The mineralogy and the origin of deep geospheres: A review

Earth-Science Reviews 113 (2012) 94–109

Contents lists available at SciVerse ScienceDirect

Earth-Science Reviews journal homepage: www.elsevier.com/locate/earscirev

The mineralogy and the origin of deep geospheres: A review D.Yu. Pushcharovsky a,⁎, Yu.M. Pushcharovsky b a b

Faculty of Geology, Lomonosov Moscow State University, Leninskie gory, Moscow, 119234, Russia Geological Institute, Russian Academy of Sciences, Pyzhevskii per. 7, Moscow, 119017, Russia

a r t i c l e

i n f o

Article history: Received 26 September 2011 Accepted 6 March 2012 Available online 23 March 2012 Keywords: Earth's model Mineral transformations Mantle mineralogy Evolution of geospheres

a b s t r a c t The structure and composition of inner geospheres are considered in light of new data on the structural transformations of minerals under high pressure. More than 100 tetrahedral complexes in silicates of the Earth's crust give way to no more than 20 structural types of minerals of this class in the Earth's mantle. The main difference in their structures is associated with the transformation of Si tetrahedra into Si octahedra. New data on the structural transformations of minerals in deep geospheres indicate that the mineralogical diversity of the Earth's crust is substantially richer than that of deep geospheres; however, mantle mineralogy is not as primitive as was supposed even twenty or thirty years ago. The results of recent seismological investigations and quantum-mechanical calculations allow the assumption that there exists a new previously unknown phase transformation under the conditions in the Earth's inner core. According to seismic tomography maps for various depth levels and available data on geophysical discontinuities, mineral composition, and phase transformations in the mantle, it is proposed that the latter can be subdivided into six geospheres. The cornerstone of the new concept lies in the recognition of middle mantle between 840 and 1700 km, which is separated from the upper and lower mantle by boundary zones, 170 and 500 km thick, respectively. The first approach to the origin and geological history of deep-seated geospheres is discussed. © 2012 Elsevier B.V. All rights reserved.

Contents 1. 2. 3. 4. 5.

Introduction . . . . . . . . . . . . . . . . . . Retrospect . . . . . . . . . . . . . . . . . . . Seismic data and the model of the Earth . . . . . Composition of the upper mantle geospheres . . Mineral transformations in the upper mantle . . . 5.1. Depths 24–410 km . . . . . . . . . . . . 5.2. Depths 410–670 km . . . . . . . . . . . 6. Deep mantle blocks . . . . . . . . . . . . . . 7. Middle mantle . . . . . . . . . . . . . . . . . 8. Division zone II and low mantle . . . . . . . . . 9. Mineral composition at the mantle–core boundary 10. Mineral and chemical compositions of the core . . 11. The origin and evolution of the geospheres . . . Acknowledgments . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (D″-layer) . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

1. Introduction The problem regarding the composition, formation and structure of deep geospheres is one of the most acute problems encountered in

⁎ Corresponding author. E-mail address: [email protected] (D.Y. Pushcharovsky). 0012-8252/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.earscirev.2012.03.004

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . .

94 95 96 98 99 99 99 100 101 103 104 105 106 107 107

geology over the past decades. This is a priority direction in geosciences. In particular, we note that more than 90% of matter in the Universe is under a pressure of more than 1 GPa. Even at present, the newest geophysical data, in combination with the results of theoretical investigations and the structural transformations of minerals, have provided a means for simulating many specific features of the structure and composition of deep geospheres, as well as the processes occurring in them. Analysis of these data favors the solution of key problems in

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

modern natural science, such as the formation and evolution of the planet, geodynamic conditions of the Earth's crust and the mantle, evaluation of the risk of deep burial of hazardous wastes, Earth's energy resources, and a number of other important problems. According to X. Mao and R. Hemley, the well-known researchers of mantle mineralogy (Mao and Hemley, 1998), the majority of minerals crystallize at high pressure conditions. It is well known that many minerals of the Earth's crust crystallize from magmatic melts or accompanying solutions formed in deep geospheres. The formation of minerals in impact zones formed under the incidence of cosmic bodies is also associated with high pressures. Moreover, the main part of the low Earth's crust and upper mantle (lithosphere) contains the minerals that lie at the high depths and exist under the conditions of strong compression. Evidently, it is these minerals that bear information on the formation and evolution of our planet and, hence, on the Earth's past and future. In this respect, investigation of the structure, properties, and transformation of minerals subjected to high pressures plays a key role in the solution of fundamental problems in the physics and chemistry of the Earth. In recent years, a large number of studies have been devoted to analysis of the mineralogical diversity of the Earth's mantle (Pushcharovsky, 1986; Navrotsky, 1994; Pushcharovsky and Pushcharovsky, 1999; Fiquet, 2001; Pushcharovsky, 2002; Pushcharovsky, 2004; Oganov et al., 2005c; Pushcharovsky and Oganov, 2006). However the array of experimental and theoretical data increases year by year, which poses the problem as to their generalization and systematization. Elucidation of this problem is the purpose of the present review. 2. Retrospect High-pressure mineralogy is an interdisciplinary field of modern science and has actively developed over the last seventy years. An English crystallographer, J. Bernal (Bernal, 1936) was the first to suggest that, in the mantle, conventional olivine becomes stable in the form of a polymorphic modification with a spinel structure whose density is 9% higher. This transformation underlays the mineralogical interpretation of the variations observed in the seismic wave velocities at a depth of 400 km, i.e., at the “upper mantle–transition zone” boundary (Jeffreys, 1937). The conclusion formulated by J. Bernal was based on the previous inferences by V.M. Goldschmidt (Goldschmidt, 1931) about the dimorphism of the Mg2GeO4 germanate (chemically similar to the olivine) crystallizing in the olivine or spinel structural type. In addition, somewhat more recently, F. Birch (Birch, 1939; Birch, 1952) studied the elastic characteristics of the mantle material and concluded that the deep zone located between 300 and 900 km is characterized by a series of phase transformations, one of which is the transformation of the olivine into a modification with a spinel structure. In subsequent investigations in the field of the high pressure crystal chemistry and mineralogy of deep geospheres, the period from the early 1950s to the middle 1960s is commonly termed exploratory. Its beginning is associated with 1953, when Coes synthesized a new silica modification (subsequently called coesite) at a pressure of 3 GPa. In 1954, crystals of synthetic diamond were prepared at a pressure of 5 GPa by General Electric researchers. At present, both minerals have been considered as indicators of a high-pressure metamorphism, thus implying a relation of rocks containing them to the depth corresponding to the upper mantle. By developing the ideas put forward by V. Goldschmidt and J. Bernal, A. Ringwood in 1959 described the transformation of the fayalite Fe2SiO4 into the polymorphic modification with a spinel structure under a high pressure. More recently, the (Mg,Fe)2SiO4 modification with a spinel structure was named ringwoodite. In 1961, Stishov (a postgraduate student of the Faculty of Geology of Moscow State University) and Popova (a researcher of the Institute of High-Pressure Physics of the USSR Academy of Sciences) synthesized a new dense silica modification with a rutile structure at a pressure higher than 8 GPa, which was named stishovite (Stishov and Popova,

95

1961). This discovery showed that conventional rocks of the Earth's crust should differ radically in their mineral composition from rocks that form the transition zone between the upper and lower mantles. The problem of the mineral composition of the Earth's core is substantially more complex than that of the Earth's mantle. Unlike many deep silicates and diamond that are stable under normal conditions, dense iron modifications, which in their density and elastic properties should be similar to the core material, can be studied only directly under high pressures (“in situ”) or, more recently, theoretically on the basis of quantum-mechanical calculations. In 1964, Takahashi and Bassett described the ε-modification of iron, which appeared to be stable at pressures higher than 13 GPa. The structure of this modification is represented by the hexagonal closest packing of iron atoms. It has been assumed that this phase is the main component of the Earth's core. Subsequent rapid progress in experimental techniques favored high-pressure investigations of quite different physical and chemical properties of a large number of minerals. The obtained results, in combination with seismological data, led to the revision of the traditional model of the Earth structure and formulation of many new problems, such as the problem regarding the change in the composition at the boundary between the lower and upper mantles or the problem of the temperature inside the core (see, for example, Vocadlo and Dobson, 1999). The development of the scientific concepts in this field allowed a number of assumptions to be made concerning materials that form deep geospheres of other planets. It has been assumed that the role of upper and lower mantles in the interior of giant planets can be played by the dense gas shells (composed of compressed hydrogen and helium gasses in Saturn and Jupiter) or ice shells (containing H2O, NH3, CH4, CO2 in Uranus and Neptune). It has turned out that, at high pressures, noble gasses are not necessarily inert and can form a number of compounds, such as He(N2)11, NeHe2, Ar(H2)2. It is remarkable that, at high pressures (>30 GPa), argon and xenon appear to be in a crystalline state at higher temperatures than iron, whose melting temperature is lower than those of these elements. This implies that, at high pressures, iron can play the role of a volatile element, whereas argon and xenon can serve as “refractory” materials. It is also surprising that Fe and Mg usually isomorphous in the minerals of the Earth's crust become ordered in their structures in deep geospheres. On the contrary, iron and potassium incompatible in general positions in structures of minerals in the Earth's crust can form the alloys under high pressures. These and other data indicate that minerals of the Earth's crust and deep geospheres differ substantially in their properties and structures. Note that, during analysis of new results obtained in high-pressure mineralogy, it is necessary to take into account one important circumstance. The majority of the conclusions associated with the solution of this extremely complex problem are made on the basis of unique experiments that can be performed by researchers of a small number of leading laboratories. F. Birch (one of the founders of the modern mineralogy of deep geospheres) (Birch, 1952) emphasized that an uninformed reader should understand the ambiguity of the inferences made in these cases. With some irony, he offered examples of terminological equivalents that are characteristic of publications concerned with complex high-pressure experiments and their interpretation at the level of standards applied to conventional investigations. In particular, Birch wrote that what is considered a “vague suggestion” in conventional publications is treated as a “positive proof” in papers with results of high-pressure investigations. Similarly, the adjective “dubious” will be substituted by “certain” and the expression “uncertain mixture of all elements” in the Earth's core will be replaced with “pure iron”. The experiments performed in the past decades revealed that chemically different minerals are characterized by different compressibility. The quantitative evaluation of the compressibility requires the use of the so-called equations of states P = f(V,T), which relate the pressure, volume, and temperature. In particular, the equation of state for an ideal gas has the form P =RT/V,

96

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

Fig. 1. Typical compressibility curves for crystals (Mao et al., 1991).

where P, V and T are the pressure, volume, and temperature of the gas, respectively; and R is the gas constant (the difference between the heat capacities at constant pressure and constant temperature) equal to 8.314 J/(mol K). For liquids and solids, it is common practice to separate the temperature and pressure effects; that is, P ðV; T Þ ¼ P 300

K ðV Þ

þ P thermal ðV; T Þ;

where P300 K is the dependence of the pressure on the volume at room temperature and Pthermal is the pressure associated with the thermal expansion. In this case, the dependence of the volume on the temperature is described by the equation: V ¼ V o ½1 þ aðT−T o Þ; where the thermal expansion coefficient α is a function of the temperature: α ¼ αо þ α1 Т þ … At relatively low pressures (lower than 1 GPa), the compressibility β = ΔV / VΔP and its reciprocal referred to as the bulk modulus

