Université Libre de Bruxelles – Département des Sciences de la Terre et de l’Environnement Musée Royal d’Afrique Centrale – Département de Géologie

The Archaean silicon cycle Insights from silicon isotopes and Ge/Si ratios in banded iron formations, palaeosols and shales

Camille DELVIGNE

Thèse présentée en vue de l’obtention du grade de Docteur en Sciences

Composition du jury : Pr. Luc André (MRAC, Belgique) - Promoteur Pr. Nadine Mattielli (ULB, Belgique) - Co-promoteur Pr. Alain Bernard (ULB, Belgique) - Président Dr. Vinciane Debaille (ULB, Belgique) - Secrétaire Pr. Marc Chaussidon (CRPG, France) Pr. Axel Hofmann (Université de Johannesburg, Afrique du Sud)

Septembre 2012

Remerciements

Arrivée au terme de cette thèse, je tiens à remercier les nombreuses personnes qui m’ont aidée à reconstituer le puzzle à l’image mystère que représente une thèse. Je tiens à adresser mes premiers remerciements à Luc André. Tout d’abord pour m’avoir donné l’opportunité de réaliser ce travail mais surtout pour son enthousiasme inébranlable, sa vision scientifique et ses encouragements. Pilier essentiel à cette thèse, un grand merci à Damien Cardinal, sans qui de nombreuses pièces du puzzle seraient encore manquantes. Merci pour ta disponibilité de tous les instants, aussi bien pour les questions scientifiques que pour les nombreux jokers « coup de fil à un ami » lors des soucis divers et variés imaginés par le célèbre MC. Et un merci supplémentaire pour être sorti de tes océans pour te frotter aux problèmes de bons vieux cailloux ! J’adresse mes remerciements sincères à Marc Chaussidon et Axel Hofmann d’avoir accepté un aller-retour sur Bruxelles pour évaluer ma thèse. Merci également à Nadine Mattielli, Alain Bernard, Vinciane Debaille d’avoir accepté de juger mon travail. Cadre incontournable d’une thèse, le labo. Je tiens à remercier Laurence, Nourdine et Jacques pour leur aide quotidienne au labo et lors des analyses ainsi que François, Harold, Pierre-Denis, Frédéric, Claire, Ginnie, Sophie et Katrin pour la vie au labo, leur bonne humeur et les échanges dans un bureau, un couloir ou au téléphone. Merci également à Suzanne pour sa gentillesse et son efficacité. René, pour son aide précieuse dans la préparation des échantillons. Et Didier, pour les jolis posters. Ensuite, je tiens à remercier en particulier Katrin, Sophie, Jean-Thomas, Ginnie, Eléanore, Vinciane, François et Loïc qui ont apporté une dimension supplémentaire aux conférences. Merci à nouveau à Axel Hofmann ainsi qu’à Phil Thurston sans qui il me manquerait tout simplement mes échantillons.

Merci également à Nadine, Vinciane, Jeroen et Ivan pour leur gestion du labo MC-ICP-MS et leur aide lors des caprices du MC.

Cette recherche n’aurait pas été possible sans le soutien financier du FNRS que je remercie pour la bourse FRIA qui m’a été accordée pendant quatre années.

Je terminerai par ceux qui me sont le plus chers, ma famille et mes amis.

Merci à mes amis, à ma belle-famille et ma famille pour leurs pensées et leur soutien lors de cette épreuve. Merci en particulier « aux filles », Marjo, Gaëlle, Nath et Marie. Il est bon de vous savoir à l’écoute lors des questionnements et des moments de doutes qui nous ont toutes taraudé l’esprit tôt ou tard et encore.

Merci à mes parents pour leur confiance sans limite dès le début et leur soutien jusqu’à la fin. Merci de m’avoir permis d’arriver jusque là, tout d’abord en éveillant ma curiosité à la moindre fleur ou chenille croisée sur un chemin, ensuite en m’ouvrant les portes de l’université et finalement en préparant de bons petits plats prêts à réchauffer pour la dernière ligne droite.

Et enfin, je ne saurai terminer ces propos sans te remercier, Pascal, toi qui partages avec moi au quotidien les hauts comme les bas. Tu as su à chaque instant avoir les mots qu’il fallait pour me remonter le moral dans les moments de doute et de remise en questions. Et pour tellement plus, tout simplement, merci.

Dilbeek, H-39

CONTENTS Abstract................................................................................................................................. 1 Résumé ................................................................................................................................. 3 General objectives ................................................................................................................. 5 Thesis outline ........................................................................................................................ 7

CHAPTER 1 ......................................................................................................... 9 General overview .............................................................................................. 9 1.1

The history of Earth................................................................................ 11

1.1.1

The Hadaean (~4.56-3.8Ga) .................................................................................................. 11

1.1.2

The Archaean (~3.8-2.5 Ga) ................................................................................................. 12

1.1.3

The Proterozoic (~2.5-0.5 Ga).............................................................................................. 13

1.2

The Archaean Earth ................................................................................ 14

1.2.1

Archaean atmosphere and the “Faint Young Sun” ............................................................ 14

1.2.2

Archaean continental crust and its weathering ................................................................. 15

1.2.3

Archaean hydrothermalism ................................................................................................ 18

1.2.4

Archaean ocean .................................................................................................................. 19

1.3

The Archaean silicon cycle .................................................................... 20

1.4

Banded Iron Formations ....................................................................... 23

1.4.1

Depositional settings ..........................................................................................................24

1.4.2

Mineralogies ....................................................................................................................... 27

1.4.3

Deposition mechanisms...................................................................................................... 27

1.5

1.4.3.1

Iron oxidation processes ................................................................................................. 28

1.4.3.2

Mechanisms of silica precipitation................................................................................... 29

1.4.3.3

BIF deposition mechanisms ............................................................................................. 30

Precambrian cherts ............................................................................... 34

1.6

Tracers of silicon cycle .......................................................................... 35

1.6.1

Silicon stable isotopes ........................................................................................................ 35 1.6.1.1

1.6.2

Germanium/Silicon ratio ..................................................................................................... 41 1.6.2.1

1.6.3

Silicon isotopic variations on Earth .................................................................................. 37

Ge/Si ratio variations on Earth......................................................................................... 41

Silicon isotopes in BIFs and cherts .................................................................................... 44 1.6.3.1

Banded iron formations .................................................................................................. 44

1.6.3.2

Precambrian cherts ......................................................................................................... 44 References .......................................................................................................................... 46

CHAPTER 2 ...................................................................................................... 59 Analytical developments ................................................................................ 59 2.1

Outline ....................................................................................................61

2.2

Developments in REE+Y and Ge analysis by HR-ICP-MS ....................... 62

2.2.1

Sample dissolution ............................................................................................................. 62

2.2.2

HR-ICP-MS analysis ............................................................................................................. 63

2.3

2.2.2.1

Detection limits............................................................................................................... 63

2.2.2.2

Correction for interferences on Eu and Gd ...................................................................... 64

2.2.2.3

Accuracy ......................................................................................................................... 66

Controlling the mass bias introduced by anionic and organic matrices

in silicon isotopic measurements by MC-ICP-MS* .......................................... 67 2.3.1

Abstract............................................................................................................................... 67

2.3.2

Introduction ....................................................................................................................... 68

2.3.3

Material and methods ....................................................................................................... 69

2.3.4

2.3.3.1

Instrumentation .............................................................................................................. 69

2.3.3.2

Material and sample preparation .................................................................................... 69

2.3.3.3

Matrix effect counter measures ...................................................................................... 71

Results and discussion ........................................................................................................ 73

2.3.5

2.3.4.1

Doped samples ............................................................................................................... 75

2.3.4.2

UV treated samples......................................................................................................... 75

Conclusion........................................................................................................................... 77 References .......................................................................................................................... 79

CHAPTER 3 ...................................................................................................... 80 Desilication in Archaean weathering processes traced by silicon isotopes and Ge/Si ratios ............................................................................................... 80 3.1

Abstract ..................................................................................................81

3.2

Introduction .......................................................................................... 82

3.3

Methods ................................................................................................ 82

3.4

Palaeosols.............................................................................................. 83

3.5

Shales .................................................................................................... 93

3.6

Conclusions ........................................................................................... 97 References .......................................................................................................................... 98

CHAPTER 4 ..................................................................................................... 101 Stratigraphic changes of Ge/Si, REE+Y and silicon isotopes as insights into the deposition of a Mesoarchaean banded iron formation ......................... 101 4.1

Abstract ................................................................................................ 103

4.2

Introduction ......................................................................................... 103

4.3

Geological setting and sample locations ............................................ 106

4.4

Analytical techniques ........................................................................... 107

4.5

Results ................................................................................................. 108

4.6

Discussion ............................................................................................. 113

4.6.1

Clastic contamination ........................................................................................................ 113

4.6.2

A common parental fluid for Fe- and Si-rich layers ........................................................... 114

4.6.3

Constraints on the common parental fluid ....................................................................... 116

4.6.4

A common siliceous ferric oxyhydroxide precursor ......................................................... 118

4.6.5

Sedimentary-diagenetic controls of the temporal δ30Si trend.......................................... 119

4.7

Conclusion ............................................................................................ 122 References ........................................................................................................................ 123

CHAPTER 5 ..................................................................................................... 129 Secular changes in the silicon cycle along the Archaean traced by silicon isotopes and Ge/Si ratios ............................................................................... 129 5.1

Abstract ................................................................................................ 131

5.2

Introduction ......................................................................................... 132

5.3

Materials and methods ........................................................................ 133

5.3.1

Samples.............................................................................................................................. 133

5.3.2

Samples preparation and analytical techniques ............................................................... 133

5.4

Results .................................................................................................. 136

5.5

Discussion ............................................................................................. 143

5.5.1

Heavier δ30Si values in chert relative to BIF ..................................................................... 143

5.5.2

Parallel increasing δ30Si trends recorded by BIF and S-chert ........................................... 145

5.5.3

A decreasing trend in Ge/Si ratios recorded by BIF .......................................................... 148

5.6

5.5.3.1

Changes in ocean inputs................................................................................................ 149

5.5.3.2

Shifts in δ30Si and Ge/Si values of continent-derived freshwaters................................... 150

Conclusions .......................................................................................... 151

References ........................................................................................................................ 152

CHAPTER 6 ..................................................................................................... 157 General conclusions, implications and perspectives.................................... 157 6.1

Insights into continent-derived Si inputs ............................................. 159

6.1.1

Paleoclimatic implications deduced from palaeosols ...................................................... 160

6.2

Insights into Si outputs ........................................................................ 161

6.2.1

Banded iron formations..................................................................................................... 161 6.2.1.1

A general model for BIF deposition? .............................................................................. 163

6.2.1.2

Implications and perspectives for the use of Ge/Si ratio as a source tracer .................... 163

6.3

Secular changes in the Si cycle along the Archaean ........................... 164

6.3.1

Constraining relative role of seawater and high-T fluids as agent of silicification with

Ge/Si ratio ...................................................................................................................................... 165

6.4

Crucial need of constrained fractionation factors ............................... 165 References ........................................................................................................................ 166

APPENDIX

169

Abstract The external silicon cycle during the Precambrian (4.5-0.5 Ga) is not well understood despite its key significance to apprehend ancient dynamics at the surface of the Earth. In the absence of silicifying organisms, external silicon cycle dramatically differs from nowadays. Our current understanding of Precambrian oceans is limited to the assumption that silicon concentrations were close to saturation of amorphous silica. This thesis aims to bring new insights to different processes that controlled the geochemical silicon cycle during the Archaean (3.8-2.5 Ga). Bulk rock Ge/Si ratio and Si isotopes (δ30Si) offer ideal tracers to unravel different processes that control the Si cycle given their sensitivity to fractionation under near-surface conditions. First, this study focuses on Si inputs and outputs to ocean over a limited time period (~2.95 Ga Pongola Supergroup, South Africa) through the study of a palaeosol sequence and a contemporaneous banded iron formation. The palaeosol study offers precious clues in the comprehension of Archaean weathering processes and Si transfer from continent to ocean. Desilication and iron leaching were shown to be the major Archaean weathering processes. The occurrence of weathering residues issued of these processes as major component in fine-grained detrital sedimentary mass (shales) attests that identified weathering processes are widely developed and suggest an important dissolved Si flux from continent to the ocean. In parallel, banded iron formations (BIFs), typically characterised by alternation of iron-rich and silica-rich layers, represent an extraordinary record of the ocean-derived silica precipitation throughout the Precambrian. A detailed study of a 2.95 Ga BIF with excellent stratigraphic constraints identifies a seawater reservoir mixed with significant freshwater and very limited amount of high temperature hydrothermal fluids as the parental water mass from which BIFs precipitated. In addition, the export of silicon promoted by the silicon adsorption onto Feoxyhydroxides is evidenced. Then, both Si- and Fe-rich layers of BIFs have a common source water mass and a common siliceous ferric oxyhydroxides precursor. Thus, both palaeosols and BIFs highlight the significance of continental inputs to ocean, generally under- estimated or neglected, as well as the close link between Fe and Si cycles. In a second time, this study explores secular changes in the Si cycle along the Precambrian. During this timespan, the world ocean underwent a progressive decrease in 1

hydrothermal inputs and a long-term cooling. Effects of declining temperature over the oceanic Si cycle are highlighted by increasing δ30Si signatures of both chemically precipitated chert and BIF through time within the 3.8-2.5 Ga time interval. Interestingly, Si isotope compositions of BIF are shown to be kept systematically lighter of about 1.5‰ than contemporaneous cherts suggesting that both depositions occurred through different mechanisms. Along with the progressive increase of δ30Si signature, a decrease in Ge/Si ratios is attributed to a decrease in hydrothermal inputs along with the development of large and widespread desilication during continental weathering.

2

Résumé Le cycle externe du silicium au précambrien (4.5-0.5 Ga) reste mal compris malgré sa position clé dans la compréhension des processus opérant à la surface de la Terre primitive. En l’absence d’organismes sécrétant un squelette externe en silice, le cycle précambrien du silicium était vraisemblablement très différent de celui que nous connaissons à l’heure actuelle. Notre conception de l’océan archéen est limitée à l’hypothèse d’une concentration en silicium proche de la saturation en silice amorphe. Cette thèse vise à une meilleure compréhension des processus qui contrôlaient le cycle géochimique externe du silicium à l’archéen (3.8-2.5 Ga). Dans cette optique, le rapport germanium/silicium (Ge/Si) et les isotopes stables du silicium (δ30Si) représentent des traceurs idéaux pour démêler les différents processus contrôlant le cycle du Si. Dans un premier temps, cette étude se focalise sur les apports et les exports de silicium à l’océan sur une période de temps restreinte (~2.95 Ga Pongola Supergroup, Afrique du Sud) via l’étude d’un paléosol et d’un dépôt sédimentaire de précipitation chimique quasicontemporain. L’étude du paléosol apporte de précieux indices quant aux processus d’altération archéens et aux transferts de silicium des continents vers l’océan. Ainsi, la désilicification et le lessivage du fer apparaissent comme des processus majeurs de l’altération archéenne. La présence de résidus issus de ces processus d’altération en tant que composants majeurs de dépôts détritiques (shales) atteste de la globalité de ces processus et suggère des flux significatifs en silicium dissout des continents vers l’océan. En parallèle, les « banded iron formations » (BIFs), caractérisés par une alternance de niveaux riches en fer et en silice, représentent un enregistrement extraordinaire et caractéristique du précambrien de précipitation de silice à partir de l’océan. Une étude détaillée d’un dépôt de BIFs permet d’identifier une contribution importante des eaux douces dans la masse d’eau à partir de laquelle ces roches sont précipitées. Par ailleurs, un mécanisme d’export de silicium via absorption sur des oxyhydroxydes de fer est mis en évidence. Ainsi, les niveaux riches en fer et riche en silice constituant les BIFs auraient une même origine, un réservoir d’eau de mer mélangée avec des eaux douces et une contribution minime de fluides hydrothermaux de haute température, et un même précurseur commun. Dès lors, tant les paléosols que les BIFs mettent en évidence

3

l’importance des apports continentaux à l’océan, souvent négligés ou sous estimés, ainsi que le lien étroit entre les cycles du fer et du silicium. Dans un second temps, cette étude explore l’évolution du cycle du silicium au cours du précambrien. Durant cette période, l’océan voit les apports hydrothermaux ainsi que sa température diminuer. Dans l’intervalle de temps 3.8-2.5 Ga, les effets de tels changements sur le cycle du silicium sont marqués par un alourdissement progressif des signatures isotopiques des cherts et des BIFs. Le fort parallélisme entre l’évolution temporelle des compositions isotopiques des deux précipités met en évidence leur origine commune, l’océan. Cependant, les compositions isotopiques des BIFs sont systématiquement plus légères d’environ 1.5‰ que les signatures enregistrées pas les cherts. Cette différence est interprétée comme le reflet de mécanismes de dépôts différents. L’alourdissement progressif des compositions isotopiques concomitant à une diminution des rapports Ge/Si reflètent une diminution des apports hydrothermaux ainsi que la mise en place d’une désilicification de plus en plus importante et/ou généralisée lors de l’altération des continents.

4

General objectives and thesis outline

General objectives This thesis aims to bring new insights to different processes that controlled Archaean Si cycle. It addresses the following key issues: (1) Did the Archaean weathering of continents induce a significant desilication? How significant was the dissolved Si riverine outflow to the global ocean? (2) What is (are) the Si source(s) of BIF? What is their deposition mechanism? (3) Did the external Si cycle change along the Archaean? To answer these questions, we have studied (1) a palaeosol (~2.95 Ga) to get insight in Archaean weathering processes and better characterise the potential Archaean Si continental runoff, (2) contemporaneous shales to quantify the relative contributions of primary and weathering-derived minerals into the detrital inputs to the global ocean, (3) contemporaneous BIF with good stratigraphic constraints to assess the Si source(s) and the deposition mechanism of BIF, (4) BIF spanning ages from 3.25 Ga to 2.5 Ga to follow the evolution of the Si cycle using a multi-tracers approach combining silicon isotopes, Ge/Si ratios and rare earth elements (REE).

5

6

Thes is outl ine This thesis is structured into six chapters.

Chapter 1 presents the current knowledge and debates on the early Earth, gives a general overview of Archaean external silicon cycle, reviews the state of the art concerning the two major Archaean Si sinks, banded iron formations and cherts and introduces the silicon isotopes and Ge/Si ratios as tracers of the silicon cycle. Chapter 2 describes the analytical procedures optimized to provide precise δ30Si compositions, REE+Y and Ge concentrations in different types of samples (mainly banded iron formation, palaeosols and shales). Chapter 3 focuses on the Archaean weathering processes and the transfer of silicon from continents to ocean. Desilication and iron leaching were shown to be major Archaean weathering processes. The occurrence of weathering residues issued of these processes as major component in fine-grained detrital sedimentary mass (shales) attests that these weathering processes are widely developed and suggests an important dissolved Si flux from the growing continents to the ocean. Chapter 4 proposed a deposition mechanism for a 2.95 Ga banded iron formation. A common precursor for silica-rich and iron-rich layers through the silicon adsorption onto Fe-oxyhydroxides forming a siliceous ferric oxyhydroxides precursor is evidenced. Besides, the parental water mass from which banded iron formation precipitated is identified as a seawater reservoir mixed with significant freshwater component and very limited amount of high temperature hydrothermal fluids. Chapter 5 explores secular changes in the Si cycle along the Precambrian. The progressive increase in δ30Si values along with decreasing Ge/Si ratios recorded by BIFs and cherts within the 3.8-2.5 Ga time interval highlights the decrease in seawater temperature and the relative increase in continental inputs. Besides, the systematic lighter δ30Si signature recorded by BIFs compared to cherts is discussed in term of different deposition mechanisms. Chapter 6 summarises the main results and conclusions of this work and discusses the perspectives open by this study.

7

CHAPTER 1 GENERAL OVERVIEW

CHAPTER 1

General overview 1.1

The history of Earth

Typically the history of Earth is divided into two supereons: the Phanaerozoic, spanning ages from today back to 500 Ma ago, and the Precambrian that covers time from 500 Ma ago back to the Earth’s formation, about 4.56 Ga. The Precambrian, which lasts about 4.0 Ga, is subdivided into three eons: the Hadaean (~4.5-4.0 Ga), the Archaean (~4.0-2.5 Ga) and the Proterozoic (~2.5-0.5 Ga). Let us describe them in a nutshell. 1.1.1

The Hadaean (~4.56-3.8Ga)

The Hadaean begins with the Earth accretion about 4.56 Ga (Patterson, 1956) and ends with the first oldest dated rocks (~3.8 Ga). In the absence of rock records on Earth, most of our knowledge of this time span comes from studies on meteorites, the Luna as well as physical and geochemical models. The Earth was formed from the solar nebula about 4.56 Ga ago. Only about 11 Ma after the formation of the solar system, the Earth achieved 99% of its current mass and already had segregated its core (e.g., Kleine et al., 2002; Yin et al., 2002). The formation of the moon by giant impact occurred appoximately 35 Ma later (Halliday, 2004). It is of common view to imagine the Hadaean Earth as covered by a magma ocean. From rare sedimentary zircons relicts (4.4-4.0 Ga), authors suggested the presence of a significant quantity of water and a transitory proto-crust (e.g., Mojzsis et al., 2001; Caro et al., 2005; Bibikova, 2010). This suggests that temperatures were quickly low enough to allow the existence of liquid water and that the magmatic ocean cooled. Then, the Earth with a core, a mantle, oceans and an atmosphere would have already existed only 150 Ma after the Earth accretion. However, Hadaean proto-crust, oceans, atmosphere were not comparable to today’s. The end of the Hadaean is marked by an intense bombardment from meteorites called “Late Heavy Bombardment” (Gomes et al., 2005) destroying any rocks and sterilizing the Earth of any life that could have already emerged.

