Chapter 3

Minerals and the interior of the Earth

The Earth is a dynamic planet, particularly when considered on the geological timescale of many millions of years. The short-term and visible evidence of this dynamism is volcanoes and earthquakes, the long-term evidence is the creation of oceans and the collision of continents. It was not until the 1950s that a theory to adequately explain this drama was propounded. The key role played by mineral magnetism in this story (as well as the role of the magnetic properties of minerals in human history) is outlined in Box 3.

Today the paradigm of plate tectonics is so widely accepted that its terminology has entered everyday speech. It is not my intention to devote space to reviewing the theory of plate tectonics; numerous eloquent accounts already cover this topic. However, the tectonic plates that are created and destroyed at different types of plate boundaries are made up of rocks with their constituent minerals, so the story of plate tectonics is also the story of the formation, transformation, and destruction of minerals. In particular, it is the story of crystallization of minerals from the molten state, and the destruction of minerals by melting. Figure 10, in a schematic and simplified cross-section, shows melts rising up from deep in the Earth along mid-ocean ridges, solidifying and forming new ocean crust, spreading apart at this constructive plate boundary. By contrast, the plate is dragged down (subducted) at a destructive

Box 3 From lodestone to spreading ocean floors

The discovery that the iron oxide mineral magnetite (Fe₃O₄) once called ‘lodestone’ (where lode means ‘way’ or ‘course’) has the distinctive properties we now associate with a particular form of magnetism was another of the great landmarks in the journey of mankind to the modern age. The ability of magnetite to attract iron was known to the ancient Greeks by around 600 bc. However, it is not until the period 1000–1200 ad that there are records of its use as a compass for navigation, helping thereafter to open the door to global exploration and to many important discoveries. The same mineral, along with a small number of other minerals with similar magnetic properties, also played a key role in a scientific revolution a thousand years later, in the 1960s.

After the Second World War, mapping of the ocean floor and its magnetic properties arising from the fossil magnetism (palaeomagnetism) frozen into the basalt rocks when they solidified on emergence at a mid-ocean ridge, provided the first evidence that the sea floor is spreading outwards on either side of the ridge. These rocks show evidence of having undergone periodic reversals in the polarity of the magnetic field in which they formed. The resulting magnetic striping of the ocean floor provided the first clear evidence for the opening of oceans, demonstrating continental drift and leading to the theory of plate tectonics. Palaeomagnetic studies, which depend on the properties of a small number of (mostly iron-containing oxide) minerals, have played a central role in unravelling the extraordinary story of opening and closing oceans and drifting continents through geological time.

plate boundary and melting occurs to form magmas which, being less dense than the surrounding rocks, rise towards the surface. If the magma reaches the surface we have volcanoes, and the magma erupts as lava or ash, but it may stop short of the surface

10. A much simplified cross-section through a part of the lithosphere

and crystallize slowly, forming a body of rock termed a pluton. Magmas may also originate at hot spots associated with mantle ‘plumes’ rising up beneath either oceanic or continental crust. Whatever the sources of magmas, mineralogy meets plate tectonics when we consider melting and crystallization processes.

Crystallization and melting

In the 1920s, an experimental scientist in the USA named Norman Bowen performed a series of classic experiments simulating the crystallization of rocks from magmas. Like others before him, Bowen found that the process of crystallization from a melt occurs over a large range of temperatures (hundreds of degrees) and that the first crystals to form have a very different chemical composition to the overall composition of the melt. As cooling continues, these crystals react with the remaining melt in quite a complicated way during the course of crystallization. For example, the first minerals to crystallize from the sort of magma that erupts at mid-ocean ridges (basaltic magma) are olivines (see Table 1). The olivines range in composition from Mg₂SiO₄ (forsterite) to Fe₂SiO₄ (fayalite) in a continuous series. The first to crystallize are magnesium (Mg) rich and they react with the melt, becoming more iron (Fe) rich as crystallization proceeds. However, with continued cooling and crystallization, these olivines are replaced, either completely or in part, through their reacting with the melt to form a new mineral, an Mg-rich pyroxene (MgSiO₃) which again reacts to become more Fe rich and can be itself replaced by an amphibole and then by a mica (biotite). Important here is the idea of incomplete reaction, because removal of the earlier formed crystals (say by settling to the bottom of a body of magma) means effectively starting with a new melt composition. This process of fractional crystallization greatly increases the range of minerals and hence types of rocks that can be produced from the starting melt.

