ANRV341-EA36-18 ARI 26 March 2008 1:58

The PostperovskiteTransitionSang-Heon ShimDepartment of Earth, Atmospheric, and Planetary Sciences, Massachusetts Instituteof Technology, Cambridge, Massachusetts 02139; email: sangshim@mit.edu

Annu. Rev. Earth Planet. Sci. 2008. 36:569–99

First published online as a Review in Advance onJanuary 30, 2008

The Annual Review of Earth and Planetary Sciences isonline at earth.annualreviews.org

This article’s doi:10.1146/annurev.earth.36.031207.124309

Copyright c© 2008 by Annual Reviews.All rights reserved

0084-6597/08/0530-0569$20.00

Key Words

mantle, core-mantle boundary, mantle dynamics, mantle structure,D′′ region

AbstractThe discovery of the postperovskite (PPv) transition has profoundimpact on our understanding of the core-mantle boundary (CMB)region. Unlike perovskite (Pv), the PPv phase has a layered structureof the SiO6 octahedra, which may lead to a large contrast in someproperties with Pv. Recent studies have proposed unusual propertiesof PPv, such as a large positive Clapeyron slope, a large sensitiv-ity of the transition depth to iron, a decrease in bulk sound speedat the transition, and a development of significant lattice preferredorientation. Many of the proposed properties can provide expla-nations for the intriguing seismic observations at the CMB region.Yet significant discrepancies still exist. However, rapid developmentsin mineral physics will continue to improve our knowledge on thechanges across the PPv transition.

569

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Perovskite (Pv): indicatesmantle phases with theperovskite-type structure.When chemicalcomposition is specified, itrefers to the perovskite-typestructure. Note that mineralperovskite has chemicalcomposition of CaTiO

1. INTRODUCTION

Earth’s core-mantle boundary (CMB) has often been compared with the interfacebetween the atmosphere and lithosphere ( Jeanloz & Williams 1998) because of thelarge contrast in the chemical compositions and physical properties between themantle (silicates and oxides) and the core (iron-nickel alloy). Some evidence existsthat the core and mantle materials may interact with each other physically and/orchemically (e.g., Knittle & Jeanloz 1991, Buffett et al. 2000, Takafuji et al. 2005). Inaddition, because of the heat flux out of the core, a thermal boundary layer exists atthe mantle side of the CMB. Therefore, the CMB has been regarded as a dynamicallyimportant region: For example, some studies proposed the CMB as the source regionof hot spots (Morgan 1971, Williams et al. 1998) and the graveyard of subductingslabs (Van der Voo et al. 1999).

Indeed seismologists have observed many intriguing structures at the bottom 200–400 km of the lower mantle, the so-called D′′ layer (for reviews, see Gurnis et al. 1998,Lay et al. 1998, Garnero 2000, Lay et al. 2004). For example, the existence of verythin layers (5–40 km) with velocity reductions of more than 10% have been reportedat the base of the mantle (ultralow velocity zones) (Garnero & Helmberger 1995,Revenaugh & Meyer 1997). Shear waves travel with different speeds, dependingon the polarization at some regions of the lowermost mantle (shear wave splitting)(Kendall & Silver 1996, Matzel et al. 1996, Garnero & Lay 1997, Fouch et al. 2001,Panning & Romanowicz 2004). Velocity discontinuities have been documented at150–250 km above the CMB (D′′ discontinuity) (Lay & Helmberger 1983, Wysessionet al. 1998, Sidorin et al. 1999, Lay et al. 2004, Helmberger & Ni 2005). The locationsof the bulk sound speed and shear wave velocity anomalies are anticorrelated at themid to lowermost mantle (Su & Dziewonski 1997, Masters et al. 2000, Trampert et al.2004). Yet the origin of these features has not been well understood until recently.

In 2004, three groups of experimentalists independently reported that MgSiO3-Pv, the dominant lower-mantle phase, transforms into a new phase, named post-perovskite (PPv), at pressure-temperature conditions similar to those at the depthof the D′′ discontinuity (Murakami et al. 2004, Oganov & Ono 2004, Shim et al.2004). Quantum mechanical calculations immediately confirmed this phase transi-tion (Iitaka et al. 2004, Tsuchiya et al. 2004b). Since then, many exciting predictionsand measurements have been made about this phase. One of the most importantquestions has been whether the PPv phase is indeed responsible for the intriguingseismic observations and whether its properties can strongly influence the dynamicsat the lowermost mantle (Wookey et al. 2005).

This article is not intended to give a thorough review on PPv as some review papersalready exist (e.g., Shim 2005, Hirose 2006). I intend to provide an introduction tothis rapidly evolving field and outstanding problems for future studies.

2. EXPERIMENTAL TECHNIQUES

Extreme pressure-temperature conditions for the stability of the PPv phase (withpressure, P ≥ 100 GPa and temperature, T ≥ 2000 K) create many challenges

570 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Infraredlaser beam

Synchrotronradiation

MonochromaticX-ray beam

X-raydiffraction

Diamondanvils

2D detectorlaser couplerThermalradiation

Gasket

Pressure medium(thermal insulation)

Sample +

Gas

per coupler

e medium

Figure 1A schematic diagram of the experimental setup for in situ X-ray diffraction measurements inthe double-sided laser-heated diamond cell at third-generation synchrotron facilities.

in experimental mineral physics. The only technique capable of maintaining thepressure-temperature conditions for a sufficiently long time for various measure-ments is the laser-heated diamond-anvil cell (DAC) (Figure 1). For measurementsover 100 GPa, the size of the diamond culet is typically less than 150 μm, which limitsthe size of the sample to less than 70 μm.

Laser heating has been an important technique for reaching mantle temperatures(>2000 K) in the DAC ( Jeanloz & Heinz 1984, Boehler 2000) (Figure 1). However,because of the small laser-beam cross section at the sample (typically <20–30 μm),large temperature gradients exist along both the radial and axial directions of theDAC during heating to high temperature. This is particularly severe for measure-ments without thermal insulation for the samples, as diamond has very high thermalconductivity. The large thermal gradients can introduce systematic uncertainties inthe determination of phase boundaries and equations of state (Shim & Duffy 2000).

In typical X-ray measurements, the samples are mixed with chemically inert metals,such as platinum or gold. These metals have known pressure-volume relations fromwhich pressure can be calculated using measured volume. These metals also couplewith infrared laser beams, such as Nd:YLF and Nd:YAG lasers, at high pressure andheat surrounding silicate and oxide samples.

In situ measurements for the tiny samples in the DAC over 100 GPa demand anextremely bright X-ray source with a tight focus (the typical volume of the sample is∼5 × 10−9 cm3 and the mass is ∼10−8 g). It is not surprising that the PPv transition

www.annualreviews.org • The Postperovskite Transition 571

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

was discovered after the construction of high-pressure beamlines at third-generationsynchrotron facilities, which provide the brightest X-ray beam at the moment. Theobservations by Murakami et al. (2004) and Oganov & Ono (2004) were made atSPring-8 in Japan (Watanuki et al. 2001), and Shim et al. (2004) observed the transi-tion at Advanced Photon Source in the United States (Shen et al. 2001). In additionto diffraction techniques, many X-ray spectroscopy techniques have recently becomeavailable for measurements in the DAC, and they have made significant contribu-tions to the understanding of the properties of PPv. Researchers have begun to applysome laser spectroscopy techniques, such as Brillouin and Raman spectroscopies,for measurements of PPv (Murakami et al. 2007, Shim et al. 2007). These tech-niques provide useful constraints on the elastic and thermodynamic properties of PPv,respectively.

3. CRYSTAL STRUCTURE

The crystal structure of PPv has not been determined experimentally owing to the dif-ficulty in synthesizing appropriate single crystals at its extreme pressure-temperaturestability field. According to some X-ray diffraction studies (Murakami et al. 2004,Shieh et al. 2006), this phase may not be quenchable to ambient pressure.

First-principles calculations are in good agreement with each other on the crystalstructure of PPv (Iitaka et al. 2004, Oganov & Ono 2004, Tsuchiya et al. 2004b). Asshown in Figure 2a,b, the calculated diffraction pattern from the proposed crystalstructure is in excellent agreement with the observed diffraction patterns of PPv.Least-squares fitting of the observed diffraction patterns, such as Rietveld refinement,has been attempted (Ono et al. 2006a). It is important to note that through the Rietveldmethod, crystal structure can be refined for a given model. However, it is difficult todetermine crystal structures from the diffraction patterns of powder samples (Young1993, Harris et al. 2001).

Study of analogy materials with the stability field of the PPv phase provides oppor-tunities to examine the proposed crystal structure. Mn2O3 transforms to a PPv phaseat 30–40 GPa and room temperature (Figure 2e) (Santillan et al. 2006). The room-temperature transition allows for high-quality diffraction intensity measurements byavoiding strong recrystallization during laser heating. NaMgF3-Pv undergoes a phasetransition to PPv (Figure 2d ) at 13–30 GPa (Liu et al. 2005; Martin et al. 2006a,b).MgGeO3-Pv transforms to PPv at 60 GPa (Figure 2c ) (Hirose et al. 2005; Kuboet al. 2006). These studies have demonstrated that the proposed structure can fit theobserved diffraction patterns well.

