TRACE ELEMENTS IN MAGMAS A Theoretical Treatment

Studying the distribution of certain elements, present in very low concentrations in igneous and metamorphic rocks, can yield important clues about the rocks’ origin and evolution. Trace elements do not give rise to characteristic minerals, but their behaviour can be modelled to provide historical information about the source magma. This book brings together the essential theory required to understand the behaviour of trace elements in magmas, and magma-derived rocks. It presents a wide range of models and mechanisms which explain trace element distribution. Trace Elements in Magmas provides an excellent resource for graduate students, petrologists, geochemists and mineralogists, as well as researchers in geophysics and materials science. Denis Shaw joined the Department of Geology at McMaster University, Ontario, in 1949, continuing his research as Professor Emeritus until 2003. Throughout his career, he taught courses in geochemistry in Canada, France and Switzerland. He worked as editor for a range of publications including Geochimica et cosmochimica acta and the Handbook of Chemistry. In 1964 he served as President of the Mineralogical Association of Canada, and received the Distinguished Service Award of the Geochemical Society in 2002. Professor Shaw passed away in October 2003 and his widow, Susan Evans Shaw, and Cambridge University Press are grateful to Professor Stuart Ross Taylor for his work in editing the final manuscript in preparation for publication. Stuart Ross Taylor, a trace element geochemist, is an emeritus professor at the Australian National Univeristy and is the author of Solar System Evolution: A New Perspective (Cambridge Univeristy Press) and several other books.

TRACE ELEMENTS IN MAGMAS A Theoretical Treatment DENIS M. SHAW Formerly of McMaster University, Ontario

Edited for publication by STUART ROSS TAYLOR

CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521822145 © S. Evans Shaw 2006 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2006 This digitally printed version 2007 A catalogue record for this publication is available from the British Library ISBN 978-0-521-82214-5 hardback ISBN 978-0-521-03634-4 paperback

Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Geochemistry is a compilation of imprecise, irreproducible and uncoordinated analyses. (i) Keep the rocks in mind, for they cannot be reduced to analytical measurements; (ii) (from O. F. Tuttle) minerals are the archives of the rocks; (iii) keep filing [your] fingernails while waving [your] arms.

Contents

Preface

page xi

1 Introduction 1.1 Defining trace elements 1.2 The quality of trace element data 1.3 Sample heterogeneity 1.4 Mineral analysis for trace elements 1.4.1 Sampling 1.4.2 Chemical analysis 1.5 Summary References

1 1 3 5 9 9 10 12 12

2 Partition coefficients 2.1 Introduction 2.2 Solutions with a common solute 2.3 Reacting solutions: law of mass action 2.4 Effects of variation of temperature and pressure 2.5 Measurement of partition coefficients 2.6 Extended theory 2.7 Major element effects 2.7.1 Olivine–melt partition 2.7.2 Redox effects 2.7.3 Volatile fluids 2.7.4 Melt structure effects 2.8 Influence of the host solid 2.8.1 Substitution site deformation 2.8.2 Influence of mineral chemistry

14 14 15 17 19 23 25 27 27 29 31 34 34 34 39

vii

viii

Contents

2.9 The Henry’s law controversy 2.10 Use of partition coefficients 2.11 Summary References

43 44 45 46

3 Crystallisation: basic trace element modelling 3.1 Introduction 3.2 Definitions 3.3 Temporal variables in a crystallising system 3.4 Equilibrium crystallisation 3.5 Fractional crystallisation 3.6 Mineral zonation 3.7 Intercumulus trapped melt 3.8 Mineral pairs 3.9 Incremental or stepped crystallisation 3.9.1 Constant melt proportion 3.9.2 Constant mass increments 3.10 Summary References

51 51 52 52 53 55 61 63 66 66 68 70 72 72

4 Crystallisation: variation of mineral proportions, partition coefficients and fluid phase proportion 4.1 Introduction 4.2 Variation in mineral proportions 4.3 Variation in partition coefficients 4.3.1 Trace elements 4.3.2 Major elements 4.4 Crystallisation in the presence of a fluid phase 4.4.1 Instantaneous degassing 4.4.2 Continued fluid release 4.4.3 Discussion 4.5 Summary References

74 74 75 78 78 83 86 87 89 90 92 92

5 Crystallisation assimilation, recharge and eruption 5.1 Introduction 5.2 Resorption or assimilation 5.3 Mass balance 5.4 Assimilation by melting and solution 5.5 Assimilation by reaction