General Overview

11

1.1.2 The Archaean (~3.8-2.5 Ga) The Archaean period streching from 3.8 to 2.5 Ga was a key time in building-up the modern Earth. This crucial period in our Earth’s history is marked by fundamentals changes: (1) the onset of a plate tectonic; (2) the growth of continents and (3) the emergence of life. As a consequence of an intense heat production, the Archaean plate tectonic was more complex than the current one, with numerous small plates moving quite fastly. To this horizontal tectonic, a vertical tectonic driven by gravity probably surperposed (e.g., Hickman, 2004; Van Kranendonk, 2007). Although how and when tectonic occured remain hotly controversial, there is a growing concessus that a form of tectonic operated on a global scale by 2.8 Ga ago (e.g., Van Kranendonk, 2004; Smithies et al., 2005). With the end of intense bombardment at 3.8 Ga, the newly formed crust survived, stabilised and continents appeared. It has been suggested that a first supercontinent, called Vaalbara, existed around 3.3 Ga ago (e.g., de Kock et al., 2009 for a recent example). Earth was mostly covered with anoxic and acidic oceans where first life appeared and evolved. The first photosynthetic organism appeared probably 3.5 Ga ago as evidenced by controversed stromatolites-like structures (see Precambrian Research, v. 158, 2007). Dating the oldest life form is intricate but it is generally accepted that first life forms are older than photosynthetic life, perhaps many hundreds of millions of years older (Schopf, 1992). Colonies of cyanobacteria (forming stromatolites) producing oxygen through photosynthesis are regarded as the builders of our oxygenated atmosphere. However, despite cyanobacteria started to produce oxygen about 3.5 Ga ago, atmosphere remained anoxic for a long time as Earth crust, reduced volcanic gasses and oceans acted as oxygen sinks. Main evidences pointing to low oxygen levels in Archaean atmosphere are (1) the occurrence of redox-sensitive minerals such as uraninite, pyrite, siderite in Archaean sediments (e.g., Ono et al., 2000; Frimmel, 2005; Hofmann et al., 2009) ; (2) the Fe-mobility in palaeosols formed prior to 2.2 Ga (see review of Rye and Holland, 1998); (3) the preservation of large mass independent fractionation of sulfur isotopes conditioned by the absence of ozone protection from UV (e.g., Farquhar et al., 2000, 2003; Bekker et al., 2004). As this study focuses on this period, a more detailed state of knowledge will be addressed in section 1.2 of this thesis.

12

Chapter 1

1.1.3 The Proterozoic (~2.5-0.5 Ga) The Proterozoic, which lasted from about 2.5 Ga to 0.5 Ga ago, was the period of early life development and explosion. Proterozoic is characterised by three main features (1) a modern-like plate tectonic; (2) the rise of oxygen in the atmosphere and (3) the development and evolution of early life. Despite having a modern-like dynamic, the movement of plates were faster than today’s as a consequence of a hotter magma. Then, frequent collisions led to the building-up of large continents and the formation of a super-continent called “Rodinia” about 1.2-1.1 Ga ago (e.g., Piper, 1976, 2000; Dalziel, 1991; Meert and Powel, 2001). Parts of this supercontinent survived and are now pieces of North America, Australia, Western Africa and Antarctica. When the various oxygen sinks (oceans, minerals, volcanic gasses) were saturated, the photosynthetically produced oxygen begun to accumulate in the atmosphere. The rise of oxygen in atmosphere about 2.4-2.3 Ga is called the “Great Oxydation Event” (GOE) (Fig. 1.1). It marks the transition between the late Archaean and the early Proterozoic. Besides production of oxygen through photosynthesis was the primary driver of oxydation during the GOE, a number of other processes such as changes in the redox potential of volcanic gases are postulated to have played a role in changing the redox state of the oceanatmosphere system (see reviews of Holland, 2009 and Pufhal and Hiatt, 2012).

Fig. 1.1 Prevailing view of atmospheric oxygen evolution over time. The large increase at 2.4 Ga is commonly known as the “Great Oxydation Event”. From Kump, 2008.

General Overview

13

Produced oxygen acted as a toxic element for anaerobic life that was wiped out forcing the evolution through aerobic life. During the Proterozoic, life evolved from anaerobic life into aerobic life, eucaryotes, and multicellulars.

1.2

The Archaean Earth

Understanding Archaean surface environments has been challenging for decades. In order to address this issue, Archaean oceans chemistry offers useful window into past processes as they intimately link atmosphere, continents and hydrothermal fluids. The next sections describe our current state of knowledge on these Archaean reservoirs. 1.2.1 Archaean atmosphere and the “Faint Young Sun” Based on observations of solar-like stars and modeling, the solar luminosity was likely reduced by 30% 4.6 Ga ago (Gough, 1981; Sagan and Mullen, 1972). If the composition of the atmosphere had remained unchanged, Earth’s surface should have been frozen before ~2.2 Ga (Sagan and Mullen, 1972; Kasting et al., 1988). However, many evidences suggested the existence of liquid water on Earth’s surface at a very early stage in its history (e.g., Valley et al., 2002; Mojzsis et al., 2001; Nutman, 2006), perhaps as early as the late stages of accretion. This apparent contradiction between solar evolution models and geological evidence for liquid water has been called the “Faint Young Sun” paradox. The most widely accepted theory to solve this problem is the presence of enhanced concentrations of greenhouse gases in early atmosphere keeping the Earth warm. CO2 and methane (CH4) are the more likely candidates for such greenhouse gasses. However, the combination and relative abundances of these greenhouse gasses are under debate (see Fig. 1.2 for CH4 and CO2 levels proposed by Lowe and Tice, 2007). Models based on CO2 alone require CO2 concentrations well above CO2 levels deduced from sediments data (Hessler et al., 2004; Rye et al., 1995; Rosing et al., 2010). Moreover, this was challenged by Sleep and Zahnle (2001) who argued that CO2 was removed from the Archean atmosphere-ocean system by carbonitization of seafloor. Then, a combined effect of CO2 and CH4 was envisaged (Rye et al., 1995; Pavlov et al., 2000; Lowe and Tice, 2004; Kasting, 2005). This was based on the premise, now widely accepted, that levels of atmospheric oxygen were low before 2.4 Ga ago. When atmospheric O2 rose, the concentrations of

14

Chapter 1

CH4 and other reduced gases should have decreased dramatically, possibly triggering the observed glaciations (e.g., Evans et al., 1997). However, it should be emphasised that the “Faint Young Sun” problem is not yet solve and other parameters such as a lower albedo contributed to keep Earth’s climate clement. Taking into account of such parameters in models alleviate the need for extreme greenhouse gasses concentrations to satisfy the Faint Young Sun paradox (von Paris et al., 2008; Rosing et al., 2010). Lowe and Tice (2004, 2007 and references therein) evidence a close link between crustal, atmospheric, climatic and biological evolutionand propose a tectonic control.

Fig. 1.2 Schematic model of fluctuations of atmospheric pCO 2 and pCH 4 between 3.5 Ga and 2.1 Ga suggested by Lowe and Tice (2007).

1.2.2 Archaean continental crust a nd its weathering The Tonalite–Trondhjemite–Granodiorite (TTG) suite composes up to 90% of the juvenile Archaean continental crust (Martin et al., 2005). These TTG are closely associated to greenstone belts made up of komatiites, basalts and sediments. The formation of TTG and komatiites are of strong interest to comprehend the Archaean Earth as their formation is restricted to the Archaean. Processes producing TTG originate in a tectonic dynamic operating with important heat flux where the subducted oceanic crust reached melting temperatures in contrast to modern tectonic where it dehydrates inducing the melting of the above mantle wedge (Martin, 1986). Besides displaying probably different composition than the modern continental crust, the continental landmass was General Overview

15

considerably smaller at this epoch. Many models, highly debated, have been proposed for the growth of continental crust over time (e.g., Reymer and Schubert, 1984; Collerson and Kamber, 1999; Condie, 1998, 2000). As illustrated by figure 1.3, the estimated volume of the Archaean continents greatly varies depending on the model. As these models are controversial, we’ll satisfy here to state that continental inputs to oceans would therefore be considerably smaller than nowadays. However, such low continental fluxes might have been compensated by an intense source-rock weathering (e.g., Fedo et al., 1996; Sugitani et al., 1996; Lowe and Tice, 2004; Hessler and Lowe, 2006).

Fig. 1.3 Different models for the growth of continental crust over time. From Lowe and Tice, 2007.

Ancient weathering processes are not well understood because ancient weathering profiles (palaeosols) are scarce and most of them underwent diagenesis and metamorphism that changed their mineralogy. Many studies report intense source-rock weathering during Archaean (e.g., Fedo et al., 1996; Hessler and Lowe, 2006). In the absence of biological effects, the factors invoked to explain intense weathering include heavy rainfall, elevated surface temperatures and higher atmospheric pCO2 (e.g., Lowe and Tice, 2004; Hessler and Lowe, 2006). Discriminating between temperature and pCO2 as the principal factor driving weathering is not yet possible and most probably both played significant roles (Hessler and Lowe, 2006). Surface conditions on early Earth are poorly

constrained

and,

despite

extensive

studies,

uncertainties

remain.

Palaeotemperature models yield large discrepancies of the results depending on the approach implemented. Oxygen isotopic composition of cherts (and recently supported 16

Chapter 1

by silicon isotopes) provide evidence for a hot Archaean ocean (70±15°C) (Knauth and Lowe, 2003; Knauth, 2005; Robert and Chaussidon, 2006). High chemical weathering indices are generally interpreted as a result of warm climatic conditions although a causeand-effect link cannot be clearly established. However, Hofmann (2005) warned that the interpretation of weathering indices requires caution as post-depositional effects such as hydrothermal alteration may distort weathering indices. Although not precluding a hot climate, a study based on quartz weathering features moderated that Archaean climate was not necessarily hot and may be consistent with modern weathering temperature (Sleep and Hessler, 2006). Besides elevated surface temperature and high pCO2, the low atmospheric pO2 induced deep changes in the weathering of redox-sensitive minerals. Significant loss of Fe is common in Precambrian palaeosols (Rye and Holland, 1998). This contrasts with constant Fe concentrations in modern weathering profiles formed under oxic conditions, where dissolved Fe2+ is oxidized to Fe3+ and remains as Fe(III)-oxides in the profile (Maynard, 1992). It is generally considered that Fe loss is typical of subaerial weathering at low pO2 prior to 2.2 Ga (Rye and Holland, 1998). However, Fe-depletion should not be considered as a typical feature of palaeosol as diagenetic processes can mimic this Fe depletion even under an oxidizing atmosphere (Ohmoto et al., 1996). In contrast, redox-sensitive minerals in a reduced state (e.g., pyrite, uraninite, siderite) remain stable under Archaean atmosphere with low pO2. This contrasts with modern oxidative weathering considered as one of the most important sulfate sources to modern oceans. This input of sulfate to oceans likely not operated during Archaean time, leading to low sulfate contents in the Archaean oceans. To sum up, Archaean continental inputs to oceans differ from nowadays through (1) intense weathering producing important flux of labile elements, typically Ca, Na, Sr; (2) reducing conditions that have promoted large Fe soils outflows but low sulfate fluxes. However, riverine supplies to the global ocean might have been partly counterbalanced by the smaller size of the continental landmass (e.g., De Wit and Hart, 1993; Collerson and Kamber, 1999; Condie, 2000).

General Overview

17

1.2.3 Archaean hydrotherma lism As nowadays the excessive heat of the mantle must have been dissipated through conduction, convection associated to the generation and subduction of oceanic crust and, convection of seawater into the oceanic crust. In response to greater heat fluxes, the total length of medio-oceanic rides (MORs) might have been twice as long as today (Isley, 1995). This might have culminated in an order of magnitude larger hydrothermal activity in the Archaean (De Wit and Hart, 1993). Although more vigorous, sea-floor hydrothermal systems appear to have operated in a relatively constant way throughout the Earth's history: early to mid-Archaean environments can be interpreted in terms of present-day settings and chemistry (e.g., de Ronde et al., 1994, 1997; de Vries and Touret, 2007; Hofmann and Harris, 2008). Then, most models rely on the assumption that chemical properties of Archaean hydrothermal fluids were similar to those of modern hydrothermal fluids. However, recent thermodynamic models pointed that this might not be valid (Shibuya et al., 2010; Wang et al., 2009). Shibuya and coworkers (2010) predicted that if the Archaean seawater was slightly acidic, CO2-rich and SO4-poor, high temperature alteration processes of the oceanic crust would have led to the generation of highly alkaline (pH>10), Si-enriched and Fe-depleted hydrothermal fluids. In contrast Wang et al. (2009) pointed out that alkaline, Si-Fe-enriched hydrothermal fluids can be merely formed from the hydrothermal leaching of komatiites. Except Shibuya et al. (2010), most studies suggest that because of reducing conditions and low sulfate concentrations of Archaean seawater, the hydrothermal delivery of Fe to the ocean was higher (Walker and Brimblecombe, 1985; Isley, 1995; Canfield, 1998; Kump and Seyfried, 2005) even under relatively low-temperature conditions (Hofmann and Harris, 2008). Therefore, hydrothermalism must have represented an important Fe flux to Archaean oceans where Fe must have been accumulated in response to the reducing conditions of the deep ocean.

18

Chapter 1

1.2.4 Archaean ocean From previous sections, it appears that many parameters controlling the chemistry of the ocean were dramatically different during the Archaean time span. Although not exhaustive, following dramatic changes between modern and Archaean conditions can be cited: (1) Archaean atmosphere was CO2-rich and O2-poor (see section 1.2.1); (2) surface temperatures were higher increasing solubility of elements (see section 1.2.1); (3) the continental fluxes were smaller (see section 1.2.2); (4) the mantle-derived inputs were likely more important (see section 1.2.3); (5) chemistry of inputs to oceans evolved through time (see section 1.2.2 and 1.2.3); (6) biogenic processes were not yet operating, or at least in a different way. Therefore, it seems difficult to imagine how Archaean ocean may not have been very different in composition from the modern ocean. In broad outline, following assumptions can be made. The relative balance between mantle-derived vs. continental-derived inputs has been the subject of much debate. Veizer et al. (1982) popularized the view that Archaean balance between continental and hydrothermal inputs could have been reversed relative to present-day with Archaean ocean dominated by hydrothermal inputs. This sounds consistent with on the one hand, reduced continental inputs reflecting the smaller continental volume at the time and, on the other hand, higher hydrothermal fluxes related to greater mantle heat flow. Numbers of studies attempted to better quantify the relative contributions using geochemical and isotopic proxies (e.g., Derry and Jacobsen, 1990; Bau et al., 1997). Recently, some authors argued that the strong dominance of hydrothermal inputs into Archaean ocean cannot be simply explained by increased Archaean hydrothermal activity and reduced continental volume (Kamber and Webb, 2001). Taking into account reduced continental inputs, they calculated that hydrothermal flux should have been ten times greater than today’s. Considering such an increase as unrealistic, Kamber and Webb (2001) further invoked a reduced erosion rate in addition to the smaller continental volume compared to nowadays. Enrichment of CO2 in the atmosphere arguably had pronounced effects on the hydrosphere because high pCO2 tends to lower the pH of solutions in contact with the atmosphere. Some studies proposed that CO2 concentrations of Archaean atmosphere might have been up to 300 times the Present Atmosphere Level (PAL) (Kasting, 1993;

General Overview

19

Lowe and Tice, 2004). Such high CO2 levels in the atmosphere would have induced a pH of the oceans between 5.4 and 8.6 (Grotzinger and Kasting, 1993). However, debates remain on how high were atmospheric CO2 levels (Kasting, 2005; Lowe and Tice, 2004; Sumner and Grotzinger, 2004; Catling et al., 2001). As the level of oxygen in atmosphere has risen over time, it is expected that the oceans too will encompass a progressive oxidation. Then, during Archaean oceans are considered to be at a reduced state in response to low atmospheric pO2. However, debate remains on the possible existence of a redox-stratified ocean where surface waters could have been oxidized. A lack of oxygen, at least in watermass below the hypothetic oxycline, allows iron to remain in its reduced state. Combined with very low levels of sulfates (very small inputs from rivers in the absence of oxidative weathering of sulfur), great amounts of iron tend to accumulate in Archaean oceans. The most widely accepted view is to assume that the Archaean oceans were anoxic and acidic with lower carbonate and sulfate contents but higher concentrations of Ca, Fe, Ba, Si, Na, Cl compared to modern oceans.

1.3

The Archaean silicon cycle

Our current understanding of Precambrian oceans is limited to the assumption that Si concentrations were close to saturation of amorphous silica (Siever, 1992). Due to the strong temperature-dependent solubility of silica in pure water (Fig. 1.4) (Gunnarsson and Arnorsson, 2000) and the absence of a consensus on temperature prevailing in Archaean times, Si concentrations of Archaean oceans remain poorly constrained. Depending on assumed temperature, Si concentrations estimates range between 30 and 60 ppm (e.g., Morris, 1993; Siever, 1992), which implies a long residence time on the order of 105 years in the Precambrian ocean (Siever, 1992).

20

Chapter 1

Fig. 1.4 Temperature-dependent solubility curves of quartz and amorphous silica at P sat. Calculated from Gunnarsson and Arnorsson, 2000.

The origin of Si in the Precambrian ocean is controversial (Maliva et al., 2005). As nowadays Si inputs were likely from hydrothermal fluids and continental-derived freshwaters. Besides a presumably reverse balance between hydrothermal and continental fluxes during the Precambrian (e.g., Bau and Möller, 1993; Bau et al., 1997; Kamber and Webb, 2001), Si concentrations of both fluxes may have been different as a result of contrasted environments. The Si concentration of modern hydrothermal fluids along modern mid-oceanic spreading centres is generally considered to be controlled by equilibrium with quartz at temperatures ~200-400°C and pressures of 100-500 bars in hydrothermal convection cells (Von Damm et al., 1985, 1991). As various pressure and temperature conditions are encountered, hydrothermal fluids may thus display different Si concentrations. Modern hydrothermal fluids have Si concentrations ranging from ~450 to ~650 ppm (Mortlock et al., 1993). According to recent studies, silicon in Archaean hydrothermal fluids can theorically reach concentration as high as 1680 ppm or 3000 ppm, which is at least 4 to 7 times higher than many modern hydrothermal fluids (Wang et al., 2009; Shibuya et al., 2010; see section 1.2.3). River influx of dissolved silicon to the Archaean oceans would strongly depend on the surface of emerged continents exposed to weathering and the weathering regime. Most

General Overview

21

of crustal growth models agree with a smaller volume of continental crust during Archaean compared to nowadays. Although weathering regimes might have been more aggressive (Fedo et al., 1996; Sugitani et al., 1996; Lowe and Tice, 2004; Hessler and Lowe, 2006), potentially increasing the dissolved Si load in rivers, it seems reasonable to assume that Si continental inputs were probably by far less than present-day due to small volume of continental crust. However, despite variances affecting Si inputs, the pivotal change between modern and Archaean Si cycles are the Si outputs. In the absence of biologically-mediated precipitation, the most likely Si removal processes are (1) the direct precipitation of silica minerals and (2) the pervasive silicification of volcanoclastic sediments and their crystalline seafloor basements. Sorption of silicon on clay minerals, Fe oxides, or organic matter could also have played a role in nucleation and/or precipitation (e.g., Perry and Lefticariu, 2003; Siever, 1992; Fischer and Knoll, 2009). Then, cherts formed either by direct precipitation (C-cherts) or silicification of a rock precursor (S-cherts) as well as cherts of banded iron formations (BIF) likely represent dominant sinks of silica in Precambrian oceans. However, for both cherts and BIFs, the source(s) of Si (continentalderived freshwaters, hydrothermal fluids, seawater or a mixture of these reservoirs) as well as their precipitation mechanisms remain unclear. Therefore the impact of their deposition on the silicon cycle is not fully understood.

22

Chapter 1

1.4

Banded Iron Formations

Banded iron formations (BIFs) are chemical precipitates characterised by the alternation of iron-rich and silica-rich layers with three scales of banding: microbands (1m) (Trendall and Blockley, 1970) (Fig. 1.5). According to the definition of James (1954), BIF typically BIF contain more than 15 wt.% of iron.

Fig. 1.5 Illustrations of the BIF banding scales. A- Outcrop of the ~2.95 Ga Mozaan BIF in the White Umfolozi Inlier (South Africa) where the macrobanding (~1m) is visible. B- Sample from the ~2.95 Ga Witwatersrand Supergroup with alternating cm-scale layers of jasper (red layers) and Fe-oxides layers (black layers); C- Sample from the ~2.5 Ga Transvaal Supergroup with alternating cm- and mm-scale layers of cherts (white layers) and Fe-oxides layers.

BIFs are largely restricted to the Precambrian with most occurrences ranging between 3.8 Ga and 1.8 Ga (Fig. 1.6). The peak in BIF deposition about ~2.5 Ga appears to correlate with major changes in the Earth’s history such as the rise of atmospheric oxygen and the change from anoxic to oxic conditions in the ocean (Canfield, 2005; Holland, 2006) (Fig. 1.6).

General Overview

23

Fig. 1.6 Schematic diagram showing the relative abundance of Precambrian BIFs through time with calculated curves for the atmospheric evolution of oxygen and carbon dioxide (Kasting,2001, 2004; Kasting and Catling, 2003; Pavlov and Kasting, 2002). Estimated abundances are relative to the Hamersley Group BIF volume taken as a maximum. From Klein, 2005.