Pressure-induced phase transitions are often accompanied by an increase in thecoordination number of cations, as best demonstrated by the postspinel transition(Mg2SiO4 ringwoodite → MgSiO3 Pv + MgO periclase) in which the coordinationnumber of Si increases from 4 to 6. Figure 3 compares the crystal structures of Pvand PPv at 120 GPa using Oganov & Ono’s (2004) first-principles results. Accordingto the proposed structure, the coordination number of Si does not increase acrossthe PPv transition (Figure 3c,d ). However, there is a profound change in the waythe SiO6 octahedra link with each other: In PPv they share edges along the a axis

572 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

22201816141210864

Two theta (angles)

Inte

nsi

ty (

arb

itra

ry u

nit

)

Re

MgSiO3-PPvSyn, 120 GPa

a

(Mg0.9Fe0.1)SiO3-PPv135 GPa and 2500 K

b

MgGeO3-PPv88 GPa and 300 K

c

NaMgF3-PPv and Pv32 GPa and 300 K

d

Mn2O3-PPv40 GPa and 300 K

e

PPv

Pv

Internal pressure calibrant

Pressure transmitting medium

Diffraction lines

Figure 2(a) Powder diffraction patterns of the postperovskite (PPv) phases in MgSiO3 synthesizedusing Oganov & Ono’s (2004) first-principles predictions. (b) (Mg0.9Fe0.1)SiO3 measured withthe gold pressure calibrant and an Ar pressure medium (S.-H. Shim, K. Catalli, J. Hustoft,A. Kubo, V.B. Prakapenka, W.A. Caldwell & M. Kunz, manuscript in review). (c) MgGeO3measured with the platinum pressure calibrant and an NaCl-B2 pressure medium (A. Kubo,B. Kiefer, S.-H. Shim, G. Shen & V.B. Prakapenka, manuscript in review). (d ) NaMgF3measured with the platinum calibrant and an Ar pressure medium [owing to incompletetransformation, perovskite (Pv) diffraction lines are still observed ( J. Hustoft, K. Catalli, S.-H.Shim, A. Kubo, V.B. Prakapenka, M. Kunz, manuscript in preparation)]. (e) Mn2O3 measuredwith an Ar pressure medium (Santillan et al. 2006). The vertical bars indicate the main diffractionlines of PPv (red ), Pv (blue), the internal pressure calibrant (black), and the pressure-transmittingmedium ( gray). The backgrounds were subtracted from the diffraction patterns. X-ray energyis 30 keV, except for in panel c and d, which were collected with 37 and 40 keV, respectively.

and corners along the c axis, resulting in a layered structure, whereas they share onlycorners to form a three-dimensional network in Pv (Figure 3a,b).

The edge-sharing SiO6 octahedra have been found in stishovite (SiO2) and akimo-toite (an ilmenite-structured MgSiO3 polymorph) at high pressures. Although edgesharing may provide more efficient packing than corner sharing, strongly chargedSi4+ ions in the adjacent edge-sharing octahedra are not as effectively shielded by

www.annualreviews.org • The Postperovskite Transition 573

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Figure 3Crystal structures of (a,c) perovskite and (b,d ) postperovskite. Crystallographic orientations of the models are indicated byunit-cell vectors at the left bottom corners. The red and orange sticks represent shorter and longer bonds, respectively,between Mg at the center and O at the corner of the MgO8 bipolar prism. The yellow sticks represent the two poles of thebipolar prism. If a longer Mg–O bond (shown in light blue) is included, the Mg–O coordination can be described as a tripolarprism. Even longer Mg–O bonds can be considered (shown in dashed lines). The structural parameters are obtained fromfirst-principles calculations at 120 GPa (Oganov & Ono 2004).

574 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

O2− ions as in the corner-sharing octahedra in Pv. In the proposed PPv structure,there exist two crystallographically distinct oxygen sites: an edge-sharing oxygensite (O2) and a corner-sharing oxygen site (O1). The proposed structure from first-principles calculations (Oganov & Ono 2004) and the refined structure of MgGeO3-PPv from diffraction measurements (A. Kubo, B. Kiefer, S.-H. Shim, G. Shen & V.B.Prakapenka, manuscript in review) indicate that the Si-O2-Si angle is larger than 90◦,and the Si-O2 bonds are longer than the Si-O1 bonds (in the case of MgGeO3, theyare the Ge-O2-Ge angle, and Ge-O2 and Ge-O1 bonds, respectively) (Figure 3d ),both of which would improve the shielding between two adjacent Si4+ (Ge4+ in caseof MgGeO3) ions. Conversely, the enhanced cation-cation interaction between theedge-sharing octahedra may influence the physical properties of the mantle PPv phasewith Fe and Al.

The coordination of Mg atoms is more complicated (Figure 3c,d ). In an idealcubic Pv structure, large cations are coordinated by 12 anions in a dodecahedron.However, Mg atoms, which are smaller than the available size of the dodecahedralsite in Pv, induce distortions in the corner-sharing SiO6 octahedra network such thatthe coordination number reduces from 12 to 8 (Horiuchi et al. 1987). As shown inFigure 3c, the Mg-O polyhedra in Pv can be described by a bipolar prism. Thesame configuration can be found in PPv but with less difference among the distancesbetween Mg and O in the prism. Overall, the Mg-O polyhedra in PPv are less distortedthan those in Pv according to the proposed structure.

4. CRYSTAL CHEMISTRY

In this section, I discuss the results that can be used for inferring the capacity ofthe PPv phase for geochemically important elements (i.e., solubility). The actualamounts of elements in mantle PPv is determined by their abundances in the bulkmantle and their partitioning among different mantle phases, MgSiO3-Pv, MgSiO3-PPv, ferropericlase (Fp), and CaSiO3-Pv. However, not many results exist for theelement partitioning involving PPv at the moment. Results on Fe partitioning arediscussed in Section 5.1.

The mantle is expected to contain approximately 10 wt% FeO. Mao et al. (2005)synthesized PPv from pyroxenes with various amounts of Fe and showed that up to 85mol% Fe can be dissolved in the PPv phase, which is a very large amount comparedwith the previously inferred Fe solubility in Pv (Mao et al. 1997, Andrault 2001).The large capacity for Fe has interesting implications, because PPv in the lowermostmantle could be in direct contact with the outer core comprising liquid Fe-Ni alloy.

McCammon (1997) and Frost et al. (2004) proposed a significant amount of Fe3+

for the lower mantle through a valence disproportionation: 2Fe2+ (in silicate) →Fe3+ (in silicate) + Fe (in metallic phase). They suggested that Al in Pv would furtherincrease the amount of Fe3+ to Fe3+/�Fe = 0.6. Sinmyo et al. (2006) have proposedthat the amount of Fe3+ in PPv is similar to that in Pv from electron energy-lossnear-edge structure spectroscopy of the sample recovered from the stable pressure-temperature conditions of PPv. Earth’s mantle is expected to contain a few wt% ofAl2O3 (Ringwood 1975). Tateno et al. (2005) have synthesized PPv with 25 mol%Al2O3.

www.annualreviews.org • The Postperovskite Transition 575

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Similar to Pv (Navrotsky et al. 2003), these trivalent cations may enter the PPvstructure through two different substitutions: (a) charge-coupled substitution,

VIIIMg2+ + VISi4+ → VIIIAl3+ + VIAl3+, (1)

and (b) oxygen defect substitution,

VISi4+ + 12

O2− →VI Al3+ + VO, (2)

where VO is an oxygen vacancy, and the superscripts before the elemental symbolsare the coordination numbers of the crystallographic sites in the structures of Pvand PPv. First-principles calculations have shown that the charge-coupled substitu-tion (Equation 1) is energetically more favorable for Fe3+ and Al in PPv than theoxygen vacancy substitution (Equation 2) at pressure-temperature conditions of thelowermost mantle (Akber-Knutson et al. 2005; Zhang & Oganov 2006a,b). This is incontrast with Mg-Pv, as it has been suggested that the oxygen vacancy substitutioncan still exist up to shallow lower-mantle conditions (30 GPa) for Al (Brodholt 2000).This difference may be significant for the mantle, as Brodholt (2000) proposed thatthe oxygen vacancy substitution can decrease bulk modulus, at least in Pv (Zhang &Weidner 1999, Daniel et al. 2001).

The stability field of PPv has been found in some sesquioxides: Al2O3 (P >

130 GPa) (Caracas & Cohen 2005b, Oganov & Ono 2005, Tsuchiya et al. 2005a),Fe2O3 (P > 60 GPa) (Ono et al. 2004, Ono & Ohishi 2005), and Mn2O3 (P > 30 GPa)(Santillan et al. 2006). In Al2O3, PPv appears after the Rh2O3-II type phase (Ono et al.2006b), and some first-principles calculations (Oganov & Ono 2005) suggest that thestability field of the Pv phase may be limited to very narrow pressure-temperatureconditions between the stability fields of Rh2O3-II and PPv at high temperature. InFe2O3, it is still unclear if the orthorhombic phase between the stability fields ofhematite and PPv has a Pbnm-Pv-type or Rh2O3-II type structure (Rozenberg et al.2002, Ono et al. 2005). In Mn2O3, the PPv stability field appears after an α-Mn2O3

phase, which also has a layered structure. Although these observations indicate thatthe sesquioxides do not follow the exact same sequence of phase transitions as sili-cates, the existence of the stability field for PPv may imply significant solubility ofthese trivalent cations in MgSiO3-PPv. Notably, Andrault & Bolfan-Casanova (2001)reported synthesis of the Pv phase with a composition of Mg3Fe2Si3O12 over 60 GPa.However, the solubility of Fe3+ has not been measured in the PPv phase. Tateno et al.(2005) reported a synthesis of pure Mg3Al2Si3O12-PPv, which supports the large sol-ubility of Al in PPv.

A large number of compounds are known to have the Pv structure, which hasbeen rationalized by the flexibility of the Pv lattice consisting of the corner sharingoctahedra (Woodward 1997). In fact, computer simulations (Corgne et al. 2003)have shown that silicate Pv has large storage capacity for trace elements. From theconnectivity among the SiO6 octahedra, it can be postulated that the PPv lattice isless flexible than the Pv lattice. However, the number of compounds with the stabilityfield of the PPv phase is rapidly growing. Therefore, the PPv lattice might also bequite flexible. Further investigation is necessary as this information is particularlyimportant for understanding the partitioning of geochemically important elements(such as tracers and radioactive elements).

576 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

5. THE POSTPEROVSKITE TRANSITION

5.1. Depth of the Transition

To confirm the existence of the PPv phase in the lowermost mantle, it is important toknow if the pressure-temperature conditions for the PPv transition agree with thoseexpected for the D′′ discontinuity (250 ± 100 km above the CMB or 115–130 GPaand 2500 K) (Lay et al. 1998). Yet it is important to realize that even the phases thatdo participate in the PPv transition (such as Fp and CaSiO3-Pv) can still influencethe transition through element partitioning. However, the effect of element parti-tioning is poorly understood for the PPv transition and the existing measurementson multicomponent systems over 100 GPa suffer larger uncertainty. More resultsexist for simple systems, such as pure MgSiO3 and binary systems (MgSiO3-FeSiO3,MgSiO3-Fe2O3, and MgSiO3-Al2O3). The results on the binary systems also allowus to understand the response of the PPv transition to compositional heterogeneities.

