 and olivine (white) Fa 74–75. OL olivine, of Canyon Diablo troilite, and up to 1% Ni. Pentlandite PX pyroxene, FO relict forsterite and SP relict spinel. was also observed where sulﬁde is very abundant. 216 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

We made EMP analyses of 9 chondrules or chondrule fragments containing pyroxene. There are two IIAB and two IIB chondrules; two II(A)B, with minor highly ferroan olivine; and three IIA(C) with augite as the only pyroxene. In the IIAB chondrules olivine compositions are Fa 37–61 and orthopyroxene En 36–42Wo 0–2. One II(A)B has olivine Fa 32–44 and orthopyroxene En 34–38Wo 1. The II(A)B illustrated in Fig. 28 c contains olivine Fa 74 and highly zoned orthopyroxene–pigeonite En 40–63Wo 1–6. The augite in the IIA(C) chondrules is Fs 9–27 Wo 36–41. Type II chondrule olivine in Paris and other CMs (Mur- chison, Mighei and Cold Bokkeveld) show the same FeO– MnO correlation (Fig. 29 ), with FeO/MnO 100, which is similar to that for CO chondrites ( Berlin et al., 2011 ), CV ( Rubin and Wasson, 1987), Tagish Lake ( Simon and Grossman, 2003; Russell et al., 2010 ), Acfer 094 (our data) and Wild 2 ( Frank et al., 2012 ). This olivine has significantly higher Fe/Mn than that for chondrules from ordinary chondrites (Berlin et al., 2011; Hewins and Zanda, 2012). However for Paris there exist distinct outliers from this distribution, reflecting reservoirs different in Fe/Mn. Ferroan relict grains connect the fields for Type II and Type I chondrules ( Fig. 29 ), and only the most forsteritic relics plot in the field for Paris Type I olivine. Fig. 30 shows a histogram of Cr 2O3 concentration in 238 ferroan olivine grains in Type II chondrules, chondrule fragments and isolated in Paris matrix. The distribution is very similar to those in histograms ( Grossman and Brear- ley, 2005) for three 3.0 CO chondrites and Acfer 094, and unlike those for two slightly re-equilibrated (CO 3.2) chondrites. Olivine generally shows normal zoning for Ca, Mn and Na, but in chondrules with chromite, it is fractionated to lower Cr, reflecting the co-crystallization of chromite and olivine (Hewins et al., 2011 ). Reverse zoning due to reheating is not observed.

6. DISCUSSION Fig. 31. (a) The PCP index for CM chondrites ( Rubin et al., 2007 ) 6.1. Degree of alteration and Paris uses the composition of tochilinite–cronstedtite intergrowths as an estimate of petrographic type. (b and c) Paris points are placed arbitrarily to fall near the curve at 2.7–3.0. Error bars 6.1.1. Mineralogy–petrology are for one standard deviation and Paris points are not significantly The abundance of metal, amorphous material and poorly different. The PCP Index, designed to measure the replacement of crystalline phyllosilicates in Paris indicates a low degree of cronstedtite, suggests that the freshest material in Paris, where little alteration. The least altered areas are close to the microstruc- cronstedtite has grown, is more altered than bulk. ture of the Meteorite Hills MET 00426 and Queen Alexandra Range QUE 99177 CR chondrites of type 3.00 ( Abreu and Brearley, 2010 ), and the ungrouped Acfer 094 chondrite ( Gre- Phyllosilicates and tochilinite are abundant in more al- shake, 1997 ). These primitive chondrites also contain nanosul- tered regions of the meteorite, and the microstructure is like fides and forsterite in association with the amorphous phase, those of other CM chondrites (e.g. Tomeoka and Buseck, as in Paris ( Figs. 9 a and 11 a, respectively). 1985; Lauretta et al., 2000 ). In these CM chondrites, alter- Paris contains up to 3 Â the metal content of the least al- ation is seen to involve the formation of serpentine (chrys- tered CM studied by Rubin et al. (2007) . Metal grains in its otile–greenalite solid solution) from Fe-rich phases, matrix show rimming with tochilinite, a reaction during the including cronstedtite, and little Fe metal has survived earliest stage of CM alteration ( Tomeoka and Buseck, (Tomeoka and Buseck, 1985; Browning et al., 1996; Rubin 1985). A typical grain in a moderately altered region of sec- et al., 2007 ). Several measures of degree of alteration at- tion 2010-4 has a rim 2–10 lm thick, as shown in Fig. 8c. tempt to quantify the degree of serpentine formation. This corresponds to alteration stage 1 of the aqueous alter- Browning et al. (1996) used the Fe 3+ /Si ratio in fine matrix ation scale 0–4 of Palmer and Lauretta (2011). It indicates to reflect the replacement of cronstedtite but phyllosilicates the beginning of replacement of metal grains by S-rich in Paris contain too many inclusions, mainly of sulfide, for water in fresh matrix. this approach to be reliable. This in itself indicates the R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 217 immature, unequilibrated nature of its aqueous alteration and Olsen, 1974; Wolf et al., 1980 ), with subsequent modi- phases (Palmer and Lauretta, 2011). fication during aqueous alteration. Importantly, the effect Rubin et al. (2007) devised a composite index to CM of aqueous alteration is seen in Paris to facilitate elemental alteration, tracking metal abundance, alteration of chon- transport on the millimeter scale between the original ma- drule phenocrysts, and in which the composition of PCP trix and refractory material, so that fine-grained regions aggregates plays an important role. The petrologic (alter- petrographically identifiable as altered matrix are now sim- ation) subtypes defined range from 2.6 (moderately altered) ilar in composition to the bulk CM. to 2.0 (totally altered). Rubin et al. (2007) showed that this The work of Clayton and Mayeda (1999) has demon- sequence correlated with