(K = 1 / β) are virtually constant and the equation of state can be written in the form V = Vo (1 − P / K). Fig. 1 shows the relative changes in the volume of the (Mg,Fe)SiO3 mineral with the perovskite structure, wüstite, and halite with an increase in the pressure. It can be seen from Fig. 1 that the compressibility of the halite is higher than those of the oxides due to the weaker interatomic bonds in the halite structure. A more general theory of the equations of state for solids has been developed in detail, but it is based on more complex approaches (Zharkov and Kalinin, 1971; Holzapfel, 2001; Oganov et al., 2002; Dorogokupets and Oganov, 2007). The analysis of the results obtained with the use of high-pressure X-ray cells has made it possible to formulate a number of highpressure high-temperature crystal chemical principles that expand scientific concepts regarding the state of matter in deep geospheres (Urusov and Pushcharovsky, 1984). In addition to these concepts, new crystal chemical data obtained in the last years allow one to relate structural transformations of minerals to seismological models (Dziewonski and Anderson, 1981; Anderson, 1989) (Figs. 2 and 3) that characterize a change in the seismic wave velocities, density, and pressure in deep geospheres. 3. Seismic data and the model of the Earth First of all, let us consider the notion of geosphere proposed by Murray. This notion, like many other fundamental notions of the Earth sciences, was frequently used by Vernadsky (2001), who wrote that geological shells and geospheres sharply differ in “material and energetic states of the matter” but are not isolated from each other because of material exchange between them; he also emphasized that not all of them are continuous. As regards the solid Earth, it is quite reasonable to subdivide it into megascopic shells with specific physicochemical and geological features, referred to as geospheres. As mentioned above, the recognition of geospheres enables a more differentiated study of the Earth's deep structure and the processes within it. The definition of the Earth's mantle arose as a response to the discovery of the Earth's core by H. Jeffreys and B. Gutenberg. Their reports were published in the beginning of the XX century. Since that time, the entire domain lying above the core with upper boundary formed by the base of the Earth's crust, i.e., the Moho discontinuity discovered in 1909, is referred to as a mantle. Thus, the trinomial

Fig.2. Change in the volume proportions of the minerals with an increase in the pressure (depth) (Ono and Oganov, 2005). Designations: DZ—division zones, CMB—core–mantle boundary, Cpx + Opx—clinopyroxenes and orthopyroxenes, HS—high-spin states of iron atoms, and LS—low-spin states of iron atoms.

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

97

was identified and interpreted as the low level of the upper mantle. Considering the different thicknesses of the Earth's crust under the continents and the oceans its upper boundary is approximated by the depth of 24 km. The discovery of the inner (solid) and the outer (liquid) cores, which was made by I. Lehmann in 1936, was of fundamental importance; it has served as a basis for all further investigations of the Earth's interior, and associated with abrupt changes in geophysical properties and composition of the mantle. Five subdivisions of the Earth were eventually recognized, i.e. the inner core, outer core, lower mantle, upper mantle, and the crust. That scheme was adopted as a basis for all further studies, including recent ones (Fig. 4a). The part of the upper mantle which lies above the asthenosphere and the crust is jointly referred to as the lithosphere. This notion is fairly uncertain as well as the relationship between the lithosphere and asthenosphere, because the latter is becoming more and more abstract notion itself. This is because the mantle contains several low-viscosity layers in some vertical sections, whereas in other sections such layers are not detected at all. The term “lithosphere” implies a solid state and rigidity of material, which is hardly acceptable taking into account the existence of heat and mass flows, magma chambers, and tectonic flow of rocks. The commonly cited lithospheric thickness of an estimated 100 km is incorrect. Asthenospheric layers beneath oceans are located at a much shallower, albeit varying depth; beneath continents, whose roots occur at various depths, sometimes as deep as 500 km, these layers are traceable at a greater depth, which also varies. The scientific works of K.E. Bullen contributed greatly to the knowledge of the Earth's deep structure. His studies began in 1940, but principal generalizations were published in 1963 and 1975; these monographs were translated into Russian in 1966 and 1978, respectively (Bullen, 1947, 1975). On the basis of seismic and density data, Bullen recognized

Fig. 3. Profiles of (a) the longitudinal (Vp) and shear (Vs) seismic wave velocities, (b) the density and (c) the pressure in the Earth's interior according to the PREM (Preliminary Reference Earth Model) (Dziewonski and Anderson, 1981).

subdivision of the Earth into the core, mantle, and crust was adopted, and it remains valid to this day. In 1904, C.R. Van Hise introduced the term “the asthenosphere”, and, in 1914, J. Barrell defined it as a layer of relatively low viscosity, hardness, and strength in the upper mantle. Subsequently, the knowledge of the asthenosphere was extended, and currently the idea of a single continuous asthenospheric layer is no longer credible. B.B. Golitsin, who initiated the seismic study of the Earth's interior, made a great contribution to the research of the mantle. He established rapid increase in seismic wave velocity at a depth of 400 km. Subsequently, the depth interval of 400–900 km was defined as the Golitsin layer. At present, this term, if used, refers to the interval of 400–670 km. In middle of last century, the boundary at the depth of 670 km discriminating between the upper and lower mantles

Fig. 4. Earth's structure models; (a) traditional currently widespread model; (b) new model based on the analysis of seismotomographic maps and data on the mantle composition and seismic boundaries. The thicknesses vary within 10% (Pushcharovsky, 1995).

98

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

seven zones within the Earth: (A) the crust, (B) zone in the interval of 33–413 km, (C) zone in the interval of 413–984 km, (D) zone in the interval of 984–2898 km, (E) 2898–4982 km, (F) 4982–5121 km, and (G) 5121–6371 km (the Earth's center). Later, Bullen subdivided zone D into zones D′ (984–2700 km) and D″ (2700–2900 km). This scheme was subsequently, modified, and only the layer D″ has been firmly accepted in the literature. Its principal specific feature is a sharp decrease in seismic velocity gradient as compared with the overlying mantle. In recent works, the interval of 410–670 km is referred to as the Golitsin layer, middle mantle, or the transitional zone (between the upper and lower mantles). Because the terminology is not yet fixed, the upper (30–670 km) and lower (670–2900 km) mantle are still regarded as the principal units of the mantle. The intermediate frontiers are fixed at different depths. The lowest one was registrated at the depth of 100 km (Pavlenkova, 1995) under the continents (level N). At the same time, a new, much more detailed scheme of mantle stratification based on seismic tomography has been elaborated since 1995 (Pushcharovsky, 1995, 1996, 1997, 1998). This scheme is based on recognition of a thick middle mantle in the interval between 840 and 1700 km, separated from the upper and lower mantles by discontinuities with a complex topography. It is now evident that the subdivision of the mantle cannot be restricted to the lower and upper mantles; more detailed subdivision is necessary for the understanding of deep tectonics and geodynamics. We now have broader foundations for a new stratification, which will be discussed below. During the last decades, seismology turned from the recognition of seismic discontinuities to the compilation of global seismotomographic maps for different depths. They yield graphical displays of deep seismic inhomogeneities, which, in combination with the data on the mantle composition and discontinuities, enable a higher degree of generalization and modeling related to the deep structure of the Earth and deep-seated processes. A.M. Dziewonski and J.H. Woodhouse in 1988 during the XXVIII Congress of IGC (Dziewonski and Woodhouse, 1988) exhibited the maps, which reflected the heterogeneities in seismic wave velocities and thus mark the mantle masses of variable viscosity. The high-velocity anomalies are inherent to the masses of elevated viscosity, whereas the low-velocities anomalies are indicative of a decrease in viscosity (Fig. 5). The first seismotomographic models were constructed for Indo-Atlantic segment of the Earth whereas the models of Pacific segment (Su et al., 1994) were created later on. Thus it became possible to separate volumetric anomalous zones and to draw their profile sections. These data give new insight on the structure of the Earth's interior which is one of the most intriguing among modern geological problems. Until recently, the Earth's mantle was believed to consist of two shells—lower and upper—separated by a boundary at a depth of 670 km. However, such a subdivision seems to be oversimplified in relation to any comprehensive understanding of the deep geodynamics of the Earth, which is a complex open system. According to seismic tomography maps for various depth levels and available data on geophysical discontinuities, mineral composition, and phase transformations in the mantle, it is proposed that the latter can be subdivided into six geospheres. The key point of the new concept lies in the recognition of the middle mantle between 840 and 1700 km, which is separated from the upper and lower mantles by boundary zones, 170 and 500 km thick, respectively. The upper mantle is subdivided into upper and lower parts with a boundary at a depth of 410 km. Both models are shown in Fig. 4. It is worthy to note that according to Saltzer et al. (2004) the correlation between vS and vP variations fixed in the middle mantle begin to degrade while the depth decreases from 1500 to 1000 km and it can be considered as an indication of mineralogical transformations in this interval. 4. Composition of the upper mantle geospheres

Fig. 5. Seismotomographic models for the depths of 1300, 2300 and 2750 km, after Dziewonski and Woodhouse (1987), which illustrate the nonhomogeneous structure of the low mantle in its traditional consideration. The scale indicates seismic wave velocity deviation from mean values, %.

same elements prevail in the other deep geospheres (Table 1). These data allow concluding, that the structure transformation of minerals but not the variability of chemical composition determine the most specific features of the deep-seated geospheres. The proposed models of geosphere compositions are largely based on differences in major element ratios (Mg/(Mg + Fe)= 0.8–0.9; (Mg + Fe)/Si = 1.2–1.9), Al abundances, and the abundances of some elements, which are not so common in mantle rocks. The models were named according to their chemical and mineral composition: (1) pyrolitic (olivine, pyroxenes, and garnet in proportion 4:2:1 as major minerals); (2) pyclogitic (pyroxene and garnet as major minerals complemented by a 40% of olivine), and (3) eclogitic with a pyroxene–

Table 1 Estimates of elemental abundances (in atoms per 1 atom Si). According to Oganov and Brodholt (2000). Element

The Universea

Whole Earthb

Earth's crustc

Mantle (increase of depth)c

Pyrolitic homogeneous mantled

O Na Mg Al Si P S Ca Cr Fe Ni

20.10 0.06 1.08 0.08 1 0.01 0.52 0.06 0.01 0.9 0.05

3.73 0.06 1.06 0.09 1 – – 0.06 – 0.9 –

2.9 0.12 0.09 0.36 1 4 · 10− 3 8 · 10− 4 0.14 1 · 10− 4 0.11 3 · 10− 5

3.63–3.63 0.03–2 · 10− 3 0.97–1.09 0.17–0.06 1 6 · 10− 4–4 · 10− 5 6·10− 4–5 · 10− 5 0.05 0.01 0.14 4 · 10− 3

3.68 0.02 1.24 0.12 1 4 · 10− 4 2 · 10− 3 0.09 0.01 0.16 3 · 10− 5

a b

According to contemporary views, a relatively small group of elements (Si, Mg, Fe, Al, Ca, and O) is predominant in the mantle. The

c d

Estimates of Anders and Ebihara (1982). Simple model based on cosmic abundances (Anderson, 1989). Recalculated from data of Anderson (1989). Recalculated from Ringwood (1991).