Although they do not provide tight constraints, earlier measurements on pureMgSiO3 (Murakami et al. 2004, Oganov & Ono 2004) and a pyrolitic composition(Murakami et al. 2005) suggest that the boundary is located at a similar depth as theD′′ discontinuity (Figure 4). Furthermore, the first-principles estimations for theMgSiO3 end member show reasonable agreement with the D′′ discontinuity (Iitakaet al. 2004, Oganov & Ono 2004, Tsuchiya et al. 2004b).

A later study by Ono & Oganov (2005) provided more dense data coverage nearthe boundary and showed that the transition depth agrees with the D′′ discontinuity

14013012011010090

Pressure (GPa)

3000

2500

2000

1500

1000

Tem

per

atu

re (

K)

0200400600800

Elevation from CMB (km)

Au(S) Au(H) Au(T)

MgO

Pt(J) Pt(H)

Perovskite

Postperovskite

Figure 4Postperovskite phase boundary determined by the gold (red lines) (Hirose et al. 2006a), MgO(black line) (Hirose et al. 2006a), and platinum (blue lines) scales (Ono & Oganov 2005). Theeffects of different pressure scales are shown for gold [Au(S) (Shim et al. 2002), Au(H) (Heinz& Jeanloz 1984), and Au(T) (Tsuchiya 2003)] and platinum [Pt( J) ( Jamieson et al. 1982) andPt(H) (Holmes et al. 1989)]. The MgO scale is by Speziale et al. (2001). The gray shaded arearepresents the depth range at which seismology studies have documented the D′′ discontinuity(Lay et al. 1998, Wysession et al. 1998).

www.annualreviews.org • The Postperovskite Transition 577

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

for pure MgSiO3 and peridotitic systems. However, the authors used different pres-sure scales for these two systems, Holmes and colleagues’ (1989) platinum scale forMgSiO3 and Jamieson and colleagues’ (1982) gold scale for the peridotitic composi-tion. The difference between these two scales can be as large as 10–20 GPa at pressuresover 100 GPa, according to Akahama et al. (2002) and Dewaele et al. (2004).

Measurements on Fe-bearing systems have suggested that the transition pressuredecreases by as much as 10–15 GPa for 10 mol% Fe (Mao et al. 2004, Shieh et al. 2006).This was subsequently supported by first-principles calculations (Caracas & Cohen2005a, Mao et al. 2005, Ono & Oganov 2005, Stackhouse et al. 2006b). Stackhouseet al. (2006b) also indicate that the width of the two-phase region would becomeas large as 200–300 km for 10 mol% Fe. However, seismologic studies have indi-cated that the thickness of the D′′ discontinuity is smaller than 50–75 km (Wysessionet al. 1998). Hirose et al. (2006b) studied pure MgSiO3, (Mg0.89Fe0.11)SiO3, andpyrolitic compositions and reported that Fe had very little effect on the transitiondepth.

Although 10%–60% of Fe in (Mg,Fe)SiO3-PPv can be Fe3+ (Sinmyo et al. 2006),the effect of Fe3+ has not been well understood. According to Ono & Ohishi (2005),the PPv transition occurs at 60 GPa in pure Fe2O3, which is significantly lower thanthat in MgSiO3. Assuming a solid-solution relation between MgSiO3 and Fe2O3, theypostulated that Fe3+ would decrease the transition pressure in mantle silicate PPv,which was supported by a later first-principles study (Ono & Oganov 2005). However,the simple assumption of solid solution may not be appropriate between MgSiO3 andFe2O3. For example, it is possible that the spin state of Fe3+ in Fe2O3-PPv is differentfrom that in (Mg,Fe3+)(Fe3+,Si)O3-PPv.

Tateno et al. (2005) found that 25 mol% Al2O3 increases the PPv transition pres-sure by 20–40 GPa, which is subsequently supported by some first-principles studies(Akber-Knutson et al. 2005, Oganov & Ono 2005, Zhang & Oganov 2006a). Akber-Knutson et al. (2005) proposed that the Pv + PPv mixed-phase region can be asthick as 150 km in an Al-bearing system. However, other first-principles calcula-tions predicted that the change in pressure is less than 1 GPa in the mantle (Caracas& Cohen 2005a, Stackhouse et al. 2005a). It should be mentioned that the formerfirst-principles calculations were conducted for Al-bearing PPv, whereas the latterproposals are based on linear interpolation of the results for pure Al2O3 and MgSiO3

end members.One of the most important sources of uncertainty over 100 GPa of pressure is the

accuracy of the internal pressure scales (Shim et al. 2001, Dewaele et al. 2004, Fei et al.2004, Shim 2005). Most pressure scales have been calibrated using a shock wave tech-nique that provides direct pressure measurements. However, the thermal effect duringshock events is difficult to separate (Vassiliou & Ahrens 1981). Elasticity measure-ments can also provide absolute pressures (Zha et al. 2000). However, this method hasyet to be applied for extreme high pressures and temperatures. Both nonhydrostatic(Akahama & Kawamura 2004) and quasi-hydrostatic (Dewaele et al. 2004) measure-ments indicate that gold and platinum, the two most popular pressure standards, candiffer by as much as 10–20 GPa at pressures over 100 GPa at 300 K (Shim 2005).The discrepancy at high temperature is still unknown at pressures over 100 GPa.

578 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Therefore, one must be cautious when comparing the results tied to different pressurescales.

Some X-ray studies (Shim et al. 2004, Mao et al. 2005, Shieh et al. 2006) haveshown that the PPv transition is sluggish. It appears that complete transformationfrom Pv to pure PPv is quite difficult to achieve below 2500 K in the laser-heated DAC,although complete transformation to PPv has been reported from the gel, glass, andpressure amorphized phases (Murakami et al. 2004, Oganov & Ono 2004). Becauseof the sluggishness of the transition, studies have often used growth and reductionin the diffraction peak intensity for the criteria. The data points for the PPv stabilityin some studies still represent a PPv + Pv mixture, not pure PPv. It is importantto note that many other factors can affect the diffraction intensity, including crystalgrowth, preferred orientation, and recrystallization. Furthermore, a binary system hasa finite pressure range for a phase transition in which both low- and high-pressurephases coexist (a two-phase region). Therefore, the former observation of the mixed-phase assemblage may indicate that the pressure and temperature conditions are stillin a two-phase region. More reliable determination is possible using the completedisappearance of either phase, and both forward (Pv → PPv) and reverse (PPv → Pv)transition paths.

As mentioned above, in the laser-heated DAC, lack of thermal insulation canresult in large gradients in pressure as well as temperature. However, many of thesemeasurements were conducted without a thermal insulation medium and/or pressure-transmitting medium. Therefore, the observation of incomplete transformation canbe also caused by large gradients in pressure and temperature.

There is some uncertainty as to the depth constrained by first-principles methods.Calculations for experimentally well-characterized phase boundaries, such as postil-menite (Wentzcovitch et al. 2004) and postspinel (Yu et al. 2007) transitions, havedemonstrated that results can differ by 6–12 GPa depending on the approximationused: local density approximation or generalized gradient approximation.

One must be careful when applying the width of the two-phase region determinedfor simple binary systems to Earth’s mantle. A fundamental difference is that themantle rocks contain multiple phases. As demonstrated for the olivine-wadsleyitephase boundary (Irifune & Isshiki 1998), element partitioning among different phases,including the phases not directly involved in the transition, can change the widthsignificantly.

The partitioning of Fe among Pv, PPv, and Fp is controversial: Murakami et al.(2005) proposed that Fe preferentially enters Pv in a pyrolitic composition, whereasKobayashi et al. (2005) suggested the opposite for an olivine composition, althoughboth studies are in agreement in that the amount of FeO is much higher in coex-isting Fp. First-principles calculations by Stackhouse et al. (2006b) suggested thatFeSiO3 preferentially enters PPv compared with Pv, which is also consistent withthe large solubility of Fe discussed in Section 4. Although the solubility of Fe inCaSiO3-Pv may be limited (Kim et al. 1991), it can store much Al (Kurashinaet al. 2004), whereas the partitioning of Al would be limited in Fp. However, thepartitioning of Al among MgSiO3-Pv, MgSiO3-PPv, and CaSiO3-Pv is not wellconstrained.

www.annualreviews.org • The Postperovskite Transition 579

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

5.2. The Clapeyron Slope

Recent seismology studies (Lay et al. 2006, van der Hilst et al. 2007) have shown thatthe Clapeyron slope of the PPv transition may provide constraints for the tempera-ture profile of the lowermost mantle. Hernlund et al. (2005) pointed out that becausethe Clapeyron slope of the PPv transition is comparable to the expected radial tem-perature gradient at the CMB, the PPv boundary may intercept the temperatureprofile twice at colder regions of the mantle: Pv → PPv at a shallower depth andPPv → Pv at a deeper depth. This provides two separate temperature measurementsat the different depths, which then can be used for the calculation of the temperaturegradient. Another parameter that can be constrained from this is the heat flux outof the core (Lay et al. 2006, van der Hilst et al. 2007). Furthermore, the sign andmagnitude of the boundary slope have a profound effect on mantle dynamics as bestdemonstrated for the postspinel transition at 660-km depth (Tackley et al. 1993).

Before the discovery of the PPv transition, Sidorin et al. (1999) refined the tomo-graphic models by including a global seismic discontinuity at the lowermost man-tle and found that the inclusion of a phase boundary with a positive Clapeyronslope better reproduces regional seismologic observations. The predicted slope was+6 MPa K−1, which is larger than the slopes of any known phase boundaries inmagnesium silicates in the upper mantle (Table 1).

The estimated slope from first-principles studies [+7 ∼ 10 MPa K−1 (Iitaka et al.2004, Oganov & Ono 2004, Tsuchiya et al. 2004b)] is consistent with the seismicprediction (Sidorin et al. 1999). Using the platinum scale (Holmes et al. 1989), Ono &Oganov (2005) suggested that the Clapeyron slope is +7 MPa K−1 for pure MgSiO3.Hirose et al. (2006a) found that the Clapeyron slope as well as the transition pressureis highly sensitive to the pressure scale used: The gold scale (Tsuchiya 2003) yields+4.7 MPa K−1, whereas the MgO scale (Speziale et al. 2001) yields +11.5 MPa K−1.

