Cerium Oxidation State in Silicate Melts and the Application to Ce … · 2015. 6. 10. · Duane J. Smythe Doctor of Philosophy Department of Earth Sciences University of Toronto

i

Cerium Oxidation State in Silicate Melts and the Application to Ce-in-Zircon Oxygen Barometry

by

Duane John Smythe

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Earth Sciences University of Toronto

© Copyright by Duane J. Smythe 2013

ii

Cerium Oxidation State in Silicate Melts and the Application to

Ce-in-Zircon Oxygen Barometry

Duane J. Smythe

Doctor of Philosophy

Department of Earth Sciences University of Toronto

2013

Abstract

Through measurements of cerium redox equilibrium in silicate melts, and an internally-

consistent model for zircon-melt partitioning of Ce, a calibration for estimating the redox

conditions during zircon crystallization is presented. To quantify Ce4+/Ce3+ in natural magmas,

Ce doped silicate glasses ranging in composition from basalt to rhyolite (± H2O) at 0.001 and 1

GPa, under fO2 conditions varying from FMQ –6.0 to +8.4 (where FMQ is the fayalite-

magnetite-quartz oxygen buffer), and temperatures from 1200 to 1500°C were synthesized.

Samples were analyzed using both synchrotron-based Ce M4,5-edge x-ray absorption near edge

structure (XANES) spectroscopy and traditional wet chemical techniques. Based on comparison

to potentiometric determinations of Ce4+/Ce3+ in digested portions of our samples, Ce4+ was

found to have a greater fluorescence yield (FLY) response to the incident radiation in the M4,5-

edge XANES region than Ce3+. A correction to M-edge determined Ce4+/Ce3+ ratios as follows:

4 4 ( 0.004) 4XANESCe / Ce 578.7[Ce / Ce 5%] 0 Ce / Ce 0.4A

where the exponent A is a function of the number of non-bridging oxygens per tetrahedrally-

coordinated cation (NBO/T) for a given glass composition, 4XANES[Ce / ΣCe] is the contribution

iii

from Ce4+ to the M4,5-edge area over the total area of the M4,5-edge, and 4Ce / ΣCe is the actual

fraction of Ce4+ in the sample.

For a given melt composition, the change in Ce4+/Ce3+ with fO2 follows the thermodynamically

predicted trend assuming simple oxides as melt species. In addition to fO2, melt composition and

water content have been found to be secondary controls on Ce4+/Ce3+, with more depolymerized

melts and hydrous compositions favoring the stabilization of Ce3+. The Ce4+/Ce3+ ratio can be

expressed through the equation,

4

23

2

Ce 1 5705( 257) NBOlog log O 0.8694( 0.005)

Ce 4 T

3.856( 0.049) H O 3.889( 0.102)

fT

x

where T is in Kelvin, and xH2O is the mole fraction of water in the melt.

Using this model, fO2 for samples from the Bishop tuff (USA), Toba tuff (Indonesia) and Nain

plutonic suite (Canada), have been calculated and are in excellent agreement (typically within

0.5 log units) with independent estimates. Using the composition of the Jack Hills Hadean

zircons, combined with estimates of parental magma composition, the fO2 during zircon

crystallization was determined to be between FMQ -1.0 to +2.5, suggesting that relatively

oxidized magmatic source regions, similar to 3.5 Ga komatiite suites, existed by ~4.4 Ga.

iv

Acknowledgments

I would first like to express my gratitude to James Brenan for introducing me to the field of

experimental petrology and for his mentorship over the course of this project. Thank you for

providing me with the opportunity to do this project, I feel privileged to have worked with such a

knowledgeable and patient teacher. Again, I’m sorry about the desk.

Secondly, I would like to thank Neil Bennett who has been a better colleague than one could

hope for. Our discussions in the lab, or over a couple pints, were invaluable and will be missed.

I am indebted to Jim Mungal, and Mike Hamilton (and Grant Henderson I suppose) for their

guidance and encouragement over throughout my PhD. Jim, the years we spent running Benny

were some of my best times at U of T.

I am grateful to George Kretschmann, Boris Foursenko, Yanan Liu, Colin Bray, Mike Gorton,

Tom Regier and the rest of the technical and office staff who made this work possible.

Thanks also to the University of Toronto and the Government of Ontario for financial support, as

well as National Science and Engineering Research Council and the Geological Society of

America for funding of this project.

To finish, I would like to thank my family and friends for all their support. There are far too

many to name, but this includes in part my sister Margo and brothers Brad and Cory, as well as

Kristy, Flanders, Jordan, Tups, Jan, Ben and Remo. In particular, thank you to my mother, who

somehow managed to work two jobs, raise four troublemaking children and get them all through

university single handed. You are truly an inspiration.

v

Table of Contents

Abstract ........................................................................................................................................... ii

Acknowledgments .......................................................................................................................... iv

Table of Contents ............................................................................................................................ v

List of Tables .................................................................................................................................. x

List of Figures ................................................................................................................................ xi

List of Appendices ....................................................................................................................... xiv

Chapter 1: Introduction ................................................................................................................... 1

1.1 Applications of Zircon in Geology ..................................................................................... 1

1.2 Zircon Structure and Chemistry .......................................................................................... 2

1.3 Zircon/Melt Partitioning ..................................................................................................... 4

1.3.1 Factors Affecting Trace Element Uptake by Zircon ............................................... 4

1.3.1.1 Equilibrium Effects ................................................................................... 4

1.3.1.2 Disequilibrium Effects .............................................................................. 5

1.3.1.3 Analytical Artifacts ................................................................................... 6

1.3.2 Summary of Zircon/Melt Partitioning Studies ........................................................ 6

1.3.2.1 Studies on Natural Samples ....................................................................... 6

1.3.2.2 Experimental Studies ................................................................................. 8

1.3.3 Ti-Thermometry ...................................................................................................... 9

1.4 Zircon as a Measure of Oxygen Fugacity ......................................................................... 10

vi

1.5 The Early Earth ................................................................................................................. 12

1.6 Contributions of this Thesis .............................................................................................. 13

1.6.1 Calibration of Cerium M4,5-edge XANES for Ce4+/Ce3+ Determinations in

Silicate Melts ........................................................................................................ 14

1.6.2 Redox Behavior of Ce in Silicate Melts ............................................................... 14

1.6.3 Derivation of a Ce-in-Zircon Oxygen Barometer ................................................. 15

1.6.4 Author Contributions ............................................................................................ 15

1.7 References ......................................................................................................................... 13

Chapter 2: Quantitative Determination of Cerium Oxidation States in Alkali-Aluminosilicate

Glasses using M4,5-edge XANES ............................................................................................. 26

2.1 Introduction ....................................................................................................................... 26

2.2 Material and Methods ....................................................................................................... 28

2.2.1 Preparation of Glasses ........................................................................................... 28

2.2.2 Analytical ............................................................................................................... 29

2.2.2.1 Electron Probe Micro-Analysis ............................................................... 29

2.2.2.2 Potentiometric Titrations ........................................................................ 30

2.2.2.3 M-edge XANES ...................................................................................... 30

2.3 Results ............................................................................................................................... 31

2.3.1 Potentiometric Titrations ...................................................................................... 31

2.3.2 XANES ................................................................................................................. 32

2.4 Discussion ......................................................................................................................... 34

vii

2.4.1 Potentiometric Method vs. XANES ...................................................................... 34

2.4.2 Cerium Valence Determinations ........................................................................... 36

2.5 Conclusions ....................................................................................................................... 37

2.6 References ......................................................................................................................... 38

Chapter 3: Determination of Cerium Oxidation State in Silicate Melts: Combined fO2,

Composition, and Temperature Effects .................................................................................... 50

3.1 Introduction and Theoretical Background ........................................................................ 50

3.2 Materials and Methods ...................................................................................................... 54

3.3 Analytical .......................................................................................................................... 56

3.3.1 Electron Microprobe ............................................................................................. 56

3.3.2 Potentiometric Titrations ...................................................................................... 56

3.3.3 Ce M4,5-edge XANES ........................................................................................... 57

3.4 Results ............................................................................................................................... 58

3.4.1 Effect of Temperature ........................................................................................... 58

3.4.2 Effect of Oxygen Fugacity .................................................................................... 59

3.4.3 Melt Composition ................................................................................................. 59

3.4.4 Water Content ....................................................................................................... 60

3.5 Discussion ......................................................................................................................... 60

3.5.1 Quench Modification of Ce Oxidation State ........................................................ 60

3.5.2 Comparison to Previous Work .............................................................................. 63

3.5.3 Geological Implications ........................................................................................ 64

viii

3.6 Conclusions ....................................................................................................................... 66

3.7 References ......................................................................................................................... 67

Chapter 4: Experimental Calibration of a Ce-in-Zircon Oxygen Barometer ................................ 90

4.1 Introduction ....................................................................................................................... 90

4.2 Samples Investigated ........................................................................................................ 92

4.3 Analytical Methods ........................................................................................................... 94

4.4 Lattice Strain Constraints on Ce Partitioning ................................................................... 95

4.5 Comparison to Independent Estimates of Magma fO2 and Temperature ........................... 96

4.6 Estimation of the Redox State of Hadean Magmas .......................................................... 97

4.7 Summary and Conclusions ............................................................................................. 100

4.8 References ....................................................................................................................... 100

Chapter 5: Conclusions ............................................................................................................... 111

Appendix 1: Methods for Potentiometric Titrations ................................................................... 113

A1.1 Sample Preparation ....................................................................................................... 113

A1.2 Preparation of Titrant .................................................................................................... 114

A1.3 Calibration .................................................................................................................... 114

A1.4 Measurement ................................................................................................................ 114

Appendix 2: Compositions of Experimental Run Products ........................................................ 117

A2.1 Calculation of NBO/T and xH2O .................................................................................. 122

Appendix 3: Curve Fitting Parameters of XANES Spectra ........................................................ 125

ix

Appendix 4: Supplementary Information for Chapter 4 ............................................................. 138

A4.1 Zircon Chemistry .......................................................................................................... 138

A4.2 Glass/Whole Rock Compositions ................................................................................. 150

A4.3 Derivation of Melt Ce4+/Ce3+ from Zircon/Melt Partitioning ....................................... 151

A4.4 Evaluation of Hadean Melt Trace Element Budget ...................................................... 153

x

List of Tables

Table 2.1 Bulk composition of experimental starting materials .................................................. 41

Table 2.2 Average curve parameters used to fit Ce M4,5-edge ..................................................... 42

Table 2.3 Results of potentiometric and XANES analyses of glasses ......................................... 43

Table 3.1 Bulk composition of experimental stating materials .................................................... 73

Table 3.2 Experiment run conditions ........................................................................................... 74

Table 3.2 Electron microprobe analyses of glass run products .................................................... 76

Table A2.1 Electron microprobe analyses of all glass run products .......................................... 117

Table A3.1 Curve parameters used to fit Ce M4,5-edge. ........................................................... 125

Table A4.1 Major and trace element data for Bishop tuff zircon .............................................. 138

Table A4.2 Major and trace element data for Toba tuff zircon .................................................. 142

Table A4.3 Major and trace element data for Umiakovik pluton zircon ................................... 147

Table A4.4 Major and trace element compositions for Bishop tuff and Toba tuff glasses and

whole rock analysis of Umiakovik pluton ............................................................................. 150

xi

List of Figures

Figure 1.1 Schematic representation of the lattice strain model showing the variation in

partition coefficient Di, for a series of isovalent cations, as a function of cation radius ......... 23

Figure 1.2 Partition coefficients for REE between zircon and melt from selected studies .......... 24

Figure 1.3 Onuma diagrams, plotting the variation in log /zircon meltiD with ionic radius for

selected zircon/melt partitioning studies .................................................................................. 25

Figure 2.1 Calibration curve for potentiometric titrations ........................................................... 44

Figure 2.2 Plot of log Ce4+/Ce3+ versus log fO2 for RH08 and BH09 compositions ................... 45

Figure 2.3 Background corrected Ce M4,5-edge XANES spectra ................................................ 46

Figure 2.4 Example of curve fitting used to determine Ce4+/Ce3+. .............................................. 47

Figure 2.5 Time series of Ce M4,5-edge under beam exposure .................................................... 48

Figure 2.6 Comparison between Ce4+ fractions determined using the potentiometric method

and the contribution to the total M-edge area from Ce4+ ......................................................... 49

Figure 3.1 Experimental assemblies for low fO2 experiments done at 1 atm and piston

cylinder experiments done at 1 GPa......................................................................................... 81

Figure 3.2 Ce M4,5-edge XANES spectra of CeF3, CeO2 and glass ............................................. 82

Figure 3.3 Arrhenius plot of the variation in the Ce redox equilibrium versus temperature ....... 83

Figure 3.4 Ce4+/Ce3+ as function of fO2, expressed relative to the fayalite-magnetite-quartz

oxygen buffer. .......................................................................................................................... 84

Figure 3.5 Compositional dependence of Ce4+/Ce3+ versus the dry NBO/T values for the

different compositions investigated in this study ..................................................................... 85

Figure 3.6 Measured versus calculated ln(Ce4+/Ce3+) comparing data determined in this

study with previous measurements on silicate glasses ............................................................. 86

xii

Figure 3.7 Plot of log (CeO2/CeO3/2) versus log fO2 determined by zircon-melt partitioning

experiments from Burnham and Berry (2012) ......................................................................... 87

Figure 3.8 Alkali content a) and Al/Si b) of the glass versus the compositional term ‘B’ ......... 88

Figure 3.9 Chondrite normalized REE diagrams showing the predicted magnitude of Ce and

Eu anomalies in zircon crystalizing from a hydrous granitic melt ........................................... 89

Figure 4.1 Representative zircon textures from the Bishop tuff ................................................ 105

Figure 4.2 Cathodoluminescence images of typical zircons from the Toba tuff ....................... 106

Figure 4.3 Representative textures of zircons from Umiakovik pluton ..................................... 107

Figure 4.4 Graphical representation of the procedure used to estimate oxygen fugacity from

a combination of zircon-melt partitioning of Ce, and the calibrated relation between melt

4 3Ce Ce/melt meltx x vs fO2 ................................................................................................................... 108

Figure 4.5 Values of log fO2 calculated using the Ce-in-zircon oxygen barometer presented

in this study as a function of temperature for zircons from the Bishop tuff, Toba tuff, and

Umiakovik pluton, compared to independent estimates ........................................................ 109

Figure 4.6 Diagram depicting the possible fO2 range of the >4.0 Ga Jack Hills zircons ........... 110

Figure A1.1 Set-up for potentiometric titrations ........................................................................ 115

Figure A1.2 Plots of electrode potential versus volume and the first-derivative cure for an

example titration..................................................................................................................... 116

Figure A4.1 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a pyrolite

parent melt composition ......................................................................................................... 153

Figure A4.2 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a trace

element parent melt composition of bulk continental crust ................................................... 154

xiii

Figure A4.3 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a trace

element parent melt composition of Archean TTG ................................................................ 155

xiv

List of Appendices

Appendix 1 Methods for Potentiometric Titrations ................................................................... 113

A1.1 Sample Preparation ............................................................................................. 107

A1.2 Preparation of Titrant .......................................................................................... 108

A1.3 Calibration ........................................................................................................... 108

A1.4 Measurement ....................................................................................................... 108

Appendix 2 Compositions of Experimental Run Products ........................................................ 117

Appendix 3 Curve Fitting Parameters of XANES Spectra ........................................................ 125

Appendix 4 Supplementary Information for Chapter 4 ............................................................. 138

A4.1 Zircon Chemistry ................................................................................................ 129

A4.2 Glass/Whole Rock Compositions ....................................................................... 144

A4.3 Derivation of Melt Ce4+/Ce3+ from Zircon/Melt Partitioning ............................. 145

A4.4 Evaluation of Hadean Melt Trace Element Budget ............................................ 145

1

Chapter 1

Introduction

1.1 Applications of Zircon in Geology

The concentration of zirconium in the silicate portion of the Earth is only 10.5 ppm (McDonough

and Sun, 1995), however, due to its high incompatibility in most common rock forming minerals,

this can be sufficient to reach saturation in magmatic systems, forming zircon or baddeleyite,

depending on silica activity. This effect accounts for the widespread occurrence of zircon in

felsic igneous rocks as well as, metamorphic rocks, some mantle xenoliths, lunar rocks and

meteorites (Hoskin and Schaltegger, 2003). Zircon is also one of the most stable heavy minerals

in the surface environment, resisting chemical and physical abrasion, often surviving multiple

episodes of weathering, transport, burial and metamorphism. During crystallization, zircon is

capable of incorporating a number of trace elements into its structure. The sluggish diffusion

rates of these elements through the structure of crystalline zircon (Cherniak and Watson, 2003)

have made it an invaluable tool in geochemical studies, significantly aiding in our current

understanding of the formation and evolution of the Earth and solar system.

One of the most fundamental tasks in geology as a scientific discipline is to establish a historical

record for the Earth. This would not be possible without the means to obtain accurate absolute

ages for geological events. As zircon commonly contains uranium in concentrations orders of

magnitude higher than its host rock, and excludes lead from its structure during crystallization, it

has been used extensively for radiometric dating. Since the time of the initial study by Strutt

(1909), who first realized the potential of zircon as a geochronometer, several methods have been

refined for the measurement of zircon crystallization ages. These primarily involve the decay

series; 238 20692 82U Pb , 235 207

92 82U Pb , and 232 20890 82Th Pb , which have halflives of 4.468 x 109,

0.738 x 109 and 14.010 x 109 years, respectively (Steiger and Jäger, 1977).

In addition to crystallization age, the composition of zircon can provide complimentary

information regarding crystallization environment. The much higher compatibility of hafnium

relative to lutetium in the zircon structure results in the lack of ingrowth of radiogenic 176Hf

(produced by the decay of 176Lu), preserving the initial Hf isotopic composition at the time of

2

crystallization. Due to the fractionation of Hf from Lu during mantle melting, this can in some

cases offer a means to trace crust/mantle differentiation (Amelin et al., 1999; Harrison et al.,

2005). The absolute and relative trace element abundances in zircon have been shown to be

sensitive to source rock type allowing for recognition of the provenance of detrital grains

(Heaman et al., 1990; Hoskin and Schaltegger, 2003). The development of secondary-ion mass-

spectrometry (SIMS) has permitted accurate and precise determinations of zircon oxygen

isotopic compositions (Valley, 2003) which can used to recognize crustal inputs to the zircon

parent melt. These advances in SIMS, as well as laser-ablation inductively-coupled-plasma

mass-spectrometry (LA-ICP-MS), have allowed for analysis of zircon with increasingly high

spatial resolution. Coupled with cathodoluminescence (CL) and back-scattered electron (BSE)

imaging, detailed crystallization histories for single zircon grains can be obtained and

interpreted, providing a powerful tool for understanding geological events over a protracted time

scale.

1.2 Zircon Structure and Chemistry

Zircon is a zirconium orthosilicate with Zr in 8-fold triangular dodecahedral coordination and Si

in tetrahedral coordination with O. The SiO4 tetrahedra are corner and edge sharing with ZrO8

dodecahedra. The dodecahedra themselves are edge sharing forming chains parallel to <100>

(Finch and Hanchar, 2003). Each oxygen atom is bonded to one Si (Si-O = 1.62 Å) and two Zr

(Zr-O = 2.13 and 2.27 Å). Zircon is tetragonal, crystalizing in the I41/amd space group, with four

ZrSiO4 per unit cell (Hazen and Finger, 1979). The mineral has a relatively open structure, with

octahedral voids connecting to form channels between the Si tetrahedra and Zr dodecahedra.

These channels have been suggested to act as interstitial sites, incorporating species that would

otherwise not easily be accommodated by the zircon structure (Speer, 1980; Finch and Hanchar,

2003).

Hafnon (HfSiO2) is the only mineral that has been shown to form a complete solid solution with

zircon (Hoskin and Rodgers, 1996). Other minerals which are isostructural with zircon

(including thorite, coffinite, xenotime) have only limited solubilities (Hoskin and Schaltegger,

2003). Due to the similar behavior of Zr and Hf in most igneous systems, the zircon-hafnon

solid solution is typically restricted to a molar Zr/Hf ratio close to that of chondrite (~19,

McDonough and Sun, 1995), although ratios ranging from 200 – 2 have been observed (Hoskin

3

and Schaltegger, 2003, and references therein). Linnen and Keppler (2002) showed that changes

in the Zr/Hf activity coefficient ratio from values close to unity, as is the case for most natural

melt compositions, can occur in a relatively narrow compositional range when a nearly fully

polymerized melt structure is approached. For peraluminous melts Linnen and Keppler (2002)

estimate Zr to be 2 to 5 times more compatible in zircon relative to Hf, consequently zircon

fractionation will decrease the Zr/Hf ratio in some granites.

Zircon is capable of incorporating a number of trace elements during crystallization, consisting

primarily of Y, P, REE, Th, U, which can be present at concentrations on the order of 10’s to

1000’s of ppm. Other elements detected in zircon by in situ techniques are typically at the 10’s of

ppm level or below, and include: Li, Be, B, F, Na, Mg, Al, Ca, Sc, Ti, V, Cr, Mn, Fe, Ga, Sr,

Nb, Ba, and Ta. The incorporation of trace elements in zircon can be accommodated through

either simple isovalent substitution or more complex coupled substitution mechanisms involving

single or multiple structural sites or interstitial sites (Finch et al., 2001; Hoskin and Schaltegger,

2003).

Examples of simple substitutions are:

M4+ = Zr4+ (where M4+ is Hf, U, Th, or Sn)

Ti4+ = Si4+

Coupled substitutions include:

One structural site

M3+ + M5+ = 2Zr4+ (where M3+ is Y, or REE and M5+ is Nb, or Ta)

Two structural sites

M3+ + P5+ = Zr4+ + Si4+ (where M3+ is Y, REE, or Sc)

Interstitial Site

Li +(int)

+ M3+ = Zr4+ (where M3+ is Y, REE, or Sc)

Zr 4+(int)

+ 4M3+ = 4Zr4+ (where M3+ is Y, REE, or Sc)

4

M 2(int)

+ 3M3+ + P5+ = 3Zr4+ + Si4+ (where M 2(int)

is Mg or Fe and M3+ is Y or REE)

M 3(int)

+ 4M3+ + P5+ = 4Zr4+ + Si4+ (where M 3(int)

is Al or Fe and M3+ is Y or REE)

More complicated substitution mechanisms involving vacancies at either O2-, Si4+ or Zr4+ sites

have also been proposed to maintain charge neutrality in cases where the total REE concentration

is lower or exceeds that of P (Hanchar et al., 2001). The incorporation of OH- in zircon has been

suggested to occur at Si vacancies through a hydrogrossular substitution (4H+ = Si4+) as well as

at Zr vacancies (Caruba et al., 1985; Trail et al., 2011a).

1.3 Zircon/Melt Partitioning

1.3.1 Factors Affecting Trace Element Uptake by Zircon

1.3.1.1 Equilibrium Effects

The Nernst partition coefficient (D) for element i is define as the equilibrium concentration of i

in phase A divided by the concentration of i in phase B ( /A BiD = iA/iB). For trace elements (i.e.

those which obey Henry’s Law) /A BiD is a constant for a given set of intensive parameters.

Knowledge of the partition coefficients can, therefore, be used as a tool to determine the trace

element composition of a melt from which a zircon crystallized.

The parabolic distribution of trace element partition coefficients versus their ionic radius was

first noted by Onuma et al. (1968), where they concluded that the crystal/melt partitioning of

trace elements was controlled primarily by the crystal structure. Jensen (1973) suggested the

shape these patterns were dependent, not only on the radius of the substituting cations, but also

the ‘elasticity’ of the crystal structure. Expanding of the work of Brice (1975), Blundy and Wood

(1994) proposed a thermodynamically based quantification for this observation where the

mineral/melt partition coefficients are related to the strain energy caused by the difference in

ionic radius of the substitution cation (ri) from that of the ideal ionic radius for a given structural

site (r0). This relation can be expressed by the equation,

5

2 30

0 0

0

14

2 3exp

A i i

i

rEN r r r r

D DRT

(1.1)

where D0 is the ‘strain compensated partition coefficient’, E is the Young’s Modulus, NA is

Avogadro’s number, R is the gas constant, and T is temperature in degrees K. A graphical

representation of this model is shown in Figure 1.1. This shows that the expected partition

coefficients for a series of isovalent cations will decrease with the difference in ri from r0 for a

given structural site. Application of this model to partitioning between zircon and melt predicts

highly selective incorporation of REEs into zircon, with the light-REEs (LREEs) being strongly

excluded due to their relatively large ionic radius.

1.3.1.2 Disequilibrium Effects

The slow diffusion rates for highly charged cations through silicate melts (Mungall, 2002) and

within the zircon structure (Cherniak and Watson, 2003; Cherniak, 2010) can result in kinetic

effects being a dominant control on the incorporation of trace elements into zircon.

Disequilibrium partitioning between zircon and melt can result from either the development of a

diffusive boundary layer, or through surface absorption and entrapment during crystal growth

(Watson, 1996). In the case of absorption, elements attached at the crystal surface will initially

be in equilibrium with the host melt. Due to the abundance of defects associated with crystal

faces, this surface layer may have a greater diversity of sites or a lower Young’s Modulus than

the crystal interior, therefore, equilibrium element concentrations at the surface may not be

equivalent to those within the crystal structure. With crystal growth, ions attached at the surface

can become buried within the structure. If the growth rate of the crystal is greater than the rate of

diffusion for a buried ion to the surface layer, it will then become trapped. As a means of

predicting the conditions at which entrapment will be important, Watson (1996) used the Péclet

number (Pe), equal to the ratio of the diffusive time scale (l2/Di; l = length, Di = diffusivity) to

the growth time scale (l/V; V = growth rate). Watson (1996) showed that for values of Pe >0.1,

surface entrapment is possible, and will be highly efficient at values >10. In typical granitic

systems Pe values will range for 3 to 12 for zircon, suggesting it will be highly susceptible to this

process.

6

Development of a diffusive boundary in the melt will occur if the rate of crystal growth is greater

than the rate of diffusion of an ion in the melt at the crystal/melt interface. This causes an

enrichment of incompatible elements and depletion in compatible elements in the melt

immediately adjacent to the growing crystal (Watson and Müller, 2009). As highly charged

cations will be most prone to this effect due to their slow diffusivities through silicate melts,

zircons grown relatively rapidly may be in disequilibrium with the bulk melt. This will result in

higher /zircon meltiD values for incompatible elements and lower /zircon melt

iD values for compatible

elements compared to by equilibrium partitioning.

1.3.1.3 Analytical Artifacts

Zircons commonly contain inclusions of other accessory phases, such as monazite, and apatite.

Each of these minerals will contain LREE in concentration several orders of magnitude higher

than that of zircon. The incorporation of such LREE-rich minerals during analysis of zircon can

therefore result in an overestimation of the apparent zircon/melt partition coefficients for these

elements (Nagasawa, 1970; Mahood and Hildreth, 1983; Murali et al., 1983; Fugimaki, 1986;

Hinton and Upton, 1991). Although large inclusions are easily avoided during zircon analysis,

Jain et al. (2001) showed that incorporation of as little as 0.01 to 0.001 vol% can significantly

affect the measured REE concentrations.

1.3.2 Summary of Zircon/Melt Partitioning Studies

1.3.2.1 Studies on Natural Samples

Studies of zircon/melt partitioning using natural samples can be divided into two groups; those

which analyze zircon separates through bulk methods, involving digestion of multiple zircon

grains, and those which analyze single grains through micro-beam techniques. Results of these

studies are shown in Figure 1.2a. The first study on partitioning of trace elements between zircon

and melt was carried out by Nagasawa (1970), by the measurement of REE concentrations of

zircons, through bulk dissolution, and their host granites and dacites. The zircon separate

measure in this study was estimated to be ~99% pure, however, as discussed previously, the

incorporation of even small amounts of other LREE-bearing phases, will result in significant

overestimation of the calculation of partition coefficients. This likely accounts for deviation of

the /zircon meltiD values for LREEs measured by Nagasawa (1970) from lattice strain predictions

7

(shown in Figure 1.3a), as discussed by Hanchar and van Westrenen (2007). Furthermore, since

zircon can potentially have a protracted growth history (Claiborne et al., 2010), regions within

single zircon grains may have crystallized from a parent melt with a different trace element

composition than the host rock. These same problems are associated with all studies involving

the bulk dissolution of multiple zircon grains such as Murali et al. (1983), Mahood and Hildreth

(1983) and Fujimaki (1986).

The introduction of SIMS and LA-ICP-MS has allowed for the in situ analysis of low

concentration trace elements in zircon with a spot resolution as small as ~15 m, significantly

improving the reliability of zircon trace element data. Bea et al. (1994) conducted a LA-ICP-MS

study on the trace element partitioning between zircon and leucosome from a peraluminous

migmatite. Their results show similarly high /zircon meltiD values for the both HREE and LREE,

which, as was the case for bulk samples, is likely the result of LREE rich phases incorporated in

the analytical volume. The results of Bea et al. (1994) illustrate that even micro-analytical

techniques can be subject to some of the same issues as bulk analysis.

Sano et al. (2002) conducted a SIMS study of zircon and apatite from the Torihama Dacite

pyroclastic pumice. The primary purpose of this study was to present a method for measurements

of REE abundance in silicate glass, apatite and zircon, using the sensitive high-resolution ion

microprobe (SHRIMP). However, a significant result of this study, was accurate estimates of

zircon/melt and apatite/melt partition coefficients for REE consistent with theoretical

expectations. Their results show the LREEs to be highly incompatible in zircon suggesting

measured concentrations have not been subject to accessory mineral contamination. Similar

findings were obtained by Thomas et al. (2002) in their SIMS study of melt inclusions hosted

within zircon from a calc-alkaline tonalite from the Quottoon Igneous Complex, British

Columbia. Other studies carried out looking at the partitioning of trace elements between zircon

and melt using micro beam techniques include Marshall et al. (2009) and Nardi et al. (2013).

Results from Nardi et al. (2013) show similar LREE enrichment as observed by Bea. et al.

(1994) suggesting incorporation of accessary phases in the zircon analysis. Partition coefficients

determined by Marshall et al. (2009) are in close agreement with the results of Sano et al. (2002).

8

1.3.2.2 Experimental Studies

The first experimental study on the partitioning of trace elements between zircon and melt was

conducted by Watson (1980). In that work, zircon was crystallized from a felsic peralkaline melt

doped at wt% levels of La Sm, Ho, and Lu. Experiments were done at 800°C and 2 kbar and run

products were characterized by electron probe micro-analysis (EPMA). Dickinson et al. (1980)

examined REE partitioning between zircon and two immiscible liquids. The starting

compositions used in this study were a 50:50 wt% mixture of the immiscible liquids produced by

the fractional crystallization of lunar KREEP basalt. Experiments were done at 1185°C and 1

atm in Mo-foil capsules sealed inside silica glass tubes. It is difficult to evaluate whether the

results of this study represent equilibrium, as no description of the samples, nor the method by

which the run products were measured, is provided. Partitioning of REE and Hf between zircon

and liquid was investigated by Okano et al. (1987). Zircons were synthesized at 1000°C and 1

GPa and separated from the matrix by first crushing followed by concentration using a Frantz

isodynamic separator. Analysis was performed using isotope dilution thermal ionization mass

spectrometry (ID-TIMS). All of these studies yielded similar results, showing an increase in

partition coefficient with increasing REE atomic number, but the degree of HREE/LREE

fractionation was smaller compared to results from studies of natural samples (Figure 1.2b). In

the study by Watson (1980) a decrease in compatibility from Ho to Lu was observed, which is

likely an analytical artifact as DLu was determined by a mass balance approach in this study. It

should also be noted that the studies by both Watson (1980) and Okano et al. (1987) failed to add

charge balancing ions, such as P5+, in their experiments to accommodate the incorporation of the

3+ REE in the zircon structure. Rubatto and Hermann (2007) measured partitioning of REEs, Hf,

Th, and U, between zircon and a hydrous granitic melt, and observed a decrease in /zircon meltiD

with increasing temperature for all elements investigated.

In more recent work, Lou and Ayers (2009) investigated zircon/melt partitioning in a hydrous

peralkaline rhyolite, similar to that studied by Watson (1980). Although their results are in broad

agreement with theoretical expectations, they concluded that the relatively subdued HREE/LREE

fractionation they observe likely arises from kinetic effects, and suggest the development of a

diffusive boundary layer in the melt. Lou and Ayers (2009) propose that these kinetic effects will

affect all experimental investigations of zircon/melt partitioning due to the slow diffusivities of

the REEs in the melt and relatively rapid zircon growth rates. These effects would result in under

9

estimations of /zircon meltiD values for compatible elements and over estimations for incompatible

elements. In a subsequent study attempting to overcome these potential disequilibrium processes,

Burnham and Berry (2012) synthesized zircon in a synthetic andesitic melt at the relatively high

temperature of ~1300C, and anhydrous conditions. The authors concluded that diffusive

boundary layers were unlikely in their experiments, in contrast to Luo and Ayers (2009).

However, the partition coefficients measured by Burnham and Berry (2012) still show a smaller

/ //zircon melt zircon meltHREE LREED D than seen in natural samples.

1.3.3 Ti-Thermometry

Substitution of Ti into the zircon structure was shown by Watson and Harrison (2005) to be a

strong function of the crystallization temperature. This has subsequently been refined by Watson

et al. (2006) and most recently by Ferry and Watson (2007) where the original assertion of Ti

substitution for Zr was reevaluated. Instead, these authors propose Ti substitutes into zircon for

Si, by the reaction:

4 2 4 2ZrSiO TiO ZrTiO SiO (1.2)

This suggest the partitioning of Ti into zircon is dependent on the activity of TiO2 (aTiO2) and

SiO2 (aSiO2). Ferry and Watson (2007) provide the following relation to calculate zircon

crystallization temperature:

2 2

4800 86log(Ti ) (5.711 0.072) log SiO log TiO

(K)Zircon a aT

(1.3)

where Ti zircon is the concentration of Ti in ppm. This allows for the thermometer to be used in

quartz and rutile under-saturated systems if aSiO2 and aTiO2 can be estimated. Fu et al. (2007),

however, in a study investigating the Ti concentration of 484 zircons from a large range of

igneous compositions, showed that temperatures calculated using the Ti-in-zircon thermometer

consistently lower than either zircon saturation temperatures or predicted crystallization

temperatures. These authors conclude that effects of aSiO2 and aTiO2 are insufficient to account

for these discrepancies, and suggest the relatively low Ti-in-zircon temperatures are the result of

other factors, including pressure, diviations from Henryès Law, errors in the calibration,

disequilibrium crystallization.

10

1.4 Zircon as a Measure of Oxygen Fugacity

A number of elements exist in multiple oxidation states over the terrestrial range of oxygen

fugacity (fO2), several of which can be incorporated into zircon during crystallization, include

Eu, Ce and U (Philpotts, 1970; Schrieber, 1980; Schrieber, 1982). If zircon effectively

discriminates between the oxidation states of these elements during crystallization, than there

exists the potential to estimate the prevailing redox conditions at the time of growth. Since both

Ce and Eu in the trivalent state are expected behave the same as the other REEs, with

partitioning varying systematically with ionic radius, deviations from this behaviour result in

either an excess or deficit in the expected abundances of these elements. Such “anomalies” in

the behavior of these elements are usually defined as the difference in chondrite-normalised

abundance relative to that expected from the interpolation between the neighbouring REE;

positive or negative anomalies denote a higher or lower abundance, respectively, than the

interpolated value. Positive chondrite-normalized Ce anomalies and negative Eu anomalies are a

common observation in natural zircons (Figure 1.2a) and result from the favored partitioning of

Ce4+ and the exclusion of Eu2+ relative to their trivalent counterparts. Recent studies have

confirmed that the partitioning of Ce and Eu into zircon is sensitive to the redox state of the

system (Trail et al., 2011b; Trail et al., 2012; Burnham and Berry, 2012). Europium anomalies in

zircon, however, will be strongly affected by the crystallization of a Ca-bearing phase, such as

plagioclase, which has a high affinity for Eu2+. Crystallization Ca-rich phases predate that of

zircon in igneous systems, and will also be present in the source region of melts eventually

reaching zircon saturation. Thus, Eu anomalies will typically be the result, at least in part, of

inheritance from the parent melt; this excludes the Eu content of zircon as a reliable redox

indicator. As no common rock forming minerals will incorporate either Ce3+ or Ce4+ in

appreciable amounts, Ce anomalies in zircon are not subject to the same problem as Eu. These

anomalies can then be used to calculate the Ce4+/Ce3+ ratio of the parent melt in the following

manner. The concentration of Ce in zircon and melt is a mixture of Ce3+ and Ce4+, expressed as:

3 4Ce Ce Cezircon zircon zircon (1.4)

3 4Ce Ce Cemelt melt melt (1.5)

11

Using partition coefficients for the individual oxidation states and substituting into Equation 1.4

yields,

3 4

3 / 4 /

Ce CeCe Ce Cezircon melt zircon melt

zircon melt meltD D (1.6)

Combining Equations 1.5 and 1.6, and solving for the Ce4+ in the melt:

3 4

4 / 4 /

Ce CeCe ( Ce Ce ) Cezircon melt zircon melt

zircon melt melt meltD D

3 4 3

/ 4 / /

Ce Ce CeCe Ce Ce ( )zircon melt zircon melt zircon melt

zircon melt meltD D D

3

4 3

/4 Ce

/ /

Ce Ce

Ce CeCe

zircon meltzircon melt

melt zircon melt zircon melt

D

D D

(1.7)

Substitution of Equation 1.7 in to 1.5 yields,

3

4 3

/3 Ce

/ /

Ce Ce

Ce CeCe Ce


melt melt zircon melt zircon melt

D

D D

Which can be solved for the Ce3+ content of the melt,

3

4 3

/3 Ce

/ /

Ce Ce

Ce CeCe Ce



D

D D

(1.8)

Dividing Equation 1.8 by Equation 1.7 then gives us the Ce4+/Ce3+ of the melt

3

4 3

3

4 3

4 3

/

Ce/ /3

Ce Ce/4

Ce/ /

Ce Ce

/ /

Ce Ce

Ce CeCe

Ce

Ce CeCe

Ce (



meltzircon melt

zircon meltmeltzircon melt zircon melt

zircon melt zircon melmelt

D

D D

D

D D

D D

3/

Ce

)1

Ce Ce

t

zircon meltzircon melt D

4

3

/3Ce

4 /

Ce

Ce CeCe

Ce Ce Ce

zircon meltmelt zirconmelt

zircon meltmelt zircon melt

D

D

12

Which can be simplified to:

3

4

/4Ce

3 /

Ce

Ce CeCe

Ce Ce Ce

zircon meltzircon meltmelt

zircon meltmelt melt zircon

D

D

(1.9)

This is now in a form in which the Ce4+/ Ce3+ in the melt can be calculated using lattice strain

constraints to estimate values of 3

/

Ce

zircon meltD and 4

/

Ce

zircon meltD . Oxygen fugacity can then be estimated

with knowledge of the variation in the Ce4+/ Ce3+ of the melt with fO2, calibrated from laboratory

experiments.