decreasing S/SiO2 and PFeO/ strated that significant oxygen isotopic differences exist be- SiO2 in PCP, i.e. the approach to a simple serpentine com- tween the composition of CM2 whole rocks and matrix position. We attempted to use these composition relations separates. Thus, Murchison has a whole rock d18 O compo- to define the petrologic subtype of Paris. Many tochili- sition of 7.30 &, whereas the matrix has a significantly hea- nite-rich intergrowths were excluded because they have less vier value of 12.7 &. However, in view of the relatively than the 10% necessary for the PCP index. We show the uniform modal content of Paris ( 47% chondrules) the compositions of Paris and CM PCP in Fig. 31 a. We used 4.5& difference between the least and most altered portions exponential fits to the composition-index data of Rubin of the meteorite cannot be explained on the basis of modal et al. (2007) as linear regressions placed Paris at values matrix variations. Instead oxygen isotope variation can be greater than the desired limit of 3.0. Paris falls on the explained by local-scale differences in the extent of alter- “FeO”/SiO2 curve near subtype 2.8 (Fig. 31 b) and on the ation. For the variations in Fig. 6, this reflects variation S/SiO2 curve near 3.0 (Fig. 31 c). Blanchard et al. (2011) re- in fluid rock ratios, but for heavily altered clast ( Fig. 7b) ported similar compositions of PCP in Paris sections 2010-5 there could have been a longer duration of alteration. and -7. As is clear from Fig. 2b analyses from Paris significantly We also used the PCP Index in an attempt to distinguish extend the CM2 oxygen isotope trend towards the CO3s, the less and more altered zones of section 2010-04, domains although a distinct gap still exists between the two groups. 1 cm in size (Fig. 6, 7a). Very few PCP analyses in the This suggests that, while the anhydrous precursor to the fresher zone were suitable as they were too poor in SiO 2 CM2s was CO3-like, the two groups are samples from dis- and cronstedtite, probably because PCP had not evolved tinct asteroidal sources. The main difference between the as much as in 2.6 CM chondrites. The data showed that two groups was presumably the content of ice/water that the fresh zone had a lower petrographic index than the accreted into their respective parent asteroids, with the more altered zone (though not statistically significant), be- CO3 being essentially anhydrous and the CM2s ice/water cause cronstedtite is more abundant in the more altered rich. The evidence from dark inclusions (Zolensky et al., zone than in the fresh zone, not less as for the most altered 1993) suggests that CM2-like material was widespread in “normal” CM chondrites for which the PCP index was the early Solar System and consequently the existence of developed by Rubin et al. (2007) . multiple CM parent bodies cannot be ruled out. However, Based on PCP intergrowths in bulk Paris, it would be the oxygen isotope data from Paris and other CM2s sug- classified as an alteration subtype 2.9 ± 0.1. This PCP index gests that if there were a large number of such bodies then suggests very little replacement of primary matrix silicates, they all experienced a remarkably similar style of aqueous and the presence of amorphous or poorly crystalline phyl- alteration. losilicates reinforces a low alteration index. A complication in the PCP index approach is that in the earliest stages of 6.2. Thermal history alteration, tochilinite and cronstedtite may form from amorphous matrix material and metal. We are studying 6.2.1. Olivine and chromite the evidence for an early increase rather than a decrease The distribution of Cr in olivine and its relationship to in cronstedtite abundance in the less altered regions of Paris primary spinel or exsolved chromite are guides to metamor- (Leroux et al., 2013 ). Parent body alteration in CM chon- phic history (Engi, 1983; Johnson and Prinz, 1991; Gross- drites probably occurred as a water-limited process and man and Brearley, 2005). Mild reheating causes there are substantial variations in alteration within short precipitation of chromite in Cr-bearing igneous olivine. distances in other CM chondrites too ( Brearley, 2006; Pal- Grossman and Brearley (2005) recommended study of Cr mer, 2009). The local abundance of metal and the amor- in olivine with > 2% Fa, but since olivine Fa 2–10 is rare phous phase in Paris suggests low ice contents with a in Paris, we applied their approach to Type II olivine. heterogeneous distribution. The mean and standard deviation of Cr 2O3 concentrations in Paris olivine are 0.33 and 0.13 wt.% Cr 2O3, respectively 6.1.2. Bulk chemical and oxygen isotopic composition lower and higher than the preliminary values given by Bou- Fresh and altered matrix differ in composition. LA-ICP- rot-Denise et al. (2010) . The mean and standard deviation MS data ( Fig. 4) for altered matrix exhibits a volatile-de- are the same as those for Cr 2O3 in olivine in Y-81020 and pleted pattern characteristic of CM chondrites, but matrix Colony, both usually classified as Type 3.0, though they in unaltered regions has undepleted abundances like CI are higher than the lowest 3.0 (Fig. 32 ). chondrite (Zanda et al., 2011a,b ). These results are compat- The partitioning of Fe/Mg between chromite and coex- ible with an origin for carbonaceous chondrites by mixing isting olivine is temperature dependent ( Engi, 1983; John- the high temperature fraction with CI dust ( Grossman son and Prinz, 1991) as well as dependent on the 218 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

the inclusions in metal may here have formed by exsolution during slow chondrule cooling.