The sign of the Clapeyron slope of a phase boundary is strongly influenced by thedifferences in crystal structure between lower and higher pressure (or temperature)

Table 1 The Clapeyron slopes of the mantle transitions

Clapeyron slope (MPa K−1)Olivine → wadsleyitea +2.5–4.0Wadsleyite → ringwooditeb +5.0–6.9Postspinelc −2.5–3.0Postilmenited −2.5–3.0Postgarnete +2.4Postperovskitef +7–10

aKatsura & Ito 1989, Morishima et al. 1994, Katsura et al. 2004.bKatsura & Ito 1989, Suzuki et al. 2000.cIto & Takahashi 1989, Irifune et al. 1998.dIto & Takahashi 1989, Kuroda et al. 2000.eHirose et al. 2001.fIitaka et al. 2004, Oganov & Ono 2004, Tsuchiya et al. 2004b.

580 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

phases. From the Clausious-Clapeyron relation,

dPdT

= �Str

�Vtr, (3)

where �Str and �Vtr are the changes in entropy and volume across a phase transition,respectively. It can be easily inferred that the positive Clapeyron slope of a high-pressure transition, which is accompanied with a volume decrease �Vtr = (–), isassociated with an entropy decrease �Str = (–).

Navrotsky (1980) predicted that the mantle transitions at pressures higher than15 GPa may have negative Clapeyron slopes in general. Her prediction is based on theassumption that high-pressure mantle transitions are associated with a coordinationnumber increase in small cations, such as Si, which results in the low-frequencyvibration of the small cations in the large sites with a high coordination numberand increased bond distances with anions. Therefore, entropy increases after thephase transition, which would make the Clapeyron slope negative. This explains thenegative Clapeyron slope of the postspinel transition.

From this prediction, the strong positive slope of the PPv transition is rather sur-prising. However, as discussed above, there is essentially no increase in coordinationnumber of Mg and Si across the PPv transition. An important change in the PPvtransition, however, is the formation of the edge-sharing linkage among the SiO6

octahedra (Figure 3).A Raman spectroscopy study of MgGeO3 has shown that the vibrational modes

of PPv are extended to much higher frequency than Pv (Shim et al. 2007). In theirentropy calculation using Kieffer’s (1979) approximation, the existence of the high-frequency modes in PPv lowers the entropy across the PPv transition. They attributedthe extended-mode frequencies in PPv to more restricted motions of edge-sharingoxygen atoms, whereas in Pv all oxygen atoms link octahedra by corners, which makesthe vibration of oxygen atoms less restricted.

The Raman mode frequencies have also been calculated for MgSiO3-PPv usingdensity functional theory (Caracas & Cohen 2006, Ono et al. 2006b). These predictedvalues are significantly different from the measurements for MgGeO3 (Shim et al.2007) even if one considers the difference in the octahedrally coordinated cations.First-principles calculations have successfully predicted Raman mode frequencieswithin �ν = ±20 cm−1 for MgSiO3-Pv (Karki et al. 2000, Caracas & Cohen 2006),MgAl2O4 spinel (Lazzeri & Thilbaudeau 2006), and Mg2SiO4 forsterite (Noel et al.2006). However, the two existing first-principles predictions for PPv are not con-sistent with each other for the Raman mode frequencies. Nevertheless, the Ramanobservation is in qualitative agreement with the predicted phonon density of statesfrom first-principles in that more density of states exists in PPv at higher frequenciesthan Pv.

The effects of Fe and Al on the slope have not been investigated. Studies of chem-ically complex systems (Murakami et al. 2005, Ono & Oganov 2005) have suggestedthat no significant change is found for the sign and the magnitude of the slope.

The large positive slope of the PPv transition coupled with the temperature gradi-ent can result in intriguing structures at the lowermost mantle. As shown in Figure 5a,

www.annualreviews.org • The Postperovskite Transition 581

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

PPv

Pv

Pv

Temperature

Dep

thD

epth

Dep

th

CMB

PvLower mantle

Pv

PPv

Fe-rich

Pv

PPv

Pv

PPv

WarmerColder

Pv

PPv

Pv

PPv

PPv

a

b

c

Fe-rich

Pv

PPv

Pv

Pv

Outer core

Colder WarmerAverage

Figure 5Schematic models ofstructures at the D′′ regioninferred from the effect of(a) temperature and(b) chemical composition.(c) A lens-shapedpostperovskite (PPv)structure predicted from thedouble crossing. The colorsin the left and middlecolumns representtemperature variations inthe mantle. The solid blacklines represent the PPvboundary, and the dashedcurves show regions with Feenrichment. The rightcolumn shows schematicphase diagrams withtemperature profiles inwarmer (red ), average(yellow), and colder (blue)regions.

because of the positive Clapeyron slope, the transition occurs at a deeper depth inwarmer regions of the mantle. It is possible that the geotherm does not even intersectthe PPv boundary at hot mantle regions (Figure 5a,c). If so, the PPv transition mightnot occur globally but rather locally at colder regions. It is not clear whether the D′′

discontinuity is a global feature because of the existence of areas with seismic signalsbelow noise level (Wysession et al. 1998).

Recent seismic studies have confirmed the existence of lens-shaped structureslikely resulting from the double crossing (Lay et al. 2006, van der Hilst et al. 2007).This type of structure would exist more likely in the colder regions of the mantle inwhich the PPv transition and mantle geotherm are very close (Figure 5c). van derHilst et al. (2007) observed structures that can be related to the double crossing atregions with the possible existence of cold subducted materials.

However, variations in chemical composition introduce some complications intothis simple picture. Provided that Fe decreases the transition depth as suggested byMao et al. (2004), the PPv boundary would be elevated at Fe-rich regions (Figure 5b).

582 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Depending on the magnitude of the effect, Fe enrichment can make Pv transformto PPv even at regions with sufficiently high temperature at which the PPv transi-tion would not occur otherwise (Figure 5b). The opposite effect is expected for Al,provided that Al increases the depth of PPv transition as suggested by Tateno et al.(2005). However, the compositional effect might be smaller in the mantle because ofthe partitioning among MgSiO3-Pv, MgSiO3-PPv, Fp, and CaSiO3-Pv.

Some mantle simulations (Nakagawa & Tackley 2004, Matyska & Yuen 2005)have shown that the proposed strong positive slope of the PPv transition may pro-mote material exchange across the boundary, resulting in developments of small-scalefeatures instead of stable stationary mantle plumes. This is not consistent with thedominant view for mantle plumes. Therefore, these simulations invoked other fac-tors, such as chemical heterogeneities and radiative heat transfer, to stabilize mantleplumes.

6. PHYSICAL PROPERTIES

6.1. Equation of State and Elasticity

Physical property changes, such as density and seismic velocities, across a seismicdiscontinuity provide important tests for an isochemical phase-transition model. A1.5%–3.0% increase in seismic velocities has been documented for the D′′ disconti-nuity with a smaller increase (or even no increase at some regions) in P wave thanS wave (Lay et al. 1998, Wysession et al. 1998). Furthermore, some tomographicstudies (Masters et al. 2000, Trampert et al. 2004) have shown that lateral variationsin the bulk sound speed and shear velocity anomalies are anticorrelated at the lowermantle. Also significant reduction in seismic velocities (ultralow velocity zones) hasbeen detected at some regions at the base of the mantle (Lay et al. 1998). However,the origin of the velocity reduction is still unclear.

The density and bulk modulus (these allow for calculation of bulk sound speed)have been obtained for PPv by fitting measured pressure-volume relations to equa-tion of state (EOS) (Table 2). Two data sets exist for MgSiO3-PPv (Ono et al. 2006a,Guignot et al. 2007) (Figure 6). However, the volume reported by Ono et al. (2006a)is larger than that reported by Guignot et al. (2007). It is even larger than thatof Fe-bearing PPv, if Shieh and colleagues’ (2006) fitted EOS is extrapolated, al-though the larger ionic size of Fe is expected to increase the volume. The bulkmodulus is in agreement in these studies, considering the fact that uncertainty in vol-ume (0.5%) is much smaller that that in bulk modulus (2%) in these measurements(Table 2).

The volume data for (Fe0.4Mg0.6)SiO3-PPv measured by Mao et al. (2006a,b)have the most dense coverage for a wide pressure range (Figure 6). The authorsfound an abnormally high compressibility below 90 GPa, which was attributedto the metastability of PPv. This behavior was also documented in MgGeO3-PPv(Kubo et al. 2006). Therefore, Mao et al. (2006a) used the data points above100 GPa for the EOS fitting. Their bulk modulus of PPv at CMB pressures ismuch higher (25%–40%) than the predictions from first principles for MgSiO3-PPv

www.annualreviews.org • The Postperovskite Transition 583

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Table 2 Volumes, bulk moduli, and pressure derivatives of bulk moduli of perovskite and postperovskite atambient and core-mantle boundary pressures

V (A3) K (GPa) K ′ P-scale Medium

Pressure (GPa) 0 125 0 125 0Experiment

MgSiO3-PPvOno et al. 2006a 162.9a 121.5 237 677 4b Au-A NaCl-B2Guignot et al. 2007 162.2b 120.4 231 671 4b MgO-S NaCl-B2(Mg,Fe)SiO3-PPvShieh et al. 2006c 164.9 120.8 215 653 4b Pt-H ArMao et al. 2006ad 157.2 124.7 272 908 6b Pt-J NoneMgSiO3-PvFiquet et al. 2000 162.3 122.0 259 665 3.7 Pt-H Ar, none

TheoryMgSiO3-PPvTsuchiya et al. 2005be 163.9 121.3 216 684 4.4Oganov & Ono 2004 162.9 122.4 232 717 4.4

(167.6) (123.0) (200) (691) (4.5)FeSiO3-PPvCaracas & Cohen 2005af 164.7 124.4 243 719 4.3

(177.5) (130.9) (212) (684) (4.3)

For theory, numbers without parenthesis are from local density approximation, and numbers in the parentheses are from generalizedgradient approximation. Au-A, gold equation of state (EOS) by Anderson et al. (1989); MgO-S, periclase EOS by Speziale et al. (2001);Pt-H, platinum EOS by Holmes et al. (1989); Pt-J, platinum EOS by Jamieson et al. (1982). PPv, postperovskite; Pv, perovskite.aFixed the volume to Oganov & Ono 2004.bFixed.cThe PPv phase is synthesized from 9 mol% Fe pyroxene, but the composition of the PPv phase is unknown.d(Mg0.6Fe0.4)SiO3.eAt 300 K.fAt 0 K. Average of antiferromagnetic and ferromagnetic phases.