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

garnet assemblage typical of eclogites supplemented by some more rare minerals, e.g., Al-bearing kyanite Al2SiO5 (up to 10 wt.%). However, all these petrological models are mainly related to the upper mantle down to ~ 670 km. As regards the bulk composition of deeper geospheres, it is only assumed that the ratio of bivalent elements (M 2 +O) versus silica (MO/SiO2) ~ 2 thus being closer to olivine, (Mg,Fe)2SiO4 than to pyroxene (Mg,Fe)SiO3, whereas dominant minerals are thought to be the perovskite-type phases (Mg,Fe) SiO3 possessing different structural deformations, magnesiowüstite (Mg,Fe)O with NaCl structure, and some other minor phases discussed below. All models are very generalized and hypothetical. The pyrolitic model of the upper mantle with olivine as the dominant mineral suggests a much greater similarity in composition with the deeper parts of the mantle. The pyclogitic model, on the contrary, implies a certain chemical contrast between the upper and lower mantles. The more specific eclogitic model allows for the presence of eclogite as isolated lenses and blocks in the upper mantle. The important contribution to our knowledge about the composition of the upper mantle and the conditions of its formation is connected with the studies of mineral and chemical constituents of xenoliths (Ancient Greek: “foreign rock”)—rock fragments which become trapped by kimberlites during the latter's development and hardening (Sobolev et al., 1975; Wirth, 2010). Kimberlite is a type of potassic magmatic rock best known for sometimes containing diamonds. Its formation occurs at depths between 150 and 450 km, from the mantle anomalously enriched by alkalis and carbon dioxide as well as other volatile components. They are erupted rapidly and violently. At the Earth's surface they form vertical cone-shaped structures known as kimberlite pipes and dykes. The inclusions in kimberlite diamonds contain ultrabasic rocks, namely, garnet peridotite, lherzolite (magmatic intrusive rock), dunite and eclogite—a coarse-grained mafic (basaltic in composition) metamorphic rock. Peridotite xenoliths and diamonds from kimberlites represent an important source of information about the composition of the continental lithosphere at depths exceeding 120–150 km. Ultramafic (or peridotitic) type of geological environment is dominating at these depths compared to eclogitic type. Olivine is the most typical mineral both of peridotite xenoliths and as diamond inclusions in most kimberlites worldwide. In spite of its simple chemical composition it contains a number of petrogenetically significant minor elements such as Ti, Al, Mn, Ca, Cr, Ni, Co in low concentrations, mostly below 0.1 wt.% of oxide except of NiO. It is concluded that all examined olivines share similar compositional features and are formed at high pressures (Sobolev et al., 2009). Basically, there are two types of micrometer-sized inclusions in diamonds from kimberlites. The inclusions of first type contain olivine, Crdiopside and pyrope, which define a peridotitic environment of diamond growth connected with the presence of olivine, pyroxenes, amphiboles and mica. Another type of inclusion contains association of omphacite ((Ca,Na)(Mg,Fe,Al)[Si2O6]), pyrope-almandine garnet and coesite, which defines an eclogite, accompanying the diamond crystallization. Currently more attention is paid to nanometer-sized inclusions which consist of silicates (phlogopite), carbonate, phosphate (apatite), chlorides (KCl, NaCl) and sulfides. Such diamonds are crystallized from a fluid during their uplift and cooling. In 1964 these data and the difference in composition of xenoliths in the kimberlite pipes of the northern part of Siberian platform allowed V.S. Sobolev and N.V. Sobolev to conclude that the upper mantle is characterized by the lateral heterogeneity.

99

of mantle minerals (Fig. 3). Note that not all the boundaries determined from the seismic data are global, even though they can be associated with the transformations of minerals. For example, the first seismic discontinuity observed at a depth of ~220 km (the so-called Lehmann discontinuity) corresponds to the transformation of the monoclinic structures of pyroxenes into the orthorhombic structures (Mendelssohn and Price, 1997). One more local seismic discontinuity recorded at a relatively small depth of ~300 km is attributed to the formation of the stishovite in eclogite rocks (Williams and Revenaugh, 2005). In this case, the stishovite can be formed as a result of the transformation of the coesite or through the transformation of clinopyroxenes according to the reaction: 2(Mg,Fe)SiO3 =(Mg,Fe)2SiO4 (wadsleyite)+SiO2 (stishovite). This reaction proceeds at a pressure of 15 GPa and a temperature of 1600–1700 K, i.e., under the conditions corresponding to a depth of ~450 km. However, at lower temperatures, the reaction can occur at lower pressures corresponding to depths of the order of 300 km. 5.2. Depths 410–670 km The global boundary at a depth of 410 km is associated with the olivine–wadsleyite structural transformation. The boundary at a depth of 520 km is attributed to the subsequent transformation of the wadsleyite into the spinel-like ringwoodite. The role of the wadsleyite, as one of the most important mantle component, was originally noted by Ringwood and Major in 1970, even though this mineral at that time was not found in nature and only its synthetic analog β-(Mg,Fe)2SiO4 (Ringwood and Major, 1970) was known. These researchers proposed to name a future mineral with the above composition as wadsleyite after Australian chemist A.D. Wadsley for his work in the field of the structural crystallography of minerals and inorganic compounds. Wadsleyite, ringwoodite and the series of Mg-silicates with the structures built of (Mg, Si, O)-octahedral layers, alternating with olivine-like and humite-like polyhedral sheets are treated as the main accumulators of water in the transition zone (410–460 km), whose volume exceeds that of the World ocean (Jacobsen et al., 2005). In this case, even insignificant water content (0.1 wt.%) in the transition zone is equivalent to a water shell of 1 km thick that covers the Earth's surface. From the crystal chemical standpoint, this specific feature of the composition of the formally anhydrous wadsleyite and ringwoodite is determined by the replacement of a number of O 2 − anions with OH − hydroxyl groups in their structures. In the wadsleyite structure, a prerequisite for this replacement is the presence of the O atom that does not participate in the formation of SiO4 tetrahedra and is coordinated only by five Mg atoms (Fig. 6) (Smyth, 1994). As a result, the bond valence sum at this anion is equal to 1.67. This result can be

5. Mineral transformations in the upper mantle 5.1. Depths 24–410 km The locations of many seismic discontinuities revealed to date in deep geospheres correlate with the depths of structural transformations

Fig. 6. Wadsleyite structure along the [100] direction. Gray spheres are the protons bonded to the O(2) atoms shared between five (Mg,Fe) octahedra.

100

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

explained by the random filling of the corresponding position with OH groups (up to 33%) and O 2 − anions (no less than 67%) in combination with a distortion of cationic coordination polyhedra. In total this very complex concentric shell related to the low part of the upper mantle is located at depths from 410 to 670 km and predominantly contains phases with the garnet-, wadsleyite-, and spineltype structures and, probably, (Mg,Al)(Si,Al)O3 solid solutions with a structure of the ilmenite (akimotoite MgSiO3) type. This assemblage of minerals below the global boundary at a depth of 670 is replaced with perovskite-like phases (which occupy approximately 70% of the volume of the mantle below 670 km and Mg wüstite (20%) with the NaCl-type structure. As secondary minerals, which form the rest 10% this region can also involve high pressure silica phases, such as the rutile-like stishovite and its orthorhombically distorted analog with a CaCl2 structure, seifertite with the α-PbO2-type structure, potassium aluminosilicate KAlSi3O8 with a hollandite structure, the MgAl2O4 and NaAlSiO4 polymorphs with structures of calcium ferrite CaFe2O4 and calcium titanite CaTi2O4. The formation of these mineral phases is connected with structure transformations of minerals within the upper mantle. The attempts to correlate structural–mineralogical and geophysical data on the upper mantle are of great interest (A. Ross, 1997; N.L. Ross, 1997). As was assumed in the last 20 years, the increase in seismic wave velocity at a depth of ~410 km is mainly related to the structural transformation of olivine α-(Mg,Fe)2SiO4 into wadsleyite β-(Mg, Fe)2SiO4 producing a more compact phase with a higher modulus of elasticity. According to geophysical data, seismic wave velocity at that depth increases by 3–5%, whereas the transformation of olivine into wadsleyite and the corresponding change in the modulus of elasticity should induce an increase in seismic wave velocity of approximately 13%. At the same time, the experimental data on olivine and the behavior of olivine–pyroxene mixture at high temperatures and pressures indicate a complete similarity between the calculated and experimental velocity increases within the interval of 200–400 km. Because olivine has approximately the same elasticity as the high-density clinopyroxenes, these data may indicate the absence of garnet in the lower zone. Garnet has a high elasticity, and its occurrence in the mantle would inevitably induce a more significant increase in seismic wave velocity. However, the concept of the garnet-free mantle contradicts the petrological models of mantle composition. All this led to the idea that the abrupt change in seismic wave velocity at a depth of 410 km is mainly due to the pyroxene–garnet transformation within the Na-rich upper mantle zones. This model implies an almost complete absence of convection in the upper mantle and thus contradicts the modern geodynamic concept. These contradictions are eliminated in a recent, more detailed model of the upper mantle, which assumes the incorporation of iron and hydrogen atoms in wadsleyite crystal structure (A. Ross, 1997; N.L. Ross, 1997). Whereas the polymorphic olivine transformation into wadsleyite does not lead to changes in chemical composition, the presence of garnet induces a reaction which yields wadsleyite enriched in Fe as compared with the initial olivine. Moreover, wadsleyite may contain more hydrogen atoms than olivine. The incorporation of Fe and H into wadsleyite crystal structure promotes a decrease in its rigidity and, hence, a decrease in the velocity of seismic wave propagation through this mineral (A. Ross, 1997; N.L. Ross, 1997). In addition, the formation of Fe-enriched wadsleyite implies a greater amount of olivine involvement in the reaction and, hence, a change in bulk rock chemistry near the 410 km interfaces. The idea of such transformations is supported by global seismic data. In general, the mineral composition of that part of the upper mantle is more or less clear. As regards the pyrolitic mineral assemblage (Table 2), its transformation down to a depth of ~800 km has been studied quite comprehensively, and is shown in a generalized form in Fig. 7 (Irifune and Ringwood, 1987). The quasiglobal seismic boundary at a depth of 520 km corresponds to wadsleyite β-(Mg,Fe)2SiO4

Table 2 Mineral composition of pyrolite (Liu, 1982). Mineral

Chemical formula

Content, vol.%

Olivine (Fo89) Orthopyroxene Clinopyroxene (omphacite) Garnet (pyrope)

(Mg, Fe)2SiO4 (Mg, Fe)SiO3 (Ca, Mg, Fe)2Si2O6–NaAlSi2O6 (Mg, Fe, Ca)3(Al, Cr)2Si3O12

57 17 12 14

transformation into ringwoodite, the γ-modification of (Mg,Fe)2SiO4 with the spinel structure. The pyroxene (Mg,Fe)SiO3 → garnet Mg3(Fe,Al,Si)2Si3O12 transformation occurs within a wider depth interval in the upper mantle. Therefore, the upper mantle at a depth of 400–600 km mainly consists of the garnet- and spinel-type phases. All models of mantle rock compositions imply the presence of ~ 4 wt.% Al2O3, which also affects the character of structural transformations. In some domains of the heterogeneous upper mantle, Al may be concentrated in a form of corundum Al2O3 or kyanite Al2SiO5. The latter at the temperature and pressure corresponding to a depth of ~ 450 km, may transform into the association of corundum and stishovite— a SiO2-modification, whose framework structure is built by SO6 octahedra. Both minerals are stable not only at the base of the upper mantle, but also at deeper levels. 6. Deep mantle blocks Seismic methods have revealed a number of vertical boundaries in the mantle. Some of them are global and others are intermediary. The Mohorovicic surface and the boundaries at 410, 520, 670, 900, 1700, as well as D″-zone (2700–2900 km) are global. The ones at 100, 300, 1000, 2000 and 2200 km are intermediary. The intermediate boundaries have different lengths. The 100-km boundary is the lower level of the Earth's division into blocks that are associated with the continents (Pavlenkova, 1998). The 300-km boundary can be tracked under the continents and island arcs (Williams and Revenaugh, 2005). The profile that cuts the West Siberian Plate and the Siberian Platform shows a multilevel seismic structure to a depth of 270 km. The same occurs in Eastern Europe, where the most clear-cut boundaries are 100 and 200 km deep; however, there are also others, more peculiar, which taken together form a picture of a layered mantle (Pavlenkova, 2003). The 520-km boundary was earlier reported only in the oceans, and so it remained intermediary. Considering the latest data on the Siberian Platform, however, it would be better regarded as global (Egorkin, 2004). The nature of many boundaries remains unknown to

Fig. 7. Volumetric proportions of minerals in pyrolite versus the pressure (depth), after Akaogi (1997): (Ol) olivine, (Gar) garnet, (Cpx) clinopyroxenes, (Opx) orthopyroxenes, (MS) “modified spinel,” or wadsleyite wadsleyite β-(Mg,Fe)2SiO4, (Sp) spinel, (Mj) majorite Mg3(Fe,Al,Si)2(SiO4)3, (Mw) magnesiowüstite (Mg, Fe)O, (Mg-Pv) Mgperovskite, (Ca-Pv) Ca-perovskite, (X) inferred Al-bearing phases with ilmenite-type, Ca-ferrite, and/or hollandite structure.