7. CONCLUSIONS

(1) Paris is classified as a CM2 chondrite, based on the petrography with chondrules in a matrix containing hydrated minerals, with confirmation from mineral compositions, major and trace element abundances, O and Cr isotope systematics, magnetic properties and noble gas data. It contains less matrix than other CM chondrites, 47%, with 8% fine-grained rims and 45% chond- Fig. 32. Plot of the standard deviation vs. the mean of the Cr 2O3 rules. It shows very fresh and more altered domains. content of ferroan olivine, after Grossman and Brearley (2005). (2) Eleven analyses of Paris show a wide range of oxygen Paris olivine plots beside that of Colony and Y-81020, two 3.0 chondrites. isotopic compositions, with less altered material plotting closer to CO chondrites than other CM chon- Cr/(Cr + Al) ratio of the chromite. We used the graphical drites, reflecting the two component mix between 16 16 approach of Johnson and Prinz (1991) to determine oliv- O-enriched anhydrous silicates and O-depleted ine–chromite equilibration temperatures (Hewins et al., hydrated phases. The regression line for Paris data 2011). The olivine–chromite assemblage in all chondrules intersects the CO3 field showing that the CM and in Paris plots near or above the 1400 °C isotherm, as it does CO groups are genetically related. for all CM chondrites (our data; Johnson and Prinz, 1991). (3) The Cr isotopic data of the minerals fall on a line 53 55 À6 In contrast, mildly reheated (3.1–3.3) CO chondrite olivine– with a slope of Mn/ Mn = (5.76 ± 0.76) Â 10 with an initial 53 Cr = À0.132 ± 0.055. If this pro- chromite pairs have their KDFe/Mg partially reset, yielding i a temperature range of 1400–600 °C ( Johnson and Prinz, vides direct chronological information, the corre- 1991). Thus the olivine in Paris resembles that in 3.0 chon- sponding age for the Paris components, particularly drites, in that there has been no visible diffusion of cations Type I and Type II chondrules, corresponds to due to reheating. Semarkona contains phyllosilicates in its 4566.44 ± 0.66 Myr, which is close to the age of matrix, yet is classified as an LL3.0. The water content CAI and puts new constraints on the early evolution was low however, because its matrix abundance is low of the solar system. and chondrule glass is preserved. The total abundance of (4) The REE pattern in Paris is extremely flat and secondary minerals in Paris is consistent with classification resembles those of most other CM chondrite falls, as a Type 2 chondrite. especially Murchison. Paris shares the depletion pattern of elements more volatile than Mg, with high 6.2.2. Sulfide and metal Na relative to K as in fresh CM falls. The LA- Hexagonal pyrrhotite contains granular pentlandite ICP-MS analysis for altered matrix exhibits the around its borders (Fig. 9a and b). Such Hpo-Pn grains cor- CM volatile-depleted pattern but that for unal- respond to category A CM chondrite sulfides (unheated tered matrix has undepleted abundances like CI group, secondary heating T < 100 °C) in the terminology chondrite. of Kimura et al. (2011) . Metal in Type I chondrules and (5) Paris contains large amounts of light noble gases with 4 20 À4 in isolated globules in matrix is kamacite (with only very He and Ne concentrations higher than 1 Â 10 - 3 À1 À7 3 À1 rare more Ni-rich exceptions). The range of Ni and Co cm g and 1 Â 10 cm g , respectively, (Fig. 25 a) corresponds to that of metal of Group A CM explained by the occurrence of a solar wind compo- chondrites (Kimura et al., 2011 ), rather than the more nent. The Q component is also present in Paris. 3 equilibrated Group B. The He c ages are 2.81 ± 0.41 and 2.75 ± 0.18 Ma 21 Inclusions of silica ( Fig. 9d), daubreélite, Cr phosphide, for two samples, in good agreement with the Ne c chromite, schreibersite and troilite occur in metal ( Bourot- ages of 3.65 ± 0.10 and 3.55 ± 0.08 Ma. Denise et al., 2010 ) often with several of these phases asso- (6) There are gradational contacts between the freshest ciated, whereas in OC and Renazzo such phases tend not to areas with approximately 3% metal and the more occur together (Zanda et al., 1994 ). There is, however, a PCP-rich areas. More altered matrix domains tendency for a given kamacite globule to contain one dom- contain 1–3% calcite. Magnetite spherules are present inant kind of precipitate, presumably related to the initial in very restricted domains. Isolated polycrystalline concentrations of the minor elements. Zanda et al. (1994) Fe–Ni sulfide grains consist of pyrrhotite rimmed distinguished two generations of inclusions in metal, those with pentlandite in a peripheral corona. Metal in formed during slow chondrule cooling and those formed matrix, sometimes with silica inclusions, shows thin during metamorphism. For Renazzo and Semarkona, rims of tochilinite in the freshest zones, but is exten- inclusions were interpreted as experiencing only a very mild sively replaced in the more altered regions. reheating, The secondary mineral assemblages of CM chon- (7) TEM images of the matrix reveal a complex fine- drites suggest low temperatures (<100 °C) and we have seen grained assemblage dominated by an amorphous no indication in sulfides or other silicates of reheating. Thus phase, with a gradation to a coarse-grained fibrous R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 219