Their abundance in the Archaean/early Proterozoic eons and their absence thereafter clearly reveal different Si and Fe cycles before 1.8 Ga ago. As BIFs represent an important sink of Si during this time span, it is of great interest to better understand the BIF deposition mechanism which has not yet reach a consensus (see section 1.4.3). Before exploring different BIF deposition mechanisms that have been proposed, a brief look at deposition settings and mineralogies is instructive. 1.4.1 Depositional settings Based on the BIF size and lithologic associations, two type of BIF are distinguished: Algoma- and Superior-type BIF (Gross, 1965). Algoma-type BIFs are relatively small, associated with volcanogenic rocks and hosted in greenstone belts. Typical lateral extensions are less than 10 km and thickness is less than 50 m. However, these characteristics do not indicate that all Algoma-type BIF were originally small, as most have been affected by deformation and tectonic dismemberment, implying that their true size and extent are likely underestimated in global compilations of BIF (Gole and Klein, 1981). Reconstruction of the original depositional settings very difficult but favored depositional environments for this type of BIF include island arc/back arc basin (Veizer, 1983) and intracratonic rift zone (Gross, 1983). BIF of Isua (~3.8 Ga, West Greenland) and Barberton (~3.5-3.1 Ga, South Africa) greenstones belts are typically representative of Algoma-type BIF. Superior-type BIFs are larger and associated with other sedimentary units. Several 24

Chapter 1

BIFs classified as Superior-type BIF have extension over 105 km² (Beukes, 1973). Deposition occurred in relatively shallow marine conditions under transgressing seas, perhaps on the continental shelves of passive margins (Gross, 1965). Superior-type sequences include the Hamersley Group (~2.5 Ga, Western Australia), the Pongola, Witwatersrand and Transvaal Supergroups (~2.5 Ga, South Africa). Note that a third type, the Rapitan-type, was defined for Neoproterozoic BIFs associated to glacio-marine deposits and resulting from anoxic conditions caused by an ice-covered ocean, the “Snowball Earth” (Gross, 1973). As this thesis focus on the Archaean, this type will not be considered further.

Fig. 1.7 Secular distributions of mantle superplume breakout events (Abbott and Isley, 2002) and BIF deposits. (From Bekker et al., 2010)

Isley (1995) and Isley and Abbot (1999) highlighted a correlation between major BIF deposition events and global plume events, suggesting a causal link between both (Fig. 1.7). Mantle superplume volcanism may promote BIF deposition by (1) increasing the Fe flux through continental weathering and/or through submarine hydrothermal processes, (2) increasing the number of tectonic environments appropriate for BIF deposition, (3) promoting global anoxic, Fe-rich hydrothermal plumes in the shallow to intermediate marine water column (Isley and Abbot, 1999). However, Huston and Logan (2004) pointed out that the relation between Superior-type BIFs and global mantle plume events is not direct. Then, Algoma-type BIFs were regarded to reflect intrabasinal pulses of magmatic and hydrothermal activity during the deposition of volcano-sedimentary General Overview

25

greenstone successions. Superior-type BIFs are considered to have formed on continental shelves during periods of global high sea levels and pulses of hydrothermal activity. If this model is correct, then a distinction between Superior- and Algoma-types of BIFs has some merit. Indeed, the geochemistry of Algoma-type BIFs may reflect local hydrothermal conditions whereas Superior-type BIFs may record the large-scale chemistry of the oceans during their formation, although a potential influence by nearby is not excluded. Although in some cases it is difficult to resolve whether a BIF belongs to Algoma- or Superior-type, Algoma-type BIFs are more common in Archaean (>2.5 Ga) whereas Superior-type BIFs appeared in the record at ca. 3.3 Ga but become significant at about 2.6 Ga ago (Fig. 1.8).

Fig. 1.8 Variations in the abundance of Algoma- and Superior-type BIF through time (from Huston and Logan, 2004).

The secular changes in the style of BIF deposition may reflect a higher mantle heat flux and a limited occurrence of continental shelves during Archaean favoring Algoma-type BIF. Increasing number of continental shelves after a major plume event at 2.75-2.70 Ga that lead to important continental crust formation allowed the widespread deposition of Superior-type BIF (Clout and Simonson, 2005; Bekker et al., 2010). There is a global consensus to attribute the end of BIF deposition at about 1.8 Ga as a consequence of the oxygenation of the atmosphere (e.g., Huston and Logan, 2004).

26

Chapter 1

1.4.2 Mineralogies Mineralogically, Algoma- and Superior-type BIF are similar. Most BIF have undergone various grades of metamorphism from the late-diagenetic to the amphibolites facies metamorphic grades. Major mineral phases in late-diagenetic and low-grade metamorphic BIFs are magnetite, hematite, pyrite, greenalite, stipnomelane and minnesotaite and carbonates including siderite and members of dolomite-ankerite series (e.g., Klein, 2005). Chert (mainly microcrystalline quartz) is ubiquitous and is almost always present. Depending on the major Fe-bearing mineral phase, iron formations are classified as oxide, carbonate, silicate and sulfide facies with several mixed facies (James, 1954; Klein and Beukes, 1993). The sulfide facies should better be considered as pyritic carbonaceous shale or slate than BIF (Bekker et al., 2010). Original precipitates of BIFs have not yet been unequivocally identified. One the one hand, some consider that chert layers formed by recrystallisation of an amorphous silica precursor (e.g., Hamade et al., 2003; Maliva et al., 2005; Posth et al., 2008; Steinhoefel et al., 2009; Wang et al., 2009). On the other hand is the hypothesis of a common precursor such as a siliceous ferric oxyhydroxide gel or Al-poor hydrous iron silicate mud (e.g., Lascelles, 2007; Fischer and Knoll, 2009). 1.4.3 Deposition mecha nisms Banded iron formations represent a style of sedimentation for which clear modern analogues do not exist, at least at a similar scale. The controls on the deposition of BIFs have long been contentious and an accepted global theory of BIF deposition is still lacking. However, it is also unclear whetever a unique theory would accomodate the diversity of BIF as their proposed deposition settings as well as mineralogies are varied. Controversy still surrounds: (1) the processes of iron precipitation; (2) the origin of chert; and (3) depositional systems.

General Overview

27

1.4.3.1 Iron oxidation processes Most studies on BIFs have focused on the origin of Fe-rich bands. It is generally accepted that Fe2+ was sourced from hydrothermal alteration of oceanic crust (e.g., Jacobsen and Pimentel-Klose, 1988; Derry and Jacobsen, 1990; Bau and Möller, 1993). Another common assumption is that the oceans (at least at depth) were anoxic and therefore capable of transporting and concentrating Fe2+ dissolved in seawater (e.g., Cloud, 1968; Holland, 1973; Drever, 1974). In addition to low oxidizing potential and high hydrothermal iron flux, low marine sulfate and sulfide concentrations (Habicht et al., 2002) allowed an amplified marine reservoir of dissoved iron. Proposed mechanisms of oxidation of Fe2+ and subsequent precipitation of ferric iron are still debated (e.g., Ohmoto et al., 2006; Beukes and Gutzmer, 2008; Bekker et al., 2010). Three main models have been proposed and are briefly described below.

Oxidation of Fe(II) by cyanobacterially-generated O2 This widely quoted model involves the inorganic oxidation of dissolved Fe(II) with photosynthetically procuded oxygen by cyanobacteria (Cloud, 1965, 1973): 4Fe2+ + O2 + 10H2O  4Fe(OH)3 + 8H+ Under an anoxic atmosphere, the produced oxygen would have confined to localised “oxygen oases” associated with cyanobacterial blooms or formed a stratified ocean with a thin oxic zone overlying an anoxic water column (e.g., Klein and Beukes, 1989). Main rebuttal against this hypothesis is that most Archaean and early Paleoproterozoic hydrogeneous sediments (e.g., carbonates and BIFs) do not show geochemical evidence of a distinct redoxcline even in shallow environments (e.g., Kamber et al., 2004; Alexander et al., 2008; Planavsky et al., 2010). This has recently been used to question the iron oxidation through mixing of oxic and anoxic waters (e.g., Planavsky et al., 2010).

Metabolic iron oxydation Despite the absence of direct evidence, it is becoming increasingly accepted that anoxygenic phototrophic bacteria (photoferrotrophy) were directly involved in the primary oxidation of Fe(II) to Fe(III) in BIF (Konhauser et al., 2002; Kappler et al., 2005; 28

Chapter 1

Posth et al., 2008). This mechanism is strengthen by the identification of modern analogue marine and freshwaters anoxygenic Fe(II)-oxidizing phototrophs (Crowe et al., 2008; Walter et al., 2009). In the past 20 years, a number of experimental studies have confirmed that various purple and green phototrophic bacteria can use Fe(II) as electron donor for CO2 fixation (Widdel et al., 1993; Heising et al., 1999; Straub et al., 1999): 4Fe2+ + CO2 + 11H2O + hy CH2O + 4Fe(OH)3 + 8H+ More recently, it has been demonstrated via experiments and calculations that these organisms would have been capable of oxidizing enough Fe(II) to account for the sedimentary ferric iron flux required to produce large BIF deposition (Konhauser et al., 2002; Kappler et al., 2005). Critical to this hypothesis is the absence of unequivocal evidence for the existence of Fe(II)-oxidizing phototrophs during Archaean.

UV photooxidation of Fe(II) UV-induced ferrous oxidation has also been advanced as an important mechanism for the oxidation of soluble Fe(II) (Cairns-Smith, 1978; Braterman et al., 1983): 2Fe2+ + 6H2O + hy → 2Fe(OH)3 + 2H2 + 4H+ Although Fe(II) can be oxidized photochemically in simple aqueous systems (Braterman et al., 1983), the efficiency of UV-dependent oxidation in more complex environments such as seawater has been questioned (e.g., Konhauser et al., 2007). 1.4.3.2 Mechanisms of silica precipitation Proposed deposition mechanisms include (1) direct precipitation of amorphous silica from seawater or hydrothermal fluids, (2) the precipitation of colloidal hydrous silicates (nontronite), (3) Si adsorption onto Fe-oxyhydroxides. In the absence of silica-secreting organisms, the Precambrian ocean may have been close to saturation with respect to amorphous silica (Siever, 1992). It then has often been suggested that an amorphous silica gel precipitated as a result of evaporative concentration (e.g., Trendall and Blockley, 1970; Garrels, 1987; Maliva et al., 2005). Considering a hydrothermal Si source, precipitation is trigger by the cooling of the fluid in contact with seawater. In these cases, chert precursor is of primary origin and its General Overview

29

recrystallisation into microquartz through dissolution-precipitation processes occurred during diagenesis (Knauth, 1994; Marin et al., 2010). Alternatively, Si reacts with Fe-hydroxides to form hydrous Al-poor Fe-rich silicates (Lascelles, 2007) or adsorbs onto Fe-oxyhydroxides forming a siliceous ferric oxyhydroxide precursor (e.g., Konhauser et al., 2005; Fischer and Knoll, 2009). The cherts of BIF should then be considered as diagenetic in origin. Then, both the parental water mass (seawater or hydrothermal fluids) and the primary or diagenetic origin of chert layers of BIFs remain debated (see section 1.6.4). 1.4.3.3 BIF deposition mechanisms As debates on the oxidation processes, the origin of silica and the primary precursor(s) are still open, the deposition and banding mechanisms also remain discussed. Main hypothesis are summarized below and schematised in Table 1.1.

Fe-oxidation processes Si source Chert-layers origin References O2 cyanobacteria photoferrotrophs oversaturation hydrothermal seawater primary diagenetic x

x x

x

x x x

x x

x x x

x x

x x

1-5 6-7 8 9 10

Table 1.1 Schematisation of different BIF deposition models depending on the Fe-oxidation process, Si source and the primary of diagenetic origin of chert layers. Beukes et al., 1990 (1); (2) Klein and Beukes, 1989; (3) Isley (1995); (4)Hamade et al. (2003); (5) Steinhoefel et al. (2009); (6) Lascelles (2007); (7) Krapez et al. (2003); (8) Wang et al. (2009); (9) Fischer and Knoll (2009); (10) Posth et al. (2008).

Two interacting water masses The most well-known model invokes the interaction of two water masses. The formation of Fe oxides layers is interpreted to reflect periods of intensive upwelling of Fe(II)-rich deep reduced waters or hydrothermal plumes into the photic zone of near-coastal waters where cyanobacteria produced oxygen, which in turn led to the precipitation of a ferric oxyhydroxide precursor (e.g., Klein and Beukes, 1989; Beukes et al., 1990; Isley, 1995; Hamade et al., 2003; Beukes and Gutzmer, 2008; Steinhoefel et al., 2009) (Fig. 1.9). Different Fe mineralogies (oxides, silicates, carbonates and sulfides) are then interpreted to reflect a redox gradient in the deposition area (Holland, 2005). Long-distance transport

30

Chapter 1

of hydrothermal solutes during oceanic anoxia (Isley, 1995) may have enabled BIF deposition on continental shelves during high sea level (Simonson and Hassler, 1996). Banding in iron formation has been argued to reflect alternating chemical precipitation of colloidal iron oxyhydroxides and silica (Garrels, 1987) or continuous evaporative silica precipitation with episodic deposition of iron (Morris, 1993). Chert layers would then record periods of hydrothermal quiescence and Fe-poor sedimentation (e.g., Morris, 1993; Steinhoefel et al., 2009).

Fig. 1.9 Simplified model for the BIF deposition implying two interacting water masses. Upwelling of a reduced Fe-rich hydrothermal plume or bottom seawater into the photic zone where cyanobacteria produced oxygen inducing the precipitation of Fe-hydroxides.

A common parental water mass and two distinct precursors Considering BIFs as analogous to modern iron-rich sediments precipitated from deep-sea smokers, Krapež et al. (2003) and Lascelles (2007) infer that both Si and Fe have a hydrothermal origin and that the precursor sediment to BIFs was granular iron-rich hydrothermal muds deposited on the flanks of submarine volcanoes. The model of Lascelles (2007) suggests that hot Fe- and Si-rich hydrothermal fluids are rapidly cooled on contact with cold ocean water, reducing the solubility of elements and producing the precipitation of colloidal particles of Al-poor hydrous iron silicate (nontronite) and iron hydroxides (Fig. 1.10 A). The rapid deposition and abundant included water formed unstable mounds around the vents (Fig. 1.10 B). Slumping of the mounds caused by General Overview

31

compaction, dewatering, gravity sliding and seismic events produced turbidity and density currents (Fig 1.10 C and D).

Fig. 1.10 Deposition of BIF according to the model of Lascelles (2007). A- Cooling of hydrothermal fluids releasing Fe2+ and H4SiO4 into ocean induced precipitation of hydrous iron silicate and iron hydroxides around vent. B-Slumping of the mounds creates turbidity and density currents C and D-deposition from turbidity and density currents. From Lascelles, 2007

Deposition from the density currents formed typical finely laminated deposits that subsequently underwent diagenesis. During diagenesis, hydrous iron silicates tend to dissociate into iron oxides and colloidal silica giving rise to the bands of chert (Fig. 1.11). The cherts of BIF are then considered as diagenetic in origin and developped during burial.

Fig. 1.11 Diagenesis of BIF considering hydrous iron silicates and Fe-hydroxides as precursors. ADeposition of iron silicates with continuous and discontinuous laminas of iron oxides; B-dissociation of iron silicate into iron oxides and colloidal silica; C-differentiation of mesobands by settling of iron oxides and accumulation of silica below oxide layers. From Lascelles, 2007

32

Chapter 1

Posth et al. (2008) proposed a radically different model for BIFs deposition despite convergences of a common parental water mass for both Si and Fe and a precipitation of these elements via their own precursors. Invoking photoferrotrophs as responsible for oxidation of the Fe in seawater, they demonstrated that the rate of ferric hydroxides formation by iron-oxidizing microbes is temperature-dependent and decoupled from silica. They showed that phototrophic Fe oxidation rate has an optimal temperature range. At temperature above or below this range, precipitation of iron hydroxides would slower (or cease) while the abiotic silica precipitation induced by silica saturation would continue (or increase in case of temperature drop). Then, natural fluctuations in the temperature of the oceanic photic zone were proposed to have led to the primary layering. A common parental water mass and a common precursor Alternatively to direct Si precipitation from seawater, Fischer and Knoll (2009) proposed a model where silicic acid adsorbed onto Fe-hydroxide (Fig. 1.12). They consider the Fe oxydation by anoxygenic photosynthesis where the ferric iron produced would rapidly undergo hydrolysis and precipitate as ferric hydroxides. Dissolved silica readily adsorbs to the ferric hydroxides surface, generating siliceous ferric hydroxides precursor that sank to the sea floor along with organic matter. In sediments, bacterial oxidation of organic matter would induce Fe reduction thus liberating silica. This silica would concentrate into pore fluids and ultimately precipitated and transformed into early diagenetic chert.

Fig. 1.12 Schematic model of BIF deposition according to the model of Fischer and Knoll (2009)

General Overview

33

Based on thermodynamic calculations, Wang et al. (2009) suggested that banding is produced by positive feedbacks that occur among the chemical reactions when hydrothermal fluids mix with ambient seawater leading to a self-organized oscillatory precipitation. A prerequisite is that hydrothermal fluids are equally Si- and Fe-rich. They showed that such fluid can be generated by the hydrothermal leaching of komatiites (limited to the Precambrian record). In this model, Fe oxidation may be linked to anoxigenic Fe(II)-oxidizing microbes or O2 generated by cyanobacteria.

1.5

Precambrian cherts

Chert is a chemically precipitated sedimentary rock, which is basically composed of silica (~ 90 %) under diverse crystalline forms: mainly of microcrystalline quartz (microquartz), chalcedonic quartz, megaquartz (Folk, 1980). In some cases, cherts may contain minor impurities of carbonates, siderite, iron oxides, sulfurs and sulfates (Knauth, 1994). Unusual variety and ubiquity of chert deposits are typical in the Precambrian sedimentary record. However, the provenance of silica and the deposition mechanism and environment remain subject to debate (e.g., Knauth and Lowe, 2003; Hofmann and Harris, 2008; Van den Boorn et al., 2007, 2010; Abraham et al., 2011). The silica may have precipitated from hydrothermal fluids as a result of a derease in silica solubility (1) consecutive to cooling of hydrothermal fluids during ascent thourgh the crust; (2) by mixing with cooler seawater at the sediment-seawater interface. As Archaean seawater was likely close to the silica saturation level, slight changes in temperature may also have induced precipitation of silica. To decipher between these processes, key issues are (1) the role of ambient seawater versus hydrothermal fluids and (2) the relative importance of direct precipitation against pervasive silicification of precursor material. A step foward was achieved through recent silicon isotopic studies (Van den Boorn et al., 2007, 2010; Abraham et al., 2010; Marin-Carbonne et al., 2011, 2012) (see section 1.6.4.2).

34

Chapter 1

1.6

Tracers of silicon cycle

From the above sections (1.4 and 1.5), it is clear that for both BIF and chert deposits, the origin of silica is central to the debate. In addition, for BIF the precursor (amorphous silica or siliceous ferric oxyhydroxide) remains to be determined. As silicon is the most common non-volatile element on Earth, tracers are needed to study Si pathways (Conley, 2002; Meunier, 2003). Recent advances in analytical capabilities have raised Si stable isotopes and germanium to silicon ratio (Ge/Si) as very useful tracers of the Si cycle. These tracers are affected by biological, chemical and/or physical processes under natural conditions. 1.6.1 Silicon stable isotopes Atoms of the same element can have a different number of neutrons (n). The different configurations for one element are called “isotopes”. Having the same number of protons (z) and electrons and configuration of the electron shale, isotopes display similar chemical properties. Because isotopes differ in mass (m=n+z), atoms of different isotopes react at different rates that result in partial separation of the light isotopes from the heavy isotopes during chemical reactions and physical processes (although all isotopes of one element take part in the reaction). This mass-dependent partitioning of isotopes between two substances or two phases is called “isotopic fractionation”. Isotopic fractionations are mainly of two types, either at kinetic or at thermodynamic equilibrium. Kinetic fractionations are usually associated to unidirectional reactions, e.g., the product formed does not react backwards with the reactant. Since the bonds involving light isotopes in the reactants are more easily broken, light isotopes react faster than heavy isotopes (Hoefs, 2009). Then, the reaction product is enriched in light isotopes whereas heavy isotopes preferentially stay in the initial reservoir. On the other hand, thermodynamic equilibrium fractionations require exchange reactions between two or more species in chemical equilibrium, meaning that products can react with the reactant. Associated to bidirectional reactions, equilibrium isotope fractionation is strictly governed by relative mass differences among different isotopes of an element. Heavy isotopes will be concentrated in the phase with the lower energy state (Hoefs, 2009).

General Overview

35

Silicon has three stable isotopes of atomic mass units (amu) (28.976495), and

28

Si (27.976927),

29

Si

30

Si (29.973770) with an average abundance of 92.23%, 4.67%, 3.10%,

respectively (Faure and Mensing, 2005). The isotopic composition of a sample is usually expressed as deviation of ratios of the heavier isotopes to the more abundant lighter isotope (e.g., 30Si/28Si) relative to the same ratio within a reference material. In the case of silicon, this reference material is the international standard quartz NBS-28. For a convenient scale, variations are expressed part per thousands (‰) using the δ-notation as δ30Si or δ29Si:   Si   

 

  Si   

 

30

29

30 30

29 29

Si

28

Si

28

Si

28

Si

28

 Si  Si

sample

NBS 28

 Si  Si

sample

NBS 28

  1  1000 

  1  1000 

A positive delta value (δ>0‰) means that the sample is enriched in the heavy isotope compared to the reference standard (heavier signature), while a negative delta value (δ0‰) precipitates silica with large initial fractionation (Δ30Si~1/T²). Consecutive to increasing temperature downward enhancing the saturation concentration level, silica precipitation stops and basalt dissolution starts producing a Si-rich high-T fluid with a basaltic δ30Si signature (-0.3‰; Abraham et al., 2008, 2011). In a second step along the upward path, this high-T fluid precipitates silica as it cools down with initially weak Δ30Si.