94 94 94 95 99 103

Contents

5.6 5.7

5.8

5.9

Assimilation-fractional crystallisation processes Magma recharge and discharge 5.7.1 Conservation of initial magma mass C+A → E → R C+A → R → E 5.7.2 Conservation of residual magma C+A → E → R C+A → R → E 5.7.3 Discussion Recharge, eruption, assimilation: the rate process model 5.8.1 Conservation of the initial magma mass 5.8.2 Magma mass M is not constrained Summary References

ix

105 105 106 108 111 112 112 113 113 116 118 119 122 123

6 Trace element evidence for crystallisation processes 6.1 Introduction 6.2 Variation diagrams 6.3 Other two-element plots 6.3.1 Crystallisation trends 6.3.2 Element ratio plots 6.3.3 Mixing and assimilation 6.4 Inversion modelling 6.5 Summary References

124 124 124 126 129 133 134 135 139 140

7 Melting: basic trace element modelling 7.1 Introduction 7.2 Melting a heterogeneous rock 7.3 Early partial melting 7.4 Definitions 7.5 Bulk partition coefficient 7.6 Trace elements in equilibrium melting 7.7 Trace elements in fractional melting 7.8 Modal and non-modal melting again 7.9 Incremental batch melting 7.10 Batch melting with retained melt 7.11 Equilibrium melting vs. fractional melting 7.12 Melting in the presence of volatiles 7.13 Disequilibrium melting

142 142 142 148 149 150 150 152 158 160 163 168 170 179

x

Contents

7.14 Accessory minerals entrained during melting 7.15 Summary References

181 183 184

8 Melting: more complex processes 8.1 Introduction 8.2 Incongruent and reaction melting 8.2.1 Simple incongruent melting 8.2.2 Reactive melting 8.2.3 Three reacting phases plus an inert phase 8.2.4 More complex reactions 8.3 Variations in mineral proportions and partition coefficients 8.3.1 Variation in mineral proportions 8.3.2 Variation in partition coefficients 8.4 Rock melting by zone refining 8.5 Summary References

187 187 187 188 194 197 199 201 202 202 205 208 209

9 Dynamic mantle melting 9.1 Introduction 9.2 Dynamic melting 9.3 Closed system model 9.4 Open system model 9.5 Discussion of models 9.6 Melt dynamics 9.6.1 One-dimensional motion 9.6.2 Two-dimensional motion 9.6.3 Percolation of melt through mantle 9.7 Summary References

211 211 211 215 220 227 232 232 233 237 239 240

Index

242

Preface

The years following World War II saw a steady improvement in the analysis of rocks for minor and trace constituents. It became clearer that trace elements were not haphazardly distributed and that chance played a minor role. To find the principles of distribution of the elements was one of the aims of geochemistry, according to V. M. Goldschmidt, and researchers began to try to understand trace element behaviour. Two main approaches developed: one was aqueous geochemistry, where the emphasis was on the oceans and mineral genesis reactions in electrolytic solutions; the other studied the igneous and metamorphic rocks and, to some extent, metallic deposits. The second, often inappropriately called hard-rock or solid-state geochemistry was, like the former, concerned with heterogeneous phase reactions, but of different kinds. Central to it is the concept of the partition coefficient, and much effort has been expended in attempts to measure such parameters or variables. Many schemes were proposed and tested to show how some observed trace element or isotopic distribution pattern could be explained in petrological terms, and such accounts are scattered throughout the literature, camouflaged under various titles. This book constitutes an attempt to gather together the wide variety of possible models or mechanisms to explain the distributions of trace elements in igneous, metamorphic and metasomatic rocks, so that they are available for application as needed. The emphasis has been quite deliberately placed on the details of the mechanisms and, as a consequence, few examples have been cited. Another reason for the paucity here of examples from the natural world is that there are, up to the present, few trace element studies available where the analytical precision is sufficient to choose among different models, although this is not the case with many isotopic systems. Much of the material here has formed part of graduate courses and I am grateful to successive waves of graduate students for helping to keep my thinking on track.

xi

xii

Preface

I have benefited from discussions with too many helpful persons to list individually. I am grateful to McMaster University for a good working environment over many years and for post-retirement services and support. It is assumed that the reader is familiar with phase petrology and modern mineralogy.

