Mass Transport in Magmatic Systems

By Bjorn Mysen

Mass Transport in Magmatic Systems describes the properties and processes of these natural occurrences, including a description and discussions of how properties can be used for quantitative description of mass and energy transport on, and in, Earth and terrestrial planets. As the experimentally obtained chemical and physical properties of magma is scattered across literature, this book provides a comprehensive volume on the topic. Moreover, links between properties and processes are rarely appreciated. This makes it challenging for a non-experimentalist to access, evaluate, and apply such data.

• Incorporates information from a range of subdisciplines, from materials science to geology, geophysics and geochemistry
• Highlights links between properties and processes of magmatic systems
• Presents chapters that can stand on their own, with practical applications and a section for non-expert readers

• Language: English
• Publisher: Elsevier Science
• Release date: Oct 9, 2022
• ISBN: 9780128232095

Bjorn Mysen

Bjorn O. Mysen. Ph.D., Senior Scientist, Geophysical Laboratory, Carnegie Institution of Washington, Washington, D.C., Editor, Proceedings in Earth and Planetary Science, General Editor, Phase Diagrams for Ceramists (American Ceramic Society), Associate Editor, Geochimica et Cosmochimica Acta and American Mineralogist, Highly-cited scientist, Thompson ISI, 2001-present

Book preview

Mass Transport in Magmatic Systems

Bjorn O. Mysen

Geophysical Laboratory, Carnegie Institution of Washington, Washington, DC, United States

Table of Contents

Cover image

Title page

Copyright

Preface

Chapter 1. Melting in the Earth's interior: solidus and liquidus relations

1.1. Introduction

1.2. Premelting

1.3. Melting of peridotite

1.4. Melting of basalt

1.5. Melting of andesite

1.6. Rhyolite melting

1.7. Concluding remarks

Chapter 2. Melting in the Earth's interior: melting phase relations between the solidus and liquidus

2.1. Introduction

2.2. Melting interval of mantle peridotite without volatiles

2.3. Melting interval of mantle peridotite with volatiles

2.4. Melting interval of basalt

2.5. Melting interval of andesite

2.6. Melting interval of granite

2.7. Concluding remarks

Chapter 3. Element distribution during melting and crystallization

3.1. Introduction

3.2. Principles

3.3. Trace element substitution in melts and minerals

3.4. Element partitioning, intensive, and extensive variables

3.5. Mineral-melt partitioning and igneous processes

3.6. Concluding remarks

Chapter 4. Energetics of melts and melting in magmatic systems

4.1. Introduction

4.2. Energetics of melting

4.3. Heat content, heat capacity, and entropy of silicate melts and magma

4.4. Thermodynamics of melts and liquidus phase relations

4.5. Concluding remarks

Chapter 5. Structure of magmatic liquids

5.1. Introduction

5.2. Glass versus melt and glass transition

5.3. Silicate melt and glass structure

5.4. Iron in magmatic liquids

5.5. Concluding remarks

Chapter 6. Structure and properties of fluids

6.1. Introduction

6.2. Fluid/melt partitioning of volatile components

6.3. Structure and properties of H2O in fluids

6.4. Solubility behavior in fluid: H2O–SiO2

6.5. Solubility behavior in fluid: H2O–SiO2–MgO

6.6. Solubility behavior in fluid: H2O–Al2O3(–NaCl–KOH–SiO2)

6.7. Minor and trace elements in aqueous fluid

6.8. Concluding remarks

Chapter 7. Water in magma

7.1. Introduction

7.2. Speciation and abundance

7.3. Principles of solubility

7.4. H2O solubility

7.5. Concluding remarks

Chapter 8. Volatiles in magmatic liquids

8.1. Introduction

8.2. Oxidized carbon species

8.3. Reduced carbon (CH4, CO, and carbide)

8.4. Sulfur solubility

8.5. Nitrogen solubility and solution mechanisms

8.6. Hydrogen solubility and solution mechanisms

8.7. Halogen solubility and solution mechanisms

8.8. Noble gas solubility and solution mechanisms

8.9. Concluding remarks

Chapter 9. Transport properties

9.1. Introduction

9.2. Relationships among transport properties

9.3. Viscosity of magmatic liquids

9.4. Viscosity of model system silicate melts

9.5. Modeling melt viscosity

9.6. Diffusion

9.7. Electrical conductivity

9.8. Concluding remarks

Chapter 10. Equation-of-state of magmatic liquids

10.1. Introduction

10.2. Equation-of-state (EOS) of glass versus melt

10.3. Functional relationships

10.4. Equation-of-state of magmatic liquids

10.5. Equation-of-state of simple system model liquids

10.6. Concluding remarks

Chapter 11. Mass transport

11.1. Introduction

11.2. Porosity, permeability, and transport

11.3. Concluding remarks

Index

Copyright

Elsevier

Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

Copyright © 2023 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-821201-1

For information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Candice Janco

Acquisitions Editor: Amy Shapiro

Editorial Project Manager: Chris Hockaday

Production Project Manager: Sreejith Viswanathan

Cover Designer: Matthew Limbert

Typeset by TNQ Technologies

Preface

The formation and evolution of the Earth and planets depend on transfer of mass and energy. Magma and fluid are integral parts of the transport processes that govern the mass and energy transfer. Mass transport property data are central to describe those processes. Mass transport is accomplished by transfer of fluids and magma and typically takes place at high temperature and pressure. Mass transport typically occurs along temperature and pressure gradients, which means that energy transport also associates with mass transport, although in this book, energy transfer is not explicitly discussed. A structure-based understanding of how transport properties reflect changes in composition, temperature, and pressure greatly enhances our ability to use property data to characterize transport and transfer processes. This knowledge not only is helpful for the materials characterization needed to describe mass transport processes in nature, it also contributes to the knowledge base of adjoining scientific disciplines including glass and materials science. The focus of this book is to describe and discuss transport properties of magma together with aspects of transport properties of fluids, and to employ such data to characterize mass transport in the interior of Earth, its moon, and the terrestrial planets.

The principal aim of this Book, therefore, is to describe mass transport by magma and fluids, what and how melt and fluid properties govern those processes, and how understanding of the structure of those transport agents, and, therefore, their chemical composition, temperature, and pressure, can be used to characterize the properties. Linkage of transport properties to structure of the transport agents is important because this understanding provides a basis for quantitative modeling of property behavior without otherwise more comprehensive and extensive experimental study of each and every composition and conditions. The latter efforts require more human and financial resources than often are available.

The main focus of this book is on transport by magma with lesser emphasis on mass transfer by fluids. Some of the reasons for this selection is that fluid property data such as density and viscosity, for example, differ greatly from those of surrounding crystalline materials to the extent that variations of those properties of fluids do not impact greatly fluid-mediated mass transport. Of course, fluid compositions, pressures, and temperatures do. These property variations, therefore, are the subject of a major chapter of the book, but have not been isolated into individual chapters as was done for silicate melt and magma properties. The variables causing petrogenetically important changes in properties of fluids also affect migration efficiency. These variables have been discussed in the last Chapter of the Book (Chapter 11). That Chapter is centered on mass transfer by fluids and magma through crystalline rock matrix together with a number of examples from natural observations that can be, or have been, interpreted in terms of the passage of melts or fluids in a rock matrix.