General Overview

45

References Abbott a nd I s le y (2002). The inte ns ity, occu r rence , a nd du ra tion of su pe rpl u me e ve nts a nd e ras ove r g e ol og ica l time . Journal of Geod ynam i c s, 34, 265 −307. Abra ha m, K. , Opfe rge l t, S. , F ripia t, F ., Ca va gna , A. , de Jong , J. , F ol ey , S. , André , L. , Ca rdina l , D. , (2008). δ 3 0 Si a nd δ 2 9 Si De te rmina tions on U SGS B HVO- 1 a nd B HVO- 2 Re fe re nce Ma teria l s with a Ne w Config u ra tion on a Nu Pl a s ma Mul ti - Col l e ctor I CP- MS. Geost and . Geoanal. Res. 32, 1 93- 202. Abra ha m, K., Hofma nn, A. , F ol ey , S. F ., Ca rdina l , D. , Ha rris, C. , Ba rth, M. , André , L . , (201 1 ). Cou ple d s il icon - oxyg e n is otope fra ctiona tion tra ce s Archa e an s il icifica tion. Eart h Pl anet . Sc i . Let t . 301, 222- 23 0. Al e xa nde r, B. , Ba u , M. , Ande rss on, P. , Dul sk i, P. , (2008). Contine nta l ly - derive d s ol u tes in s ha ll ow Arche an s ea wa te r: Ra re e a rth e l eme nt a nd Nd is otope e vide nc e in iron forma tion fro m the 2. 9 Ga Pong ola Su pe rg rou p, Sou th Africa . Geoc hi m . C osm oc hi m . Act a. 7 2, 378- 394. Al l e man, L . Y. , D. Ca rdina l , C. Cocqu y t, P. - D. Pl is nie r, J. P. Des cy , I. Kimire i, D. Siny inza , L . André, (2005 ). Sil icon is ot opic fra ctiona tion in L a k e Ta nga ny ik a and its ma in tribu ta rie s , J. Great Lakes Res. , 31, 5 09- 5 1 9. André , L. , Ca rdina l , D., Al le ma n, L . , Moorba th, S. , 2006. Sil icon is otope s in~ 3. 8 Ga We s t Gree nl and rock s as clu e s to the Eoa rcha e an s u pra cru s ta l Si cy cl e. Ea rth Pl a ne t. Sci. Le tt . 245 , 1 62- 1 73. Army ta ge R. M. G. , R. B . Ge org , P. S. Sa va ge , H. M. Wil l ia ms , A. N. Ha l l iday (201 1 ) Sil icon is otope s in me te orite s a nd pl a ne ta ry core forma tion, Geoc hi m . C osm oc hi m. Ac t a, 7 5, 3662- 3676. B a u M. and Mol l e r P. (1 993). Ra re e a rth e l e ment s y s te ma tics of the che mica l l y pre cipita te d compone nt in e a rl y Pre ca mbria n iron forma tions a nd the e vol u tion of te rre s tria l a tmos phe re hy dros phe re - l ithos phe re sy s te m. Geoc him i c a et C osm oc hi m i c a Ac t a, 57, 2239- 2249. B a u M. , Höhndorf A. , Du l s k i P. , and B eu ke s N. J. (1 99 7) Sou rce s of ra re - e a rth el e ments a nd iron in Pa l e oprote rozoic ir on - forma tions from the Tra ns va a l Su pe rg rou p, Sou th Africa : Evide nce from ne ody miu m is oto p e s . J. Geol . , 105, 1 21 –1 29. B a re ill e , G., La bra che rie , M., B e rtrand, P. , La be ry ie, L. , L a va u x, G. , and Dig na n, M. (1 998). Gl a cia l - inte rg la cia l cha nge s in the a ccumu l a tion ra tes of ma jor bi og e nic compone nts in Sou the rn I ndia n Oce an se dime nts . Journal of M ari ne Syst em s, 17 , 5 27- 5 39. B a s ile - Doe ls ch, I. , Meu nie r, J. D. , Pa rron, C. , (2005 ). Anothe r contine nta l pool in the te rre s trial s il icon cy cle . N at ure 4 33 , 399- 402. B e kk e r A., H. D. Hol l a nd, P. L. Wa ng, D. Ru mbl e , H. J. Ste in, J. L . Ha nna h, L . L. Coe tze e , N. J. B e uk e s (2004). Da ting the ris e of a tmos phe ric oxy g e n, N at ure 4 27 11 7–1 20. B e kk e r, A., Sl a ck, J. F ., Pl a na vsk y , N. , Kra pe ž, B. , Hofma nn, A., Konhau se r, K. O. , Rou xe l, O. J. (201 0). I ron forma tion: the s e dime nta ry produ ct of a compl e x inte rpl a y a mong ma ntle , te ctonic, oce a nic, a nd bios phe ric proce s se s. Ec on. Geol . 105, 467–5 08. B e u che r, C. P. , M. A. , B rze zins k i, J. L ., Jone s , (2008). Sou rce s and biol og ica l fra ctiona tion of s il icon is ot ope s in the Ea s te rn Equ a toria l Pa cific. Geoc hi m. C osm oc hi m . Act a 7 2, 3063‐ 3073. B e uk es , N. J. , (1 973). Pre ca mbria n iron - forma tions of Sou the rn Africa . Ec onom i c Geol ogy 68 , 960–1 004 B e uk es , N. J. , and Gu tzme r, J. , (2008). Orig in a nd pa l e oenvironme nta l s ignifica nce of ma jor iron forma tions a t the Arche a n- Pa le oprote rozoic b ou nda ry . Rev i ews i n Ec onom i c Geol ogy , 15, 5 −47.

46

Chapter 1

B e uk es , N. J. , Kle in, C. , Kau fma n, A. J., a nd Ha ye s , J. M., (1 990). Ca rbona te pe trog ra phy , k e rog en dis tribu tion, a nd ca rbon a nd oxy g en is otope va ria tions in a n Ea rl y Prote rozoic tra ns ition from l ime s tone to ir on - forma tion de pos iti on: Tra ns va a l Su pe rg rou p, Sou th Africa . Ec onom i c Geol ogy , 85, 663–690. B ibik ova , E. V. , (201 0). I s otope - g e oche mical cons tra ints on the forma tion of the e a rl y Ea rth crus t. Pet rol ogy, 18 , 482- 488. B ra te rman, P. S. , Ca irns- Smith, A. G. , a nd Sl ope r, R. W. , (1 983). Photo oxida ti on of hy dra te d F e 2+: Sig nifica nce for ba nde d iron for ma tions . N at ure, 303 , 1 63−1 64. Ca irns - Smith, A. G. , (1 978). Pre ca mbria n s ol u tion phot oche mis try , inve rs e se g re g a tion, a nd ba nde d iron forma tions . N at ure, 7 6, 807 –808. Ca nfie l d D. E. , (1 998). A ne w mode l for Prote rozoic oce a n che mis try , N at ure, 396, 45 0– 45 3. Ca nfie l d D. E. (2005 ) The e a rl y his tory of a tmos phe ric oxy g en: homa g e to Robe rt M. Ga rre ls . Ann. Re v. Eart h Pl anet. Sc i . 33 , 1 –36. Ca rdina l , D., L. Al l e ma n, F . Deha irs, N. Sa voy e , T. Tru ll , L . André , (2005 ). Rel e vance of s il icon is otope s to Si - nu trie nt u til iza tion a nd Si - s ou rce a ss e ss me nt in Anta rctic wa te rs , Gl ob al Bi ogeoc hem . C yc l es, 19, GB 2007 Ca rdina l , D., N. , Sa voy e , T. W. , Tru ll , F ., De hairs , E. E. , Kopczy ns k a, F . , F ripia t, J. ‐ L . ., Tis on, L . , André , (2007 ). Sil icon is otope s in s pri ng Sou the rn Oce a n dia toms : l a rge zona l cha nge s des pite homog e ne ity a mong s ize fra ctions . M ar. C hem. 106, 46‐ 62. Ca rdina l , D. , J. Ga il la rde t, H. J. Hu ghe s, S. Opfe rg e l t, L. André , (201 0). Contra s ting s il icon is oto pe s ig na tu re s in rive rs from th e Cong o B a s in a nd the s pe cific be ha viou r of org a nic - rich wa te rs, Geop hys. Res. Let t ., 37 , L 1 2403 Ca ro G. , B ou rdon B . , Wood B . J. a nd Corg ne A. (2005 ). Tra ce -e l e me nt fra ctiona tion in Ha de a n mantl e g e ne ra te d by me l t se g re g a tion from a ma g ma oce a n. N at ure, 4 36, 246- 249. Ca tl ing D. C. , Zahnle K. J. , a nd McKa y C. P. (2001 ). B iog enic me tha ne , hy drog e n e s ca pe , a nd the irre ve rs ibl e oxida tion of e a rl y Ea rth . Sc i enc e, 293 , 839- 843. Cha k ra ba rti R. a nd Ja cobs e n S. B . (201 0) Sil icon is oto pe s in the inne r Sol a r Sy s te m: impl ica tions f or core f orma tion, s ol a r ne bu l ar proce s s e s a nd pa rtia l mel ting . Geoc hi m . C osm oc hi m. Ac t a , 7 4, 6921 –6933. Cl ou d, P. E., (1 965 ). Sig nifica nce of Gu nfl int (Pre ca mbria n) microfl ora —photos y nthe tic oxy g e n ma y ha ve ha d importa nt l oca l e ff e cts be fore be coming a ma jor a tmos phe ric g a s . Sc i enc e, 148 , 27−35 . Cl ou d Jr. , P. E. , (1 968). Atmos phe ric a nd hy dros phe ric e vol u tion on the primitive Ea rth. Sc i enc e, 160, 729–736. Cl ou d, P. , (1 973), Pa l e oe col og ical s ig nifica nce of ba nde d iron - f orma tion: Econo mic Geol ogy, 68 , 1 1 35 −1 1 43. Cl ou t J. M. F . a nd Simons on B . M. (2005 ) Pre ca mbria n iron forma tion a nd iron forma tion hos te d ir on ore de pos its . Ec onom i c Geol ogy , 64 3- 679 Col l e rs on K. D. a nd Ka mbe r B . S. (1 999). Evol u tion of the contine nts a nd the a tmos phe re infe rre d from Th - U - Nb s y s te ma tics of the de p l e te d ma ntle . Sc i enc e, 283 , 1 5 1 91 5 22. Condie K. C. (1 998). Epis odic contine ntal growth a nd s u pe rcontine nts : a ma ntle a va l anche conne ction? Eart h and Pl anet ary Sc i enc e Let t ers , 163 , 97 - 1 08. Condie K. C. (2000). Epis odic contine nta l g rowth mode l s : a fte rthou g hts a nd e xte ns ions . Tec t onop hysi c s, 322, 1 5 3- 1 62. Conl e y , D. J. (2002). Te rre s tria l e cosy s te ms a nd the g l oba l biog e oche mical s il ica cy cl e. Gl oba l B iog e oche mica l Cy cl e s, 1 6(4): 1 1 21 doi: 1 0.1 029/2002GB 001 894.

General Overview

47

Crowe , S. A. J. , Ka ts e v, C. , Ma ge n, S. , O’ Ne il l , C. , Stu rm, A. H. , Ca n fie l d D. E. , Ha ffne r, G. D. , Mu cci, A. , Su ndby , B ., a nd F owle , D. A. , (2008). Photo fe rrotrophs thrive in a n Archea n oce a n ana l ogu e. Proc eed i ngs of the N at i onal Ac ad em y of Sc i enc es , 105, 1 5 937 −1 5 943. Da l zie l , I . W.D. , (1 991 ). Pa cific ma rg ins of L au re ntia and Ea s t Anta rctica a s a conju ga te rift pa ir: e vide nce a nd impl ica ti ons for a n Eoc a mbria n s u pe rcontine nt. Geol ogy 19, 5 98– 601 . De Kock , M. O., Eva ns , D.A. D., Be uk e s, N. J. (2009). Va l ida ting the e xis te nce of Va a l ba ra in the Ne oa rchea n . Prec am b ri an Researc h , 17 4, 1 45 - 1 5 4. De L a Rocha , C. L . , M. A. B rze zins k i, M. J. De Niro, (1 997 ). F ra ctiona tion of s il icon is otope s by ma rine dia toms du ring biog e nic s il ica forma tion, Geoc hi m. C osm oc hi m . Act a, 61, 5 051 - 5 05 6 De L a Rocha , C. L. , M. A. , B rze zinsk i, M. J. , De Niro, (1 998). Sil icon ‐ is otope c omp os ition of dia toms a s a n indica tor of pa s t oce a nic cha ng e. N at ur, e 395, 680‐ 683. De L a Rocha , C. L . , M. A. Brze zinsk i, M. J. De Niro, (2000). A f irs t l ook a t the dis tribu tion o f the s ta bl e is otope s of s il ic on in na tu ra l wa ters , Geoc hi m . C osm oc hi m . Act a, 64, 2467 - 2477 De L a Rocha , C. , (2003). Sil icon is ot ope fra ctiona tion by ma rine s pong e s a nd the re cons tru ction of the s il icon is oto pe com p os ition of a ncie nt de e p wa te r, Geol ogy, 31, 423- 4 26 De Ronde , C. E. J. , De Wit, M. J. , Spoone r, E. T. C. , (1 994). Ea rly Arche a n (>3. 2 Ga ) F eoxide - rich, hy drothe rma l dis cha rge ve nts in the B a rbe rton Gree ns tone B el t, Sou th Africa . Geol . Soc . Am er. Bul l . 106, 86–1 04. De Ronde , C. E. J. , Cha nne r, D. M., De , R. , F au re , K. , B ra y, C. J. , Spoone r, E. T. C. , (1 997 ). F l u id che mis try of Archea n s e a fl oor hy drotherma l ve nts: impl ica tions for the compos it ion o f circa 3. 2 Ga s e a wa ter. Geoc him . C osm oc hi m . Ac t a , 61, 4025 –4042. De rry L. A. and Ja cobs e n S. B . (1 990) The chemica l e vol u tion of Pre ca mbria n s e a wa te r: Evide nce from REEs in ba nde d iron f orma tion s . Geoc hi m. C osm oc hi m. Ac t a , 54 , 2965 –2977. De rry , L. A. , Ku rtz, A. C. , Zie gl e r, K., and Cha dwick , O. A. (2005 ). B iol og ica l control of te rre s trial s il ica cy cl ing and e xport u xe s to wa te rs he ds . N at ure, 4 33 , 728- 731 . De Vrie s , S. T., Tou re t, J. L. R. , (2007 ). Ea rly Archa ea n hy drothe rma l flu ids ; a s tu dy of incl u s ions from the 3. 4 B u ck Ridg e Che rt, Ba be rton Gree ns tone B el t, Sou th Africa . C hem ic al Geol ogy 237 , 289– 302. De Wit M. J. a nd Ha rt R. A. (1 993). Ea rth’ s ea rl ies t contine nta l l ithos phe re , hy drothe rmal fl u x a nd crus ta l re cy cl ing . Li t hos, 30, 309- 35 5 . Ding , T. , Jia ng , S. , Wa n, D. , L i, Y. , L i, J. , Song , H. , L iu , Z. , Ya o, X. , (1 996). Sil icon is otope g e oche mis try. Ge ol og ica l Pu bl ishing Hou se , Be ijing , China , 1 25 pp. Ding , T. , D. Wa n, C. Wa ng , F . Zhang , (2004). Sil icon is otope comp os itions of dis s ol ve d a nd s us pe nde d ma tter in the Ya ng tze , Geoc him . C osm oc hi m . Ac t a, 68, 205 - 21 6 Ding , T. P. , Ma, G. R. , Shu i, M. X. , Wa n, D.F ., a nd L i, R. H. , (2005 ). Sil icon is otope s tu dy on rice pl a nts from the Zhe jia ng province , China . Chem . Geol. 218 , 41 - 5 0. Dou thitt, C. B . (1 982). The g e oche mis try of the s ta bl e is otope s of s il icon. Geoc him i c a et C osm oc hi m i c a Ac t a , 4 6, 1 449- 1 45 8. Dre ve r, J. I. , (1 974). Ge oche mica l mode l for the orig in of Pre ca mbria n ba nde d iron forma tions . Geol ogi c al Soc i et y of Am eri c a Bul l eti n 35, 1 099–1 1 06. Eng s tröm, E. , I . Rodu s hk in, J. I ng ri, D. B a xte r, F. Ecke , H. Ös terl und, B. Öhl ande r, (201 0). Te mpora l is otopic va ria ti ons of dis s o l ve d s il icon in a pris tine bore al rive r, C hem . Geol . , 27 1, 1 42- 1 5 2

48

Chapter 1

Eva ns D.A. , N. J. B eu ke s , J.L . Kirs chvink , (1 997) L ow- l a titu de gl a cia tions in the Pa l a e oprote rozoic e ra , N at ure 38 6, 262–266. Eva ns , M. J. , De rry, L . A. , (2002). Qua rtz control of hig h g e rma niu m/s il icon ra tios in g e othe rma l wa te rs. Geol ogy. 30, 1 01 9- 1 022. F a u re , F . a n d Me ns ing , T. M. (2005 ). I s otope s: Principl e s and Appl ica tions . Wil ey and s ons . F a rqu ha r J. , H. M. Ba o, M. Thie me ns , (2000) Atmos phe ric infl u e nce of Ea rth’s e a rl ie s t s u l fu r cy cle , Sc i enc e 28 9 75 6–75 8. F a rqu ha r J. , B . A. Wing , (2003) Mu l tipl e su l fu r is otope s a nd the e vol u tion of the a tmos phe re , Eart h Pl anet . Sc i . Let t . 213 , 1 – 1 3. F e do, C. M., Erik s s on, K. A., a nd Krog s ta d, E. J., (1 996). Ge oche mis try of s ha le s from the Arche an (~3. 0Ga ) Bu hwa g re e ns tone be l t, Zimba bwe : I mpl ica tions for prove na nce and s ou rce- a re a wea the ring . Geochi m i c a et C osm oc hi m ic a Ac t a , 60, 1 75 1 –1 763. F is che r, W. W. And Knol l , A. H. (2009). An iron s hu ttl e for de e pwa te r s il ica in L a te Arche an a nd ea rly Pa l e oprote rozoic iron forma tion. Geol . Soc . Am . Bul l . 121, 222– 235 . F itou s s i C. , B ou rdon B . , Kl e ine T. , Obe rl i F. a nd Re y nol ds B . C. (2009) Si is otope s y s te ma tics of me te orite s a nd te rre s tria l peridotite s : impl ica ti ons for Mg / Si fra ctiona tion in the s ola r ne bu la and for Si in the Ea rth’ s core . Eart h Pl anet. Sc i . Let t . 287 , 77–85. F itou s s i C. a nd B ou rdon B . (201 2) Sil icon I s otope Evide nce Ag a ins t an Ens ta tite Chondrite Ea rth, Sc i enc e, 335, 1 477- 1 480 F ripia t, F ., A.- J. Ca vag na, N. Sa voye , F . De ha irs , L. Andre , D. Ca rdina l , (201 1 ). I s otopic cons tra ints on the Si - biog e oche mica l cy cl e of the Anta rct ic Zone in the Ke rgu e le n a re a (KEOPS), M ar. C hem. , 123 , 1 1- 22 F ol k , R. L ., (1 980). Pe trol ogy of Se dime nta ry Rock s , He mphil l Pu bl is hing Compa ny, Au s tin, 1 85 pp F rimme l , H. E. , (2005 ). Archa ea n a tmos phe ric e volu tion: e vide nce from the Witwa te rs ra nd g ol d fie l ds , Sou th Africa . Eart h Sc i . Rev . 7 0, 1 –46. F roe l ich, P. N. , Ha mbrick , G. A. , Andre ae , M. O. , and Mortl ock , R. A. (1 985). The g e oche mis try of inorg a nic g e rmaniu m in na tu ra l wa te rs. Journal of p hysi c al researc h, 90, 1 1 3 3- 1 1 41 . F roe l ich, P. N. , Bl a nc, V. , Mor tl ock , R. A., Chil lru d, S. N., Du ns ta n, W. , U domk it, A., and Pe ng , T. H. (1 992). Rive r u xes of dis s ol ve d s il ica to the oce an we re hig her du ring g l a cial s : Ge /Si in dia toms , rive rs , a nd oce ans . Pal eoc eanograp hy , 7 , 739- 767. F ry er B . J. , F y fe W. S. , and Ke rrich R. (1 979). Arche an vol ca nog enic oce a ns . Che mica l Geol ogy, 24 , 25 - 33. Ga rre ls , R. M., (1 987 ), A mode l for the de pos ition o f the micro - ba nde d Pre ca mbria n iron - forma tions . Am eri c an Journal of Sc i enc e , 28 7 , 81 –1 06. Ge org , R. B ., B. C. Re ynol ds , M. F rank , A. N. Ha l l ida y, (2006). Me cha nis ms control l ing the s il icon is otopic c omp os itions o f rive r wa t e rs , Eart h Pl anet . Sc i. Lett . , 24 9, 290- 306 Ge org , R.B . , Ha l l iday , A. N., Schau bl e, E.A. , Re y nol ds , B . C. , (2007a ). Sil icon in the Ea rth’ s core. N at ure 44 7, 1 1 02- 1 1 06. Ge org , R. B . , B . C. Re ynol ds , A. J. We s t, K. W. B u rton, A. N. Ha l l iday , (2007b). Sil icon is otope va ria ti ons a ccompa ny ing ba s a l t we a the ring in I ce l and, Eart h Pl anet . Sc i. Let t . , 261, 476- 490