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

date, even though these boundaries in many cases correspond to the structural transformations of minerals. A number of recently revealed transformations are given below. 670–850 km: (Ca0.5 Mg0.5)Al2Si2O8 (anorthite-type structure) → hollandite-type structure (Gautron and Madon, 1994); 850–900 km: Mg3Al2Si3O12 pyrope → (Mg,Fe)SiO3 (perovskite); Mg3Al2Si3O12 pyrope →Al2O3 (corundum)+ (Mg,Fe)SiO3 (ilmenite); MgAl2O4 (spinel-type structure) → MgAl2O4 (CaFe2O4type structure) (Irifune et al., 1991); ~1200 km: MgAl2O4 (spinel-type structure) → MgAl2O4 (CaTi2O4type structure) (Funamori et al., 1998); ~ 1500 km: SiO2–stishovite → SiO2 (CaCl2-type structure) (Kingma et al., 1995; Oganov et al., 2005a); FeO (metallization of interatomic bonds) (Knittle and Jeanloz, 1986); ~ 1900 km: MgCO3 (magnesite) → C2/m: with C3O9-rings (Oganov et al., 2008) ~ 2300 km: SiO2 (CaCl2-type structure) → seifertite (Dubrovinsky et al., 1997; Oganov et al., 2005a); ~ 2700 km: MgSiO3 (perovskite-type structure) → MgSiO3 (CaIrO3type structure); ~ 2850 km: MgSiO3 (CaIrO3-type structure) → MgSiO3 (perovskitetype structure); CaCO3: postaragonite → s.t. pyroxene MgCO3: C2/m → P21: with C3O9-rings Let us comment only the very recent data referred to the phase transitions listed above. The geosphere underlying the upper mantle in the depth interval of 670–850 km is readily distinguishable. This interval is marked by transformation of the mineral (Ca0.5Mg0.5) Al2Si2O8 with the anorthite structure into a mineral with the hollandite-type structure (Gautron and Madon, 1994). In our model

101

of the mantle structure, this geosphere is designated as the zone of division I that separates the upper and middle mantles. The upper part of the middle mantle (840–900 km) is distinctly distinguished by abrupt phase transformations. In particular the rearrangement of the electronic structure of alkali cations that transforms them from s- to d-elements is established already at 30 GPa (~800 km). It can be ruled out that, under conditions of the middle mantle, the Fe-bearing phases accumulate alkali metals, including K (McMahon et al., 2006). Fig. 8 reproduces the main mineral transformations in middle mantle and adjacent deep zones.

7. Middle mantle The term middle mantle was first introduced in 1995 (Pushcharovsky, 1995) and subsequently used by other authors, for example, in the papers of Niu and Kawakatsu (1997), Kaneshima and Helffrich (1999), Shen et al. (2003) etc. Pressure and temperature in the middle mantle vary appreciably from roof to bottom. According to Sorokhtin and Ushakov (2002), the pressure changes from 35.2 GPa near the upper edge to 73.0 GPa at the bottom, thus yielding a difference of 37.8 GPa (Table 3). As follows from a simple calculation, the temperatures at the upper and lower edges are 2200 and 2590 K; the difference is 390 K (Table 4). The temperature contrast between the coldest and hottest regions of the mantle, at a given depth, to be about 800 K at 1,000 km, 1,500 K at 2,000 km, and possibly over 2,000 K at the core–mantle boundary (Oganov et al., 2001). The lower boundary of the middle mantle corresponds to the seismic level 1700, which was mentioned by A. Dziewonski and coauthors in 1992 (Dziewonski et al., 1992). As a low frontier of middle mantle it was for the first time considered by Yu. Pushcharovsky (Pushcharovsky, 1996). The experimental studies of phase transformations at the above mentioned pressure and temperature parameters are not numerous.

Fig. 8. The main mineral transformations in the middle mantle and adjacent zones: s.t.—structure type; Pv—perovskite; Ilm—ilmenite; HS—high spin state; LS—low spin state; dotted lines—intergeosphere boundaries.

102

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

Table 3 Pressure variations at different depths in the mantle, after Sorokhtin and Ushakov (2002). Depth, km Pressure, GPa

200 6.55

430 13.8

670 24.72

800 30.57

1000 39.77

1200 49.17

However, the results available allow us to outline mineralogical features of the middle mantle (Pushcharovsky and Oganov, 2006). The mineral transformations in the upper part of the middle mantle (840–900 km) are very specific thus underlying its separation. Pyrope Mg3Al2Si3O12 transforms into (Mg,Fe)SiO3 with perovskite structure and apart from that—into corundum Al2O3 + (Mg,Fe)SiO3 with ilmenite structure. At the same depth spinel MgAl2O4 is replaced by its polymorph with CaFe2O4 structure type (Irifune et al., 1991). At pressures corresponding to a depth of 1200 km, the phase with the MgAl2O4 spinel composition crystallizes in a structure of the Ca titanate type. Until recently, the jump with an increase in the seismic velocity at a depth of 1200 km was associated with the structural transformation of the stishovite (Stishov and Popova, 1961) into the phase with a structure of the CaCl2 type, which is an orthorhombically distorted structural analog of the rutile (Kingma et al., 1995). The SiO2 poststishovite modification (CaCl2 analog) is stable at pressures higher than ~ 50 GPa. At room temperature, this pressure corresponds to a depth of ~ 1200 km; however, in combination with an increase in the temperature in the mantle, conditions arising at the given pressure are equivalent to those at depths of ~ 1500 km (Oganov et al., 2005a). The recently revealed structural transformations of the most important carbonates, which apparently represent the main carbon minerals in the mantle, occur at pressures corresponding to the middle mantle. It is known that the magnesite, calcite, and dolomite are the most abundant carbonates in the Earth's crust. N. Ross, G. Fiquet and co-authors (N.L. Ross, 1997; Fiquet et al., 2002) demonstrated that the magnesite is stable to a pressure of at least 80 GPa, which corresponds to depths of ~1900 km. This depth correlates with division zone II. The results of thermodynamic studies indicate that magnesite does not dissociate on periclase MgO and CO2 even at pressures of 100–130 GPa, i.e. at parameters which correspond to the low mantle and outer core conditions (Dorogokupets, 2007). Recently two new structure types of MgCO3 were predicted to be stable in the relevant pressure range: one at 82–138 GPa (1900–2900 km) and the other above 138 GPa (>2900 km) (Oganov et al., 2008). Both phases contain rings of corner-sharing CO4-tetrahedra and occur outside the middle mantle. Their predictions were largely confirmed by experiments (Oganov et al., 2008) which revealed complex high-pressure chemistry of MgCO3. Calcite transforms into aragonite in the upper mantle. Moreover, dolomite decomposes into magnesite and aragonite also in the same geosphere. Until recently, the specific features of the behavior of these minerals at depths larger than 1000 km remained unknown in many respects. Isshiki et al. (2004) found the phase transition of the magnesite at pressures exceeding 110 GPa; however, the experimental data were insufficient for determining the structure of the new phase of the MgCO3 compound. More recently, Ono et al. (2005b) revealed that the aragonite at pressures higher than ~ 40 GPa transforms into a new phase, whose structure could not be determined either experimentally or with the use of conventional methods for the theoretical simulation. Recently, Oganov and Glass proposed a new simulation method that provides a way of predicting crystal structures at any specified pressure on the basis only of the chemical composition.

1400 58.78

1600 68.6

1800 78.63

2200 99.49

2600 121.62

2886 138.4

3000 150.3

This approach was used successfully to determine the structure of the CaCO3 postaragonite phase (Oganov et al., 2006). The structure belongs to a new structural type and allows one to explain all the experimentally measured characteristics of this phase (the X-ray powder diffraction pattern, the compressibility, and the stability field). The structures of the calcite, aragonite, and new postaragonite phase are compared in Fig. 9a–c. Note that the coordination number of Ca atoms increases from 9 for the aragonite to 12 for the postaragonite phase. According to the calculations, the postaragonite phase at 137 GPa should transform into a new modification with a structure that is similar to pyroxene structures (Fig. 9d) characterized by the space group C2221, the formation of chains from CO4 tetrahedra, and a rather large coordination number of Ca atoms (equal to 10). A similar structure is also possible for the high-pressure MgCO3 phase, even though other variants of the atomic arrangement for the magnesium carbonate at these pressures cannot be ruled out. Recognition of intermediate seismic boundaries below the 670 km level correlates with the data on the structural transformations of mantle minerals, whose forms may be rather diverse, including changes in chemical bonds and electronic structure of transitional metals (transition from a high-spin (HS) to low-spin (LS) state). According to Knittle and Jeanloz (1986), the rearrangement of ionic–covalent bonds in wüstite, which are related to the metallic type of interactions between atoms under a pressure of 70 GPa (~1700 km) and revealed by experiments, illustrates the changes in many properties of various crystals under deep mantle conditions. In deep zones of the middle mantle, one may assume a change of the electronic structure of some atoms as well as their deformation and violation of sphericity. For example, similar to wüstite ferropericlase (Mg0.83Fe0.17)O reveals the transition of Fe atoms from the high-spin to low-spin states at 60–70 GPa, i.e., a depth of ~1700 km (Badro et al., 2003). This transition depends on the Fe content. While the Fe content is increasing, the pressure drops to 25 GPa. Another iron oxide, α-Fe2O3 (hematite), is transformed at 30 GPa (~800 km) into a perovskite-like modification. Under compression above 50 GPa (1200 km), this phase acquires the recently established postperovskite structure of CaIrO3 (Oganov and Ono, 2004; Ono and Oganov, 2005). At a lower temperature and a depth of ~800 km, α-Fe2O3 undergoes orthorhombic distortion and crystallizes in the s.t. of Rh2O3 (Rozenberg et al., 2002). Close to the low boundary of the middle mantle (1700 km) Dubrovinsky et al. (2001) reported the reaction between iron and corundum which led to the formation of wüstite and Fe,Al alloy: Fe + Al2O3 → FeO+ (Fe,Al)-alloy. The structural transformation of troilite FeS partly occurring at a pressure corresponding to the middle mantle (>40 GPa) is no less attractive (Ono et al., 2008). At room conditions, troilite with antiferromagnetic properties and space group P-62c crystallizes in the s.t. of NiAs. At 3.5 GPa, the symmetry of this phase changes: The hexagonal space group gives way to the orthorhombic space group Pnma; thus, modification FeS II with a structure of the MnP type is formed. The next transition at 7 GPa violates the antiferromagnetic structure and gives rise to the formation of monoclinic FeS III. The abovementioned transformations pertain to the upper mantle. The next two transformations into FeS IV and FeS V are related to the increase

Table 4 Temperature distribution in the Earth's interior, after Sorokhtin and Ushakov (2002). Depth, km Temperature, K

0 288

200 1770

430 1940

600 2130

670 2170

1000 2260

1200 2360

1400 2450

1600 2540

1800 2640

2200 2820

2600 3010

2886 3130

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

103

Fig. 9. Structures of the (a) calcite, (b) aragonite, (c) postaragonite phases stable at pressures higher than 42 GPa and (d) the CaCO3 pyroxene-like phase formed at pressures higher than 137 GPa (Oganov et al., 2006). Large spheres are the calcium atoms, intermediate spheres are the carbon atoms, and small spheres are the oxygen atoms.

in temperature, and both phases demonstrate similarity with NiAs (FeS V) or its superstructural modification (FeS IV). Modification FeS VI is related to the conditions of the middle mantle (P = 40–135 GPa and a depth below ~900 km). In contrast to FeS II, this phase is nonmagnetic but belongs to the MnP s.t. The lower boundary of

its stability corresponds to zone D″. At a greater depth, FeS VI is transformed into FeS VII with space group Pmmn and a distorted structure of the NaCl type. It has to be mentioned that FeS VII is stable up to 400 GPa and might be stable in the inner core of the Earth. As a conclusion to this part it can be noted that the new results and the performed geochemical and crystal chemical analyses primarily indicate a mineralogical variability along the Earth radius, i.e., along the vertical. However, there is a horizontal variability because geospheres are inhomogeneous in the lateral direction (Pushcharovsky and Pushcharovsky, 1999). This aspect of the problem, as before, calls for further investigation. 8. Division zone II and low mantle

Fig. 10. Structure of the MgSiO3 postperovskite modification. Magnesium atoms are located between layers formed by the silicon octahedra (Oganov and Ono, 2004).