The transport properties of magmatic liquids, often substantiated with information from compositionally simpler model system, are the main focus of this Book. For this purpose, the Book is organized in a petrogenetically evolutionary sense beginning with melting and crystallization of rock-forming materials to form and evolve magma (Chapters 1 and 2). Within this evolution, which leads to a wide variety of magma compositions and greatly variable transport properties, we follow the melting and crystallization behavior from the most primitive magma created by partial melting of peridotitic parental rocks in the Earth's mantle to a finish where melting and crystallization of the most evolved magmatic liquids, such as those of rhyolite and granite composition, are presented. Roles of volatiles, in particular H2O and CO2, were incorporated as appropriate. The compositional variations of the magmatic liquids in those environments can cause their transport properties to vary over many orders of magnitude.

Element distribution among melts, fluids, and minerals, and how this distribution is affected by their composition and structure, is central to characterization of mass transport in the Earth. Bulk composition of magma and crystalline minerals together with element, oxide, and isotopic solubility in and partitioning between these phases are sensitive to temperature, pressure, and redox conditions of the formation and evolution of magmatic liquids and the environment in which partitioning occurs. Element partitioning is described and discussed in Chapter 3, which follow naturally, therefore, from phase equilibrium melting and crystallization behavior presented in Chapters 1 and 2. The focus of Chapter 4 is thermodynamic data needed for characterization of the properties and processes discussed in Chapters 1–3. This chapter highlights existing thermodynamic data and how such information aids our understanding of the behavior of magmatic systems. This includes melting and crystallization behavior, element partitioning, and how thermodynamic data can be employed to characterize transport properties (viscosity, diffusion, and electrical conductivity) of silicate melts and magmatic liquids. Thermodynamics, therefore, not only help us to understand melting, crystallization, and element distribution behavior, such information can be employed directly to model transport properties of magma. Of course, ultimately, thermodynamic data and other melt and fluid property data are manifestations of the structure of the materials of interest.

Structural information forms the basis, therefore, for characterization of transport properties of magmatic liquids and of fluids. Structural data and how those data are linked to transport and associated properties such as described in Chapters 1–4, obtained for the most part from experimental studies, are contained in Chapters 5–8. Those four chapters are separated into a basic description of structural principles necessary to describe silicate melt structure and can be found in Chapter 5 for melt and in Chapter 6 for fluid structure.

In Chapter 6, in addition to structure discussions, other properties of fluids, including partitioning of the fluid components (H2O, CO2, CH4, H2, halogen-, N-, and S-containing fluid species, both reduced and oxidized) between fluids and melts fill out the initial sections of the discussion. This is followed by description of solubility behavior of major, minor, and trace compositions in fluids of various relevant compositions. The solubility and solution mechanisms of volatiles in magmatic liquids and model simple-system silicate melts are discussed in Chapters 7 and 8. This presentation was intended to follow naturally from the structure data provided in Chapters 5 and 6. Many facets of melt and fluid structure affect their transport properties, some of which also can be found in these chapters.

The remaining chapters (Chapters 9–11), focus directly on how mass transport (properties and processes governed by properties) by magma and fluid and of magma- and fluid-bearing systems depends on intensive and extensive parameters. Transport properties such as viscosity, diffusion, and conductivity together with how these may be linked together, can be found in Chapter 9. This chapter also offers several examples of how transport properties affect mass transport processes in the Earth and terrestrial planets.

Mass transport in planetary interiors is affected critically by the equation-of-state (EOS) of magmatic liquids as discussed in Chapter 10. The EOS information includes density, volumes, thermal expansion, and compressibility of chemically complex magmatic liquids. Similar data reported for the simpler model system are employed for a more thorough understanding of EOS of magma and (fluid) at high temperature and pressure.

The last chapter (Chapter 11) deals with actual movement of fluids and magma through rock matrices at pressures and temperatures exceeding those above which open cracks can be supported by the rock strength. Characterization of these properties and how they are affected by intensive and extensive parameters are critical for characterization of mass transport in planetary interiors. In this chapter, there not only is a discussion of some of the main variables governing fluid and melt migration, Chapter 11 also includes assessment of which melt and fluid properties can affect movement of those liquids through a crystalline matrix. Moreover, this chapter contains summaries of how liquid distribution and composition in a crystalline matrix affects geophysically and geochemically important properties of rocks often with small volume fractions of magma or fluid, and how such knowledge helps interpretation of natural geochemical and geophysical data.

The creation of a book such as this requires input from a wide range of specialties, many of which might not always have been in the center of the author's research activities. It has been very important, therefore, to garner input from friends and colleagues and, perhaps, most important of all, access and help in accessing published literature from a wide variety of scientific disciplines and subdisciplines. The assistance from our library and its two members, Shaun Hardy and M. O. O'Donnell, has been invaluable in this regard. This book could not have been produced without their exceptional professionalism, efficiency, and cheerful assistance. This is particularly so as this book was written while the COVID-19 pandemic was raging here and elsewhere in the world. Hence, much of the work was carried out electronically because person-to-person contact was difficult. Moreover, COVID-19-related technical problems such as, for example, production of graphics were overcome thanks in no little part to the assistance and support of my wife, Susana, who assisted in the generation of many of the diagrams used in the text.

There is, of course, much more data and understanding needed before we can claim an understanding of all transport processes governing the mass and energy transfer associated with the formation and evolution of Earth, its moon, and the terrestrial planets. I hope, however, that the information that is offered in this book will help pointing not only to what we believe we know, but also, and perhaps more importantly, what we do not know. It provides, therefore, an overview of current understanding of mass transport in petrogenetic processes. A major aim also is to develop suggestions for where future research activities might be the most useful. Those objectives can be reached not by what may be the fancy of the day, but with concerted and integrated efforts and inputs from natural observations, from systematic laboratory experiments, and by numeric modeling and integration.

Chapter 1: Melting in the Earth's interior

solidus and liquidus relations

Abstract

Magma is the dominant mass transport agent in the Earth and likely also terrestrial planets. Following formation of the magma ocean during the Earth's early history, primary magma was and is formed by partial melting of peridotite mantle.

The melting temperature of a volatile-free peridotite mantle ranges from about 1350°C near the bottom of a continental crust to near 2000°C at the top of the transition zone. The partial melt formed at the solidus of peridotite in this pressure range is basaltic.

The peridotite-H2O solidus decreases to between 800and 900°C at 1–2GPa pressure followed by a gradual temperature increase at greater pressure (depth). The melt in this initial pressure range is andesitic and gradually becomes increasingly mafic with increasing depth.

The peridotite-CO2 solidus can be profoundly different from that of volatile-free peridotite. At pressures less than ∼2.5GPa, CO2 only has a minor impact on melting temperature, and the melt is alkali basaltic. At greater pressures, the solidus temperature decreases very rapidly by as much as about 400°C. The solidus magma under these pressure conditions is carbonatitic.

Keywords

Basalt; Liquidus; Partial melting; Peridotite; Solidus; Volatiles

1.1. Introduction

Mass transport in the Earth and terrestrial planets is by magma (silicate melts) and by fluids compositionally in the system C–O–H–N–S. Generation of magma is the focus of this chapter.

Magma exists from ambient pressure and high temperature to the pressures and temperatures corresponding to the core/mantle boundary (Labrosse et al., 2007; Andrault et al., 2014; French and Romanowicz, 2015). Magma can, therefore, serve as a mass (and energy) transport medium throughout the pressure range of the silicate Earth (136GPa). Details of magma transport are presented in Chapter 9.