morphology compatible with a serpentine structure. (13) The least altered areas contain abundant amorphous FIB samples also contain elongated grains of cronsted- material and poorly crystalline phyllosilicates and tite sometimes with fine interstratified tochilinite. Cron- have a microstructure like those of primitive chon- stedtite and fibrous material are coarser in the matrix of drites of type 3.00. They contain up to 3x the metal more altered zones. Tochilinite seen in TEM is mainly a content of the least altered of the classic CM chon- sulfur-rich fibrous phase, locally associated with an iron drites. Their metal has thin tochilinite rims, corre- oxide phase, probably magnetite. Rounded forsterite sponding to stage 1 of metal. Using an exponential and enstatite, micron to sub-micron in size, and hollow fit to the PCP Index ( Rubin et al., 2007 ) we obtained carbon globules are observed in matrix. a value of 2.9 ± 0.1. This approach did not indicate (8) Serpentine and cronstedtite compositions (EMPA) that the freshest material was less altered. Cronsted- run from about Chrysotile60 –Greenalite40 to about tite and tochilinite were being formed in the more Chrysotile20 –Cronstedtite80 but ‘serpentine’ must a altered regions of Paris whereas in normal CM chon- mixture of the amorphous phase with phyllosilicates. drites they were being transformed to serpentine. Compositions (ATEM) of the abundant amorphous Aqueous alteration in Paris transported elements phase and the finest fibers are comparable to the ser- out of high temperature material into altered matrix, pentine, but more Si-rich. With ATEM, S is detect- changing its composition from CI-like to CM-like. able in fibrous phyllosilicates though sulfide grains (14) The 4.5& difference in d18 O between the least and are not present. Platy cronstedtite is sometimes int- most altered portions can be explained by local-scale erstratified with thin layers of tochilinite. With differences in the extent of alteration reflecting varia- EPMA, some PCP includes magnetite as well as cron- tion in fluid rock ratios. stedtite and tochilinite. (15) The mean and standard deviation of Cr 2O3 concen- (9) Magnetic susceptibility measurements show the Ver- trations in Paris olivine are the same as those for wey transition and a Curie temperature indicating the Cr 2O3 in olivine in Y-81020 and Colony, both usually presence of magnetite, except for the sample that was classified as Type 3.0. The partitioning of Fe/Mg specifically sampled in the less aqueously altered between chromite and coexisting olivine is not reset lithology of the meteorite. The average saturation from chondrule temperatures. Exsolution of inclu- 2 magnetization (M S) of 5.9 Am /kg for Paris indicates sions in metal may be due to slow chondrule cooling a maximum magnetite content of 6.4 wt.%, and a or a very mild reheating. Pyrrhotite–pentlandite tex- maximum metal content of 2.7 wt.%. tures and compositions of Fe–Ni metal indicate (10) The CAI are mainly fine-grained spinel-rich inclusions, unheated group A of CM chondrites. with a condensation sequence of hibonite, perovskite, (16) Paris is an essentially unmetamorphosed CM chon- grossite, spinel, gehlenite, anorthite, diopside/fassaite drite and it is partially altered. It is conceivably a and forsterite. Hibonite-rich CAI and melilite-rich CM2.9, though we have no systematic way to index CAI are present. RMN, with similar compositions to low degrees of alteration and the alteration is those in Murchison, occur exclusively inside CAI or heterogeneous. attached to disrupted CAI minerals. The most refractory alloys occur in hibonite, with compositions dominated by Os and Ir; those in perovskite + Al-spinel ACKNOWLEDGEMENTS CAI are Os–Ir–Mo–Ru alloys; those hosted in spinel are enriched in Pt and Rh. We thank Michel Fialin and Fre´de´ric Couffignal for invaluable (11) Type I chondrules contain olivine and pyroxene are assistance with electron probe analysis, Omar Boudouma for SEM Fa and Fs Wo . Ca decreases and cartography, David Troadec for the FIB sections, prepared at IEMN, 0.9±0.7 2.3±1.7 1.4±1.3 University Lille 1, and Laurette Piani for help with image analysis. Cr increases sharply in olivine going from Fa to 0 We are indebted to A. Elmaleh, E. Palmer and M. Zolensky for dis- Fa 1. Metal-rich chondrules are abundant, with a cussion. We thank the Agence Nationale de la Recherche for grants small range of Ni contents in metal, mostly 4 to ANR-09-BLAN-436 0042 (C.C. and J.G.), and ANR-08-Blan-0260- 8 wt.%. Inclusions of silica, daubreélite, Cr phos- CSD6 (C.G.) We thank the PNP (programme national de plane´tolo- phide, chromite, schreibersite and troilite were also gie) for support of the ATEM work (H.L.), and NASA for grant observed in metal. There is a correlation of Fa vs. NNX10AI37G (M.H.) in support of laser ablation measurements. “Cr 2O3” in the most forsteritic olivine (Fa 0.2 to 1) We thank STFC for grant ST/I001964/1 (R.C.G. and I.A.F.) in sup- and a negative correlation between Cr in metal and port of oxygen isotope laser fluorination work. Drs. N. Abreu and W. Fujiya, and A. Krot, are thanked for detailed constructive reviews Cr 2O3. Cr concentrations in metal are depleted relative to Ni partly due to oxidation. and editorial comments, respectively, which led to major improve- ments to the paper. (12) Type II chondrules present are IIA (PO,BO), IIAB and IIB, but also II(A)B, with minor highly ferroan olivine, and IIA(C) with augite as the only pyroxene. REFERENCES The range of melt-grown olivine compositions is Fa 7 to Fa 76 and of orthopyroxene En 34–63. Type II chon- Amelin Y. (2008) U–Pb ages of angrites. Geochim. Cosmochim. drule olivine in Paris has an FeO–MnO ratio of Acta 72 , 221–232. 