Under a higher pressure (~96 GPa) and at a temperature of 800 K (depth ~ 2150 km), FeO polytypism has been established. This phenomenon is related to the formation of niccolite-type (NiAs) structural fragments (B8) alternating with antiniccolite domains where Fe atoms are located in the positions of As atoms, and O atoms—in the positions of Ni atoms (Mazin et al., 1998). In the same zone near the low boundary of division zone II at a pressure of ~ 90 GPa (~2100 km), the Al2O3 oxide with a corundum structure can transform into the phase with an orthorhombic structure of the Rh2O3 (II) type with a decrease in the molar volume by 4% (Funamori and Jeanloz, 1997). The specific feature of this transformation under pressure, like the transformation of the SiO2 oxide into a structure of the pyrite type (Oganov et al.,

104

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

2005a), is the disappearance of the closest anion packing and the formation of phases that contain no closest atomic packings and are characterized by a disturbance of the sphericity of ions and their strong distortion. However, the presence of the Al2O3 oxide as an individual phase in the Earth's mantle is highly improbable and the inclusion of the above transformation is important from the crystallographic rather than from the geophysical standpoint. It was demonstrated that, at a higher pressure of ~ 100 GPa, SiO2modification (CaCl2 analog) transforms into the phase with an αPbO2 structure (Dubrovinsky et al., 1997). This phase, called seifertite, has been found in meteorites. Reasoning from the results of investigations of similar compounds and theoretical simulation of the structure of silica modifications, Prewitt (2003) recently assumed that silicon oxide SiO2 with a pyrite structure can be synthesized in the immediate future, even though the formation of this phase according to the theoretical estimates requires extremely high pressures (~ 210 GPa). Actually, Kuwayama et al. (2005) synthesized this phase and refined its structure by the Rietveld method at a pressure above 200 GPa. The stability field of the phase under consideration is beyond the range of the physicochemical parameter characteristic of the Earth's mantle. At present, there are no grounds to believe that silica can occur in the Earth's core. However, this phase can be considered as a probable mineral in rocky cores of giant planets, where the pressures can be as high as a few thousand GPa. The structure of the silicon oxide SiO2 under the conditions corresponding to the lower mantle is of interest not only due to the possible formation at large depths but also due to the possible effect on the hypothetical reaction of decomposition of the MgSiO3 perovskite into the MgO periclase and SiO2 silica. However, according to the recent experimental (Murakami et al., 2004; Oganov and Ono, 2004) and theoretical (Oganov and Price, 2005) data, the given reaction cannot proceed under the mantle conditions. It should be noted that, in Oganov et al. (2005a), the authors express serious doubt on the presence of stishovite, its orthorhombic analog with a CaCl2 structure, and other highpressure silicon oxide forms in the mantle. According to the results of the theoretical simulation of the phase diagram of the SiO2 oxide at pressures and temperatures possible at large depths and petrological investigations of ultrabasic rocks of the peridotite, harzburgite, and pyrolite compositions at high pressures, the inference has been made that the content of free SiO2 in the mantle is very insignificant. This oxide can occur in considerable amounts only in specific mantle regions close to subduction zones of basalt blocks. In the majority of the mantle, the ratio MgO/SiO2 > 1 holds true and the SiO2 oxide enters into the reaction with the MgO oxide to form the mineral MgSiO3 with a perovskite structure. Therefore, the presence of the silica is excluded in the lower mantle, unlike the upper mantle, where, as was noted above, stishovite can be formed. In recent years, data on a change in the electronic structure of iron atoms in perovskite-like phases of the (Mg0.9Fe0.1)SiO3 composition have appeared. In this case, the content of iron in the low-spin state increases from a pressure of ~70 GPa (~1850 km) and reaches a maximum at 120 GPa (~ 2700 km), i.e., in the vicinity of the D″ layer (Badro et al., 2004). This two-step transition of iron atoms from the high-spin state to the lowspin state in perovskite-like phases (unlike the ferropericlase) can be explained by the filling of both cation positions with iron atoms in the MgSiO3 phase. Since the symmetry and crystal field forces (as well as the cation–anion distances) in both positions are different, the spin transitions in cations occupying these positions are also different. It was experimentally established (Frost et al., 2004) that the Fe3 + valence state (rather than the Fe2 + valence state, as was believed earlier) of Fe impurities in the MgSiO3 phase with a perovskite structure is dominant. Furthermore, since Fe 2 + ions continuously transfer from the upper mantle and transition zone, the formation of the MgSiO3 perovskite-like phase in the lower mantle should be accompanied by the reaction 3Fe2 + (perovskite) → 2Fe3 + (perovskite) + Fe0 (metal), which was experimentally confirmed by

Frost et al. (2004). Recent calculations (Zhang and Oganov, 2006) also confirmed this reaction and showed that Fe3 + cations in the perovskite structure (as well as in the postperovskite structure in which Fe3 + ions are also dominant) tend to occupy the nearest Mg and Si positions. According to these calculations, the reaction described in Frost et al. (2004) is attended by a strong exothermic effect and can increase the temperature of the entire mantle by 100 K. As a general conclusion we can emphasize that the mineralogical diversity of the mantle is much lower as compared with the Earth's crust. For example, hundreds of structure types attributed to the crust silicates are replaced by two dozens in the transition zones and in the upper, middle and low mantles. However the new data indicate that the mantle mineralogy is not so simple as it was anticipated two or three decades ago and there can be a lot of discoveries in this field in the nearest future. 9. Mineral composition at the mantle–core boundary (D″-layer) The core–mantle boundary is treated as the most pronounced manifestation of the differences between the material properties in the Earth's interior. Analysis of the experimental data demonstrates that, beginning with this depth (beginning with temperatures of 3000 K and pressures of 25–70 GPa), the iron melt and the (Mg,Fe) SiO3 phase with a perovskite structure interact to form stishovite and an iron alloy containing silicon and oxygen. According to Badro et al. (2003), the FeSi compound, in addition to all the minerals inherent in the middle and lower mantles, can occur at the core–mantle boundary. The most unexpected recent finding revealed from quantummechanical calculations and direct experiments is the transformation of the MgSiO3 perovskite phase into the phase with a structure of the CaIrO3 type at pressures of 125–127 GPa and temperatures of 2500–3000 K corresponding to depths of 2700–2740 km, i.e., to the upper boundary of the D″ layer (Fig. 6) (Oganov and Ono, 2004). The density of this phase, which is MgSiO3 phase with a CaIrO3-type structure is the main frequently termed the postperovskite phase, is 1.2% higher than that of the perovskite phase (Fig. 10). Note that the mineral of the D″ layer covers the depth range of 2700– 2890 km. The same inferences were independently and almost simultaneously made by Murakami et al. (2004)) on the basis of the performed experiments. It should be noted that the idea of the possibility of crystallizing minerals that have the above structure and are characteristic of the core–mantle boundary was put forward on the basis of the experiments on compression of the Fe2O3 compound (Ono et al., 2004). At a pressure of 30 GPa, α-Fe2O3 (hematite) transforms into the perovskite-like modification, which with a further compression to pressures higher than 50 GPa acquires the aforementioned postperovskite structure (Oganov and Ono, 2004; Ono et al., 2005a). In Oganov et al. (2005b), it was demonstrated that the perovskite and postperovskite modifications of the MgSiO3 compound are the extreme terms of the polytypic series of structures (Fig. 11). The intermediate structures of this series are metastable at T = 0 K. However, since their enthalpies are only slightly higher than those of the perovskite and postperovskite, they can be stabilized by impurities and temperature effects. This circumstance enables us to assume that the phases with these structures can occur as secondary minerals in the lower mantle and the D″ layer. Moreover, the given structures correspond to distortions of the atomic packing in the main motifs of the perovskite and postperovskite and, according to Oganov et al. (2005b), are formed under plastic deformations. These data, in combination with the predicted elastic constants of the postperovskite, made it possible to interpret the observed unusual character and a pronounced anisotropy of the D″ layer. The Al2O3 oxide with a CaIrO3 structure stable at pressures above 130 GPa is the final product of the corundum transformation (Oganov and Ono, 2005). As was noted above, the corundum at a pressure of 90 GPa crystallizes in a structure of the Rh2O3 (II) type, in which,

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

Fig. 11. Polytypes of the MgSiO3 compound (Oganov et al., 2005b): (a) perovskite, (b) a new type of structure consisting of 2 × 2 octahedral units, (c) a new type of structure consisting of 3 × 1 octahedral units, and (d) postperovskite. Only octahedral motifs are represented. Magnesium atoms are not shown. Arrows indicate the most probable displacements under plastic deformations.

unlike many high-pressure phases, the closest packing of oxygen atoms is absent. According to the geophysical data, the deep zones of the lower mantle are characterized by a very high electrical conductivity, whose origin is not quite clear under the assumption that (Mg,Fe)SiO3, (Mg,Fe)O, and CaSiO3 are the main minerals of these zones. However, the previously established sharp increase in the electrical conductivity of the Al2O3 oxide at a pressure of 130 GPa and a temperature of 1500 K, in combination with the transformation of this phase under similar conditions into a CaIrO3-type structure, permits us to assume that a high electrical conductivity should be observed for the MgSiO3 postperovskite modification and, hence, be responsible for the electrical conductivity of the entire zone containing this phase. An important potential component of the lower mantle is the FexO wüstite, whose composition corresponds to a stoichiometric coefficient x b 0.98. This implies that the wüstite structure involves both Fe 2 + and Fe 3 + ions. According to the experimental data obtained by Boehler (1993), the melting temperature of the wüstite at the boundary between the lower mantle and the D″ layer is estimated at ~5000 K. This is considerably higher than the anticipated temperatures (~4000–4300 K) at these depths (the mean adiabatic temperature of the mantle is approximately equal to 2700 K, and the temperature at the bottom of the lower mantle can be higher by approximately 1400 K). Consequently, the wüstite can be retained in the solid form at this depth. The concept of the phase boundary between the solid lower mantle and the liquid outer core requires a more flexible approach and does not mean that there is a sharp boundary between them. 10. Mineral and chemical compositions of the core The inner core 1225 km in radius exists in a solid state and has a high density of 12.5 g/cm 3. The outer core exists in a liquid state