Largely overlapping with the crystallization of Mg–Fe silicates from a basaltic magma, there is also crystallization of feldspars, beginning with calcium-rich species (CaAl₂Si₂O₈) and with continuous substitution of calcium by sodium, and of part of the aluminium by silicon to reach the composition NaAlSi₃O₈. With the feldspars, the same crystal structure is retained throughout in a process described by Norman Bowen as a continuous reaction series and contrasting with the discontinuous reaction series in which the species of mineral changes as well as its chemistry (olivine→pyroxene→amphibole→mica and ultimately →potassium feldspar and quartz). What we see here is the way in which a magma can ‘evolve’ via the process of fractional crystallization with the changes now known as Bowen’s Reaction Series. Indeed, Bowen entitled his classic book describing his work The Evolution of the Igneous Rocks.

Although Bowen suggested that all igneous rocks could be derived from a parent basaltic magma (the most common magma) by fractional crystallization, we now know that such extreme ‘differentiation’ is uncommon. Even so, fractional crystallization is extremely important in understanding how minerals crystallize from melts, and in explaining many phenomena observed in minerals. One of these is zoning in crystals such as those of olivine or feldspar (particularly in certain plagioclase feldspars formed at comparatively low temperatures). A zoned crystal exhibits ‘onion-like’ layers between the core and the rim. For example, the innermost core of a crystal of feldspar (the part which must have crystallized first) will commonly have a calcium-rich composition and successive layers (zones) have compositions that are more sodium rich. This is because the earlier formed feldspar crystals have not reacted completely with the melt before crystallization of later (more sodium-rich) feldspar. The minerals of the discontinuous reaction series, from olivine through to quartz, involve an increase in silica content from first formed to last formed. Also, this series shows successive changes in silicate crystal structure from island (olivine) to single chain (pyroxene) to double chain (amphibole) to layer (mica) to framework (quartz) forms.

Many of the concepts applicable when we talk about crystallization of magmas are important when we consider melting of rocks; in a sense they apply in reverse. So, just as crystallization takes place over a wide range of temperature, melting also takes place over hundreds of degrees. When melting begins, it is only a small fraction of the rock which becomes liquid, that fraction which would be last to crystallize if we were cooling instead of heating. Because only part of the rock melts, the process is called by the obvious name of partial melting. This is a very important process because it plays a major role in the generation of magmas as part of the plate tectonic cycle. Much of this magma generation arises from partial melting of material deep within the Earth, beneath mid-ocean ridges or where a lithospheric plate is subducted. Partial melting also occurs when ascending magmas react with the rocks they pass through en route to the surface. In order to take these ideas further, we need to make a journey. But, before that, we need to say a few words about two other topics of importance for the formation and evolution of the continental crust: granites and their formation, and the change in form associated with metamorphism.

About granites and continents

By definition, granites are rocks formed from melts which have cooled relatively slowly at depth in the Earth. For this reason the minerals they contain have formed relatively large crystals. The essential minerals in granites are quartz, potassium feldspar, and plagioclase feldspar with relatively small amounts of biotite or muscovite mica and/or amphibole. Granite rocks form a major part of the continental crust of the Earth, occurring as both relatively small bodies and as the very large plutons now often exposed as the cores of great mountain ranges. Currently known only on planet Earth, granite is the most abundant ‘basement’ rock of our planet, underlying the relatively thin sedimentary veneer of the continents. So, why is granite so abundant?