100, typical of CM and CO chondrites but there Abreu N. M. and Brearley A. J. (2010) Early solar system processes exist distinct outliers from this distribution. recorded in the matrices of two highly pristine CR3 carbona- 220 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

ceous chondrites, MET 00426 and QUE 99177. Geochim. Croat T. K., Berg T., Bernatowicz T., Groopman E. and Jhadav Cosmochim. Acta 74 , 1146–1171. M. (2013) Refractory metal nuggets within presolar graphite: Barrat J.-A., Zanda B., Moynier F., Bollinger C., Liorzou C. and first condensates from acircumstellar environment. Meteor. Bayon G. (2012) Geochemistry of CI chondrites: major and Planet. Sci. 48 , 686–699. trace elements, and Cu and Zn isotopes. Geochim. Cosmochim. Dunlop D. J. and O ¨ zdemir O¨ . (1997) Rock Magnetism: fundamen- Acta 83 , 79–92. tals and Frontiers. Cambridge Univ. Press, Cambridge . Berg T., Maul J., Scho¨nhense G., Marosits E., Hoppe P., Ott U. Elmaleh A., Tarantino S. C., Zema M., Zanda B., Devouard B. and Palme H. (2009) Direct evidence for condensation in the and Fialin M. (2011) Low-T magnetic properties of 5 CM2 early solar system and implications for nebular cooling rates. chondrites: comparison with the signature of cronstedtite and Astrophys. J. 702, L172–L176. with preliminary results from the Paris chondrite. Meteor. Berlin J., Jones R. H. and Brearley A. J. (2011) Fe–Mn systematics Planet. Sci. 45 , #5492 (abstr.). of type IIA chondrules in unequilibrated CO, CR, and ordinary Elmaleh A., Tarantino S. C., Zema M., Devouard B. and Fialin M. chondrites. Meteor. Planet. Sci. 45 , 513–533. (2012) The low-temperature magnetic signature of Fe-rich Blanchard I., Gounelle M., Bourot-Denise M. and Kearsley A. serpentine in CM2 chondrites, comparison with terrestrial (2011) Secondary processes within the Paris CM chondrite. cronstedtite and evolution with the degree of alteration. Meteor. Planet. Sci. 46 , #5322 (abstr.). Geochem. Geophys. Geosyst. 13 , Q05Z42. Blander M., Fuchs L. H., Hurowitz C. and Land R. (1980) Engi M. (1983) Equilibria involving Al–Cr spinel: I. Mg–Fe Primordial refractory metal particles in the Allende meteorite. exchange with olivine. Experiments, thermodynamic analysis, Geochim. Cosmochim. Acta 44 , 217–223. and consequences for geothermometry. Amer. J. Sci. 283A, 29– Bourot-Denise M., Zanda B., Marrocchi Y., Greenwood R. C., 71. Pont S., Hewins R. H., Franchi I. A. and Cornen G. (2010) Eugster O. and Michel Th. (1995) Commo asteroid break-up events Paris: the slightly altered, slightly metamorphosed CM that of eucrites, diogenites, and howardites and cosmic-ray produc- bridges the gap between CMs and COs. Lunar Planet. Sci. XLI. tion rates for noble gases in achondrites. Geochim. Cosmochim. Lunar Planet. Inst., Houston. #1683 (abstr.). Acta 59 , 177–199. Brearley A. J. (1993) Matrix and fine-grained rims in the Eugster O., Herzog G. F., Marti K. and Caffee M. W. (2006) unequilibrated CO3 chondrite, ALHA77307: origins and evi- Irradiation records, cosmic-ray exposure ages, and transfer dence for diverse, primitive nebular dust components. Geochim. times of meteorites. In Meteorites and the early solar system II Cosmochim. Acta 57 , 1521–1550. (eds. D. S. Lauretta and , Jr.H. Y. McSween). Arizona, The Brearley A. J. (2006) The role of microchemical environments in University of Arizona, Tucson, pp. 829–851 . the alteration of CM carbonaceous chondrites. Lunar Planet. Eugster O., Lorenzetti S., Kra¨henbu¨hl U. and Marti K. (2007) Sci. XLI. Lunar Planet. Inst., Houston. #2074 (abstr.). Comparison of cosmic-ray exposure ages and trapped noble Brennecka G. A. and Wadwha M. (2012) Uranium isotope gases in chondrule and matrix samples of ordinary, enstatite, and compositions of the basaltic angrite meteorites and the chro- carbonaceous chondrites. Meteor. Planet. Sci. 42 , 1351–1371. nological implications for the early solar system. Proc. Nat. Frank D. R., Zolensky M. E. and Le L. (2012) Using the Fe/Mn Acad. Sci. U. S. A 109, 9299–9303. ratio of FeO-rich olivine in Wild 2, chondrite matrix, and type Browning L. B., McSween, Jr., H. Y. and Zolensky M. E. (1996) IIA chondrules to disentagle their histories. Lunar Planet. Sci. Correlated alteration effects in CM carbonaceous chondrites. XLIII. Lunar Planet. Inst., Houston. #2748 (abstr.). Geochim. Cosmochim. Acta 60 , 2621–2633. Friedrich J. M., Wang M.-S. and Lipschutz M. E. (2002) Busemann H., Baur H. and Wieler R. (2000) Primordial noble Comparison of the trace element composition of Tagish Lake gases in “phase Q” in carbonaceous and ordinary chondrites with other primitive carbonaceous chondrites Meteor. Planet. studied by closed system etching. Meteor. Planet. Sci. 35 , 949– Sci. 37 , 677–686. 