Although uranium can exist in 4+, 5+ and 6+ oxidation states in magmatic systems (Beattie

1993a,b; Latourrette and Burnett, 1992), and is not appreciably concentrated by any major rock-

forming minerals, the partitioning of pentavalent and hexavalent cations into zircon is not well

constrained. Therefore, estimation of magma redox from U partitioning, although potentially

useful, is not currently possible.

1.5 The Early Earth

The Hadean eon (>4.0 Ga) marks a period in the Earth’s history from which little pristine solid

material has survived. However, it is clearly important to understand this part of Earth’s history,

as it provides boundary conditions for models of subsequent Earth evolution. Despite this dearth

of material, an internally consistent picture of this eon has been emerging over the past two

decades, owed almost entirely to the preservation of detrital zircon grains. These samples,

preserved in the ca. 3.0 Ga conglomerates from the Jack Hills in the Narryer Gneiss Complex,

Western Australia, yield ages ranging from 3.0 to as old as 4.404 ± 0.008 Ga (Wilde et al.,

2001). Such material thus provides a record of events occurring ~150 Ma after terrestrial

accretion.

Since the initial discovery of ancient zircons in the Jack Hills by Froude et al. (1983), a

significant research effort has been focused on their study. Measurements of oxygen isotopic

composition have shown that a majority of pre 4.0 Ga grains are isotopically heavier than the

mantle range of O18 ( 18 16 18 16(O / O ) / (O / O )sample standard *1000) values, implying interaction

with a hydrosphere (Mojzsis et al., 2001). As the Lu/Hf ratio of zircon is typically less than

0.01, the 176Hf/177Hf ratio remains unaffected by ingrowth from 176Lu decay, and reflects that of

13

melt at the time of crystallization. Measurements of the hafnium isotopic composition of Jack

Hills zircons have shown they crystalized from melts with source regions having Lu/Hf ratios

which are fractionated with respect to chondritic values, suggesting early separation of the

silicate Earth into mantle and crustal reservoirs (Amelin et al., 1999; Harrison et al., 2005). As

the inclusion population of the Hadean-aged grains is dominated by quartz and muscovite

(Cavosie et al., 2004) and Ti-thermometry indicates crystallization temperatures of ~700C

(Watson and Harrison, 2005), such zircons were interpreted to have formed from anatectic melts

of pelagic sediment. Other evidence suggests that perhaps not all grains have formed in this

manner, however. For example, if aTiO2 is assumed to be <1, then higher zircon crystallization

temperatures are indicated (Coogan and Hinton, 2006). Also, evidence suggests that part of the

“granitic” inclusion suite could be younger than the zircon host, and are the product of

replacement (Rasmussen et al., 2011). The following chapters will demonstrate that, with

constraints on the variation in Ce4+/Ce3+ with fO2 in silicate melts, the redox state of the magmas

which formed these ancient zircons can be determined, providing further constraints on the redox

evolution of the Earth.

1.6 Contributions of this Thesis

The primary goal of this thesis is to calibrate a method for the measurement of magma redox

state using the magnitude of cerium anomalies in zircon. Owing to the disequilibrium effects

which appear to govern laboratory experiments involving zircon growth, direct determination of

zircon-melt partitioning of Ce was not attempted. Instead, the method of Ballard et al., (2002)

was adopted, where measurement of partitioning between zircon and host rock is used to

estimate 4

/

Ce

zircon meltD and 3

/

Ce

zircon meltD from lattice strain constraints, from which 4 3Ce Ce/melt meltx x can be

estimated. As there was little preexisting information on the redox behavior of Ce in magmatic

systems, this required an investigation of the variation in 4 3Ce Ce/melt meltx x as a function of fO2, T, and

melt composition, from which estimates of the fO2 of zircon-forming melts could be made. This

thesis is divided into three parts: the first describing a calibration of Ce M4,5-edge x-ray

absorption near-edge structure (XANES) spectroscopy for quantitative measurement of

Ce4+/Ce3+ in silicate glass; the second is an experimental investigation of cerium redox behavior

in magmatic systems; the third outlines a method for applying the experimental calibration of

4 3Ce Ce/melt meltx x to determine the redox state of zircon-forming systems whose fO2 is known

14

independently, then estimating the fO2 of Hadean magmas, as represented by the zircons from

Jack Hills.

1.6.1 Calibration of Cerium M4,5-edge XANES for Ce4+/Ce3+ Determinations in Silicate Melts

Investigation of the redox behavior of an element first requires an appropriate method to measure

the abundance of each oxidation state. For the case of cerium in silicate melts, we are

significantly limited in the number of suitable techniques. This is primarily due to the low

Ce4+/Ce3+ ratios in silicate glasses, which result in the Ce4+ signal being obscured by the much

greater response from Ce3+ in most spectroscopic techniques. Another complicating factor is that

the Ce4+/Ce3+ ratio can be potentially modified by beam sample interactions involving electron

and high energy x-ray-based spectroscopic techniques. Cerium M4,5-edge XANES, however, is

highly sensitive to the presence of Ce4+, and the low energy of the M-edge (~900 eV) reduces the

potential for Ce4+/Ce3+ modification during analysis.

In Chapter 2 of this thesis a calibration of Ce M4,5-edge XANES for quantitative measurements

of Ce redox state in silicate glass is presented. Based on comparison to potentiometric

determinations of Ce4+/Ce3+, Ce4+ was found to have a greater fluorescence yield (FLY) response

to the incident radiation in the M4,5-edge XANES region than Ce3+. A correction to M-edge

determined Ce4+/Ce3+ ratios was provided, and takes the form:

4 4 ( 0.004) 4XANESCe / Ce 578.7[Ce / Ce 5%] 0 Ce / Ce 0.4A (1.10)




fraction of Ce4+ in the sample. This calibration allows for rapid and accurate determination of Ce

redox state even in samples with low Ce4+/Ce3+ values.

1.6.2 Redox Behavior of Ce in Silicate Melts

Silicate glasses have been widely used as both chemical and structural analogs of the melt phase

for several decades (Henderson, 2005). Although in situ measurement of silicate melts at

temperature is desirable, these analyses are difficult. Instead, since silicate glasses are solid at

15

room temperature, this allows for information on the melt structure to be more easily extracted.

In Chapter 3 of this thesis, the Ce4+/Ce3+ ratios of silicate glasses, quenched from melts

equilibrated over varying fO2, T, and composition, are determined using the calibration presented

in Chapter 2. With this information, a model is presented for the redox behavior of Ce in natural

silicate melts, where the Ce4+/Ce3+ ratio can be expressed through the equation:

22

3/2

2

CeO 1 5705( 257) NBOlog log O 0.8694( 0.005)

CeO 4 T

3.856( 0.049) H O 3.889( 0.102)

fT

x

(1.11)

Where T is in Kelvin, NBO/T is the proportion of non-bridging oxygen to tetrahedrally

coordinated cations, and xH2O is the mole fraction of water dissolved in the melt. These results

show that at terrestrial oxygen fugacities, Ce will predominantly be in a 3+ oxidation state,

although Ce4+ will be present in trace amounts even under relatively reducing conditions.

1.6.3 Derivation of a Ce-in-Zircon Oxygen Barometer

The oxygen fugacity of igneous systems will have a significant influence on the saturation and

composition of mineral phases, speciation in the gas phase, physical properties of magmas, such

as viscosity, as well as the generation of a variety of ore deposits. Constraints on fO2 are

therefore necessary to fully understand the evolution and emplacement of magmatic systems as

well as terrestrial evolution as a whole.

Using the empirical relation for cerium redox equilibrium presented in Chapter 3, combined with

lattice strain constraints on 4

/

Ce

zircon meltD and 3

/

Ce

zircon meltD , Chapter 4 offers a calibration for the

measurement of magma redox state from Ce partitioning between zircon and melt. Oxygen

fugacities calculated from the method (expressed relative to the fayalite-magnetite-quartz, FMQ,

oxygen buffer) for samples from the Bishop tuff, California (+1.2 ± 0.7), Toba tuff, Indonesia

(+1.0 ± 1.1), and the Nane plutonic suite, Labrador (-2.4 ± 0.6), are in excellent agreement with

independent estimates for these suites. Application of this technique to >4.0 Ga zircons from the

Jack Hills, Australia, give estimates of fO2 between -1.0 and +2.5 FMQ, during zircon

crystallization, assuming saturation at conditions typical of modern systems and a parent melt

composition similar to Archean tonalite-trondhjemite-granodiorite (the oldest preserved felsic

igneous rocks). Whereas initial core-mantle equilibrium requires oxygen fugacities of FMQ -4 or

16

so, these results suggest that the oxidation fugacity of Earth’s mantle increased approximately

five orders of magnitude within the first 150 Ma of Earth history.

1.6.4 Author Contributions

Both D .J. Smythe and J. M. Brenan contributed to the design of this study. All experiments

described in Chapters 2 and 3 were carried out by D. J. Smythe. Sample analysis described in

Chapters 2, 3, and 4 were done by D. J. Smythe with the following exceptions: x-ray absorption

spectra for experimental run products catalogued beginning with DS09-C2, which were collected

by N. R. Bennett, and DS11, collected by T. Regier; electron probe micro-analysis of run product

glasses for Mn, Mo, Cr and V, done by J. M. Brenan; and whole rock analysis of sample EC87-

86 from the Umiakovik pluton, done by Activation Laboratories Ltd. in Ancaster, ON. All

calculations and modeling were carried out by D. J. Smythe. D. J. Smythe and J. M. Brenan

contributed to the preparation of the written component of the thesis. Chapter 2 of this thesis has

been published in the Journal of Non-Crystaline Solids (Smythe et al. 2013) and Chapters 3 and

4 have been written for submission to the journals Geochimica et Cosmochimica Acta and

Geology respectively.

1.7 References

Amelin, Y., Lee, D.-C., Halliday, A.N., Pidgeon, R.T., 1999. Nature of the Earth's earliest crust

from hafnium isotopes in single detrital zircons. Nature 399, 252-255.

Bea, F., Pereira, M.D., Stroh, A., 1994. Mineral/leucosome trace-element partitioning in a

peraluminous migmatite (a laser ablation-ICP-MS study). Chemical Geology 117, 291-312.

Beattie, P., 1993a. The generation of uranium series diseqilibria by partial melting of spinel

peridotite: Constraints from partitioning studies. Earth and Planetary Science Letters, 117, 379-

391.

Beattie, P., 1993b. Uranium thorium disequilibrium and partitioning on melting of garnet

peridotite. Nature, 363, 63-65.

Blundy, J., Wood, B., 2003. Mineral-melt partitioning of uranium, thorium and their daughters,

Reviews in Mineralogy and Geochemistry 52, 59-123.

17

Blundy, J.D., Wood, B.J., 1994. Prediction of crystal-melt partition coefficients from elastic

moduli. Nature 372, 452-454.

Brice, J.C., 1975. Some thermodynamic aspects of the growth of strained crystals. Journal of

Crystal Growth 28, 249-253.

Burnham, A.D., Berry, A.J., 2012. An experimental study of trace element partitioning between

zircon and melt as a function of oxygen fugacity. Geochimica et Cosmochimica Acta 95, 196-

212.

Caruba, R., Baumer, A., Ganteaume, M., Iacconi, P., 1985. An experimental study of hydroxyl

groups and water in synthetic and natural zircons: a model of the metamict state. American

Mineralogist 70, 1224-1231.

Cavosie, A.J., Wilde, S.A., Liu, D., Weiblen, P.W., Valley, J.W., 2004. Internal zoning and U–

Th–Pb chemistry of Jack Hills detrital zircons: a mineral record of early Archean to

Mesoproterozoic (4348–1576 Ma) magmatism. Precambrian Research 135, 251-279.

Cherniak, D.J., 2010. Diffusion in accessory minerals: zircon, titanite, monazite and xenotime.


Cherniak, D.J., Watson, E.B., 2003. Diffusion in Zircon. Reviews in Mineralogy and

Geochemistry 53, 113-143.

Claiborne, L.L., Miller, C.F., Flanagan, D.M., Clynne, M.A., Wooden, J.L., 2010. Zircon reveals

protracted magma storage and recycling beneath Mount St. Helens. Geology 38, 1011-1014.

Coogan, L.A., Hinton, R.W., 2006. Do the trace element compositions of detrital zircons require

Hadean continental crust? Geology 34, 633.

Dickinson, J.E., Hess, P.C., Rutherford, M.J., 1980. REE partitioning between zircon,

whitlockite and two liquids. EOS Transactions of the American Geophysical Union 61, 397.

Drake, M.J., 1975. The oxidation state of europium as an indicator of oxygen fugacity.

Geochimica et Cosmochimica Acta 39, 55-64.

18

Ferry, J.M., Watson, E.B., 2007. New thermodynamic models and revised calibrations for the Ti-

in-zircon and Zr-in-rutile thermometers. Contributions to Mineralogy and Petrology 154, 429-

437.

Finch, R. J., Hanchar, J. M., Hoskin, P. W. O., Burns, P. C., 2001. Rare-earth elements in

synthetic zircon: Part 2. A single-crystal X-ray study of xenotime substitution. American


Finch, R.J., Hanchar, J.M., 2003. Structure and Chemistry of Zircon and Zircon-Group Minerals.


Froude, D.O., Ireland, T.R., Kinny, P.D., Williams, I.S., Compston, W., Williams, I.R., Myers,

J.S., 1983. Ion microprobe identification of 4,100-4,200 Myr-old terrestrial zircons. Nature 304,

616-618.

Fujimaki, H., 1986. Partition coefficients of Hf, Zr, and REE between zircon, apatite, and liquid.

Contributions to Mineralogy and Petrology 94, 42-45.

Hanchar, J.M., van Westrenen, W., 2007. Rare earth element behavior in zircon-melt systems.

Elements 3, 37-42.

Hanchar, J.M., Finch, R.J., Hoskin, P.W.O., Watson, E.B., Cherniak, D.J., Mariano, A.N., 2001.

Rare earth elements in synthetic zircon: Part 1. Synthesis, and rare earth element and phosphorus

doping. American Mineralogist 86, 336-680.

Harrison, T.M., Blichert-Toft, J., Muller, W., Albarede, F., Holden, P., Mojzsis, S.J., 2005.

Heterogeneous Hadean hafnium: evidence of continental crust at 4.4 to 4.5 ga. Science 310,

1947-1950.

Hazen, R.M., Finger, L.W., 1979. Crystal structurea and compressibility of zircon at high

pressure. American Mineralogist 64, 196-201.

Heaman, L. M., Bowins, R., Crocket, J., 1990. The chemical composition of igneous zircon

suites: implications for geochemical tracer studies. Geochimica et Cosmochimica Acta 54, 1597-

1607.

19

Henderson, G.S., 2005. The structure of silicate melts: a glass perspective. The Canadian


Hinton, R.W., Upton, B.G.J., 1991. The chemistry of zircon: Variations within and between large

crystals from syenite and alkali basalt xenoliths. Geochimica et Cosmochimica Acta 55, 3287-

3302.

Hoskin, P.W.O., Rodgers, K.A., 1996. Ramen spectral shift in the isomorphous series (Zr1-

xHfx)SiO4. European Journal of Solid State Inorganic Chemistry 33, 1111-11121.

Hoskin, P.W.O., Schaltegger, U., 2003. The composition of zircon and igneous and metamorphic

petrogenesis. Reviews in Mineralogy and Geochemistry 53, 27-62.

Jain, J.C., Neal, C.R., Hanchar, J.M., 2001. Problems Associated with the Determination of Rare

Earth Elements of a “Gem” Quality Zircon by Inductively Coupled Plasma-Mass Spectrometry.

Geostandards Newsletter 25, 229-237.

Jensen, B.B., 1973. Patterns of trace element partitioning. Geochimica et Cosmochimica Acta 37,

2227-2242.

Latourrette, T.Z., Burnett, D.S., 1992. Experimental determination of U-partitioning and Th-

partitioning between clinopyroxene and natural and synthetic basaltic liquid. Earth and Planetary

Science Letters, 110, 227-244.

Linnen, R.L., Keppler, H., 2002. Melt composition control of Zr/Hf fractionation in magmatic

processes. Geochimica et Cosmochimica Acta 66, 3293-3301.

Luo, Y., Ayers, J.C., 2009. Experimental measurements of zircon/melt trace-element partition

coefficients. Geochimica et Cosmochimica Acta 73, 3656-3679.

Mahood, G., Hildreth, W., 1983. Large partition coefficients for trace elements in high-silica

rhyolites. Geochimica et Cosmochimica Acta 47, 11-30.

Marshall, A.S., Macdonald, R., Rogers, N.W., Fitton, J.G., Tindle, A.G., Nejbert, K., Hinton,

R.W., 2009. Fractionation of Peralkaline Silicic Magmas: the Greater Olkaria Volcanic

Complex, Kenya Rift Valley. Journal of Petrology 50, 323-359.

20

McDonough, W.F., and Sun, S.-s., 1995. The composition of the Earth. Chemical Geology 120,

223-253.

Mojzsis, S.J., Harrison, T.M., Pidgeon, R.T., 2001. Oxygen-isotope evidence from ancient

zircons for liquid water at the Earth's surface 4,300 Myr ago. Nature 409, 178-181.

Mungall, J.E., 2002. Empirical models relating viscosity and tracer diffusion in magmatic silicate

melts. Geochimica et Cosmochimica Acta 66, 125-143.

Murali, A.V., Parthasarathy, R., Mahadevan, T.M., Das, M.S., 1983. Trace element

characteristics, REE patterns and partition coefficients of zircons from different geological

environments-A case study on Indian zircons. Geochimica et Cosmochimica Acta 47, 2047-

2052.

Nagasawa, H., 1970. Rare earth concentrations in zircons and apatites and their host dacites and

granites. Earth and Planetary Science Letters 9.

Nardi, L.S.V., Formoso, M.L.L., Muller, I.F., Fontana, E., Jarvis, K., Lamarao, C., 2013.

Zircon/rock partition coefficients of REEs, Y, Th, U, Nb, and Ta in granitic rocks: Uses for

provenance and mineral exploration purposes. Chemical Geology 335.

Nasdala, L., Zhang, M., Kempe, U., Panczer, G., Gaft, M., Andrut, M., Plötze, M., 2003.

Spectroscopic methods applied to zircon. Reviews in Mineralogy and Geochemistry 53, 427-467.

Okano, O., Watson, E.B., Tatsumoto, M., 1987. Partition coefficents for REE and Hf between

zircon and liquid: Inferences for lunar granite petrogenesis. Lunar and Planetary Science

Absrtacts XVIII, 740-741.

Onuma, N., Higuchi, H., Wakita, H., Nagasawa, H., 1968. Trace element partitioning between

two pyroxenes and the host lava. Earth and Planetary Science Letters 5, 47-51.

Philpotts, J.A., 1970. Redox estimation from a calculation of Eu2+ and Eu3+ concentrations in

natural phases. Earth and Planetary Sceince Letters 9, 257-268.

21

Rasmussen, B., Fletcher, I.R., Muhling, J.R., Gregory, C.J., Wilde, S.A., 2011. Metamorphic

replacement of mineral inclusions in detrital zircon from Jack Hills, Australia: Implications for

the Hadean Earth. Geology 39, 1143-1146.

Rubatto, D., 2002. Zircon trace element geochemistry: partitioning with garnet and the link

between U-Pb ages and metamorphism. Chemical Geology 184, 123-138.

Rubatto, D., Hermann, J., 2007. Experimental zircon/melt and zircon/garnet trace element

partitioning and implications for the geochronology of crustal rocks. Chemical Geology 241, 38-

61.

Sano, Y., Terada, K., and Fukuoka, T., 2002, High mass resolution ion microprobe analysis of

rare earth elements in silicate glass, apatite and zircon: lack of matrix dependency: Chemical

Geology, v. 184, p. 217-230.

Schreiber, H.S., 1982. The chemistry of uranium in glass-forming melts: redox interactions of

U(VI)-U(V)-U(IV) with cerium in aluminosilicates. Journal of Non-Crystaline Solids 49, 189-

200.

Schreiber, H.D., Lauer, H.V., Jr., Thanyasiri, T., 1980. The redox state of cerium in basaltic

magmas: an experimental study of iron-cerium interactions in silicate melts. Geochimica et

Cosmochimica Acta 44, 1599-1612.

Speer, J.A., 1980. Zircon. Reviews in Mineralogy and Geochemistry 5, 67-112.

Steiger, R.H., Jäger, E., 1977. Subcommission on geochronology: Convention on the use of

decay constants in geo- and cosmochronology. Earth and Planetary Science Letters 36, 359-362.

Strutt, R.J., 1909. The accumulation of helium in geologic time III. Proceedings of the Royal

Society of London, A 83.

Thomas, J.B., Bodnar, R.J., Shimizu, N., Sinha, A.K., 2002. Determination of zircon/melt trace

element partition coefficients from SIMS analtsis of melt inclusions in zircon. Geochimica et


22

Trail, D., Bruce Watson, E., Tailby, N.D., 2012. Ce and Eu anomalies in zircon as proxies for the

oxidation state of magmas. Geochimica et Cosmochimica Acta 97, 70-87.

Trail, D., Thomas, J.B., Watson, E.B., 2011a. The incorporation of hydroxyl into zircon.

American Mineralogist 96, 60-67.

Trail, D., Watson, E.B., Tailby, N.D., 2011b. The oxidation state of Hadean magmas and

implications for early Earth’s atmosphere. Nature 480, 79-82.

Valley, J.W., 2003. Oxygen isotopes in zircon. Reviews in Mineralogy and Geochemistry 53,

343-385.

Watson, E.B., Wark, D.A., Thomas, J.B., 2006. Crystallization thermometers for zircon and

rutile. Contributions to Mineralogy and Petrology 151, 413-433.

Watson, E.B., 1980. Some experimentally determined zircon/liquid partition coefficients for the

rare earth elements. Geochimica et Cosmochimica Acta 44, 895-987.

Watson, E.B., 1996. Surface enrichment and trace-element uptake during crystal growth.


Watson, E.B., Harrison, T.M., 2005. Zircon thermometer reveals minimum melting conditions

on earliest Earth. Science 308, 841-844.

Watson, E.B., Müller, T., 2009. Non-equilibrium isotopic and elemental fractionation during

diffusion-controlled crystal growth under static and dynamic conditions. Chemical Geology 267,

111-124.

Wilde, S.A., Valley, J.W., Peck, W.H., and Graham, C., 2001. Evidence from detrital zircons for

the existence of continental crust and oceans on the Earth 4.4 Gyr ago. Nature 409, no. 6817, p.

175-178.

23

Figure 1.1 Schematic representation of the lattice strain model showing the variation in partition

coefficient Di, for a series of isovalent cations, as a function of cation radius. D0 is the ‘strain

compensated partition coefficient’, E is the Young’s Modulus, and r0 is the ideal ionic radius for

the structural site. Adapted from Blundy and Wood (2003). The arrows indicate the tightness of

the parabola will be a function of E. Higher values of E will result in a more rapid decrease in

partitioning with deviation from r0.

24

Figure 1.2 Representative partition coefficients for REE between zircon and melt. a) Results

from natural samples, and b) results from laboratory experiments. Results from experimental

studies have a consistently lower / /HREE LREE/zircon melt zircon meltD D than those from studies on natural

samples. The REEs are plotted in order of increasing atomic number. Anomalous values of D for

Ce and Eu can be seen in the results from natural samples. Adapted from Hanchar and van

Westrenen (2007).

25

.

Figure 1.3 Onuma diagrams, plotting the variation in log D with ionic radius, for data from a)

Nagasawa (1970), b) Sano et al. (2002), c) Luo and Ayers (2009), and d) Burnham and Berry

(2012). Contamination from LREE-bearing phases can been seen in the results of Nagasawa

(1970) and appear absent in those of Sano et al. (2002). The experimental results in c) and d) can

be modeled using lattice stain, however, the resulting parabolas are significantly wider than those

from natural samples. Adapted from Hanchar and van Westrenen (2007).

26

Chapter 2

Quantitative Determination of Cerium Oxidation States in Alkali-

Aluminosilicate Glasses using M4,5-edge XANES

A calibration method for quantitative determination of Ce4+/Ce3+ in alkali-aluminosilicate glasses

using Ce M4,5-edge X-ray absorption near edge structure (XANES) spectroscopy is presented.

Samples of varying alkali oxide content synthesized over a range of temperatures and oxygen

fugacities (fO2) were analyzed using both synchrotron-based XANES and traditional wet

chemical techniques. Based on comparison to potentiometric determinations of Ce4+/Ce3+ in

digested portions of our samples, Ce4+ was found to have a greater fluorescence yield (FLY)

response to the incident radiation in the M4,5-edge XANES region than Ce3+. This results in an

overestimation of the Ce4+/Ce3+ ratio when directly comparing the contributions of the two

oxidation states to the total M-edge area, as has been used for crystalline materials. We offer a

correction to M-edge determined Ce4+/Ce3+ ratios as follows:

4 4 ( 0.004) 4XANESCe / Ce 578.7[Ce / Ce 5%] 0 Ce / Ce 0.4A




fraction of Ce4+ in the sample. The FLY response of Ce4+ was found to systematically increase

with the degree of polymerization of the glass, suggesting that in addition to the local electronic

environment, the FLY response of Ce is also dependent on longer range interactions.

2.1 Introduction

Silicate glasses containing minor concentrations of cerium have been of interest for the past

several decades, initially due to the role of Ce as a coloring agent in the presence of Ti (Paul,

1976) and to prevent the formation of optic centers under radiation exposure (Stroud, 1962) and

electron bombardment (Kilbourn, 2011). Subsequently, the luminescent properties of Ce

resulted in its extensive use in scintillating glasses (Chiodini et al., 2002; Zanella et al., 1994).

Recently, there has been a renewed interest in cerium as it has potential use in white light

27

emitting diodes, due to the rapid decay rate of Ce relative to other phosphors (Andrade et al.,

2009), as well as uses in photo-thermo-refractive glasses (Efimov et al., 2011), and substitution

in bioactive glasses (Salinas et al., 2011). In silicate glasses, as well as other oxide glasses, Ce

ions can exist in either 3+ or 4+ oxidation states. The relative proportions of Ce3+ and Ce4+ will

depend on the bulk solution properties of the glass, in addition to the preparation conditions

(Paul and Douglas, 1965). Although it is known that the properties of Ce are strictly dependent

on its oxidation state in silicate glass, determination of Ce4+/Ce3+ in glasses is limited to only a

small number of studies. Furthermore, in natural silicate melts, there is extensive evidence for

the presence of both Ce3+ and Ce4+ (Schreiber, 1980; Takahashi et al., 2002), however, the redox

behaviour of Ce in such materials is poorly understood. This dearth of information is in part due

to a lack of in-situ analytical techniques appropriate for determination of Ce oxidation state in

these materials. The purpose of this communication is to provide a calibration method for the

determination of Ce4+/Ce3+ in alkali-aluminosilicate glasses using synchrotron based M4,5-edge

X-ray Absorption Near Edge Structure (XANES) spectroscopy, a relatively rapid and non-

destructive technique which does not modify the Ce4+/Ce3+ ratio under beam exposure.

Synchrotron based determinations of Ce4+/Ce3+ in materials have traditionally used L3-edge

XANES which involves a transition of a 2p electron to an unoccupied 5d state, and is located

between 5710 and 5750 eV. The ground state of Ce3+ and Ce4+ are [Xe] 4f1 and [Xe] 4f0,

respectively. Differences in the coulomb interaction between the 2p core hole and the valence

band of these ground states result in a shift in the L3-edge from ~5726 eV for Ce3+, to ~5740 eV

for Ce4+. Additionally, there is the appearance of a second peak at ~5730 eV in the Ce4+ spectra

resulting from a 4f1L ground state ionic configuration, where L is a ligand O 2p hole (Bianconi et

al., 1987). These differences can then be used to quantify the proportion of Ce3+ and Ce4+ in a

material (Darab et al., 1998; Shahin et al., 2005; Takahashi et al., 2002).

Ce M4,5-edge XANES occurs between 875 and 910 eV and involves a 3d 4f transition,

directly probing the occupancy of the 4f orbitals. This transition is located approximately 5 eV

below the true edge arising from 3d electron excitations to the Fermi level. As the spin-orbit

interaction of the 3d9 hole is much larger than the 3d94fn+1 exchange interaction, the x-ray

absorption spectra of Ce in the M-edge region is characterized by two distinct line groups

corresponding to 3d3/24f5/2 (M4) and 3d5/24f7/2 (M5) which are separated by approximately 17 eV.

Interaction between the core hole and the 4f electrons in the excited state results in final states

28

with different symmetry and energy leading to complicated multiplet features within the M4 and

M5 line groups (Bonnelle et al., 1974). In addition to changes in the spectral features, the

position of the Ce M4,5-edge has been shown to systematically shift to higher energy by

approximately 2 eV with changes in oxidation state from 3+ to 4+ (Bonnelle et al., 1974; Jo and

Kotani, 1988; Kaindl et al., 1984).

Previous determinations of Ce oxidation state using the Ce M4,5-edge have focused exclusively

on crystalline materials whereas in this study we investigate the behaviour in the amorphous

state. Furthermore, this is the first attempt at calibrating the Ce M-edge for quantitative

measurements of Ce oxidation state using an independent measure of the Ce4+/Ce3+ ratio.

Synchrotron based techniques appear best suited for these determinations as electron energy loss

spectroscopy (EELS) which uses inelastic electron scattering, a complimentary process to X-ray

absorption, has been shown to modify the Ce4+/Ce3+ ratio (Garvie and Buseck, 1999). The

relatively low energy of the Ce M4,5-edge, relative to the L3-edge, further minimizes the potential

for modification of the Ce4+/Ce3+ ratio under exposure to the X-ray beam. As we will show, the

M4,5-absorption edge also has the benefit of being highly sensitive to Ce4+, particularly at low

concentrations.

2.2 Material and Methods

2.2.1 Preparation of Glasses

Experiment starting materials consisted of glasses, prepared from weighed proportions of

crystalline reagent grade oxides and carbonates (>99.99% purity), including SiO2, TiO2, Al2O3,

MgO, CaCO3, Na2CO3, K2CO3, Ca10(OH)2(PO4)6, and CeO2, which were ground under ethanol

with an agate mortar and pestle and calcined for approximately 12 hours at 1000 C . To ensure

homogeneity within a given composition, sample powders were then fused twice at 1450 C in a

platinum crucible for 0.5 h, quenched in cold water then ground twice to dryness under ethanol.

The chemical compositions of the glass starting materials are listed in Table 2.1.

To achieve variable values of Ce4+/Ce3+, samples were synthesized under controlled atmosphere

and temperature in a vertical tube furnace. Approximately 90 mg of powdered starting material

was mixed with polyvinyl alcohol solution and applied to a loop of either Pt or Pt10%Rh wire,

which was suspended in the furnace by a fused quartz rod. The furnace atmosphere was

29

controlled using either pure O2 or a CO - CO2 mixture. Temperature was monitored using a Pt-

Pt10%Rh thermocouple and fO2 was monitored using a SIRO2 C700+ solid zirconia electrolyte

oxygen sensor purchased from Ceramic Oxide Fabricators©. Samples were allowed to

equilibrate for ~12 h in the furnace atmosphere before being quenched in cold water. Details on

experimental run conditions are provided in Table 2.3. To ensure the run products had quenched

to a single homogeneous phase, portions of the glass were crushed and mounted on a glass slide

along with index oil and visually inspected under cross polarized light. If any crystalline phase

was observed samples were rejected.

2.2.2 Analytical

2.2.2.1 Electron Probe Micro-Analysis

Portions of the experimental run products were mounted in epoxy and polished with diamond

grit down to 1 m followed by 0.3 m alumina. Compositional analysis was carried out with a

Cameca SX50 electron microprobe housed in the Department of Earth Sciences at the University

of Toronto. Analyses were done at 15 kV using a 20 m defocused beam and 5 nA beam

current. Alkali elements were analyzed at the beginning of the acquisition, so as to minimize

their migration under the electron beam during analysis. Standards employed were natural

basaltic glass (Mg, Ca), obsidian (Si, Al), albite glass (Na), sanidine (K), TiO2 (Ti), and CePO4

(P, Ce). Raw count rates were converted to element concentrations using the ZAF correction

routine.

The concentrations of the redox-sensitive trace elements, Mn, Mo, V, and Cr, were also

measured in run-product glasses. Standards used were bustamite (Mn), MoO3 (Mo), Cr2O3 (Cr)

and V2O5 (V). Wavelength scans were performed on the run-product glasses over the spectral

region corresponding to the x-ray peak used for quantification (Mn K , Mo L , Cr K and V

K ) to assess background positions and possible interferences. A Ti K peak was observed

near the V K peak, but was well-resolved, and very weak, due to only small amounts (<1

wt%) of Ti in sample glasses. Beam conditions employed were 100 nA, 20 kV and a 120 second

on-peak count time. Trace elements were measured in a separate analytical session, but used the

major element glass composition determined previously to convert raw intensities to

concentrations using the ZAF method. For the conditions employed, the calculated detection

30

limits are Mn (35 ppm), Mo (95 ppm), Cr (50 ppm), and V (75 ppm). In all run-product glasses,

concentrations of these elements are below detection.

2.2.2.2 Potentiometric Titrations

Potentiometric titrations to determine the Ce4+/Ce3+ in glasses were done using a modified

version of the method described by Paul and Douglas (1965). Run product glasses were crushed

in an agate mortar under air, then approximately 20 mg of sample was placed in a Savillex

container and digested in 1.0 mL of a 12% HF – 7% H2SO4 mixture for 4 h under constant

stirring. All samples were visually inspected under a microscope to determine if the dissolution

had gone to completion. After digestion, 250.0 mg of boric acid was added to each sample to

complex with the fluorine ions, stirred for an additional 30 min, followed by the addition of 35

mL of deionized water. The entire digestion procedure was done with samples contained in an

ice-water bath.

Voltage measurements of the solution were made using a Pt pin indicator electrode and an epoxy

body Ag-AgCl reference electrode connected to an Orion model 525a+ pH meter. Calibration of

the pH meter was carried out using 220 mV and 427 mV redox buffer solutions. Samples were

titrated with 0.0001 N ammonium Fe (II) sulfate solution.

Calibration of this technique was carried out using three different Ce4+ compounds: Ce(SO4)2,

(NH4)4Ce(SO4)4 2H2O, and (NH4)3Ce(NO3)6. These materials were mixed with approximately

20 mg of Ce-free blank glass. The calibration curve was then constructed by plotting the mass of

Ce4+ added versus the volume of titrant added at the equivalence point (Figure 2.1). Standard

titrations were run before and after the samples to ensure oxidation of the ammonium Fe (II)

sulphate solution had not occurred over the course of the analyses.

2.2.2.3 M-edge XANES

Synchrotron-based Ce M4,5-edge XANES spectra were collected on the Spherical Grating

Monochromator (SGM) undulator beamline at the Canadian Light Source, Saskatoon,

Saskatchewan. Coarse glass chips were mounted on stainless steel disks using carbon tape and

then placed in the Solid State Absorption Spectroscopy Chamber endstation under high vacuum

for analysis. Samples were analyzed at room temperature. The incident beam was 1000 m x

100 m and the sample faced approximately 45 toward the beam. Spectra were collected in the

31

region of 870 to 920 eV with a step size of 0.1 eV over the edge region. X-ray absorption

spectra were collected in both fluorescence yield (FLY) and total electron yield (TEY) modes.

The FLY was measured using a microchannel plate detector and the TEY was determined by the

sample current. Due to the insulating nature of the glasses, the TEY spectra was found to have a

unsatisfactorily low signal to noise ratio, hence FLY spectra have been used for determination of

the Ce oxidation state in glass samples. Between two and three spectra were collected on all

samples moving the sample position with respect to the beam after each scan. X-ray energies

were calibrated using crystalline CeO2 as a reference.

2.3 Results

2.3.1 Potentiometric Titrations

The equivalence points for all titrations were located at approximately 750 mV, consistent with

the maximum and minimum potentials around 1050 mV and 450 mV. Small local minima were

consistently observed after the equivalence point in E V versus V plots of the titrations.