General Overview

49

Ge org , R. B . , C. Zhu, B . C. Re y nol ds , A. N. Ha ll ida y , (2009). Sta bl e s il icon is otope s o f g rou ndwa te r, fe l ds pa rs , a nd cl ay coa ting s in the Na va jo Sa nds tone a qu ife r, B l a ck Mes a , Arizona , U SA, Geoc hi m. C osm oc him. Ac t a , 73 , 2229- 2241 Gol e , M. J., a nd Kl e in, C. , (1 981 ). Ba nde d iron - forma tion throu g h mu ch of Pre ca mbria n time . Journal of Geol ogy , 8 9, 1 69−1 83. Gome s , R. , L e vins on, H. F ., Ts iga nis , K. And Morbide l l i, A. (2005 ). Orig in of the ca ta cl ys mic La te He a vy B omba rdme nt pe riod of the te rre s trial pl ane ts. N at ure 4 35 , 466–469 Gou g h, D. O., (1 981 ). Sol a r inte rior s tru ctu re and l u m inos ity va ria tions . Sol ar Phys. 7 4 , 21 –34. Gu nnars s on, I . & Arnors s on, S. (2000). Amorphou s s il ica s olu bil ity a nd the the rmody na mic prope rtie s of H4Si O4 in the ra nge of 0° to 35 0° C a t P s a t Geoc hi m ic a et C osm oc hi m ic a Ac t a , 64 , 2295 - 2307 Gros s , G. A. , (1 965 ) . Ge ol og y of iron de pos its in Ca na da, Vol . 1 . Ge ne ra l g e ol ogy a nd e va l ua tion of ir on de pos its . Ca na dia n Ge ol og ica l Su rve y, Economic Ge ol og y Re port. Gros s G. A. (1 973) The de pos itiona l e nvironment of the princip ia l ty pe s of Pre ca mbria n ironforma ti ons . I n Gene s is of Pre ca mbria n I ron a nd Ma ng a ne se De pos its (e ds . ). U NESCO, Pa ris , Ea rth Scie nces 9. pp. 1 5 - 21 . Gros s , G.A. , (1 983). Te ctonic s ys te ms a nd the de pos ition of ironf orma tion. Prec am b ri an Researc h 2 0, 1 71 –1 87. Grotzing e r J. P. a nd Ka s ting J. F. (1 993). Ne w cons tra ints on Pre ca mbria n oce an compos it ion. Journal of Geol ogy , 1 01, 235 - 24 3. Ha bicht, K. S. , Ga de , M. , Tha mdru p, B ., Be rg, P. , a nd Ca nfie l d, D. E., (2002). Ca l ibra tion of s u l fa te l e ve l s in the Archea n oce a n. Sc i enc e, 298 , 2372− 2374. Ha l l ida y A. N. , (2004). Mixing, vol a tile l oss and compos iti ona l chang e du ring impa ct drive n a ccre tion of the Ea rth, N ature 4 27 5 05 –509. Ha ma de T. , Konha us e r K., Ra is we l l R. , Gol ds mith S. a nd Morris R. (2003) U s ing Ge /Si ra tios to de cou pl e iron a nd s il ica flu xes in Preca mbria n ba nde d iron foma tions . Geol ogy 31, 35 - 38. Ha mmond, D. E. , McMa nu s , J. , Be rel s on, W. M. , Me re dith, C. , Kl ink ha mme r, G. P. , a nd Coa l e , K. H. (2000). Dia g e ne tic fra ctiona tion o f Ge a nd Si in re du cing s e dime nts : the miss ing Ge s ink a nd a pos s ibl e me chanis m to ca u se g l a cial /inte rg la cia l va ria tions in oce a nic Ge /Si. Geoc hi m i c a et C osmoc hi m i c a Ac t a , 64, 245 3- 2465 . Ha qq- Mis ra , J. D. , Doma ga l - Gol dma n, S. D. , Ka s ting, P. J. & Ka s ting , J. F. (2008) Ast rob i ol ogy 8 , 1 1 27 - 1 1 37. He ck , P. R. , J. M. Hu be rty , N. T. Kita , T. U s hikubo, R. Ko zdon, J. W. Va l l e y (201 1 ). SI MS a na ly se s of s il icon a nd oxy g en is otope ra tios for qu a rtz from Archea n a nd Pa l e oprote rozoic ba nde d iron f orma tions . Geoc hi m . C osm oc hi m . Ac t a 75, 5 8795 891 . He is ing , S. , Richte r, L. , Lu dwig , W., a nd Schink, B . , (1 999), Chl orobiu m fe rrooxida ns s p. nov. , a phototrophic g re e n su l fu r ba cte rium tha t oxidize s fe rrou s iron in cocu l tu re with a “Ge os piril lu m” s p. Stra in . Arc hi v es of M i c robi ol ogy , 17 2, 1 1 6– 1 24. He s s le r, A., L owe , D., Jone s , R. , B ird, D. , (2004). A l owe r l imit for a tmos phe ric ca rbon dioxide l e ve l s 3. 2 bil l ion y e a rs ag o. N at ure 4 28, 736–738. He s s le r A. M. a nd L owe D. R. (2006). Wea the ring and s e dime nt g e ne ra tion in the Arche an: An inte g ra te d s tu dy of the e vol u tion of s il icicl a s tic s e dime nta ry rocks of the 3. 2 Ga Mo odie s Grou p, B a rbe rton Gre e ns tone B e l t, Sou th Africa . Prec am b ri an Researc h , 151, 1 85 –21 0.

50

Chapter 1

Hick ma n, A. H. , (2004). Two c ontra s ting g ranite - g ree ns tone s te rranes in the Pil ba ra Cra ton, Aus tra l ia : e vide nce for ve rtical and hori zonta l te ctonic re g imes prior to 2900 Ma . Prec am b ri an Res. , 131, 15 3–1 72. Hoe fs J. (200 4) Sta bl e I s otope Ge oche mis try . Spring e r. Hofma nn, A. , (2005 ). The g e oche mis try of s e dime nta ry rock s from the F ig Tre e Grou p, B a rbe rton g ree ns tone be l t: impl ica tions for tectonic, hy drothe rmal and su rfa ce proce s s e s du ring mid - Archae a n time s. Prec am bri an Researc h 14 3 , 23–49. Hofma nn, A. , Ha rris , C. , (2008). Sil ica a l te ra tion zone s in the B a rberton g ree ns tone be l t: A window into s u bse a fl oor proce s se s 3. 5 - 3. 3 Ga a g o. C hem . Geol. 257, 221 239. Hofma nn, A. , B ek ke r, A. , Rou xel , O. , Ru mbl e, D. , Ma s te r, S. , (2009). Mu l tipl e s u l phu r a nd iron is otope com pos iti on of de trita l py rite in Archa ea n se dime nta ry rock s : a ne w tool for prove na nce a na ly s is i. Eart h and Planet ary Sc i enc e Let t er s 28 6, 436– 445 . Hol l a nd, H. D. (1 973) The Oce a ns : a pos s ibl e s ou rce for iron in iron - forma ti on. Economic Geol ogy, 68 , 1 1 69. 1 172. Hol l a nd, H. D. (2005 ), Se dime nta ry mine ra l de pos its a nd the e vol u tion of Ea rth’ s ne a rs u rfa ce e nvironme nts. Ec onom i c Geol ogy , 100, 1 489−1 5 09. Hol l a nd H. D. (2006) The oxyg e na tion of the atmos phe re a nd ocea ns. Phil . Trans. Roy. Soc . B 361, 903–91 5 . doi: 1 0. 1 098/ rs tb. 2006. 1 838. Hol l a nd, H. D., (2009). Why the a tmos phe re beca me oxy g ena te d: a propos a l . Geoc him . C osm oc hi m . Act a 7 3 , 5 241 –5 25 5 . Hu g hes , H. J. , F . Sonda g , C. Cocqu y t, A. L a ra que , A. Pandi, L. André, D. Cardina l , (201 1 ). Effe ct of s e a s onal biog e nic s il ica va ria tions on dis s ol ve d s il icon flu xe s a nd is otopic s ig na tu re s in the Cong o Rive r, Li m nol ogy and Oc eanograp hy , 56, 5 5 1- 5 61. Hu s ton, D.L . , a nd L og an, G.A. , (2004), B a rite , BI F s a nd bug s: Evide nce for the e vol u tion of the Ea rth’ s ea rl y hy dros phe re . Eart h and Pl anet ary Sc i enc e Lett ers , 220, 41 −5 5 . I s l ey , A. E. , (1 995 ). Hy drothe rma l pl u mes and the del ive ry of iron to ba nde d iron forma tion . Jou rnal of Geol ogy , 103 , 1 69−1 85 . I s l ey , A. E., a nd Abbott, D. H. , (1 999). Plu me - re la te d ma fic vol ca nis m a nd the de pos ition of ba nde d iron f orma tion . Jou rnal of Geop hysi c al Researc h , 1 04, 1 5 , 461 −1 5 , 477. Ja cobs e n, S. B ., Pime nte l- Kl ose , M. R. , (1 988). Nd is otop ic va ria tions in Pre ca mbria n ba nde d iron forma tions . Geop hys. Res. Let t . 15, 393- 396. Ja me s , H. L . (1 95 4) Se dime nta ry fa cie s of iron - forma tion. Ec onom i c Geol ogy, 4 9, 235 . 293. Ka mbe r, B . S. , We bb, G. E. , (2001 ). The g e oche mis try of l a te Archae a n microbia l ca rbona te : impl ica tions for oce a n che mis try a nd contine nta l e ros ion his tory . Geoc hi m . C osm oc hi m. Ac t a . 65, 25 09- 25 25 . Ka mbe r, B . S. , B ol ha r, R. , We bb, G. E. , (2004). Ge oche mis try of l a te Archa e an s troma tol ite s from Zimba bwe : e vide nce for micro bia l l ife in re s tricte d e picontine nta l se a s. Prec am b ri an Res. 132, 379-399. Ka ppl e r, A. , Pas qu e ro, C. , Konha us e r, K. O., and Ne wma n, D. K. , (2005 ), De pos ition of ba nde d iron forma tions by a noxy g e nic phototrophic F e (I I ) - oxidizing ba cte ria : Geol ogy, 33 , 865 –868 Ka s ting, J. F . , (1 988). Ru na wa y a nd mois t g re enhou s e a tmos phe re s a nd the e vol u tion of Ea rth a nd Ve nu s. Ic arus 7 4 , 472–494. Ka s ting J. F. (1 993). Ea rth’ s ea rly a tmos phe re. Sc i enc e, 259, 920- 926. Ka s ting (2001 ) The rise of a tmos phe ric oxy g e n. Sc i enc e, 293 , 81 9. 820. Ka s ting (2004) Whe n me tha ne ma de cl ima te . Sci ent iÞ c Am eri c an , 291, 78. 85. Ka s ting J. F. (2005 ) Me tha ne and cl ima te du ring the Pre ca mbria n e ra . Prec am b . Res. 137 , 1 1 9–1 29.

General Overview

51

Ka s ting, J. F . a nd Ca tl ing , D. (2003) Evol u tion of a ha bita bl e pl a ne t. Annual Rev i ew of Ast ronom y and Ast rop hysi c s, 4 1, 429. 46 3. King , S. L ., Froe l ich, P. N., and Ja hnk e, R. A. (2000). Ea rly dia ge ne s is of g e rma niu m in s e dime nts of the Anta rctic Sou th Atl a ntic: in se a rch of the mis s ing Ge s ink . Geoc hi m ic a et C osm oc hi m ic a Ac t a , 64 , 1 375 -1 390. Kl e in C. (2005 ) Some Pre ca mbria n ba nde d iron - forma tions (B I Fs ) from a rou nd the worl d: The ir ag e , ge ol og ic s e tting, mine ra l ogy , me ta morphis m, g e oche mis try , a nd orig in. Am eri c an Mi neral ogi st 90, 1 473- 1 499. Kl e in, C., a nd Be uk es , N. J. , (1 989). Ge ochemis try a nd s e dimentol og y of a fa cie s tra ns ition from l ime s tone to iron - forma ti on d e pos ition in the Ea rl y Prote rozoic Tra ns va al Su pe rg rou p, Sou th Africa . Ec onom i c Geol ogy , 8 4 , 1 733−1 774. Kl e in a nd B eu k es (1 993) Prote rozoic ir on - for ma tion. I n K. C. Condie , Ed. , Prote rozoic Cru s ta l Evolu tion, p. 383. 41 8. El s e vie r, Ams te rda m. Kl e ine T. , C. Mu nk e r, K. Me zg er, H. Pa l me. (2002). Ra pid a ccre tion a nd ea rl y core forma tion on a s te roids a nd the te rre s tria l pl a ne ts from Hf –W chronome try , N at ure, 4 18 , 95 2–95 5. Kna u th, L . P. , (1 994). Pe trog e ne s is of che rt, in: He a ne y, P. J. , Pre witt, C. T., Gibbs , G. V. (Eds . ), Sil ica : Phys ica l B eha vior, Ge oche mis try a nd Ma te ria l s Appl ica tions . Re vie ws in Mine ral og y ,Mine ra l og ica Socie ty of Ame rica , Wa shing ton, D. C. , 29, pp. 23 3- 25 8. Kna u th, L . P. , L owe , D. R. , (2003). Hig h Archea n cl ima tic te mpe ra tu re infe rre d from oxy g e n is otope g e oche mis try of che rts on the 3. 5 Ga Swa zil a nd Su pe rg rou p, Sou th Africa . GSA Bul l . 115, 5 66-5 80. Kna u th L . P. (2005 ). Te mpe ra tu re a nd sa l inity his tory of the Pre ca mbria n oce a n: impl ica tions for the cou rs e of micro b ia l e vol u tion. Pal eogeograp hy, Pal eoc l i m at ol ogy, Pal eoec ol ogy , 219, 5 3- 69. Konha us e r, K. O. , Ha ma de , T. , Morris , R. C., F e rris , F . G., Sou tha m, G. , Ra is wel l , R. , a nd Ca nfie l d, D. , (2002). Cou l d ba cte ria ha ve forme d the Pre ca mbria n ba nde d iron forma tions ? Geol ogy, 30, 1 079−1 082. Konha us e r, K. O., Ne wma n, D. K. , a nd Ka pple r, A. , (2005 ). F e(I II ) re du ction in B IF s : The pote ntia l s ig nifica nce of microbia l F e (II I ) re du ction du ring de pos ition of Pre ca mbria n ba nde d iron forma tions . Geob i ol ogy, 3 , 1 67 –1 77. Konha us e r, K. O. , Amsk ol d , L . , La l onde , S. V. , Pos th, N. R. , Ka ppl e r, A. , a nd Anba r, A., (2007a ), De cou pl ing photoche mica l Fe (II ) oxida tion from s ha l l ow - wa te r B IF de pos ition . Eart h and Pl anet ary Sc i enc e Lett ers , 258 , 87−1 00. Kra pe ž, B . , B a rl ey , M. E., and Pick a rd, A.L . , (2003). Hy dr othe rma l a nd res e dime nte d orig ins of the pre cu rs or s e diments to ba nde d iron forma ti ons : Se dime ntol og ica l e vide nce from the ea rly Pa l ae oprote rozoic B ro ck ma n Su pe rs e que nce of We s te rn Au s tral ia : Sed im ent ol ogy , 50, 979−1 01 1. Ku mp L . R. a nd Sey frie d W. E. (2005 ). Hy drothe rma l Fe fl u xe s du ring the Pre ca mbria n: Effe ct of l ow oce a nic s u l fa te conce ntra tions a nd l ow hy dros ta tic pre ss u re on the compos ition of bl a ck s mok e rs . Eart h and Planet ary Sc i enc e Let t ers , 235, 65 4662. Ku rtz, A. C. , De rry , L . A. , a nd Cha dwick , O. A. (2002). Ge rma niu m- s il icon fra ctiona tion in the we a the ring environme nt. Geoc him i c a et C osm oc hi m i c a Ac t a , 66, 1 5 25- 1 5 37. Ku rtz, A. C. and De rry , L . A. (2004). Tra cing s il ica te wea the ring and te rre s tria l s il ica cy cl ing with Ge /Si ra tios . I n Wa nty, R. B . a nd Se a l , R. R. , e ditors , Proc . 11t h Int . Sym p . on Wat er Roc k Int erac t i on , 833- 836, The Ne the rla nds .B al ke ma Pu bs . L a s ce l le s , D.F ., (2007). B l a ck s mok e rs and de ns ity cu rrents : A u niformita ria n mode l for the g ene s is of ba nde d iron - forma tions . Ore Geol ogy Rev i ews. 32, 381 - 41 1 .

52

Chapter 1

L owe D. R. a nd Tice M. M. (2004). Ge ol og ic e vide nce for Arche an a tmos phe ric a nd cl ima tic e vol u tion: F l u ctu a ting le ve ls of CO2 a nd CH4, a nd O2 with a n ove rriding te ctonic control . Geol ogy, 32, 493- 49 6. L owe D. R. a nd Tice M. M. (2007). Te ctonic c ontrol s on a tmos phe ric, cl ima tic, a nd biol og ica l e vol u tion 3. 5 - 2. 4 Ga . Prec am b ri an Researc h, 158 , 1 77 - 1 97. Ohmoto, H. , Wa ta na be, Y. , Ya ma gu chi, K. E., Na ra ok a , H. , Ha ru na, M., Ka k eg a wa , T. , Ha y a shi, K. , a nd Ka to, Y. , (2006), Che mica l a nd biol og ica l e vol u tio n o f e a rl y Ea rth: Cons tra ints from ba nde d iron f orma tio ns . Geol ogi c al Soc i et y of Am eric a M em oi r 198 , 291 −331 . Ma rin, J. , M. Chau ss idon, F . Robe rt (201 0). Micros ca l e oxy ge n is otope va ria tions in 1 . 9 Ga Gunfl int che rts : as s es s ments of diag e nes is e ffe cts and impl ica tions for oce a nic pa l e ote mpe ra tu re re cons tru ctions . Geoc hi m . C osm oc hi m. Ac t a 74 , 11 6– 1 30. Ma rin - Carbonne , J. , M. Chau ss idon, M. - C. B oiron, F . Robe rt (201 1 ). A combine d in s itu oxy g e n, s il icon is otopic a nd fl u id incl u s ion s tu dy of a che rt s a mpl e from Onve rwa cht Grou p (3. 35 Ga , Sou th Africa ) : Ne w cons tra ints on fl u id circu l a tion. C hem . Geol. 28 6, 5 9- 71 . Ma rin - Carbonne , J. , M. Chau s s idon, F . Robe rt (201 2). Microme te r - s ca le che mical a nd is otopic crite ria (O a nd Si) on the or ig in a nd his tory of Pre ca mbria n che rts : I mpl ica tions for pa l e o- te mpe ra tu re re cons tru ctions . Geoc hi m. C osm oc hi m. Ac t a . 92, 1 29- 1 47. Mé heu t, M. , L a zze ri, M., Ba la n, E. , Mau ri, F. , (2007 ). Equ il ibriu m is otopic fra ctiona ti on in the ka ol inite , qua rtz, wa ter s ys te m: Pre diction from firs t - principl e s de ns ityfu nctiona l the ory. Geoc hi m. C osm oc him . Act a 7 1, 31 70- 31 81 . Mé heu t, M., L a zze ri, M., B a la n, E., Ma uri, F . , (2009). Stru ctu ra l control ove r e qu il ibriu m s il icon a nd oxy g en is otopic fra ctio na tion: A firs t - principl e s dens ity fu nctiona l the ory s tu dy. C hem . Geol . 258, 28- 37. Me unie r, J. D. (2003). L e rôl e de s pl a nte s da ns l e tra ns fe rt du s il ici u m à la s u rfa ce des contine nts . C omp t es Rend us Geosc i enc e, 335, 1 199- 1 206. Morris R. C. (1 993) Ge ne tic mode l l ing for ba nde d iron - forma tion of the Ha me rs l e y Grou p, Pil ba ra Cra ton, We s tern Aus tra l ia. Prec am b ri an Researc h 60, 24 3- 286. Opfe rg e l t, S. , D. Ca rdi na l , C. He nrie t, L. André , B. Del va u x, (2006). Sil icon is otope fra ctiona tion be twe e n pl a nt pa rts in ba na na : in s itu vs . in vitro, J. Geoc hem . Exp l or. , 88 , 224- 227 Opfe rg e l t, S., Ca rdina l , D. , André, L. , De l vig ne , C. , Bre mond, L. and Del va u x, B. (201 0). Va ria tions of δ30S i a nd Ge /Si with we a the ring a nd biog e nic inpu t in tropica l ba s a l tic a sh s oil s u nde r monocu l tu re . Geoc him . C osm oc hi m . Act a 7 4, 225 - 240. Ma l iva R. G. , Knol l A. H., a nd Simons on B . M. (2005 ). Se cul a r cha nge s in the Pre ca mbria n s il ica cy cl e: Ins ig h ts from che rt pe trol og y . Geologi c al Soc i et y of Am eri c a Bul l eti n , 117 , 835 - 845 . Ma rtin, H. , (1 986). Effe ct of s te e pe r Archea n g e othe rma l g ra dient on g e oche mis try of s u bdu ction - zone ma g mas . Geol ogy 14 , 75 3– 75 6Ma rtin F . , Il de fons e P. , Ha ze mann J. - L . , Pe tit S. , Gra u by O., and De ca rrea u A. (1 996) Ra ndom dis tribu tion of Ge a nd Si in s y nthe tic ta l c a nd EXAF S a nd F TI R s tu dy . Eur. J. Mi neral . 8 , 289–299. Ma rtin, H., Smithie s , R. H. , Ra pp, R., Moye n, J. - F . a ndCha mpion, D. , (200). An ove rvie w of a da k ite , TTG a nd s a nu k it oid: rel a tions hips and s ome impl ica tions for cru s tal e vol u tion. Li t hos, 7 9, 1 –24. Ma y na rd, J. B . (1 992). Che mis try of mode rn s oil s a s a g u ide to inte rpre ting Pre ca mbria n pa l e os ol s . Jour. Geol ogy 100, 279- 289