105

and has a density of 11 g/cm 3. The jumps in the longitudinal wave velocity and the density are observed at the core–mantle boundary. In the mantle, the density decreases to 5.5 g/cm 3. The D″ layer is in direct contact with the outer core, which has a substantial (especially, thermal) effect on the D″ layer because the temperature in the core is considerably higher than the mantle temperature. At some locations, mantle plumes, i.e., great upward flows of a hot material from the Earth's mantle into the crust, originate at the D″ layer. The inner core is surrounded by the liquid outer core 3480–3490 km in radius. The density of the outer core is approximately 10% lower than the density of molten iron. This allows one to make the inference that the composition of the outer core involves a number of light elements (Si, S, O, C, H). The boundary between the outer and inner cores is sharp. The velocity jump is as large as 0.78 km/s. Comparison of viscosity values of the Earth's outer core in its external zone (10 3–10 5 P) with one in the internal zone (10 7–10 10 P) which is contiguous with the hard inner core allows to conclude that the substances of external zone should be in viscous-flow and high-plastic state. According to F.A. Letnikov such low viscosity of this zone is mainly determined by its high saturation by fluids and does not depend on composition of the outer core. Due to the Earth's one rotation per day all its geospheres formed by hard rocks also repeat this cycle. The hard rocks of the core and of the mantle are characterized by viscosity >10 20 P and consequently they cannot be mutually displaced. Another situation is anticipated in the outer core where the density increases with the depth from 9.9 to 12.5 g/cm 3, and the viscosity changes in large interval. Thus the outer liquid core is located between the hard mantle and the hard inner core. As it was mentioned above the whole system exhibited one rotation per day. The difference in rheological properties between the outer core and the mantle on one side and hard and liquid cores on the other side lead to the deceleration of their media which occurs at their boundaries. This hypothesis was confirmed by geophysical results obtained by Avsyuk et al. (2001) who experimentally showed the difference in the rotation speed within the boundary zone between the liquid core and the hard mantle. This mismatch in the process of geosphere's rotation must lead to the release of heat connected with the friction along all these boundaries. It's possible that this phenomenon at the mantle–core frontier (D″ layer) determines the smallest value of viscosity. It is believed that the inner core consists of an iron–nickel alloy and contains a considerably smaller amount of light elements. The nickel content is estimated at 5–15 wt.%. As is known, Fe–Ni alloys with a nickel content of up to 25 at.% under room conditions have a structure with an I cell. At higher nickel contents, alloys have structures with an F cell. According to experimental data, the Fe0.8Ni0.2 alloy with an I cell at room temperature under compression transforms into a structure with the hexagonal closest packing of metal atoms. This structure is formed at a pressure of 7–14 GPa and retained up to 260 GPa. The density of the formed phase is equal to 14.35 g/cm 3 and close to the density of the ε-Fe phase (14.08 g/cm 3), which also has the hexagonal closest packing and, as was noted above, dominates in the Earth's core. The recent results of the static compressibility of Fe confirm these conclusions (Tateno et al., 2010). Tateno and co-authors revealed that the structure of Fe with hexagonal closest packing is stable at a pressure of 377 GPa and temperature of 5700 K. These values correspond to the conditions in the inner core. A small change of the c/a ratio at these parameters in the iron structure indicates its elastic anisotropy. The latter statement should be considered in connection with a conclusion that the Earth's inner core is elastically anisotropic (Belonoshko et al., 2008). The velocity of sound waves along the Earth's spin axis is higher than in the equatorial plane. This anisotropy has been also explained by a preferred orientation of the ε-Fe hexagonal crystals. However hexagonal iron becomes less anisotropic with an increase in pressure and temperature. An alternative explanation, confirmed by experiments, is that iron loses hexagonal closest packing

106

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

and transforms to body-centered cubic form in the inner core. The results obtained by molecular dynamic simulations show that this phase is extremely anisotropic to sound waves despite its high symmetry. Direct simulations of seismic wave propagation reveal its anisotropy of 12%, a value adequate to explain the anisotropy of the inner core (Belonoshko et al., 2008). In Alfè et al. (2002, 2007), the quantum-mechanical calculations were used to determine the differentiated composition of the Earth's core: 8.5 mol% Si + S and 0.2 mol% O in the inner core and 10 mol% Si + S and 8 mol% O in the outer core (the Si and S contents were not separated in these calculations). The temperature at the boundary between the inner and outer cores was estimated at 5400–5700 K. Recent seismological observations indicate that, inside the inner core, there is a seismic boundary that can be explained by the transformation of the iron structure. At present, it has been assumed that the most stable form of pure iron under conditions of the Earth's core should be a modification based on the hexagonal closest packing of Fe atoms. However, the modification with a body-centered cell turns out to be more stable for iron containing Si and S impurities at temperatures close to the melting temperature (Belonoshko et al., 2003; Vocadlo et al., 2003). In the inner core, the presence of stoichiometric phases in the Fe–S and Fe–Si systems for example, the Fe3S and Fe7S phases (Fig. 12) cannot be ruled out. The possible structures of iron silicides studied theoretically to a pressure of ~ 400 GPa revealed the exceptional stability of FeSi with CsCl structure type. The other members of this series (Fe3Si, Fe2Si, Fe5Si3, FeSi2 and FeSi3) decompose on Fe + FeSi or FeSi + Si. Thus an impurity of Si promotes the stability of the body-centered cubic cell of the iron which is an important constituent of the inner core (Zhang and Oganov, 2010). The presence of hydrogen in the core has been debated for a long time owing to the low solubility of hydrogen in iron at an atmospheric pressure. However, the experimental results (Badding et al., 1992) demonstrated that the iron hydride FeH can be formed at high temperatures and pressures and appears to be stable at pressures higher than 62 GPa, which corresponds to a depth of ~1600 km. In this respect the presence of a significant amount of H in the core seems very reasonable. The H-atoms are incorporated in the iron hydride FeH with hexagonal close packing and with 17% increase of its unit cell volume in comparison with α-Fe. The development of the scientific ideas related with mineralogical and chemical compositions of the deep-seated geospheres allowed expressing the new hypothesis about the structure of the deep shells of some other planets. It is anticipated that equivalents of the Earth's mantle in the deep parts of gigantic planets like Saturn and Jupiter could be dense gaseous (composed from compressed gasses, namely

Fig. 12. Tetragonal structure of Fe7S—a possible component of the inner (solid) core, after Sherman (1997).

hydrogen and helium) or ice (containing H2O, NH3, CH4—in case of Uranus or Neptune) shells. In the cores of these gigantic planets the pressure increases to 600 GPa and the temperature achieves several thousand degrees (K) (Gao et al., 2010). The high-pressure forms of methane, CH4, are of special interest. At high pressure Gao and coauthors recently described its transforms with the formation of three compounds. Remarkably, under high pressure, methane becomes unstable and dissociates into ethane—C2H6—at 95 GPa, butane—C4H10— at 158 GPa, and further, into carbon—diamond—and hydrogen above 287 GPa at zero temperature. These results support the idea of diamond formation in the interiors of giant planets such as Neptune. 11. The origin and evolution of the geospheres The geological history of geospheres is connected with the origin of the Earth which is estimated as 4.6 billion years ago. The first periods in this process are considered in Voitkevich and Bessonov (1986) and Condie (1994, 2004). “The chemical composition of the Earth resulted from condensation of a solar nebula and subsequent accumulation of condensates into compact masses. All these final condensates were rock-forming minerals of the Earth and meteorites” (Voitkevich and Bessonov, 1986, p. 14). The nonvolatile refractory elements were first to condense. Subsequently, the poorly volatile elements and, at the very end, the most volatile elements and their compounds were subjected to condensation. This process was primarily controlled by the temperature, while pressure played a subordinate role. According to Voitkevich and Bessonov (1986, p. 37), “the growth of the primary Earth began with coalescence of metallic particles that made up a primary metallic nucleus, which had a mass sufficient for further gravitational capture of later condensates from the environment.” In the end of accretion the mantle was separated from the primary core (with inner and outer shells) by the marginal layer D″. The formation of protocrust presumably of basic– ultrabasic composition belongs to this time (Fig. 13a). Next important period in the mantle differentiation appeared 4.5 Ga ago. It is reasonable to start its consideration with the Sm–Nd isotopic characteristics of supracrustal rocks from the Isua area

Fig. 13. Earth's mantle: the stages of its formation.

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

(western Greenland) (Harper and Jacobsen, 1992). These results indicate that the further differentiation could have begun 4.44–4.54 Ga ago. The primordial Earth's surface was enriched in LREE at this stage and formed the basis of the future Earth's crust. The study of the 142Nd/ 144Nd ratio in metamorphosed gabbro and basalts from the same area (Boyet et al., 2003) yielded similar results, which led to the conclusion that further mantle differentiation began at this epoch. It is connected with formation of the lower mantle separated from residual mantle by transition zone II (division zone II) (Fig. 13b). The D″ layer is considered as transitional because lower mantle geosphere is located over it and the core—below it. However this frontier is diffuse. It allows considering their formation as isochronic in the geological history of the Earth. This stage corresponds to the mantle evolution and spread the next 600 Ma till 3.9 Ga (Eoarchean). During the same time the protocrust with basic composition was replaced by the crust of sialitic composition. The formation of the present-day structure of the Earth belongs to the later time. The variations in the thermal regime of the Earth were one of the critical factors responsible for mantle differentiation. Fig. 14 shows the average temperature changes of the mantle from the beginning of its differentiation (1600 °C) till nowadays. The temperature fluctuations may be related, in our opinion, to the nonuniform heat transfer between the upper and lower mantles (in traditional model). It should be noted, however, that the modeling of thermal evolution throughout Earth's history is an extremely complex problem. Various approaches may lead to the different solutions. For the last 900 Ma, the value of ~ 1300 °C is assumed. In total the graph reproduced from Griffin et al. (2003) and shown on Fig. 14 demonstrates the cooling of the upper and whole mantles throughout geologic history. All these data on temperature variations and magmatic activity during the geological history provide a basis for scientific ideas related with origin of mantle geospheres. Other geospheres were formed later with correspondence to the main boundaries in the magmatism evolution. The Archean/Proterozoic boundary (2.5 billion years ago) was probably one of the most important in this respect and marked the above-mentioned drastic changes in the geological history of the Earth. It marks the cessation of intense tectonomagmatic activity and development of acid magmatism in the Earth's upper shells during the cratonic stage. The formation of the upper mantle and the transition zone I (division zone I) is connected with this epoch. The recognition of the middle mantle as a separate geosphere makes it necessary to define the time of its origin in the course of the

Fig. 14. Temperature variations in the two-layer model of the Earth, simplified after Davies (1995). Temperature fluctuations are determined by insignificant heat transfer between the upper and lower mantles (in a traditional sense). In the burning hot early Earth, the heat was transferred into the lower mantle faster than it was lost by the upper mantle, thereby giving rise to the intermittent convective flows, in which geotherms in the lower mantle cross the anhydrous peridotite solidus (dotted line), thus inducing melting at a depth of ≥150 km. This circumstance provided accumulation of the subcontinental lithospheric mantle in the Archean and, eventually, the termination of this process owing to the Earth's cooling in the post-Archean time.