Graphs of this type should be symmetric about the equivalence point, this asymmetry therefore

suggesting the presence of trace amounts of another redox sensitive element. The potential for

3 2Mn Mne is close to that of 4 3Ce Cee and Mn is a known impurity in the

starting materials. Though Mn was measured to be below 35 ppm in our samples, concentrations

on the order of 10 ppm could account for this feature. It therefore seems likely that the local

minima result from trace amounts of Mn in our samples.

In a silicate melt Ce will be present as oxide species. This allows for Ce3+ and Ce4+ to be

expressed as dissolved oxides (CeO3/2 and CeO2, respectively). Changes in the oxidation state of

Ce can therefore be expressed through the reaction

3/2 2 2

1CeO O CeO

4melt melt melt (2.1)

The corresponding equilibrium constant (K) for this reaction is

2

3/2

CeO

1/4CeO 2O

aK

a f

(2.2)

32

where 2CeOa and

3/2CeOa are the activities of CeO2 (Ce4+) and CeO3/2 (Ce3+) in the melt,

respectively, and fO2 is the imposed oxygen fugacity. Since the activity term for a component (i)

is the product of the mole fraction (xi) and the activity coefficient (i) of the component, Equation

2.2 can be rearranged by taking the logarithm of both sides to yield,

3/22

3/2 2

CeOCeO2

CeO CeO

1log log O log

4

xf K

x

(2.3)

At a fixed temperature, pressure and composition a plot of 2 3/2CeO CeOlog( / )x x versus log fO2

should then yield a straight line with a slope of 1/4. As this relationship with fO2 is observed in

the Ce4+/Ce3+ ratio determined using the potentiometric method (Figure 2.2), it is concluded this

represents the actual oxidation state of Ce in glasses investigated in this study. The error

associated with the fraction of Ce4+ ( 4Ce / Ce ) determined by this method was taken as the

amount of Ce4+ corresponding to the volume range of the width at half the minimum in the first

derivative curve for the titration, plus the uncertainty in the total amount of Ce in the analyzed

solution.

2.3.2 XANES

For cerium in a 3+ oxidation state the M4,5-edge (shown in Figure 2.3 by CeF3) consists of two

main peaks centered at ~882.5 and ~899.8 eV (features C and C’, respectively) with smaller

satellite peaks at approximately 879.6 (A), 881.5 (B), 896.6 (A’) and 898.4 eV (B’). The Ce4+

M4,5-edge (CeO2 in Figure 2.3) occurs at a slightly higher energy and is comprised of main

maxima at ~884.1 (D) and ~902.0 eV (D’) with smaller features at ~889.0 (E) and ~907.0 eV

(E’). The curve fitting analysis of the glass samples, which contain both Ce3+ and Ce4+, was

therefore carried out using five curves to model each of the M4 and M5-edge spectra, positioned

approximately at the energies mentioned above, using PeakFit® v4.12. Voigt functions V(x)

were used to account for detector broadening (Gaussian) and the natural line shape (Lorentzian).

In addition, an arctangent function T(x) was also required to fit the M4-edge in order to account

for the step in the featureless continuum. The inflection point of the arctangent function was

positioned at ~900.5 eV. Although there are uncertainties associated with the exact energy at

which the inflection point should be positioned, the chosen energy is within the overlap between

the M4-edge for Ce3+ and Ce4+ allowing for this to be kept constant over the entire range of

33

Ce4+/Ce3+ investigated. A step in the continuum is also expected to be associated with the M5-

edge, however, this was found to be small and a linear correction was sufficient. The calculated

fit to the measured XANES spectra can be expressed through Equations 2.4-2.6 (Temme, 2010).

( ) ( ) ( )nn

F x V x T x (2.4)

2

2

2

exp( )( ) ( ) ( )

n nn

nn

n

h z tV x L x t G t dt dt

x Ez t

d

(2.5)

1( ) tan2

h x ET x

B

(2.6)

Where F(x) is the calculated Ce M4,5-edge XANES spectrum, x is the x-ray energy, h is the

amplitude of the curve, E is the energy position of the curve, d is the Gaussian width, and z is

proportional to the Lorentzian/Gaussian width ratio. Curve parameters are given in Table 2.2

and examples of the curve fitting analysis of the M4 and M5-edge spectra are shown in Figure

2.4. The contributions from Ce3+ and Ce4+ to the M4,5-edge area can be determined by,

3XANES '[Ce ] [ ( ) ( ) ( ) ( ) ( ) ( )]A B C A B CV x V x V x V x V x V x dx

(2.7)

4XANES[Ce ] [ ( ) ( ) ( ) ( )]D E D EV x V x V x V x dx

(2.8)

and the relative contribution of Ce4+ to the total M4,5-edge area can be calculated by

44

XANES4 3

XANES XANESXANES

[Ce ]Ce

Ce [Ce ] [Ce ]

(2.9)

allowing for direct comparison to the results from wet chemistry.

The complex structure of the Ce M4,5-edge in samples containing both oxidation states of Ce

result in uncertainties in the curve parameters given in Table 2.2. The uncertainties, though

small, result in a relatively large range of 4XANES[Ce / Ce] values for a given spectrum. These

34

variations are greater than the standard errors calculated by the curve fitting software, which are

simply based on the residuals between the modeled and observed spectrum. We have therefore

chosen to use a conservative relative error of 5% for the 4XANES[Ce / Ce] values determined by

this technique. This covers the range of values that can be obtained through manipulation of the

curve parameters within the uncertainties (Table 2.3).

Modification of element redox states by exposure to the X-ray beam during XANES data

acquisition has been shown to occur for sulfur in silicate glasses as analyzed by S K-edge

XANES (Wilke et al., 2008). Additionally, Ce valence in pure crystalline CeO2 has also been

shown to be modified during analysis by Ce M4,5-edge EELS (Garvie and Buseck, 1999). To

investigate the possibility of modification of the redox pair in our samples during exposure to the

beam, a series of six rapid scans were carried out on one of the glass samples. Scans were taken

between 870 and 920 eV with a step size of 0.2 eV and a 0.5 second dwell time. The sample

exposure time to the beam over each scan was approximately 6 min. Before the final scan the

sample was exposed to the full beam for 3 min. by opening the exit slit to 400 m at 920 eV.

The results are shown in Figure 2.5 and although the signal to noise ratio is too low for precise

Ce4+/Ce3+ determination, all scans agree within error, suggesting beam damage has not occurred.

2.4 Discussion

2.4.1 Potentiometric Method vs. XANES

Previous valence determinations using the Ce M4,5-edge have primarily employed the modified

Anderson impurity model proposed by Gunnarsson and Schönhammer (1983) where the

intensity ratio of the peaks resulting from 4f1 and 4f0 ground states in the XANES spectrum are

corrected to yield the 4f occupancy of the initial state. This is not suited for determination of

cerium oxidation state as the 4f1 initial state will be a combination of Ce3+ 4f1 and Ce4+ 4f1L

initial states. Determinations of oxidation state using this model would require that both the

relative absorption efficiencies between Ce3+ and Ce4+ as well as the partitioning of Ce4+

between 4f0 and 4f1L initial states be known. As this is not the case with our samples, oxidation

states cannot be determined by this method.

The M5:M4 intensity ratio determined by EELS has also been used previously to quantify the Ce

valence in electron beam damaged CeO2 (Garvie and Buseck, 1999). This technique was not

35

used here due to the uncertainties involved with the continuum subtraction, which could have a

substantial effect on the calculated oxidation state proportions. Furthermore, the intensity ratio

between different Ce3+ and Ce4+ standards varied significantly, preventing reliable calibration.

The approach taken here for the XANES-based Ce4+/Ce3+ determinations compares the relative

contributions from each oxidation state to the total area of the background-corrected M4,5-edge.

Comparison of the area ratio to the Ce4+/Ce3+ determined potentiometrically reveals that the

XANES-based measurements systematically over estimate the amount of Ce4+ (Table 2.3).

There are a number of possible explanations for this difference. These include modification of

the Ce4+/Ce3+ ratio during sample digestion for potentiometric analysis, differences in the

efficiency of the fluorescence yield of the two species, preferential self-absorption of Ce3+

fluorescent photons, and higher x-ray absorption efficiency for Ce4+ relative to Ce3+.

For several reasons, modification of the Ce4+/Ce3+ ratio during sample preparation for

potentiometric analysis is an unlikely explanation for the discrepancy between the two methods.

First, between samples of the same composition, but synthesized at different fO2, the Ce4+/Ce3+

ratio follows the variation predicted from thermodynamics (e.g. Figure 2.2). Second, there are

no other multivalent cations present in these samples at the concentrations required to shift the

Ce4+/Ce3+ equilibrium through charge transfer. The only other added component which has been

found to exist in more than one oxidation state in silicate melts is Ti (Ti4+ and Ti3+), however, the

conditions at which these samples were synthesized the concentration of Ti3+ would be

negligible (Moringa et al., 1994). The concentration of likely multivalent cation impurities (Mn,

Mo, Cr, and V) were all determined to be below detection (<90 ppm). Such levels could only

result in 4Ce / Ce errors of at most 1%. Thus, for the remainder of this communication we

assume that the Ce4+/Ce3+ ratio determined with the wet chemical technique represents the actual

ratio of the sample.

If the two species have different fluorescence yields this would imply that Ce3+ generates Auger

electrons more efficiently than Ce4+ to accommodate atomic relaxation. Given that Auger decay

is dominant in the soft x-ray region (Alford et al., 2007) one would expect that variations in the

fluorescence yield efficiencies would be obscured in the TEY spectra of a sample. Though the

samples analyzed in this study are insulating and therefore generally produce poor TEY spectra,

satisfactory TEY spectra were obtained on a small number of powdered samples. Comparison

36

between the FLY and TEY spectra of these samples show similar, if not slightly higher, Ce4+

contributions to the TEY spectra. This suggests that differences in the FLY efficiency of the two

species may exist, however, it cannot explain the offset between the XANES and wet chemical

determinations of oxidation state, as this would lower the Ce4+ response in the FLY spectra.

Self-absorption of fluorescence photons from Ce3+ preferentially over those from Ce4+ would

result in an overestimation of Ce4+/Ce3+ of a sample. If this was occurring in these samples, a

high degree of asymmetry would be expected in the curves used to fit the FLY M4,5-edge spectra,

which is not the case. Furthermore, this should result in relatively higher estimations of Ce3+

from the TEY spectra, which is again contrary to observation.

Differences in the x-ray capture efficiency of the two cerium oxidation states in our samples, and

possibly in the crystalline materials investigated by previous authors (Kaindl et al., 1984; Yagci,

1985), appear to be necessary to explain the differences between the potentiometric

determinations of Ce4+/Ce3+ and the Ce M4,5 XANES spectra. Although valence state in most

cases does not have a profound effect on the absorption coefficient, a similar phenomenon has

been observed at the Ce L3-edge in cerium alkali-borosilicate glasses where Ce valence was also

determined through wet chemical methods (Darab et al., 1998). In that case the Ce4+/Ce ratios

determined by Ce L3-edge XANES were consistently lower, by as much as 0.12, than those

determined by wet chemistry. Although this is opposite to what we observe with the M4,5-edge

there is no reason to expect Ce4+ to have larger absorption coefficients over all energies.

2.4.2 Cerium Valence Determinations

As the response to the incident radiation from Ce4+ appears to be larger than that of Ce3+, we

offer a calibration for quantitative determinations of the ceric-cerous ratio in silicate glasses

using Ce M4,5-edge XANES. Figure 2.6 shows a comparison between the fractions of Ce4+

determined by the chemical method and using the area ratio of the M4,5-edge. Even at low

concentrations the contribution of Ce4+ to the M4,5-edge remains significant. As suggested in

Figure 2.6, the response of Ce to the incident radiation also shows a distinct dependence on glass

composition. This dependence has been expressed here as a function of the ratio of non-bridging

oxygens to tetrahedrally coordinated cations (NBO/T), although other melt parameters such as

optical basicity, also work well. For Ce4+/Ce values below 0.4 (the range investigated here) the

37

offset in the contribution of Ce4+ to the total Ce M-edge area from the actual proportion of Ce4+

in the glass can be expressed as:

4 4 ( 0.004) 4XANESCe / Ce 578.7[Ce / Ce 5%] 0 Ce / Ce 0.4 A (2.10)

where the exponent A is a function of the NBO/T value of the glass expressed as:

4 3

2

log ( 0.004) 0.754 log NBO/T 1.119 log NBO/T

0.577 log NBO/T 0.119 log NBO/T 0.709

A

(2.11)

4XANES[Ce / Ce] is the contribution of Ce4+ to the total area of the M4,5-edge, and 4Ce / Ce is

the actual fraction of Ce in a 4+ oxidation state in the glass. The 4Ce / Ce and log (Ce4+/Ce3+)

determined potentiometrically and by XANES for each glass, as well as the values of ‘A’ for

each composition are summarized in Table 2.3. Equation 2.10 is only calibrated to Ce4+

fractions up to 0.07 (BH09), 0.12 (AA08), 0.11 (DA09), and 0.32 (RH08). Evaluation of these

relations at higher Ce4+/Ce3+ would require oxygen pressures significantly exceeding 1 bar.

The compositional dependence to the XANES response, as implied by Equation 2.10, is a

somewhat surprising result, since the 4f electrons lie within the closed 5s and 5p shells, and

should therefore be well shielded from crystal field and bonding effects. However, this may be

explained by the Ce4+ peaks at 884.1 (D) and 902.0 eV ( D ) which appear to result from Ce4+ in

a 4f1L ground state configuration, analogous to the lower energy peak observed in the Ce4+ L3-

edge. The local environment of the oxygen would have an effect on the formation of the ligand

2p hole and long-range interactions would then be expected to affect the Ce M4,5-edge XANES

spectra.

2.5 Conclusions

For the samples analyzed in this study we have found Ce4+ to have a larger contribution to the

total area of the Ce M4,5-edge versus Ce3+ in silicate glasses than predicted by chemical methods.

This results in overestimations of the Ce4+ content when simply summing the contributions of the

individual species to the total area of the M4,5-edge. This effect appears to be more pronounced

in highly polymerized melts suggesting a dependence on long range interactions involving the

38

local environment of neighboring oxygen. The correction offered here allows for accurate and

rapid determinations of Ce4+/Ce3+ in alkali-aluminosilicate glasses of a range of compositions

even at Ce4+ concentrations below detection by chemical methods.

2.6 References

Alford, T.L., Feldman, L.C., Mayer, J.W., 2007. Fundamentals of Nanoscale Thin Film Analysis.

Springer, Boston, MA, USA.

Andrade, L.H.C., Lima, S.M., Novatski, A., Steimacher, A., Rohling, J.H., Medina, A.N., Bento,

A.C., Baesso, M.L., Guyot, Y., Boulon, G., 2009. A step forward toward smart white lighting:

Combination of glass phosphor and light emitting diodes. Applied Physics Letters 95.

Bianconi, A., Marcelli, A., Dexpert, H., Karnatak, R.C., Kotani, A., Jo, T., Petiau, J., 1987.

Specific intermediate-valence state of insulating 4f compounds detected by L3 x-ray absorption.

Physical Review B 35, 806-812.

Bonnelle, C., Karnatak, R.C., Sugar, J., 1974. Photoabsorption in the vicinity of 3d absorption

edges os La, La2O3, Ce, and CeO2. Physical Review A 9, 1920-1923.

Chiodini, N., Fasoli, M., Martini, M., Rosetta, E., Spinolo, G., Vedda, A., Nikl, M., Solovieva,

N., Baraldi, A., Capelletti, R., 2002. High-efficiency SiO2 : Ce3+ glass scintillators. Applied

Physics Letters 81, 4374-4376.

Darab, J.G., Li, H., Vienna, J.D., 1998. X-ray absorption spectroscopic investigation of the

environment of cerium in glasses based on complex cerium alkali borosilicate compositions.

Journal of Non-Crystalline Solids 226, 162-174.

Efimov, A.M., Ignat'ev, A.I., Nikonorov, N.V., Postnikov, E.S., 2011. Spectral components that

form UV absorption spectrum of Ce3+ and Ce(IV) valence states in matrix of

photothermorefractive glasses. Optics and Spectroscopy 111, 426-433.

Garvie, L.A.J., Buseck, P.R., 1999. Determination of Ce4+/Ce3+ in electron-beam-damaged CeO2

by electron energy-loss spectroscopy. Journal of Physics and Chemistry of Solids 60, 1943-1947.

39

Gunnarsson, O., Schonhammer, K., 1983. Electron spectroscopies for Ce compounds in the

model Physical Review B 28, 4315-4341.

Jo, T., Kotani, A., 1988. Effect of valence mixing on multiplet structure in core photoabsorption

spectra for Ce compounds. Physical Review B 38, 830-833.

Kaindl, G., Kalkowski, G., Brewer, W.D., Perscheid, B., Holtzberg, F., 1984. M-edge X-ray

absorption-spectroscopy of 4f instabilities in rare-earth systems. J. Appl. Phys. 55, 1910-1915.

Kilbourn, B.T., 2011. Cerium and Cerium Compounds, Kirk-Othmer Encyclopedia of Chemical

Technology. , pp. 1-23.

Moringa, K., Yoshida, H., Takebe, H., 1994. Compositional Dependence of Absorption Spectra

of Ti3+ in Silicate, Borate, and Phosphate Glasses. Journal of the American Ceramic Society 77,

3113-3118.

Paul, A., 1976. Cerium - titanium yellow color in glass. Physics and Chemistry of Glasses 17, 7-

9.

Paul, A., Douglas, R.W., 1965. Cerous-ceric equilibrium in binary alkali borate and alkali

silicate glasses. Physics and Chemistry of Glasses 6, 212-215.

Salinas, A.J., Shruti, S., Malavasi, G., Menabue, L., Vallet-Regi, M., 2011. Substitutions of

cerium, gallium and zinc in ordered mesoporous bioactive glasses. Acta Biomaterialia 7, 3452-

3458.




Shahin, A.M., Grandjean, F., Long, G.J., Schuman, T.P., 2005. Cerium LIII-edge XAS

investigation of the structure of crystalline and amorphous cerium oxides. Chemistry of

Materials 17, 315-321.

Stroud, J.S., 1962. Color centers in a cerium-containing silicate glass. Journal of Chemical

Physics 37, 836-841.

40

Takahashi, Y., Sakami, H., Nomura, M., 2002. Determination of the oxidation state of cerium in

rocks by Ce LIII-edge X-ray absorption near-edge structure spectroscopy. Analytica Chimica

Acta 468, 345-354.

Temme, N. M. 2010. Voigt function, in Olver, Frank W. J., Lozier, D. M., Boisvert, R. F., Clark,

C. W., NIST Handbook of Mathematical Functions, Cambridge University Press.

Wilke, M., Jugo, P.J., Klimm, K., Susini, J., Botcharnikov, R.E., Kohn, S.C., Janousch, M.,

2008. The origin of S4+ detected in silicate glasses by XANES. American Mineralogist 93, 235-

240.

Yagci, O., 1985. The M4,5 photo-absorption spectra of cerium in CeO2 and oxidation of metallic

cerium. Journal of Physics C-Solid State Physics 19, 3487-3495.

Zanella, G., Zannoni, R., Dalligna, R., Locardi, B., Polato, P., Bettinelli, M., Marigo, A., 1994. A

new cerium scintillating glass for X-ray detection. Nuclear Instruments & Methods in Physics

Research Section a-Accelerators Spectrometers Detectors and Associated Equipment 345, 198-

201.

41

Table 2.1 Bulk composition of starting materials in weight percent (excluding Ce).

BH09 AA08 DA09 RH08

SiO2 51.7 61.5 70.6 72.1

TiO2 1.1 0.6 0.4 0.3

Al2O3 16.8 18.7 15.5 14.8

MgO 11.9 4.2 1.5 0.7

CaO 15.9 11.5 4.3 1.7

Na2O 2.6 2.8 4.6 6.4

K2O 0.7 3.1 3.9

P2O5 0.1

NBO/T1 0.78 0.27 0.10 0.08

*All compositions contained 1.0-1.2 wt.% CeO2 1 Non-bridging oxygen vs. tetrahedrally coordinated

cations, calculated as (2O - 4T)/T, where T = Si + Ti + Al + P in atomic percent, and O is the atomic percent of oxygen. Al was assigned to T since Al < Na + K + 2Ca + 2Mg in all compositions.

42

Table 2.2 Curve parameters used to fit Ce M4,5-edge. E = the energy position, d = Gaussian width, and z = the Lorentzian-Gaussian

width ratio. Letters A-E (and A’-E’) correspond to features shown in Figure 2.3. Errors given in brackets.

M5 M4

Curve Type E (eV) d (eV) z Curve Type E (eV) d (eV) z A Voigt 879.7 (0.9) 0.37 (0.48) 1.38 (0.04) A' Voigt 896.5 (0.1) 0.16 (0.03) 0.00 (0.00)B Voigt 881.5 (0.2) 0.30 (0.15) 1.42 (0.17) B' Voigt 898.3 (0.1) 0.32 (0.07) 0.82 (0.29)C Voigt 882.5 (0.1) 0.29 (0.22) 1.29 (0.82) C' Voigt 899.7 (0.1) 0.25 (0.03) 2.85 (0.40)D Voigt 884.0 (0.2) 0.20 (0.27) 3.67 (0.61) D' Voigt 901.9 (0.1) 0.36 (0.03) 3.42 (0.18)E Voigt 889.0 (0.3) 0.63 (0.06) 0.80 (0.03) E' Voigt 907.0 (0.3) 0.44 (0.02) 2.88 (0.09)

Arctan 900.2 (0.4) 0.24 (0.00) -

43

Table 2.3 Results of potentiometric and XANES analyses of glasses and values for ‘A’ as calculated by Equation 2.11 for the different

melt compositions. Errors given in brackets.

Sample Comp. Temp. (C) log fO2 Chemical XANES (Area Ratio) A

Ce4+/Cea log Ce4+/Ce3+ Ce4+/Ceb log Ce4+/Ce3+

DS09-C3-10 RH08 1298 -2.73 0.095 (0.005) -0.979 (0.030) 0.358 (0.018) -0.254 (0.034) 8.096

DS11-C36 RH08 1500 0.00 0.133 (0.007) -0.814 (0.032) 0.333 (0.017) -0.301 (0.033) 8.096

DS09-C3-19 RH08 1399 -0.02 0.200 (0.010) -0.602 (0.029) 0.386 (0.019) -0.202 (0.035) 8.096

DS09-C2-13 RH08 1299 0.00 0.281 (0.005) -0.408 (0.011) 0.397 (0.020) -0.181 (0.036) 8.096

DS08-C1-20 RH08 1100 -0.56 - - 0.215 (0.021) -0.563 (0.056) 8.096

DS11-C35 DA09 1500 0.00 0.111 (0.006) -0.904 (0.033) 0.231 (0.012) -0.523 (0.028) 6.050

DS09-C3-25 DA09 1399 -0.02 0.051 (0.006) -1.270 (0.072) 0.228 (0.011) -0.531 (0.028) 6.050

DS11-C34 AA08 1500 0.00 0.086 (0.013) -1.026 (0.095) 0.173 (0.009) -0.679 (0.026) 5.023

DS09-C3-18 AA08 1399 -0.02 0.118 (0.006) -0.874 (0.030) 0.179 (0.009) -0.662 (0.026) 5.023

DS11-C21 BH09 1799 -2.40 0.011 (0.006) -1.954 (0.319) 0.116 (0.006) -0.882 (0.025) 5.018

DS09-C3-07 BH09 1299 -2.73 0.034 (0.005) -1.453 (0.094) 0.143 (0.007) -0.779 (0.025) 5.018

DS09-C2-12 BH09 1299 0.00 0.075 (0.009) -1.091 (0.074) 0.168 (0.008) -0.694 (0.026) 5.018

44

Figure 2.1 Calibration curve depicting the volume of ammonium Fe(II) sulfate required to reach

the potentiometric inflection point as a function of the mass of Ce4+ added to the sample solution.

45

Figure 2.2 Plot of log Ce4+/Ce3+ versus log fO2 for RH08 and BH09 compositions synthesized at

the same temperature. Lines are of 1/4 slope as predicted for a one electron exchange reaction.

Errors for RH08 are smaller than the data symbols.

46

Figure 2.3 Background corrected Ce M4,5-edge XANES spectra collected in the region of 870 to

915 eV of CeF3, CeO2 and sample DS09-C2-13 (composition RH08). Letters A-E and A’-E’

mark the main features of the M5 and M4 edges, respectively.

47

Figure 2.4 Example of curve fitting used to determine contributions to the total area from Ce4+

and Ce3+ for the Ce M5- (a) and Ce M4-edge (b). Black symbols are measured values, while grey

curves are the model used to fit the data. Spectrum shown is of sample DS09-C2-13.

48

Figure 2.5 Time series of Ce M4,5-edge under beam exposure. Before the last scan the sample

was exposed to the full beam for three min. Estimations of 4Ce / Ce given below durations of

beam exposure. Errors (given in brackets) are estimated as 10% due to the relatively poor signal

to noise ratio. No evidence of modification of the spectra is observed.

49

Figure 2.6 Comparison between Ce4+ fractions determined using the potentiometric method and

the contribution to the total M-edge area from Ce4+. Black curves show the results from

Equation 2.10 using the NBO/T values for the given melt compositions.

50

Chapter 3

Determination of Cerium Oxidation State in Silicate Melts:

Combined fO2, Composition, and Temperature Effects

To quantify the relative proportions of Ce3+ and Ce4+ in natural magmas, we have synthesized a

series of Ce doped glasses ranging in composition from basalt to rhyolite (± H2O) at 0.001 and 1

GPa, under fO2 conditions varying from FMQ –6.0 to FMQ +8.4, and temperatures from 1200 to

1500°C. The Ce4+/Ce3+ ratio in the experimental run products was determined both

potentiometrically and in situ, using Ce M4,5-edge x-ray absorption near-edge structure (XANES)

spectroscopy. For a given melt composition, the change in Ce4+/Ce3+ ratio with fO2 follows the

trend predicted from the reaction stoichiometry assuming simple oxides as melt species. In

addition to fO2, melt composition and water content have been found to be secondary controls on

Ce4+/Ce3+, with more depolymerized melts and hydrous compositions favouring the stabilization

of Ce3+. The Ce4+/Ce3+ ratio can be expressed through the equation,

22

3/2

2

CeO 1 5705( 257) NBOlog log O 0.8694( 0.005)

CeO 4 T

3.856( 0.049) H O 3.889( 0.102)

fT

x

where T is in Kelvin, NBO/T is the proportion of non-bridging oxygen to tetrahedrally

coordinated cations, and xH2O is the mole fraction (calculated using molecular oxides, e.g.

Al2O3, Na2O) of water dissolved in the melt. Results indicate that even at relatively low oxygen

fugacity, trace amounts of Ce4+ will be present in most terrestrial igneous systems, suggesting

that Ce partitioning could be a sensitive indicator of fO2.

3.1 Introduction and Theoretical Background

Cerium is unique among the rare-earth elements (REEs) as it can exist as 4+ in addition to the 3+

valence state common to this element group. Although Ce4+ has been typically associated with

Earth’s surface environment, thermodynamic data for the component oxides (Zinkevich et al.,

2006) and measurements of Ce redox state in silicate glasses (Darab et al., 1998; Paul and

Douglas, 1965) suggest that both Ce3+ and Ce4+ should also be present in igneous systems under

conditions experienced by terrestrial magmas. This prediction is supported by the observation of

51

positive Ce anomalies relative to other REE in igneous phases which have a higher affinity for

tetravalent cations, such as zircon and cassiterite (Hanchar and van Westrenen, 2007; Jiang et al.,

2004). Laboratory partitioning studies (Burnham and Berry, 2012; Luo and Ayers, 2009; Trail et

al., 2011) have also observed enhanced partitioning of Ce into zircon under highly oxidizing

conditions, presumably reflecting the stabilization of Ce4+ at high fO2. As described by Ballard

et al (2002), detailed knowledge of Ce4+/Ce3+ in silicate melts, combined with mineral/melt

partitioning models, offers a potentially powerful means to assess magma redox state. A

fundamental limitation to this pursuit, however, is information on the Ce4+/Ce3+ variation with

fO2 and composition in melts.

The redox state of Ce in molten silicate can be expressed via the homogeneous reaction:

3/2 2 2

1CeO O CeO


The corresponding equilibrium constant, K, for this reaction is:

3/2

2

1/4CeO

C O

2

e

Ome

m lt

lt

e

Ka f

a

(3.2)

where 2CeO

melta and 3/2CeO

melta are the activities of the CeO2 and CeO3/2 components in the melt. Taking

the natural logarithm of both sides of this equation yields the expression:

2

3/2

CeO2

CeO

1log log log O

4

melt

melt

aK f

a (3.3)

which, by separating the activity terms into their constituents yields:

2 2

3/2 3/2

CeO CeO2

CeO CeO

1log log log log O

4

melt melt

melt melt

xK f

x (3.4)

where x and are the mole fractions and activity coefficients, respectively, of CeO2 and CeO3/2

in the melt. Rearrangement of Equation 3.4 results in the expression:

52

2 2

3/2 3/2

CeO CeO2

CeO CeO

1log log O log log

4

melt melt

melt melt

xf K

x (3.5)

For low concentrations of Ce it may be assumed that 3/2CeO

melt and 2CeO

melt are constants for a given

melt composition (Henry’s Law). Since both log K and log 2 3/2CeO CeO

melt melt are constants under set

conditions, plotting the logarithm of the mole ratio verses log fO2 should then yield a straight

line, with a slope of 1/4, and an intercept corresponding to the sum of – log 2 3/2CeO CeO

melt melt and log

K. Both K and the ratio of the activity coefficients are implicit functions of temperature and melt

composition. The temperature dependence of K can be explicitly expressed with the form,

2.30259 log

r

r r

G RT K

H T S (3.6)

where rG, rH, and rS are the standard state Gibbs energy, enthalpy, and entropy of the

reaction, R is the gas constant, and T is temperature in Kelvin. Based on heat capacity (CP) data

for the component Ce oxides, the change in log K due to rCP over the temperature range

investigated can be considered to be negligible relative to the rH. Therefore,

ln2.30259 2.30259

r rH S

KRT R

(3.7)

Assuming both rH and rS are constant with T, this can then be expressed as:

lna

K bT

(3.8)

where a and b are constants for a particular melt composition. Previous work has shown K may

also be a function of melt composition (Sack et al., 1980) in addition to fO2 and T, fully

described by the empirical expression:

2

3/2

CeO2

CeO

1log log O

4

melt

i imelti

x af b x c

x T (3.9)

where bi and c are constants, and xi is the mole fraction of component i in the melt.

53

In this study we employ two different methods to measure Ce valence state. The first involves

digestion of the sample in a HF – H2SO4 solution then titrating against a standard reducing agent,

in this case ammonium Fe(II) sulfate. In systems where other redox sensitive elements are

present, analyses through wet chemical techniques become problematic. Apart from the

potential for charge transfer on quench, re-equilibration of the redox pairs in the aqueous

solution during digestion is an unavoidable consequence of the analytical method. To obviate

these complexities, experiments were done with Fe-free analogues of natural silicate melts.

Furthermore, the detection limits for Ce4+/Ce3+ determined by this technique were found to be

~0.01 at best. Evaluation of Ce oxidation state at low fO2 was therefore not possible through wet

chemical analysis.

To complement the wet chemical method, determinations of Ce redox state have also been done

using synchrotron based x-ray absorption. Measurement of Ce4+/Ce3+ using synchrotron

radiation typically exploits the Ce L3-edge X-ray absorption near-edge structure (XANES)

spectra, which involves a transition of a 2p electron to an unoccupied 5d state. Depending on

oxidation state the L3-edge shifts from ~5723 eV for Ce3+ to ~5740 for Ce4+, resulting from

differences in the coulomb interaction between the 2p core hole and the valence band. In the

present study determination of Ce4+/Ce3+ was done using Ce M4,5-edge XANES. The Ce M4,5

absorption edge results from a 3d → 4f electron transition and is characterized by two distinct

line groups located at 902.4 (M4) and 883.8 eV (M5) corresponding to 3d3/24f5/2 and 3d5/24f7/2,

respectively. This transition has the advantage of being at a lower energy, minimizing the

potential for modification of the Ce oxidation state under beam exposure (Shimizugawa et al.,

2001; Wilke et al., 2008) which is of concern, particularly when analyzing hydrous samples.

Past experimental work by Schreiber et al. (1980) suggested that Ce4+ was unstable in magmatic

systems and would be converted to Ce3+ by oxidation of Fe2+. This result, however, is clearly at

odds with the inferred presence of Ce4+ in natural igneous rocks. It now seems probable that

Ce4+ was eliminated in the glasses produced in the Schreiber et al. (1980) study as a result of

electron exchange with Fe2+ during quench, as shown for Cr oxidation state in Fe-bearing melts

(Berry et al., 2003).

In this study we use both wet chemical methods and XANES analysis to determine the redox

state of Ce over an fO2 interval which spans 15 orders of magnitude. Melt compositions were

54

both hydrous and anhydrous and vary from rhyolite to basalt, but are free of other multivalent

cations (with the exception of H in the 1 GPa runs) to prevent charge transfer during quench.

The resulting dataset allows for the accurate prediction of Ce valence over a range of T, fO2 and

composition relevant to magmas generated on Earth and other solar system bodies.

3.2 Materials and Methods

Experiment starting materials were prepared from reagent grade oxides and carbonates (>99.99%

purity), first ground using an agate mortar and pestle under ethanol until dry before being

calcined at 1100C for 8 to 12 hours. The mixtures were then fused twice at 1500C in a

platinum crucible for approximately 30 min and quenched in water. After each fusion the

resulting glass was ground twice under ethanol until dry, so as to ensure homogenous starting

materials. Bulk compositions of the starting materials are provided in Table 3.1, which consist

of synthetic basalt (BH09), andesite (AA08), and rhyolite (RH08).

Experiments done at atmospheric pressure used a vertical-tube gas mixing furnace, with oxygen

fugacity controlled using either CO-CO2 gas mixtures or pure O2. Owing to space limitations

within the furnace, T and fO2 were not monitored during each experiment. Instead, temperature

within the furnace was checked before and after each experiment using a Pt-Pt10%Rh

thermocouple. The fO2 of the furnace atmosphere was also measured before and after each

experiment using a SIRO2 C700+ solid zirconia electrolyte oxygen sensor purchased from

Ceramic Oxide Fabricators©.

A typical sample contained approximately 100 mg of glass starting material mixed with

polyvinyl alcohol and applied to a wire loop made from either Pt or Pt10%Rh, then dried. The

loop + glass were initially suspended within the cool end of the sealed furnace from a fused silica

glass rod, which was lowered into the hot spot after the furnace atmosphere had been allowed to

equilibrate. Experiments were terminated by quenching the melt bead in cold H2O. In some

cases up to four samples were run simultaneously using a “chandelier” arrangement made from

fused silica rod.

Loss of alkali elements through volatilization is a common problem in one atmosphere high

temperature experiments done at low fO2, as the dominant alkali species in the gas phase are

monatomic Na and K (O'Neill, 2005). To prevent the loss of alkalis from our experiments under

55

these conditions we have employed the use of a silicate alkali reservoir melt. This technique has

been shown to fix the activity of alkali metal oxides in experiments done at one atmosphere

(Borisov et al., 2006; O'Neill, 2005). The configuration for these experiments is shown in Figure

3.1a and consists of a Pt crucible containing ~1 g of reservoir melt suspended under the sample.

The sample was positioned within the walls of the crucible and approximately 3 cm above the

reservoir melt. Initial experiments using the alkali-disilicate melt employed by Borisov et al.

(2006) resulted in an overabundance of alkali elements in experimental run products. In

subsequent experiments, we found that reservoir melts of approximately the same composition as

the sample, but enriched in Na and K by 10%, resulted in run products with alkali contents

varying by less than 0.1 wt% from their initial content.

To determine the run durations necessary for the Ce oxidation state to reach equilibrium values,

we performed a time series on the RH08 composition, as this is the most polymerized

composition and likely the slowest to react. Samples were run at 1200C under pure oxygen for

durations of 2, 12, and 48 hours. Run products were analyzed spectroscopically (see below) and

Ce4+/Ce3+ was found to be consistent for all durations investigated. In light of these results, all

subsequent experiments were equilibrated for ~12 h.

We also performed a series of hydrous experiments at 1 GPa using a piston cylinder apparatus.

Starting materials were weighed to 100 mg and thoroughly mixed with a noble metal oxide

buffer (Ru-RuO2, Ir-IrO2; O'Neill and Nell, 1997) to promote rapid equilibration at the imposed

fO2. To prevent exhaustion of the buffer due to H2 diffusion into the experimental charge,

additional buffer was added to the top and bottom of the glass starting material. Samples were

sealed in Pt capsules containing 2-6 L of deionized H2O. Capsules were placed vertically into

holes drilled into a graphite slug, which was contained in a fired pyrophyllite cup to prevent

contact with the graphite furnace. The piston-cylinder apparatus employed a 1.905 cm bore

pressure vessel with pressure cells consisting of MgO filler pieces and a graphite furnace fit into

concentric sleeves of Pyrex (inner) and NaCl (outer, Figure 3.1b). After quenching, samples

were re-weighed to check for H2O loss.

Run products were optically transparent and varied from clear to pale yellow in colour. The

yellow colour was most intense in glasses with the highest Ti content, as would be expected

given that Ce-Ti interaction is a known colouring agent in glass (Paul, 1976). All run products

56

were inspected optically to ensure that the samples were crystal-free, thus indicating that the

equilibration temperature was above the liquidus for the given composition. A summary of

experimental run conditions is provided in Table 3.2.