General Overview

53

McMa nu s , J. , Ha mmond, D. E. , Cu mmins , K. , Kl ink ha mme r, G. P. , a nd B ere ls on, W. M. (2003). Dia g e ne tic Ge - Si fra ctiona tion in contine nta l ma rg in e nvironme nts : fu rthe r e vide nce for a nonopa l Ge s ink . Geochi m i c a et C osm oc hi m ic a Ac t a , 64 , 45 45 - 45 5 7. Me e rt, J. G. , Powe l l , C. McA., (2001 ). I ntrodu ction to the s pe cia l vol u me on the a s se mbl y a nd brea ku p of R odinia . Prec am b ri an Res. 110, 1 –8. Mil l ig a n, A. J. , D. E., Va re la , M. A., B rze zins k i, F.M. M., More l , (2004). Dy ma nics of s il icon me ta bol is m a nd s il icon is ot opic dis crimina tion in a ma rine dia to m a s a fu nction of pCO 2. Li m nol . Oc eanogr. 4 9, 322‐ 329. Mojzs is , S. J. , Ha rris on, T. M. , Pidg e on, R. T., (2001 ). Oxy ge n is otope e vide nce from a ncie nt zircons for l iqu id wa te r a t the Ea rth’s s u rfa ce 4, 300 My r a g o. N at ure 4 09, 1 78–1 81 . Mortl ock , R. A. , a nd P. N. F roe l ich (1 987), Contine nta l wea the ring of g e rma niu m: Ge /Si in the g l oba l rive r dis charg e, Geoc hi m. C osm oc him. Ac t a , 51, 2075- 2082 Mortl ock , R. A. , F roel ich, P. N. , F ee l y, R. A. , Ma s s oth, G. J. , B u tte rfie l d, D. A. , a nd L u pton, J. E. (1 993). Sil ica a nd g e rma niu m in Pa cific - oce a n hy drothe rma l ve nts a n d pl u me s. Eart h and Pl anet ary Sc i enc e Let t ers , 119, 365 - 378. Morris , R. C. , (1 993). Gene tic mode l l ing for ba nde d iron - forma tion of the Ha me rs l e y Grou p, Pil ba ra Cra ton, We s te rn Au s tral ia . Prec am b ri an Res. 60, 243- 28 6. Mu rna ne, R. J. a nd Sta l l a rd, R. F . (1 990) . Ge rma niu m a nd s il icon in rive rs of the Orinoco dra ina ge - ba s in. N at ure, 34 4, 749- 75 2. . Nu tma n, A. P., (2006). Antiqu ity of the oce a ns and contine nts . El em ent s 2, 223– 227. Ohmoto H. (1 996). Evide nce in pre –2. 2 Ga pa l e os ol s for the e a rly e vol u tion of a tmos phe ric oxy g e n a nd terre s tria l biota . Geol ogy, 24 , 1 1 35 –1 1 38. Ono, S. , F a ye k, M., Ohmoto, H. , (2000). Orig in of u ra ninite in the El l iot l a ke dis trict, Ca na da a nd the a tmos phe ric oxyg e n le ve l of 2. 3 Ga Earth. Ev ol ut i on of t he At m osp here, Oc ean, C rust and Bi osp here, 67 –74. Pa tte rs on, C. , (1 95 6). Age of me te orites and the e a rth. Geoc him . C osm oc him . Ac t a 10, 230– 2 37. Pa vl ov A. A. , J.F . Ka s ting, L. L . B rown, K. A. Ra g es , R. F ree dma n, (2000) . Gree nhous e wa rming by CH4 in the a tmos phe re of e a rl y Ea rth, J. Geop hys. Res. 105, 11 981 – 1 1 990. Pa vl ov, A. A. and Ka s ting , J. J. (2002) Ma ss - indepe nde nt fra ctiona tion of s u l fu r is otope s in Arche a n se dime nts: Strong e vide nce for a n a noxic Arche an a tmos phe re . Ast rob i ol ogy, 2, 27. 41 0. Pe rry Jr. E. C. a nd Le ftica riu L . (2003). F orma tion a nd g e oche mis try of pre ca mbria n che rts . I n: Tre a tis e on g e oche mis try , Vol . 7 (e d. H. D. Hol l a nd a nd K.K. Tu re k ian), El s e vie r- Pe rga mon, pp. 99 - 1 1 4. Pipe r, J. D. A. , (1 976). Pa la e oma gne tic e vidence for a Prote roz oic s u pe rcontine nt. Phi l os. Trans. R. Soc . Lond . A280, 469 –490. Pipe r, J. D. A. , (2000). The Ne oprote rozoic s u pe rcontine nt: Rodinia or Pa l ae opa ng ea ? Eart h Pl anet . Sc i. Lett . 17 6, 1 31 –1 46. Pl a na vsk y, N. , B e kk e r, A. , Rou xe l , O. J. , Ka mbe r, B . , Hofma nn, A. , Knu ds en, A. , Ly ons , T. W. , (201 0). Ra re Ea rth El e me nt a nd y ttrium comp os itions of Arche a n a nd Pa l e oprote rozoic F e forma tions re vis ite d: Ne w pe rs pe ctive s on the s ig nifica nce a nd me chanis ms of de pos ition. Geoc hi m . C osm oc hi m . Ac t a . 74 , 6387 –6405 . Pok rovs k y , O. S. , Pok rovs k i, G. S. , Schott, J. , and Ga ly , A. (200 6). Expe rime nta l s tu dy of g e rma niu m a ds orption on g oe thite a nd g e rma niu m copre cipita tion w ith iron hy droxide : X- ray a bs orption fine s tru ctu re and ma cros copic cha ra cte riza tion. Geoc hi m ic a et C osm oc hi m ic a Ac t a , 7 0, 3325 - 3341 .

54

Chapter 1

Pos th, N. R. , He g l e r, F. , Konhau se r, K. O., Ka ppl e r, A. , (2008). Al te rna ting Si a nd F e de pos ition ca u s e d by te mpe ra tu re flu ctu a tions in Pre ca mbria n oce a ns . N at . Geosc i . 1, 703–708. Pu fa hl P. K., E. E. Hia tt (201 2) Oxy g e na tion of t he Ea rth’s a tmos phe re - oce a n sy s te m: A re vie w of phy s ica l a nd c he mica l s e dimentol og ic re s pons e s , M ari ne and Pet rol eum Geol ogy 32, 1 - 20 Re y me r A. a nd Schu bert G. (1 984). Pha ne rozoic a ddition ra te s to the contine nta l cru s t a nd cru s ta l g rowth. Tec t oni c s, 3 , 63- 77. Re y nol ds , B. , R. Ge org , F . Obe rl i, U . Wie che rt, A. Ha l l id a y , (2006). Re - as se s s me nt of s il icon is otope re fe re nce ma te ria ls u s ing hig h -re s ol u tion mu l ti - col le ctor I CP- MS, J. Anal . At . Sp ec t rom. , 21, 266- 269 Robe rt, F . & Cha us s idon, M. A (2006). Pa la e ote mpe ra tu re cu rve for the Pre ca mbria n oce a ns ba se d on s il icon is o tope s in che rts , N ature, 4 4 3 , 969- 972. Ros ing , M. T. , B ird, D. K. , Sl e e p, N. H. & B je rrum, (201 0) C. J. N at ure 4 64 , 744- 747. Ry e , R. , Ku o, P. H., Hol l a nd, H. D., (1 995). Atmos phe ric ca rbon dioxide c once ntra tions be fore 2. 2 bil l i on y e a rs ag o. N at ure 37 8 , 603–605 . Ry e , R. & Hol l and, H. D. (1 998). Pa le os ol s and the e vol u tion of a tmos phe ric oxy g e n: a critica l re vie w. Am eri c an Journal of Sc i enc e 298, 621 - 672. Sa g a n, C., a nd Mu ll e n, G. , (1 972). Ea rth a nd Ma rs : Evol u tion of a tm os phe re s a nd s u rfa ce te mpe ra tu re s. Sc i enc e, 177 , 5 2–5 6. Sa va g e P. S. , Ge org R. B. , Army tag e R. M. G. , Wil l ia ms H. M. a nd Ha ll ida y A. N. (201 0) Sil icon is ot ope hom og e ne ity in the ma ntle . Eart h Pl anet. Sc i . Let t . 295, 1 39–1 46. Sa va g e P. S. , Ge org R. B . , Will ia ms H. M., Bu rton K. W. , Ha ll ida y A. N (201 1 ) Sil icon is oto pe fra ctiona tion du ring ma g ma tic diffe rentia tion. Geoc hi m ic a et C osm oc hi m i c a Ac t a, 7 5, 61 24- 61 39 Schopf, J. W. , a nd Wa l ter, M. R. (1 983). Arche an microfos s il s : Ne w e vide nce of ancie nt microbe s . I n J. W. Schopf (Ed. ), Ea rth' s Earl ie s t B ios phe re , I ts Orig in a nd Evol u tion (Prince ton, NJ: Prince ton U niv. Pre s s), pp. 21 4 - 239. Schopf J. W. (1 992) Cra dl e of L ife: The Dis cove ry of Ea rth' s Ea rl ie s t F os s il s, Prince ton U nive rs ity Pres s , 336pp. Scribne r, A. M. , Ku rtz, A. C., a nd Cha dwick , O. A. (2006). Ge rma niu m se qu es tra tion by s oil : Ta rg e ting the rol e s of s e conda ry cla y s a nd Fe oxy hy droxide s. Eart h and Pl anet ary Sc i enc e Lett ers, 24 3 , 760- 770. She n B . , Le e C.- T. A. , Xia o S. (201 1 ) Ge rma niu m/s il ica ra tios in dia g e ne tic che rt nodu le s from the Edia ca ra n Dou s ha nt u o F orma tion, Sout h C hi na C hem . Geol . , 28 0, 323335 . Shibu y a T. , Komiy a T., Na ka mu ra K., Ta ka i K. , Ma ruy a ma S. (201 0), Hig hly a lk al ine , hig hte mpe ra tu re hy drothe rma l flu ids in the e arl y Arche an oce an, Prec am b ri an Researc h 18 2, 2 30–238 Sie ve r, R. (1 992). The s il ica cy cl e in the Pre ca mbria n. Geoc hi m . C osm oc him . Ac t a 56, 3265 –327 2 Simons on, B . M. , a nd Ha ss le r, S. W. , (1 996). Wa s the de pos ition o f l a rg e Pre ca mbria niron forma tions l ink e d to ma jor ma rine tra ns g res s ions ? Journal of Geol ogy , 104 , 665 −676. Sl e e p, N. H. , Za hnl e, K. , (2001 ). Ca rbon dioxide cy cl ing a nd impl ica tions for cl i ma te on a ncie nt ea rth. J. Geop hys. Res. 106, 1 373–1 399. Sl e e p N. H. a nd Hes s le r A.M. (2006). Wea thering of qu a rtz a s a n Archea n cl ima tic indica tor . Eart h and Pl anet ary Sc i enc e Lett ers , 24 1, 5 94- 602. Smithie s , R. H. , Cha mpion, D. C. , Va n Kra ne ndonk , M. J., Howa rd, H. M. a nd Hick man, A. H. , (2005 ). Mode rn -s tyl e su bdu ction proce s s es in the Me s oa rchae a n:

General Overview

55

g e oche mica l e vide nce from the 3. 1 2 Ga Whu ndo intra oce a nic a rc. Eart h Pl anet . Sc i . Let t. , 231, 221 –237. Ste inhoe fe l , G. , Horn, I. , von B l ancke nbu rg , F. , (2009). Micro - s ca l e tra cing of F e a nd Si is otope s ig na tu re s in ba nde d iron forma ti on u s ing fe mtos e cond l a se r a bl a tion. Geoc hi m . C osm oc hi m. Acta 7 3, 5 343–5 360. Ste inhoe fe l , G. , von B la ncke nbu rg , F. , H orn, I ., Konha u se r, K. O. , Be uk es , N. J. , Gu tzme r, J. , (201 0). De ciphe ring forma tion proce s s es of ba nde d iron forma tions from the Tra ns va al a nd the Ha me rs le y s u cce s s ions by combine d Si a nd F e is otope a na l ys is u s ing U V fe mtos e cond l as e r a bl a tion. Geoc him . C osm oc hi m . Act a 7 4, 2677 –2696. Ste inhoe fe l , G. , B reu e r, J. , von B l a nck e nbu rg , F . , Horn, I ., Ka czore k, D. & Somme r, M. (201 1 ). Microme te r s il icon is otope dia g nos tics of s oil s by U V fe mtos e cond l as e r a bl a tion. C hem ic al Geol ogy 28 6, 280- 289. Stra u b, K.L . , Ra in e y, F. A., and Widde l , F ., (1 999). Rhodovu l u m iodos u m s p. nov. a nd Rhodovu l u m robig inos u m s p. nov. , two ne w ma rine phototrophic fe rrou s - iron oxidiz ing pu rpl e ba cte ria . Int ernati onal Journal of Sy st em at i c Bac t eri ol ogy , 4 9, 729−735 . Su g ita ni, K., Horiu chi, Y. , Ada chi, M. , a nd Su g is a k i, R. , (1 996). Anoma l ou s ly l ow Al 2O3/TiO 2 va l u e s for Arche an che rts from the Pil ba ra bl ock , We s te rn Au s tral ia —Pos s ibl e e vide nce for e xte ns ive che mica l wea the ring on the ea rl y Ea rth . Prec amb ri an Re searc h, 8 0, 49–76. Su mne r D. Y. a nd Grotzing e r J. P. (2004). I mpl ica tions for Ne oa rcha e a n oce a n che mis try from prima ry ca rbona te mine ra l og y of the Ca mpbe l l ra nd -Ma l ma ni Pl a tform, Sout h Afri c a. Sed i m ent ol ogy , 51, 1 273- 1 299. Tré g u er P. , Nel s on D.M. , va n Be nne nk om A. J. , De Mas te r D. J., Le y na e rt A. , Qu é gu ine r B . (1 995 ). The s il ica bal a nce in the worl d ocea n : a ree s tima te. Sc i enc e, 268 , 375 379. Tre nda l l, A.F . a nd Bl ock l ey , J. G. (1 970) The I ron -F orma tions of the Pre ca mbria n Ha me rs l ey Grou p, Wes te rn Aus tra l ia. Geol ogi c al Surv ey West ern Aust ral i a Bul l et i n 119, 366 p. Va l l ey , J. W. , Pe ck , W. H., King , E. M. , Wil de , S. A. , (2002). A cool e a rl y e a rth. Geol ogy 30, 35 1 –35 4. Va n de n B oorn, S. H. J. M., Va n B e rg en, M. J. , Nijma n, W., Vroon, P. Z., (2007 ). Du al rol e of s e a wa te r a nd hy drothe rmal fl u ids in Ea rl y Arche a n che rt forma tion: Evide nce from s il icon is otope s . Geol ogy 35, 939- 942. Va n de n B oorn, S. H. J. M, va n B e rg en, M. J. , Vroon, P. Z. , de Vrie s , S. T, Nijma n, W. , (201 0). Sil icon is ot ope a nd tra ce e l e me nt cons tra ints on the orig in of ~3. 5 g a che rts: impl ica tions f or e a rl y Archae an ma rine e nvironme nts . G eoc hi m. C osm oc him . Ac t a 7 4 , 1 077 - 1 1 03. Va n Krane ndonk , M. J. , (2004). Archa e an te ctonics : a re vie w. Prec am b ri an Res. , 131, 1 43– 151. Va n Kra ne ndonk, M. J. , Smithie s, R. H. , Hick man, A. H. , Cha mpion, D. C. , (2007). Re vie w: s e cul a r te ctonic e vol u tion of Arche a n conti ne ntal cru s t: inte rpla y be twee n horizonta l a nd ve rtica l proce s s e s in the forma tion o f the Pil ba ra Cra ton, Au s tral ia . Terra N ov a 19, 1 –38. Va re l a, D. E. , C. J., Pride , M. A., B rze zins k i, (2004). B iol og ica l fra ctiona tion of s il icon is otope s in S ou the rn Oce a n su rfa ce wa te rs. Gl ob al Bi ogeoc hem . C yc l es 18 : GB 1 047, doi. 1 0. 1 029/2003GB 0021 40. Ve ize r J. , Comps ton W. , Hoe fs J. , a nd Ne l se n H. (1 982) Ma ntl e bu ffe ring of the e arl y oce a ns . N at urwi ssensc haft en 69, 1 73–1 80.

56

Chapter 1

Ve ize r J. (1 983) Ge ol og ic e v ol u tion of the Archa e an - Ea rl y Prote rozoic e a rth. I n Schopf. J. W. , e d. , Ea rth’s Ea rl ies t B ios phe re , Prince ton, N. J. , Prince ton Unive rs ity Pre ss , 240- 25 9. Von Da mm K. L ., Edmond J. M. , Gra nt B. G. a nd Me as u re s C. I . (1 985 ) Che mis try of s u bma rine hy drothe rma l s olu tions a t 21 _N, Ea s t Pa cific Ris e . Geoc him . C osm oc hi m . Act a 4 9, 21 97 –2220. Von Da mm K. L . , B is choff J. L . a nd Ros enba u e r R. J. (1 991 ) Qu a rtz s ol u bil ity in hy drothe rma l s e a wa te r: a n e xpe rime nta l s tu dy a nd e qua tion de s cribing qu a rtz s ol u bil ity for u p to 0. 5 M Na Cl s ol u tions . Am . J. Sc i . 291, 977 –1 007. von Pa ris , P. , Rau e r, H., L e e Grenfe l l, J. , Pa tzer, B. , He de l t, P. , Stra cke , B ., Trau tma nn, T. , Schre ie r, F ., (2008). Wa rming the e a rl y e a rth - CO2 re cons id e re d. Pl anet . Sp ac e Sc i . 56, 1 244–1 25 9. Wa l ke r J. C. G., P. B rimbl e combe , I ron and su l fur in the pre biol og ic oce a n, Pre ca mbria n Re s . 28 (1 985 ) 205 – 222. Wa l te r, X.A. , Pica zo, A. , Mira cl e, R. M., Vice nte , E. , Ca ma cho, A., Arag no, M. , a nd Zopfi, J. , (2009), Anae robic microbia l iron ox ida tion in a n ironme romictic l a k e [a bs . ]: Gol ds chmidt2009 C onfe re nce, Da vos , Switze rl a nd: Ge ochimica e t Cos mochimica Acta , v. 73, Su ppl e me nt 1 , p. A1 405. Wa ng , Y., Xu , H. , Me rino, E., Konis hi H. , (2009). Ge nera tion of ba nde d iron f orma ti ons by inte rna l dy na mics a nd le a ching of oce anic cru s t. N at ure Geosc i enc e , 2, doi: 1 0. 1 038/NGEO65 2. Widde l , F. , Schne l l, S. , He is ing , S., Ehre nre ich, A. , Ass mu s , B ., and Schink , B ., (1 993). F e rrous iron oxida tion by a noxy g e nic phototrophic ba cte ria . N at ure, 362, 834−8 36. Yin Q. Z. , S.B . Ja cobs en, K. Ya mas hita , J. B l ichert - Toft, P. Te l ou k , F . Al ba re de , (2002). A s hort time s ca l e for te rre s tria l pl ane t forma tion fro m H f –W chrono me try of me te orite s , N at ure 4 18 949– 95 2. Zie g le r, K., Cha dwick , O. A. , B rze zinsk i, M. A. & Ke l ly , E. F . (2005 a ). Na tu ra l va ria tions of δ 3 0 Si ra tios du ring prog re s s ive ba s a l t we a the ring , Ha wa iian I sl a nds. Geoc him . C osm oc hi m . Act a 69, 45 97 –461 0. Zie g le r, K. , Cha dwick , O. A. , White , A. F . & Brze zins k i, M. A. (2005 b). δ 3 0 Si s y s te ma tics in a g ra nitic s a prol ite , Pue rto Rico. Geol ogy 33 , 81 7 –820. Zie g le r K., You ng E. D. , Scha u ble E. A. a nd Wa s s on J. T. (201 0) Me ta l –s il ica te s il icon is otope fra ctiona tion in e ns ta tite me te orite s a nd cons tra ints on Ea rth’ s core forma tion. Ea rt h Pl anet . Sc i. Lett . 295, 487 –496.

General Overview

57

CHAPTER 2 ANALYTICAL DEVELOPMENTS

CHAPTER 2

Analytical developments

2.1

Outline

To get insight into the Archaean Si cycle, we investigated palaeosols, shales and BIFs for which Si- and Fe-rich layers were analysed separately. Then, various samples with contrasted matrices were analysed for silicon isotopes, Ge/Si ratios and Rare Earth Elements (REE). In BIF, the peculiar Fe-oxide rich matrix of Fe-rich layers combined with low REE+Y concentrations (down to ppb level in rock) required adaptation of the “classical” alkaline fusion and a correction for polyatomic interferences. The section 2.2 will focus on the developments of the REE+Y and Ge analysis by HR-ICP-MS. In addition, we improved an existing sample preparation method for Si isotopes by controlling the mass bias induced by anionic and organic matrices on silicon isotopic measurements by MC-ICP-MS. This new analytical method for Si isotope measurements is described in section 2.3.