The answers to this question have been provided by decades of experiments at high temperatures along with our evolving ideas about the dynamic processes associated with plate tectonics. The experimental work has shown that magmas will evolve so that the last melt compositions (products of differentiation of the kind associated with the Bowen Reaction Series) will be granite melts. Conversely, the first magma compositions to be produced on partial melting of many rock types will be granitic. Such magmas have either evolved or been formed by melting at or near what is termed a eutectic point (eutectic being a word derived from the Greek meaning ‘easily melting’). This is a mixture of minerals in fixed proportions that melts (or solidifies) at a single temperature which is a lower melting temperature than that of the separate minerals, or any other mixture of them.

In the context of plate tectonics, the magmas formed by partial melting of subducted lithosphere or of material at depth in the continental crust will be granitic in most cases (see Figure 10). A very important finding from experiments is that the presence of small amounts of water can greatly lower the temperature at which melting can occur (by several hundred degrees) such that melting can occur in the deeper parts of the continental crust. The melts formed at depth will be less dense than the surrounding rocks and slowly rise towards the Earth’s surface to produce volcanoes if they do reach the surface, or if not, to solidify as plutons. Granites are classified on the basis of their detailed mineralogy, which in turn provides clues as to the processes by which they have formed. For example, whether a granite has been formed from melting of an igneous or a sedimentary source rock (hence I-type and S-type granites).

Metamorphism

When rocks are subjected to increases in temperature or pressure and temperature, changes occur in the minerals they contain. This may involve changes in the sizes and shapes of mineral grains without the minerals themselves changing. A good example here is that of a marble produced by the metamorphism of a pure limestone. In both rocks, the only mineral present is calcite (CaCO₃) but the calcite grains may be coarser and more closely interlocking and compact in the marble than in the limestone. Marble has been a favoured raw material for architects and sculptors since ancient times because it is softer and more homogeneous than many alternatives. The material favoured by Michelangelo for his finest sculptures was a pure white marble from Carrara in the Italian Alps. Most metamorphism, however, involves changes in the minerals present in the rock. There are a number of kinds of metamorphism; two of the most important are contact metamorphism and regional metamorphism. As the name suggests, contact metamorphism occurs where a magma is brought into contact with the rocks into which it has been emplaced (‘intruded’), thus raising temperatures and, in some cases, introducing new chemical components to form new minerals. These changes are only local and diminish away from the ‘intrusion’.

Regional metamorphism has affected very large areas of the crust, notably where the rocks are parts of either present-day or ancient mountain belts. These are rocks that have been subjected to increases in both temperature (typically somewhere between 100°C and roughly 900°C, above which rocks begin to melt) and pressure. It was in the ancient mountain belts of the Scottish highlands that the British geologist George Barrow established, in work published in 1893, the idea of metamorphic zones (later known as ‘Barrovian’ zones). Here, each zone is an area of rocks characterized by the appearance of a key mineral or minerals produced by changes in the original rock caused by the higher temperatures and pressures of metamorphism. With increasing temperatures and pressures, which were later attributed to increasing depth of burial, Barrow identified zones he named chlorite, biotite, garnet, staurolite, kyanite, and sillimanite after the key minerals in rocks that were originally mudstones. Kyanite and sillimanite are typical metamorphic minerals. They are minerals of the same chemical composition (Al₂SiO₅) but different crystal structures. Whereas sillimanite only forms at relatively high temperatures, kyanite forms at relatively high pressures.

Metamorphic zones were later defined in much more detail by the great Finnish geologist Pentti Eskola who identified the ranges of pressure and temperature leading to different groups (‘assemblages’) of minerals being produced from the same initial rock composition. He observed the sequence of minerals produced during metamorphism of basaltic rocks, leading to the concept of metamorphic facies. At a specific range of temperature and pressure (a specific facies) a metamorphosed basaltic rock will contain a specific assemblage of minerals. If the temperature and/or pressure changes, the mineral assemblage changes. Most regional metamorphism is also accompanied by deformation due to tectonic stresses, which produce an alignment of the minerals formed as a result of reactions due to increased temperature and pressure. This can produce rocks such as slates, where the alignment of minerals means that the rock can be easily broken (‘cleaved’) in one direction to produce the very thin flat sheets widely used as roofing materials. Slates are produced by the regional metamorphism of mudstones.

