Earth science

Lower mantle may be rich in silica


There have been conflicting claims about the composition of Earth's lower mantle. The latest Brillouin-spectroscopy data suggest that this section of the planet's interior may contain more silica than the upper mantle. See Letter p.90

In this issue, Murakami et al.1 describe pioneering opto-acoustic measurements of transverse, or shear, elastic waves on polycrystalline specimens of the dominant lower-mantle minerals silicate perovskite, (Mg,Fe)SiO3, and ferropericlase, (Mg,Fe)O. Crucially, the researchers performed the measurements under high-pressure and high-temperature conditions comparable to those of the deepest parts of Earth's mantle.

Murakami and colleagues (page 90) find1 that the shear-wave speeds for both minerals vary more mildly with pressure than had been suggested from extrapolations of data obtained at lower pressures. Consequently, the variation of wave speed with pressure for the perovskite mineral, which displays the higher wave speed, is close to that determined seismologically2 for the lower mantle, thereby excluding the possibility of a substantial volume fraction of ferropericlase, which has a much lower wave speed (see Fig. 4b of the paper1). Accordingly, the authors conclude that the lower mantle is perovskitic in composition — markedly more silica-rich than the upper mantle. A silica-enriched lower mantle would reconcile the major-element chemical composition of the bulk-silicate Earth with that of carbonaceous chondrite meteorites, which are commonly thought to approximate the composition of the solar nebula from which Earth accreted.

The elegant opto-acoustic technique of Brillouin spectroscopy has, over the past three decades, revolutionized our knowledge of mineral elasticity. In this method, the inelastic scattering of monochromatic light in a crystal, by lattice waves of thermal origin, results in a frequency shift of the scattered radiation that is proportional to the speed of the lattice wave. Systematic variation of the crystal orientation relative to that of the incident light allows determination of the full set of constants describing the elasticity of a crystal of arbitrarily low structural symmetry. Introduction of this technique to the Earth sciences allowed the single-crystal elasticity of high-pressure silicate minerals to be determined for the first time — from Brillouin spectra measured on 100-micrometre-sized crystals3.

Brillouin spectroscopy is also ideally suited to measurements of microcrystals confined at very high pressures within a diamond-anvil high-pressure apparatus4,5. Murakami and colleagues' application of the technique to polycrystalline samples of minerals with strongly direction-dependent (anisotropic) elasticity, under conditions of extreme pressure and high temperature, breaks new ground.

Analyses of the elasticity, chemical composition and temperature of the lower mantle date from the classic work of Birch6, who boldly suggested that “dense high-pressure modifications of the ferro-magnesian silicates, probably close-packed oxides ... are required to explain the high elasticity of the deeper part of the mantle”. This hypothesis was confirmed by subsequent demonstration of the transformation of upper-mantle silicates first to spinel, garnet and related crystal structures7 at pressures corresponding to mantle depths of 250–500 kilometres, and, ultimately — at depths greater than 660 kilometres — to a mixture of silicate perovskites based on SiO6 octahedral building blocks, and coexisting ferropericlase8. However, the question of whether or not there is also a change in chemical composition between the upper and lower mantle has not, until now, been answered definitively.

Increasingly detailed knowledge of the pressure–temperature conditions required for the stability of the relevant minerals, and of their densities and elastic properties, has provided the basis for numerous analyses of the elasticity, composition and temperature of the lower mantle. The result has been a series of conflicting claims over whether a contrast in chemical composition between the upper and lower mantle is required to match the seismological models. Diverse inferences9,10,11,12 have been drawn concerning the need for lower-mantle enrichment or depletion in silica, iron oxide and other components, and/or for substantial departures from the adiabatic temperature–depth gradient — expected to develop in a fluid subject to thermal convection without exchange of heat between ascending and descending parcels. Such contrasting conclusions are attributable to differences in methodology, trade-offs between chemical composition and temperature, and to residual uncertainties in the thermo-physical database for the high-pressure minerals.