Analytical developments

61

2.2

Developments in REE+Y and Ge analysis by HR-ICP-MS

Analysing REE+Y of BIFs is challenging in three aspects: (1) the dissolution of Fe-oxides rich samples; (2) the low REE+Y contents of BIFs (down to ppb levels in rock) and (3) the polyatomic interferences on Eu and Gd. First, a complete dissolution of Fe-oxides-rich layers of BIF is difficult to achieve. Using lithium metaborate fusion, a complete dissolution requires increased amount of alkaline flux compared to preconised sample to flux ratios (Abraham et al., 2008). Second, BIF display very low contents of REE+Y compared to commonly analysed rocks such as basalts or shales. Since we have to deal with increasing amounts of flux to allow dissolution of Fe-rich layers this decreases the REE+Y contents in solution and worsens blanks which then requires improve detection limits. Efforts in improving usual techniques were then oriented in two directions: (1) the determination of the optimal sample to flux ratio for Fe-rich layers dissolution and (2) the test of different flux purity in order to improve blanks. A third aspect concerns the correction of polyatomic ion interferences on heavy REE (HREE) caused by the production of Ba and light REE (LREE) oxides. 2.2.1 Sample dissolution A classical and rapid method of rock dissolution is based on the rock fusion with a lithium metaborate (LiBO2) flux. Powdered rock samples is mixed with excess LiBO2 powder in a platinium crucible and fused for 1h at 1000°C in a muffle furnace. The fusion beads were dissolved in 5% HNO3 and the solutions were analysed for major elements by ICP-AES or ICP-MS and Ba, Ge and REE+Y contents by HR-ICP-MS (Element 2). A sample to flux ratio of 1/3 is generally optimum (Crock et al., 1984) but Abraham et al. (2008) preconised a 1/6 ratio. A complete dissolution of paleaosols, shales and Si-rich samples was achieved using a 1/3 sample to flux ratio. On the other hand, Fe-oxides are particularly refractory minerals and Fe-rich layers of BIF were never dissolved using a 1/3 sample to flux ratio and not systematically using a 1/6 sample to flux ratio, probably depending on the Fe-oxides content. Then, a series of tests were performed to determine the appropriate sample to flux ratio to achieve complete dissolution of the Fe-oxides rich samples. Two BIF reference materials, IF-G and FeR-1 (standards from Service d’Analyse 62

Chapter 2

des Roches et des Minéraux and Geological Survey of Canada, respectively), were then fused using sample to flux ratios from 1/6 to 1/12 (Table 2.1). Whereas IF-G standard was dissolved with a 1/6 ratio, the dissolution of FeR-1 standard was achieved from a sample to flux ratio of 1/8. This is likely due to the higher Fe contents of FeR-1 compared to IF-G (Fe2O3 TOT of 75.86% and 55.85%, respectively). Si and Fe recoveries attest that dissolution was complete at a sample to flux ratio of 1/6 for IF-G and 1/8 for FeR-1 (Table 2.1). Then, dissolution of Fe-rich layers samples was performed using a 1/8 sample to flux ratio. IF-G IF-G IF-G

sample/flux ratio n Si recovery (%) Fe recovery (%) 1/6 2 103±2 101±1 1/8 3 99±5 100±3 1/10 3 98±12 96±11

FeR-1 FeR-1 FeR-1

1/8 1/10 1/12

4 5 5

97±2 97±8 95±3

96±3 95±5 93±7

Table 2.1 Si and Fe recoveries for fusions of IF-G and FeR-1 (BIF reference materials) using sample to flux ratio varying between 1/6 and 1/12.

2.2.2 HR-ICP-MS analysis The solutions from LiBO2 fusions were analysed for Ba, Ge and REE+Y concentrations by HR-ICP-MS (Element 2) in low-resolution mode with indium (In) as internal standard (Robinson et al., 1999). 2.2.2.1 Detection limits The consequence of the increased alkaline flux required to dissolve Fe-oxides in Fe-rich layers of BIF, is the increase of REE+Y detection limits. To minimize the blanks, we opted for a LiBO2 flux with a 99.999% purity (American Element) (Table 2.2).

American Element Alfa Aesar Merck

flux purity 99.999% 99.997% 98%

Ba La Ce 63 50 1.5 49 198 26.1 97 689 1894.1

Pr 0.13 1.60 1.95

Nd 0.45 6.97 5.61

Sm Eu Gd Tb Dy Y Ho Er Yb Lu Ge 0.15 0.6 2.7 0.10 0.10 3.2 0.08 2.4 0.05 1.24 - 5.5 n.a. 5.1 - 0.8 - 44 0.85 0.1 3.3 n.a. 0.12 18.9 0.061 0.61 0.1 0.06 8

Table 2.2 REE+Y, Ba and Ge concentrations (ppt) of procedural blanks using LiBO 2 fluxes of different purity.n.a. = not analysed

Analytical detection limits (ADL) are defined as ten times the standard deviation of 11 consecutive procedural blanks measurements (Table 2.3). As Si-rich and Fe-rich layers of Analytical developments

63

BIF were dissolved using different sample to flux ratios (1/3 and 1/8, respectively), both types of layers have their own ADL. Besides, both types of layers were not analysed on the same day. Then, we prefer to report ADL relevant to the day of analysis rather than average ADL. Except for Yb (as well as La and Gd for Fe-rich samples), ADL obtained always lie more than a factor ten below the abundance in the sample (Table 2.3). Sample/flux Type of layer Ba* La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Yb Lu Ge*

ADL 1/3 Si-rich 1.8 33 43 1.18 3.7 1.1 0.63 4.9 0.50 0.52 4.8 0.23 0.71 17.5 0.34 0.089

PG-1 bulk 1/3 Si-rich 7764 341 511 63 251 60 30 59 11 71 817 17 44 25 6.6 3.2

ADL 1/8 Fe-rich 2.6 2016 30 2.19 8.6 34.4 2.44 129.9 5.23 4.74 107.0 0.59 2.69 394.5 1.78 0.132

PG-4A 1/8 Fe-rich 2630 3081 4898 538 2041 434 198 491 65 374 2780 95 240 192 35.5 5.4

Table 2.3 Analytical detection limits as rock equivalent (ADL) calculated as ten times the standard deviation of 11 consecutive procedural blanks during one analytical session with a sample to flux ratio of 1/3 for Si-rich mesobands and 1/8 for Fe-rich mesobands. Concentrations are expressed in ppb except for elements marked by * (Ba and Ge) which are in ppm. For comparison, typical Si-rich and Fe-rich samples are reported.

2.2.2.2 Correction for interferences on Eu and Gd It has been well documented that oxides (and to a lesser extent hydroxides) of Ba and LREE can interfere with the HREE (Dulski, 1994; Robinson et al, 1999; Aries et al, 2000; Nakamura and Chang, 2007). These studies reported that most interferences are minor and that a correction was only necessary for 157

135

Ba16O on

151

Eu,

141

Pr16O and

140

Ce16OH on

Gd when the oxide (or hydroxide) contribution is more than 5% compared to the

analyte content. These interferences were corrected following the method of Aries et al. (2000) described below. For each element the equation 2.1 can be written. m

64

I  mX  

m 16

Y 16O   m 17 Z 16OH  Equation 2.1

Chapter 2

where mI is the measured intensity at mass m, mX+ is the contribution of the analyte X of mass m,

m-16 16

Y O+ and

m-17 16

Y OH+ are the contributions of the polyatomic interference,

oxides and hydroxides respectively. Then to determine the contribution of the analyte (X+) in the measured signal, the contribution of the interferences (YO+ or ZOH+) have to be quantified. For any element Y and W and at given plasma conditions, the following relations between oxide production ratios can be assumed to be constant during HR-ICP-MS analysis (Equation 2.2) (Lichte et al., 1987).

YO







Y  WO W   K Equation 2.2

where K is a constant. This equation remains valid for hydroxide production but we consider here only the case of oxide production. Then, for each sample, the contribution of YO+ to the analyte X+ can be calculated using the following equation 2.3:

 WO  YO  sample  K    W

   Y  sample  sample

Equation 2.3

The constant K can easily be deduced from measurements in monoelemental (here Ba 100ppb, Pr 1ppb and Th 1ppb) standard solutions. Thorium (Th) is generally chosen as normalizing element (W) as it presents two advantages: (1) 232Th concentrations in natural samples are negligible and (2) no natural element display the mass 248 implying that what is measured at this mass is exclusively 232Th16O. Then, a 0.5ppb Th spike was added to all samples and the contribution of the interference can be quantified using equation 2.4:

YO  sample

 YO     Y   ThO   Th 

   ThO        Th   s tan dard

   Y  sample  sample

Equation 2.4

Specific oxide formation rates during HR-ICP-MS analysis were about 0.25% for BaO+/Ba+, 0.35% for BaOH+/Ba+, 0.09% for CeOH+/Ce+, 4% for PrO+/Pr+ and 11% for ThO+/Th+. In most cases, interferences of analyte) whereas

135

Ba16O,

135

Ba16OH and

140

Ce16OH were negligible ( 8), a cation-specific resin can be used to separate silicon from positively charged ions (e.g., BioRad's DOWEX 50W-X12 200–400 mesh in H+ form). This method has been applied on different kinds of natural samples such as river water or rocks (Georg et al., 2006; Van den Boorn et al., 2009). It offers the advantages of being fast and allows the processing of very small quantities of sample (i.e., just the required amount for one isotopic analysis) with an excellent silicon recovery, without requiring HF. However, this method removes neither anionic species nor any other species that are not positively charged. Although it was first thought that the presence of sulphate in samples does not influence isotopic analyses (Georg et al., 2006), it has recently been reported by Van den Boorn et al. (2009) that the presence of sulphate in samples can induce an offset in silicon isotope measurements that becomes significant above SO42-/Si weight ratios of ca. 0.02. 68

Chapter 2

Such ratios are easily reached in sulphur-rich rocks or in river waters. To eliminate the interference caused by the presence of SO42-, Van den Boorn et al. (2009) proposed an additional purification step to remove sulphur by ignition under a constant stream of oxygen. Though quite effective, this method is valid solely for rock samples and requires specific equipment. We propose a simple methodology that extends the one described in Georg et al. (2006) to be valid for sulphate rich rocks and all river water samples. It is a combination of cationic purification, with anionic matrix additions and organic matter mineralization. Indeed, it is shown here that the presence of dissolved organic matter – referred below as dissolved organic carbon (DOC) – also causes non-isobaric matrix effects. 2.3.3 Material and methods 2.3.3.1 Instrumentation Si isotopes were measured using a Nu Plasma MC-ICP-MS (ULB, Brussels) in dry plasma mode with a Cetac Aridus II desolvating sample introduction system equipped with a PFA nebulizer and spraychamber. Mass bias and instrumental drift were corrected using a combined external Mg doping and standard-sample bracketing approach (Cardinal et al., 2003). All measurements were done in 2 mg L-1 Si and the Mg concentration was adjusted on a daily basis to get a signal intensity similar to that of the Si. Silicon isotopic ratios were measured relative to NBS28 silica sand standard (or in house standard: pro analysi Quartz from Merck). Data acquisition was done in medium resolution mode as in Abraham et al. (2008) 14N16O interference on the side of the

30

Si peak was resolved by measuring at the low-mass

30

Si peak, which is not affected by the interference (“pseudo high

resolution”). Data were acquired from numerous sessions spread over several months. 2.3.3.2 Material and sample preparation Two silicon isotopic references (BHVO-1 and Diatomite) and five natural samples with contrasted matrix (2 rock standards and 3 river waters) have been used to test the new procedure. Rock samples include a shale (SGR-1 standard from the U.S. Geological Survey) and an iron formation (FeR-1 standard from the Geological Survey of Canada). This Analytical developments

69

standard was also chosen owing to the growing interest in Si isotopes for banded iron formations (André et al., 2006; Steinhoefel et al., 2010). Diatomite (opal standard) and BHVO-1 (basalt) – two reference materials of known Si isotopic values (Abraham et al., 2008; Reynolds et al., 2007) with low sulphate content – were analysed to check the accuracy. The three river samples were selected from various environments owing to their contrasted SO42-/Si ratios and DOC concentrations (Table 2.4): Tana in Kenya (TN21), Congo River (CNG4/07), and Vuilbeek 10/08 (a small stream running in a forested region in Belgium). River samples were filtered through 0.2µm or 0.45µm filters immediately after sampling. Si content and SO42-/Si ratios of the samples are provided in Table 2.4. SO42-/Si ratios (wt.) in the selected samples vary up to 0.65, far above the 0.02 threshold reported by Van den Boorn et al. (2009) for the appearance of an offset in Si isotopic values. Anions and DOC contents were measured respectively by ionic chromatography and by total organic carbon analyzer on a highly modified Thermo HiPerTOC coupled to IRMS (Bouillon et al., 2006). Table 2.4 Natural ratios between sulphates, nitrate, chloride, and dissolved organic carbon (DOC) and Si—n.a. means not applicable

Rocks

Si concentration Natural SO42- / Si ratio Natural NO3- / Si ratio (by weight) (by weight) Si % (weight) Diatomite 46.7 ~0 ~0

Natural Cl / Si ratio Natural DOC/Si ratio (by weight) (by weight)

~0

~0

BHVO-1 SGR-1 FeR-1

23.3 13.2 7.9

< 0.01 0.35 0.1

n.a. n.a. n.a.

< 0.01 < 0.01 < 0.01

n.a. n.a. n.a.

0.65 0.16 0.26

0.13 < 0.02 n.a.

1.01 0.14 0.54

0.14 1.75 0.69

-1

mg L Si River waters TN21 8.3 CNG 4/07 5.5 Vuilbeek 10/08 12.1

Bracketing quartz standards and Diatomite were prepared following a digestion technique adapted from Georg et al. (2006) About 5 mg of powdered rocks were melted with a sodium hydroxide flux at 730°C for 10 minutes in silver crucibles (Silver Boats from Elemental Microanalysis, Ref. D5035), then immerged overnight in 50 ml of unacidified MilliQ water. Na+ was then removed by using a cation exchange resin as described by Georg et al. (2006). Proceeding so allowed us to produce matrix-free standards and references. For rock samples (SGR-1, BHVO-1 and FeR-1), the same digestion technique was used, but the recovery of the digestion was achieved in 30ml of water and acidified with HCl, following the recommendation of Fitoussi et al. (2009) to adjust the pH of solution between 2 and 2.4 prior loading on cationic resin, thereby preventing 70

Chapter 2

precipitation of other cations such as iron for FeR-1. As first underlined by Fitoussi et al. (2009), it is recommend to use a ratio of 1-5 mg sample for ~200 mg of solid NaOH pellet during the fusion process for Fe-rich rocks like FeR-1 to reach complete rock dissolution. Indeed, a lower flux-to-sample ratio was not always sufficient to dissolve such particularly resistant rock types, as part of Fe-oxides remained undissolved. While using the flux-tosample ratio from Fitoussi et al. (2009), the recovery of the digestion was 97 ±7 % for all samples and 99 ±3 % for FeR-1. After NaOH fusion, the solution was diluted to decrease the Na+ concentration at about 1.5 g L-1 before loading it on the cationic resin as incomplete cation removal was sometimes observed at higher concentrations. Here the cation concentration is probably too high for the adsorption kinetics of the resin. Samples with no previously published 30Si value (rivers, FeR-1 and SGR-1) have also been prepared by triethylamine-molybdate (TEA-Moly) co-precipitation, which removes all elements but Si and O (De La Rocha et al., 1996), and redissolved in dilute HF-HCl (Cardinal et al., 2003). Isotopic ratios measured after TEA-Moly purifications were used as reference values and compared with the values obtained with our new method. 2.3.3.3 Matrix effect counter measures After cationic exchange chromatography following Georg et al. (2006) and prior to MCICP-MS analyses, sulphuric acid (H2SO4, Merck Suprapur) is added to both the standard and the samples in order to reach the same SO42-/Si ratio, e.g., 20 mg L-1 or 100 mg L-1 SO42for 2 mg L-1 Si (i.e., 10/1 and 50/1 SO42-/Si mass ratios). Added sulphuric acid is thereby in large excess in comparison with the naturally occurring SO42- in the sample (i.e., at least ca. ten times the natural concentration). This imposes a similar SO42-/Si ratios in the sample and the standard solutions. Although the natural NO3- levels in river water samples are low (Table 2.4), NO3- has been added in a similar way up to 100 mg L-1 with bidistilled nitric acid in addition to the SO42- doping. Exceptionnally, Vuilbeek 10/08 sample had been acidified with HNO3 during sampling and was therefore doped with up to 1500 mg L-1 HNO3 for analyses, as well as its bracketing standards (Table 2.5). Note that a strong HNO3 matrix had already been used by Van den Boorn et al. (2006) without any noticeable increase of the 14N16O interference. Similarly, the presence of Cl- in rivers is resolved by the use of hydrochloric acid as a solvent in the solution (Merck Suprapur, HCl Analytical developments

71

2000 mg L-1, equivalent to 0.6% HCl by volume). SGR-1 has also been analysed after cationic purification without SO42- doping in order to test the importance of the offset on uncorrected samples in our analytical configuration. Table 2.5 Anion concentrations in samples after acid doping NO3

-

mg L Diatomite BHVO-1 SGR-1 FeR-1 TN21 CNG 4/07 Vuilbeek 10/08 Evaporated Diatomite

100 100 100 1500 100

Cl -1

-

mg L 2000 2000 2000 2000 2000 2000 2000 2000

SO4 -1

2-

mg L

-1

20 or 100 20 or 100 20 or 100 20 or 100 100 100 100 100

A possible analytical bias induced by dissolved organic matter cannot be resolved by balancing the contaminant in both the sample and the standard, as a doping solution reproducing the organic matrix would be too complex to create. Moreover, adding an important organic matrix in the solutions is not recommended, as it may clog the membrane of the desolvating unit, perturb analyses, and cause soot deposit in the mass spectrometer. It is therefore necessary to decompose the organic matter prior to the analyses. Various methods exist to mineralize the organic matrix (see Matusiewicz, 2010 for an overview), we opted here for photo-ozonolysis. Dissolved oxygen is transformed into ozone under the action of the UV-C and both ozone and UV-C will act to decompose DOC in CO2 that will not affect the analyses. For that purpose, 7.5 ml of the river sample is transferred in a 15ml Savillex® PFA vials and submitted to UV-C radiations (254nm and 185nm, low pressure mercury lamp from Heraeus Noblelight, reference GPH287T8VH/4) under constant oxygen bubbling during 0.5 to 3 hours. The UV lamp was placed as close as possible of the surface of the sample (~2 cm). PFA vials were chosen owing to their resistance to UV. The mineralization of the organic matter is carried out prior to the cationic purification to avoid contaminating the resin. Hydrogen peroxide (H2O2), often used for DOC destruction, was avoided as it was feared that this strong oxidizing agent would damage the resin. After the oxidation treatment, the river samples were processed as the previous ones, i.e. purified with the cationic resin and doped with SO42- and NO3before analysis. The heat of the UV lamp produces a slight evaporation of the sample of 72

Chapter 2

about 100 µL h-1. Since Si is not a volatile element, no isotopic fractionation due to this evaporation was expected. However, this has been checked through an evaporation test carried out on the Diatomite Si isotopic reference (Reynolds et al., 2006): 5mL of a solution of 5 mg L-1 Si has been evaporated down to 2.5mL by warming on a hot-plate at 60°C. The volume was readjusted to 5 mL with mQ water. This evaporation step was repeated 3 times. The standard was then purified for isotopic analyses following the same procedure as the other samples.

2.3.4 Results and discussion Isotopic results are presented in Table 2.6. For rock samples with no previously published 30Si value, the 30Si signatures were measured after TEA-Moly purification. SGR-1 exhibits a 30Si signature of +0.03 ± 0.06 ‰ (±1σ), within the range of crustal derived components (Ding et al., 1996). For the FeR-1, the 30Si signature is -0.17 ± 0.09 ‰, within the heaviest range of values published so far for banded iron formations (André et al., 2006; Steinhoefel et al., 2010). Among the river samples, the CNG 4/07 has a 30Si signature of +0.90 ‰, which is consistent with the previously published values for the Congo River (Cardinal et al., 2010). Vuilbeek 10/08 and TN21 present higher 30Si signatures of +2.03 ‰ and +1.79 ‰, respectively. If we now compare the results for SGR-1 with both the TEA-Moly method and the cationic purification method as described by Georg et al. (2006) (i.e., not corrected for the presence of SO42-), an offset of δ30Si of +0.28 ‰ is observed for a SO42-/Si ratio of 0.35. Although this offset is significant, it is much smaller than the large offsets observed by Van den Boorn et al. (2009) at similar SO42-/Si ratio (more than +1 ‰ for δ30Si). Several factors could explain this difference. It is possible that the Nu Plasma MC-ICP-MS used here and by Georg et al. (2006) is less sensitive to a sulphate matrix effect than the ThermoFinnigan Neptune MC-ICP-MS used by Van den Boorn et al. (2009) Another possibility is that the acid matrix has an impact on the induced bias. Indeed, these samples were analysed in a 0.6% HCl matrix while Van den Boorn et al. (2009) used a 1% HNO3 matrix. In the two following sections, results show that the δ30Si offsets observed are indeed due to the presence of SO42- and/or DOC and that our proposed doping method is able to reconcile TEA-Moly vs. cationic purifications. Note that we also tested Analytical developments

73

74

0.05 0.06 0.09 0.09 0.05 0.06

-0.17 0.01 -0.09 0.92 0.47 1.05

0.65

29

δ Si ‰

0.03 0.07 0.05 0.05 0.03 0.06

0.07

SD publ. 5 9 5 5 4

publ.

n

0.31

30

0.13

δ Si ‰ SD

Cationic Resin

0.15

29

0.07

δ Si ‰ SD

5

n -0.32 0.03 -0.24 1.80 1.50 2.23 1.24

1.26

30

δ Si ‰ 0.05 0.09 0.09 0.05 0.12 0.15 0.08

0.07

SD -0.16 0.02 -0.09 0.93 0.82 1.15 0.60

0.03 0.03 0.06 0.05 0.06 0.11 0.06

0.67 0.04

29

δ Si ‰ SD

Cationic Res. + doping

8 6 5 5 4 10 7

11

n

0.85 1.94

30

0.09 0.08

δ Si ‰ SD

0.43 1.01

29

0.04 0.07

δ Si ‰ SD

4 6

n

Cationic Res. + dop. + mineralization

Table 2.6 Measured δ30Si and δ29Si values for the different purification techniques : the TEA-Moly removes all elements but Si and O and is used as a reference to compare with the other values. The cationic resin columns list results obtained by purification of the samples on the cationic resin only, and with the addition of the additional steps resolving the matrix effects of anionic species (SO 42- doping for rocks and SO42- and NO3- dopings for rivers) and organic matter (mineralization). Note that the doped Diatoite was analysed either with sulfate doping or with sufate and nitrate doping. Published values (italic) for standards Diatomite and BHVO standards are from the literature (Reynolds et al., 2006; Abraham et al., 2008, respectively). The standard error is given as one standard deviation (SD) on the n measurements.