The large-scale processes involved in regional metamorphism are now understood in the framework of plate tectonics and its role in mountain building. In what are amongst the most remarkable achievements in geological research over the past half-century, it has been possible to unravel the detailed geological histories of great mountain belts such as the Himalayas. This has been possible by combining large-scale field studies with the information gained from laboratory investigations of the stabilities of metamorphic minerals and their uses as indicators of temperatures and pressures at their time of formation (as so-called ‘geothermometers’ and ‘geobarometers’). Another important contribution to the ‘detective work’ needed to unravel these histories has come from advances in the age dating of minerals. A good example of such work involves the mineral zircon (ZrSiO₄) which is a common minor mineral in many igneous and metamorphic rocks. This mineral contains trace amounts of uranium and thorium, which are slowly decaying radioactive elements used in radiometric dating (i.e. determining the age of the mineral since formation by determining the extent of such radioactive decay).

Journey to the centre of the Earth

Although Hollywood movies and the classic fiction of Jules Verne might suggest that humans can directly explore the Earth’s interior, the deepest mines have barely reached 4 km beneath the surface and deepest boreholes only 11 km, whereas the average distance from the surface to the centre of the Earth is 6,371 km. Increases in temperature and pressure with depth impose practical limitations, with pressures at the centre of the Earth calculated to be nearly 4 million times greater than pressure at the surface and temperatures estimated at 4,300°C. Therefore, our understanding of the interior is based on indirect evidence, and on experiments or computer modelling.

The most important evidence comes from the study of earthquake waves. The velocity at which an earthquake wave travels through a material such as a mineral is a fundamental property which can be measured in the laboratory; it is a property directly related to its ‘stiffness’ and its density. The stiffer the material the faster the earthquake (seismic) wave velocity. The waves from a major earthquake pass right through the Earth and are detected at hundreds of seismic stations scattered over the globe. The data acquired from these stations are used to determine how seismic velocities change with depth in the Earth. As we would expect, there is an increase in seismic velocity with depth attributable to the increase in density and stiffness as the materials are compacted under the enormous pressures which must prevail at greater depths.

Two other key observations from such studies are that there are relatively abrupt changes in velocity with depth, showing that the Earth is a layered body made up of concentric shells, and that the layer that surrounds a solid inner core is actually a liquid outer core. We know this because a certain type of seismic wave (a shear or ‘S’ wave) is not transmitted through it. The shells identified in this way and their thicknesses are shown in Figure 11a and, in Figure 11b, how seismic wave velocity changes with depth in the Earth is also shown. The shells consist of a solid inner core and liquid outer core, a solid mantle divided into a lower mantle, a transition zone and an upper mantle, the uppermost part of which is plastic (the asthenosphere), all overlain by a rigid lithosphere (the upper part of which is called the crust). The crust is also different on the continents and beneath the oceans, with the average compositions of rocks beneath the oceans being richer in iron and magnesium and hence essentially basalt, and the continents richer in silica and therefore essentially granite. Given this evidence for an Earth made up of concentric layers, what are the changes in mineralogy that take place when we pass from one layer into another?

11. The interior of the Earth: (a) showing the concentric shell structure

11. (b) showing phase transformations in the mantle and evidence from changes in seismic wave velocities

The seismic properties offer valuable information to guide our ideas about the minerals making up the rocks that form these successive shells. Other clues come from a number of sources. In what has been described as perhaps the most important experiment of 18th-century science, Henry Cavendish determined the universal gravitational constant. He did this by measuring the force of attraction between objects of large mass (large lead spheres). From this and the Earth’s radius, determined from measuring the curvature of the Earth, it was possible to determine the mean density of the Earth as 5.52 (grams per cubic centimetre). As rocks at the surface have an average density of 2.7, much of the interior must have a density greater than 6.0, in line with our ideas of an internally layered body. Although we cannot directly sample rocks from depths greater than a few kilometres (km), samples of rocks from the upper mantle can be carried to the surface by volcanic activity. Commonly these are brought up as lumps of rock termed xenoliths (literally ‘foreign rocks’) in a magma and are dominated by magnesium and iron silicate minerals of the olivine and pyroxene families.