3.3 Analytical

3.3.1 Electron Microprobe

All experimental run products were analyzed for major and minor elements using a Cameca

SX50 electron microprobe housed in the Department of Earth Sciences, University of Toronto.

Samples were prepared for analysis by mounting in epoxy, then polishing with diamond grit

down to 1 m, followed by 0.3 m alumina. Beam conditions employed for analyses were 15 kV

acceleration voltage, 20 m defocused beam and 5 nA beam current. Standards used for

calibration were natural basaltic glass (Mg, Ca), obsidian (Si, Al), albite glass (Na), sanidine (K),

TiO2 (Ti), and CePO4 (P, Ce). To prevent their underestimation caused by migration under the

electron beam, alkali elements were analyzed at the beginning of the acquisition cycle. The ZAF

correction routine was used to convert raw count rates to element concentrations. Typically, 8 to

10 spots were measured on each sample. Table 3.3 provides the averages of these analyses.

3.3.2 Potentiometric Titrations

Potentiometric determinations of the Ce4+ content of samples were done using the method

described in detail in Chapter 2 which is briefly summarized here. Approximately 20 mg of

sample was crushed in an agate mortar then dissolved in 1.0 mL of a 12% HF – 7% H2SO4

mixture which was cooled in an ice bath. Following the digestion, 250 mg of boric acid was

added to the sample solution to complex with the excess HF, preventing the gradual reduction of

Ce4+. The resulting solution was then diluted with 35 mL of deionized H2O and titrated against a

0.0001 N ammonium Fe (II) sulfate solution. The voltage of the solution was monitored using a

Pt-pin platinum indicator electrode and a Ag-AgCl reference electrode which were connected to

an Orion model 525a+ pH meter.

Calibration of this technique was done by mixing ~20 mg of Ce-free blank glass with measured

amounts of Ce standard compounds, which included Ce(SO4)2, (NH4)4Ce(SO4)4·2H2O, and

(NH4)3Ce(NO3)6. A calibration curve was then constructed by plotting the volume of titrant

added at the equivalence point verses the mass of Ce4+ added. Titrations of standard material

57

were run both before and after the unknowns to ensure oxidation of the ammonium Fe (II)

sulphate solution had not occurred over the course of the analyses.

3.3.3 Ce M4,5-edge XANES

Cerium M4,5-edge XANES spectra were collected on the Spherical Grating Monochromator

(SGM) undulating beamline at the Canadian Light Source, University of Saskatchewan.

Samples consisted of freshly cleaved coarse glass chips which were mounted on stainless steel

disks with carbon tape. The disks were placed in the absorption chamber under a vacuum of

<10-7 Torr. The samples were faced approximately 45 toward the 1000 m x 100 m incident

beam. Spectra were collected at room temperature in the region between 870 and 920 eV with a

0.2 eV sampling step size over the edge region in both fluorescence yield (FLY) and total

electron yield (TEY) modes. The TEY spectra were found to have significantly higher noise than

the FLY spectra resulting from charge buildup due to the insulating nature of the glasses. As a

result, only the FLY spectra were used for redox measurements. X-ray energies were calibrated

using CeO2 as a standard.

Three spectra were recorded per sample, which were then normalized and averaged. Some

averages consisted of only two spectra if one was found to be too noisy. Although this was not a

significant problem, approximately 10% had signal to noise ratios on the order of 10:1 or lower,

in which case the spectra were not used for redox determinations. The Ce3+ M-edge consists of

two edges centered at ~882.5 (M5) and ~899.8 eV (M4) with resolvable satellite peaks located at

approximately 879.6, 881.5, 896.6 and 898.4 eV. The Ce4+ M-edge is shifted to higher energies

by approximately 2 eV with main peaks at ~884.1 (M5) and ~902.0 eV (M4) and smaller features

at ~889.0 and ~907.0 eV (Figure 3.2). Details of the curve fitting procedure can be found in

Chapter 2. We found that the Ce4+/Ce3+ obtained from M4,5-edge area measurements somewhat

overestimates the true Ce4+/Ce3+, as determined potentiometrically. The M4,5-edge

measurements are corrected for this effect using the empirical relation:

4 4 ( 0.004) 4XANESCe / Ce 578.7[Ce / Ce 5%] 0 Ce / Ce 0.4A (3.10)

where A is a function of the NBO/T value of the glass expressed as:

58

4 3

2

log ( 0.004) 0.754 log NBO/T 1.119 log NBO/T

0.577 log NBO/T 0.119 log NBO/T 0.709

A

(3.11)

4XANES[Ce / Ce] is the Ce4+contribution to the total area of the M4,5-edge, and 4+Ce /ΣCe is the

actual fraction of Ce4+ in the glass.

3.4 Results

3.4.1 Effect of Temperature

To investigate the dependence of Ce oxidation state on the equilibration temperature a series of

experiments were done from 1300 to 1500C and constant fO2. For most of these experiments,

pure O2 was chosen for the furnace atmosphere to yield the highest concentration of Ce4+,

insuring that the Ce4+/Ce3+ ratio would be well above the detection limit of the wet chemical

technique. Two experiments were also done at more reducing conditions (log fO2 = -2.4). The

temperature dependence for a given melt composition can then be evaluated through Equation

3.9, which can be re-written as:

2

3/2

CeO

CeO

log melt

melt

x aB

x T (3.12)

In which B is a constant for a given melt composition at fixed fO2.

Over the temperature range investigated, log Ce4+/Ce3+ varies linearly with the reciprocal

temperature for all compositions (Figure 3.3). The slope of the resulting regression yields the

enthalpy of the reaction, determined using the Van’t Hoff relation. Values are negative, as is

expected for an exothermic oxidation reaction. There appears to be no discernible compositional

or fO2 dependence on the enthalpy of the redox reaction, with values for all samples being within

error of -109.2 (± 4.8) kJ/mol. This similarity of reaction enthalpy differs from results obtained

in binary alkali-oxide silica systems, in which enthalpy correlates strongly with melt basicity

(Paul and Douglas, 1965). The average enthalpy determined in this study is substantially higher

than values for melts of Na2O·2SiO2 (33.5 kJ/mol; Johnston, 1965), 3Li2O·7SiO2 (76 kJ/mol) and

alkali-borate melt compositions (-10 to -76 kJ/mol; Paul and Douglas, 1965). This is not

surprising, however, given the markedly different compositions in these investigations. Of the

59

compositions investigated by Paul and Douglas (1965), those most similar in composition

(3Na2O·7SiO2 and 3K2O·7SiO2) to the range investigated in this study agree fairly well (-100 and

-130 kJ/mol, respectively) with the enthalpy determined here. The close agreement between

these alkali-silicate melts and the more complex alkali-poor compositions of this study suggests

the enthalpy of the reaction in Equation 3.1 should remain essentially constant to alkali contents

extending beyond those investigated here.

3.4.2 Effect of Oxygen Fugacity

As discussed previously, Equation 3.5 predicts a linear relation between 2 2/3CeO CeOlog /melt meltx x and log

fO2 with an expected slope of 1/4 for data obtained at constant temperature and a given melt

composition. For the compositions analyzed in this study, the regressed slopes are 0.225 (±

0.025), 0.231 (± 0.031) and 0.274 (± 0.028), for RH08, AA08 and BH09, respectively. The

model, therefore, holds true as a slope of 1/4 can be fit through the data within error (Figure 3.4).

Such results are in agreement with the observations from earlier work involving simple silicate-,

and borate-based glass compositions (Johnston, 1965; Paul and Douglas, 1965; Schreiber et al.,

1980).

3.4.3 Melt Composition

In addition to redox conditions and temperature, the redox state of Ce was also found to change

systematically with the composition of the host melt. This is seen in Figures 3.3 and 3.4 where

Ce4+/Ce3+ in the basaltic composition BH09 is lower relative to the more silicic compositions

AA08 and RH08 under the same T – fO2 conditions. Given the number of components present in

the three compositions investigated here, regression analysis to determine values of bi in

Equation 3.9 is not possible. To apply a model of the form of Equation 3.9 would require

investigation of at least as many compositions as components. Instead, we sought a single

compositional parameter which accounts for the bulk solution properties of the melt. A number

of melt parameterizations have been put forth to describe the solution properties of silicate melts,

including optical basicity (Duffy and Ingram, 1971), M, the cation ratio (Na + K + 2Ca)/(Al · Si),

(Watson and Harrison, 1983), metal to oxygen ratio (M/O) (Henderson et al., 1985) as well as

the number of non-bridging oxygens (NBO) to tetrahedrally-coordinated cations (T), NBO/T

(Virgo et al., 1980). Here we have chosen to use NBO/T as this yields the best fit to our data out

60

of the parameters mentioned (Figure 3.5). By expressing the melt composition through NBO/T,

Equation 3.9 becomes:

2

3/2

CeO2

CeO

1 NBOlog log O

4 T

melt

melt

x af b c

x T (3.13)

in which values of b and c are constants determined through regression analysis. Fitting all of

the anhydrous data to Equation 3.13 yields b = -0.8694 (±0.005) and c = -3.889 (±0.102).

3.4.4 Water Content

Determination of the Ce4+/Ce3+ in hydrous glasses using potentiometric methods is complicated

by possible reaction with the metal + oxide added to buffer fO2 during synthesis. Although the

noble metal oxide buffers have very low solubilities in silicate melts (discussed in Section 3.5.1),

these were finely dispersed through the glass to facilitated equilibration. Consequently, we use

the Ce M4,5-edge XANES to evaluate the effect of H2O on Ce oxidation state. Based on the

XANES analysis, the presence of H2O was found to result in a decrease in the proportion of Ce4+

in the melt (Figure 3.5). This result is expected given the known depolymerizing effect of OH- on

silicate melt structure (Mysen and Virgo, 1986). However, the resulting increase in NBO/T

caused by the addition of H2O is insufficient to explain the decrease in Ce4+ concentration using

Equation 3.13. We therefore calculate NBO/T on an anhydrous basis, and treat water separately

from the other melt components, through the introduction of an additional H2O term to Equation

3.13:

2

3/2

CeO2 2

CeO

1 NBOlog log O H O

4 T T

melt

melt

x af b c d x

x (3.14)

Where d is a constant, determined by regression, to equal -3.856 (±0.049), and xH2O is the mole

fraction of total H2O dissolved in the melt.

3.5 Discussion

3.5.1 Quench Modification of Ce Oxidation State

A possible complication when determining the redox state of Ce in hydrous glasses is the

modification of the Ce4+/Ce3+ by reaction between dissolved Ce and either the buffer material or

61

dissolved H during quench. For the case of the buffer, two types of reactions seem likely. The

first is between Ce and the dispersed solid oxide and metal, constituting a heterogeneous reaction

involving either decomposition of the metal oxide (Ce3+ oxidation) or oxidation of the metal

(Ce4+ reduction). Both reactions require bulk transport of oxygen through the melt. The diffusive

lengthscale for this process, as estimated from data on oxygen diffusion in molten silicate (Dunn,

1982), is ~2.1 m over the ~10 seconds required to cool the experiment below the glass

transition (~700C). Given this short length scale, any modification of the Ce oxidation state in

the melt will be negligible. The second type of reaction involving the buffer is the homogeneous

equilibrium involving dissolved components in the melt. Under the oxidizing conditions

investigate in the study Ru3+ (Borisov and Nachtweyh, 1998) and Ir2+ (O'Neill et al., 1995), or

possibly Ir3+ (Borisov and Palme, 1995), are likely melt species. Oxidation of Ce3+ may then

occur through the electron exchange reaction:

3 4Ce Me Ce Menmetaln n (3.15)

where Me is Ru or Ir. The estimated solubilities of Ru and Ir in the melt at the fO2 of our

experiments are ~85 ppm Ru (Borisov and Nachtweyh, 1998), and ~1 ppm Ir (Brenan and

McDonough, 2009), which are 103 to 104 times less than the amount of dissolved Ce, indicating

that the Ce4+/Ce3+ will not be shifted significantly by this effect.

Charge transfer between Ce and H is more of a potential problem and would take place through

the reaction,

2 2 2 3/2

1 1CeO ( ) H ( ) H O( ) CeO ( )

2 2melt melt melt melt (3.16)

The redox pair with the lower enthalpy for the oxidation reaction will be reduced during the

quenching process. The enthalpy of the Ce redox reaction has been determined in this study,

whereas the equivalent data for hydrogen in silicate melts is not well constrained. The enthalpy

of formation of H2O under the P, T conditions of our experiments ranges from -256 to -261

kJ/mol (Dow Chemical Co., 1971). Assuming this is an approximation for the equivalent

homogeneous reaction involving hydrogen in silicate melts, the oxidation of H2 would then be

more exothermic than that for Ce. This would suggest that if charge transfer were occurring on

62

quench, it would be through the reduction of Ce4+, requiring the consumption of H2. The amount

of H2 in our experiments will be governed by the dissociation of H2O through the reaction:

2 2 2H O 1/ 2O H (3.17)

The extent of H2O dissociation can be calculated through the equation,

log2.30259

r

d

GK

RT (3.18)

where rG is the free energy of the reaction in the standard state, and Kd is the equilibrium

constant for the dissociation reaction,

1/2

2 2

2

O H

H O

d

f fK

f (3.19)

Using values for Kd from the JANAF thermochemical tables (Dow Chemical Co., 1971) and

assuming an H2O activity of unity, the fH2 at known fO2 can be determined. All of the hydrous

samples reported in this study were equilibrated under highly oxidizing conditions, in which the

maximum fH2 would be 0.0723. The H2O fugacity at saturation under these conditions would be

~2.1 GPa , giving a molar hydrogen to water ratio of 10-5.46. This translates to a concentration of

~0.5 ppm hydrogen in our experiments, which should be taken as a maximum value as it

assumes H2O saturation (all of the hydrous experiments in this study were water under-

saturated). Given the low availability of H2, any reduction of Ce4+ would, therefore, not shift the

measured Ce4+/Ce3+ by any resolvable amount.

If the application of the standard state enthalpy of formation of H2O to the oxidation reaction is

not valid for silicate melt systems and it is in fact lower than the enthalpy of the Ce oxidation,

charge transfer would then be expected to result in the oxidation of Ce3+. Increased water

content would then result in an intensification of this process. Contrary to this expectation, we

have found that samples with higher water contents trend to lower Ce4+ (Figure 3.5) suggesting

that the Ce4+/Ce3+ ratio established at temperature was not shifted during quench.

63

3.5.2 Comparison to Previous Work

A dependence of Ce oxidation state on melt composition similar to that documented in this study

has also been noted in previous work (Patra et al., 2000; Paul and Douglas, 1965; Schreiber et

al., 1980). The compositional range studied in earlier work differs greatly from that of the

present study, however, making it difficult to replicate past results within the context of the

model presented. Some general trends are in agreement, however. For example, our results

show the stabilization of Ce4+ with increasing silica content, which is consistent with the

observation that Ce4+ is nearly the sole Ce species in Ce-doped SiO2 glass under an atmosphere

of pure oxygen (Patra et al., 2000). In detail our model would predict that the SiO2 glass

synthesized by Patra et al. (2000) should have Ce4+/Ce of 0.80 ± 0.05, which is in remarkably

good accord, given that pure SiO2 is well outside of the composition range investigated.

Furthermore, a subset of the results from Paul and Douglas (1965), Darab et al. (1998), and

Johnston (1965), and one of the compositions of Schreiber et al. (1980) are within error of the

Ce4+/Ce3+ calculated by the model presented here (Figure 3.6).

In an experimental study investigating zircon/melt trace element partitioning, Burnham and

Berry (2012) estimated the Ce4+/Ce3+ of the melt based on the variation in partitioning with fO2.

These experiments were done at ~1300C and 1 atm using an alkali-earth aluminosilicate

composition containing ~56 wt% SiO2 (approximating a synthetic andesite, but with ~2.9 wt%

ZrO2), over an fO2 range of 15 log units. Using the run conditions and composition investigated

by Burnham and Berry (2012), we calculate the Ce4+/Ce3+ predicted by the model presented here

(Figure 3.7). These results are in excellent agreement, especially considering the difference in

melt composition compared to the range investigated here. Furthermore, since the results of

Burnham and Berry (2012) are based on the partitioning of Ce between zircon and melt, the

calculated Ce4+/Ce3+ are not subject to quench effects. This validates the inherent assumption

that the Ce4+/Ce3+ ratios measured in our glasses are representative of those at temperature.

There are some apparent discrepancies between our results and previously published data which

warrant discussion. The compositional dependence on Ce redox state observed by Paul and

Douglas (1965), and Schreiber et al. (1980) has the opposite correlation with NBO/T to that

observed for our samples. In the glasses investigated by Paul and Douglas (1965), those with the

highest alkali contents deviate the most from the model presented in this study (Figure 3.6), with

64

much higher Ce4+/Ce3+ than predicted. It is difficult to evaluate the role alkalis play in our

experiments, primarily due to the relatively restricted range investigated, and the co-variation of

alkalis with other melt components. However, values for the compositional term B show a strong

correlation with alkali content (Figure 3.8a), with our data plotting at the low alkali end of the

trend defined by this past work.

The source of the inconsistency between our results and those from Schreiber et al. (1980) is

uncertain, but may be linked in part to differences in Al content, as more Al in the melt correlates

with a decrease in the proportion of Ce4+ (Patra et al., 2000). For the range of compositions

evaluated in this study, we observe a weak negative correlation between Ce4+/Ce3+ and Al/Si

(Figure 3.8b), which can also be seen in the compositions analyzed by Schreiber et al. (1980).

One of the compositions in the study by Schreiber et al. (1980) which only contained 2.3 wt%

Al2O3, well below the natural range of aluminum contents in silicate melt systems, and showed a

relatively high Ce4+ concentration compared to our results. Since both Si and Al are network

formers this seems to suggest that the NBO/T parameter used here to describe the compositional

dependence of Ce valence may in fact be an over simplification. The oxidation state of Ce in

silicate melts may therefore be controlled by more complicated site volume factors and steric

effects, similar to those that have been observed in some crystalline materials (Capehart et al.,

1993). Application of this model to compositions with low Al2O3 contents, such as komatiitic

and picritic melts, should therefore be approached with caution.

Lastly, values of Ce4+/Ce3+ measured in the glasses produced by Darab et al. (1998) are generally

higher than predicted by our model. However, the glass compositions they synthesized contain

high concentrations of boron (7 to 12 wt% B2O3) and alkalis (15.5 to 19.5 wt% A2O) both of

which have been shown to stabilized Ce4+ in glasses (Paul and Douglas, 1965).

3.5.3 Geological Implications

Anomalous chondrite normalized concentrations of Ce relative to its neighboring REEs have

been documented in whole rock analyses of igneous lithologies from a variety of tectonic

settings (Fodor et al., 1992; Frey and Green, 1974; Shimizu et al., 1992). The presence of Ce4+

in natural systems would not be expected to affect rare earth element patterns in whole rock

analyses, however, as our results indicate that the proportion of Ce in a 4+ oxidation state would

typically be 1% or less. Therefore, Ce anomalies observed in these samples are likely imparted

65

through alteration in the surface environment (Cotten et al., 1995; Marsh, 1991) due to the

greater solubility of Ce3+ in aqueous fluids or through recycling of sediments carrying a positive

or negative Ce anomaly (Neal and Taylor, 1989). This also implies that element ratios

commonly used as indicators of source region characteristics, such as Ce/Pb, would not be

greatly affected by subtle variations in fO2, and hence Ce oxidation state.

Our results predict ~0.03-0.29% of Ce will be present as Ce4+ in granitic compositions over the

range of terrestrial fO2 (taken to be FMQ ±2). Based on lattice strain predictions (Blundy and

Wood, 1994), mineral phases such as zircon, baddeleyite, and cassiterite, which have a relatively

large 4+ cation as a major component, should efficiently discriminate between the two oxidation

states of Ce. As the estimated difference in the mineral/melt partitioning of Ce4+ and Ce3+ in

these phases is as large as 6 orders of magnitude (Ballard et al., 2002), Ce anomalies would be

expected in the above phases even if they crystalized at conditions well below FMQ. This

explains the ubiquitous presence of positive Ce anomalies observed in zircon (Ballard et al.,

2002).

Contemporaneous positive Ce anomalies (resulting from the presence of Ce4+) and negative Eu

anomalies (caused by the stabilization of Eu2+) are a common observation in terrestrial zircons.

This has posed somewhat of a paradox, as Ce4+, interpreted to represent oxidizing conditions,

and Eu2+, which is stable under reducing conditions, were thought to not coexist (Hoskin and

Schaltegger, 2003; Schreiber et al., 1980). This leads to the suggestion that other mechanisms,

such as the fractionation of plagioclase, prior to or during zircon crystallization, are required to

form simultaneous Ce and Eu anomalies (Hoskin et al., 2000). Using the model presented here

for Ce oxidation state, and that of Drake (1975) for Eu, it is possible to estimate the coexisting

Ce and Eu anomalies in zircon at equilibrium (Figure 3.9a). This shows the high affinity of

zircon for Ce4+, coupled with the stabilization of trace amounts of Ce4+ at relatively reducing

conditions (where a significant portion of Eu will exist in a 2+ oxidation state) can also account

for this observation. This is not to say that the observed Eu anomalies in zircon are strictly the

result of the prevailing redox conditions, or that one would expect fO2s calculated from

coexisting Ce and Eu anomalies to be concordant, as the crystallization of plagioclase would

indeed have large effect on the budget of Eu in a melt (shown in Figure 3.9b). Nevertheless, this

does demonstrate that the simultaneous Ce and Eu anomalies can form under equilibrium

66

conditions. A similar conclusion was reached by Burnham and Berry (2012) based on direct

measurements of zircon-melt partitioning over a range of fO2.

Analysis of synthetic glasses containing both Fe and Ce done by Schreiber et al. (1980) showed

evidence for interaction between the Ce3+ – Ce4+ and Fe2+ – Fe3+ redox pairs. The authors

suggest that in the presence of Fe2+, Ce4+ will be reduced through the reaction,

2 4 3 3Fe ( ) Ce ( ) Fe ( ) Ce ( )melt melt melt melt (3.20)

The authors argued that, since Fe2+ will always be more abundant than Ce4+ in natural magmas,

Ce4+ will not be a stable melt species. If true, this would require that the observed Ce anomalies

in phases like zircon are the result of a process other than the differences in partitioning of the

two species. Numerous studies have shown that zircons with positive Ce anomalies form from

magmas that lack such anomalies. It is difficult to envision a process other than the preferential

uptake of Ce4+ to account for these observations. On examination of the data from Schreiber et

al. (1980) it appears that the experiments containing Fe and Ce from which this conclusion was

based may have been subject to charge transfer on quench through the reaction in Equation 3.20.

Although the enthalpy of the Fe oxidation reaction in most natural silicate melts is lower than

that of Ce, values can vary significantly with melt composition (Kress and Carmichael, 1991;

Mysen et al., 1985; Sack et al., 1980). Schreiber et al. (1980) did not measure the enthalpy of the

Fe2+ → Fe3+ reaction in their experiments, but their compositions are close to those investigated

by Mysen et al. (1985). These authors studied Fe3+/Fe2+ equilibria by Mössbauer spectroscopy in

the CaO-MgO-Al2O3-SiO2-FeO system and found reaction enthalpies as high as 147.3 kJ/mol.

This is considerably more exothermic than that of the Ce3+ → Ce4+ reaction determined in this

study, and indicates that charge transfer during quenching was a possible mechanism to reduce

Ce4+ in the experiments of Schreiber et al. (1980). It is of note that Schreiber et al. (1980) infer

the presence of Ce-O-Fe “complexes” in their run-product glasses, based on electron

paramagnetic resonance (EPR) measurements and optical absorption spectroscopy. Thus, the

close proximity of these cations would serve to facilitate the charge transfer process.

3.6 Conclusions

The redox state of cerium has been determined in a series of geologically relevant melt

compositions. The measured oxidation state of Ce was found to have a dependence on fO2

67

predicted from the reaction stoichiometry. The variation in Ce4+/Ce3+ with temperature indicates

the reaction is exothermic with an enthalpy of -109.2 (± 4.8) kJ/mol, which is constant between

all investigated melt compositions. The average oxidation state of Ce was observed to increase

with polymerization of the melt and decrease with the addition of water.

These results show that Ce will exist primarily in a 3+ oxidation state under terrestrial magmatic

conditions, however, Ce4+ will be present in low abundances over a significant range of fO2. The

low proportions of Ce4+ in silicate melts would not be expected to affect the whole rock REE

patterns, as most rock forming minerals will not effectively discriminate between Ce4+ and Ce3+

and small amounts of Ce4+ will be obscured in a log cycle, typically used to evaluate REE

abundances. The Ce4+ concentrations predicted by this study for terrestrial igneous systems will,

however, result in elevated Ce content in accessory phases such as zircon which have a much

greater affinity for 4+ cations.

3.7 References

Aigner-Torres, M., Blundy, J., Ulmer, P., Pettke, T., 2007. Laser Ablation ICPMS study of trace

element partitioning between plagioclase and basaltic melts: an experimental approach.


Ballard, J.R., Palin, M.J., Campbell, I.H., 2002. Relative oxidation states of magmas inferred

from Ce(IV)/Ce(III) in zircon: application to porphyry copper deposits of northern Chile.


Berry, A.J., Shelley, J.M.G., Foran, G.J., O'Neill, H.S., Scott, D.R., 2003. A furnace design for

XANES spectroscopy of silicate melts under controlled oxygen fugacities and temperatures to

1773 K. Journal of Synchrotron Radiation 10, 332-336.

Blundy, J.D., Wood, B.J., 1994. Prediction of crystal-melt partition coefficients from elastic


Borisov, A.A., Lahaye, Y., Palme, H., 2006. The effect of sodium on the solubilities of metals in

silicate melts. American Mineralogist 91, 762-771.

68

Borisov, A.A., Nachtweyh, K., 1998. Ru solubility in silicate melts: experimental results in

oxidizing region. Lunar and Planetary Science Absrtacts XXIX, 1320.

Borisov, A.A., Palme, H., 1995. The solubility of iridium in silicate melts: New data from

experiments with Ir10Pt90 alloys. Geochimica et Cosmochimica Acta 59, 481-485.

Brenan, J.M., McDonough, W.F., 2009. Core formation and metal–silicate fractionation of

osmium and iridium from gold. Nature Geoscience 2, 798-801.

Burnham, A.D., Berry, A.J., 2012. An experimental study of trace element partitioning between

zircon and melt as a function of oxygen fugacity. Geochimica et Cosmochimica Acta 95, 196-

212.

Capehart, T.W., Mishra, R.K., Meisner, G.P., Fuerst, C.D., Herbst, J.F., 1993. Steric variation of

the cerium valence in Ce2Fe14B and related compounds. Applied Physics Letters 63, 3642-3644.

Co., D.C., 1971. JANAF Thermochemical Tables, U.S. National Bureau of Standards,

Washington D.C.

Cotten, J., Le Dez, A., Bau, M., Caroff, M., Maury, R.C., Dulski, P., Fourcade, S., Bohn, M.,

Brousse, R., 1995. Origin of anomalous rare-earth element and yttrium enrichments in

subaerially exposed basalts: evidence from French Polynesia. Chemical Geology 119, 115-138.

Darab, J.G., Li, H., Vienna, J.D., 1998. X-ray absorption spectroscopic investigation of the

environment of cerium in glasses based on complex cerium alkali borosilicate compositions.

Journal of Non-Crystalline Solids 226, 162-174.

Drake, M.J., 1975. The oxidation state of europium as an indicator of oxygen fugacity.


Duffy, J.A., Ingram, M.D., 1971. Establishment of an optical scale for lewis basicity in inorganic

oxyacids, molten salts, and glasses. Journal of the American Chemical Society 93, 6448-6454.

Dunn, T., 1982. Oxygen diffusion in three silicate melts along the join diopside-anorthite.


69

Fodor, R.V., Frey, F.A., Bauer, G.R., Clague, D.A., 1992. Ages, rare-earth element enrichment,

and petrogenesis of tholeiitic and alkalic basalts from Kahoolawe Island, Hawaii. Contributions

to Mineralogy and Petrology 110, 442-462.

Frey, F.A., Green, D.H., 1974. Mineralogy, geochemistry and origin of lherzolite inclusions in

Victorian basanites. Geochimica et Cosmochimica Acta 38, 1023-1059.

Hanchar, J.M., van Westrenen, W., 2007. Rare earth element behavior in zircon-melt systems.

Elements 3, 37-42.

Henderson, P., Nolan, J., Cunningham, G.C., Lowry, R.K., 1985. Structural controls and

mechanisms of diffusion in natural silicate melts. Contributions to Mineralogy and Petrology 89,

263-272.

Hoskin, P.W.O., Kinny, P.D., Wyborn, D., Chappell, B.W., 2000. Identifying accessory mineral

saturation during differentiation in granitoid magmas: an integrated approach. Journal of

Petrology 41, 1365-1396.

Hoskin, P.W.O., Schaltegger, U., 2003. The composition of zircon and igneous and metamorphic

petrogenesis, in: Hanchar, J.M., Hoskin, P.W.O. (Eds.), Zircon. Mineralogical Soc America,

Washington, pp. 27-62.

Jiang, S.-Y., Yu, J.-M., Lu, J.-J., 2004. Trace and rare-earth element geochemistry in tourmaline

and cassiterite from the Yunlong tin deposit, Yunnan, China: implication for migmatitic–

hydrothermal fluid evolution and ore genesis. Chemical Geology 209, 193-213.

Johnston, W.D., 1965. Oxidation-reduction equilibria in molten Na2O2SiO2 glass. Journal of the

American Ceramic Society 48, 184-&.

Kress, V.C., Carmichael, I.S.E., 1991. The compressibility of silicate liquids containing Fe2O3

and the effect of composition, temperature, oxygen fugacity and pressure on their redox states.


Luo, Y., Ayers, J.C., 2009. Experimental measurements of zircon/melt trace-element partition

coefficients. Geochimica et Cosmochimica Acta 73, 3656-3679.

70

Marsh, J.S., 1991. REE fractionation and Ce anomalies in weathered Karoo dolerite. Chemical

Geology 90, 189-194.

Mysen, B.O., Virgo, D., 1986. Volatiles in silicate melts at high-pressure and temperature. 1,

Interaction between OH groups and Si4+, Al3+, Ca2+, Na+ and H+. Chemical Geology 57, 303-

331.

Mysen, B.O., Virgo, D., Neumann, E.R., Seifert, F.A., 1985. Redox equilibrium and the

structural states of ferric and ferrous iron in melts in the system CaO-MgO-Al2O3-SiO2-Fe-O -

Relationships between redox equilibria, melt structure and liquidus phase-equilibria. American


Neal, C.R., Taylor, L.A., 1989. A negative Ce anomaly in a peridotite xenolith: evidence for

crustal recycling into the mantle or mantle metasomatism? Geochimica et Cosmochimica Acta

53, 1035-1040.

O'Neill, H.S.C., 2005. A method for controlling alkali-metal oxide activities in one-atmosphere

experiments and its application to measuring the relative activity coefficients of NaO0.5 in silicate

melts. American Mineralogist 90, 497-501.

O'Neill, H.S.C., Dingwell, D.B., Borisov, A.A., Spettel, B., Palme, H., 1995. Eperimental

petrochemistry of some highly siderophile elements at high temperatures, and some implications

for core formation and the mantle's early history. Chemical Geology 120, 255-273.

O'Neill, H.S.C., Nell, J., 1997. Gibbs free energies of formation of RuO2, lrO2, and OsO2: a high-

temperature electrochemical and calorimetric study. Geochimica et Cosmochimica Acta 61,

5279-5293.

Patra, A., De, G., Kundu, D., Ganguli, D., 2000. Preparation and characterization of Al and B co-

doped cerium containing sol-gel derived silica glasses. Materials Letters 42, 200-206.

Paul, A., 1976. Cerium - titanium yellow color in glass. Physics and Chemistry of Glasses 17, 7-

9.

Paul, A., Douglas, R.W., 1965. Cerous-ceric equilibrium in binary alkali borate and alkali

silicate glasses. Physics and Chemistry of Glasses 6, 212-215.

71

Sack, R.O., Carmichael, I.S.E., Rivers, M., Ghiorso, M.S., 1980. Ferric-ferrous equilibria in

natural silicate liquids at 1 bar. Contributions to Mineralogy and Petrology 75, 369-376.

Sano, Y., Terada, K., Fukuoka, T., 2002. High mass resolution ion microprobe analysis of rare

earth elements in silicate glass, apatite and zircon: lack of matrix dependency. Chemical Geology

184, 217-230.




Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomic

distances in halides and chaleogenides. Acta Crystallogr B 32, 751-767.

Shimizu, H., Sawatari, H., Kawata, Y., Dunkley, P.N., Masuda, A., 1992. Ce and Nd isotope

geochemistry on island arc volcanic rocks with negative Ce anomaly: existence of sources with

concave REE patterns in the mantle beneath the Solomon and Bonin island arcs. Contributions to

Mineralogy and Petrology 110, 242-252.

Shimizugawa, Y., Umesaki, N., Hanada, K., Sakai, I., Qui, J., 2001. X-ray induced reduction of

rare earth ion doped in Na2O-Al2O3-B2O3 glasses. Journal of Synchrotron Radiation 8, 797-799.

Trail, D., Watson, E.B., Tailby, N.D., 2011. The oxidation state of Hadean magmas and


Virgo, D., Mysen, B. O., Kushiro, I., 1980. Anionic constitution of 1-atmosphere silicate melts:

Implications for the structure of igneous melts. Science 20, 1371-1373.

Watson, E.B., Harrison, T.M., 1983. Zircon saturation revisited: temperature and composition

effects in a variety of crustal magma types. Earth and Planetary Science Letters 64, 295-304.

Wilke, M., Jugo, P.J., Klimm, K., Susini, J., Botcharnikov, R.E., Kohn, S.C., Janousch, M.,

2008. The origin of S4+ detected in silicate glasses by XANES. American Mineralogist 93, 235-

240.

72

Zinkevich, M., Djurovic, D., Aldinger, F., 2006. Thermodynamic modelling of the cerium–

oxygen system. Solid State Ionics 177, 989-1001.

73

Table 3.1 Bulk composition of stating materials in weight percent oxide.

BH09 AA08 RH08 SiO2 51.7 61.5 72.1TiO2 1.1 0.6 0.3 Al2O3 16.8 18.7 14.8 MgO 11.9 4.2 0.7 CaO 15.9 11.5 1.7 Na2O 2.6 2.8 6.4 K2O 0.7 3.9 P2O5 0.1All compositions contained 1.0-1.2 wt% CeO2

74

Table 3.2 Experiment run conditions.

Sample Composition Temperature (K)

log fO2a Pressure

(GPa) wt% H2O

b

DS11-C66 RH08 1672 -9.31 0.001 - DS11-C28 RH08 1772 -6.77 0.001 - DS09-C3-32 RH08 1672 -6.31 0.001 - DS09-C3-24 RH08 1672 -4.30 0.001 - DS09-C3-05 RH08 1571 -6.02 0.001 - DS11-C24 RH08 1772 -2.40 0.001 - DS09-C3-10 RH08 1571 -2.73 0.001 - DS11-C36 RH08 1773 0.00 0.001 - DS09-C3-19 RH08 1672 -0.02 0.001 - DS09-C2-13 RH08 1572 0.00 0.001 - DS11-C41c RH08 1478 -0.02 0.001 - DS11-C38c RH08 1476 -0.02 0.001 - DS11-C40c RH08 1474 -0.02 0.001 - DS11-C44c RH08 1473 -0.02 0.001 - DS09-C3-37 RH08 1573 -1.30 (Ru) 1 7.3 DS08-C1-02 RH08 1373 -2.72 (Ru) 1 9.1 DS11-C63 AA08 1672 -9.31 0.001 - DS09-C3-29 AA08 1674 -6.30 0.001 - DS09-C3-15 AA08 1574 -7.28 0.001 - DS09-C3-23 AA08 1674 -4.29 0.001 - DS09-C3-06 AA08 1574 -6.01 0.001 - DS11-C22 AA08 1772 -2.40 0.001 - DS09-C3-08 AA08 1571 -2.73 0.001 - DS11-C34 AA08 1773 0.00 0.001 - DS09-C3-18 AA08 1672 -0.02 0.001 - DS09-C2-09 AA08 1572 0.00 0.001 - DS11-C50 AA08 1574 -0.01 0.001 - DS09-C3-22 AA08 1673 1.54 (Ir) 1 6.0 DS08-C1-09 AA08 1573 1.08 (Ir) 1 6.1 DS09-C2-06 AA08 1573 1.08 (Ir) 1 7.2 DS08-C1-20 AA08 1373 -0.06 (Ir) 1 8.1 DS11-C64 BH09 1673 -9.31 0.001 - DS09-C3-28 BH09 1674 -6.30 0.001 - DS09-C3-14 BH09 1575 -7.28 0.001 - DS09-C3-13 BH09 1574 -7.28 0.001 - DS09-C3-20 BH09 1672 -4.30 0.001 - DS09-C3-03 BH09 1571 -5.32 0.001 - DS11-C21 BH09 1772 -2.40 0.001 -

75

DS09-C3-07 BH09 1572 -6.29 0.001 - DS11-C33 BH09 1773 0.00 0.001 - DS09-C3-17 BH09 1672 -0.04 0.001 - DS11-C49 BH09 1574 -0.02 0.001 - DS09-C2-12 BH09 1572 0.00 0.001 - DS09-C3-09 BH09 1573 -3.00 (Ru) 1 2.8 DS09-C3-04 BH09 1573 2.48 (Ir) 1 3.6 a Ru and Ir indicate the noble metal oxide buffer used in piston cylinder experiments.b Determined by difference from EMP analysisc Run as part of a time series. DS11-C38 = 2 hrs; DS11-C41 = 6 hrs; DS11-C44 = 12 hrs, DS11-C40 = 48 hrs.