In this chapter, we will discuss how to generate magma in the Earth with a focus on the variables that govern the melting (solidus) and crystallization (liquidus) temperature/pressure coordinates. The phase relations that describe equilibria between minerals and melt the temperature interval between initial melting and complete melting will be discussed in Chapter 2. Here, after a brief discussion of premelting phenomena, we will describe the relationships at or near the solidus and the liquidus of the dominant silicate rocks in the Earth.

1.2. Premelting

A phenomenon known as premelting is detected by discontinuities in heat capacity versus temperature trajectories (Fig. 1.1). To date premelting has been observed only in laboratory experiments using endmember minerals (Richet and Fiquet, 1991; Courtial et al., 2000; Richet et al., 1996, 1998). The lack of information from chemically complex natural systems may simply be because the relevant experiments have not been carried out.

Macroscopically, premelting is represented by a rapid increase of the heat capacity as the melting temperature of a crystal is approached (Fig. 1.1). The heat capacity discontinuity begins from 80 to 250°C below actual melting temperatures. Enthalpy and entropy effects representi from 7% to 22% of the enthalpies and entropies of fusion (Richet and Fiquet, 1991; Thiéblot et al., 1999; Courtial et al., 2000; Nerád et al., 2013).

Premelting has been reported in synthetic diopside, CaMgSi2O6, together with other synthetic metasilicates (Richet et al., 1996). For diopside, the onset of premelting coincides with discontinuous changes in properties such as Ca diffusion (Dimanov and Ingrin, 1995) and electrical conductivity (Bouhifd et al., 2002). In this case, premelting has been inferred to be a reflection of temperature-dependent (Ca, Mg) structural disorder as the structural mechanism for the premelting phenomenon (Richet et al., 1996). In other metasilicates, incipient breakup of the silicate chain structure has been proposed (Richet et al., 1998; Nesbitt et al., 2017). For aluminosilicate crystals such as anorthite (CaAl2Si2O8), (Al,Si) disordering accounts for the premelting effect (Richet et al., 1994). Possible effects of solid solutions such as diopside-hedenbergite and anorthite-albite on premelting have not been addressed as yet.

Figure 1.1  Mean heat capacity of crystalline diopside (CaMgSi2O6) and anorthite (CaAl2Si2O8) as a function of temperature. Shaded region shows temperature interval of actual temperature range of premelting. Modified after Richet et al. (1996).

1.3. Melting of peridotite

Partial melting of peridotite in the Earth's mantle is the principal source of primary magma. Following melting, magma aggregates and ascends toward the surface of the Earth either to form shallow-depth magma chambers, perhaps governed by the principle of neutral buoyancy (Ryan, 1987), where crystal fractionation can alter the magma composition, or magma ascends directly ascent to or near the Earth's surface.

Peridotite melting may take place under essentially volatile-free conditions (e.g., Kushiro, 1969; Falloon et al., 1988; Zhang and Herzberg, 1994; Walter, 1998) or it occurs in the presence of volatiles such as H2O (Grove et al., 2006; Kawamoto and Holloway, 1997), CO2 (Canil and Scarfe, 1990; Brey et al., 2008), or mixtures of CO2 and H2O (Mysen and Boettcher 1975a,b; Wyllie, 1977; Ulmer and Sweeney, 2002). Under redox conditions equal to or more reducing than that corresponding to the iron-wüstite (IW) oxygen buffer ¹ reduced species such as H2 and CH4 can also play important roles during melting (Eggler and Baker, 1982; Luth and Boettcher, 1986; Taylor and Green, 1988). Such conditions likely were more common during the Earth's early history.

1.3.1. Peridotite melting without volatiles

Notwithstanding the common occurrence of mantle melting with volatiles such as either H2O or CO2, or both, melting of a peridotite lithosphere also takes place without volatiles (Herzberg et al., 1990; Hirose and Kushiro, 1993; Asimow et al., 2001). Early experimental studies on peridotite melting using natural peridotite starting material were those of Green and Ringwood (1967) and Kushiro et al. (1968). As can be seen in Fig. 1.2, the ambient pressure solidus of a typical peridotite is near 1150°C. The solidus temperature increases with increasing pressure at a rate of about 150°C/GPa. Within experimental error of the Kushiro et al. (1968) study, the solidus curve is linear. However, given the change of solidus mineral assemblage from olivine+orthopyroxene+clinopyroxene+spinel to olivine+orthopyroxene+clinopyroxene+garnet at pressures between 2 and 3GPa, one would expect a change of the slope of the solidus curve. From the Claussius-Clapeyron expression

(1.1)

the volume change, ΔV, will change as the mineral assemblage changes with increasing pressure. Such a volume change would be expected near 3GPa at the solidus temperature shown in Fig. 1.2 as this is approximately where the garnet-to-spinel transition is located. Evidently, this kink is within experimental error in the early data shown in Fig. 1.2. In more recent experimental studies, there are distinctive kinks of the solidus curve as a phase transformation is encountered although no kinks in the solidus were reported where the spinel-to-garnet is located (Herzberg et al., 2000; see also Fig. 1.3). The results summarized in Fig. 1.3 do, however, show kinks of the solidus near 15, 20, and 23GPa. These kinks and change in solidus slope reflect transformation from olivine to β-spinel phase (β-Mg2SiO4), Ca-perovskite (CaSiO3), to magnesiowüstite+Mg-perovskite (MgO and MgSiO3). Obviously, these changes in phase assemblages will also affect the composition of the melts on the solidus. These latter issues will be discussed in detail in Chapter 2.

Figure 1.2  Pressure/temperature of peridotite melting (solidus) in the absence of volatiles. Modified after Kushiro et al. (1968).

Figure 1.3  Pressure/temperature trajectory of peridotite melting to pressures near the interface of the transition zone to the lower mantle. Modified after Herzberg and Zhang (1996).

Although there is little disagreement as to the general nature of solidus phase assemblages of peridotite in the Earth's mantle, details of these phase assemblages as well as the pressure/temperature coordinates of the solidus curve and of the phase changes remain open to some discussion (see, for example, a review of those data by Herzberg et al., 2000). Some of the differences, seen, for example, in the various solidus temperatures reported in the literature (Fig. 1.4) are the result of different peridotite compositions.

The most obvious compositional effect on the peridotite solidus temperature is from changes in the Mg/(Mg+Fe) ratio of the peridotite. This ratio ranges from near 0.95 to less than 0.85 in mantle peridotite. From a compilation of 3GPa data from various experimentally determined peridotite solidus temperatures, Hirschmann (2000) found there to be a 90–100°C range in temperatures as a function of the bulk melt Mg/(Mg+Fe) of the peridotite (Fig. 1.5). This effect is not surprising given the relationship between Mg/(Mg+Fe) and solidus temperatures of the peridotite mineral phases (olivine, pyroxenes, spinel, and garnet). The Mg/(Mg+Fe) ratio also affects the pressure of the spinel-to-garnet transformation garnet on the peridotite solidus (Mysen and Boettcher, 1975a).

Another composition variable affecting solidus temperatures of terrestrial mantle peridotite is the alkali content (Na+K) (Fig. 1.6). This probably happens because alkali elements are incompatible in peridotite mineral assemblages and, therefore, enters the melt phase almost exclusively, at least under upper mantle conditions. Increasing Na and K, or both, results in solidus temperature depression.