An important example of a rock originating in the mantle and brought to the surface by volcanic activity is kimberlite. This rock is also important because it is the main source of diamonds (and is named after the famous Kimberley diamond mines in South Africa). Coming from depths of around 150 km, these rocks contain a wide range of xenoliths, including olivines and pyroxenes from the mantle, as well as material entrained from surrounding rocks at shallower depths. Kimberlites are mostly found in the form of volcanic chimney structures (called pipes) which record the routes along which they blasted a way to the surface, probably powered by admixed volatile gases made up largely of steam and carbon dioxide. The presence of diamonds in some kimberlites further testifies to their origins at great depths, diamond being a dense and compact form of carbon that can only be synthesized at very high pressures, as we have already discussed (see Figure 1c). The origins of kimberlites, and why only some of them contain diamonds, remains something of an enigma. This is partly because there are no examples of such rocks actually being formed at the present day.

More indirect clues to the minerals making up the interior of our planet come from the study of meteorites. These fragments left over from the material that accreted (‘clumped together’) to form the rocky planets of our Solar System sometimes survive the journey from elsewhere in the solar system, such as the asteroid belt, and crash to Earth. There are two main types of meteorites. The stony meteorites are made up of silicate minerals, mostly olivine and pyroxene, and so resemble the rocks of our upper mantle. The iron meteorites are largely made up of iron alloyed with some nickel. They provide strong indirect evidence for an Earth’s core that is dominantly composed of iron, probably with some nickel.

There are two additional complications we need to address in any attempt to understand the make-up of the core; the first is the seismic evidence for a liquid outer core which is also required to explain the Earth’s magnetic field. It is suggested that the magnetic field arises from the motions of a layer of molten iron acting as a dynamo. Only in this way can the complexities of the Earth’s field be explained. These complexities include changes in the direction and intensity of the magnetic field at Earth’s surface from place to place and on a timescale of only hundreds of years, and the reversals of the polarity of the Earth’s magnetic field (magnetic north becoming south and vice versa) which have happened on numerous occasions over geological time (magnetic reversals). The second complication is that the seismic evidence points to an inner core which has a density somewhat less than iron itself, so that there must be another element (or elements) present in the core. The identity of the light element contribution to the core is still a matter of debate; sulphur, silicon, and potassium have all been suggested.

Experiments at high pressures

One of the most important areas of research on minerals concerns experiments at high pressures. Because of the links with geophysics and with the physical properties of materials, this area is a major part of the field often termed mineral physics. The importance of this area will be obvious, given that the overwhelming majority of minerals are at high pressures in the Earth’s interior. Topics of particular interest are pressure-induced changes at the boundaries between the different ‘shells’ making up the regions of the Earth’s interior, and changes in mineralogy when lithospheric material is subducted into the mantle at a destructive plate boundary. Less dramatic, but no less important, are the measurements of physical properties such as seismic wave velocities at high pressures in suspected mantle minerals.

There is a distinguished history of research in high pressure mineralogy closely linked with materials research at high pressures. The earlier work involved equipment such as a hydraulic press forcing anvil jaws together, and with the sample surrounded by a furnace to simulate the high temperatures at depth in the Earth (see Figure 12a). Such equipment and more sophisticated rigs developed from the early apparatus were capable of simulating conditions at depths of about 300 km, roughly halfway down through the upper mantle. Much greater pressures, reaching those of the core, were also achieved for very brief periods of time by using explosives to generate shock waves, or by firing a bullet into a mineral target. Although the compression in such experiments lasts for only a few millionths of a second, it is enough time to determine density, pressure, and the parameters that provide information on seismic wave velocities.