76

Table 3.3 Electron microprobe analyses of glass run products (values in wt%). Values of NBO/T

calculated as anhydrous for all experiments. Errors given in brackets.

Sample DS08-C1-02 DS08-C1-09 DS08-C1-20 DS09-C2-06 DS09-C2-09 n 10 8 10 10 10 SiO2 65.44 (0.5) 57.36 (0.38) 67.42 (0.22) 56.11 (0.19) 59.70 (0.44) TiO2 0.31 (0.14) 0.54 (0.16) 0.26 (0.14) 0.64 (0.18) 0.55 (0.2) Al2O3 13.19 (0.2) 17.01 (0.13) 13.38 (0.13) 16.57 (0.14) 17.76 (0.23) MgO 0.64 (0.04) 3.72 (0.09) 0.63 (0.04) 3.56 (0.1) 3.88 (0.1) CaO 1.62 (0.03) 10.87 (0.11) 1.72 (0.04) 10.68 (0.12) 11.27 (0.13) Na2O 5.11 (0.16) 2.65 (0.18) 5.82 (0.23) 3.34 (0.18) 2.72 (0.12) K2O 3.40 (0.09) 0.66 (0.01) 3.43 (0.1) 0.65 (0.05) 0.67 (0.05) P2O5 0.11 (0.05) - 0.10 (0.12) - - Ce2O3 1.03 (0.11) 1.17 (0.09) 1.05 (0.15) 1.08 (0.07) 1.08 (0.12) Total 90.89 (0.81) 93.95 (0.53) 93.85 (0.47) 92.67 (0.54) 97.64 (0.83) NBO/T 0.065 (0.003) 0.276 (0.009) 0.080 (0.002) 0.295 (0.009) 0.273 (0.012) log Ce4+/Ce3+ -1.47 (0.11) -0.94 (0.11) -0.56 (0.11) -1.09 (0.11) -0.44 (0.11)

Sample DS09-C2-12 DS09-C2-13 DS09-C3-03 DS09-C3-04 DS09-C3-05 n 10 8 9 10 10 SiO2 51.88 (0.35) 72.34 (0.42) 51.52 (0.3) 48.94 (0.35) 71.42 (0.27) TiO2 1.01 (0.16) 0.37 (0.11) 0.95 (0.17) 1.06 (0.23) 0.22 (0.14) Al2O3 16.39 (0.16) 14.35 (0.19) 16.26 (0.17) 16.18 (0.23) 14.16 (0.14) MgO 11.46 (0.12) 0.66 (0.05) 11.40 (0.15) 10.95 (0.14) 0.68 (0.05) CaO 16.29 (0.12) 1.83 (0.07) 16.24 (0.15) 15.32 (0.1) 1.90 (0.06) Na2O 2.11 (0.13) 6.06 (0.11) 2.28 (0.14) 2.84 (0.12) 6.36 (0.21) K2O - 3.55 (0.1) - - 3.76 (0.09) P2O5 - 0.09 (0.05) - - 0.12 (0.06) Ce2O3 1.13 (0.12) 1.17 (0.10) 1.12 (0.1) 1.13 (0.04) 1.17 (0.13) Total 100.31 (0.58) 100.48 (0.74) 99.8 (0.62) 96.44 (0.6) 99.84 (0.27) NBO/T 0.766 (0.002) 0.074 (0.002) 0.774 (0.029) 0.772 (0.036) 0.090 (0.002) log Ce4+/Ce3+ -1.09 (0.07) -0.41 (0.01) -2.09 (0.11) -1.12 (0.10) -1.63 (0.13)

77

Sample DS09-C3-06 DS09-C3-07 DS09-C3-08 DS09-C3-09 DS09-C3-10 n 10 10 10 10 9 SiO2 60.84 (0.37) 51.65 (0.27) 61.01 (0.29) 50.16 (0.42) 71.44 (0.32) TiO2 0.58 (0.19) 1.05 (0.15) 0.69 (0.15) 0.99 (0.2) 0.25 (0.1) Al2O3 17.90 (0.14) 16.28 (0.14) 17.96 (0.16) 15.86 (0.16) 14.36 (0.13) MgO 4.09 (0.1) 11.53 (0.18) 4.02 (0.07) 11.10 (0.2) 0.7 (0.04) CaO 11.75 (0.19) 16.28 (0.12) 11.74 (0.15) 15.81 (0.22) 1.84 (0.04) Na2O 2.83 (0.19) 2.24 (0.15) 2.80 (0.09) 2.18 (0.13) 6.39 (0.2) K2O 0.67 (0.06) - 0.66 (0.03) - 3.65 (0.14) P2O5 - - - - 0.09 (0.04) Ce2O3 1.16 (0.1) 1.11 (0.1) 1.15 (0.1) 1.05 (0.05) 1.10 (0.09) Total 99.87 (0.68) 100.19 (0.61) 100.05 (0.48) 97.21 (0.59) 99.87 (0.44) NBO/T 0.290 (0.011) 0.777 (0.026) 0.284 (0.009) 0.771 (0.037) 0.084 (0.002)

log Ce4+/Ce3+ -2.01 (0.11) -1.46 (0.09) -1.26 (0.11) -1.60 (0.11) -0.99 (0.03)

Sample DS09-C3-13 DS09-C3-14 DS09-C3-15 DS09-C3-17 DS09-C3-18 n 10 10 10 10 10 SiO2 51.57 (0.37) 51.32 (0.34) 60.64 (0.27) 51.35 (0.33) 60.99 (0.34) TiO2 1.09 (0.19) 1.00 (0.11) 0.60 (0.09) 1.06 (0.13) 0.60 (0.13) Al2O3 16.60 (0.15) 16.48 (0.23) 18.32 (0.15) 16.52 (0.25) 18.08 (0.14) MgO 11.61 (0.18) 11.46 (0.13) 4.13 (0.05) 11.54 (0.08) 4.11 (0.06) CaO 16.14 (0.14) 16.15 (0.17) 11.55 (0.13) 16.09 (0.13) 11.74 (0.17) Na2O 2.17 (0.09) 2.23 (0.1) 2.86 (0.14) 2.30 (0.08) 2.88 (0.16) K2O - - 0.65 (0.04) - 0.65 (0.04) P2O5 - - - - - Ce2O3 1.10 (0.1) 1.12 (0.09) 1.20 (0.11) 1.12 (0.1) 1.21 (0.11) Total 100.30 (0.39) 99.81 (0.59) 99.96 (0.38) 100.04 (0.59) 100.29 (0.49) NBO/T 0.765 (0.031) 0.768 (0.031) 0.280 (0.007) 0.769 (0.031) 0.288 (0.009) log Ce4+/Ce3+ -2.83 (0.11) -2.98 (0.11) -1.93 (0.11) -1.04 (0.11) -0.87 (0.03)

78

Sample DS09-C3-19 DS09-C3-20 DS09-C3-22 DS09-C3-23 DS09-C3-24 n 10 10 10 10 10 SiO2 72.01 (0.42) 52.23 (0.37) 57.3 (0.31) 61.21 (0.36) 72.09 (0.37) TiO2 0.30 (0.1) 1.07 (0.16) 0.60 (0.08) 0.65 (0.14) 0.30 (0.1) Al2O3 14.29 (0.15) 16.53 (0.15) 16.87 (0.17) 18.17 (0.17) 14.38 (0.17) MgO 0.69 (0.04) 11.52 (0.09) 3.73 (0.08) 4.08 (0.09) 0.71 (0.03) CaO 1.89 (0.07) 16.35 (0.1) 10.95 (0.26) 11.72 (0.13) 1.91 (0.07) Na2O 6.04 (0.22) 1.94 (0.12) 2.91 (0.22) 2.79 (0.06) 6.02 (0.2) K2O 3.51 (0.09) - 0.60 (0.03) 0.67 (0.06) 3.64 (0.12) P2O5 0.08 (0.05) - - - 0.11 (0.14) Ce2O3 1.18 (0.08) 1.14 (0.15) 1.07 (0.13) 1.21 (0.13) 1.18 (0.1) Total 100.04 (0.56) 100.82 (0.55) 94.07 (0.54) 100.54 (0.6) 100.38 (0.39) NBO/T 0.077 (0.002) 0.757 (0.029) 0.287 (0.01) 0.282 (0.009) 0.078 (0.003) log Ce4+/Ce3+ -0.60 (0.03) -2.47 (0.11) -1.02 (0.11) -1.72 (0.11) -1.58 (0.13)

Sample DS09-C3-28 DS09-C3-29 DS09-C3-32 DS09-C3-37 DS11-C21 n 10 9 10 9 10SiO2 51.62 (0.26) 60.91 (0.25) 71.83 (0.27) 66.63 (0.97) 51.57 (0.26) TiO2 1.08 (0.16) 0.60 (0.11) 0.28 (0.14) 0.34 (0.1) 1.02 (0.11) Al2O3 16.25 (0.16) 17.89 (0.12) 14.27 (0.12) 13.18 (0.3) 16.28 (0.24) MgO 11.60 (0.13) 4.10 (0.1) 0.72 (0.04) 0.58 (0.1) 11.40 (0.14)CaO 16.32 (0.22) 11.69 (0.13) 1.92 (0.08) 1.61 (0.26) 16.23 (0.14) Na2O 2.11 (0.15) 2.74 (0.17) 6.43 (0.26) 5.57 (0.22) 2.12 (0.08) K2O - 0.64 (0.03) 3.53 (0.08) 3.56 (0.11) - P2O5 - - 0.06 (0.07) - - Ce2O3 1.14 (0.05) 1.13 (0.11) 1.17 (0.09) 1.01 (0.17) 1.15 (0.11) Total 100.17 (0.34) 99.74 (0.4) 100.26 (0.57) 92.62 (0.53) 99.80 (0.61) NBO/T 0.779 (0.027) 0.286 (0.009) 0.087 (0.002) 0.075 (0.005) 0.768 (0.028) log Ce4+/Ce3+ -2.80 (0.11) -2.33 (0.11) -2.09 (0.13) -1.62 (0.11) -1.96 (0.32)

79

Sample DS11-C22 DS11-C24 DS11-C28 DS11-C33 DS11-C34 n 10 9 10 9 10 SiO2 51.58 (0.24) 71.23 (0.39) 72.16 (0.44) 51.91 (0.45) 60.96 (0.34) TiO2 1.20 (0.2) 0.33 (0.11) 0.29 (0.11) 1.04 (0.22) 0.68 (0.13) Al2O3 16.27 (0.12) 14.59 (0.16) 14.55 (0.15) 16.50 (0.13) 18.01 (0.2) MgO 11.35 (0.12) 0.65 (0.02) 0.66 (0.03) 11.32 (0.19) 3.93 (0.09) CaO 16.22 (0.22) 1.79 (0.06) 1.75 (0.11) 16.18 (0.16) 11.77 (0.22) Na2O 2.14 (0.15) 5.93 (0.27) 5.37 (0.2) 2.15 (0.16) 2.83 (0.13) K2O - 3.64 (0.1) 3.53 (0.13) - 0.72 (0.05) P2O5 - 0.13 (0.11) 0.07 (0.09) - - Ce2O3 1.09 (0.1) 1.02 (0.1) 0.98 (0.11) 1.09 (0.12) 0.96 (0.1) Total 99.88 (0.41) 99.37 (0.31) 99.41 (0.47) 100.23 (0.77) 99.89 (0.46) NBO/T 0.764 (0.026) 0.067 (0.002) 0.052 (0.002) 0.753 (0.035) 0.280 (0.01) log Ce4+/Ce3+ -1.68 (0.11) -1.58 (0.13) -1.83 (0.13) -1.28 (0.08) -1.03 (0.09)

Sample DS11-C36 DS11-C38 DS11-C40 DS11-C41 DS11-C44 n 10 10 10 9 10 SiO2 72.11 (0.56) 71.05 (1.06) 71.23 (0.27) 71.15 (1.09) 71.60 (0.95) TiO2 0.34 (0.12) 0.32 (0.13) 0.25 (0.15) 0.18 (0.13) 0.25 (0.12) Al2O3 14.52 (0.16) 14.62 (0.67) 14.37 (0.16) 14.56 (0.72) 14.30 (0.67) MgO 0.65 (0.05) 0.66 (0.04) 0.64 (0.04) 0.64 (0.03) 0.63 (0.06) CaO 1.82 (0.06) 1.79 (0.08) 1.83 (0.06) 1.81 (0.08) 1.74 (0.08) Na2O 6.16 (0.27) 6.57 (0.28) 6.49 (0.15) 6.36 (0.35) 6.45 (0.39) K2O 3.59 (0.17) 3.79 (0.09) 3.76 (0.09) 3.71 (0.13) 3.73 (0.08) P2O5 0.02 (0.04) 0.14 (0.14) 0.09 (0.12) 0.11 (0.07) 0.14 (0.13) Ce2O3 0.98 (0.09) 0.92 (0.08) 0.94 (0.12) 0.97 (0.07) 0.94 (0.1) Total 100.20 (0.8) 99.90 (0.73) 99.63 (0.51) 99.55 (0.38) 99.82 (0.62) NBO/T 0.070 (0.003) 0.082 (0.007) 0.083 (0.002) 0.078 (0.007) 0.081 (0.007) log Ce4+/Ce3+ -0.82 (0.03) -0.15 (0.13) -0.05 (0.13) -0.19 (0.13) -0.15 (0.13)

80

Sample DS11-C49 DS11-C50 DS11-C63 DS11-C64 DS11-C66 n 10 10 10 10 10 SiO2 51.86 (0.48) 61.53 (0.28) 61.36 (0.33) 51.99 (0.19) 71.88 (0.71) TiO2 1.17 (0.15) 0.58 (0.14) 0.56 (0.15) 0.92 (0.19) 0.3 (0.1) Al2O3 16.38 (0.18) 18.42 (0.17) 18.09 (0.11) 16.4 (0.13) 14.76 (0.3) MgO 11.28 (0.14) 3.96 (0.1) 3.94 (0.11) 11.61 (0.16) 0.73 (0.07) CaO 16.30 (0.15) 11.59 (0.12) 11.7 (0.11) 16.32 (0.16) 1.99 (0.26) Na2O 2.14 (0.07) 2.66 (0.15) 2.42 (0.15) 1.6 (0.15) 5.62 (0.2) K2O - 0.7 (0.05) 0.72 (0.05) - 3.56 (0.08) P2O5 - - - - 0.07 (0.08) Ce2O3 1.12 (0.19) 0.95 (0.09) 0.96 (0.12) 1.14 (0.09) 1.03 (0.15) Total 100.29 (0.47) 100.43 (0.5) 99.79 (0.49) 100.01 (0.34) 99.97 (0.45) NBO/T 0.758 (0.035) 0.264 (0.008) 0.267 (0.008) 0.758 (0.023) 0.063 (0.003) log Ce4+/Ce3+ -0.93 (0.02) -0.57 (0.02) -3.04 (0.11) -3.57 (0.11) -2.57 (0.13)

81

Figure 3.1 Experimental assemblies for a) low fO2 experiments at 1 atm employing an alkali-

reservoir melt, and b) piston cylinder experiments done at 1 GPa.

82

Figure 3.2 Ce M4,5-edge XANES spectra of Ce3+ (CeF3) and Ce4+ (CeO2) standards and a mixed

valence run product (DS09-C2-13).

83

Figure 3.3 Arrhenius plot of the variation in the Ce redox equilibrium versus temperature.

Shaded symbols were done at log fO2 = 0 (pure O2) and open symbols run at log fO2 ~-2.4.

Squares, diamonds and triangles correspond to compositions RH08, AA08 and BH09,

respectively. All samples measured potentiometrically. Lines are fit to the data by linear

regression.

84

Figure 3.4 Plot of Ce4+/Ce3+ as function of fO2, expressed in log units relative to the fayalite-

magnetite-quartz oxygen buffer (FMQ). Lines through the data have slope of 1/4 in log-log

space (as per Equation 3.3) with the intercepts determined by linear regression. Ce4+/Ce3+

determined potentiometrically (open symbols) and by XANES (shaded symbols). Symbol

shapes as in Figure 3.3.

85

Figure 3.5 Values for the compositional term ‘B’ ( i ii

b X from Equation 3.9) versus the dry

NBO/T values for the different compositions investigated in this study. Symbol shapes as in

Figure 3.3. Shaded symbols = anhydrous compositions, open = hydrous experiments. Dashed

lines show the displacement in Ce redox equilibria with the mole percent of H2O calculated for

the experiment (numbers beneath lines).

86

Figure 3.6 Measured versus calculated log(Ce4+/Ce3+) comparing data determined in this study

with previous measurements on silicate glasses. Symbols are as follows: X’s, this study; light

and dark grey diamonds are FAS and FAD compositions, respectively (Schreiber et al., 1980);

light, medium and dark circles are Li, Na, and K silicate glasses, respectively (Paul and Douglas,

1965); light and dark triangles are Pu10S-I and Pu10S-II compositions, respectively (Darab et

al., 1998); and squares are Na-silicate glasses (Johnston, 1965). Dashed arrows in inset are in the

direction of increasing alkali content for compositions from Paul and Douglas (1965). Error bars

omitted for clarity. Errors on calculated log(Ce4+/Ce3+) are approximately 0.5 log unit.

87

Figure 3.7 log (CeO2/CeO3/2) versus log fO2 determined by zircon-melt partitioning experiments

from Burnham and Berry (2012, grey triangles). Grey line shows best fit to the data, solid black

line shows the predicted values for log (CeO2/CeO3/2) from the model presented in this study for

the melt composition and run conditions of the partitioning experiments. Results are in excellent

agreement, plotting within error in many cases.

88

Figure 3.8 Alkali content (A = Li, Na or K) (a) and Al/Si (b) of the glass versus the

compositional term B. Open symbols are average values for the compositions investigated in this

study (shapes as in Figure 3.3); shaded circles and diamonds as in Figure 3.6.

89

Figure 3.9 Chondrite normalized REE diagrams showing the predicted magnitude of Ce and Eu

anomalies in zircon crystalizing from a hydrous granitic melt with a) no plagioclase

crystallization, and b) 20 percent plagioclase fractionation. Zircon/melt partition coefficients for

3+ REEs from Sano et al. (2002) and Ce4+ from Ballard et al. (2002). Plagioclase/melt partition

coefficients from Aigner-Torres (2007). Partitioning of Eu2+ into zircon was considered to be

negligible relative to Eu3+.

90

Chapter 4

Experimental Calibration of a Ce-in-zircon Oxygen Barometer

Using a newly-calibrated relation for cerium redox equilibrium in silicate melts, and an

internally-consistent model for zircon-melt partitioning of Ce, we provide a method to estimate

the prevailing redox conditions during crystallization of zircon-saturated magmas. With this

approach, oxygen fugacities are calculated for samples from the Bishop tuff (USA), Toba tuff

(Indonesia) and the Nain plutonic suite (Canada), which are in excellent agreement (typically

within 0.5 log units or better) with independent estimates. With the success of reproducing the

fO2 of well-constrained igneous systems, we have applied our Ce-in-zircon oxygen barometer to

estimating the redox state of Earth’s earliest magmas. Using the composition of the Jack Hills

Hadean zircons, combined with estimates of their parental magma composition, we determine

the fO2 during zircon crystallization to be between FMQ -1.0 to +2.5 (where FMQ is the fayalite-

magnetite-quartz buffer). Of the parental magmas considered, Archean tonalite-trondhjemite-

granodiorite (TTG) compositions yield zircon-melt partitioning most similar to well-constrained

modern suites (e.g., Sano et al., 2002). Although broadly consistent with previous redox

estimates from the Jack Hills zircons, our results provide a more precise determination of fO2,

narrowing the range for Hadean parental magmas by more than 8 orders of magnitude. Results

suggest that relatively oxidized magmatic source regions, similar in oxidation state to that of 3.5

Ga komatiite suites, existed by ~4.4 Ga.

4.1 Introduction

The mineral zircon incorporates a variety of trace elements during crystallization, including in

part the rare-earth elements (REE), uranium, thorium, and titanium, which makes this phase

particularly useful for tracing magma chemistry as well as for geochronometry, and

geothermometry (Hoskin and Schaltegger, 2003; Watson and Harrison, 2005). A nearly

ubiquitous, but until recently underutilized, feature of terrestrial zircon (though generally lacking

in lunar and meteoritic samples) is the anomalously high chondrite-normalized concentrations of

Ce relative to neighboring REEs. In most terrestrial magmas, Ce will exist in both 3+ and 4+

oxidation states (Chapter 3), so anomalous Ce concentrations in zircon result from the favoured

91

partitioning of Ce4+ relative to Ce3+ into the zircon structure, since Ce4+ has the same charge and

is a close match in ionic radius to Zr4+ (Shannon, 1976). The change in Ce oxidation state in

silicate melts can be expressed via the homogeneous reaction,

3/2 2 2

1CeO O CeO


This implies that the relative proportions of Ce3+ (CeO3/2) and Ce4+ (CeO2) in a silicate melt will

be a function of the oxygen fugacity (fO2) of the system, and the magnitude of the resulting Ce

anomaly in zircon will then be a record of the prevailing redox conditions during crystallization.

In principle, it might be possible to measure the Ce4+/Ce3+ in zircon directly as an indicator of

magma oxidation state. However, alpha recoil events during radioactive decay of U and Th

induce structural changes in zircon (Trachenko et al., 2002) which allow for post-crystallization

changes in Ce4+/Ce3+, making this approach unreliable. Instead, because the sluggish diffusion

rates of 3+ and 4+ cations in zircon (Cherniak et al., 1997a; Cherniak et al., 1997b) prevent

modification of absolute Ce concentrations, Ce anomalies produced during crystal growth

potentially constitute a robust record of magma redox state.

Recently, Trail et al. (2012; 2011) provided an experimental calibration which relates the

magnitude of the Ce anomaly in zircon, termed (Ce/Ce*)D (in which Ce* is the estimated value

for 3

/

Ce

zircon meltD ), to fO2 and temperature. Values of (Ce/Ce*)D can be calculated from natural

samples from knowledge of the bulk partition coefficient for Ce, combined with the estimated

value for 3

/

Ce

zircon meltD as interpolated from partition coefficients for the neighboring light REEs

(LREEs), La and Pr. This approach is a valuable first step in exploiting the Ce anomaly as a

redox sensor, but seems to yield relatively imprecise estimates for magma fO2 (possible reasons

for this are discussed later). As an alternative, we have taken the approach of Ballard et al.

(2002), in which the bulk partition coefficient for Ce, /Cezircon meltD , along with values of 3

/

Ce

zircon meltD

and 4

/

Ce

zircon meltD are related to the mole fractions of Ce4+ ( 4Ce

meltx ) and Ce3+ ( 3Ce

meltx ) in the melt by the

equality:

4 3

3 4

/

Ce Ce/

Ce Ce

Ce Ce

Ce Ce

melt zircon meltzircon melt

melt zircon meltmelt zircon

x D

x D

(4.2)

92

where Ce is the total concentrations of Ce in zircon or melt (see Appendix 4 for a full

derivation). Using results from our previous work on the redox behavior of Ce in silicate melts

(Chapter 3), the relative proportions of Ce4+ and Ce3+ can be related to melt fO2 through the

equation:

4

3

Ce2

Ce

2

1 13136( 591) NBOln ln O 2.064( 0.011)

4 T

8.878( 0.112) H O 8.955( 0.091)

melt

melt

xf

x T

x

(4.3)

where T is in Kelvin and can be determined using the Ti-in-zircon thermometer (Ferry and

Watson, 2007), NBO/T is the proportion of non-bridging oxygens to tetrahedrally coordinated

cations (Virgo et al., 1980) calculated on an anhydrous basis, and xH2O is the mole fraction of

water dissolved in the melt. Zircon/melt partition coefficients for Ce3+ 3

/

Ce( )zircon meltD and Ce4+

4

/

Ce( )zircon meltD can be calculated for individual zircon-melt pairs using the lattice strain model of

Blundy and Wood (1994), and are constrained by the partition coefficients for the other REEs as

well as Hf, Th, and U (as described below). In this communication, we demonstrate the accuracy

of this new oxygen barometer through application to three zircon-saturated igneous suites whose

redox state has been determined by independent methods. We then show how the method can be

used to estimate the fO2 during crystallization of magmas formed within 500 million years of

terrestrial accretion, as represented by the Hadean-aged Jack Hills zircons.

4.2 Samples Investigated

Samples investigated in this study were taken from three different localities for which fO2 has

been previously estimated, namely: the Bishop tuff (BT), California, Toba tuff (TT), Indonesia,

and Umiakovik pluton (UP), Nain plutonic suit, Labrador. A full summary of zircon textures, fO2

and T estimates are provided in Appendix 1. Here we combine the fO2-T estimates to express the

redox state of samples relative to the fayalite-magnetite-quartz (FMQ) reference buffer, in terms

of log fO2 units, where FMQ = log fO2 (sample) – log fO2 (FMQ), both calculated at the same

temperature. Eruption of BT occurred at 760 ka and the unit consists of several packages of fall

deposits and ignimbrite. Samples of BT were obtained from the Ig2E unit as defined by Wilson

93

and Hildreth (1997), kindly supplied by John Hanchar, and consist of high-silica rhyolite

pumices containing <5% crystal fragments, primarily of quartz, in a glassy matrix. Fourier

transform infrared (FTIR) spectroscopy of melt inclusions in quartz have determined the H2O

content of the pre-eruptive magma to be between 5.1 and 6.8 wt% (Anderson et al., 1989;

Wallace et al., 1999). Oxygen fugacity and temperature for BT samples, calculated from Fe-Ti-

oxide thermobarometry, ranges from FMQ of +0.5 to +2.0 and 700 to 840C (Hildreth and

Wilson, 2007). The TT sample, acquired with the aid of Nurcahyo Basuki, is a potassic high-

silica rhyolitic welded tuff which erupted at 74 ka, belonging to the youngest unit in the Toba

volcanic package (YTT as defined by Chesner, 1998). The sample is primarily composed of

~25% crystal fragments (consisting of quartz, plagioclase, sanidine, and biotite) in a partially

devitrified glass matrix. Oxygen fugacity and temperature for TT samples, also calculated from

Fe-Ti-oxide thermobarometry, yield FMQ of 0 to +1.5 and 680 to 850C (Chesner, 1998).

Melt inclusions analyzed by FTIR spectroscopy constrains the pre-eruptive H2O content of the

TT to between 4.0 and 5.5 wt%. The sample from UP (Geological Survey of Canada sample

number EC87-86) is a pyroxene-bearing quartz monzodiorite which intruded Archean and

Proterozoic gneiss between 1311-1320 Ma (Emslie and Loveridge, 1992). The fO2 and

temperature of crystallization for the UP, estimated by the intersection of isopleths defined by

the hematite content of ilmenite and the ferric-ferrous ratio of biotite, ranges from FMQ of -1 to

-4 and 710 to 840C. The water content of the UP magma is somewhat less straightforward to

estimate, as primary melt inclusions have not been documented. Instead, the fH2O has been

calculated, from biotite-melt phase relations, to be ~250 bars (Emslie and Stirling, 1993),

equivalent to ~4.5 wt% H2O. Zircons obtained from BT are clear, and generally have a prismatic

crystal habit, ranging in size from ~50 to 200microns. Inclusions of apatite were found in nearly

all of the BT zircons. Back-scattered electron (BSE) and cathodoluminescence (CL) images

show relatively simple chemical zonation with sector zoning superimposed on oscillatory growth

zoning (Figure 4.1). Approximately half of the zircons from BT contain cores with a relatively

high CL response. Zircons from TT are clear, highly elongate prisms up to 300 microns in length

with abundant apatite and melt inclusions (Figure 4.2). From BSE and CL imaging, the TT

zircons appear to have a relatively complex growth history, showing in some cases multiple

episodes of growth and resorption. The outer domains in all TT zircons have oscillatory zoning

occasionally with faint sector zoning. UP zircons are typically 100 to 500 microns in size with

tabular to prismatic morphologies (Figure 4.3). Grains vary from colourless to pale brown and

94

generally lack mineral inclusions. Imaging shows broad and oscillatory zoning in the outer

domains of most samples, cored by regions with convolute zoning or lacking any internal

structure suggesting recrystallization. Chemical analyses of all zircons were restricted to pristine

regions with oscillatory growth zoning. Central domains interpreted to be regions of early growth

were avoided.

4.3 Analytical Methods

The whole-rock analysis of the UP sample was done on a powdered aliquot obtained from the

Geological Survey of Canada using solution ICP-MS at Activation Laboratories in Ancaster ON.

Accuracy and precision of trace element analysis are within 10 and 4% of the measured values,

respectively, with detection limits of 0.01 ppm or better. Major and trace element analysis of

zircons and natural glass were done in the Department of Earth Sciences, University of Toronto.

Major element compositions were determined using a Cameca SX50 electron microprobe.

Analyses of glasses were done at 15 kV using a 20 m defocused beam and 5 nA beam current.

Alkali elements were measured at the beginning of the acquisition, so as to minimize their

migration under the electron beam. Standards employed were natural basaltic glass (Mg, Ca, Mn,

Fe), obsidian (Si, Al), albite glass (Na), sanidine (K), TiO2 (Ti), and CePO4 (P, Ce). Major

element analysis of zircon was done using a 20 kV accelerating voltage, 50 nA beam current and

1 m beam size. Standards used include synthetic zircon (Zr, Si), hafnon (Hf), and YPO4 (Y, P).

Raw count rates were converted to element concentrations using the ZAF correction routine.

The trace element content of the glasses from BT and TT, and zircons from all samples, were

determined by LA-ICP-MS. The system at University of Toronto employs a Thermo Elemental

(VG) PlasmaQuad PQ ExCell ICP-MS coupled to a Nu-Wave UP-213 laser and ablation cell.

NIST 610 glass was used to quantify trace element concentrations using the following isotopes: 31P, 44Ca, 48Ti, 89Y, 139La, 140Ce, 141Pr, 146Nd, 147Sm, 153Gd, 159Tb, 163Dy, 165Ho, 166Er, 169Tm, 172Yb, 175Lu, 178Hf, 232Th, and 238U. 44Ca and 180Hf were used to monitor ablation yields in glass

and zircon, respectively (178Hf was also measured in zircon to check for potential interferences

from REE-oxides). On-peak dwell times of 10 ms were used for all elements, except La and Pr

in zircon, in which counts were measured for 30 ms, due to low concentrations. Glass and zircon

analyses were done using 50 and 20 micron spots, respectively, with a laser repetition rate of 10

Hz. Beam fluence was adjusted to optimise aerosol production. During zircon analysis, the

95

signals from 27Al and 44Ca were also monitored to screen for contamination by other phases.

Major and trace element analyses are compiled in Appendix 4.

Measurements of glass and whole rock compositions are in close agreement with previous

determinations, with the exception of glass from TT samples, which on average contained ~6.7

wt% potassium, considerably higher than values of ~5.2 wt% quoted for the same unit in earlier

studies (Chesner, 1998; Chesner and Luhr, 2010). As the TT glasses analyzed here are partially

devitrified, and previous analysis are of pristine glass and melt inclusions this discrepancy is

likely due to alteration of our samples. For this reason we choose to use the major and trace

element values for quartz- hosted melt inclusion from Chesner and Luhr (2010) as it seems likely

that this more closely represents the melt from which zircon crystallized.

4.4 Lattice Strain Constraints on Ce Partitioning

Using the estimations of melt composition and the measured trace element concentrations in

zircon, values of /Cezircon meltD can be calculated. Estimates for 3

/

Ce

zircon meltD and 4

/

Ce

zircon meltD can be made

after the method of Ballard et al (2002). In this approach, partition coefficients for the trivalent

REEs and the quadrivalent series Hf, Th, and U are used to constrain 3

/

Ce

zircon meltD , 4

/

Ce

zircon meltD ,

respectively. Blundy and Wood (1994) showed that the mineral melt partition coefficient for a

cation i can be related to the lattice strain energy created by substituting a cation whose ionic

radius (ri) differs from the optimal value for that site (r0). They provide the expression,

2 30

0 0

0

14

2 3exp

A i i

i

rEN r r r r

D DRT

(4.4)

where D0 is the ‘strain compensated partition coefficient’, E is the Young’s Modulus, NA is

Avogadro’s number, R is the gas constant, and T is in degrees K. Rearranging Equation 4.4 and

taking the logarithm of both sides, yields the relation:

200 0

4ln ln

3 6iA

i i

r rEND D r r

RT

(4.5)

96

Plotting ln Di against the term (ri/3 + ro/6)(ri-ro)2 therefore yields a linear relation for an isovalent

series of cations. With knowledge of the ionic radii for Ce3+ and Ce4+, partition coefficients for

these species can be determined by interpolation. Since Ce will be a mixture of Ce3+ and Ce4+,

the value of /Cezircon meltD will lie between these two partition coefficient end-members, and through

combination of Equations 4.2 and 4.3 be used to determine the fO2 during crystallization. A

graphical example of this procedure for estimating the fO2 of the TT samples is portrayed in

Figure 4.4.

Ionic radii for the REEs and Hf are taken from the compilation of Shannon (1976), whereas

values for Th and U are from David and Vokhmin (2003). In practice, we excluded the LREEs,

La and Pr, in the fitting procedure for 3+ cations as these elements are present at very low levels

in natural zircons, and concentrations are highly susceptible to contamination by accessory

mineral inclusions. Eu was also excluded, as it can exist as both Eu2+ and Eu3+ species, the

former being sensitive to plagioclase fractionation. As uranium may exist in multiple oxidation

states (Schreiber, 1982), and Zr partitioning into zircon is not a trace element substitution (and

therefore subject to non-Henryan partitioning), the 4+ trend can only be modeled with

confidence using Hf and Th. Values of r0 used for 3+ and 4+ cations are 0.93 Å (determined by

regression) and 0.83 Å (8-fold coordinated Hf), respectively.

4.5 Comparison to Independent Estimates of Magma fO2 and Temperature

Oxygen fugacity and temperature calculated for the BT, TT and UP suites are portrayed in

Figure 4.5. Temperatures were calculated using Ti content of zircon, using the expression (Ferry

and Watson, 2007):

2 2

4800 86log(Ti ) (5.711 0.072) log SiO log TiO

zircon a a

T (4.6)

Where Tizircon is the concentration of Ti in zircon in ppm, T is degrees K, and ai is the ratio of the

concentration of component i in the melt over the concentration of component i in the melt at

saturation. For the samples under investigation, aSiO2 can be assumed to be unity, as all systems

reached quartz saturation, and aTiO2 was calculated using the TiO2 solubility model of Hayden

and Watson (2007).

97

As is typical for volcanic suites undergoing prolonged cooling and eruption, the data for the BT

and TT suites describe linear arrays in fO2-T space, roughly parallel to the FMQ buffer

(Carmichael, 1991). Significantly, the arrays for the BT and TT suites defined by

thermobarometry involving either zircon or the Fe-Ti oxides overlap, reflecting excellent

agreement between the two methods. The higher degree of scatter for the arrays defined by

zircon thermobarometry may not be too surprising, given possible variations in melt chemistry

and temperature which may occur during protracted zircon growth in the crust (e.g., Claiborne et

al., 2010). Indeed, there is evidence for compositional zoning in the pre-eruptive magmas of

both the BT and TT systems (Chesner, 1998; Hildreth and Wilson, 2007). In our calculations,

we assume that the host melt is the composition from which the zircon formed, whereas this only

strictly applies to the outermost zircon rim, if at all. As for the UP, calculated zircon

crystallization temperatures are generally higher than those determined from biotite, indicating

that zircon likely formed first, and records an earlier stage in the fO2-T evolution of the UP.

There is, however, overlap in the fO2-T variation determined by the different methods, and

importantly, the generally reduced nature of the UP is well-captured by the zircon redox sensor.

4.6 Estimation of the Redox State of Hadean Magmas

Detrital zircon grains occurring in ca. 3 Ga conglomerates from Jack Hills in the Narryer Gneiss

Complex, Western Australia, have U-Pb ages ranging from 3.0 Ga to as old as 4.404 ± 0.008 Ga

(Wilde et al., 2001), the oldest of which crystallized within 150 Ma of terrestrial accretion. The

REE patterns of the Hadean grains have been shown to fall into one of two catagories, a LREE

depleted group (type-1), taken to represent pristine magmatic zircon, and a LREE enriched group

(type-2), interpreted to be hydrothermally altered (Hoskin, 2005). Application of the model

presented here to estimate the fO2 of the magma parental to the type-1 Hadean zircons from Jack

Hills is a challenge, as care must be made in the assumptions regarding the major and trace

element chemistry, as well as water content, of the Hadean zircon-forming melt. As outlined by

Watson and Harrison (1983), conditions that will promote zircon crystallization in magmatic

systems are a general increase in melt SiO2 content, as well as a decrease in temperature.

Consistent with this notion are constraints from both oxygen isotopic data (Wilde et al., 2001),

mineral inclusions (Maas et al., 1992), and Ti thermometry (Watson and Harrison, 2005)

suggesting that the Hadean Jack Hills zircons crystallized from relatively low temperature,

hydrous, felsic magmas.

98

What is more difficult to constrain is the relative trace element abundances of the parent melt

from which the Jack Hills zircons crystallized. One approach is to simply use published

experimental or empirical values of zircon-melt partitioning. However, results of experimental

investigation of zircon-melt partitioning has shown that the uptake of REE with partition

coefficients which differ significantly from unity can be over- or under-estimated (depending if

/zircon meltiD <1 or >1, respectively) due to slow diffusion in the melt phase (Luo and Ayers, 2009).