1 Introduction

1.1 Defining trace elements The intent of this book is to examine processes which lead to the accumulation of trace elements in magmatic and metamorphic minerals and rocks, so at the beginning we must consider and define the terms to be encountered. The first task is to examine what is meant by a trace element. In a literal sense it is an element which is present in a rock, mineral or fluid at a low concentration. It is usual in the field of geochemistry to define major elements as those which give the sample whatever distinctive character it has, such as its mineralogical make-up; for example, the major elements of cherty limestone would include Ca, C, Si and O. In the case of most common rocks the major elements would include Si, Ti, Al, Fe, Mn, Mg, Ca, Na and K, with abundances in excess of perhaps 1% (in this book % will always be taken as a weight ratio, unless otherwise indicated). Note that O is not usually listed, because the other elements are bound to it. A number of minor elements occur at concentrations usually below 1%; in part they correspond to the presence of accessory minerals such as, apatite (P), zircon (Zr), fluorite (F) etc. Elements at low concentrations, but which do not give rise to characteristic minerals are classed as trace elements. Their manner of occurrence is to be discussed later. The terms used for concentration are %, ppm (1 ppm = 10−6 gg−1 = 0.0001%) and gt−1 (1 gt−1 = 1 ppm). So far, this is the language of the chemical analyst, and a statement of the kind ‘the trace element rubidium comprises 25 ppm of the rock’ gives no problem about its intended message: in 1 tonne of the rock there are 25 g of the metal Rb. Sometimes, however, the meaning is not so clear; e.g. to speak of ‘the partition of Ni between olivine and clinopyroxene’ raises the question of whether Ni, as a metallic element, can somehow occur in the silicate olivine or, in other words, what form Ni must take in order for it to be involved in some sort of reaction or equilibrium between

1

2

Introduction 1 000 000

83 Granitoid rocks

100 000 Ranges for silica and potash

Boron ppm

10 000 1000 100 10

Ranges for boron and lithium

1 1

10

100

1000

10 000 100 000 1 000 000

Lithium ppm Fig. 1.1 Analyses of 83 granitoid rocks from the Pe˜na Negra complex in central Spain (Pereira and Shaw, 1997) show how trace elements Li and B vary in concentration over two orders of magnitude, whereas the major components SiO2 and K2 O vary much less. The trace element concentrations are more sensitive to variations in conditions of origin.

the two minerals; is it present as Ni metal, or as Ni2+ , or as NiO, or in some other state? Without further elaboration the statement is not very precise, although useful as a kind of shorthand. Since the behaviour of trace elements in rocks is the major topic of this book it is desirable to indicate why they merit such attention. In many geochemical studies the approach used is exploratory and, with a particular goal in mind, elemental analyses of rock or sediment samples may serve as variables to test different alternatives and reach conclusions about an origin or a historical evolution. All the elements chosen may, of course, provide useful information, but there are at least three reasons why trace elements are often given special attention in such research: (i) the lower the concentration of an element, the more likely it is that its behaviour will be regular (ideal, in the language of the solution chemist) and not subject to effects linked to its absolute abundance; it may therefore provide information regarding external variables governing the evolution of the rock; (ii) the range of concentrations is not as restricted or interdependent as major elements (see Fig. 1.1); the latter must sum to 100% and therefore their concentrations are not independent of each other; (iii) trace elements exhibit a wider range of chemical behaviour as exhibited by their position in the periodic table, compared with the more restricted range of major element chemistry.

1.2 The quality of trace element data

3

50

Analytical precision for W-1

Coefficient of variation (%)

Mn

40

30

Mean and coefficient of variation of analyses from 24 laboratories

H

P K

20 Ti Na

10

Mg Al Fe

Ca

Si

0 0.1

1

10

100

Oxide weight (%)

Fig. 1.2 Mean values and coefficient of variation (percentage) for some major and minor elements (expressed as oxides) in the diabase W-1, using data supplied by 24 laboratories (Fairbairn et al., 1951, Table 14).

1.2 The quality of trace element data Trace elements are inherently at low abundances and consequently they are difficult to analyse with precision and accuracy.1 Few quantitative data were available before World War II and the quality of analyses has changed greatly since ∼1950. In the 1940s a project was initiated in the United States to calibrate two rocks (a granite G-1 and a diabase W-1) by cooperative analyses for major, minor and trace elements from a number of participating laboratories, so that the rocks could serve as standards for precision and accuracy. The results of the project (see Fairbairn et al., 1951) were disappointing, because the analyses from the participating laboratories showed many discrepancies (for major as well as trace components), leading to a confirmation of the view that rock analysis is a difficult art. For example, five laboratories respectively reported the Sr concentration in G-1 as 900, 120, 250, 280 and 450 ppm, with similar wide ranges for other trace elements. The variation in such results may be expressed by the coefficient of variation, which is the ratio of the standard deviation (sd) of the results to their mean value, and is usually expressed as a percentage. For major and minor elements (or oxides) Fig. 1.2 1

It is necessary to bear in mind that precision (Ger: zuf¨alliger Fehler, Reproduzierbarkeit; Fr: r´eproductibilit´e) refers to the ability to reproduce an analytical result by multiple measurements using a particular method, whereas accuracy (Ger: Genauigkeit; Fr: pr´ecision) is concerned with the ability to measure the ‘true’ value, without systematic bias dependent on the analytical method in use. The uncertainties introduced from these two causes are sometimes referred to, respectively, as random error and systematic error.