973. Fujiya W., Sugiura N., Sano Y. and Hiyagon H. (2013) Mn–Cr Caillet Komorowski C., Boudouma O., Reynard B., Van de ages of dolomites in CI chondrites and the Tagish Lake Moorte`le B. and El Goresy A. (2011) Insights on the thermal ungrouped carbonaceous chondrite. Earth Planet. Sci. Lett. history of the Paris CM chondrite based on the phase relations 362, 130–142. of its opaque assemblages. Meteor. Planet. Sci. 45 , A35, #5289 Gaboardi M. and Humayun M. (2009) Elemental fractionation (abstr.). during LA-ICP-MS analysis of silicate glasses: implications for Chizmadia L. J. and Brearley A. J. (2008) Mineralogy, aqueous matrix-independent standardization. J. Anal. At. Spectrom. 24 , alteration, and primitive textural characteristics of fine-grained 1188–1197. http://dx.doi.org/10.1039/b900876d. rims in the Y-791198 CM2 carbonaceous chondrite: TEM Glavin D. P., Kubny A., Jagoutz E. and Lugmair G. W. (2004) observations and comparison to ALHA81002. Geochim. Cos- Mn–Cr isotope systematics of the D’Orbigny angrite. Meteor. mochim. Acta 72 , 602–625. Planet. Sci. 39 , 693–700. Clayton R. N. and Mayeda T. K. (1984) The oxygen isotope record Go¨pel C., Birck J.-L. and Zanda B. (2011) Mn/Cr systematics in in Murchison and other carbonaceous chondrites. Earth Planet. carbonaceous chondrites: mineral isochrons versus stepwise Sci. Lett. 67 , 151–166. dissolution. Min. Mag. 75 , 936. Clayton R. N. and Mayeda T. K. (1999) Oxygen isotope studies of Go¨pel C., Birck J.-L., Cartigny P., Assayag N. and Zanda B. (2012) carbonaceous chondrites. Geochim. Cosmochim. Acta 63 , 2089– 54 Cr and 17 O in carbonaceous chondrites and an old 53 Cr/53 Mn 2104. age of the Paris meteorite. Min. Mag. 76 , 1775. Connelly J. N., Bizzarro M., Krot A. N., Nordlund A., Wielandt Go¨pel C., Birck J. -L., Cartigny P., Assayag N. and Zanda B. D. and Ivanova M. A. (2012) The absolute chronology and (2013) 54 Cr and ( D17 O in carbonaceous chondrites and an old thermal processing of solids in the solar protoplanetary disk. 53 Cr/53 Mn age for the Paris meteorite (submitted for Science 338, 651–655. publication). Consolmagno G. J., Britt D. J. and Macke R. J. (2008) The Greenwood R. C. and Franchi I. A. (2004) Alteration and significance of meteorite density and porosity. Chemie der Erde metamorphism of CO3 chondrites: evidence from oxygen and 68 , 1–29. carbon isotopes. Meteor. Planet. Sci. 39 , 1823–1838. R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222 221

Greenwood R. C., Franchi I. A., Gibson J. M. and Benedix G. K. Keller L. P. and Messenger S. (2012) Formation and processing of (2012) Oxygen isotope variation in primitive achondrites: the amorphous silicates in primitive carbonaceous chondrites and influence of primordial, asteroidal and terrestrial processes. cometary dust. Lunar Planet. Sci. XLIII. Lunar Planet. Inst., Geochim. Cosmochim. Acta 94 , 146–163. Houston. #1880 (abstr.). Greshake A. (1997) The primitive matrix components of the unique Kimura M., Grossman J. N. and Weisberg M. K. (2011) Fe–Ni carbonaceous chondrite ACFER 094: a TEM study. Geochim. metal and sulfide minerals in CM chondrites, an indicator for Cosmochim. Acta 61 , 437–452. thermal history. Meteor. Planet. Sci. 46 , 431–442. Grossman J. N. and Zipfel J. (2001). Meteor. Planet. Sci. 36 , A293– Kita N. T. and Ushikubo T. (2012) Evolution of protoplanetary A322. disk inferred from 26 Al chronology of individual chondrules. Grossman J. N. and Brearley A. J. (2005) The onset of metamor- Meteor. Planet. Sci. 47 , 1108–1119. phism in ordinary and carbonaceous chondrites. Meteor. Kring D. A. (1986) The petrology of meteoritic chondrules: Planet. Sci. 40 , 87–122. evidence for fluctuating conditions In the solar nebula. Ph. D. Grossman L. and Olsen E. (1974) Origin of the high-temperature thesis, Harvard Univ., p. 346. fraction of C2 chondrites. Geochim. Cosmochim. Acta 38 , 173– Lauretta D. S., Hua X. and Buseck P. R. (2000) Mineralogy of fine- 187. grained rims in the ALH 81002 CM chondrite. Geochim. Haack H., Grau T., Bischoff A., Horstmann M., Wasson J., Cosmochim. Acta 64 , 3263–3273. Sørensen A. N., Laubenstein M., Ott U., Palme H., Gellissen Leroux H., Rietmeijer F. J. M., Velbel M. A., Brearley A. J., Jacob M., Greenwood R. C., Pearson V. K., Franchi I. A., Gabelica D., Langenhorst F., Bridges J. C., Zega T. J., Stroud R. M., Z. and Schmitt-Kopplin P. (2012) Maribo – a new CM fall from Cordier P., Harvey R. P., Lee M., Gounelle M. and Zolensky Denmark. Meteor. Planet. Sci. 47 , 30–50. M. E. (2008) A TEM study of thermally modified Comet 81P/ Harries D., Berg T., Langenhorst F. and Palme H. (2012) Wild 2 dust particles by interactions with the aerogel matrix Structural clues to the origin of refractory metal alloys as during the Stardust capture process. Meteor. Planet. Sci. 43 , condensates of the solar nebula. Meteor Planet. Sci. 47 , 2148– 97–120. 2159. Leroux H., Cuvillier P., Zanda B. and Hewins R. H. (2013) A TEM Heber V., Wieler R., Baur H., Olinger C., Friedmann T. A. and investigation of the fine-grained matrix of the Paris CM Burnett D. S. (2009) Noble gas composition of the solar wind as chondrite. Lunar Planet. Sci. XLIV. Lunar Planet. Inst., collected by the Genesis mission. Geochim. Cosmochim. Acta 73 , Houston. #1528 (abstr.). 