Figure 1.4  Pressure/temperature trajectories of various peridotite solidii in the absence of volatiles. Modified after Herzberg et al. (2000) with the sources of individual curves indicated on individual solidii.

Figure 1.5  Solidus temperature of volatile-free peridotite as a function of their Mg/(Mg+Fe). Modified after Hirschmann et al. (2000).

Figure 1.6  Solidus temperature of volatile-free peridotite as a function of their total alkali content. Modified after Hirschmann et al. (2000).

1.3.2. Solidus phase assemblage and pressure

The peridotite solidus mineral assemblage governs the composition of initial melts. This assemblage and, therefore, the melt composition on the peridotite solidus, is a function of pressure (see also Chapter 2 for discussion of melting and crystallization mineral assemblages). Up to pressures near 15GPa, olivine, orthopyroxene, clinopyroxene, and one or more aluminous phases (plagioclase, spinel, and garnet) form the solidus mineral assemblage. At pressures below approximately 1GPa, plagioclase is the principal aluminous phase and initial melt is similar to midocean ridge basalt (Yoder and Tilley, 1962; Presnall et al., 2002). For typical terrestrial peridotite, aluminous spinel is on the solidus from near 1GPa to somewhere between 2 and 3GPa above which pressure garnet becomes the aluminous phase on the solidus of volatile-free peridotite. Garnet and aluminous spinel can coexist over a pressure range up to as much as 1.5GPa for the most Fe-rich peridotites (Bertka and Holloway, 1994; Walter, 1998; Grove et al., 2013). There is also a pressure range between about 1 and 1.5GPa where spinel and plagioclase coexist. In this pressure range, plagioclase becomes increasingly anorthite-rich as pressure increases until the plagioclase endmember, anorthite, finally disappears via the melting reaction (Presnall et al., 2002):

(1.2)

At pressures near 2GPa, spinel begins to react out to form a garnet+spinel peridotite mineral assemblage with spinel finally disappearing at pressures near 2.5GPa for typical peridotite compositions such as illustrated in Fig. 1.7. The pressure range with only garnet on the solidus can sometimes be as wide as 10GPa, which corresponds to the depth range ∼300km in the upper mantle (Takahashi, 1986; Herzberg and Zhang, 1996). The garnet in this pressure range not only changes its Mg/(Mg+Fe) but also its Al/(Al+Si) ratio because the concentration of the silicate perovskite component in garnet increases with increasing pressure (Irifune, 1994; Okamoto and Maruyama, 2004).

Figure 1.7  Pressure and temperature ranges of spinel and garnet peridotite mineralogy on the peridotite solidus. Modified after Walter (1998).

At pressures near 15GPa, olivine on both the solidus and liquidus of peridotite compositions undergoes a transformation to denser β-(Mg, Fe)2SiO4 (Fei et al., 1992). This phase is transformed to γ-(Mg,Fe)2SiO4 with a further pressure increase before silicate perovskite is stabilized at pressures near 20GPa. Magnesiowüstite [(Mg, Fe)O] becomes the solidus phase at even higher pressures (Irifune, 1994; Herzberg and Zhang, 1996).

The composition of the initial melt at these latter very high pressures (>20GPa) is not well known. Most likely, this lack of information results from challenges associated with temperature-quenching of melt without crystallization of quench phases at these very high pressures.

As noted earlier in the description of the results in Fig. 1.3, the pressure/temperature trajectory of the solidus curve shows a distinctive changes or kinks in slope as changes in solidus phase assemblages take place. These kinks reflect the volume change of melting as new mineral phases appear on the solidus.

1.3.3. Peridotite melting with volatiles

The principles that describe congruent melting of any rock in the presence of H2O or any other volatile in the C–O–H–N–S system are illustrated in the isobaric, low-pressure schematic representation in Fig. 1.8. In this figure, the solidus temperature, f-d, is fixed regardless of the amount of H2O in the system unless all the H2O is bound in hydrated minerals such as chlorite phases, amphiboles, mica minerals, or epidote. The solidus terminates at d because there is a finite solubility of rock materials in the H2O fluid (see Chapter 6). The liquidus topology, on the other hand, depends on the amount of H2O present in the system. For any bulk composition between f and b, the initial melt is at b. This melt is saturated with H2O. By increasing temperature above the undersaturated liquidus, a-b, an H2O-undersaturated melt will form. As drawn in Fig. 1.8, it is assumed that the H2O solubility in the melt decreases with increasing temperature, a feature commonly observed in experiments (Holtz et al., 1995; see also Chapter 7 for discussion of H2O solubility behavior in magmatic liquids). This means that by increasing the temperature until the H2O-saturated liquidus, c-b, is reached, H2O will exsolve. It is even possible to reach a condition below the H2O-saturated liquidus where the melt will exsolve H2O and will also partially crystallize. A further increase will eventually reach the vaporous. Details on solubility of silicate (rock) in the vapor (or fluid) can be found in Chapter 6.

Figure 1.8  Schematic representation of rock-H2O phase relations from temperatures above their vaporous to subsolidus conditions at low pressure (see text for detailed discussion).

1.3.3.1. Peridotite-H2O

Melting of peridotite in the presence of H2O at high pressure such as the deep crust, upper mantle, and beyond occurs at lower temperature than melting of peridotite in the absence of H2O (Fig. 1.9). When there is excess H2O over that which may be bound in hydrous minerals (amphibole, phlogopite, and chlorite, for example) or if temperatures and pressures are outside the stability field of hydrous phases in a peridotite-H2O system (Mysen and Boettcher, 1975a; Grove et al., 2006; Till et al., 2012), the isobaric hydrous solidus temperature is the same regardless of total H2O content.

As is always the case for melting of rocks in the presence of H2O, its solidus temperature decreases from its coincidence with H2O-free melting at ambient pressure to minimum temperature at pressures in the 2–4GPa range (Fig. 1.9). The coincidence at ambient pressure occurs because the H2O solubility in magma at ambient pressure is only a small fraction of wt% (see Chapter 7) and does, therefore, have no discernible effects on the solidus temperature. The exact pressure/temperature trajectory of the H2O-saturated solidus depends on the particular bulk composition of the peridotite. Mysen and Boettcher (1975a) found, for example, that depending of Mg/(Mg+Fe) ratio and alkali content, the hydrous peridotite temperature can vary by as much as 150°C at pressures near 3GPa (Fig. 1.9).

Figure 1.9  Solidus pressure/temperature trajectories of different peridotite compositions in the presence of excess H2O. The individual curves are from peridotite with varying Mg/(Mg+Fe) and total alkali content. Modified after Mysen and Boettcher (1975a) and Grove et al. (2006). Also shown in the solidus trajectory of volatile-free peridotite from Kushiro et al. (1968).

The temperature/pressure trajectory of the hydrous peridotite differs significantly among various published experimental studies. At the minimum temperature between 2 and 4GPa, solidus temperatures have been reported to be from 1000°C (Hirose and Kawamoto, 1995; Kawamoto and Holloway, 1997) to less than 800°C (Mysen and Boettcher, 1975a; Grove et al., 2006; Till et al., 2012). The reason for such a large variation in experimentally determined solidus temperatures is not clear. It is even further puzzling in light of the fact that for other rock types ranging from basalt/gabbro+H2O to granite/rhyolite+H2O, there is little disagreement between the published experimental data (see discussion of those experimental data below). A ∼200°C difference in reported solidus temperatures for hydrous peridotite is important as this affects the depth in the mantle where melting of hydrous peridotite may take place by perhaps 25km depending on the geotherm.