Recent decades have seen great advances in high pressure experimentation. These have been particularly associated with a device called the diamond anvil cell (see Figure 12b). Small enough to be held in one hand, this device contains two gem quality diamonds, cut so that two perfectly flat faces can be mounted face to face with a microscopic mineral sample between them. A little thumbscrew is then tightened to apply the pressure. Because all of the pressure is applied to a tiny area of sample, enormous pressures (more than 3 million times atmospheric pressure) can be achieved. Visible light, X-rays, or other forms of radiation can be transmitted through the diamonds and through the sample, which can also be heated using a laser beam.

12. High pressure experiments: (a) a simplified drawing of a ‘traditional’ high pressure rig

12. (b) a diamond cell

High pressure experiments, such as those involving the diamond cell, have been especially successful in providing ideas as to the mineralogy of the mantle, and the causes of the discontinuous changes in seismic wave velocities with depth (as seen in Figure 11b). Olivine is believed to be the major component of the upper mantle; some estimates suggest it comprises as much as 58 per cent of the mineral inventory. A sample of olivine compressed in a diamond anvil cell will undergo changes associated with a reorganization of its constituent atoms (phase transformations). In such transformations, there is no change in chemical composition, but there is a rearrangement of the atoms to produce a denser and more compact crystal structure. At a pressure equivalent to a depth in the Earth of 410 km, such a transition takes place to a mineral phase called wadsleyite, β-(Mg,Fe)₂SiO₄, or simply the ‘β phase’.

This transformation to a more compact structure involves a 6 per cent increase in density. Experiments show that wadsleyite then transforms at a pressure equivalent to 500–550 km depth to a mineral called ringwoodite, also referred to as γ-(Mg,Fe)₂SiO₄. This phase has the same structure as the mineral spinel, MgAl₂O₄, but there is only a small density increase of ~2 per cent associated with this transition, an increase which is not generally enough to produce a resolvable seismic discontinuity (see Figure 11b). Wadsleyite and ringwoodite are found in meteorites, adding further weight to the idea that they might be important phases in the mantle (and providing the natural occurrences essential for them to be given mineral names).

At 660 km depth in the Earth, there is another major discontinuity with a large increase in both seismic wave velocity and in density (5 per cent). Again, experiments at pressures equivalent to this depth show a change in mineralogy, but this time ringwoodite (spinel) changes to a mixture of two phases. One is a silicate, (Mg,Fe)SiO₃, with the same structure as the mineral perovskite, CaTiO₃, and the other is not a silicate but a simple iron and magnesium oxide mineral called magnesiowüstite, (Mg,Fe)O, related to periclase (MgO) and having the very simple and compact ‘halite structure’ found also in galena (see Figure 1a). A point to emphasize here is that these are experimentally verified phase transformations. They occur at pressures equivalent to exactly where there are abrupt changes in seismic wave velocities within the Earth. Although it cannot be proved beyond all doubt, this strongly suggests that the seismic discontinuities at 410 km and 660 km depth are caused by changes in crystal structure rather than a reaction leading to an entirely new assemblage of mantle minerals.

The high pressure phase transformations of olivine, commonly referred to as the olivine → ‘spinel’ transition (despite the role of wadsleyite, which actually has a different structure to spinel) are amongst the most important in all Earth sciences. Olivine, (Mg,Fe)SiO₄, an ‘island silicate’ mineral, has quite an open structure in which the SiO₄ tetrahedral units do not share any of their oxygens with other tetrahedra but are held together by Mg and Fe atoms. The spinel structure of γ-(Mg,Fe)₂SiO₄ is more dense and compact; it takes up about 8 per cent less space than olivine. The transformations to a mixture of silicate perovskite and magnesiowüstite, represent further compaction. This involves certain of their atoms being bonded to, and therefore surrounded by, a larger number of oxygen atoms (which are more compressible).