7414–7432. Liu M.-C., Chaussidon M., Go¨pel C. and Lee T. (2012) A Hewins R. H. and Zanda B. (2011) Type II chondrule origin in O heterogeneous solar nebula as sampled by CM hibonite grains. and C chondrites. Meteor. Planet. Sci. 45 , A94, #5144 (abstr.). Earth Planet. Sci. Lett. 327, 75–83. Hewins R. H. and Zanda B. (2012) Origin of FeO–MnO olivine Marrocchi Y., Derenne S., Marty B. and Robert F. (2005) reservoirs in chondrules. Meteor. Planet. Sci. 46 , #5274 Interlayer trapping of noble gases in insoluble organic matter (abstr.). of primitive meteorites. Earth Planet. Sci. Lett. 236, 569–578. Hewins R. H., Zanda B. and Bourot-Denise M. (2011) The Marrocchi Y., Marty B., Reinhardt P. and Robert F. (2011) formation of type II chondrules in CM chondrites: the view Adsorption of xenon ions onto defects in organic surfaces: from Paris. Lunar Planet. Sci. XLII. Lunar Planet. Inst., implications for the origin and the nature of organics in Houston. #1914 (abstr.). primitive meteorites. Geochim. Cosmochim. Acta 75 , Humayun M. (2012) Chondrule cooling rates inferred from 6255–6266. diffusive profiles in metal lumps from the Acfer 097 CR2 Marty B., Kelley S. and Turner G. (2010) Chronology and shock chondrite. Meteor. Planet. Sci. 47 , 1191–1208. history of the Bencubbin meteorite: a nitrogen, noble gas, and Humayun M., Simon S. B. and Grossman L. (2007) Tungsten and Ar–Ar investigation of silicates, metal and fluid inclusions. hafnium distribution in calcium-aluminum inclusions CAIs Geochim. Cosmochim. Acta 74 , 6636–6653. from Allende and Efremovka. Geochim. Cosmochim. Acta 71 , McSween, Jr., H. Y. (1977) Chemical and petrographic constraints 4609–4627. on the origin of chondrules and inclusions in carbonaceous Humayun M., Davis F. A. and Hirschmann M. M. (2010) Major chondrites. Geochim. Cosmochim. Acta 41 , 1843–1860. element analysis of natural silicates by laser ablation ICP-MS. McSween, Jr., H. Y. (1979) Alteration in CM carbonaceous J. Anal. At. Spectrom. 25 , 998–1005. http://dx.doi.org/10.1039/ chondrites inferred from modal and chemical variations in C001391A. matrix. Geochim. Cosmochim. Acta 43 , 1761–1770. Humbert F., Libourel G., France-Lanord C., Zimmermann L. and Menyeh A. and O’Reilly W. (1991) The magnetization process in

Marty B. (2000) CO 2-laser extraction-static mass spectrometry monoclinic pyrrhotite (Fe7S8) particles containing few analysis of ultra-low concentrations of nitrogen in silicates. domains. Geophys. J. Int. 104, 387–399. Geostand. Newsl. 24 , 255–260. Merouane S., Djouadi Z., Le Sergeant d’Hendecourt L., Zanda Jarosewich E. (1990) Chemical analyses of meteorites: a compila- B. and Borg J. (2012) Hydrocarbon materials of likely tion of stony and iron meteorite analyses. Meteoritics 25 , 323– interstellar origin from the Paris meteorite. Astrophys. J. 756, 337. 154. Johnson C. and Prinz M. (1991) Chromite and olivine in type II Metzler K., Bischoff A. and Sto¨ffler D. (1992) Accretionary dust chondrules in carbonaceous and ordinary chondrites: implica- mantles in CM chondrites: evidence for solar nebula processes. tions for thermal histories and group differences. Geochim. Geochim. Cosmochim. Acta 56 , 2873–2897. Cosmochim. Acta 55 , 893–904. Miller M. F., Franchi I. F., Sexton A. S. and Pillinger C. T. (1999) Kallemeyn G. W. and Wasson J. T. (1981) The compositional High precision d17 O isotope measurements of oxygen from classification of chondrites- I. The carbonaceous chondrite silicates and other oxides: methods and applications. Rapid groups. Geochim. Cosmochim. Acta 45 , 1217–1230. Commun. Mass Spectrom. 13 , 1211–1217. Kallemeyn G. W., Rubin A. E. and Wasson J. T. (1994) The Mittlefehldt D. W. and Wetherill G. W. (1979) Rb–Sr studies of CI compositional classification of chondrites: VI. The CR carbo- and CM chondrites. Geochim. Cosmochim. Acta 43 , 201–206. naceous chondrite group. Geochim. Cosmochim. Acta 58 , 2873– Mostefaoui S. (2011) The search for presolar oxides in Paris. 2888. Meteor. Planet. Sci. 45 , A94, #5170 (abstr.). 222 R.H. Hewins et al. / Geochimica et Cosmochimica Acta 124 (2014) 190–222

Nakamura N. (1974) Determination of REE, Ba, Fe, Mg, Na and Simon S., Grossman L., Hutcheon I., Phinney D., Weber P. and K in carbonaceous and ordinary chondrites. Geochim. Cosmo- Fallon S. (2006) Formation of spinel-, hibonite-rich inclusions chim. Acta 38 , 757–775. found in CM2 carbonaceous chondrites. Am Mineral 91 , 1675– Nakamura T. (2005) Post-hydration thermal metamorphism of 1687. carbonaceous chondrites. J. Mineral. Petrol. Sci. 100, 260–272. Tatsumoto M., Unruh D. M. and Desborough G. A. (1976) U–Th– Nazarov M. A., Kurat G., Brandstaetter F., Ntaflos T., Chaussi- Pb and Rb–Sr systematics of Allende and U–Th–Pb systematics don M. and Hoppe P. (2009) Phosphorus-bearing sulfides and of Orgueil. Geochim. Cosmochim. Acta 40 , 617–634. their associations in CM chondrites. Petrology 17 , 101–123. Tomeoka K. and Buseck P. R. (1985) Indicators of aqueous Ozima M. and Podosek F. A. (2002) Noble Gas Geochemistry. alteration in CM carbonaceous chondrites: microtextures of a Cambridge Univ. Press, Cambridge, p. 286 . layered mineral containing Fe, S, O and Ni. Geochim. Cosmo- Palme H. and Wlotzka F. (1976) A metal particle from a Ca, Al chim. Acta 49 , 2149–2163. rich inclusion from the meteorite Allende, and the condensation Trinquier A., Birck J.