The mineral assemblages on the hydrous peridotite solidus in the continental lithosphere are the same as for anhydrous peridotite except that the pressures at which the transformation of plagioclase to spinel and spinel-to-garnet occurs on the H2O-saturated solidus is lower because the pressures and temperatures of the hydrous solidus is lower than for anhydrous peridotite and the spinel-to-garnet transformation as a positive dT/dP slope (see Fig. 1.9, for example). Garnet appears near and below 2GPa, for example (Taylor and Green, 1988), whereas for anhydrous melting, garnet on the peridotite solidus appears above 2.5–3.0GPa (Takahashi et al., 1993; Walter, 1998).

The stability relations of hydrous phases in continental lithosphere are profoundly different from their stability relations in the peridotite wedge in subduction zones. This difference is governed by the release of hydrous fluids saturated in silicate components from the descending slab in subduction zones, whereas no such source of H2O and silicate components can be found in lithospheric mantle. Hydrous phases such as amphibole, mica, and chlorite on the hydrous peridotite solidus wedge in subduction zones can occur over a range of pressures and temperatures (Mysen and Boettcher, 1975a; Grove et al., 2006; Till et al., 2012). During melting of continental lithosphere, on the other hand, the near absence of H2O in the melting region results in lack of significant contribution of hydrous phases to the peridotite melting.

It is generally agreed that at least to pressures near 2GPa, the initial melt on the solidus of hydrous peridotite is quartz normative and resembles andesitic compositions (Kushiro, 1972; Grove et al., 2006). It is less well known how that melt composition may change at higher pressures. It seems reasonable to assume that the melt compositions may eventually take on an olivine normative character (Condamine et al., 2016).

1.3.3.2. Dehydration on the peridotite solidus

Whenever the total H2O of hydrous peridotite is contained in hydrous phases, initial melting in limited pressure ranges could be controlled by the dehydration of the hydrous mineral(s). This may be the situation in the continental lithosphere where the H2O contents are on the order of hundreds of ppm (Jambon, 1994). This H2O likely is contained in a few hydrous phases and in nominally anhydrous phases. Among these hydrous phases, their detailed stability field depends on the peridotite composition. For the Hart and Zindler (1986) primitive peridotite used in the experiments by Till et al. (2012), the relationships between hydrous phase stability and dehydration melting are shown in Fig. 1.10.

Figure 1.10  Example of solidus pressure/temperature trajectory with all H2O bound in hydrous phases (amphibole and chlorite) in their stability range on the solidus. Modified after Fig. 1.5: Solidus temperature of volatile-free peridotite as a function of their Mg/(Mg+Fe) (Modified after Till et al. (2012).

The Mg/(Mg+Fe) of mantle peridotite is in the range 0.85–0.94 and total alkali concentrations ranging between, 0.1 and 0.9wt%. A range in amphibole stability over about 75°C and 0.3–0.4GPa temperature and pressure range is the result (Fig. 1.11). The alkali concentration, which is important for the amphibole stability (Allen et al., 1975), is uncertain in the peridotite wedge above descending plates in subduction zones as dehydration of the plate materials likely will release an aqueous fluid enriched in alkali metals (Mysen, 2002; Manning, 2004).

Alkali metal concentration also is important in defining mica stability field as a K-rich phase such as phengite can be stable to pressures near 10GPa, for example (Poli and Schmidt, 1998; Trönnes, 2002). K-rich amphiboles have been reported stable to near 10GPa at temperatures near the hydrous peridotite solidus (Sudo and Tatsumi, 1990; Trönnes, 2002). Such amphiboles and micas may not be found in typical upper mantle, but can be stable in metasomatized peridotite wedge above subducting slabs. In the deeper mantle of subduction zones, the H2O content likely is so low that all H2O is bound in such hydrous phases. The peridotite solidus under this circumstances can then be governed by dehydration of these phases (Fig. 1.12).

Figure 1.11  Pressure/temperature trajectories of amphibole dehydration solidi of peridotite with all H2O bound in amphibole when stable on the solidus. Modified after Green (1973) and Mysen and Boettcher (1975a).

Figure 1.12  Pressure/temperature trajectory of dehydration solidus of peridotite with K-richterite and phlogopite as dehydration phases. Modified after Trönnes (2002). H2O-saturated solidus with extrapolation to high pressure in dashed lines from Mysen and Boettcher (1975a).

Another important variable in peridotite mantle is the oxygen fugacity. This will affect not only amphibole stability, but likely also pressure/temperature stability range of biotite (Ernst, 1968; Niida and Green, 1999; Fumagali and Poli, 2005). Unfortunately, less is known about the relationship between bulk composition and biotite stability than of the amphibole stability. We do know, however, the phlogopite stability tends to exceed that of amphibole (Fig. 1.12; see also Kushiro et al., 1967; Trönnes, 2002; Frost, 2006; Condamine et al., 2016). However, in certain circumstances K-richterite stability can exceed that of phlogopite (Konzett et al., 1997; Konzett and Ulmer, 1999; Konzett and Fei, 2000; Trönnes, 2002). Note that in Fig. 1.12, the stability ranges of dense, hydrous magnesian phases (DHMS) are not included (see Ulmer and Trommsdorf, 1995,1998). Some of these may indeed be stable to temperatures above the H2O-saturated solidus at pressures exceeding 25GPa. It is important to note, however, that the stability field of some of these phases depend on their Mg/(Mg+Fe) (Konzett and Ulmer, 1999). Such phases may be found along the hydrous peridotite solidus at least to the bottom of the transition zone.

The stability range of hydrous phases as a function of bulk composition, not to mention fluid composition and fluid abundance, is important for peridotite melting because with all H2O tied up in one or more hydrous phases, their pressure/temperature breakdown curves also define the solidus pressure and temperature conditions of peridotite hydrated in this manner. Those relations, in addition to the local geotherm, will then govern the pressure stability and, therefore, the depth of peridotite melting. For the situation illustrated in Fig. 1.12, melting in the presence of H2O will occur at pressures above about 14GPa if all H2O were contained in K-richterite.

However, for most mantle situations, the K-concentration of the peridotite is such that phlogopite would be the highest-pressure stable H2O-bearing phase on the hydrous mantle solidus. This corresponds to a depth of about 180km. The relationship between this maximum pressure and the stability of DHMS will govern whether or not H2O might be transported to greater depth in the mantle.

1.3.3.3. Peridotite-CO2 melting

Carbon dioxide is the dominant C-species in the mantle to depth of perhaps the bottom of the upper mantle (up to ∼15GPa). At greater depths, the mantle appears sufficiently reducing (Wood et al., 1990; Frost and McCammon, 2008; Kaminsky et al., 2015) so that reduced carbon species are stable. Melting phase relations of the mantle under these conditions are, therefore, governed by carbon species different from CO2. This difference, in turn, results in changed melting phase relations (Litasov et al., 2014; see also further discussion of melting in a reduced mantle later in the chapter).