(Oganov & Ono 2004, Tsuchiya et al. 2004b) and for a hypothetical Fe end member(FeSiO3-PPv) (Caracas & Cohen 2005a, Tsuchiya & Tsuchiya 2006, Stackhouseet al. 2006b) (Table 2). The lack of data on Pv with the same composition makes itdifficult to infer the physical property changes at the PPv transition from this study.

Shieh et al. (2006) measured PPv synthesized from (Fe0.09Mg0.91)SiO3 pyrox-ene in an Ar medium. In this measurement, the coexistence of PPv and Pv dueto an incomplete transition allows for a direct comparison of these two phases for awide pressure range (12–106 GPa). The authors found that the volume and bulksound speed of PPv decrease by 1.1 ± 0.2% and 2.3 ± 2.1%, respectively, atthe PPv transition, which is in agreement with first-principles calculations (Iitakaet al. 2004, Oganov & Ono 2004, Tsuchiya et al. 2004a, Wentzcovitch et al. 2006)(Table 2).

However, the amounts of Fe in Pv and PPv are not known in the measurementsby Shieh et al. (2006), and based on recent studies (Mao et al. 2004), it is unlikely

584 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

170

160

150

140

130

120

Un

it-c

ell v

olu

me

(Å3 )

160140120100806040200

Pressure (GPa)

Experiments on MgSiO3-PPvOno et al. (2006a)Guignot et al. (2007)

Experiments on (Mg,Fe)SiO3-PPvMao et al. (2006a)Mao et al. (2006b)Shieh et al. (2006)

Theory on MgSiO3-PPvOganov and Ono (2004) LDAOganov and Ono (2004) GGATsuchiya et al. (2004b) LDA

MgSiO3-PvFiquet et al. (2000)

Figure 6The volume of MgSiO3-PPv (blue symbols) and (Mg,Fe)SiO3-PPv (red symbols). Thefirst-principles results are shown in orange curves. The equation of state of MgSiO3-Pv isshown in a gray curve for comparison.

that Fe partitions into Pv and PPv in equal amounts. Also, because of the largegap in the data between 40 and 80 GPa (Figure 6), it is not clear whether the samemetastable behavior observed by Mao et al. (2006a) exists in the PPv data by Shieh et al.(2006).

First-principles calculations are in agreement in that Al and Fe3+ do not changethe bulk modulus of PPv significantly (Akber-Knutson et al. 2005, Caracas & Cohen2005a, Oganov & Ono 2005, Tsuchiya et al. 2005a, Stackhouse et al. 2006b, Tsuchiya& Tsuchiya 2006). For example, Caracas & Cohen (2005b) showed that the changedoes not exceed 5% for both Fe3+ and Al with mantle concentrations.

Constraints on the shear modulus of PPv have been provided mostly by first princi-ples. Existing calculations are in agreement that the shear modulus would increase by5%–7% across the PPv transition (Iitaka et al. 2004, Oganov & Ono 2004, Tsuchiyaet al. 2004a, Wentzcovitch et al. 2006). Some first-principles studies (Stackhouse et al.2005a, Tsuchiya & Tsuchiya 2006) have found that the effect of chemical compositionis small for the change in shear modulus.

If the bulk modulus decreases as suggested by Shieh et al. (2006) and the shear mod-ulus increases as suggested by first-principles calculations across the PPv transition,

www.annualreviews.org • The Postperovskite Transition 585

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

the lateral variations in the contents of PPv and Pv should result in anticorrelation be-tween bulk sound speed and shear wave velocity anomalies. Furthermore, the decreasein bulk sound speed also can explain a smaller P-wave velocity increase observed atthe D′′ discontinuity. However, if Mao and colleagues’ (2006a,b) result is used, bulkmodulus should increase at the PPv transition, provided that Fe does not have alarge effect on the bulk modulus as suggested by some first-principles calculations(Table 2).

Mao et al. (2006b) obtained some indirect constraints on the shear modulus ofPPv by combining data from nuclear resonant inelastic X-ray scattering spectroscopyand X-ray diffraction of (Fe0.4Mg0.6)SiO3-PPv. Their data indicate large reductionsin the P- and S-wave velocities by Fe. They further argued that the Fe enrichment inPPv would provide an alternative explanation for the ultralow velocity zones (Maoet al. 2006b).

This proposal assumes that PPv exists at the bottommost mantle, i.e., where ul-tralow velocity zones exist. However, if the core temperature is sufficiently high andthe existence of PPv is limited to colder regions, Pv may be the dominant phase athot regions of the lowermost mantle. If the solubility of Fe is also very high in Pv atthe lowermost mantle as proposed by Tateno et al. (2007), it is worthwhile to conductsimilar measurements on Fe-rich Pv.

Murakami et al. (2007) have conducted Brillouin spectroscopy of powder MgSiO3-PPv in its stability field and found that the increase in shear-wave velocity across theboundary would not exceed 0.5%, which is much smaller than the 2%–3% observedfor the D′′ discontinuity. As discussed below, PPv develops strong lattice preferredorientation (LPO), which could affect the Brillouin measurements of aggregates. Itis also important to investigate the effect of Fe for the shear moduli of Pv and PPv.

Tsuchiya et al. (2005b) have investigated the effect of temperature on the EOSof PPv using first-principles calculations. They predicted that thermal expansivitydoes not change significantly across the PPv transition. Recent high-temperatureX-ray diffraction measurements on MgSiO3-PPv (Guignot et al. 2007) support thisresult. However, because of their limited pressure-temperature coverage, constraintsfrom first principles were necessary for data interpretation. A recent Raman spec-troscopy study on MgGeO3-Pv and PPv suggested that the thermal expansivity andGruneisen parameter decrease by 25% ± 10% across the PPv transition (Shim et al.2007).

The pressure-scale inconsistency problem can introduce systematic uncertainty inthe measured compressibility. Moreover, it is difficult to compare the results obtainedwith different pressure scales. This makes it difficult to use existing data to extract thecompositional effect or change across the PPv transition. For example, most studiesfor Pv used the platinum EOS ( Jamieson et al. 1982, Holmes et al. 1989), whereasOno and colleagues’ (2006a) Mg end member PPv EOS is tied to Anderson and col-leagues’ (1989) gold scale, and Guignot and colleagues’ (2007) EOS is tied to Spezialeand colleagues’ (2001) MgO scale. In addition, some diffraction measurements wereconducted in a rigid pressure medium or even no medium. This results in severepressure gradients and deviatoric stresses in the sample, which introduce systematicuncertainties in the fitted parameters.

586 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

6.2. Elastic Anisotropy and Lattice Preferred Orientation

Seismologists have observed that the horizontally polarized S-wave (VSH) arrivesearlier than the vertically polarized one (VSV ) beneath the circum-Pacific regions ofthe D′′ layer, whereas it is the opposite in the central Pacific (Kendall & Silver 1998,Lay et al. 1998, Panning & Romanowicz 2004). The magnitude of the splitting can beas high as 10 s, which corresponds to approximately 3% of polarizational anisotropyat the D′′ layer (Lay et al. 1998). It appears that the observations can be describedby transverse isotropy (Kendall & Silver 1998). This anisotropy has been associatedwith the large-scale horizontal shear flow at the CMB (Panning & Romanowicz 2004).However, it has not been clear whether the LPO or the shape preferred orientationin the D′′ layer is responsible for the observations (Kendall & Silver 1998). Existingmeasurements and theoretical calculations indicate that elastic anisotropy of the Pvaggregate cannot explain the D′′ anisotropy (Wentzcovitch et al. 1998, Cordier et al.2004, Stackhouse et al. 2005b).

From the two-dimensional nature of the PPv crystal structure (Figure 3), investi-gators have predicted that this structure is elastically more anisotropic than Pv. Somefirst-principles studies predicted that the PPv structure is most compressible alongthe b axis (Iitaka et al. 2004, Tsuchiya et al. 2004b), which is perpendicular to the SiO6

octahedral layers. From the largest compressibility along the b axis, Iitaka et al. (2004)and Oganov & Ono (2004) inferred that the octahedral layers would align parallel tothe horizontal shear flow direction, which is parallel to the CMB. The anisotropy forthis alignment was calculated to be 3%–4%, which is comparable to the seismic obser-vations. However, no significant change is found in measurements of the axial ratioswith pressure (Shieh et al. 2006), which indicates that the compressibility of the b axisis not particularly higher than the other axes. Tsuchiya et al. (2004b) have exploredother crystallographic orientations and found that the anisotropy is much larger forthe alignment of the (001) planes (10%) than that of the (010) plane (2%–3%).

Stackhouse et al. (2005b) conducted ab initio molecular dynamics calculations toinvestigate the effect of temperature on single-crystal elastic constants. They foundthat temperature has little effect on the single-crystal elastic anisotropy, and thealignment of the (001) planes remains to yield the highest anisotropy (15%) at man-tle temperatures. Although alignment of the (010) planes is a still viable explanation,this would require a higher degree of preferred orientation for the seismically ob-served anisotropy (Stackhouse et al. 2005b). In contrast, Wentzcovitch et al. (2006)found that the magnitude of the anisotropy of PPv decreases with temperature and be-comes similar to that of Pv using a first-principles calculation with the quasi-harmonicapproximation.