0.10

1.25

SD

BHVO-1 -0.33 SGR-1 0.03 FeR-1 -0.17 TN21 1.79 CNG 4/07 0.90 Vuilbeek 10/08 2.03 Evaporated Diatomite

Diatomite

δ Si ‰

30

TEA-Moly or published value

the possibility of removing anionic species from the solution by means of a weak anion-

exchange resin in Cl- form (BioRad’s AG4-X4), but we could not achieve complete SO42-

removal (SO42-/Si > 0.02) despite the complete Si recovery (data not presented).

Chapter 2

2.3.4.1 Doped samples For the two SO42--free reference materials (Diatomite and BHVO-1), we observe that the doping in SO42- does not induce any bias in comparison with the well constrained published values, since the difference in the mean δ30Si is less than 0.01 ‰, much smaller than the standard deviation (1σ = 0.07 ‰). This result confirms that if SO42- is present in a similar quantity in both the sample and the bracketing standards, its effects are nil. Similarly, the good result for the diatomite confirms that the addition of NO3- does not cause a problematic increase of the 14N16O interference on the

30

Si peak when using a

desolvating introduction system and measuring at the low-mass side of the 30Si peak. For the other rock samples and the TN21 river water (low DOC/Si ratio) the comparison between values obtained after TEA-Moly purification with the values measured on the sample purified with cationic resin and with SO42-doping shows no significant difference. The isotopic ratios being identical within error, confirms that the doping of SO42efficiently corrects the offset induced by the natural presence of SO42-. The combined mean δ30Si values from TEA-Moly and doped resin-purified samples for SGR-1 are +0.03 ±0.08 ‰ (n = 11) and for FeR-1 -0.20 ±0.10 ‰ (n = 14). Similarly, the doping with NO3- in the TN21 river water also corrects for natural nitrate-induced mass bias. However, δ30Si ratios measured on river water samples with high DOC/Si ratios (Vuilbeek 10/08 and CNG4/07) remain far too high in comparison with the isotopic signatures measured after TEA-Moly purification (shift of δ30Si signatures of +0.29‰ and +0.65‰, respectively) probably as a result of their high DOC. Counteracting the effect of high DOC concentrations on mass bias, was investigated further in the following section. 2.3.4.2 UV treated samples The test of DOC mineralization was carried out on the river samples CNG 4/07 and Vuilbeek 10/08 with initial DOC concentrations of 9.64 mg L-1 C and 8.37 mg L-1 C (±0.13), respectively. Results show that most of the DOC had disappeared after only 0.5 hours of exposure to UV-C and O3 (Fig 2.1), and reached a stable low concentration after one hour. The DOC/Si ratios is then below 0.05. Since DOC can exhibit various complex forms it is recommended however to expose samples for two hours at least, which should be Analytical developments

75

enough for most river samples, including black rivers (e.g., the Rio Negro is mostly below 26.6 mg L-1 C; Day and Davies, 1986). The method for DOC mineralization presented here is only one oxidation process among many others (Matusiewicz, 2010). It has been chosen because it does not require the addition of any contaminating reactant. However, despite good results for river samples, this oxidation process is rather mild and is inappropriate to mineralize very high charges of organic matter such as those of soil solutions, where the amount of humic acid can be high and very resistant to oxidation. Indeed, this oxidation step was tested on soil solutions and in some samples, DOC contents were still above 100 mg L-1 after 2 hours exposure time (data not shown). After the mineralization step, δ30Si results (Table 2.6) are in agreement with those obtained after TEA-Moly purification, confirming that the DOC was responsible for the bias reported in the results from section 1.3.4.1. Furthermore, such bias can be avoided by mineralizing most of the DOC initially present using UV-C and ozone. With a lower DOC/Si ratio of 0.14, the δ30Si signature measured in the untreated TN21 river sample was not affected by the presence of DOC. However, it seems difficult to predict the influence of each specific organic matrix and it is recommend to treat every sample containing significant amounts of DOC relative to Si systematically with this method. Average precision on long term isotopic measurements expressed as ±1σ standard deviation are ±0.07‰ and ±0.04‰ (Diatomite, n = 11) for 30Si and 29Si respectively. No clear difference was observed in average precision between the samples purified with the TEA-Moly method and samples purified by resin and treated against DOC. After the evaporation test on the diatomite, a δ30Si composition of +1.24 ±0.08 ‰ (n = 7) was measured, which is similar to the recommended value of the diatomite (+1.25 ‰). This shows that the three evaporation steps have no effect on the isotopic ratio of the diatomite, thus evaporation induced by the heat of the UV lamp is not an issue for δ30Si measurements. Moreover, this implies that evaporation can be used to concentrate samples in which Si concentration is too low for an accurate isotopic measurement.

76

Chapter 2

10.00 10

DOC - Vuilbeek 10/08 DOC - CNG 4/07 DOC/Si - Vuilbeek 10/08

1.00

DOC/Si

DOC (mg/L)

8

DOC/Si - CNG 4/07

6

4

0.10

2

0.01

0 0

1

2

3

Exposure time (h)

Fig. 2.1 DOC concentration and DOC/Si ratio in the Vuilbeek and the Congo water samples after exposure to UV-C and ozone treatment.

2.3.5 Conclusion Results confirm that sulphate matrix effects may bias silicon isotopic measurements when only using cationic chromatography purification. This is also the first time published results show a matrix effect due to the presence of dissolved organic matter, a common component of waters, particularly in tropical rivers and rivers under high anthropogenic pressure. These results provide a strong evidence that the purification step in the Si isotopes analyses is of extreme importance and also that the purification does not need to be fully complete as long as the level of contaminant is known and can be balanced both in the sample and the bracketing standard. This can be easily solved for sulphate by adding sulphuric acid. Similarly, the presence of NO3- in the river samples seems to be efficiently corrected by adding nitric acid, with this method likely being used for other contaminating anions. Results also show that the important mass offset induced by dissolved organic matter in river samples can be remedied by decomposing with UV-C and oxygen bubbling. Nevertheless, our results confirm that the extent of matrix effect cannot be directly transferable from one laboratory to another, probably because of a combination of instrument and sample types, analytical settings, and/or purification processing (e.g., Georg et al., 2006 vs. Van den Boorn et al., 2009 vs. this study). Therefore we recommend that each lab to test the applicability of the doping ratios we used to their specific routine procedure. Finally, the first δ30Si values of two contrasted geostandards are given, making them potential rock reference materials once values Analytical developments

77

have been confirmed by further independent laboratories: SGR-1 (shale) with a δ30Si ratio of +0.03 ±0.08 ‰ and and FeR-1 (banded iron formation) with a δ30Si ratio of -0.20 ±0.10 ‰.

78

Chapter 2

References Abra ha m K. , S. Opfe rg e l t, F. F ripia t, A. J. Ca vag na , J. de Jong , S. F . F ol e y, L . André a nd D. Ca rdina l, Ge os ta nd. Ge oa nal . Re s. , 2008, 32, 1 93 - 202. André L . , D. Ca rdinal , L . Y. All e man and S. Moorba th, Ea rth Pla ne t. Sci. Le tt. , 2006, 245 , 1 62- 1 73. Arie s S. , Va l la don M. , Pol vé M. a nd Du pré B . , Ge os ta nda rd. Ne wl e tt., 2000, 24, 1 9 - 31 . B a rl ing J. a nd D. We is , J. Ana l . At. Spe ctrom. , 2 008, 23, 1 01 7 - 1 025 . B ou il l on S. , M. Kor ntheu e r, W. B ae ye ns a nd F. De ha irs, L imnol . Oce anog r. : Me thods , 2006, 4, 21 6- 2 26. Ca rdina l D., L . Y. Al le ma n, J. D. Jong , K. Zie gle r a nd L. André, J. Ana l. At. Spe ctrom. , 2003, 1 8, 21 3- 21 8. Ca rdina l D. , J. Ga ill a rde t, H. J. Hu ghe s, S. Opfe rg el t a nd L . Andr é , Ge ophy s. Re s . L e tt. , 201 0, 37, L 1 2403. Crock J. G. , L ichte F .E. , and Wil de ma n T. R. , Chem. Ge ol . , 1 984, 45 , 1 49 - 1 63. Da y J. A. a nd B . R. Da vies , e d. B . R. Da vies a nd K. F . Wa lk e r, Monog ra phiae B iol og ica e 60. W. Ju nk , Dordre cht. 793 pp. 1 986, ch. The Ama z on Ri ve r Sy s te m, pp. 289 31 8. De L a Rocha C. L ., M. A. B rze zinsk i a nd M. J. DeNiro, Ana l . Che m., 1 996, 68, 3746 - 375 0. Du l sk i P. F res e n. J. Ana l. Che m. , 1 994, 35 0, 1 94 -203. Ding T. , S. Jia ng, D. Wan, Y. L i, J. L i, H. Song , Z. L iu a nd X. Ya o, Sil icon Is otop e Ge oche mis try , Ge ol og ica l Pu bl is hing Hou s e , B e ijing , China . 1 996. F itou s s i C. , B . B ou rdon, T. Kl e ine , F . Obe rl i a nd B . C. Re y nol ds , Ea rth Pl a ne t. Sci. L e tt. , 2009. Ge org R. B . , B. C. Re ynol ds , M. F ra nk a nd A. N. Ha l l ida y, Che m. Ge ol ., 2006, 235 , 95 - 1 04. L ichte F .E. , Me ie r A.L . and Crock J. G. , Anal . Chem. , 1 987, 5 9, 11 5 0 -1 1 5 7. Na k a mu ra K. and Cha ng Q. Ge os ta nda rd. Ne wlett. , 2007, 31 , 1 85 - 1 97. Ma tu s ie wicz H. , e d. J. Na mie snik a nd P. Sze fer, Ta yl or a nd F ra ncis Grou p - CRC Pre ss . 201 0, ch. Mine ral iza tion te chni qu es u se d in the s a mpl e pre pa rtion s te p, pp. 95 1 02 Re y nol ds B . C., J. Ag g a rwal , L . André , D. Ba xte r, C. B eu che r, M. A. Brze zinsk i, E. Eng s tröm, R. B . Ge org, M. La nd, M. J. L eng , S. Opfe rg e l t, I . Rodu shk in, H. J. Sl oa ne , S. H. J. M. va n de n B oorn, P. Z. V roon a nd D. Ca rdina l , J. Anal . At. Spe ctrom. , 2007, 2 2, 5 61 - 5 68. Robins on P. , Towns e nd A. T. , Yu Z. a nd Mü nk er C. Ge os ta nda rd. Ne wle tt. , 1 999, 23, 31 46. Ste inhoe fe l G. , F. von B l a ncke nbu rg , I. Horn, K. O. Konha u se r, N. J. B eu ke s a nd J. Gu tzme r, Ge ochim. C os mochim. Acta , 201 0, 74, 2677 - 2696. DOI : DOI 1 0. 1 01 6/j. g ca. 201 0. 01 . 028. Va n de n B oorn S. H. J. M, P. Z. Vr oon a nd M. J. B e rge n, J. Anal . At. Spe ctrom. , 2009, 24, 1 1 1 1 -1 1 1 4. Va n de n B oorn S. H. J. M. , P. Z. Vroon, C. C. va n Be l le , B . va n de r Wa g t, J. Schwi e te rs a nd M. J. va n B e rge n, J. Ana l. At. Spe ctrom. , 2006, 21 , 734 - 742. DOI : Doi 1 0. 1 039/B 600933f.

Analytical developments

79

CHAPTER 3 DESILICATION IN ARCHAEAN WEATHERING PROCESSES TRACED BY SILICON ISOTOPES AND GE/SI RATIOS

CHAPTER 3

Desilication in Archaean weathering processes traced by silicon isotopes and Ge/Si ratios

3.1

Abstract

The geochemical characteristics of the Archaean ocean and their relation with continental weathering are poorly understood. Here we combine for the first time Si isotopes and Ge/Si ratios on an Archaean palaeosol (~2.95 Ga) as proxies of the weathering processes and Si mass transfers at the early Earth’s surface. Neoformation of secondary clay minerals in modern soils is associated with fractionation of Si isotopes and Ge/Si ratios with increasing weathering degree and desilication (e.g., Cornelis et al., 2010; Opfergelt et al., 2010; Steinhoefel et al., 2011). Likewise, we show that the observed δ30Si and Ge/Si palaeosol fingerprints were controlled by the neoformation of secondary minerals during Si mobilization and the leaching of Fe under low pO2 conditions. This potentially released fluids to stream waters with peculiar δ30Si and Ge/Si signatures estimated between -0.09 to +0.61‰ and 1.33 to 1.96µmol/mol, respectively. We further demonstrate that a combination of δ30Si and Ge/Si tracers in 2.95 Ga Archaean shales can estimate the relative proportions of desilicated soil-derived clays resulting from previously identified weathering processes. The important contribution of these clays in shales is indicative of a high degree of weathering of the source. Therefore, we suspect an important Si flux to the rivers and ultimately to the ocean at that time.

* Adapted from Delvigne C., Opfergelt S.., Cardinal D. and André L. Desilication in Archaean weathering processes traced by silicon isotopes and Ge/Si ratios. To be submitted

Desilication during Archaean weathering

81

3.2

Introduction

It is commonly argued that in the Archaean ocean hydrothermal inputs dominated over continental inputs in contrast to the modern days. However, recent studies demonstrated the strong impact of continental inputs especially on epicontinental seas geochemistry (e.g., Alexander et al., 2008). Recent advances in analytical capabilities have added Si isotopes and Ge/Si ratios to the tracers of modern weathering. Relative to crustal rocks displaying little variations (Ge/Si between 1 and 3µmol/mol and δ30Si about 0±0.4‰; Douthitt, 1982; Mortlock and Froelich, 1987; Ding et al., 1996), neoformed clay minerals incorporate preferentially light Si isotopes and Ge relative to Si, resulting in lighter δ30Si values from -2.59 to -0.53‰ and higher Ge/Si ratios (4.8-6.3µmol/mol) (e.g., Ziegler et al., 2005b; Opfergelt et al., 2010; Cornelis et al., 2010; Lugolobi et al., 2010). δ30Si and Ge/Si values in modern soils are therefore controlled by the relative proportion of primary minerals and secondary clay-sized minerals, i.e. the degree of weathering (Ziegler et al., 2005b; Opfergelt et al., 2010). These tracers have not yet been employed to evaluate palaeoweathering despite the advantage that Si isotopes are insensitive to metamorphic processes (André et al., 2006). The Pongola Supergroup in South Africa offers the opportunity to investigate Archaean weathering processes in a 2.95 Ga palaeosol and to evaluate their significance in near-contemporaneous shales.

3.3

Methods

Major element concentrations in palaeosols and shales samples were analysed by XRF spectrometry (Philips X’Unique XRFS) at University of KwaZulu-Natal. The accuracy of the major element determinations was checked against international standard NIM-G and is better than 3 %. For Si isotopes and Ge/Si ratios determinations, shale samples were milled in agate while palaeosol samples were milled in carbon steel. About 50mg of powdered rock samples were dissolved by fusion with lithium metaborate fusion (99.999% purity; American Element) at 1000°C for 1h in platinum crucibles with a 1:8 flux to sample ratio. The fusion beads were dissolved in 5% HNO3 and the solutions were analysed for Ge contents. Ge 82

Chapter 3

concentrations were analysed by HR-ICP-MS (Element 2) in low resolution mode with indium (In) as internal standard. Typical accuracy for Ge is ±2% as checked with geostandards and 0.1 ppb Ge artificial solution. Digestion procedure, chromatic purification and instrumentation settings carried out to measure Si isotope compositions are detailed by Hughes and coworkers (2011a). Si isotopic ratios were measured at least in duplicate (when possible) using a Nu Plasma multicollector plasma source mass spectrometer (MC-ICP-MS) at ULB (Brussels, Belgium). A SO4 spike (20ppm) was added both in samples and standards to avoid offsets due to potentially high sulphur contents (Hughes et al., 2011a). Silicon isotope data are reported relative to the international quartz reference material NBS-28 using delta notation δ30Si or δ29Si. The Diatomite secondary reference material was included in every sequence to check for accuracy and precision, and yielded a δ30Si of +1.26 ± 0.07 (1SD, n=11) well in agreement with recommended value (+1.25 ± 0.13‰, Reynolds et al., 2007).

3.4

Palaeosols

Palaeosols in the c. 2.98 to 2.87 Ga Pongola Supergroup have been suspected since long (Button and Tyler, 1981). A well preserved palaeoweathering profile was recently identified in a drill core (TSB07-26) 1.5 km north of Denny Dalton Mine (28°16'6.20"S, 031°13'24.33"E) in the White Mfolozi Inlier immediately below the unconformity that separates the Mozaan Group from the Nsuze Group (Fig. 3.1). This alteration profile was investigated by Nhleko (2003) in a different borehole who concluded that it conforms to all criteria that should identify a palaeosol. It developed on amygdaloidal basaltic andesite of the Nsuze Group and was subjected to greenschist-facies metamorphism. The uppermost part of the altered volcanic rock to about 190 cm below the palaeosurface consists mainly of sericite occurring with minor chlorite and rutile (sericite-dominated zone, ZS). Immediately below the ZS there is a chlorite-dominated zone (ZC). This zonation is known from other Precambrian palaeosols, e.g. the Mt Roe and Hekpoort palaeosols (Macfarlane et al., 1994; Rye and Holland, 2000) where the presence of two distinct zones is likely inherited from pedogenic processes.

Desilication during Archaean weathering

83

Shale and interlayed sandstone

Dolerite

Medium grained sandstone

Borehole TSB 07-26

a

Green shale BIF Black shale Shale samples Grit Small pebble conglomerate Grit Pebbly quartzite Dolerite Pebbly quartzite Palaeosol Lavas

b

Fig. 3.1 Location of the studied area: (a) Geological map of the study area showing the extent of Pongola Supergroup, location of White Mfolozi Inlier and the borehole TSB 07-26. (b) Stratigraphic log of the drill core under study (TSB 07-26) indicating sampling depths for studied shale and palaeosol samples.

While kaolinite or smectite would be expected during modern andesite weathering, postweathering processes likely explain the association of sericite and chlorite. Sericitization is attributed to reverse weathering reactions in which kaolinite and/or gibbsite are replaced metasomatically through the percolation of K-rich waters during burial diagenesis or low-temperature metamorphism (Macfarlane et al., 1994; Rye and Holland, 2000). Greenalite and/or Fe(II)-rich smectite are considered as pedogenic precursor minerals that transformed into chlorite plus quartz during diagenesis and greenschist facies metamorphism (Macfarlane et al., 1994; Rye and Holland, 2000). The chemical composition of the palaeosol varies with depth (see Box 3.1). The ZC is characterised by major depletion of mobile elements such as Ca, Na and K, by slight increases of Fe and Ge contents, and a limited desilication. In contrast, abrupt changes in chemistry mark the ZS characterised by a complete loss of Fe, and to a lesser extent, Ge, by a strong Kenrichment and a significant desilication. From elemental loss or enrichment calculations (τ index, see Box 3.1), it appears that the mobility of Si and Ge throughout the palaeoprofile are consistent with weathering processes, which supports the use of δ30Si and Ge/Si values as tracers of weathering processes.

84

Chapter 3

Box 3.1 Elemental mobility in the palaeosol Elemental loss or enrichment of the palaeosol profile were estimated through elemental mass balance calculation using the immobile index element approach (e.g., Kurtz et al., 2000). As zircon is present as a trace mineral in the parental andesitic material, Ti was preferred over Zr as an immobile element. The mass fraction (τ) of a mobile element (j) lost from or added to a soil horizon relative to Ti, originally present in the parent material is calculated from the equation 3.1: τ j,w = ([j]w [Ti]p / [Ti]w [j]p )-1 Eq. 3.1 where [j] is concentration of an element in the weathered (w) and parent (p) materials. Positive and negative values of τ indicate gain or loss of a mobile element, respectively. This calculation accounts for changes in bulk soil density and collapse or dilatation of the horizon during soil development (Kurtz et al., 2000).

0

meters below palaeosol top

-2 -4 -6 -8

Si vs depth Fe vs depth Ge vs depth Al vs depth K vs depth

-10 -12 -14 -16 -1,0

-0,5

0,0

0,5

1,0

2,0

3,0

j

Fig. 3.2. Depletion index (τ j) of major elements (j = Si, Fe, Ge, Al and K) relative to Ti in the palaeosol profile studied. Parental andesite, chloritic and serictitic zones are represented by dark-, mid-grey and light grey area, respectively.

The calculations of the τj index for the different elements (j = Si, Fe, Ge, Al and K) show that Si, Fe and Ge display very similar patterns throughout the profile where 62% of Si, 98% of Fe and 76% of Ge present in the parent rock have been lost within the uppermost 1m of the palaeoweathering profile (Fig. 3.2). Significant loss of Fe is common in Precambrian palaeosols (Rye and Holland, 1998), but contrasts with constant Fe concentrations in modern weathering profiles formed under oxic conditions, where dissolved Fe2+ is oxidized to Fe3+ and remains as Fe(III)-oxides in the profile (Maynard, 1992). It is generally considered that Fe loss is typical of subaerial weathering at low pO2 prior to 2.2Ga (Rye and Holland, 1998). However, diagenetic processes can mimic this Fe depletion even under an oxidizing atmosphere (Ohmoto et al., 1996). This can be excluded here as Fe enrichment in the upper part of the ZC supports a pedogenic process (Maynard, 1992). A strong correlation between τFe and τGe shows the siderophile behaviour of Ge (r²=0.96; pvalue60% Fe2O3 TOT) or as Si-Fe-rich mesobands (subequal contents of SiO2 and Fe2O3 TOT). Al2O3 and MnO are present as minor elements with contents from 0 to 1.24% and 0.04 to 2.33%, respectively. Mg, Ca, Na, Ti and P are trace elements, whereas K2O was not measurable (Table 4.1).

108

Chapter 4

Insights into BIF deposition

109

30

n.a.