Melting in a CO2-bearing mantle is distinctly dependent on depth of melting because the solidus has a pronounced kink at pressure/temperature conditions approximately those of the decarbonation reaction (Eggler, 1976, 1978);

(1.3)

This feature originally was noted (Eggler, 1978) from simple system experiments in the system CaO–MgO–SiO2–CO2 (Fig. 1.13) where the relationship between the solidus temperature depression and the univariant equilibrium (1.3) is illustrated. In this system, the melting temperature at pressures below about 2.5GPa is so high that instead of diopside+enstatite, pigeonite is stable. At higher pressure, the temperature of the peridotite-CO2 solidus decreases by nearly 400°C over a ∼0.5GPa pressure interval (Fig. 1.13). In his groundbreaking study of the role of CO2 in upper mantle melting processes, Eggler (1976) also found, for example, that in the simple system model mantle, CaMgSi2O6 – Mg2SiO4 – SiO2, the pressure/temperature coordinates of the melt in equilibrium forterite+enstatite shifts from near 3.5GPa at 1800°C absent CO2 to near 1.7GPa and 1650°C in the presence of CO2 (Fig. 1.14).

Figure 1.13  Pressure/temperature trajectory of the solidus in the system CaO–MgO–SiO2–CO2. Modified after Eggler (1978).

The effect of CO2 on melting relationships in simple model mantle systems later was demonstrated also to operate in chemically more complex natural peridotite systems (Falloon and Green, 1989, Canil and Scarfe, 1990; Brey et al., 2008; Ghosh et al., 2009; see also Fig. 1.15). Several of those studies also extended the experiments to pressures near the bottom of the upper mantle (Canil and Scarfe, 1990; Ghosh et al., 2009). In these experimental studies, the nature and composition of the carbonate phase at the solidus also were monitored (Dasgupta et al., 2007) as well as the CO2 concentration in the melt at the solidus (Brey et al., 2008; Ghosh et al., 2014).

From his experimental data in the model mantle system CaO–MgO–SiO2–CO2, Eggler (1978) observed that with increasing pressure, the carbonate phase changed from calcite to dolomite and finally magnesite. This is the same trend as reported by Dasgupta et al. (2007) from their experiments with natural peridotite compositions where, however, the absolute Ca/(Ca+Mg) in the carbonate from natural peridotite+CO2 is lower (Fig. 1.16). A linear fit to the experimental data from Dasgupta et al. (2007) yielded the relationship for the carbonate phase as a function of pressure:

(1.4)

Figure 1.14  Phase relations in pressure/temperature space in the system CaO–MgO–SiO2 with and without CO2 near the invariant point, forsterite+enstatite+liquid+vapor. Modified after Eggler (1976).

Figure 1.15  Pressure/temperature trajectory of the peridotite-CO2 solidus. Modified from Falloon and Green (1989).

Figure 1.16  Ca/(Ca+Mg) of carbonate mineral in peridotite-CO2 and CaO–MgO–Al2O3–SiO2–CO2. Modified after Dasgupta et al. (2007).

1.3.3.4. Peridotite-C-O-H melting

Fluid in the C–O–H system under oxidizing conditions are H2O–CO2. This redox regime extends to oxygen fugacity conditions near or slightly below that corresponding to the magnetite-wüstite oxygen buffer (MW). At lower oxygen fugacity conditions, the dominant C-bearing species in the C–O–H is CH4 (see discussion of these relationships in Chapter 8). Absent hydrogen, the carbon is stored as graphite to pressures near 200km or diamond at greater depth (see also Kennedy and Kennedy, 1976; for data on the graphite-to-diamond transition). Under extremely reducing conditions, it has been suggested that carbon may substitute for oxygen in silicate polyhedra. This phenomenon has been reported in simple system silicate+C and in crystalline solid solutions between metal nitride and metal carbonate (Renlund et al., 1991).

1.3.3.4.1. Peridotite: H2O–CO2

As noted above, under oxidizing conditions, H2O and CO2 are the two fluid species in the C–O–H system. These are the conditions in the uppermost 100–150km in the modern Earth. At greater depth, the redox conditions are such that reduced carbon species may be stable (Stagno et al., 2013; Hammouda and Keshav, 2015).

By adding CO2 to the peridotite-H2O environment, the solidus temperature increases in a systematic manner (Fig. 1.17; see also Mysen and Boettcher, 1975a). The transformation of spinel-to-garnet peridotite at the solidus also increased as the CO2/H2O ratio increases because the temperature of intersection of the spinel-to-garnet transformation with the solidus increases with increasing CO2/(CO+H2O) ratio.

In analyses of the solidus phase relations in peridotite-H2–CO2 (Eggler, 1976, 1978; Wylllie, 1977), it was demonstrated that at higher pressures, above the of the intersection of the reaction (1.3) with the vapor-saturated peridotite–H2O–CO2 solidus, the initial melt remains carbonatitic to at least where the mol fraction in the vapor, CO2/(H2O+CO2), reaches 0.25. This melt composition remained even though the temperatures of the solidus may change significantly (Wyllie, 1977).

Figure 1.17  Pressure/temperature trajectories of peridotte–H2O–CO2 solidii as a function of H2O/(H2O+CO2) of the system. Modified from Mysen and Boettcher (1975a). The volatile-free solidus is from Kushiro et al. (1968).

Melting of an upper mantle with both CO2 and H2O depends significantly on whether the volatiles are contained completely in a crystalline phase such as amphibole, phlogopite, or carbonate, or whether there is excess volatiles over that which can be contained in crystalline phases (Eggler, 1978). With excess fluid, the composition of partial melt at 2–3GPa total pressure varies continuously from basanite or nephelinite in an H2O-free environment and then changes gradually ending up as andesitic melt in a CO2-free environment (see also Mysen and Boettcher, 1975a, and the topological analysis of this situation by Eggler, 1978; see also Fig. 1.18A). However, with only a small amount of fluid such that all H2O is contained in amphibole (less than about 0.4wt% fluid for a typical peridotite upper mantle), this amphibole is present in the solid assemblage over a CO2/(CO2+H2O)-range of 0–0.8. in this CO2/(CO2 +H2O)-range, the melt composition, of melilitic composition, is invariant (Mysen and Boettcher, 1975b; Eggler, 1978). This fluid region was termed Zone of Invariant Composition (ZIVC) (Fig. 1.18B). Analogous analyses may be carried out with other hydrated phases such as, for example, phlogopite. In such cases, however, the pressures and melt compositions of the zone of invariant melting would be different.

Figure 1.18  Melting phase relations in the system CaO–MgO–Al2O3–SiO2–H2O–CO2 as a function of CO2/(CO2+H2O). (A) Phase relations with excess fluid. (B) Phase relations with all H2O bound in amphibole to create the Zone of Invariant Vapor Composition (ZIVC). Modified after Eggler (1978).

1.3.3.4.2. Peridotite-C-O-H under reducing conditions

Experiments under reducing conditions involve equilibria between peridotite mineral assemblages and CH4-rich vapor (see also Chapter 8). The first attempt to determine melting of silicates in the presence of CH4 was that of Eggler and Baker (1982) where it was shown that CH4 reduced the melting temperature of diopside (CaMgSi2O6) by about 100°C at pressures near 2GPa. Taylor and Green (1988) reported what appears to be the first experimental data on peridotite–CH4–H2O melting at oxygen fugacity conditions defined by the tungsten-tungsten oxide buffer (which is about one log unit more oxidizing than the more commonly used iron-wüstite buffer). They reported a solidus temperature depression in the 1–3.5GPa to be in the 100–150°C range (Fig. 1.19). The spinel-to-garnet transition occurs near 2GPa on the solidus. The solidus mineral assemblage in the Taylor and Green (1988) study showed amphibole-bearing spinel lherzolite to maximum pressures of about 3GPa and phlogopite-bearing garnet lherzolite at higher pressures. Amphibole was not stable on the solidus with the activity of H2O in aH2O, less than about 0.5. The maximum pressure of amphibole stability decreased as aH2O decreased.