Although the transformations of olivine are the best known and most researched of those believed to involve mantle minerals, they are certainly not the only such possible transformations. Garnets and pyroxenes are believed to be important phases in the upper mantle (possibly comprising 10–20 per cent) and these transform to denser phases; for example, at pressures equivalent to depths of 350–450 km, pyroxenes transform to a garnet-structured mineral known as majorite garnet with a density increase of about 6 per cent. This majorite garnet transforms to a perovskite structure at similar pressures to the transformation of ringwoodite (spinel) to perovskite (in this case equivalent to a depth of 650–680 km). Each of these transformations entails a progressive increase in the number of oxygen atoms surrounding the silicon atoms from 4 to 6, and hence a more compact, denser structure.

There is one other important part of the story of high pressure phase transformations. It concerns the mechanism and the kinetics (i.e. the ‘speed’) at which a crucial phase transformation such as that of olivine → ‘spinel’ takes place. At destructive plate margins, olivine-bearing rocks are dragged down into the mantle (subducted) and, at some depth, the olivine will transform to a ‘spinel’-type high pressure form of greater density.

Ultrafine grained material forming as a transient product of this reaction (before grain growth occurs) may facilitate deformation by a process of sliding along grain boundaries, producing fault zones. This process is thought to be responsible for deep-focus earthquakes. The point at which a high pressure phase transformation starts is determined by both the depth (pressure) and temperature. It is regarded as unlikely that nucleation and growth of the new phase would occur at temperatures lower than about 700°C, the expected temperature at the top of the transition zone. A subducted slab of lithosphere is likely to be colder than the surrounding mantle rocks, and it is likely that a metastable wedge of olivine-containing rock may persist well into the transition zone. As the slab heats up, the transition may occur very rapidly, with the weakening due to the formation of ultrafine reaction products leading to a release of energy as seismic waves, detected as deep focus earthquakes in the downgoing slab of lithosphere.

The deeper parts of the Earth’s interior have an important role in the plate tectonic cycle. Although the mantle is solid rock, it is hot and weak enough below the lithosphere to flow like a viscous liquid. Heat from the decay of radioactive elements and from the still molten outer core can cause local heating at depth. The heated mass of rock expands, becomes less dense, and rises very slowly. To compensate for this rising mass, rock that is cooler and denser must sink downwards in a process of convection (see the arrows in Figure 10). The very large scale of these processes is indicated by seismic data used to ‘map’ the mantle. In seismic tomography, data can be obtained for the Earth’s deep interior rather as we now obtain data on human organs using a body scanner. In fact, the mantle is found to be very heterogeneous. Also, the measured rate at which heat reaches the Earth’s surface can only be accounted for if this heat comes from the deep mantle and core.

There is still much to learn about the deeper parts of the Earth and the fate of subducted slabs of lithosphere. For example, just above the core is a layer, perhaps 100–400 km thick, termed the D″ layer. It is very heterogeneous, both vertically and laterally, and suggested to be the remains of lithospheric slabs that have sunk to the base of the mantle. Here, mixing between molten iron from the core and high-pressure silicates could take place, and experiments suggest vigorous reactions producing a mixture of magnesium perovskite, wüstite (FeO), plus a high pressure form of silica, and iron silicide (FeSi).

Convection in the mantle is regarded as the main driving force for the creation and destruction of lithospheric plates, causing the upwelling of magma to form new crust at spreading centres, and its destruction at subduction zones. Even a long way from these plate boundaries, there appear to be so-called mantle plumes, where hot material slowly rises from depth and causes a ‘hot spot’ within a plate rather than at a plate margin. The Hawaiian Islands, located in the middle of the Pacific Plate, are an example of hot spot volcanism.

The D″ layer is also thought to be the source of the mantle plumes that give rise to hot spots, the deep mantle source of such plumes contrasting, it has been proposed, with magmas at ocean ridges fed from the uppermost mantle.

The role of minerals in the plate tectonic cycle is evident from the above discussion. The unending formation and the destruction of the lithosphere, driven by energy from the interior of the Earth, is part of a cycling of rocks and minerals that is characteristic of our dynamic planet. But this is not the whole story. To complete the story of the cycling of minerals, we now need to consider transformations driven largely by energy coming from the sun.