Studies on natural systems have yielded a wide range in zircon/melt partition coefficients

(Fujimaki, 1986; Hinton and Upton, 1991; Nagasawa, 1970; Sano et al., 2002; Thomas et al.,

2002), with values for the REEs differing by up to four orders of magnitude. Although it is

possible that some of this variation is the result of contamination by REE-rich phases during

zircon analysis (particularly the LREEs), other factors, such as the availability of charge

compensating ions (e.g., phosphorous), temperature, or composition of the crystallization

environment, are also likely to play an important role. Despite this shortcoming, values for

/zircon meltiD determined by Sano et al. (2002) seem to provide an accurate estimate of zircon-hosted

melt compositions (Hanchar and van Westrenen, 2007; Luo and Ayers, 2009) for the 3+ REEs.

In this case, we use the partition coefficients measured by Sano et al (2002) as a basis for

comparison. We have calculated values for /zircon meltiD by assuming that zircons have equilibrated

with three different types of parental magma compositions: pyrolite (i.e., equivalent to unmelted

upper mantle; McDonough and Sun, 1995), bulk continental crust (Rudnick and Gao, 2003), and

Archean tonalite-trondhjemite-granodiorite (TTG, Martin, 1995). These values, combined with

measured concentrations in Hadean zircons, have been used to estimate magma fO2. In this case,

we have restricted our zircon dataset to spot analyses of oscillatory zone domains with magmatic

REE patterns, for which U-Pb age, oxygen isotopic composition, and crystallization temperature

have been determined, totaling 23 analyses (Cavosie et al., 2006; Fu et al., 2008). Comparison

of partition coefficients calculated from these different melt reservoirs, to values measured by

Sano et al. (2002), suggest a close similarity of values using the Archean TTG as the parent melt

of the Jack Hills zircons (Figure A4.5). This allows for calculation of /zircon meltiD values for 3+

and 4+ cations, and from the method outlined above determine 4 3Ce Ce/melt meltx x . This is not to say that

the Hadean melts from which these zircons crystallized are necessarily the same as Archean

TTG, only that the relative trace element abundances of interest here are similar.

99

By estimating the anhydrous NBO/T values for felsic magmas (e.g., broadly dacite-rhyolite) to

range between 0.01 and 0.20, and assuming water contents of typical modern felsic melts (2.5 to

6.5 wt%) we approximate the likely range for the compositional terms in Equation 4.6 (shown by

the horizontal bar in Figure 4.6). Within one standard deviation of the calculated values of

4 3Ce Ce/melt meltx x (vertical bar in Figure 4.6), we estimate the possible range of fO2 for the Jack Hills

zircon to lie between FMQ -1.0 to +2.5. As shown in Figure 4.6, estimates of fO2 are quite

sensitive to changes in H2O and NBO/T. If a completely anhydrous composition was assumed,

the resulting fO2 would be 2.5 log units lower than the same melt with ~5 wt% H2O. On the other

hand if a basaltic composition was used in place of a granitic composition the resulting change in

the NBO/T term would increase the calculated fO2 by ~3 log units. As mentioned earlier, melts

of this nature do not typically saturate in zircon, nor appear to pertain to the parent melts of the

Jack Hills zircons, therefore, these are not included in the estimated compositional range.

These results, although broadly consistent with previous redox estimates from these zircons

(Trail et al., 2011), are on average slightly more oxidizing. Furthermore, Trail el al. (2011)

observe a range in fO2 spanning 12 log units, whereas using an identical data set, the range

calculated in this study is only ~2 log units, assuming a uniform melt composition. The lower

precision results from this earlier study may be the outcome of inaccuracies involved in the

estimation of (Ce/Ce*)D, since this parameter is sensitive to the concentration of LREE in zircon,

and hence prone to underestimation if a LREE-bearing contaminant contributes to the zircon

analysis. Moreover, it is possible that the experiments used to calibrate (Ce/Ce*)D, were subject

to the same disequilibrium effects as documented by Luo and Ayers (2009), as Dzircon for the

LREE are similar between the two studies, and larger than the measurements of Sano et al.

(2002).

The range in fO2 calculated here for Hadean zircons is strikingly similar to that of Archean

komatiite as determined by vanadium partitioning into olivine (Canil, 1997). Although the

currently favored model for the formation of at least the majority of these grains is through

subduction and re-melting of hydrated mafic crust, Trail et al. (2011) note that a small subset of

the Hadean zircons have oxygen isotopic compositions within the mantle range. We observe no

correlation between fO2 and the oxygen isotopic composition, suggesting that interaction with

the surface environment did not alter the redox state of early basaltic crust. The similarity

between the crustal and mantle derived zircons is not necessarily surprising, as the mantle and

100

surface would be close to redox equilibrium in Hadean times, as appreciable oxidation of surface

environment would not occur until much later (Farquhar et al., 2000). These results suggest that

the inception of relatively oxidizing conditions in the Earth’s mantle to be within the first 150

Ma of the Earth’s history. As the Earth’s mantle would have been considerably more reducing at

the time of core formation (Wood et al., 2006), a mechanism is required for its relatively rapid

oxidation, possibly through the arrival of oxidized material during the late stages of accretion

(O’Neill, 1991) or through the loss of H2 immediately after core formation (Bali et al., 2013).

4.7 Summary and Conclusions

Using a recently developed model for the redox behavior of Ce in silicate melts, combined with

lattice strain constraints on 3

/

Ce

zircon meltD and 4

/

Ce

zircon meltD , we present a method for calculating the fO2

at the time of crystallization from Ce anomalies in zircon. Oxygen fugacities calculated using

this method are in excellent agreement with independent estimates for BT, TT and UP,

demonstrating that zircon can be an accurate tool in the evaluation of magma redox state.

Application of this method to the Hadean zircons from the Jack Hills, Australia, result in fO2

estimates between FMQ -1.0 and +2.5, indicating that the Earth’s mantle reached its current

redox state by ~4.4 Ga .

4.8 References

Anderson, A.T., Newman, S., Williams, S.N., Druitt, T.H., Skirius, C., and Stolper, E., 1989.

H2O, CO2, Cl, and gas in Plinian and ash-flow Bishop rhyolite. Geology 17, 221-225.

Bali, E., Audetat, A., Keppler, H., 2013. Water and hydrogen are immiscible in Earth/'s mantle.

Nature 495, 220-222.

Ballard, J.R., Palin, M.J., and Campbell, I.H., 2002. Relative oxidation states of magmas inferred

from Ce(IV)/Ce(III) in zircon: application to porphyry copper deposits of northern Chile.


Blundy, J.D., and Wood, B.J., 1994. Prediction of crystal-melt partition coefficients from elastic


101

Cavosie, A.J., Valley, J.W., and Wilde, S.A., 2006. Correlated microanalysis of zircon: Trace

element, δ18O, and U–Th–Pb isotopic constraints on the igneous origin of complex >3900 Ma

detrital grains. Geochimica et Cosmochimica Acta 70, 5601-5616.

Cherniak, D.J., Hanchar, J.M., and Watson, E.B., 1997a. Diffusion of tetravalent cations in

zircon. Contributions to Mineralogy and Petrology 127, 383-390.

Cherniak, D.J., Hanchar, J.M., and Watson, E.B., 1997b. Rare-earth diffusion in zircon.

Chemical Geology 134, 289-301.

Chesner, C.A., 1998. Petrogenesis of the Toba tuffs, Sumatra, Indonesia. Journal of Petrology

39, 397-438.

Chesner, C.A., and Luhr, J.F., 2010. A melt inclusion study of the Toba Tuffs, Sumatra,

Indonesia. Journal of Volcanology and Geothermal Research 197, 259-278.

Claiborne, L.L., Miller, C.F., Flanagan, D.M., Clynne, M.A., and Wooden, J. L., 2010. Zircon

reveals protracted magma storage and recycling beneath Mount St. Helens. Geology 38, 1011-

1014.

David, F.H., and Vokhmin, V., 2003. Thermodynamic properties of some tri- and tetravalent

actinide aquo ions. New Journal of Chemistry 27, 1627-1632.

Emslie, R.F., and Loveridge, W.D., 1992. Fluorite-bearing early and middle Proterozoic granites,

Okak Bay area, Labrador: Geochronology, geochemistry and petrogenesis. Lithos 28, 87-109.

Emslie, R.F., and Stirling, J.A.R., 1993. Rapakivi and related granitoids of the Nain plutonic

suite: Geochemistry, mineral assemblages and fluid equilibria. Canadian Mineralogist 31, 821-

847.

Farquhar, J., Bao, H., Thiemens, M., 2000. Atmospheric influence of Earth's earliest sulfur cycle.

Science 289, 756-758.

Ferry, J.M., and Watson, E.B., 2007. New thermodynamic models and revised calibrations for

the Ti-in-zircon and Zr-in-rutile thermometers. Contributions to Mineralogy and Petrology 154,

429-437.

102

Fujimaki, H., 1986. Partition coefficients of Hf, Zr, and REE between zircon, apatite, and liquid.


Hanchar, J.M., and van Westrenen, W., 2007. Rare earth element behavior in zircon-melt

systems. Elements 3, 37-42.

Hayden, L.A., and Watson, E.B., 2007. Rutile saturation in hydrous siliceous melts and its

bearing on Ti-thermometry of quartz and zircon. Earth and Planetary Science Letters 258, 561-

568.

Hildreth, W., and Wilson, C.J.N., 2007. Compositional Zoning of the Bishop Tuff. Journal of

Petrology 48, 951-999.

Hinton, R.W., and Upton, B.G.J., 1991. The chemistry of zircon: Variations within and between

large crystals from syenite and alkali basalt xenoliths. Geochimica et Cosmochimica Acta 55,

287-3302.

Hoskin, P.W.O., 2005. Trace-element compositions of hydrothermal zircon and the alteration of

Hadean zircon from the Jack Hill, Australia. Geochimica et Cosmochimica Acta 69, 637-648.

Luo, Y., and Ayers, J.C., 2009. Experimental measurements of zircon/melt trace-element

partition coefficients. Geochimica et Cosmochimica Acta 73, 3656-3679.

Martin, H., 1995. The Archaean grey gneisses and the genesis of the continental crust, in Condie,

K. C., ed., The Archaean Crustal Evolution, Elsevier, p. 205–259.

McDonough, W.F., and Sun, S.-s., 1995. The composition of the Earth. Chemical Geology 120,

223-253.

Nagasawa, H., 1970. Rare earth concentrations in zircons and apatites and their host dacites and

granites. Earth and Planetary Science Letters 9, 359-364.

O’Neill, H.St C., 1987. Quartz-fayalite-iron and quartz-fayalite-magnetite equilibria and the free

energy of formation of fayalite (Fe2SiO4) and magnetite (Fe3O4). American Mineralogist 72, 67-

75.

103

O’Neill, H.St C., 1988. Systems Fe-O and Cu-O: Thermodynamic data for the equilibria Fe-

"FeO," Fe-Fe3O4, "FeO"-Fe3O4, Fe3O4-Fe2O3, Cu-Cu2O, and Cu2O-CuO from emf

measurements. American Mieralogist 73, 470-486.

O’Neill, H.St C., 1991. The origin of the moon and the early history of the earth—A chemical

model. Part 2: The earth. Geochimica et Cosmochimica Acta 55, 1159-1172.

Rudnick, R.L., and Gao, S., 2003. Composition of the Continental Crust, in Editors-in-

Chief: Heinrich, D. H., and Karl, K. T., eds., Treatise on Geochemistry: Oxford, Pergamon, p. 1-

64.

Sano, Y., Terada, K., and Fukuoka, T., 2002. High mass resolution ion microprobe analysis of

rare earth elements in silicate glass, apatite and zircon: lack of matrix dependency. Chemical

Geology 184, 217-230.

Schreiber, H.D., 1982. The chemistry of uranium in glass-forming melts: Redox interactions of

U(VI)-U(V)-U(IV) with cerium in aluminosilicates. Journal of Non-Crystalline Solids 49, 189-

200.

Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomic

distances in halides and chaleogenides. Acta Crystallographica. Section B, Structural science 32,

751-767.

Thomas, J.B., Bodnar, R.J., Shimizu, N., and Sinha, A.K., 2002. Determination of zircon/melt

trace element partition coefficients from SIMS analtsis of melt inclusions in zircon. Geochimica

et Cosmochimica Acta 66, 2887-2901.

Trachenko, K., Dove, M.T., and Salje, E.K.H., 2002. Structural changes in zircon under alpha-

decay irradiation. Physical Review B 65, 180102(R).

Trail, D., Watson, E.B., and Tailby, N.D., 2011. The oxidation state of Hadean magmas and


Virgo, D., Mysen, B.O., Kushiro, I., 1980. Anionic constitution of 1-atmosphere silicate melts:

Implications for the structure of igneous melts. Science 20, 1371-1373.

104

Wallace, P.J., Anderson, A.T., and Davis, A.M., 1999. Gradients in H2O, CO2, and exsolved gas

in a large-volume silicic magma system: Interpreting the record preserved in melt inclusions

from the Bishop Tuff. Journal of Geophysical Research: Solid Earth 104, 20097-20122.

Watson, E.B., 1996. Surface enrichment and trace-element uptake during crystal growth.


Watson, E.B., and Harrison, T.M., 1983. Zircon saturation revisited: temperature and

composition effects in a variety of crustal magma types. Earth and Planetary Science Letters 64,

295-304.

Watson, E.B., and Harrison, T.M., 2005. Zircon thermometer reveals minimum melting

conditions on earliest Earth. Science 308, 841-844.

Wilde, S.A., Valley, J.W., Peck, W.H., and Graham, C.M., 2001. Evidence from detrital zircons

for the existence of continental crust and oceans on the Earth 4.4 Gyr ago. Nature 409, 175-178.

Wood, B.J., Walter, M.J., Wade, J., 2006. Accretion of the Earth and segregation of its core.

Nature 441, 825-833.

105

Figure 4.1 Representative zircon textures from the Bishop tuff. Images a) and b) are CL and

BSE images, respectively, of 12-BT2-9, c) through f) are CL images of 12-BT-1, 12-BT2-3, 12-

BT2-5, 12-BT2-2, respectively. Prominent sector and oscillatory zoning are present in all

zircons. Samples in a) and d) contain partially resorbed cores which appear bright in CL, and

dark in BSE. Apatite (ap) inclusions are common and are found in most grains. Scale bars = 50

m.

106

Figure 4.2 Cathodoluminescence images of typical zircons from the Toba tuff. The outer regions

of all samples show oscillatory zoning. Some grains contain cores which may be either euhedral

(e) or partially resorbed (a). Apatite inclusions are common and occur in almost all grains.

Images shown are of samples TT-49 (a), TT-27 (b), TT-40 (c), TT-55 (d), TT-23 (e) and TT-36

(f). Scale bars = 50 m.

107

Figure 4.3 Representative textures of zircons from Umiakovik pluton. Images are of samples 12-

UB-Z11 (a, – CL, and b, – BSE), 12-UB-Z48 (c, – CL), 12-UB-Z6 (d, – CL), 12-UB-Z18 (e, –

CL), and 12-UB-Z4 (f, – CL). Zircons from this suite typically have outer domains with either

broad (a, b, and d) or oscillatory zoning (c and e). Inner domains often contain xenocrystic cores

(c), and tend to have convolute zoning (e). In some cases zircons appear to be partially, or

entirely recrystallized (f). Mineral inclusions are rare in UP zircons, however, bubble inclusions

(bi) are fairly common. Scale bars = 50 m.

108

Figure 4.4 Graphical representation of the procedure used to estimate oxygen fugacity from a

combination of zircon-melt partitioning of Ce, and the relation between melt 4 3Ce Ce/melt meltx x vs fO2

calibrated in this thesis (Chapter 3). In this example, values of ln /zircon meltiD are calculated from

individual zircon and glass analyses from samples of the Toba tuff. Average values are shown as

points, whereas lines connect results from individual analyses. The data display linear arrays

when plotted as a function of the Blundy and Wood (1994) lattice strain parameter,

20 0( / 3 / 6)( )i ir r r r , for 3+ (left) and 4+ (center) cations. The interpolated values of 4

/

Ce

zircon meltD

and 3

/

Ce

zircon meltD are used to define the upper and lower asymptotes for the change in /Cezircon meltD

with 4 3Ce Ce/melt meltx x . The measured /

Cezircon meltD lies between these limits and can then be used to

calculate a value of 4 3Ce Ce/melt meltx x , and hence fO2. Values for /

Uzircon meltD are shown in middle panel

although were not used to constrain 4

/

Ce

zircon meltD .

109

Figure 4.5 Values of log fO2 calculated using the Ce-in-zircon oxygen barometer presented in

this study as a function of temperature for zircons from the Bishop tuff (blue diamonds), Toba

tuff (red circles), and Umiakovik pluton (green triangles) compared to independent estimates

(equivalently-coloured fields). Temperatures are calculated using zircon Ti contents and the

calibration of Ferry and Watson (2007; see Section 4.5). Independent estimates of fO2 and T for

the Bishop and Toba volcanic samples are derived from Fe-Ti oxide thermobarometry (Chesner,

1998; Hildreth and Wilson, 2007). Values for the Umiakovik pluton are estimated by the

intersection of isopleths defined by the hematite content of ilmenite and the ferric-ferrous ratio of

biotite (Emslie and Stirling, 1993). Typical error in calculated fO2 is shown in the upper left

corner. The dashed black curves correspond to the variation in fO2 with T along the FMQ buffer,

calculated from O’Neill (1987), and the magnitite-hematite (MH) and iron-wustite (IW) buffers

calculated from O’Neill (1988).

110

Figure 4.6 Diagram depicting the compositional terms from Equation 4.3 as a

function of the calculated values for log 4 3Ce Ce/melt meltx x (Equation 4.2) for analyses of

the >4.0 Ga Jack Hills zircons. Analyses selected were from oscillatory zone regions, with

magmatic REE patterns, for which oxygen isotopic composition and crystallization temperature

had been determined. The probability density functions for calculated log 4 3Ce Ce/melt meltx x are plotted

on the right hand vertical axis, corresponding to different parent melt trace element

concentrations (pyrolite = coarse dashed curve; bulk continental crust = fine dashed curve;

Archean TTG = solid curve). Diagonal dashed lines are fO2 isopleths, labeled according the log

fO2 relative to the FMQ buffer at 750oC. The horizontal grey bar corresponds to the likely range

in the melt composition parameter for zircon crystallizing melts. The vertical gray bar is one

standard deviation in the calculated values of 4 3Ce Ce/melt meltx x determined assuming an Archean TTG

parental magma. The box defined by the intersection of the two bars denotes a range in fO2

estimates for Hadean zircons of approximately FMQ -1.0 to +2.5.

111

Chapter 5

Conclusions

A detailed investigation of the redox behavior in silicate melts was conducted with the objective

to derive an oxygen barometer using the magnitude of Ce anomalies in zircon. This first involved

the calibration of the Ce M4,5-edge x-ray absorption near edge structure spectrum for quantitative

measurement of Ce redox state through comparison with potentiometric determinations of

Ce4+/Ce3+. It was found that Ce4+ has a greater fluorescence yield (FLY) response to the incident

radiation than Ce3+. It was suggested that this was the result of Ce4+ having a greater x-ray

capture cross-section. This effect appears to be more pronounced in highly polymerized melts

and implies a dependence on long range interactions involving the local environment of

neighboring oxygen.

Evaluation of Ce redox behavior in terrestrial igneous systems was carried out through

experiments involving the synthesis of Ce-doped glasses. Samples ranging in composition from

basalt to rhyolite (± H2O) were equilibrated at 0.001 and 1 GPa, under fO2 conditions varying

from FMQ –6.0 to FMQ +8.4, and temperatures from 1200 to 1500°C. It was found that for a

given melt composition, the change in Ce4+/Ce3+ ratio with fO2 follows the trend predicted from

the reaction stoichiometry assuming simple oxides as melt species. In addition to fO2,

temperature, melt composition and water content have been found to be secondary controls on

Ce4+/Ce3+. The variation in Ce4+/Ce3+ with temperature indicates the reaction is exothermic with

an enthalpy of -109.2 (± 4.8) kJ/mol, which is constant for all melt compositions investigated. It

was also found that more depolymerized melts and hydrous compositions favor the stabilization

of Ce3+. These results indicate that even at relatively low oxygen fugacity, minor to trace

amounts of Ce4+ will be present in most terrestrial igneous systems, suggesting that Ce

partitioning could be a sensitive indicator of fO2.

Using the empirical model for Ce4+/Ce3+ in silicate melt as a function of fO2, T and melt

composition, combined with the method of Ballard et al. (2002) for estimation of melt Ce4+/Ce3+,

a method for calculating the fO2 at the time of crystallization from Ce anomalies in zircon was

derived. Oxygen fugacities calculated using this method are in excellent agreement with

independent estimates for the Bishop tuff, California (+1.2 ± 0.7), Toba tuff, Indonesia (+1.0 ±

112

1.1), and the Nane plutonic suite, Labrador (-2.4 ± 0.6), demonstrating that zircon can be an

accurate tool in the evaluation of magma redox state.

Values for /zircon meltiD were calculated by assuming that zircons have equilibrated with three

different types of parental magma compositions: pyrolite (i.e., equivalent to unmelted upper

mantle; McDonough and Sun, 1995), bulk continental crust (Rudnick and Gao, 2003), and

Archean tonalite-trondhjemite-granodiorite (TTG, Martin, 1995). Comparison of partition

coefficients calculated from these different melt reservoirs, to values measured by Sano et al.

(2002), suggest a close similarity of values using the Archean TTG as the parent melt of the Jack

Hills zircons. The resulting fO2 estimates from these zircons range between FMQ -1.0 and +2.5.

These results suggest that oxidation fugacity of the Earth’s mantle reached its current state by the

time of Hadean zircon crystallization, implying an increase of approximately five log units in fO2

within 150 Ma after segregation of the Earth’s core.

The results of this study would be greatly complimented by further constraints on the

incorporation of 4+ cations into zircon. Although there are few appropriate cations for such an

investigation Sn, or possibly Po, may serve this purpose. Furthermore, it was concluded that the

experimental investigation of Ce redox behavior in Fe-bearing silicate melts was subject to

charge transfer on quench. However the presence of Fe in a silicate melt may still effect the Ce

redox equilibrium. This may be evaluated in the future through in situ spectroscopic

measurement at temperature. Lastly, it should be noted that this technique is not restricted to

redox measurements using zircon. The Ce redox model presented here can be applied to any

phase, crystalizing within the calibrated compositional range, which preferentially incorporates

Ce4+ into its structure. Potential mineral phases that may warrant future study include rutile,

baddeleyite, and cassiterite.

113

Appendix 1 Methods for Potentiometric Titrations

Potentiometric titrations to determine the Ce4+/Ce3+ in glasses were done using a modified

version of the method described by Paul and Douglas (1965). In general this method involves

the dissolution of a sample in an HF – H2SO4 mixture. The resulting solution is titrated against a

standard reducing agent, in the case of these measurements ammonium Fe (II) sulphate solution

was used. The amount of titrant needed to reach the equivalence is linearly related to the

proportion of Ce4+ present in the sample solution. Measurement of the total Ce concentration

(Ce) then allows for Ce4+/Ce3+ [Ce4+/ (Ce-Ce4+)] to be determined. This same method can be

used for measurements of the concentration of other oxidized species or reduced specie using a

standard oxidizing agent, with the proper selection of reference electrode and titrant. The

following Appendix will describe this technique in detail.

A1.1 Sample Preparation

Run product glasses were crushed to ~0.1 mm in an agate mortar, which was first cleaned with

silica sand followed by rinsing with concentrated HCl. Samples were then weighed to ~20 mg on

aluminum foil and added to the Savillex digestion container (Figure A1.1c). The aluminum foil

was weighed after this to insure accuracy. 1.0 mL of a 12% HF – 7% H2SO4 was added to the

digestion container followed by a stir bar. The container would then be sealed, and the container

walls would then be washed with the acid solution to ensure that all of the glass was immersed in

the acid mixture. Samples were then digested in an ice bath for 4 h under constant stirring.

Following digestion all samples were visually inspected under a microscope to determine if the

dissolution had gone to completion. After digestion, 250.0 mg of boric acid was added to each

sample to complex with the fluorine ions, and stirred for an additional 30 min under an ice bath.

After the digestion procedure samples were transferred to a larger Savillex container. The

digestion container was washed with 35 mL of deionized H2O which was then added to the

sample solution. The entire digestion procedure was done with samples contained in an ice-water

bath.

114

A1.2 Preparation of Titrant

For the measurement of Ce redox state ammonium Fe (II) sulphate solution was chosen for the

reducing agent. Based on the sample size low concentrations of Ce4+ anticipated, the

concentration of ammonium Fe (II) sulphate used for the titrant was 5.0 x 10-5 N. The

ammonium Fe (II) sulphate used for these measurements was purchased from Fisher Scientific

quoted to be 2.82 x 10-3 N. This was, therefore, diluted with 554 mL of deionized H2O for every

10 mL of ammonium Fe (II) sulphate solution. The resulting solution was stored in Nalgene

bottle and refrigerated before use.

A1.3 Calibration

Calibration of this technique was carried out using three different Ce4+ compounds including:

Ce(SO4)2, (NH4)4Ce(SO4)4 2H2O, and (NH4)3Ce(NO3)6. Calibration with CeO2 was also

attempted but was found to be insoluble in the acid mixture. These materials were weighed by

the same method as the samples an mixed with approximately 20 mg of Ce-free blank glass with

the same major element concentrations as the andesite composition investigated. The calibration

curve was then constructed by plotting the mass of Ce4+ added versus the volume of titrant added

at the equivalence point (Figure 2.1). Standard titrations were run before and after the samples to

ensure oxidation of the ammonium Fe (II) sulphate solution had not occurred over the course of

the analyses.

A1.4 Measurement

Voltage measurements of the solution were made using a Pt pin indicator electrode and an epoxy

body Ag-AgCl reference electrode which were connected to an Orion model 525a+ pH meter,

calibrated using 220 mV and 427 mV redox buffer solutions. Voltage was monitored as a

function of the volume of titrant added to the sample solution. As Ce4+ begins to become

exhausted in the solution Fe2+ will become a stable species. The point in the titration where the

concentrations of these two species are equal is the equivalence point. The volume needed to

reach the equivalence point is directly proportional to the concentration of Ce4+ in the initial

solution. A representative plot of one of these titrations is shown in Figure A1.2.

115

Figure A1.1 a) and b) set-up for potentiometric titrations. c) container for digestions.

116

Figure A1.2 Titration results for sample DS09-C3-10. a) Pt-electrode potential versus volume of

titrant added. b) First-derivative curve for the titration. Dashed line shows the equivalence point.

117

Appendix 2 Compositions of Experimental Run Products

Table A2.1 Electron microprobe analyses of glass run products (values in wt%). Values of

NBO/T calculated as anhydrous for all experiments. Errors given in brackets.

Sample DS08-C1-02 DS08-C1-09 DS08-C1-20 DS09-C2-06 DS09-C2-09n 10 8 10 10 10 SiO2 65.44 (0.5) 57.36 (0.38) 67.42 (0.22) 56.11 (0.19) 59.70 (0.44)TiO2 0.31 (0.14) 0.54 (0.16) 0.26 (0.14) 0.64 (0.18) 0.55 (0.2) Al2O3 13.19 (0.2) 17.01 (0.13) 13.38 (0.13) 16.57 (0.14) 17.76 (0.23)MgO 0.64 (0.04) 3.72 (0.09) 0.63 (0.04) 3.56 (0.1) 3.88 (0.1) CaO 1.62 (0.03) 10.87 (0.11) 1.72 (0.04) 10.68 (0.12) 11.27 (0.13)Na2O 5.11 (0.16) 2.65 (0.18) 5.82 (0.23) 3.34 (0.18) 2.72 (0.12) K2O 3.40 (0.09) 0.66 (0.01) 3.43 (0.1) 0.65 (0.05) 0.67 (0.05) P2O5 0.11 (0.05) - 0.10 (0.12) - - Ce2O3 1.03 (0.11) 1.17 (0.09) 1.05 (0.15) 1.08 (0.07) 1.08 (0.12) Total 90.89 (0.81) 93.95 (0.53) 93.85 (0.47) 92.67 (0.54) 97.64 (0.83)NBO/T 0.065 (0.003) 0.276 (0.009) 0.080 (0.002) 0.295 (0.009) 0.273 (0.012)log Ce4+/Ce3+ -1.47 (0.11) -0.94 (0.11) -0.56 (0.11) -1.09 (0.11) -0.44 (0.11)

Sample DS09-C2-12 DS09-C2-13 DS09-C3-03 DS09-C3-04 DS09-C3-05n 10 8 9 10 10 SiO2 51.88 (0.35) 72.34 (0.42) 51.52 (0.3) 48.94 (0.35) 71.42 (0.27)TiO2 1.01 (0.16) 0.37 (0.11) 0.95 (0.17) 1.06 (0.23) 0.22 (0.14) Al2O3 16.39 (0.16) 14.35 (0.19) 16.26 (0.17) 16.18 (0.23) 14.16 (0.14)MgO 11.46 (0.12) 0.66 (0.05) 11.40 (0.15) 10.95 (0.14) 0.68 (0.05) CaO 16.29 (0.12) 1.83 (0.07) 16.24 (0.15) 15.32 (0.1) 1.90 (0.06) Na2O 2.11 (0.13) 6.06 (0.11) 2.28 (0.14) 2.84 (0.12) 6.36 (0.21) K2O - 3.55 (0.1) - - 3.76 (0.09) P2O5 - 0.09 (0.05) - - 0.12 (0.06) Ce2O3 1.13 (0.12) 1.17 (0.10) 1.12 (0.1) 1.13 (0.04) 1.17 (0.13) Total 100.31 (0.58) 100.48 (0.74) 99.8 (0.62) 96.44 (0.6) 99.84 (0.27)NBO/T 0.766 (0.002) 0.074 (0.002) 0.774 (0.029) 0.772 (0.036) 0.090 (0.002)log Ce4+/Ce3+ -1.09 (0.07) -0.41 (0.01) -2.09 (0.11) -1.12 (0.10) -1.63 (0.13)

118

Sample DS09-C3-06 DS09-C3-07 DS09-C3-08 DS09-C3-09 DS09-C3-10n 10 10 10 10 9 SiO2 60.84 (0.37) 51.65 (0.27) 61.01 (0.29) 50.16 (0.42) 71.44 (0.32)TiO2 0.58 (0.19) 1.05 (0.15) 0.69 (0.15) 0.99 (0.2) 0.25 (0.1) Al2O3 17.90 (0.14) 16.28 (0.14) 17.96 (0.16) 15.86 (0.16) 14.36 (0.13)MgO 4.09 (0.1) 11.53 (0.18) 4.02 (0.07) 11.10 (0.2) 0.7 (0.04) CaO 11.75 (0.19) 16.28 (0.12) 11.74 (0.15) 15.81 (0.22) 1.84 (0.04) Na2O 2.83 (0.19) 2.24 (0.15) 2.80 (0.09) 2.18 (0.13) 6.39 (0.2) K2O 0.67 (0.06) - 0.66 (0.03) - 3.65 (0.14) P2O5 - - - - 0.09 (0.04) Ce2O3 1.16 (0.1) 1.11 (0.1) 1.15 (0.1) 1.05 (0.05) 1.10 (0.09) Total 99.87 (0.68) 100.19 (0.61) 100.05 (0.48) 97.21 (0.59) 99.87 (0.44)NBO/T 0.290 (0.011) 0.777 (0.026) 0.284 (0.009) 0.771 (0.037) 0.084 (0.002)

log Ce4+/Ce3+ -2.01 (0.11) -1.46 (0.09) -1.26 (0.11) -1.60 (0.11) -0.99 (0.03)

Sample DS09-C3-13 DS09-C3-14 DS09-C3-15 DS09-C3-17 DS09-C3-18n 10 10 10 10 10 SiO2 51.57 (0.37) 51.32 (0.34) 60.64 (0.27) 51.35 (0.33) 60.99 (0.34)TiO2 1.09 (0.19) 1.00 (0.11) 0.60 (0.09) 1.06 (0.13) 0.60 (0.13) Al2O3 16.60 (0.15) 16.48 (0.23) 18.32 (0.15) 16.52 (0.25) 18.08 (0.14)MgO 11.61 (0.18) 11.46 (0.13) 4.13 (0.05) 11.54 (0.08) 4.11 (0.06) CaO 16.14 (0.14) 16.15 (0.17) 11.55 (0.13) 16.09 (0.13) 11.74 (0.17)Na2O 2.17 (0.09) 2.23 (0.1) 2.86 (0.14) 2.30 (0.08) 2.88 (0.16) K2O - - 0.65 (0.04) - 0.65 (0.04) P2O5 - - - - - Ce2O3 1.10 (0.1) 1.12 (0.09) 1.20 (0.11) 1.12 (0.1) 1.21 (0.11) Total 100.30 (0.39) 99.81 (0.59) 99.96 (0.38) 100.04 (0.59) 100.29 (0.49)NBO/T 0.765 (0.031) 0.768 (0.031) 0.280 (0.007) 0.769 (0.031) 0.288 (0.009)log Ce4+/Ce3+ -2.83 (0.11) -2.98 (0.11) -1.93 (0.11) -1.04 (0.11) -0.87 (0.03)

119

Sample DS09-C3-19 DS09-C3-20 DS09-C3-22 DS09-C3-23 DS09-C3-24

n 10 10 10 10 10

SiO2 72.01 (0.42) 52.23 (0.37) 57.3 (0.31) 61.21 (0.36) 72.09 (0.37)

TiO2 0.30 (0.1) 1.07 (0.16) 0.60 (0.08) 0.65 (0.14) 0.30 (0.1)

Al2O3 14.29 (0.15) 16.53 (0.15) 16.87 (0.17) 18.17 (0.17) 14.38 (0.17)

MgO 0.69 (0.04) 11.52 (0.09) 3.73 (0.08) 4.08 (0.09) 0.71 (0.03)

CaO 1.89 (0.07) 16.35 (0.1) 10.95 (0.26) 11.72 (0.13) 1.91 (0.07)

Na2O 6.04 (0.22) 1.94 (0.12) 2.91 (0.22) 2.79 (0.06) 6.02 (0.2)

K2O 3.51 (0.09) - 0.60 (0.03) 0.67 (0.06) 3.64 (0.12)

P2O5 0.08 (0.05) - - - 0.11 (0.14)

Ce2O3 1.18 (0.08) 1.14 (0.15) 1.07 (0.13) 1.21 (0.13) 1.18 (0.1)

Total 100.04 (0.56) 100.82 (0.55) 94.07 (0.54) 100.54 (0.6) 100.38 (0.39)

NBO/T 0.077 (0.002) 0.757 (0.029) 0.287 (0.01) 0.282 (0.009) 0.078 (0.003)

log Ce4+/Ce3+ -0.60 (0.03) -2.47 (0.11) -1.02 (0.11) -1.72 (0.11) -1.58 (0.13)

Sample DS09-C3-25 DS09-C3-28 DS09-C3-29 DS09-C3-32 DS09-C3-37n 8 10 9 10 9 SiO2 70.92 (0.78) 51.62 (0.26) 60.91 (0.25) 71.83 (0.27) 66.63 (0.97)TiO2 0.41 (0.11) 1.08 (0.16) 0.60 (0.11) 0.28 (0.14) 0.34 (0.1) Al2O3 15.10 (0.18) 16.25 (0.16) 17.89 (0.12) 14.27 (0.12) 13.18 (0.3) MgO 1.35 (0.05) 11.60 (0.13) 4.10 (0.1) 0.72 (0.04) 0.58 (0.1) CaO 4.37 (0.10) 16.32 (0.22) 11.69 (0.13) 1.92 (0.08) 1.61 (0.26) Na2O 4.56 (0.13) 2.11 (0.15) 2.74 (0.17) 6.43 (0.26) 5.57 (0.22) K2O 2.86 (0.04) - 0.64 (0.03) 3.53 (0.08) 3.56 (0.11) P2O5 - - - 0.06 (0.07) - Ce2O3 1.27 (0.29) 1.14 (0.05) 1.13 (0.11) 1.17 (0.09) 1.01 (0.17)

Total 100.90 (0.88) 100.17 (0.34) 99.74 (0.4) 100.26 (0.57) 92.62 (0.53)NBO/T 0.133 (0.031) 0.779 (0.027) 0.286 (0.009) 0.087 (0.002) 0.075 (0.005)log Ce4+/Ce3+ -1.27 (0.07) -2.80 (0.11) -2.33 (0.11) -2.09 (0.13) -1.62 (0.11)

120

Sample DS11-C21 DS11-C22 DS11-C24 DS11-C28 DS11-C33 n 10 10 9 10 9 SiO2 51.57 (0.26) 51.58 (0.24) 71.23 (0.39) 72.16 (0.44) 51.91 (0.45)TiO2 1.02 (0.11) 1.20 (0.2) 0.33 (0.11) 0.29 (0.11) 1.04 (0.22) Al2O3 16.28 (0.24) 16.27 (0.12) 14.59 (0.16) 14.55 (0.15) 16.50 (0.13)MgO 11.40 (0.14) 11.35 (0.12) 0.65 (0.02) 0.66 (0.03) 11.32 (0.19)CaO 16.23 (0.14) 16.22 (0.22) 1.79 (0.06) 1.75 (0.11) 16.18 (0.16)Na2O 2.12 (0.08) 2.14 (0.15) 5.93 (0.27) 5.37 (0.2) 2.15 (0.16) K2O - - 3.64 (0.1) 3.53 (0.13) - P2O5 - - 0.13 (0.11) 0.07 (0.09) - Ce2O3 1.15 (0.11) 1.09 (0.1) 1.02 (0.1) 0.98 (0.11) 1.09 (0.12)


Sample DS11-C34 DS11-C35 DS11-C36 DS11-C38 DS11-C40 n 10 10 10 10 10 SiO2 60.96 (0.34) 70.39 (0.43) 72.11 (0.56) 71.05 (1.06) 71.23 (0.27)TiO2 0.68 (0.13) 0.39 (0.11) 0.34 (0.12) 0.32 (0.13) 0.25 (0.15) Al2O3 18.01 (0.2) 15.08 (0.22) 14.52 (0.16) 14.62 (0.67) 14.37 (0.16)MgO 3.93 (0.09) 1.35 (0.07) 0.65 (0.05) 0.66 (0.04) 0.64 (0.04) CaO 11.77 (0.22) 4.22 (0.08) 1.82 (0.06) 1.79 (0.08) 1.83 (0.06) Na2O 2.83 (0.13) 4.57 (0.19) 6.16 (0.27) 6.57 (0.28) 6.49 (0.15) K2O 0.72 (0.05) 2.88 (0.10) 3.59 (0.17) 3.79 (0.09) 3.76 (0.09) P2O5 - - 0.02 (0.04) 0.14 (0.14) 0.09 (0.12) Ce2O3 0.96 (0.1) 1.21 (0.11) 0.98 (0.09) 0.92 (0.08) 0.94 (0.12)


121

Sample DS11-C41 DS11-C44 DS11-C49 DS11-C50 DS11-C63 n 9 10 10 10 10 SiO2 71.15 (1.09) 71.60 (0.95) 51.86 (0.48) 61.53 (0.28) 61.36 (0.33)TiO2 0.18 (0.13) 0.25 (0.12) 1.17 (0.15) 0.58 (0.14) 0.56 (0.15) Al2O3 14.56 (0.72) 14.30 (0.67) 16.38 (0.18) 18.42 (0.17) 18.09 (0.11)MgO 0.64 (0.03) 0.63 (0.06) 11.28 (0.14) 3.96 (0.1) 3.94 (0.11) CaO 1.81 (0.08) 1.74 (0.08) 16.30 (0.15) 11.59 (0.12) 11.7 (0.11) Na2O 6.36 (0.35) 6.45 (0.39) 2.14 (0.07) 2.66 (0.15) 2.42 (0.15) K2O 3.71 (0.13) 3.73 (0.08) - 0.7 (0.05) 0.72 (0.05) P2O5 0.11 (0.07) 0.14 (0.13) - - - Ce2O3 0.97 (0.07) 0.94 (0.1) 1.12 (0.19) 0.95 (0.09) 0.96 (0.12)


Sample DS11-C64 DS11-C66 n 10 10 SiO2 51.99 (0.19) 71.88 (0.71) TiO2 0.92 (0.19) 0.3 (0.1) Al2O3 16.4 (0.13) 14.76 (0.3) MgO 11.61 (0.16) 0.73 (0.07) CaO 16.32 (0.16) 1.99 (0.26) Na2O 1.6 (0.15) 5.62 (0.2) K2O - 3.56 (0.08) P2O5 - 0.07 (0.08) Ce2O3 1.14 (0.09) 1.03 (0.15)

Total 100.01 (0.34) 99.97 (0.45) NBO/T 0.758 (0.023) 0.063 (0.003) log Ce4+/Ce3+ -3.57 (0.11) -2.57 (0.13)

122

A2.1 Calculation of NBO/T and xH2O

The ratio of non-bridging oxygen (NBO) to tetrahedrally coordinated cations (T) was calculated

as follows:

1) Compositional data obtained using EPMA were first converted from wt.% oxide to

atomic mole percent.