Litasov et al. (2014) extended the pressure range of peridotite-CH4 melting to about 20GPa. They found that, compared with the volatile-free melting curve, CH4 depressed the peridotite solidus by between 100 and 400°C in this pressure range (Fig. 1.20). The solidus mineral assemblage from about 3GPa to near 13GPa is olivine+orthopyroxene+clinopyroxene+garnet. At pressures below ∼3GPa, garnet is replaced by spinel. From about 13GPa to near 18GPa, olivine on the solidus is replaced by wadsleyite (β-Mg2SiO4). At pressures above ∼18GPa, wadsleyite is replaced by ringwoodite (γ-Mg2SiO4). To the extent it was possible to determine the melt composition on the solidus, Litasov et al. (2014) reported it to be of basaltic composition.

Figure 1.19  Melting phase relations of peridotite-C-O-H under redox conditions at IW+1. A. Comparison of the reduced peridotite-C-O-H solidus with the volatile-free and the peridotite-H2O solidus. B. Peridotite-C-O-H solidi as a function of the activity of H2O, aH2O. Modified after Taylor and Green (1988).

Figure 1.20  Melting phase relations of peridotite-C-O-H under redox conditions at IW oxygen buffer to pressures in the transition zone of the mantle. Note that this solidus does not show a change in dT/dP slope when the olivine-to-wadsleyite and wadsleyite-to-ringwoodite univariant lines are crossed. This probably reflects lack of experimental precision when determining the solidus curve under these high-pressure conditions. Modified after Litasov et al. (2014).

1.3.3.5. Magmatic processes in peridotite-C-O-H environments

The dominant fluid in most tectonic settings of the upper mantle where peridotite melting occurs, likely is dominated by C-bearing volatiles (Zhang and Duan, 2009). The bulk C content of the silicate Earth is, however, only on the order of 200ppm (Jambon, 1994).

Low-degree partial melt at the depth of melt separation from its mantle residue, suggested to be at 30–50km depth (Presnall et al., 2002), could then yield a tholeiitic to alkali basaltic magma with less than a few wt% CO2 (Brooker et al., 2001; Iacono-Marziano et al., 2012). Melting of a CO2-rich upper mantle, on the other hand, will result in alkali basalt (Mysen and Boettcher, 1975b).

The exception to the statement above that CO2 tends to be the most abundant volatile in the melting region of the upper mantle can be found in island arcs. Here H2O concentrations in the mantle wedge can reach upwards of 1wt% (Scambelluri and Philippot, 2001; Grove et al., 2002; Till et al., 2012). This H2O is stored in hydrous phases such as amphibole, mica, chlorite, and DHMS (Konzett and Ulmer, 1999; Fumagalli and Poli, 2005; Melekhova et al., 2015). Dehydration of these minerals with migration of this H2O into the overlying mantle wedge can result in H2O contents in the initial partial melts between 5 and 10wt%.

The composition of melt on the hydrous peridotite solidus to depths near 100km is quartz normative and has been considered andesitic by many experimentalists (Kushiro, 1972; Mysen and Boettcher, 1975b; Grove et al., 2002, 2006). This conclusion would not be significantly affected by addition of some CO2 to the H2O fluid least at depth less than 100km because the CO2 is retained in the solidus mineral assemblage as carbonate (Kerrick and Connolly, 2001; Poli et al., 2009). The CO2/H2O ratio of the released fluid does, however, increase with additional depth so that melting beneath back arc regions would yield nepheline-normative magma (Mysen and Boettcher, 1995b; Eggler, 1978; Dasgupta et al., 2007).

The oxygen fugacity (fO2) is a variable that can affect melting phase relations in the upper mantle significantly mostly because it affects the speciation of the C–O–H volatiles. The fO2 is, however, a function of depth in the mantle, as concluded, for example, by Frost and McCammon (2008). The fO2-values in the uppermost 100+ km of the mantle are near those corresponding to the QFM oxygen buffer. These latter oxygen fugacity conditions are those under which midocean basalt last equilibrates with upper mantle mineral assemblages (Davis and Cottrell, 2018). As the pressure increases with a further increase of mantle depth, the oxygen fugacity decreases to about four orders of magnitude below QFM at 7GPa (about 250km depth) (Frost and McCammon, 2008; see also Fig. 1.21).

Melting of peridotite-CO2 in a cratonic mantle begins at depths near 150km (Sleep, 2009; see also Fig. 1.22). The initial melt at this depth would be carbonatitic under oxidizing conditions. However, if the conditions were sufficiently reducing so that the C will exist as CH4, melting along the continental platform geotherm will take place at slightly greater at depths, near 200km (Fig. 1.22). These melting conditions yield a melt of basaltic composition (Falloon and Green, 1989). This contrasts with melt compositions under oxidizing conditions where the melt on the CO2 and CO2 +H2O solidus is carbonatitic at this depth. Of course, any warmer geotherm will be consistent with melting in either peridotite-CO2 or peridotite-CH4 environments depending on whether the depth at which the geotherm and the peridotite solidus intersect. Moreover, if H2O is present, the solidus temperatures, whether with CO2–H2O or CH4–H2O, may decrease by as much as 200°C depending on C/H ratio (Mysen and Boettcher, 1975a; Taylor and Green, 1989). This will also result in melting at shallower depth and, therefore, possibly under more oxidizing conditions with the attending oxidation of the C–O–H fluid phase and consequent changes in solidus mineral assemblages and chemical composition of partial melts.

Figure 1.21  Oxygen fugacities for a range of different peridotite samples from different depth (pressure) in the upper mantle. Modified after Frost and McCammon (2008).

Figure 1.22  Pressure/temperature trajectories of solidii of peridotite-CH4 and peridotite-CO2. Modified after Sleep (2009).

1.4. Melting of basalt

In this section, we will cover basalt sensu strictu as well gabbro, which is, of course, the intrusive form of basalt. In addition, basalt can be metamorphosed to eclogite in the deep continental crust and upper mantle. Eclogite melting will, therefore, also be discussed in this section. The mineral assemblage of eclogite is garnet+clinophroxene±quartz±olivine, whereas at lower pressure in the crust, the mineral assemblage that will undergo initial melting is that of gabbro (plagioclase+olivine+pyroxene±quartz). With H2O, rocks of basaltic composition could be amphibolite with various amphiboles as the dominant solidus mineral. Basalt at or near the surface typically has olivine+plagioclase as liquidus phases.