It is gradually becoming accepted that an internally consistent framework13 should be used for the evaluation and assimilation of thermoelastic data from laboratory experiments and computer models of mineral behaviour, and for extrapolation to higher pressures and temperatures. Application of such a framework to diverse experimental data for periclase (MgO) highlighted and resolved14 minor tensions between different data sets. The result was a robust compromise model of the material's thermoelastic behaviour, including a well-constrained value for the pressure derivative — at zero pressure — of the shear modulus, or rigidity, from which the shear-wave speed is calculated.

A markedly lower value (by about 20%) of the zero-pressure pressure derivative was deduced by Murakami and colleagues from ultra-high-pressure Brillouin spectroscopy on polycrystalline MgO (ref. 15). Possible explanations for this discrepancy include a systematic error in pressure calibration in the ultra-high-pressure diamond-anvil experiments and uncertainty concerning the average wave speed that is determined by Brillouin scattering of incident light by the individual crystallites within a polycrystalline specimen.

Given the systematically low values of the pressure derivatives of shear modulus obtained from all of the recent ultra-high-pressure Brillouin spectroscopic measurements, and the consequences for inferred lower-mantle composition, it is vital that the technical issues surrounding pressure calibration and Brillouin scattering from polycrystalline material — ideally of both shear and compressional waves — be resolved. More precise ab initio calculations of elastic properties and laboratory measurements independent of an empirical pressure scale16 may help to explain and to eliminate the discrepancy.

Notwithstanding the impressive experiments of Murakami et al., we are probably still awaiting the final word on the chemical composition and thermal regime of Earth's lower mantle.


  1. 1

    Murakami, M., Ohishi, Y., Hirao, N. & Hirose, K. Nature 485, 90–94 (2012).

  2. 2

    Dziewonski, A. M. & Anderson, D. L. Phys. Earth Planet. Inter. 25, 297–356 (1981).

  3. 3

    Sawamoto, H., Weidner, D. J., Sasaki, S. & Kumazawa, M. Science 224, 749–751 (1984).

  4. 4

    Duffy, T. S., Zha, C. S., Downs, R. T., Mao, H. K. & Hemley, R. J. Nature 378, 170–173 (1995).

  5. 5

    Sinogeikin, S. V & Bass, J. D. Phys. Earth Planet. Inter. 120, 43–62 (2000).

  6. 6

    Birch, F. J. Geophys. Res. 57, 227–286 (1952).

  7. 7

    Ringwood, A. E. & Major, A. Phys. Earth Planet. Inter. 3, 89–108 (1970).

  8. 8

    Liu, L. Earth Planet. Sci. Lett. 31, 200–208 (1976).

  9. 9

    Li, B. & Zhang, J. Phys. Earth Planet. Inter. 151, 143–154 (2005).

  10. 10

    Khan, A., Connolly, J. A. D. & Taylor, S. R. J. Geophys. Res. 113, B09308 (2008).

  11. 11

    Matas, J., Bass, J., Ricard, Y., Mattern, E. & Bukowinski, M. S. T. Geophys. J. Int. 170, 764–780 (2007).

  12. 12

    Cobden, L. et al. J. Geophys. Res. 114, B11309 (2009).

  13. 13

    Stixrude, L. & Lithgow-Bertelloni, C. Geophys. J. Int. 162, 610–632 (2005).

  14. 14

    Kennett, B. L. N. & Jackson, I. Phys. Earth Planet. Inter. 176, 98–108 (2009).

  15. 15

    Murakami, M., Ohishi, Y., Hirao, N. & Hirose, K. Earth Planet. Sci. Lett. 277, 123–129 (2009).

  16. 16

    Li, B., Woody, K. & Kung, J. J. Geophys. Res. 111, B11206 (2006).

Download references

Author information

Correspondence to Ian Jackson.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Jackson, I. Lower mantle may be rich in silica. Nature 485, 51–52 (2012).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.