For the D′′ anisotropy, it is also important to know the dominant slip systemsin PPv. Oganov et al. (2005) have identified polytypes with energy intermediatebetween Pv and PPv through first-principles metadynamics. These structures canbe derived by stacking faults developing parallel to the (110) plane of PPv. Based onthis, the authors predicted that the dominant slip plane would be the (110) plane,and this would require less degree of LPO to explain the D′′ anisotropy. However,a theoretical study with a different approach, the density functional theory with the

www.annualreviews.org • The Postperovskite Transition 587

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Peierls-Nabarro model (Carrez et al. 2007), found that the dominant slip system inPPv is [001](010). These calculations were conducted at static conditions (0 K).

Merkel et al. (2006, 2007) have conducted radial X-ray diffraction measurementsfor the plastically deformed polycrystalline MgGeO3 and (Mg,Fe)SiO3-PPv in theDAC at their stable pressure conditions. They proposed that the dominant slip planeof PPv would be (100) or (110), which is consistent with Oganov and colleagues’(2005) computer simulations. Merkel et al. (2007) calculated the shear wave splittingof transversely isotropic aggregate using their LPO observations together with first-principles predictions of the single-crystal elastic constants. They found that neitherof the two theoretical results (Stackhouse et al. 2005b, Wentzcovitch et al. 2006)provides satisfactory results to explain the D′′ anisotropy. Yamazaki et al. (2006) arguedthat the strain rates in the DAC are not sufficiently high. Another important point isthat the differential stress is quite high in the radial diffraction measurements in theDAC ranging between 4 and 9 GPa (Merkel et al. 2007).

CaIrO3 is isostructural to PPv and stable at ambient conditions. The transmissionelectron microscopy observation of this material suggested that the dominant slipplane would be the (010) plane (Miyajima et al. 2006). This has been confirmed bysome recent deformation studies (Yamazaki et al. 2006, Walte et al. 2007).

Santillan et al. (2006) have studied the PPv phase in Mn2O3. Because Mn2O3

transforms to the PPv phase at room temperature and high pressure, the authorswere able to measure the LPO with and without heating. They found that the LPOchanges from (010) to (110) after heating. They argued that many different factors caninfluence the LPO of the PPv phase in laboratory measurements, such as deviatoricstress, phase transition, and thermal annealing.

Although it has been argued that crystal structure is a more important factorthan chemical composition for LPO (Karato 1989), it is still difficult to rule out thepossibility that the compositional difference is responsible for the observed discrep-ancy among different PPv structured materials. Furthermore, Yamazaki et al. (2006)pointed out that the existence of a mechanically much weaker phase, Fp (Yamazaki& Karato 2002), together with the PPv can change the LPO of multiphase aggregatesignificantly.

7. SUMMARY AND OUTSTANDING ISSUES

For the past three years, remarkable progress has been made on the PPv phase.However, many important properties of PPv are still under debate. Yet rapid devel-opments in computational power and experimental techniques at third-generationsynchrotron facilities will help to better understand this important phase at extremepressure-temperature conditions of the lowermost mantle.

One of the most important questions is whether the existence of PPv is global orlocal in the lowermost mantle. If the double crossing is ubiquitous or the temperatureat the base of the mantle is sufficiently high, Pv can still be the dominant phase atthe bottom of the mantle except for colder regions (such as regions with subductingslabs). In other words, the phase that is in direct contact with the core could still bePv. This can be examined by studying the global distribution of the D′′ discontinuity

588 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

and the double crossing. Also accurate determination of the transition depth andClapeyron slope will help resolve this issue.

Some studies indicate that the mixed-phase region (the coexistence of Pv and PPv)can be thick, >100 km, which is not compatible with seismic observations of the D′′

discontinuity. Because the PPv transition occurs only a few hundred kilometers abovethe CMB, Pv may not transform completely to PPv in the mantle if the two-phaseregion is indeed very thick (Akber-Knutson et al. 2005). Some studies indicate thatthe effects of Fe and Al would compensate for each other and result in a thinnertwo-phase region (Murakami et al. 2005). However, more accurate determination ofthe boundary is necessary. In addition, one of the most important parameters to bedetermined is the partitioning of Fe and Al among different mantle phases, such asMgSiO3-Pv, MgSiO3-PPv, Fp, and CaSiO3-Pv, which can significantly change thethickness of the mixed-phase region. It is also important to measure the partitioningof radioactive elements and geochemical tracers among the mantle phases at thelowermost mantle.

The large magnitude of the Clapeyron slope of the PPv transition can lead tosignificant topography of the discontinuity and lateral variations in mineralogy, inresponse to lateral variations in temperature. However, it is important to realize thatlateral variations in chemical composition, such as Fe and Al, can affect the topographyand mineralogy as well. It is essential, however, to improve the data on the effects ofFe and Al on the transition pressure and physical properties of PPv to use this as aviable tool to search for chemical heterogeneities.

As pointed out by Ono et al. (2005), the secular cooling of the mantle wouldresult in gradual thickening of the PPv structures at the lowermost mantle. It is evenpossible that the PPv transition did not exist in the early mantle if the content of Fe isnot much higher than the present mantle. If the PPv structures play important rolesfor the dynamics of the lowermost mantle, it would be important to investigate theinfluence of this time-dependent change in the thickness of the PPv structures on themantle dynamics.

Recently it has been shown that the spin state of Fe in Fp and Pv changes at mantlepressures (Badro et al. 2003, 2004). The spin transition may influence importantproperties, which has been confirmed for Fp (Lin et al. 2005, Speziale et al. 2005,Fei et al. 2007). Possible difference in the spin states of Fe among Pv, PPv, and Fp isparticularly interesting because it would affect Fe partitioning among these phases.Also if the spin state influences the optical and transport properties of Fe-bearingphases (Goncharov et al. 2006) and the spin state is different in mantle phases, changesin mineralogy would lead to variations in these properties at the lowermost mantle,which will have important dynamic consequences. However, the spin state of Fe inPPv needs to be investigated experimentally, although there exist some theoreticalpredictions (Stackhouse et al. 2006a, Zhang and Oganov 2006b).

Intriguingly, PPv can provide explanations for some enigmatic seismic observa-tions, such as anticorrelation between bulk sound speed and shear wave velocityanomalies and seismic anisotropy at the lowermost mantle. However, it is impor-tant to know the depth extent of the seismic structures. If indeed they are associatedwith the PPv transition, their existence should be limited to below the depth of the

www.annualreviews.org • The Postperovskite Transition 589

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

PPv transition, and the geographic distributions of these features should be consistentwith expected thermal and chemical responses of the PPv transition. Global studies ofimportant seismic features (including the ultralow velocity zones, the D′′ anisotropy,the D′′ discontinuity, and velocity anomalies) would certainly help us understand theexistence and distribution of the PPv transition.

In mineral physics, it is important to continue improving our knowledge on theproperties of Mg-Pv as well as PPv. As discussed above, Pv can still be an importantphase at the lowermost mantle. To understand radial and lateral variations found inseismology, we must understand the contrast in properties between Pv and PPv.

DISCLOSURE STATEMENT

The authors are not aware of any biases that might be perceived as affecting theobjectivity of this review.

ACKNOWLEDGMENTS

J. Hustoft and A. Kubo provided X-ray diffraction patterns of NaMgF3 and MgGeO3,respectively. Discussions with T. Duffy, R. Jeanloz, and an anonymous reviewer im-proved this manuscript. I thank K. Catalli, J. Hustoft, J. Santillan, S. Lundin, S. Rekhi,S. Slotznick, G. Shen, V. Prakapenka, A. Kubo, A. Tikhonov, S. Clark, M. Kunz, andW. Caldwell for their experimental assistance at synchrotron facilities. Part of thiswork is supported by the National Science Foundation (NSF) (EAR0337156 andEAR0337005) and the Wade fund. Some unpublished results presented here wereobtained at the GSECARS and HPCAT sectors of Advanced Photon Source andbeamline 12.2.2 of Advanced Light Source. GSECARS is supported by the NSF andDepartment of Energy (DOE). HPCAT is supported by DOE-BES, DOE-NNSA,NSF, and the W.M. Keck Foundation. The 12.2.2 beamline is supported by theCOMPRES under NSF. Use of APS and ALS is supported by the DOE.

LITERATURE CITED

Akahama Y, Kawamura H. 2004. A comparison of volume compressions of silver andgold up to 150 GPa. J. Appl. Phys. 95:4767–71

Akahama Y, Kawamura H, Singh AK. 2002. Equation of state of bismuth to 222 GPaand comparison of gold and platinum pressure scales to 145 GPa. J. Appl. Phys.92:5892–97

Akber-Knutson S, Steinle-Neumann G, Asimow PD. 2005. Effect of Al on the sharp-ness of the MgSiO3 perovskite to postperovskite phase transition. Geophys. Res.Lett. 32:L14303

Anderson OL, Isaak DG, Yamamoto S. 1989. Anharmonicity and the equation ofstate for gold. J. Appl. Phys. 65:1534–43

Andrault D. 2001. Evaluation of (Mg,Fe) partitioning between silicate perovskite andmagnesiowustite up to 120 GPa and 2300 K. J. Geophys. Res. 106:2079–87

590 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Andrault D, Bolfan-Casanova N. 2001. High-pressure phase transformations in theMgFe2O4 and Fe2O2-MgSiO3 systems. Phys. Chem. Mineral. 28:211–17

Badro J, Fiquet G, Guyot F, Rueff JP, Struzhkin VV, et al. 2003. Iron partitioningin Earth’s mantle: toward a deep lower mantle discontinuity. Science 300:789–91

Badro J, Rueff J, Vanko G, Monaco G, Fiquet G, Guyot F. 2004. Electronic transitionsin perovskite: possible nonconvecting layers in the lower mantle. Science 305:383–86

Boehler R. 2000. High-pressure experiments and the phase diagram of lower mantleand core materials. Rev. Geophys. 38:221–45

Brodholt JP. 2000. Pressure-induced changes in the compression mechanism of alu-minous perovskite in the earth’s mantle. Nature 407:620–22

Buffett BA, Garnero EJ, Jeanloz R. 2000. Sediments at the top of Earth’s core. Science290:1338–42

Caracas R, Cohen RE. 2005a. Effect of chemistry on the stability and elasticity of theperovskite and postperovskite phases in the MgSiO3-FeSiO3-Al2O3 system andimplications for the lowermost mantle. Geophys. Res. Lett. 32:L16310