2) All Si, Ti, and P were assigned to T

3) If Al < Na + K + 2Ca +2Mg than all Al was assigned to T. If Al > Na + K + 2Ca +2Mg

than the proportion of Al in T = Na + K + 2Ca + 2Mg

4) For samples containing Fe the proportion of Fe3+ assigned to T = Na + K - Al if Fe3+ >

Na + K – Al.

5) NBO =2O - 4T

Example:

wt % oxide mole % atomic

SiO2 70.4 → 24.3 Si

TiO2 0.4 → 0.1 Ti

Al2O3 15.1 → 6.1 Al

MgO 1.4 → 0.7 Mg

MnO 0.0 → 0.0 Mn

FeO 0.0 → 0.0 Fe2+

Fe2O3 0.0 → 0.0 Fe3+

CaO 4.2 → 1.6 Ca

NiO 0.0 → 0.0 Ni

Na2O 4.6 → 3.1 Na

K2O 2.9 → 1.3 K

P2O5 0.0 → 0.0 P

Ce2O3 1.2 → 0.2 Ce

H2O 0.0 → 0.0 H

62.6 O

Total 100.1 100 Total

123

Al < Na + K + 2Mg + 2Ca

6.1 <3.1+1.3+2(0.7)+2(1.6)

6.1<8.7 →True

T = Si + Ti + Al + P

T = 24.3 + 0.1 + 6.1 + 0

T = 30.5

NBO = 2O - 4T

NBO = 2(62.6) - 4(30.5)

NBO = 3.2

NBO/T = 3.2/30.5

NBO/T = 0.10

124

The mole fraction of H2O (xH2O) in hydrous experiments were calculated by converting the

compositional data obtained using EPMA from wt. % oxide to mole fraction oxide. The resulting

values for HO1/2 were then divide by 2 to yield the mole fraction of H2O

Example:

mole % atomic

mole fraction

oxide

Si 17.5 → 0.406 SiO2

Ti 0.1 → 0.003 TiO2

Al 6.1 → 0.141 AlO3/2

Mg 1.7 → 0.039 MgO

Mn 0.0 → 0.000 MnO

Fe2+ 0.0 → 0.000 FeO

Fe3+ 0.0 → 0.000 FeO3/2

Ca 3.6 → 0.083 CaO

Ni 0.0 → 0.000 NiO

Na 1.7 → 0.040 NaO1/2

K 0.2 → 0.005 KO1/2

P 0.0 → 0.000 PO5/2

Ce 0.1 → 0.003 CeO3/2

H 12.1 → 0.280 HO1/2

O 56.8

Total 100 1.00 Total

xHO1/2 = 0.280 → xH2O = 0.140

125

Appendix 3 Curve Fitting Parameters of XANES Spectra

Table A3.1 Curve parameters used to fit Ce M4,5-edge. E = the energy position, d = Gaussian

width, and z = the Lorentzian-Gaussian width ratio. Letters A-E (and A’-E’) correspond to

features shown in Figure 2.3. Errors given in brackets.

DS08-C1-02

M5 M4

Curve Type E d z Area Curve Type E d z Area

A Voigt 879.7 0.93 1.36 0.03 A' Voigt 896.5 0.17 0.00 0.07

B Voigt 881.5 0.45 1.42 0.76 B' Voigt 898.4 0.31 0.84 0.11

C Voigt 882.5 0.04 1.12 1.36 C' Voigt 899.7 0.22 2.95 1.36

D Voigt 884.0 0.48 3.81 0.12 D' Voigt 901.8 0.37 2.94 0.43

E Voigt 889.0 0.59 0.80 0.01 E' Voigt 907.1 0.45 2.51 0.09

Arctan 900.4 0.24 - 0.93

DS08-C1-09

M5 M4

Curve Type E d z Area Curve Type E d z AreaA Voigt 879.5 0.37 1.36 0.02 A' Voigt 896.6 0.17 0.00 0.10B Voigt 881.5 0.45 1.39 0.79 B' Voigt 898.3 0.32 0.80 0.12C Voigt 882.5 0.04 1.09 1.24 C' Voigt 899.7 0.22 2.97 1.36D Voigt 884.0 0.37 3.96 0.23 D' Voigt 901.9 0.37 2.94 0.52E Voigt 888.7 0.61 0.80 0.03 E' Voigt 906.7 0.44 2.58 0.05

Arctan 900.3 0.24 - 1.40

DS08-C1-20M5 M4


Arctan 900.4 0.24 - 0.57

126

DS09-C2-06M5 M4


Arctan 900.5 0.24 - 1.342

DS09-C2-09

M5 M4


Arctan 900.4 0.24 - 0.814

DS09-C2-12

M5 M4


Arctan 900.3 0.24 - 1.316

DS09-C2-13

M 5 M 4


Arctan 900.4 0.24 - 0.961

127

DS09-C3-03

M 5 M 4


Arctan 900.1 0.24 - 0.858

DS09-C3-04

M5 M4


Arctan 900.4 0.24 - 0.361

DS09-C3-05

M 5 M 4


Arctan 900.4 0.24 - 1.124

DS09-C3-06

M 5 M 4


Arctan 900.1 0.24 - 0.902

128

DS09-C3-07

M 5 M 4


Arctan 900.0 0.24 - 0.835

DS09-C3-08

M 5 M 4


Arctan 900.0 0.24 - 0.774

DS09-C3-09

M5 M4


Arctan 900.3 0.24 - 0.881

DS09-C3-10

M 5 M 4


Arctan 900.1 0.24 - 0.429

129

DS09-C3-13

M 5 M 4


Arctan 900.0 0.24 - 1.358

DS09-C3-14

M 5 M 4


Arctan 900.0 0.24 - 1.295

DS09-C3-15

M 5 M 4


Arctan 900.0 0.24 - 0.930

DS09-C3-17

M 5 M 4


Arctan 900.3 0.24 - 1.216

130

DS09-C3-18

M 5 M 4


Arctan 900.2 0.24 - 0.544

DS09-C3-19

M 5 M 4


Arctan 900.2 0.24 - 0.783

DS09-C3-20

M 5 M 4


Arctan 900.1 0.24 - 0.934

DS09-C3-22

M5 M4


Arctan 901.1 0.24 - 1.020

131

DS09-C3-23

M 5 M 4


Arctan 900.1 0.24 - 1.102

DS09-C3-24

M 5 M 4


Arctan 900.3 0.24 - 0.819

DS09-C3-25

M5 M4


Arctan 900.3 0.24 - 0.512

DS09-C3-28

M5 M4


Arctan 900.1 0.24 - 0.717

132

DS09-C3-29

M 5 M 4


Arctan 900.1 0.24 - 0.411

DS09-C3-32

M 5 M 4


Arctan 900.3 0.24 - 0.722

DS09-C3-37

M5 M4


Arctan 900.3 0.24 - 0.499

DS11-C21

M 5 M 4


Arctan 900.1 0.24 - 0.329

133

DS11-C22

M5 M4


Arctan 900.0 0.24 - 0.838

DS11-C24

M 5 M 4


Arctan 900.0 0.24 - 0.324

DS11-C28

M 5 M 4


Arctan 900.0 0.24 - 0.755

DS11-C34

M 5 M 4


Arctan 900.1 0.24 - 0.880

134

DS11-C35

M 5 M 4


Arctan 900.1 0.24 - 0.439

DS11-C36

M 5 M 4


Arctan 900.1 0.24 - 0.331

DS11-C38

M5 M4


Arctan 900.4 0.24 - 0.942

DS11-C40

M5 M4


Arctan 900.1 0.24 - 1.160

135

DS11-C41

M5 M4


Arctan 900.6 0.24 - 0.934

DS11-C44

M5 M4


Arctan 900.5 0.24 - 0.706

DS11-C63

M 5 M 4


Arctan 900.0 0.24 - 1.381

DS11-C64

M 5 M 4


Arctan 900.0 0.24 - 0.883

136

DS11-C66

M 5 M 4


Arctan 900.0 0.24 - 0.796

Damage Series - 1

M5 M4


Arctan 900.0 0.24 - 1.310

Damage Series - 2

M5 M4


Arctan 900.1 0.24 - 1.234

Damage Series - 3

M5 M4


Arctan 900.0 0.24 - 1.156

137

Damage Series - 4

M5 M4


Arctan 900.0 0.24 - 1.391

Damage Series - 5

M5 M4


Arctan 900.0 0.24 - 1.357

Damage Series - 6M5 M4Curve Type E d z Area Curve Type E d z AreaA Voigt 881.0 1.10 1.36 0.016 A' Voigt 896.5 0.14 0.00 0.082B Voigt 881.5 0.46 1.42 0.657 B' Voigt 898.2 0.32 0.80 0.120C Voigt 882.6 0.04 1.12 1.168 C' Voigt 899.6 0.23 2.91 1.209D Voigt 884.2 0.49 3.80 0.273 D' Voigt 901.9 0.38 2.89 0.530E Voigt 888.6 0.60 0.80 0.042 E' Voigt 906.6 0.45 2.51 0.055

Arctan 900.0 0.24 - 2.261

138

Appendix 4 Supplementary Information for Chapter 4

A4.1 Zircon Chemistry

Table A4.1 Major and trace element data for Bishop tuff zircon (n = number of EPMA analysis)

Samplen

SiO2 32.15 (0.10) 32.24 (0.10) 32.14 (0.10) 32.27 (0.10) 32.22 (0.10)

P2O5 0.27 (0.03) 0.32 (0.03) 0.20 (0.03) 0.17 (0.03) 0.20 (0.03)

ZrO2 64.44 (0.55) 63.99 (0.55) 63.10 (0.55) 63.92 (0.55) 63.73 (0.55)

HfO2 1.20 (0.03) 1.16 (0.03) 1.15 (0.03) 1.18 (0.03) 1.14 (0.03)

Y2O3 0.61 (0.05) 0.58 (0.05) 0.74 (0.05) 0.60 (0.05) 0.52 (0.05)

Total 98.73 (0.82) 98.34 (0.82) 97.40 (0.79) 98.22 (0.79) 97.86 (0.83)

Ti 11.2 (0.6) 10.5 (0.7) 10.1 (0.6) 10.2 (0.6) 11.0 (0.7)Y 2291 (118) 1878 (98) 3750 (197) 3201 (178) 1559 (89)La 0.11 (0.02) 0.08 (0.02) 2.22 (0.10)Ce 43.6 (1.8) 46.6 (1.9) 55.7 (2.3) 33.8 (1.4) 51.6 (2.2)Pr 0.09 (0.01) 0.15 (0.02) 0.12 (0.01) 0.07 (0.01) 0.97 (0.04)Nd 1.68 (0.12) 1.83 (0.18) 2.19 (0.18) 1.48 (0.09) 6.61 (0.32)Sm 5.62 (0.37) 6.98 (0.56) 7.02 (0.49) 5.30 (0.27) 6.67 (0.39)Eu 0.07 (0.02) 0.16 (0.04) 0.13 (0.03) 0.07 (0.01) 0.11 (0.02)Gd 42.3 (1.6) 46.5 (2.0) 48.7 (2.0) 36.6 (1.4) 40.7 (1.6)Tb 18.2 (0.6) 18.3 (0.7) 20.7 (0.7) 15.6 (0.5) 17.4 (0.6)Dy 225.0 (7.5) 222.8 (7.6) 252.1 (8.5) 197.0 (6.5) 209.7 (7.0)Ho 92.5 (3.2) 89.8 (3.2) 101.9 (3.6) 79.6 (2.8) 82.8 (3.0)Er 392.2 (13.1) 378.0 (12.8) 436.0 (14.7) 348.6 (11.8) 355.5 (12.1)Tm 88.5 (3.0) 83.4 (2.9) 95.1 (3.3) 78.1 (2.7) 78.3 (2.7)Yb 788.6 (25.9) 735.5 (24.4) 852.3 (28.2) 709.7 (23.4) 696.2 (23.1)Lu 159.3 (5.2) 146.5 (4.8) 166.8 (5.4) 138.3 (4.5) 136.3 (4.4)Hf 10185 (322) 9867 (313) 9772 (309) 10042 (318) 9693 (307)Th 1101 (39) 1586 (56) 2387 (85) 1109 (40) 1371 (49)U 1883 (74) 2050 (82) 2656 (106) 1967 (81) 2002 (84)

EPMA (wt.%)

LA-ICP-MS (ppm)

n.d. n.d.

2 2 2 2 1BTZ1-1 BTZ1-2 BTZ1-3 BTZ1-6 BTZ1-7

139

Samplen

SiO2 32.63 (0.11) 32.51 (0.11) 32.33 (0.11) 32.31 (0.11) 32.49 (0.11) 31.81 (0.10) 32.56 (0.11) 32.80 (0.11)

P2O5 0.34 (0.03) 0.18 (0.03) 0.17 (0.02) 0.30 (0.03) 0.29 (0.03) 0.19 (0.03) 0.17 (0.02) 0.02 (0.02)

ZrO2 63.70 (0.55) 63.75 (0.55) 64.63 (0.55) 64.69 (0.55) 65.67 (0.56) 63.22 (0.55) 63.62 (0.55) 64.86 (0.56)

HfO2 1.18 (0.03) 1.20 (0.03) 1.13 (0.03) 1.14 (0.03) 1.23 (0.03) 1.16 (0.03) 1.23 (0.03) 1.30 (0.03)

Y2O3 0.53 (0.05) 0.48 (0.05) 0.42 (0.04) 0.37 (0.04) 0.35 (0.04) 0.71 (0.05) 0.53 (0.05) 0.11 (0.04)

Total 98.45 (0.79) 98.23 (0.84) 98.68 (0.76) 98.86 (0.84) 100.09 (0.83) 97.15 (0.79) 98.18 (0.70) 99.13 (0.79)

Ti 11.8 (1.3) 10.7 (0.7) 10.3 (0.7) 10.7 (0.7) 11.6 (0.8) 10.3 (1.0) 9.7 (0.5)Y 2504 (146) 1780 (106) 1573 (100) 2019 (132) 2545 (181) 1825 (134) 1901 (80) 996 (58)La 0.07 (0.04) 2.40 (0.12) 0.52 (0.06)Ce 52.7 (2.6) 45.3 (2.0) 36.7 (1.7) 28.7 (1.4) 28.0 (1.4) 43.3 (2.4) 37.2 (1.3)Pr 0.04 (0.03) 0.08 (0.01) 0.09 (0.01) 0.94 (0.05) 0.15 (0.01) 0.33 (0.04) 0.08 (0.01)Nd 2.10 (0.41) 1.60 (0.12) 1.80 (0.16) 5.34 (0.31) 2.96 (0.19) 3.14 (0.36) 1.41 (0.12)Sm 8.42 (1.27) 5.99 (0.39) 4.91 (0.39) 5.58 (0.40) 9.17 (0.51) 6.92 (0.82) 4.76 (0.35) 23.41 (10.51)Eu 0.13 (0.09) 0.12 (0.03) 0.14 (0.04) 0.09 (0.03) 0.14 (0.03) 0.09 (0.05) 0.15 (0.03)Gd 53.5 (3.5) 42.0 (1.7) 37.4 (1.6) 34.1 (1.5) 49.5 (2.0) 43.5 (2.5) 36.1 (1.5)Tb 21.3 (1.0) 16.7 (0.6) 15.7 (0.6) 14.2 (0.5) 19.8 (0.7) 18.2 (0.8) 14.8 (0.5) 8.0 (3.5)Dy 260.3 (10.1) 202.4 (6.8) 193.1 (6.7) 180.6 (6.2) 238.9 (8.2) 220.0 (8.2) 183.0 (6.1) 82.5 (16.7)Ho 101.5 (4.0) 79.4 (2.9) 79.5 (3.0) 74.5 (2.8) 94.0 (3.6) 86.8 (3.6) 72.5 (2.6) 35.6 (5.8)Er 463.8 (16.9) 348.2 (12.0) 349.2 (12.2) 323.5 (11.4) 406.4 (14.6) 384.8 (14.4) 312.4 (10.4) 140.8 (17.9)Tm 95.3 (3.7) 74.9 (2.6) 76.9 (2.8) 73.1 (2.7) 88.6 (3.3) 83.1 (3.2) 69.8 (2.3) 46.2 (6.2)Yb 856.4 (30.2) 686.9 (23.0) 722.5 (24.4) 692.7 (23.5) 802.2 (27.5) 779.9 (27.6) 617.3 (20.7) 509.9 (45.2)Lu 159.2 (5.6) 133.9 (4.4) 141.1 (4.6) 135.7 (4.5) 154.1 (5.0) 144.8 (4.9) 123.0 (4.3) 105.2 (10.0)Hf 10005 (319) 10175 (322) 9586 (304) 9647 (305) 10398 (329) 9870 (313) 10405 (329) 10982 (413)Th 3610 (132) 1347 (49) 828 (31) 726 (27) 702 (27) 934 (36) 1283 (52) 182 (16)U 3559 (152) 1883 (82) 1416 (64) 1238 (57) 1218 (59) 1561 (78) 1901 (67) 1400 (64)

n.d.

EPMA (wt.%)

LA-ICP-MS (ppm)

n.d.n.d.n.d.

n.d.

n.d.

n.d. n.d. n.d. n.d. n.d.

1 1 1 1 1 1 2 1BTZ1-8 BTZ1-9 BTZ1-11 BTZ1-12 BTZ1-15 BTZ1-16 BTZ1-22 BTZ1-23

140

Samplen

SiO2 32.27 (0.11) 32.40 (0.11) 32.52 (0.11) 32.01 (0.10) 32.17 (0.07) 32.35 (0.08) 32.50 (0.08) 32.27 (0.16)

P2O5 0.33 (0.03) 0.29 (0.03) 0.14 (0.02) 0.25 (0.03) 0.17 (0.05) 0.16 (0.09) 0.09 (0.01) 0.11 (0.07)

ZrO2 65.73 (0.56) 63.86 (0.55) 64.68 (0.55) 63.09 (0.55) 64.80 (0.58) 64.85 (0.43) 64.75 (0.22) 64.72 (0.89)

HfO2 1.11 (0.03) 1.15 (0.03) 1.19 (0.03) 1.16 (0.03) 1.23 (0.02) 1.23 (0.03) 1.18 (0.05) 1.28 (0.07)

Y2O3 0.41 (0.04) 0.40 (0.04) 0.45 (0.04) 0.68 (0.05) 0.45 (0.09) 0.44 (0.02) 0.39 (0.12) 0.48 (0.35)

Total 99.92 (0.81) 98.13 (0.79) 99.09 (0.83) 97.23 (0.85) 98.86 (0.61) 99.07 (0.45) 98.95 (0.23) 98.91 (0.73)

Ti 9.1 (0.5) 9.1 (0.7) 16.3 (1.4) 12.9 (0.9) 6.8 (0.4) 6.6 (0.4) 6.0 (0.6) 5.4 (0.5)Y 1857 (82) 1855 (83) 2834 (132) 3586 (170) 1823 (81) 1445 (65) 1368 (63) 4874 (230)La 0.04 (0.01) 0.07 (0.02) 0.20 (0.05) 0.15 (0.05)Ce 34.9 (1.2) 20.3 (0.8) 45.1 (1.8) 56.6 (2.0) 29.3 (1.2) 23.4 (1.0) 24.2 (1.1) 30.8 (1.4)Pr 0.09 (0.01) 0.11 (0.02) 0.13 (0.04) 0.20 (0.03) 0.05 (0.01) 0.04 (0.02) 0.13 (0.03) 0.28 (0.04)Nd 1.41 (0.11) 1.84 (0.23) 2.02 (0.35) 3.02 (0.30) 1.26 (0.18) 0.65 (0.15) 1.09 (0.33) 4.32 (0.54)Sm 4.82 (0.32) 5.26 (0.61) 7.05 (0.98) 9.99 (0.87) 4.88 (0.39) 3.38 (0.33) 2.29 (0.54) 12.52 (1.13)Eu 0.08 (0.03) 0.12 (0.05) 0.09 (0.05)Gd 37.1 (1.5) 32.6 (1.8) 48.3 (2.9) 63.0 (2.9) 28.2 (1.4) 22.9 (1.2) 25.0 (1.9) 62.9 (3.3)Tb 15.0 (0.5) 14.0 (0.6) 19.4 (0.9) 27.9 (1.1) 11.6 (0.5) 9.3 (0.4) 8.5 (0.5) 22.7 (1.0)Dy 187.5 (6.2) 176.1 (6.3) 251.9 (9.3) 340.3 (11.6) 145.8 (5.0) 124.7 (4.3) 122.7 (4.9) 288.2 (10.3)Ho 75.9 (2.8) 68.7 (2.6) 104.3 (4.1) 138.4 (5.3) 57.8 (2.1) 48.0 (1.8) 51.4 (2.1) 104.5 (4.0)Er 325.0 (11.0) 300.2 (10.5) 446.7 (16.0) 599.8 (20.9) 255.2 (9.7) 217.5 (8.3) 216.6 (8.8) 492.7 (19.5)Tm 73.0 (2.4) 68.2 (2.4) 101.3 (3.6) 133.3 (4.6) 56.1 (1.9) 48.1 (1.7) 47.3 (1.8) 98.8 (3.5)Yb 646.2 (21.9) 635.1 (22.0) 930.9 (33.1) 1189.5 (41.4) 536.7 (18.0) 447.8 (15.1) 424.0 (15.2) 902.6 (31.0)Lu 124.3 (4.4) 122.0 (4.5) 188.3 (7.2) 225.5 (8.5) 96.2 (3.5) 89.8 (3.2) 84.8 (3.3) 166.7 (6.2)Hf 9419 (298) 9770 (310) 10112 (322) 9849 (312) 10457 (331) 10428 (330) 10020 (319) 10874 (345)Th 1294 (55) 544 (24) 1125 (51) 1736 (80) 856 (29) 467 (16) 550 (19) 929 (32)U 1734 (62) 948 (34) 1860 (68) 2896 (106) 1552 (55) 999 (36) 1043 (38) 1236 (45)

EPMA (wt.%)

LA-ICP-MS (ppm)


n.d. n.d. n.d. n.d.

3 3 31 1 1 1 3BTZ1-26 BTZ1-27 BTZ1-30 BTZ1-31 12-BT2-Z1 12-BT2-Z2 12-BT2-Z3 12-BT2-Z5

141

Samplen 3

SiO2 32.33 (0.05) 32.23 (0.14) 32.25 (0.15) 32.31 (0.06) 32.09 (0.07) 32.16 (0.16)

P2O5 0.23 (0.07) 0.11 (0.02) 0.19 (0.08) 0.23 (0.08) 0.23 (0.07) 0.18 (0.10)

ZrO2 64.91 (0.44) 65.12 (0.27) 64.87 (0.30) 64.46 (0.49) 65.31 (0.72) 65.00 (0.21)

HfO2 1.20 (0.01) 1.22 (0.04) 1.23 (0.06) 1.25 (0.00) 1.23 (0.02) 1.29 (0.05)

Y2O3 0.45 (0.07) 0.50 (0.10) 0.58 (0.11) 0.54 (0.20) 0.48 (0.03) 0.48 (0.14)

Total 99.16 (0.44) 99.21 (0.35) 99.17 (0.33) 98.83 (0.33) 99.41 (0.73) 99.13 (0.23)

Ti 9.2 (0.8) 5.7 (0.4) 6.1 (0.5) 6.0 (0.4) 6.2 (0.5) 7.2 (0.5)Y 2844 (137) 1704 (85) 1609 (82) 1440 (75) 1524 (82) 2307 (127)La 0.51 (0.07) 0.07 (0.03)Ce 54.2 (2.3) 16.8 (0.8) 28.4 (1.3) 13.5 (0.6) 24.9 (1.2) 35.9 (1.7)Pr 0.30 (0.04) 0.11 (0.02) 0.05 (0.02) 0.05 (0.01) 0.11 (0.02)Nd 3.11 (0.41) 1.70 (0.21) 1.04 (0.17) 0.86 (0.16) 0.94 (0.17) 1.57 (0.17)Sm 6.72 (0.73) 5.97 (0.51) 3.58 (0.38) 2.70 (0.30) 3.64 (0.37) 5.26 (0.48)EuGd 47.5 (2.6) 34.7 (1.8) 26.0 (1.4) 19.1 (1.1) 26.9 (1.5) 38.7 (2.0)Tb 19.3 (0.9) 13.2 (0.6) 11.2 (0.5) 8.9 (0.4) 10.3 (0.5) 15.1 (0.7)Dy 251.9 (8.9) 158.2 (5.6) 145.8 (5.2) 117.0 (4.2) 133.9 (4.8) 196.0 (6.9)Ho 99.1 (3.8) 60.7 (2.4) 56.7 (2.2) 49.5 (2.0) 54.2 (2.2) 76.0 (3.1)Er 451.8 (18.0) 271.1 (11.0) 260.3 (10.8) 235.5 (9.9) 235.3 (10.0) 336.3 (14.5)Tm 93.4 (3.3) 56.6 (2.0) 56.1 (2.0) 51.9 (1.8) 49.8 (1.8) 72.0 (2.6)Yb 880.8 (30.0) 511.5 (17.5) 523.1 (18.0) 494.4 (17.1) 455.3 (15.8) 666.4 (23.1)Lu 175.6 (6.5) 96.0 (3.6) 99.4 (3.8) 94.4 (3.6) 84.1 (3.3) 124.6 (4.9)Hf 10193 (323) 10382 (329) 10472 (332) 10611 (336) 10463 (331) 10905 (345)Th 1295 (45) 417 (15) 656 (23) 405 (14) 1230 (44) 1955 (71)U 2299 (85) 665 (25) 1268 (49) 891 (35) 2315 (91) 3018 (120)

n.d.

EPMA (wt.%)

LA-ICP-MS (ppm)

n.d.


n.d. n.d. n.d. n.d.

3 3 3 3 312-BT2-Z6 12-BT2-Z8 12-BT2-Z9 12-BT2-Z10 12-BT2-Z11 12-BT2-Z12

142

Table A4.2 Major and trace element data for Toba tuff zircon (n = number of EPMA analysis)

Samplen

SiO2 32.63 (0.11) 32.88 (0.11) 32.79 (0.11) 32.87 (0.11) 32.68 (0.11) 32.87 (0.11) 32.69 (0.11) 32.49 (0.11)

P2O5 0.09 (0.02) 0.04 (0.02) 0.11 (0.03) 0.08 (0.03) 0.09 (0.02) 0.07 (0.02) 0.11 (0.02) 0.18 (0.03)

ZrO2 64.76 (0.55) 64.74 (0.55) 64.87 (0.55) 65.14 (0.56) 64.77 (0.55) 65.57 (0.56) 65.04 (0.55) 64.11 (0.55)

HfO2 1.40 (0.03) 1.08 (0.03) 0.92 (0.03) 1.09 (0.03) 1.13 (0.03) 1.16 (0.03) 1.06 (0.03) 0.82 (0.03)

Y2O3 0.22 (0.04) 0.08 (0.04) 0.30 (0.04) 0.12 (0.04) 0.17 (0.04) 0.32 (0.04) 0.24 (0.04) 0.41 (0.04)

Total 99.14 (0.79) 98.86 (0.79) 99.03 (0.80) 99.33 (0.79) 98.90 (0.77) 100.05 (0.79) 99.17 (0.82) 98.01 (0.77)

Ti 14.6 (0.9) 13.6 (0.9) 12.9 (0.8) 13.5 (0.9) 15.1 (1.1) 10.1 (0.7) 10.4 (0.7) 8.6 (0.7)

Y 1646 (62) 553 (21) 772 (29) 546 (21) 1316 (50) 618 (24) 553 (21) 511 (20)La 0.09 (0.01) 0.07 (0.01) 0.02 (0.00) 0.20 (0.04) 0.17 (0.02) 0.27 (0.02) 0.36 (0.03)Ce 9.7 (0.4) 7.7 (0.3) 7.6 (0.3) 7.3 (0.3) 16.6 (0.8) 7.9 (0.4) 9.9 (0.5) 7.1 (0.3)Pr 0.05 (0.01) 0.04 (0.01) 0.20 (0.02) 0.04 (0.01) 0.36 (0.05) 0.06 (0.01) 0.10 (0.01) 0.18 (0.02)Nd 0.62 (0.08) 0.66 (0.08) 3.17 (0.21) 0.66 (0.10) 5.05 (0.49) 0.53 (0.06) 0.84 (0.07) 1.43 (0.13)Sm 2.14 (0.15) 1.38 (0.12) 6.00 (0.28) 1.62 (0.17) 8.37 (0.61) 1.25 (0.10) 1.26 (0.08) 1.94 (0.12)Eu 0.34 (0.03) 0.35 (0.03) 1.64 (0.07) 0.35 (0.04) 2.22 (0.16) 0.18 (0.02) 0.25 (0.02) 0.38 (0.02)Gd 17.0 (1.2) 8.4 (0.7) 30.2 (2.2) 8.7 (0.7) 38.1 (3.1) 8.2 (0.7) 7.9 (0.7) 11.0 (1.0)Tb 7.5 (0.7) 3.3 (0.3) 10.3 (1.0) 3.5 (0.3) 13.3 (1.3) 3.8 (0.4) 3.3 (0.4) 4.5 (0.5)Dy 103.5 (7.3) 45.5 (3.3) 116.7 (8.5) 47.2 (3.6) 159.7 (12.5) 52.0 (4.3) 45.3 (3.8) 57.7 (5.2)Ho 42.5 (3.8) 19.6 (1.8) 42.8 (3.9) 19.6 (1.9) 62.7 (6.2) 23.0 (2.4) 20.3 (2.2) 25.2 (2.9)Er 198.2 (8.1) 90.5 (3.8) 174.2 (7.3) 91.4 (4.0) 254.7 (11.2) 112.7 (5.1) 99.6 (4.6) 120.0 (5.7)Tm 44.0 (5.4) 21.0 (2.6) 33.9 (4.4) 21.2 (2.9) 54.7 (7.7) 25.2 (3.8) 23.7 (3.7) 27.1 (4.6)Yb 414.9 (34.6) 216.2 (18.4) 309.0 (26.8) 219.6 (20.0) 517.7 (48.5) 252.2 (25.0) 236.7 (24.2) 266.0 (29.1)Lu 80.9 (2.9) 45.1 (1.6) 60.7 (2.2) 44.4 (1.6) 99.0 (3.6) 53.8 (2.0) 51.5 (1.9) 56.4 (2.1)Hf 11832 (374) 9134 (289) 7820 (247) 9205 (291) 9609 (304) 9874 (312) 8984 (284) 6974 (221)Th 251 (21) 84 (7) 137 (12) 103 (9) 333 (30) 161 (16) 203 (20) 135 (14)U 296 (30) 87 (9) 93 (10) 111 (12) 217 (24) 238 (28) 221 (27) 161 (21)

LA-ICP-MS (ppm)

EPMA (wt.%)

TTZ1-5 TTZ1-6 TTZ1-8 TTZ1-9TTZ1-1 TTZ1-2 TTZ1-3 TTZ1-1112 2 2 2 2 1 1

n.d.