Melting of basalt/gabbro takes place in the continental crust. In subduction zone settings, basalt melting (or its metamorphic equivalents, amphibolite and eclogite) occurs in upper portions of descending slabs. Melting of amphibolite occurs to depth near 75–100km. Below this depth (equivalent to 2.5–3GPa), amphibole (hornblende) breaks down and the rock is transformed to eclogite with or without hydrous minerals that are stable to higher pressures than amphibole (Ernst, 1968; Allen et al., 1975). The exact depth where this occurs is, however, significantly dependent on the amphibole composition and redox conditions. In this latter environment, the basaltic compositions often are those that were metamorphosed in the presence of recirculating ground water at high temperature when the basaltic magma was intruded or extruded near their original setting at midocean ridges (Poli and Schmidt, 2002).

1.4.1. Basalt/gabbro melting without volatiles

At or near ambient pressure, basalt typically melts between about 1050°C and 1350°C. The primary control on the melting temperature is the Mg/(Mg+Fe) of the rock (Yoder and Tilley, 1961; see also Fig. 1.23). Many more recent data could be added to the results summarized in Fig. 1.23. However, none of this would affect the general conclusion that links the bulk Mg/(Mg+Fe) of basalts to their melting temperatures.

Figure 1.23  Relationship between liquidus temperatures of basalt and their Mg/(Mg+Fe). Modified after Tilley et al. (1964).

The relationship illustrated in Fig. 1.23 is because the dominant minerals in basalt are olivine, orthopyroxene, and clinopyroxene, the melting temperatures of which are dominated by their Mg/(Mg+Fe) (Bowen and Schairer, 1935). For olivine, which is also the first phase to crystallize from most basaltic magmas upon cooling at crustal depths, there is a simple solid solution from the highest-temperature forsterite endmember melting near 1890°C to the lowest-temperature melting fayalite with a melting point near 1200°C (Bowen and Schairer, 1935). Of course, this endmember behavior is not directly applicable to the more complex basaltic magma where, for example, olivine is in a reaction relationship with orthopyroxene (Bowen and Anderson, 1914). Moreover, the plagioclase composition, which is the common second phase to crystallize from tholeiitic magma and the first to crystallize from high-alumina basalt magma, also is quite variable. The plagioclase crystallization temperature in basaltic composition environment can cover a wide range temperatures depending on the Na/(Na+Ca) (albite-anorthite solid solution of plagioclase) with melting temperatures from ∼1120°C for albite to ∼1550°C for anorthite (Tuttle and Bowen, 1950). The liquidus temperatures of pyroxenes can be more complex because of several possible reaction relations. However, this is not directly relevant to liquidus temperatures of basalt because at and near ambient pressure, pyroxenes do not crystallize on the liquidus. Pyroxene reaction relations will, however, affect the solidus conditions as well as high-pressure melting relations. Details of reaction relations such as those, as well as more complex relationships, will be discussed in more detail in Chapter 2.

With increasing pressure, olivine as the solidus phase found in most basalts is replaced by clinopyroxene (cpx) over a rather narrow pressure interval (Fig. 1.24A). At pressure near 1GPa there is a near-invariant point with olivine, orthopyroxene, clinopyroxene, and an aluminous phase (spinel or plagioclase) appearing with a few degrees of the liquidus (see Bender et al., 1978; Fuji and Bougalt, 1983; Draper et al., 1992; see also Fig. 1.24A). Clinopyroxene is the liquidus phase at high pressure followed by garnet. Interestingly, it is high-alumina basalt melts that crytallize the four-phase assemblage, olivine+orthopyroxene+clinopyroxene+spinel and/or plagioclas in the 1–1.2GPa pressure range (Yoder and Tilley, 1962; Fujii and Kushiro, 1977; Johnson, 1986), whereas what is referred to as primitive tholeiite does not show this near-invariant behavior at any pressure (Fig. 1.24B). On the other hand, Yoder and Tilley (1962) in their seminal paper on phase relations of basalt did indeed observe near-invariant crystallization of the mineral assemblage olivine+orthopyroxene+clinoppyroxee+spinel+plagioclase on the tholeiite liquidus at 1.1±0.05. GPa. These differences leave open the question as to whether merely crystallizing a four-phase peridotite-type mineral assemblage means that this rock was in fact equilibrated with upper mantle peridotite under relevant pressure and temperature conditions. In addition, there is also the question whether volatiles such as CO2 and/or H2O might play a role. Finally, we need to remember that the variations in chemical compositions are not great and the different small variations may be within experimental error.

1.4.2. Basalt/gabbro-H2O

The high water content of igneous rocks found at or near convergent plate boundary commonly is ascribed to partial melting of descending oceanic sediments, of hydrated oceanic basalt (amphibolite), and/or to H2O-triggered partial melting of the overlying peridotite wedge (Mysen and Boettcher, 1975b; Poli and Schmidt, 2002; Mitchell and Grove, 2015). In general, the great majority of subducted and altered crust samples has H2O/CO2 >1 (Fig. 1.25). The depth of devolatilization and the concentration of the volatiles together with their H2O/CO2 abundance ratio reflect the pressure-temperature stability fields of the individual hydrous and carbonated minerals (Kerrick and Connolly, 2001; Poli and Schmidt, 2002). In general, the H2O-rich hydrous minerals epidote, chlorite, and amphibole tend to dehydrate at lesser depth than that at which decarbonation of carbonate minerals takes place (Poli and Schmidt, 2002). In fact, even in a water-rich CO2–H2O environment, CO2 is retained in the crystalline carbonate phases such as calcite and dolomite while the released fluid is quite H2O-enriched (Eggler, 1978; Yaxley and Green, 1994; Molina and Poli, 2000). The melting phase relations to depth near 100km may, therefore, be illustrated by phase relations in hydrous basalt (Fig. 1.26). Those phase relations are the basis for the H2O budget in subduction zones as a function of depth (Fig. 1.27).

Figure 1.24  Examples of phase relations near the liquidus of primitive basalt (A) High-alumina basalt. (B) Olivine tholeiite. (A) Modified after Fujii and Bougault (1983), (B) See Gust and Perfit (1987).

Melting of subducting oceanic crust involves melting of hydrous basalt, overlying altered sediments, and perhaps a combination of the two. Melting of basalt with excess H2O results in a solidus temperature depression of 600°C–700°C compared with the basalt solidus in the absence of volatiles [For an early review of experimental data, Lambert and Wyllie (1972, 1974) and summary in Fig. 1.28]. More recent data than those of Lambert and Wyllie (1972) and the papers references therein do not show significant differences from their original data source for hydrous melting of basalt.

Figure 1.25  Estimate H2O and CO2 abundance of crustal materials being subducted in various subduction zones. Modified after Poli and Schmidt (2002). See this reference for original sources of data.

Figure 1.26  Pressure/temperature trajectory of melting relations of Mid-Ocean Ridge Basalt-H2O and stability hydrous phases near the solidus. Modified after Poli and Schmidt (2002). See this reference for original sources of data.

Figure 1.27  H2O content available for release from subducting slab beginning with 6.2wt% H2O of the slab at the onset of subduction. Modified from Poli and Schmidt (2002).

Figure 1.28  Pressure/temperature trajectory of basalt with excess H2O and under anhydrous conditions. Modified after Hill and Boettcher (1970).

A striking feature of the data in Fig. 1.28, and also in other published experimental data, is the observation that the magnitude of the temperature depression of basalt+H2O is near 50% greater than for peridotite+H2O. In fact, in general the extent of the temperature depression of the solidus caused by H2O at any pressure increases as the magma becomes more (SiO2+Al2O3)-rich.

An interesting feature of the H2O-saturated solidus curve of basalt (as well those of more felsic rocks,