Caracas R, Cohen RE. 2005b. Prediction of a new phase transition in Al2O3 at highpressures. Geophys. Res. Lett. 32:L06303

Caracas R, Cohen RE. 2006. Theoretical determination of the Raman spectra ofMgSiO3 perovskite and postperovskite at high pressure. Geophys. Res. Lett.33:L12S05

Carrez P, Ferre D, Cordier P. 2007 . Implications for plastic flow in the deep mantlefrom modelling dislocations in MgSiO3 minerals. Nature 445:68–70

Corgne A, Allan NL, Wood BJ. 2003. Atomistic simulations of trace element in-corporation into the large site of MgSiO3 and CaSiO3 perovskites. Phys. EarthPlanet. Inter. 139:113–27

Cordier P, Ungar T, Zsoldos L, Tichy G. 2004. Dislocation creep in MgSiO3 per-ovskite at conditions of the Earth’s uppermost lower mantle. Nature 428:837–40

Daniel I, Cardon H, Fiquet G, Guyot F, Mezouar M. 2001. Equation of state ofAl-bearing perovskite to lower mantle pressure conditions. Geophys. Res. Lett.28:3789–92

Dewaele A, Loubeyre P, Mezouar M. 2004. Equations of state of six metals above94 GPa. Phys. Rev. B 70:094112

Fei Y, Li H, Hirose K, Minarik W, Orman JV, et al. 2004. A critical evaluation ofpressure scales at high temperatures by in situ X-ray diffraction measurements.Phys. Earth Planet. Inter. 143–144:515–26

Fei Y, Zhang L, Corgne A, Watson H, Ricolleau A, et al. 2007. Spin transition andequation of state of (Mg,Fe)O solid solution. Geophys. Res. Lett. 34:L17307

Fiquet G, Dewaele A, Andrault D, Kunz M, Bihan TL. 2000. Thermoelastic prop-erties and crystal structure of MgSiO3 perovskite at lower mantle pressure andtemperature conditions. Geophys. Res. Lett. 27:21–24

Fouch MJ, Fischer KM, Wysession ME. 2001. Lowermost mantle anisotropy beneaththe Pacific: imaging the source of the Hawaiian plume. Earth Planet. Sci. Lett.190:167–80

www.annualreviews.org • The Postperovskite Transition 591

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Frost DJ, Liebske C, Langenhorst F, McCammon CA, Trennes RG, Rubie DC. 2004.Experimental evidence for the existence of iron-rich metal in the Earth’s lowermantle. Nature 428:409–12

Garnero EJ. 2000. Heterogeneity of the lowermost mantle. Annu. Rev. Earth PlanetSci. 28:509–37

Garnero EJ, Helmberger DV. 1995. A very slow basal layer underlying large-scalelow-velocity anomalies in the lower mantle beneath the Pacific: evidence fromcore phases. Phys. Earth Planet. Inter. 91:161–76

Garnero EJ, Lay T. 1997. Lateral variations in lowermost mantle shear waveanisotropy beneath the north Pacific and Alaska. J. Geophys. Res. 102:8121–35

Goncharov AF, Stuzhkin VV, Jacobsen SD. 2006. Reduced radiative conductivity oflow-spin (Mg,Fe)O in the lower mantle. Science 312:1205–8

Guignot N, Andrault D, Morard G, Bolfan-Casanova N, Mezouar M. 2007. Ther-moelastic properties of postperovskite phase MgSiO3 determined experimentallyat core-mantle boundary P-T conditions. Earth Planet. Sci. Lett. 256:162–68

Gurnis M, Wysession ME, Knittle E, Buffett BA. 1998. The Core-Mantle BoundaryRegion. Washington, DC: Am. Geophys. Union

Harris KDM, Tremayne M, Kariuki BM. 2001. Contemporary advances in the useof powder X-ray diffraction for structure determination. Angew. Chem. Int. Ed.Engl. 40:1626–54

Heinz DL, Jeanloz R. 1984. The equation of state of the gold calibration standard.J. Appl. Phys. 55:885–93

Helmberger DV, Ni S. 2005. Seismic modeling constraints on the south Africansuper plume. See van der Hilst et al. 2005, pp. 63–81

Hernlund JW, Thomas C, Tackley PJ. 2005. A doubling of the postperovskite phaseboundary and structure of the Earth’s lowermost mantle. Nature 434:882–86

Hirose K. 2006. Postperovskite phase transition and its geophysical implications. Rev.Geophys. 44:RG3001

Hirose K, Fei Y, Ono S, Yagi T, Funakoshi K-I. 2001. In situ measurements of thephase transition boundary in Mg3Al2Si3O12: implications for the nature of theseismic discontinuities in the Earth’s mantle. Earth Planet. Sci. Lett. 184:567–73

Hirose K, Karato S-I, Cormier VF, Brodholt JP, Yuen DA. 2006a. Unsolved problemsin the lowermost mantle. Geophys. Res. Lett. 33:L12S01

Hirose K, Kawamura K, Ohishi Y, Tateno S, Sata N. 2005. Stability and equation ofstate of MgGeO3 postperovskite phase. Am. Mineral. 90:262–65

Hirose K, Sinmyo R, Sata N, Ohishi Y. 2006b. Determination of postperovskite phasetransition boundary in MgSiO3 using Au and MgO pressure standards. Geophys.Res. Lett. 33:L01310

Holmes NC, Moriarty JA, Gathers GR, Nellis WJ. 1989. The equation of state ofplatinum to 660 GPa (6.6 Mbar). J. Appl. Phys. 66:2962–67

Horiuchi H, Ito E, Weidner DJ. 1987. Perovskite-type MgSiO3: single-crystal X-raydiffraction study. Am. Mineral. 72:357–60

Iitaka T, Hirose K, Kawamura K, Murakami M. 2004. The elasticity of the MgSiO3

postperovskite phase in the earth’s lowermost mantle. Nature 430:442–45Irifune T, Isshiki M. 1998. Iron partitioning in a pyrolite mantle and the nature of

the 410 km seismic discontinuity. Nature 392:702–5

592 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Irifune T, Nishiyama N, Kuroda K, Inoue T, Isshiki M, et al. 1998. The postspinelphase boundary in Mg2SiO4 determined by in situ X-ray diffraction. Science279:1698–700

Ito E, Takahashi E. 1989. Postspinel transformations in the system Mg2SiO4-Fe2SiO4

and some geophysical implications. J. Geophys. Res. 94:10637–46Jamieson JC, Fritz JN, Manghnani MH. 1982. Pressure measurement at high tem-

perature in X-ray diffraction studies: gold as a primary standard. In High-PressureResearch in Geophysics, ed. S Akimoto, MH Manghnani, pp. 27–48, Tokyo: Cent.Acad. Publ. Jpn.

Jeanloz R, Heinz DL. 1984. Experiments at high temperature and pressure: laserheating through the diamond cell. J. Phys. (Paris) 45:C8-83–92

Jeanloz R, Williams Q. 1998. The core-mantle boundary region. In Ultrahigh-PressureMineralogy: Physics and Chemistry of the Earth’s Deep Interior, ed. RJ Hemley,pp. 241–59. Washington, DC: Mineral. Soc. Am.

Karato S-I. 1989. Plasticity-crystal structure systematics in dense oxides and its im-plications for the creep strength of the Earth’s deep interior: a preliminary result.Phys. Earth Planet. Inter. 55:234–40

Karki BB, Wentzcovitch RM, de Gironcoli S, Baroni S. 2000. Ab initio lattice dy-namics of MgSiO3 perovskite at high pressure. Phys. Rev. B 62:14750–56

Katsura T, Ito E. 1989. The system Mg2SiO4-Fe2SiO4 at high pressures and temper-atures: precise determination of stabilities of olivine, modified spinel, and spinel.J. Geophys. Res. 94:15663–70

Katsura T, Yamada H, Nishikawa O, Song M, Kubo A, et al. 2004. Olivine-wadsleyitetransition in the system (Mg,Fe)2SiO4. J. Geophys. Res. 109:B02209

Kendall J-M, Silver PG. 1996. Constraints from seismic anisotropy on the nature ofthe lowermost mantle. Nature 381:409–12

Kendall J-M, Silver PG. 1998. Investigating causes of D ′′ anisotropy. See Gurniset al. 1998, pp. 97–118

Kieffer SW. 1979. Thermodynamics and lattice vibrations of minerals: 1. Mineralheat capacities and their relationships to simple lattice vibrational models. Rev.Geophys. Space Phys. 17:1–19

Kim YH, Ming LC, Manghnani MH, Ko J. 1991. Phase transformation studies on asynthetic hedenbergite up to 26 GPa at 1200◦C. Phys. Earth Planet. Inter. 17:540–44

Knittle E, Jeanloz R. 1991. Earth’s core-mantle boundary: results of experiments athigh pressures and temperatures. Science 251:1438–43

Kobayashi Y, Kondo T, Ohtani E, Hirao N, Miyajima N, et al. 2005. Fe-Mg parti-tioning between (Mg,Fe)SiO3 postperovskite, perovskite, and magnesiowustitein the Earth’s lower mantle. Geophys. Res. Lett. 32:L19301

Kubo A, Kiefer B, Shen G, Prakapenka VB, Cava RJ, Duffy TS. 2006. Stability andequation of state of the postperovskite phase in MgGeO3 to 2 Mbar. Geophys.Res. Lett. 33:L12S12

Kurashina T, Hirose K, Ono S, Sata N, Ohishi Y. 2004. Phase transition in Al-bearingCaSiO3 perovskite: implications for seismic discontinuities in the lower mantle.Phys. Earth Planet. Inter. 145:67–74

www.annualreviews.org • The Postperovskite Transition 593

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Kuroda K, Irifune T, Nishiyama N, Miyashita M, Funakoshi K, Utsumi W. 2000. De-termination of the phase boundary between ilmenite and perovskite in MgSiO3

by in situ X-ray diffraction and quench experiments. Phys. Chem. Mineral.27:523–32

Lay T, Garnero EJ, Russell SA. 2004. Lateral variation of the D′′ discontinuity be-neath the Cocos plate. Geophys. Res. Lett. 31:L15612

Lay T, Helmberger DV. 1983. A lower mantle S-wave triplication and the velocitystructure of D′′. Geophys. J. R. Astron. Soc. 75:799–837