107

evolution of mantle shells (Pushcharovsky and Pushcharovsky, 2007). On the basis of comprehensive evidence, above all, with allowance for mineralogical and petrological data, it was shown that mantle geospheres were formed for a long time and during different time spans. The origination of the middle mantle is determined by the formation of transition zone I (discontinuity I) which belongs to the period of the Neoarchean and Paleoproterozoic (2.5 Ga) (Fig. 13c). The problem discussed in this part is very broad. We have proposed only a preliminary approach to its solution and would be quite contented if our paper arouses interest in its elaboration. Acknowledgments The authors are grateful to Professors. M. Flower, A. Oganov and N. Koronovsky for their helpful discussion and comments. This work was supported in part by the Russian Scientific Foundation (grant 12-05-00250) and by the Foundation of the President of Russian Federation (grants NSh-2883.2012.5 and NSh-5177.2012.5). References Akaogi, M., 1997. High-pressure high-temperature transformations and thermodynamic properties of mantle minerals. Journal of the Mineralogical Society of Japan 26 (1), 15–24. Alfè, D., Gillan, M.J., Price, G.D., 2002. Composition and temperature of the Earth's core constrained by combining ab initio calculations and seismic data. Earth and Planetary Science Letters 195 (1–2), 91–98. Alfe, D., Gillan, M.J., Price, G.D., 2007. Temperature and composition of the Earth's core. Contemporary Physics 48 (2), 63–80. Anders, E., Ebihara, M., 1982. Solar system abundances of the elements. Geochem. Cosmochim. Acta 46, 2363–2380. Anderson, D.L., 1989. Theory of the Earth: Boston. Blackwell Scientific Publications. 366 pp. Avsyuk, Yu.N., Adushkin, V.V., Ovchinnikov, V.N., 2001. Multidisciplinary study of the mobility of the Earth's inner core. Izvestia, Physics of Solid Earth 37 (8), 673–683. Badding, J.V., Mao, H.K., Hemley, R.J., 1992. High-pressure crystal structure and equation of state of iron hydride: implications for the Earth's core. In: Syono, Y., Manghnan, M.H. (Eds.), High-Pressure Research: Application to Earth and Planetary Sciences. American Geophysical Union, Washington, pp. 363–371. Tokyo, Terrapub. Badro, J., Fiquet, G., Guyot, F., Rueff, J.-P., Struzhkin, V.V., Vankó, G., Monaco, G., 2003. Iron partitioning in the Earth's mantle: toward a deep lower mantle discontinuity. Science 300 (5620), 789–791. Badro, J., Rueff, J.P., Vanko, G., Monaco, G., Fiquet, G., Guyot, F., 2004. Electronic transitions in perovskite: possible nonconvecting layers in the lower mantle. Science 305 (5682), 383–386. Belonoshko, A.B., Ahuja, R., Johansson, B., 2003. Stability of the body-centered-cubic phase of iron in the Earth's inner core. Nature 424 (6952), 1032–1034. Belonoshko, A.B., Skorodumova, N.V., Rosengren, A., Johansson, B., 2008. Elastic anisotropy of Earth's inner core. Science 319 (5864), 797–800. Bernal, J.D., 1936. Discussion. Observatory 59, 268. Birch, F., 1939. The variation of seismic velocities with a simplified earth model in accordance with the theory of finite strain. Bulletin of the Seismological Society of America 29 (3), 463–479. Birch, F., 1952. Elasticity and constitution of the Earth's interior. Journal of Geophysical Research 57 (2), 227–286. Boehler, R., 1993. Temperatures in the Earth's core from melting-point measurements of iron at high static pressures. Nature 363 (6429), 534–536. doi:10.1038/363534a0. Boyet, M., Blichert-Toft, J., Rosing, M., Storey, M., Télouk, Ph., Albarède, Francis, 2003. 142 Nd evidence for early Earth differentiation. Earth and Planetary Science Letters 214 (3–4), 427–442. Bullen, C.E., 1947. Introduction to the Theory of Seismology. Cambridge University Press, Cambridge. 276 pp. Bullen, K.E., 1975. The Earth's Density. J. Wiley & Sons, New York. Condie, K.C., 1994. Introduction. In: Condie, K.C. (Ed.), Archean Crustal Evolution. Elsevier, Amsterdam, pp. 1–10. Condie, K.C., 2004. Earth as an Evolving Planetary System. Elsevier Press. 350 pp. Davies, G.F., 1995. Punctuated tectonic evolution on the Earth. Earth and Planetary Science Letters 136 (3–4), 363–379. Dorogokupets, P.I., 2007. Equation of state of magnesite for the conditions of the Earth's lower mantle. Geochemistry International 45 (6), 561–568. Dorogokupets, P.I., Oganov, A.R., 2007. Ruby, metals, and MgO as alternative pressure scales: a semiempirical description of shock-wave, ultrasonic, X-ray, and thermochemical data at high temperatures and pressures. Physics Reviews B75 art. 024115. Dubrovinsky, L.S., Saxena, S.K., Lazor, P., Ahuja, R., Eriksson, O., Wills, J.M., Johansson, B., 1997. Experimental and theoretical identification of a new high-pressure phase of silica. Nature 388 (6640), 362–365. Dubrovinsky, L.S., Annersten, H., Dubrovinskaia, N.A., Westman, F., Harryson, H., Fabrichnaya, O., Carison, S., 2001. Chemical interaction of Fe and Al2O3 as a source of heterogeneity at the Earth's core–mantle boundary. Nature 412 (6846), 527–529.

108

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109

Dziewonski, A.M., Anderson, D.L., 1981. Preliminary reference Earth model. Physics of the Earth and Planetary Interiors 25 (4), 297–356. Dziewonski, A.M., Woodhouse, J.H., 1987. Global images of the Earth's interior. Science 236 (4797), 37–48. Dziewonski, A.M., Woodhouse, J.H., 1988. Three-dimensional Earth structure and mantle convection. XXVIII International Geological Congress: Abstracts. Washington (D. C.), 1, pp. 427–428. Dziewonski, A.M., Su, W.J., Woodward, R.L., 1992. A change in the spectrum of lateral heterogeneity at about 1700 km depth. EOS, Transactions of the American Geophysical Union 73 (45), 577 Supplement. Egorkin, A.V., 2004. The mantle structure of the Siberian platform. Izvestia, Physics of Solid Earth 40 (5), 385–394. Fiquet, G., 2001. Mineral phases of the Earth's mantle. Zeitschrift für Kristallographie 216 (5), 248–271. Fiquet, G., Guyot, F., Kunz, M., Matas, J., Andrault, D., Hanfland, M., 2002. Structural refinements of magnesite at very high pressure. American Mineralogist 87 (8–9), 1261–1265. Frost, D.J., Lebske, C., Langenhorst, F., McCammon, C.A., Trønnes, R.G., Rubie, D.C., 2004. Experimental evidence for the existence of iron-rich metal in the Earth's lower mantle. Nature 428 (6981), 409–412. Funamori, N., Jeanloz, R., 1997. High-pressure transformation of Al2O3. Science 278 (5340), 1109–1111. Funamori, N., Jeanloz, R., Nguyen, J.H., Kavner, A., Caldwell, W.A., Fujino, K., Miyajima, N., Shinmei, T., Tomioka, N., 1998. High-pressure transformations in MgAl2O4. Journal of Geophysical Research 103 (9), 20813–20818. Gao, G., Oganov, A.R., Ma, Y., Wang, H., Li, P., Li, Y., Iitaka, T., Zou, G., 2010. Dissociation of methane under high pressure. Journal of Chemical Physics 133, 144508-1–144508-5. Gautron, L., Madon, M., 1994. A study of the stability of anorthite in the PT-conditions of Earth's transition zone. Earth and Planetary Science Letters 125 (1–4), 281–291. Goldschmidt, V.M., 1931. Zur Kristallchemie des Germaniums. Nachrichten der Akademie der Wissenschaften in G"ottingen. II. Mathematisch-Physikalische Klasse 2, 184–190. Griffin, W.L., O' Reilly, S.Y., Abe, N., Aulbach, S., Davies, R.M., Pearson, N.J., Doyle, K., Kivi, B.J., 2003. The origin and evolution of Archean lithospheric mantle. Precambrian Research 127 (1–3), 19–41. Harper Jr., C.L., Jacobsen, S.D., 1992. Evidence from coupled 147Sm–143Nd and 146Sm–142Nd systematics for very early (4.5-Gyr) differentiation of the Earth's mantle. Nature 360 (6406), 728–732. Holzapfel, W.B., 2001. Equations of state for solids under strong compression. Zeitschrift für Kristallographie 216 (9), 473–488. Irifune, I., Ringwood, A.E., 1987. Phase transformations in primitive MORB and pyrolite compositions to 25 GPa and some geophysical implications. In: Manphani, M.H., Syono, Y. (Eds.), High-pressure Research in Geophysics. American Geophysical Union, Washington DC, pp. 231–242. Irifune, T., Fujino, K., Ohtani, E., 1991. A new high-pressure form of MgAl2O4. Nature 349 (6308), 409–411. Isshiki, M., Irifune, T., Hirose, K., Ono, S., Ohishi, Y., Watanuki, T., Nishibori, E., Takata, M., Sakata, M., 2004. Stability of magnesite and its high-pressure form in the lowermost mantle. Nature 427 (6969), 60–63. Jacobsen, S.D., Demouchy, S., Frost, D.J., Ballaran, T.B., Kung, J., 2005. A systematic study of OH in hydrous wadsleyite from polarized FTIR spectroscopy and single-crystal X-ray diffraction: oxygen sites for hydrogen storage in Earth's interior. American Mineralogist 90 (1), 61–70. Jeffreys H. 1937. On the materials and density of the Earth's crust. Monthly Notices Royal Astronomical Society. Geophys. Suppl. V. 4. P. 50–61. Kaneshima, S., Helffrich, G., 1999. Dipping low-velocity layer in the mid-lower mantle: evidence for geochemical heterogeneity. Science 283, 1888–1891. Kingma, K.J., Cohen, R.E., Hemley, R.J., Mao, H.-K., 1995. Transformation of stishovite to a denser phase at lower-mantle pressure. Nature 374 (6519), 243–245. Knittle, E., Jeanloz, R., 1986. High-pressure metallization of FeO and implications for the Earth's core. Geophysical Research Letters 13 (13), 1541–1544. Kuwayama, Y., Hirose, K., Sata, N., Ohishi, Y., 2005. The pyrite-type high-pressure form of silica. Science 309 (5736), 923–925. Liu, L.-G., 1982. Chemical inhomogeneity of the mantle: geochemical considerations. Geophysical Research Letters 9 (2), 124–126. Mao, H.K., Hemley, R.J., 1998. New windows on the Earth's deep interior. In: Hemley, R.J. (Ed.), Ultrahigh-pressure Mineralogy: Physics and Chemistry of the Earth's Deep Interior. Reviews in Mineralogy and Geochemistry, vol. 37. Mineralogical Society of America, Washington, D.C., pp. 1–32. Mao, H.K., Hemley, R.J., Fei, Y., Shu, J.F., Chen, L.C., Jephcoat, A., Wu, Y., Bassett, W., 1991. Effect of pressure, temperature, and composition on lattice parameters and density of (Fe,Mg)SiO3–Perovskites to 30 GPa. Journal of Geophysical Research 96 (B5), 8069–8079. Mazin, I.I., Fey, Y., Downs, R., Cohen, R., 1998. Possible polytypism in FeO at high pressure. American Mineralogist 83 (5–6), 451–457. McMahon, M.I., Nelmes, R.J., Schwarz, U., Syassen, K., 2006. Composite incommensurate K-III and a commensurate form: study of a high-pressure phase of potassium. Physical Review B74 (14) 140102(R), 4 pp. Mendelssohn, M.J., Price, G.D., 1997. Computer modelling of a pressure induced phase change in clinoenstatite pyroxenes. Physics and Chemistry of Minerals 25 (1), 55–62. Murakami, M., Hirose, K., Kawamura, K., Sata, N., Ohishi, Y., 2004. Post-perovskite phase transition in MgSiO3. Science 304 (5672), 855–858. Navrotsky, A., 1994. Physics and Chemistry of Earth Materials. Cambridge University Press. 413 pp. Niu, F., Kawakatsu, H., 1997. Depth variation of the mid-mantle seismic discontinuity. Geophysical Research Letters 24 (4), 429–432.