143

Samplen

SiO2 24.81 (0.09) 26.60 (0.10) 26.94 (0.10) 22.21 (0.09) 32.80 (0.11) 21.29 (0.09) 30.52 (0.10) 32.61 (0.11)

P2O5 0.07 (0.02) 0.08 (0.02) 0.03 (0.02) 0.02 (0.02) 0.08 (0.02) 0.30 (0.03) 0.08 (0.02) 0.27 (0.02)

ZrO2 50.59 (0.49) 57.30 (0.52) 55.72 (0.51) 53.39 (0.50) 65.06 (0.55) 44.66 (0.46) 60.09 (0.53) 64.48 (0.55)

HfO2 0.75 (0.03) 0.96 (0.03) 0.89 (0.03) 0.75 (0.03) 1.20 (0.03) 1.17 (0.03) 1.23 (0.03) 0.81 (0.03)

Y2O3 0.83 (0.05) 0.15 (0.04) 0.01 (0.03) 0.31 (0.04) 0.04 (0.03) 0.40 (0.04) 0.23 (0.04) 0.74 (0.05)

Total 77.05 (0.70) 85.09 (0.71) 83.63 (0.72) 76.70 (0.71) 99.28 (0.83) 67.82 (0.64) 92.17 (0.75) 98.96 (0.80)

Ti 8.5 (0.7) 9.2 (0.5) 9.4 (0.6) 7.2 (0.6) 12.3 (0.7) 9.8 (0.5) 9.5 (0.5) 7.5 (0.5)

Y 489 (19) 587 (23) 319 (13) 533 (21) 800 (32) 380 (15) 681 (27) 620 (25)La 0.01 (0.01) 0.09 (0.01) 0.02 (0.01) 0.77 (0.05) 0.51 (0.02) 0.04 (0.01) 0.30 (0.02) 0.02 (0.01)Ce 21.0 (1.0) 9.1 (0.4) 5.2 (0.2) 8.1 (0.4) 11.6 (0.5) 6.0 (0.2) 7.3 (0.3) 9.0 (0.4)Pr 0.10 (0.02) 0.05 (0.00) 0.02 (0.01) 0.24 (0.02) 0.19 (0.01) 0.04 (0.00) 0.11 (0.01) 0.08 (0.01)Nd 1.36 (0.18) 0.55 (0.05) 0.43 (0.07) 1.60 (0.18) 1.37 (0.08) 0.45 (0.04) 0.76 (0.05) 1.40 (0.09)Sm 3.14 (0.29) 1.21 (0.10) 0.70 (0.13) 2.35 (0.34) 2.25 (0.13) 0.86 (0.09) 1.51 (0.11) 2.41 (0.18)Eu 0.67 (0.06) 0.21 (0.02) 0.12 (0.03) 0.51 (0.08) 0.37 (0.03) 0.24 (0.03) 0.13 (0.02) 0.44 (0.04)Gd 14.5 (1.4) 8.2 (0.3) 4.8 (0.3) 10.7 (0.7) 11.8 (0.5) 5.5 (0.3) 9.3 (0.4) 11.4 (0.5)Tb 5.9 (0.7) 3.5 (0.1) 2.2 (0.1) 4.0 (0.2) 4.9 (0.2) 2.3 (0.1) 4.3 (0.2) 4.6 (0.2)Dy 77.5 (7.2) 47.2 (1.6) 26.9 (1.1) 46.9 (1.9) 65.0 (2.2) 30.1 (1.1) 56.4 (2.0) 54.4 (1.9)Ho 31.3 (3.8) 20.7 (0.7) 11.6 (0.4) 18.7 (0.7) 28.2 (0.9) 13.4 (0.5) 24.5 (0.8) 23.5 (0.8)Er 149.3 (7.3) 97.5 (3.2) 54.2 (1.9) 87.2 (3.1) 135.9 (4.4) 65.6 (2.2) 117.6 (3.8) 107.2 (3.5)Tm 36.0 (6.3) 23.3 (0.8) 12.4 (0.4) 19.8 (0.7) 32.8 (1.1) 15.9 (0.5) 27.1 (0.9) 25.1 (0.9)Yb 331.8 (37.6) 228.9 (8.0) 123.0 (4.4) 188.4 (6.9) 327.1 (11.4) 158.0 (5.6) 262.9 (9.3) 243.9 (8.9)Lu 68.9 (2.7) 48.2 (1.6) 24.2 (0.8) 38.9 (1.4) 70.4 (2.3) 34.6 (1.1) 53.9 (1.8) 51.3 (1.7)Hf 6348 (201) 8116 (257) 7560 (239) 6391 (203) 10140 (321) 9962 (315) 10416 (329) 6833 (216)Th 387 (42) 191 (7) 69 (2) 102 (4) 159 (5) 144 (5) 181 (6) 206 (7)U 227 (31) 242 (9) 78 (3) 82 (3) 188 (7) 146 (6) 258 (10) 147 (6)

EPMA (wt.%)

LA-ICP-MS (ppm)

TTZ1-17 TTZ1-19 TTZ1-20 TTZ1-23TTZ1-12 TTZ1-13 TTZ1-14 TTZ1-161 1 1 11 1 1 1

144

Samplen

SiO2 32.94 (0.11) 32.54 (0.11) 32.62 (0.11) 32.90 (0.11) 24.35 (0.09) 28.00 (0.10) 30.53 (0.10) 25.04 (0.09)

P2O5 0.04 (0.02) 0.07 (0.02) 0.06 (0.02) 0.07 (0.02) 0.03 (0.02) 0.04 (0.02) 0.12 (0.02) 0.01 (0.02)

ZrO2 65.48 (0.56) 64.48 (0.55) 65.41 (0.56) 64.79 (0.55) 50.86 (0.49) 58.36 (0.53) 61.99 (0.54) 56.55 (0.52)

HfO2 1.12 (0.03) 1.13 (0.03) 0.99 (0.03) 1.36 (0.03) 1.11 (0.03) 0.98 (0.03) 1.03 (0.03) 0.97 (0.03)

Y2O3 0.14 (0.04) 0.19 (0.04) 0.11 (0.04) 0.19 (0.04) 0.17 (0.04) 0.21 (0.03) 0.20 (0.04) 0.08 (0.03)

Total 99.77 (0.78) 98.48 (0.79) 99.25 (0.79) 99.41 (0.82) 76.55 (0.69) 87.65 (0.74) 93.90 (0.79) 82.68 (0.71)

Ti 9.7 (0.5) 8.9 (0.5) 8.9 (0.6) 10.4 (0.6) 15.1 (0.9) 11.4 (0.6) 10.6 (0.5) 10.5 (0.5)

Y 478 (19) 1011 (41) 357 (14) 1125 (46) 1756 (72) 1082 (41) 686 (26) 641 (25)La 0.28 (0.01) 0.30 (0.02) 0.17 (0.02) 0.11 (0.01) 0.02 (0.01)Ce 6.3 (0.3) 8.4 (0.4) 6.5 (0.3) 12.0 (0.5) 21.1 (0.9) 11.3 (0.4) 10.9 (0.4) 9.5 (0.3)Pr 0.09 (0.01) 0.06 (0.01) 0.02 (0.00) 0.15 (0.01) 0.37 (0.02) 0.26 (0.02) 0.04 (0.01) 0.05 (0.00)Nd 0.64 (0.04) 1.06 (0.06) 0.38 (0.05) 1.44 (0.11) 5.98 (0.28) 4.01 (0.21) 0.70 (0.07) 0.98 (0.06)Sm 0.99 (0.08) 2.75 (0.16) 0.81 (0.10) 3.09 (0.24) 10.07 (0.51) 6.68 (0.39) 1.78 (0.17) 1.86 (0.13)Eu 0.15 (0.02) 0.45 (0.03) 0.12 (0.02) 0.33 (0.04) 2.74 (0.13) 1.27 (0.08) 0.31 (0.04) 0.39 (0.03)Gd 6.2 (0.3) 16.4 (0.6) 5.3 (0.3) 16.9 (0.7) 49.7 (1.8) 29.9 (1.2) 11.1 (0.5) 10.7 (0.4)Tb 3.0 (0.1) 7.0 (0.2) 2.2 (0.1) 7.4 (0.3) 17.5 (0.6) 10.4 (0.4) 4.5 (0.2) 4.4 (0.2)Dy 39.8 (1.4) 93.0 (3.3) 30.1 (1.1) 93.4 (3.4) 199.0 (7.1) 116.1 (3.9) 56.3 (1.9) 55.0 (1.8)Ho 17.3 (0.6) 39.1 (1.3) 12.8 (0.4) 39.6 (1.3) 76.1 (2.5) 45.6 (1.5) 25.0 (0.8) 23.2 (0.8)Er 82.7 (2.7) 175.5 (5.7) 61.0 (2.0) 185.0 (6.0) 325.0 (10.6) 190.6 (6.2) 117.3 (3.8) 108.6 (3.5)Tm 19.9 (0.7) 39.4 (1.4) 14.8 (0.5) 42.2 (1.5) 68.4 (2.4) 39.5 (1.3) 27.4 (0.9) 25.5 (0.8)Yb 196.7 (7.2) 370.2 (13.6) 152.0 (5.7) 408.3 (15.3) 622.4 (23.5) 365.4 (12.1) 267.8 (8.9) 248.7 (8.2)Lu 40.2 (1.3) 74.8 (2.5) 31.6 (1.1) 84.2 (2.8) 121.5 (4.1) 71.7 (2.3) 56.2 (1.8) 53.1 (1.7)Hf 9492 (300) 9611 (304) 8390 (265) 11508 (364) 9433 (298) 8283 (262) 8770 (278) 8220 (260)Th 118 (4) 228 (8) 104 (4) 340 (12) 376 (13) 233 (8) 120 (4) 119 (4)U 170 (7) 339 (14) 120 (5) 366 (15) 216 (9) 138 (5) 134 (5) 115 (4)

EPMA (wt.%)

LA-ICP-MS (ppm)

TTZ1-30 TTZ1-31TTZ1-24 TTZ1-25 TTZ1-26 TTZ1-27 TTZ1-28 TTZ1-291 1 11 1 1 1 1

n.d. n.d. n.d.

145

Samplen

SiO2 20.21 (0.08) 25.56 (0.09) 32.73 (0.11) 32.88 (0.11) 19.44 (0.08) 25.48 (0.09) 24.36 (0.09) 28.77 (0.10)

P2O5 0.01 (0.02) 0.06 (0.02) 0.02 (0.02) 0.08 (0.02) 0.03 (0.03) 0.01 (0.02) 0.02 (0.02) 0.06 (0.02)

ZrO2 47.76 (0.47) 56.58 (0.51) 65.61 (0.56) 65.48 (0.56) 45.43 (0.46) 58.89 (0.52) 55.45 (0.51) 60.37 (0.53)

HfO2 0.87 (0.03) 0.66 (0.03) 1.47 (0.03) 1.04 (0.03) 0.95 (0.03) 0.92 (0.03) 0.85 (0.03) 1.00 (0.03)

Y2O3 0.11 (0.03) 0.39 (0.04) 0.05 (0.03) 0.13 (0.04) 0.13 (0.03) 0.17 (0.04) 0.16 (0.04) 0.18 (0.04)

Total 68.96 (0.64) 83.28 (0.76) 99.96 (0.78) 99.67 (0.79) 66.01 (0.56) 85.47 (0.70) 80.87 (0.70) 90.40 (0.78)

Ti 8.5 (0.4) 7.1 (0.3) 13.6 (0.7) 10.6 (0.5) 9.5 (0.6) 9.8 (0.5) 10.8 (0.6) 8.9 (0.7)

Y 801 (31) 1364 (53) 1434 (55) 902 (35) 779 (31) 818 (32) 876 (35) 506 (20)La 0.10 (0.01) 0.06 (0.01) 0.08 (0.01) 0.03 (0.01) 0.07 (0.01) 0.05 (0.01) 0.62 (0.09)Ce 20.4 (0.7) 41.9 (1.4) 13.9 (0.5) 9.0 (0.3) 10.5 (0.4) 7.9 (0.3) 10.0 (0.4) 6.6 (0.4)Pr 0.09 (0.01) 0.12 (0.01) 0.14 (0.01) 0.16 (0.01) 0.07 (0.01) 0.09 (0.01) 0.25 (0.02) 0.22 (0.05)Nd 1.29 (0.07) 1.87 (0.08) 2.34 (0.14) 2.57 (0.14) 1.39 (0.13) 1.40 (0.08) 4.61 (0.27) 1.27 (0.27)Sm 2.78 (0.17) 3.93 (0.18) 4.79 (0.32) 4.39 (0.26) 3.30 (0.31) 2.98 (0.19) 5.71 (0.43) 1.31 (0.30)Eu 0.49 (0.03) 0.70 (0.04) 0.87 (0.07) 1.11 (0.07) 0.61 (0.07) 0.61 (0.04) 1.74 (0.13) 0.15 (0.06)Gd 15.2 (0.6) 22.3 (0.8) 26.7 (1.1) 20.9 (0.8) 19.9 (0.9) 16.3 (0.6) 25.6 (1.2) 6.7 (0.7)Tb 6.3 (0.2) 8.9 (0.3) 10.6 (0.4) 7.4 (0.3) 8.4 (0.3) 6.3 (0.2) 8.7 (0.3) 3.0 (0.2)Dy 74.4 (2.5) 108.4 (3.6) 134.6 (4.5) 86.1 (2.9) 106.2 (3.7) 77.7 (2.6) 92.7 (3.3) 36.1 (1.6)Ho 30.9 (1.0) 43.4 (1.4) 55.9 (1.8) 33.2 (1.1) 45.2 (1.5) 32.1 (1.1) 35.1 (1.2) 16.1 (0.7)Er 141.7 (4.6) 198.1 (6.4) 251.6 (8.2) 147.8 (4.8) 215.7 (7.1) 137.9 (4.5) 149.7 (5.0) 78.9 (2.9)Tm 33.3 (1.1) 45.3 (1.5) 56.1 (1.9) 33.2 (1.1) 51.9 (1.8) 31.1 (1.0) 31.3 (1.1) 19.7 (0.8)Yb 318.2 (10.6) 435.5 (14.5) 532.7 (17.8) 323.0 (10.9) 510.1 (17.4) 296.6 (10.1) 285.2 (10.0) 188.9 (6.8)Lu 66.4 (2.1) 85.1 (2.7) 109.7 (3.6) 65.7 (2.1) 108.2 (3.5) 56.7 (1.8) 55.4 (1.8) 39.6 (1.4)Hf 7347 (232) 5601 (177) 12492 (395) 8857 (280) 8023 (254) 7815 (247) 7250 (230) 8471 (269)Th 672 (24) 1572 (57) 351 (13) 195 (7) 332 (12) 160 (6) 153 (6) 95 (4)U 345 (13) 482 (19) 332 (13) 148 (6) 283 (12) 175 (7) 86 (4) 143 (5)

LA-ICP-MS (ppm)

EPMA (wt.%)

TTZ1-36 TTZ1-39 TTZ1-40TTZ1-34 TTZ1-35 TTZ1-42 TTZ1-44 TTZ1-451 1 21

n.d.

1 1 11

146

Samplen

SiO2 32.75 (0.11) 32.78 (0.11) 32.66 (0.11) 32.18 (0.10) 32.36 (0.10) 32.62 (0.11) 32.58 (0.11) 33.20 (0.11)

P2O5 0.08 (0.02) 0.04 (0.03) 0.08 (0.02) 0.05 (0.02) 0.08 (0.02) 0.08 (0.02) 0.11 (0.02) 0.08 (0.02)

ZrO2 65.80 (0.56) 66.14 (0.56) 65.39 (0.55) 65.38 (0.55) 64.41 (0.55) 65.75 (0.56) 65.02 (0.56) 64.81 (0.55)

HfO2 1.07 (0.03) 1.27 (0.03) 1.39 (0.03) 0.77 (0.03) 1.08 (0.03) 1.04 (0.03) 1.09 (0.03) 1.21 (0.03)

Y2O3 0.22 (0.04) 0.04 (0.04) 0.24 (0.04) 0.34 (0.04) 0.17 (0.04) 0.17 (0.04) 0.22 (0.04)

Total 99.96 (0.78) 100.31 (0.77) 99.82 (0.82) 98.76 (0.78) 98.15 (0.77) 99.70 (0.79) 99.13 (0.79) 99.46 (0.75)

Ti 12.5 (0.6) 11.0 (0.5) 13.2 (0.7) 7.2 (0.3) 14.4 (0.7) 11.0 (0.7) 8.3 (0.5) 18.1 (1.0)

Y 1077 (41) 222 (9) 1314 (51) 297 (12) 963 (37) 637 (25) 379 (15) 1713 (68)La 0.11 (0.02) 0.08 (0.02) 0.04 (0.01) 0.11 (0.03)Ce 11.4 (0.4) 4.9 (0.2) 16.5 (0.6) 3.7 (0.1) 8.6 (0.3) 9.6 (0.4) 4.5 (0.2) 18.3 (0.7)Pr 0.17 (0.02) 0.12 (0.02) 0.02 (0.01) 0.12 (0.02) 0.06 (0.02) 0.03 (0.01) 0.27 (0.04)Nd 2.60 (0.17) 0.19 (0.05) 1.94 (0.23) 0.27 (0.05) 2.35 (0.19) 0.75 (0.16) 0.53 (0.11) 4.12 (0.35)Sm 4.72 (0.27) 0.45 (0.07) 4.68 (0.39) 0.57 (0.07) 4.93 (0.31) 1.95 (0.28) 0.80 (0.16) 8.69 (0.59)Eu 0.96 (0.06) 0.07 (0.02) 0.61 (0.07) 0.19 (0.02) 1.19 (0.08) 0.49 (0.07) 0.12 (0.04) 1.61 (0.13)Gd 23.5 (0.9) 3.1 (0.2) 24.2 (1.1) 4.3 (0.2) 23.1 (0.9) 11.7 (0.7) 6.6 (0.5) 36.2 (1.5)Tb 8.6 (0.3) 1.2 (0.1) 9.8 (0.4) 1.8 (0.1) 8.2 (0.3) 4.4 (0.2) 2.7 (0.1) 13.9 (0.6)Dy 100.4 (3.3) 17.0 (0.6) 120.5 (4.2) 24.3 (0.9) 98.7 (3.3) 53.3 (2.0) 34.5 (1.3) 154.1 (5.3)Ho 39.9 (1.5) 7.7 (0.3) 49.4 (2.0) 10.6 (0.4) 38.5 (1.6) 22.6 (1.0) 14.0 (0.6) 60.5 (2.7)Er 174.0 (5.6) 37.5 (1.3) 224.5 (7.3) 51.3 (1.7) 163.6 (5.3) 109.4 (3.7) 67.3 (2.3) 267.6 (8.7)Tm 38.0 (1.3) 9.2 (0.3) 48.3 (1.6) 12.5 (0.4) 35.1 (1.2) 25.7 (0.9) 15.3 (0.6) 60.7 (2.1)Yb 359.6 (11.8) 91.9 (3.1) 473.9 (15.8) 125.6 (4.2) 324.1 (10.8) 259.4 (8.9) 153.0 (5.3) 560.6 (18.9)Lu 72.0 (2.3) 19.2 (0.6) 91.0 (3.0) 28.5 (0.9) 63.9 (2.1) 56.4 (1.9) 31.2 (1.1) 115.7 (3.8)Hf 9113 (288) 10747 (340) 11758 (372) 6488 (205) 9146 (289) 8802 (279) 9244 (293) 10255 (325)Th 207 (8) 48 (2) 382 (14) 61 (2) 124 (5) 130 (5) 71 (3) 347 (14)U 164 (6) 72 (3) 334 (12) 73 (3) 88 (3) 121 (5) 110 (4) 235 (9)

LA-ICP-MS (ppm)

EPMA (wt.%)

TTZ1-56 TTZ1-57 TTZ1-58TTZ1-49 TTZ1-50 TTZ1-52 TTZ1-53TTZ1-48

n.d.

n.d. n.d. n.d.

n.d.

n.d.

1 1 11 2 1 1 1

147

Table A4.3 Major and trace element data for Umiakovik pluton zircon (n = number of EPMA analysis).

Samplen

SiO2 32.67 (0.11) 32.72 (0.13) 32.71 (0.10) 32.85 (0.13) 32.92 (0.11) 32.74 (0.08) 32.76 (0.06) 32.81 (0.05)

P2O5 0.023 (0.017) 0.024 (0.014) 0.028 (0.022) 0.043 (0.017) 0.022 (0.015) 0.048 (0.006) 0.037 (0.019) 0.040 (0.011)

ZrO2 66.30 (0.57) 65.44 (0.73) 65.84 (0.28) 66.06 (0.61) 65.71 (0.82) 65.75 (0.27) 65.96 (0.64) 65.73 (0.27)

HfO2 1.01 (0.06) 0.98 (0.03) 0.98 (0.10) 0.99 (0.07) 1.11 (0.12) 1.02 (0.04) 1.01 (0.04) 1.03 (0.06)

Y2O3 0.105 (0.103) 0.099 (0.051) 0.149 (0.097) 0.079 (0.015) 0.098 (0.026) 0.049 (0.026) 0.067 (0.006) 0.221 (0.084)

Total 100.16 (0.55) 99.30 (0.78) 99.75 (0.29) 100.10 (0.55) 99.89 (0.77) 99.63 (0.29) 99.87 (0.71) 99.85 (0.31)

Ti 27.0 (1.5) 23.5 (1.7) 26.3 (1.6) 22.5 (1.5) 28.3 (1.1) 25.3 (1.0) 25.9 (1.1) 23.4 (1.0)

Y 299 (10) 1647 (57) 1587 (55) 508 (18) 364 (13) 473 (17) 575 (21) 1756 (64)La 0.39 (0.15) 0.39 (0.17) 0.43 (0.16) 0.21 (0.08)Ce 1.6 (0.2) 7.1 (0.5) 7.1 (0.5) 4.1 (0.4) 2.5 (0.2) 3.2 (0.2) 2.3 (0.2) 8.1 (0.5)Pr 0.61 (0.11) 0.55 (0.11) 0.14 (0.04) 0.85 (0.06)Nd 9.70 (1.31) 8.18 (1.07) 2.18 (0.71) 1.37 (0.30) 2.04 (0.35) 10.04 (0.76)Sm 14.15 (1.87) 11.72 (1.54) 1.79 (0.54) 2.09 (0.59) 14.67 (1.06)Eu 0.82 (0.32) 1.22 (0.29) 0.50 (0.14)Gd 6.6 (1.3) 58.2 (3.8) 53.5 (3.1) 12.0 (1.7) 8.1 (0.8) 10.0 (0.9) 15.4 (1.0) 62.5 (2.8)Tb 2.3 (0.2) 17.8 (1.0) 15.7 (0.8) 4.3 (0.3) 3.0 (0.2) 3.6 (0.2) 4.8 (0.3) 17.8 (0.8)Dy 27.2 (1.7) 186.7 (8.3) 164.8 (6.8) 46.7 (2.7) 34.8 (1.6) 40.3 (1.8) 54.2 (2.4) 188.8 (7.7)Ho 10.5 (0.6) 58.3 (2.7) 56.9 (2.4) 17.9 (1.0) 12.5 (0.6) 16.1 (0.7) 20.4 (0.9) 65.2 (2.9)Er 44.7 (2.3) 232.2 (10.5) 220.5 (9.6) 77.0 (3.9) 55.6 (2.6) 69.6 (3.2) 84.8 (3.9) 250.2 (11.5)Tm 9.1 (0.5) 43.9 (2.0) 43.1 (1.8) 16.0 (0.9) 11.7 (0.5) 15.0 (0.6) 16.9 (0.7) 48.0 (2.0)Yb 90.6 (4.2) 375.2 (15.6) 355.9 (14.0) 141.5 (6.6) 108.4 (4.4) 131.7 (5.3) 147.6 (6.0) 389.9 (15.4)Lu 16.1 (0.8) 60.6 (2.5) 60.5 (2.2) 26.7 (1.2) 19.8 (0.8) 24.3 (0.9) 27.6 (1.0) 64.6 (2.3)Hf 8606 (274) 8351 (268) 8328 (266) 8436 (270) 9372 (297) 8623 (274) 8550 (271) 8770 (278)Th 10 (1) 73 (3) 66 (3) 47 (2) 12 (1) 25 (1) 23 (1) 98 (4)U 11 (1) 43 (2) 40 (2) 32 (1) 13 (1) 21 (1) 19 (1) 55 (2)

EPMA (wt.%)

LA-ICP-MS (ppm)

12-UB-Z1 12-UB-Z83 3 3 3 3 3 3

12-UB-Z2 12-UB-Z3 12-UB-Z4 12-UB-Z5 12-UB-Z6 12-UB-Z73

n.d. n.d. n.d. n.d.

n.d. n.d.n.d. n.d. n.d. n.d.

n.d. n.d. n.d.n.d. n.d. n.d. n.d. n.d.

148

Samplen

SiO2 32.79 (0.16) 32.92 (0.06) 32.82 (0.17) 32.65 (0.11) 32.67 (0.06) 32.85 (0.02) 32.91 (0.03) 32.91 (0.06)

P2O5 0.023 (0.020) 0.029 (0.021) 0.046 (0.030) 0.028 (0.010) 0.026 (0.013) 0.038 (0.024)

ZrO2 65.98 (0.34) 66.05 (0.60) 66.20 (0.27) 65.96 (0.26) 66.12 (0.05) 65.84 (0.50) 66.53 (0.26) 66.28 (0.57)

HfO2 1.03 (0.02) 1.05 (0.01) 0.99 (0.12) 1.02 (0.11) 1.08 (0.07) 1.03 (0.01) 1.09 (0.07) 1.02 (0.05)

Y2O3 0.046 (0.031) 0.096 (0.035) 0.164 (0.172) 0.106 (0.109) 0.036 (0.015) 0.051 (0.009) 0.039 (0.023) 0.091 (0.099)

Total 99.92 (0.36) 100.18 (0.63) 100.27 (0.36) 99.83 (0.16) 99.98 (0.06) 99.83 (0.45) 100.65 (0.33) 100.36 (0.57)

Ti 23.6 (1.0) 26.9 (1.1) 26.3 (1.1) 28.3 (1.5) 26.5 (1.3) 25.9 (1.2) 22.4 (1.0) 26.7 (1.3)

Y 344 (13) 368 (14) 1308 (50) 1545 (60) 293 (10) 376 (13) 269 (9) 1639 (58)La 0.21 (0.07) 0.23 (0.08) 0.33 (0.13) 0.18 (0.07)Ce 2.3 (0.2) 2.3 (0.2) 5.8 (0.4) 6.2 (0.5) 2.5 (0.2) 2.3 (0.2) 2.3 (0.1) 5.9 (0.3)Pr 0.54 (0.06) 0.54 (0.71) 0.58 (0.06)Nd 1.24 (0.28) 9.50 (0.65) 7.69 (1.00) 0.90 (0.37) 1.49 (0.36) 8.71 (0.88)Sm 1.44 (0.50) 12.07 (0.88) 12.44 (1.35) 2.51 (0.58) 1.87 (0.55) 10.06 (1.05)Eu 0.68 (0.17) 0.96 (0.19)Gd 6.6 (0.8) 7.4 (0.9) 42.0 (2.1) 55.5 (3.3) 7.0 (1.0) 10.1 (1.0) 4.7 (0.7) 53.6 (3.0)Tb 2.7 (0.2) 2.8 (0.2) 13.1 (0.6) 15.9 (0.9) 2.3 (0.2) 2.7 (0.2) 2.0 (0.1) 15.6 (0.7)Dy 31.0 (1.5) 34.1 (1.7) 145.2 (6.2) 163.8 (7.6) 26.7 (1.4) 32.6 (1.5) 22.0 (1.0) 169.3 (6.5)Ho 12.2 (0.6) 13.0 (0.7) 49.6 (2.3) 57.4 (2.9) 10.8 (0.5) 12.9 (0.6) 9.5 (0.4) 58.7 (2.3)Er 56.1 (2.7) 58.0 (3.0) 193.3 (9.5) 217.0 (11.2) 47.7 (2.0) 55.2 (2.2) 39.9 (1.6) 230.4 (8.3)Tm 10.8 (0.5) 12.7 (0.6) 37.3 (1.6) 44.6 (2.1) 10.1 (0.5) 10.7 (0.5) 8.6 (0.4) 42.0 (1.6)Yb 97.9 (4.1) 116.4 (5.1) 318.9 (13.1) 375.9 (16.3) 90.5 (3.7) 112.2 (4.4) 82.9 (3.2) 360.6 (13.2)Lu 19.2 (0.7) 21.5 (0.9) 51.6 (1.8) 60.9 (2.4) 15.1 (0.7) 20.0 (0.8) 15.7 (0.6) 66.2 (2.5)Hf 8771 (278) 8902 (283) 8407 (267) 8678 (277) 9186 (292) 8705 (277) 9229 (293) 8634 (275)Th 13 (1) 12 (1) 55 (2) 62 (3) 16 (1) 14 (1) 22 (1) 65 (2)U 14 (1) 12 (1) 37 (2) 41 (2) 19 (1) 14 (1) 27 (1) 39 (2)

12-UB-Z1612-UB-Z9 12-UB-Z10 12-UB-Z11 12-UB-Z12 12-UB-Z13 12-UB-Z14 12-UB-Z153 3 3 3 3

n.d. n.d.

n.d.

3 3 3EPMA (wt.%)

LA-ICP-MS (ppm)

n.d. n.d.n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d.

n.d. n.d. n.d. n.d. n.d. n.d.n.d. n.d.

149

Samplen

SiO2 32.79 (0.14) 32.73 (0.13) 32.95 (0.07) 32.71 (0.16) 32.84 (0.12) 32.95 (0.13)

P2O5 0.054 (0.019) 0.038 (0.036) 0.026 (0.018) 0.010 (0.004) 0.024 (0.008) 0.042 (0.020)

ZrO2 65.73 (0.47) 66.10 (0.46) 65.92 (0.75) 66.43 (0.44) 66.22 (0.83) 66.27 (0.68)

HfO2 1.00 (0.08) 1.05 (0.03) 1.04 (0.08) 1.16 (0.05) 1.00 (0.09) 1.04 (0.07)

Y2O3 0.174 (0.092) 0.081 (0.046) 0.056 (0.015) 0.033 (0.021) 0.076 (0.079) 0.141 (0.093)

Total 99.77 (0.64) 100.05 (0.52) 100.03 (0.73) 100.39 (0.34) 100.20 (0.83) 100.47 (0.81)

Ti 24.2 (1.2) 24.3 (1.2) 23.6 (1.1) 14.4 (0.6) 13.7 (0.6) 14.4 (0.7)

Y 1095 (39) 354 (13) 392 (14) 203 (7) 214 (8) 336 (13)La 0.23 (0.09) 0.05 (0.02)Ce 4.9 (0.3) 2.5 (0.2) 2.6 (0.2) 1.6 (0.1) 1.4 (0.1) 1.8 (0.1)Pr 0.42 (0.06) 0.16 (0.04) 0.08 (0.01) 0.06 (0.01)Nd 5.35 (0.61) 1.02 (0.32) 0.87 (0.32) 0.41 (0.09) 0.61 (0.09) 0.88 (0.11)Sm 7.75 (0.88) 1.50 (0.52) 3.09 (0.49) 0.96 (0.14) 0.84 (0.14) 1.46 (0.19)Eu 0.51 (0.18) 0.12 (0.04) 0.28 (0.04)Gd 31.2 (1.9) 6.7 (0.9) 9.4 (0.8) 4.0 (0.3) 4.3 (0.3) 7.5 (0.4)Tb 10.3 (0.5) 2.5 (0.2) 2.7 (0.2) 1.5 (0.1) 1.6 (0.1) 2.7 (0.1)Dy 110.5 (4.2) 32.0 (1.4) 33.1 (1.5) 18.3 (0.7) 19.2 (0.7) 30.8 (1.2)Ho 40.3 (1.5) 12.2 (0.5) 13.4 (0.6) 7.4 (0.3) 7.4 (0.3) 11.6 (0.5)Er 156.4 (5.6) 51.2 (2.0) 59.4 (2.3) 31.6 (1.1) 32.4 (1.2) 50.7 (1.8)Tm 29.7 (1.1) 10.6 (0.5) 13.1 (0.5) 6.4 (0.2) 6.4 (0.2) 10.1 (0.4)Yb 244.6 (8.8) 99.4 (3.8) 114.8 (4.4) 61.8 (2.2) 59.4 (2.1) 90.0 (3.2)Lu 45.0 (1.7) 19.1 (0.8) 21.3 (0.8) 11.8 (0.4) 11.5 (0.4) 17.5 (0.6)Hf 8468 (269) 8942 (284) 8830 (280) 9857 (312) 8480 (268) 8790 (278)Th 47 (2) 18 (1) 23 (1) 12 (0) 9 (0) 15 (1)U 29 (1) 18 (1) 26 (1) 13 (0) 8 (0) 12 (0)

12-UB-Z17 12-UB-Z18 12-UB-Z193 3 4

n.d.n.d. n.d. n.d.

n.d. n.d. n.d.

n.d. n.d.

3 3 312-UB-Z20 12-UB-Z27 12-UB-Z48

EPMA (wt.%)

LA-ICP-MS (ppm)

150

A4.2 Glass/Whole Rock Compositions

Table A4.4 Major and trace element compositions for Bishop tuff and Toba tuff glasses and

whole rock analysis of Umiakovik pluton. Major element analyses for glasses measured by

EPMA, trace elements by LA-ICP-MS. Umiakovik major elements measured by x-ray fluoresce,

trace elements by solution ICP-MS. n = number of analysis (EPMA on right).

Samplen

SiO2 73.12 (1.81) 76.49 (1.03) 59.63 (0.12)

TiO2 0.02 (0.14) 0.04 (0.10) 1.48 (0.01)

Al2O5 12.75 (1.15) 12.39 (0.68) 11.61 (0.07)

MgO 0.30 (0.02) 0.46 (0.00)

MnO 0.02 (0.03) 0.02 (0.07) 0.19 (0.00)

FeO 1.39 (0.06) 0.30 (0.16) 15.13 (0.08)CaO 1.11 (0.17) 0.46 (0.19) 5.93 (0.02)

Na2O 3.49 (0.14) 2.57 (0.28) 2.51 (0.01)

K2O 4.31 (0.25) 6.62 (0.96) 2.04 (0.01)

P2O5 0.10 (0.05) 0.01 (0.09) 0.49 (0.49)

Total 96.69 (1.35) 98.92 (1.14) 100.50 (0.32)

Y 21.95 (1.86) 26.08 (2.06) 50 (0.25)Zr 47.69 (4.00) 56.29 (4.81) 572 (13)La 24.59 (1.99) 33.37 (2.57) 46.5 (12.5)Ce 49.26 (3.46) 63.00 (4.70) 106.0 0.1Pr 5.18 (0.32) 6.13 (0.24) 14.20 (0.00)Nd 18.72 (1.58) 20.15 (0.73) 65.00 (2.18)Sm 3.94 (0.24) 3.68 (0.45) 14.00 (3.55)Eu 0.24 (0.04) 0.50 (0.08) 3.83 (2.41)Gd 3.43 (0.36) 3.28 (0.34) 12.80 (3.32)Tb 0.61 (0.07) 0.57 (0.05) 1.96 (1.11)Dy 3.57 (0.32) 3.73 (0.37) 11.20 (3.17)Ho 0.75 (0.04) 0.87 (0.06) 2.16 (1.31)Er 2.14 (0.17) 2.51 (0.25) 6.06 (0.35)Tm 0.34 (0.03) 0.39 (0.05) 0.894 (0.633)Yb 2.47 (0.22) 2.98 (0.41) 6.07 (1.28)Lu 0.39 (0.05) 0.49 (0.05) 1.06 (1.33)Hf 3.03 (0.27) 2.95 (0.34) 11.90 (0.59)Th 46.89 (1.75) 26.37 (1.52) 1.47 (0.03)U 28.98 (1.07) 4.87 (0.57) 0.17 (0.02)

EPMA (wt.%)

LA-ICP-MS (ppm)

n.d.

Bishop Toba Umiakovik4,9 23,8 2

151

A4.3 Derivation of Melt Ce4+/Ce3+ from Zircon/Melt Partitioning

The concentrations of Ce in zircon and melt will be the sum of the concentrations of Ce4+ and

Ce3+. This can be expressed as,

3 4Ce Ce Cezircon zircon zircon (A4.1)

3 4Ce Ce Cemelt melt melt (A4.2)

Using partition coefficients for the individual oxidation states and substituting into Equation

A4.1 yields,

3 4

3 / 4 /

Ce CeCe Ce Cezircon melt zircon melt

zircon melt meltD D (A4.3)

Combining Equations A4.2 and A4.3 we can then solve for Ce4+.

3 4

3 3 4

3 4

4 / 4 /

Ce Ce

/ 4 / 4 /

Ce Ce Ce

/ 4 /

Ce Ce

Ce ( Ce Ce ) Ce

Ce Ce Ce

Ce Ce (

zircon melt zircon meltzircon melt melt melt

zircon melt zircon melt zircon meltmelt melt melt

zircon melt zircon meltmelt melt

D D

D D D

D D

3

/

Ce)zircon meltD

3 4 3

/ 4 / /

Ce Ce CeCe Ce Ce ( )zircon melt zircon melt zircon melt

zircon melt meltD D D

3

4 3

/4 Ce

/ /

Ce Ce

Ce CeCe



D

D D

(A4.4)

Substitution of Equation A4.4 in to A4.2 gives us,

3

4 3

/3 Ce

/ /

Ce Ce

Ce CeCe Ce



D

D D

This can be rearranged to yield,

3

4 3

/3 Ce

/ /

Ce Ce

Ce CeCe Ce



D

D D

(A4.5)

152

Dividing Equation A4.5 by Equation A4.4 then gives us the Ce4+/Ce3+ ratio of the melt.

3

4 3

3

4 3

4 3

/

Ce/ /3

Ce Ce/4

Ce/ /

Ce Ce

/ /

Ce Ce

Ce CeCe

Ce

Ce CeCe

Ce (



meltzircon melt

zircon meltmeltzircon melt zircon melt

zircon melt zircon melmelt

D

D D

D

D D

D D

3

4 3

3

/

Ce

/ /

Ce Ce/

Ce

)1

Ce Ce

Ce Ce1

Ce Ce

t


zircon melt zircon meltmelt melt


D

D D

D

3 4 3

3

3/ / /

4Ce Ce Ce

/

Ce

Ce( Ce Ce ) Ce Ce

Ce

Ce Ce

zircon melt zircon melt zircon meltmeltzircon melt melt melt

melt


D D D

D

4

3

/3Ce

4 /

Ce

Ce CeCe

Ce Ce Ce

zircon meltmelt zirconmelt

zircon meltmelt zircon melt

D

D

3

4

/4Ce

3 /

Ce

Ce CeCe

Ce Ce Ce

zircon meltzircon meltmelt

zircon meltmelt melt zircon

D

D

(A4.6)

This is now in a form where terms on the right hand side of the equation can be determined,

allowing for the Ce4+/Ce3+ ratio of the melt to be calculated.

153

A4.4 Evaluation of Hadean Melt Trace Element Budget

Approximation of the trace element composition for the Hadean parent melt of the Jack Hills

zircons was done through calculation of theoretical partition coefficients ( /zircon meltiD ) which were

then compared to the /zircon meltiD values of Sano et al. (2002). Partition coefficients were

calculated using compositions for the >4.0 Ga Jack Hills zircons and values for primitive mantle,

modern bulk continental crust, and Archean TTG (shown in Figures A4.4, A4.5, and A4.6,

respectively. The calculated partition coefficients best reproduce those of Sano et al. (2002)

when assuming a TTG source. This allows for approximation of the Hf, Th, and U budget of the

zircon crystalizing Hadean melts.

Figure A4.1 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a pyrolite

(McDonough and Sun, 1995) parent melt composition (grey lines). Also plotted are partition

coefficients from Sano et al. (2002, black diamonds). REE are plotted in order of increasing

atomic number.

154

Figure A4.2 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a trace element

parent melt composition of bulk continental crust (Rudnick and Gao, 2003). Symbols as in

Figure A4.4.

155

Figure A4.3 Plot of log /zircon meltiD calculated for the Jack Hills zircons assuming a trace element

parent melt composition of Archean TTG (Martin, 1995). Symbols as in Figure A4.4.

Documents

Cerium Oxidation State in Silicate Melts and the Application to Ce … · 2015. 6. 10. · Duane J. Smythe Doctor of Philosophy Department of Earth Sciences University of Toronto