-L. and Alle`gre C.-J. (2007) Widespread 54 Cr of refractory siderophile elements. Earth Planet. Sci. Lett. 33 , heterogeneity in the inner solar system. Astrophys. J. 655, 1179– 45–60. 1185. Palme H., Hutcheon I. D. and Spettel B. (1994) Composition and Wasson J. T. and Kallemeyn G. W. (1988) Composition of origin of refractory-metal-rich assemblages in a Ca, Al-rich chondrites. Philos. Trans. 325, 535–544. Allende inclusion. Geochim. Cosmochim. Acta 58 , 495–513. Wieler R. and Baur H. (1994) Krypton and xenon from the solar Palmer E. E. (2009) Volatiles on solar system objects: carbon wind and solar energetic particles in two lunar ilmenites of dioxide on Iapetus and aqueous alteration in CM chondrites. different antiquity. Meteoritics 29 , 570–580. Ph. D. thesis, Arizona Univ., p. 234. Wolf D. and Palme H. (2001) The solar system abundances of Palmer E. E. and Lauretta D. S. (2011) Aqueous alteration of phosphorus and titanium and the nebular volatility of phos- kamacite in CM chondrites. Meteor. Planet. Sci. 46 , 1587–1607. phorus. Meteor. Planet. Sci. 36 , 559–571. Remusat L., Bernard S., Le Guillou C. and Brearley A. (2011) Wolf R., Richter G., Woodrow A. and Anders E. (1980) Chemical Parent body influence on organic matter: the message from fractionations in meteorites. XI–C2 chondrites. Geochim. Cos- Paris. Meteor. Planet. Sci. 45 , #5327 (abstr.). mochim. Acta 44 , 711–717. Rochette P., Gattacceca J., Bonal L., Bourot-Denise M., Chevrier Yin Q.-Z., Jacobsen B., Moynier F. and Hutcheon I. D. (2007) V., Clerc J.-P., Consolmagno G., Folco L., Gounelle M., 53 Mn–53 Cr chronometry for chondrules. Astrophys. J. 662, Kohout T., Pesonen L., Quirico E., Sagnotti L. and Skripnik A. L43–L46. (2008) Magnetic classification of stony meteorites: 2. Non- Yin Q. -Z., Yamashita K., Yamakawa A., Tanaka R., Jacobsen B., ordinary chondrites. Meteor. Planet. Sci. 43 , 959–980. Ebel D., Hutcheon I. D. and Nakamura E. (2009) 53 Mn– 53 Cr Rubin A. E. and Wasson J. T. (1987) Chondrules, matrix and systematics of allende chondrules and epsilon 54 Cr–D17 O coarse-grained chondrule rims in the Allende meteorite: origin, correlation in bulk carbonaceous chondrites. Lunar Planet. interrelationships and possible precursor components. Geochim. Sci. XL . Lunar Planet. Inst., Houston. #2006 (abstr.). Cosmochim. Acta 51 , 1923–1937. Zanda B., Bourot-Denise M., Perron C. and Hewins R. H. (1994) Rubin A. E., Trigo-Rodrıguez J. M., Huber H. and Wasson J. T. Origin and metamorphic redistribution of silicon, chromium (2007) Progressive aqueous alteration of CM carbonaceous and phosphorus in the metal of chondrites. Science 265, 1846– chondrites. Geochim. Cosmochim. Acta 71 , 2361–2382. 1849. Russell S. D. J., Longstaffe F. J., King P. L. and Larson T. E. Zanda B., Bourot-Denise M., Hewins R. H., Barrat J.-A. and (2010) The oxygen-isotope composition of chondrules and Gattacceca J. (2010) The Paris CM chondrite yields new isolated forsterite and olivine grains from the Tagish Lake insights on the onset of parent body alteration. Meteor. Planet. carbonaceous chondrite. Geochim. Cosmochim. Acta 74 , 2484– Sci. 44 , A222, #5312 (abstr.). 2499. Zanda B., Humayun M., Barrat J. -A., Bourot-Denise M. and Scherer P. and Schultz L. (2000) Noble gas record, collisional Hewins R. H. (2011a) Bulk and matrix composition of Paris. history, and paring of CV CO., CK, and other carbonaceous Inferences on parent-body alteration and the origin of matrix- chondrites. Meteor. Planet. Sci. 35 , 145–153. chondrule complementarity. Lunar Planet. Sci. XLII. Lunar Schulze H. (2007) Mineralogy and mineral chemistry of noble Planet. Inst., Houston. #2040 (abstr.). metal grains in R chondrites. Lunar Planet. Sci. XXX. Lunar Zanda B., Humayun M., Barrat J.-A., Bourot-Denise M. and Planet. Inst., Houston. #1720 (abstr.). Hewins R. H. (2011b) Chemistry of carbonaceous chondrite Schwander D., Berg T., Ott U., Scho¨nhense G. and Palme H. matrices: parent-body alteration and chondrule-matrix com- (2011) Refractory metal nugetts in carbonaceous chondrites plementarity. Meteor. Planet. Sci. 45 , #5358 (abstr.). and early solar nebula condensates. Lunar Planet. Sci. #9070 Zega T. J., Nittler L. R., Busemann H., Hoppe P. and Stroud R. (abstr.). M. (2007) Coordinated isotopic and mineralogic analyses of Schwander D., Berg T., Ott U., Scho¨nhense G. and Palme H. planetary materials enabled by in situ lift-out with a focused ion (2012) Formation of refractory metal alloys and their occur- beam scanning electron microscope. Meteor. Planet. Sci. 42 , rence in CAIs. Meteor. Planet. Sci. 47 , A348, #5203 (abstr.). 1373–1386. Schwander D., Berg T., Ott U., Scho¨nhense G. and Palme H. Zolensky M., Barrett R. and Browning L. (1993) Mineralogy and (2013) Clues to formation of refractory metal nuggets. Lunar composition of matrix and chondrule rims in carbonaceous Planet. Sci. XLIV. Lunar Planet. Inst., Houston. #1959 (abstr.). chondrites. Geochim. Cosmochim. Acta 57 , 3123–3148. Simon S. B. and Grossman L. (2003) Petrography and mineral chemistry of the anhydrous component of the Tagish Lake Associate editor: Alexander N. Krot carbonaceous chondrite. Meteor. Planet. Sci. 38 , 813–825.