Oganov, A.R., Brodholt, J.P., 2000. High-pressure phases in the Al2SiO5 system and the problem of aluminous phase in the Earth's lower mantle: ab initio calculations. Physics and Chemistry of Minerals 27, 430–439. Oganov, A.R., Ono, S., 2004. Theoretical and experimental evidence for a post-perovskite phase of MgSiO3 in Earth's D″ layer. Nature 430 (6998), 445–448. Oganov, A.R., Ono, S., 2005. The high-pressure phase of alumina and implications for Earth's D″ layer. Proceedings of the National Academy of Sciences USA 102 (31), 10828–10831. Oganov, A.R., Price, G.D., 2005. Ab initio thermodynamics of MgSiO3 perovskite at high pressures and temperatures. Journal of Chemical Physics 122 art. 124501-6. Oganov, A.R., Brodholt, J.P., Price, G.D., 2001. The elastic constants of MgSiO3 perovskite at pressures and temperatures of the Earth's mantle. Nature 411 (6840), 934–937. Oganov, A.R., Brodholt, J.P., Price, G.D., 2002. Ab initio theory of thermoelasticity and phase transitions in minerals. In: Gramaccioli, C.M. (Ed.), Energy Modelling in Minerals, EMU Notes in Mineralogy, vol. 4, pp. 83–170. Oganov, A.R., Gillan, M.J., Price, G.D., 2005a. Structural stability of silica at high pressures and temperatures. Physical Review 71 (6) 064104(8). Oganov, A.R., Martoňák, R., Laio, A., Raiteri, P., Parrinello, M., 2005b. Anisotropy of Earth's D″ layer and stacking faults in the MgSiO3 post-perovskite phase. Nature 438 (7071), 1142–1144. Oganov, A.R., Price, G.D., Scandolo, S., 2005c. Ab initio theory of planetary materials. Zeitschrift für Kristallographie 220 (5–6), 531–548. Oganov, A.R., Glass, C.W., Ono, S., 2006. High-pressure phases of CaCO3: crystal structure prediction and experiment. Earth and Planetary Science Letters 241 (1–2), 95–103. Oganov, A.R., Ono, S., Ma, Y., Glass, C.W., Garca, A., 2008. Novel high-pressure structures of MgCO3, CaCO3 and CO2 and their role in Earth's lower mantle. Earth and Planetary Science Letters 273 (1–2), 38–47. Ono, S., Oganov, A.R., 2005. In situ observations of phase transition between perovskite and CaIrO3-type phase in MgSiO3 and pyrolitic mantle composition. Earth and Planetary Science Letters 236 (3–4), 914–932. Ono, S., Kikegawa, T., Ohishi, Y., 2004. High-pressure phase transition of hematite, Fe2O3. Journal of Physics and Chemistry of Solids 65 (8–9), 1527–1530. Ono, S., Funakoshi, K., Ohishi, Y., Takahashi, E., 2005a. In situ X-ray observation of the phase transformation of Fe2O3. Journal of Physics: Condensed Matter 17, 269–276. Ono, S., Kikegawa, T., Ohishi, Y., Tsuchiya, J., 2005b. Post-aragonite phase transformation in CaCO3 at 40 GPa. American Mineralogist 90 (4), 667–671. Ono, S., Oganov, A.R., Brodholt, J.P., Vocadlo, L., Wood, I.G., Glass, C.W., Cote', A.S., Price, G.D., 2008. High-pressure phase transformations of FeS: novel phases at conditions of planetary cores. Earth and Planetary Science Letters 272 (1–2), 481–487. Pavlenkova, N.I., 1995. Structural regularities in the lithosphere of continents and plate tectonic. Tectonophysics 243 (3–4), 223–239. Pavlenkova, N.I., 1998. Seismic models of the Earth's crust and upper mantle and their geological interpretation. In: Karyakin, Yu.V. (Ed.), Tectonics and Geodynamics: General and Regional Aspects, vol. 2. GEOS, Moscow, pp. 72–75. Pavlenkova, G.A., 2003. Velocity heterogeneity of the upper mantle of Eastern Europe. Geophysics of the XXI Century: 2002; Nauchnyi Mir, Moscow, pp. 45–49. [in Russian]. Prewitt, Ch.T., 2003. Mineral physics: looking ahead. Journal of Mineralogy and Petrological Sciences 98 (1), 1–8. Pushcharovsky, D., 1986. Structural Mineralogy of Silicates and their Synthetic Analogues. Nedra, Moscow. 169 pp. [in Russian]. Pushcharovsky, Yu.M., 1995. On three paradigms in geology. Geotektonics 29 (1), 2–8. Pushcharovsky, Yu.M., 1996. Seismic tomography and the mantle structure: tectonic aspects. Transactions (Doklady) of the Russian Academy of Sciences, Earth Science Sections 351A (9), 1424–1426. Pushcharovsky, Yu.M., 1997. Major tectonic asymmetry of the Earth: the Pacific and Indo-Atlantic segments and their relationships. In: Yu, Leonov (Ed.), Tectonic and Geodynamic Phenomena: Geological Institute, Russian Academy of Sciences, Transactions, vol. 505, pp. 8–24. Moscow, Nauka. Pushcharovsky, Yu.M., 1998. Seismic tomography, tectonics, and deep geodynamics. Transactions (Doklady) of the Russian Academy of Sciences/Earth Science Sections 360 (4), 514–517. Pushcharovsky, D., 2002. Minerals of deep geospheres. Physics-Uspekhi 45 (4), 480–485. Pushcharovsky, D.Yu., 2004. Minerals of the deep geospheres. Izvestia, Earth’s Sci. Section, Rus. Acad. of Nat. Sci. N11, 69–76 (special issue). Pushcharovsky, D., Oganov, A.R., 2006. Structural transformations of minerals into deep geospheres: a review. Crystallography Reports 51 (5), 767–777. Pushcharovsky, Yu., Pushcharovsky, D., 1999. Geospheres of the Earth's mantle. Geotectonics 33 (1), 1–11. Pushcharovsky, Yu., Pushcharovsky, D., 2007. An approach to the geologic history of the Earth's mantle geospheres. Geotectonics 41 (1), 4–12. Ringwood, A.E., 1991. Phase transformations and their bearing on the constitution and dynamics of the mantle. Geochem. Cosmochim. Acta 55, 2083–2110. Ringwood, A.E., Major, A., 1970. The system Mg2SiO4–Fe2SiO4 at high pressures and temperatures. Physics of the Earth and Planetary Interiors, v. 3, 89–108. Ross, A., 1997. The Earth's mantle remodeled. Nature 385 (6616), 490–491. Ross, N.L., 1997. The equation of state and high-pressure behavior of magnesite. American Mineralogist 82 (7–8), 682–688. Rozenberg, G.Kh., Dubrovinsky, L.S., Pasternak, M.P., Naaman, O., Le Bihan, T., Ahuja, R., 2002. High-pressure crystal structural studies of hematite Fe2O3. Physical Review B 65, 064112-1–064112-8. Saltzer, R.L., Stutzmann, E., Hilst, R.D., 2004. Poisson's ratio in the lower mantle beneath Alaska: evidence for compositional heterogeneity. Journal of Geophysical Research. Solid Earth 109, B06301 doi: 1029/2003 JB002712. Shen, Y., Wolfe, C.J., Solomon, S.C., 2003. Seismological evidence for a mid-mantle discontinuity beneath Hawaii and Iceland. Earth and Planetary Science Letters 214 (1–2), 143–151.

D.Y. Pushcharovsky, Y.M. Pushcharovsky / Earth-Science Reviews 113 (2012) 94–109 Sherman, D.M., 1997. The composition of the Earth's core: constraints on S and Si vs. temperature. Earth and Planetary Science Letters 153 (3–4), 149–155. Smyth, J.R., 1994. A crystallographic model for hydrous wadsleyite: an ocean in the Earth's interior? American Mineralogist 79 (9–10), 1021–1024. Sobolev, V.S., Dobretsov, N.L., Sobolev, N.V. (Eds.), 1975. Deep-seated Xenoliths and the Upper Mantle. Nauka, Novosibirsk. 269 pp. (in Russian). Sobolev, N.V., Logvinova, A.M., Zedgenizov, D.A., Pokhilenko, N.P., Malygina, E.V., Kuzmin, D.V., Sobolev, A.V., 2009. Petrogenetic significance of minor elements in olivines from diamonds and peridotite xenoliths from kimberlites of Yakutia. Lithos 112 (Supplement 2), 701–713. Sorokhtin, O.G., Ushakov, S.A., 2002. The Earth's Evolution. Moscow State University, Moscow. 500 pp. [in Russian]. Stishov, S.M., Popova, S.V., 1961. A new dense modification of silicon oxide No.10 Geokhimiya 837–839 [in Russian]. Su, W., Woodward, R.L., Dziewonski, A.M., 1994. Degree 12 model of shear velocity heterogeneity in the mantle. Journal of Geophysical Research 99 (B4), 6945–6980. Tateno, Sh., Hirose, K., Ohishi, Y., Tatsumi, Y., 2010. The structure of iron in Earth's inner core. Science 330 (6002), 359–361. Urusov, V.S., Pushcharovsky, D., 1984. Crystal-chemical principles of high-pressure. Mineralogical Journal, Naukova Dumka, Kiev 6 (3), 23–36. Vernadsky, V.I., 2001. Chemistry of the Earth's Biosphere and Environment. Nauka, Moscow. 371 pp. [in Russian].

109

Vocadlo, L., Dobson, D., 1999. The Earth's deep interior: advances in theory and experiment. Philosophical Transactions of Royal Society, London A 357 (1763), 3335–3357. Vocadlo, L., Alfe, D., Gillan, M.J., Wood, I.G., Brodholt, J.P., Price, G.D., 2003. Possible thermal and chemical stabilization of body-centred-cubic iron in the Earth's core. Nature 424 (6948), 536–539. Voitkevich, G.V., Bessonov, O.A., 1986. Chemical Evolution of the Earth. Nedra, Moscow. 212pp. [in Russian]. Williams, Q., Revenaugh, J., 2005. Ancient subduction, mantle eclogite, and the 300 km seismic discontinuity. Geology 33 (1), 1–4. Wirth, R., 2010. Focused ion beam (FIB): site-specific sample preparation, nanoanalysis, nano-characterization and nano-machining. In: Brenker, F.E., Jordan, G. (Eds.), Nanoscopic Approaches in Earth and Planetary Sciences: EMU Notes in Mineralogy, vol. 8, pp. 1–21. ch.1. Zhang, F., Oganov, A.R., 2006. Valence state and spin transitions of iron in Earth's mantle silicates. Earth and Planetary Science Letters 249 (3), 436–443. Zhang, F., Oganov, A.R., 2010. Iron silicides at pressures of the Earth's inner core. Geophysical Research Letters 37, L02305. doi:10.1029/2009GL041224 (4 pp). Zharkov, V.N., Kalinin, V.A., 1971. Equations of State of Solids at High Pressures and Temperatures. Consultants Bureau, New York. 257 pp.