Lay T, Garnero EJ, Williams Q. 2004. Partial melting in a thermo-chemical boundarylayer at the base of the mantle. Phys. Earth Planet. Inter. 146:441–67

Lay T, Hernlund J, Garnero EJ, Thorne MS. 2006. A postperovskite lens and D′′

heat flux beneath the central pacific. Science 314:1272–76Lay T, Williams Q, Garnero EJ. 1998. The core-mantle boundary layer and deep

Earth dynamics. Nature 392:461–68Lazzeri M, Thilbaudeau P. 2006. Ab initio Raman spectrum of the normal and dis-

ordered MgAl2O4 spinel. Phys. Rev. B 74:140301Lin JF, Struzhkin VV, Jacobsen SD, Hu MY, Chow P, et al. 2005. Spin transition of

iron in magnesiowustite in the Earth’s lower mantle. Nature 436:377–80Liu H-Z, Chen J, Hu J, Martin CD, Weidner DJ, et al. 2005. Octahedral tilting evo-

lution and phase transition in orthorhombic NaMgF3 perovskite under pressure.Geophys. Res. Lett. 32:L04304

Mao H-K, Shen G, Hemley RJ. 1997. Multivariable dependence of Fe-Mg partition-ing in the lower mantle. Science 278:2098–100

Mao WL, Mao H-K, Prakapenka VB, Shu J, Hemley RJ. 2006a. The effect of pressureon the structure and volume of ferromagnesian postperovskite. Geophys. Res. Lett.33:L12S02

Mao WL, Mao H-K, Sturhahn W, Zhao J, Prakapenka VB, et al. 2006b. Iron-richpostperovskite and the origin of ultralow-velocity zones. Science 312:564–65

Mao WL, Meng Y, Shen G, Prakapenka VB, Campbell AJ, et al. 2005. Iron-richsilicates in the Earth’s D′′ layer. Proc. Natl. Acad. Sci. USA 102:9751–53

Mao WL, Shen G, Prakapenka VB, Meng Y, Campbell AJ, et al. 2004. Ferromag-nesian postperovskite silicates in the D′′ layer of the Earth. Proc. Natl. Acad. Sci.USA 101:15867–69

Martin CD, Crichton WA, Liu H, Prakapenka V, Chen J, Parise JB. 2006a. Phasetransitions and compressibility of NaMgF3 (neighborite) in perovskite- andpostperovskite-related structures. Geophys. Res. Lett. 33:L11305

Martin CD, Crichton WA, Liu H, Prakapenka V, Chen J, Parise JB. 2006b. Ri-etveld structure refinement of perovskite and postperovskite phases of NaMgF3

(neighborite) at high pressures. Am. Mineral. 91:1703–6Masters G, Laske G, Bolton H, Dziewonski AM. 2000. The relative behavior of shear

velocity, bulk sound speed, and compressional velocity in the mantle: implica-tions for chemical and thermal structure. In Earth’s Deep Interior: Mineral Physicsand Seismic Tomography from the Atomic to the Global Scale, ed. S-I Karato, AMForte, RC Liebermann, G Masters, L Stixrude, pp. 63–87. Washington, DC:Am. Geophys. Union

594 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Matyska C, Yuen DA. 2005. The importance of radiative heat transfer on superplumesin the lower mantle with the new postperovskite phase change. Earth Planet. Sci.Lett. 234:71–81

Matzel E, Sen MK, Grand SP. 1996. Evidence for anisotropy in the deep mantlebeneath Alaska. Geophys. Res. Lett. 23:2417–20

McCammon C. 1997. Perovskite as a possible sink for ferric iron in the lower mantle.Nature 387:694–96

Merkel S, Kubo A, Miyagi L, Speziale S, Duffy TS, et al. 2006. Plastic deformationof MgGeO3 postperovskite at lower mantle pressures. Science 311:644–46

Merkel S, McNamara AK, Kubo A, Speziale S, Miyagi L, et al. 2007. Deformationof (Mg,Fe)SiO3 postperovskite and D′′ anisotropy. Science 316:1729–32

Miyajima N, Ohgushi K, Ichihara M, Yagi T. 2006. Crystal morphology and dislo-cation microstructures of CaIrO3: a TEM study of an analogue of the MgSiO3

postperovskite phase. Geophys. Res. Lett. 33:L12302Morgan WJ. 1971. Convection plumes in the lower mantle. Nature 230:42–43Morishima H, Kato T, Suto M, Ohtani E, Urakawa S, et al. 1994. The phase boundary

between α-Mg2 SiO4 and β-Mg2SiO4 determined by in-situ X-ray observation.Science 265:1202–3

Murakami M, Hirose K, Kawamura K, Sata N, Ohishi Y. 2004. Post-perovskite phasetransition in MgSiO3. Science 304:855–58

Murakami M, Hirose K, Sata N, Ohishi Y. 2005. Post-perovskite phase transitionand mineral chemistry in the pyrolitic lowermost mantle. Geophys. Res. Lett.32:L03304

Murakami M, Sinogeikin SV, Bass JD, Sata N, Ohishi Y, Hirose K. 2007. Soundvelocity of MgSiO3 postperovskite phase: A constraints on the D ′′ discontinuity.Earth Planet. Sci. Lett. 259:18–23

Nakagawa T, Tackley PJ. 2004. Effects of a perovskite–post perovskite phase changenear core-mantle boundary in compressible mantle convection. Geophys. Res.Lett. 31:L16611

Navrotsky A. 1980. Lower mantle phase transitions may generally have negativepressure-temperature slopes. Geophys. Res. Lett. 7:709–11

Navrotsky A, Schoenitz M, Kojitani H, Xu HW, Zhang JZ, et al. 2003. Alu-minum in magnesium silicate perovskite: formation, structure, and energeticsof magnesium-rich defect solid solutions. J. Geophys. Res. 108:2330

Noel Y, Catti M, D’Arco P, Dovesi R. 2006. The vibrational frequencies of forsteriteMg2SiO4: an all-electron ab initio study with the CRYSTAL code. Phys. Chem.Mineral. 33:383–93

Oganov AR, Martonak R, Laio A, Raiteri P, Parrinello M. 2005. Anisotropy ofEarth’s D ′′ layer and stacking faults in the MgSiO3 postperovskite phase. Nature438:1142–44

Oganov AR, Ono S. 2004. Theoretical and experimental evidence for a postperovskitephase of MgSiO3 in Earth’s D′′ layer. Nature 430:445–48

Oganov AR, Ono S. 2005. The high-pressure phase of alumina and implications forEarth’s D′′ layer. Proc. Natl. Acad. Sci. USA 102:10828–31

Ono S, Funakoshi K, Ohishi Y, Takahashi E. 2005. In situ X-ray observation of thephase transformation of Fe2O3. J. Phys. Condens. Matter 17:269–76

www.annualreviews.org • The Postperovskite Transition 595

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Ono S, Kikegawa T, Ohishi Y. 2004. High-pressure phase transition of hematite,Fe2O3. J. Phys. Chem. Solids 65:1527–30

Ono S, Kikegawa T, Ohishi Y. 2006a. Equation of state of CaIrO3-type MgSiO3 upto 144 GPa. Am. Mineral. 91:475–78

Ono S, Oganov AR. 2005. In situ observations of phase transition between perovskiteand CaIrO3-type phase in MgSiO3 and pyrolitic mantle composition. EarthPlanet. Sci. Lett. 236:914–32

Ono S, Oganov AR, Koyama T, Shimizu H. 2006b. Stability and compressibility ofthe high-pressure phases of Al2O3 up to 200 GPa: implications for the electricalconductivity of the base of the lower mantle. Earth Planet. Sci. Lett. 246:326–35

Ono S, Ohishi Y. 2005. In situ X-ray observation of phase transformation in Fe2O3

at high pressures and high temperatures. J. Phys. Chem. Solids 66:1714–20Panning M, Romanowicz B. 2004. Inferences on flow at the base of Earth’s mantle

based on seismic anisotropy. Science 303:351–53Revenaugh J, Meyer R. 1997. Seismic evidence of partial melt within a possibly

ubiquitous low-velocity layer at the base of the mantle. Science 277:670–73Ringwood AE. 1975. Composition and Petrology of the Earth’s Mantle. New York:

McGraw-HillRozenberg GK, Dubrovinsky LS, Pasternak MP, Naaman O, Bihan TL, Ahuja R.

2002. High-pressure structural studies of hematite Fe2O3. Phys. Rev. B 65:64112Santillan J, Shim S-H, Shen G, Prakapenka VB. 2006. High-pressure phase transition

in bixbyite: application for the crystal structure and properties of the CaIrO3-type. Geophys. Res. Lett. 33:L15307

Shen G, Rivers ML, Wang Y, Sutton SR. 2001. Laser heated diamond cell system atthe Advanced Photon Source for in situ X-ray measurements at high pressureand temperature. Rev. Sci. Instrum. 72:1273–82

Shieh SR, Duffy TS, Kubo A, Shen G, Prakapenka VB, et al. 2006. Equation of stateof the postperovskite phase synthesized from a natural (Mg,Fe)SiO3 orthopy-roxene. Proc. Natl. Acad. Sci. USA 103:3039–43

Shim S-H. 2005. Stability of MgSiO3 perovskite in the lower mantle. See van derHilst et al. 2005, pp. 261–82

Shim S-H, Duffy TS. 2000. Constraints on the P-V-T equation of state of MgSiO3

perovskite. Am. Mineral. 85:354–63Shim S-H, Duffy TS, Jeanloz R, Shen G. 2004. Stability and crystal structure of

MgSiO3 perovskite to the core-mantle boundary. Geophys. Res. Lett. 31:L10603Shim S-H, Duffy TS, Kenichi T. 2002. Equation of state of gold and its application

to the phase boundaries near the 660-km depth in the mantle. Earth Planet. Sci.Lett. 203:729–39

Shim S-H, Duffy TS, Shen G. 2001. The postspinel transformation in Mg2SiO4 andits relation to the 660-km seismic discontinuity. Nature 411:571–74

Shim S-H, Kubo A, Duffy TS. 2007. Raman spectroscopy of perovskite and postper-ovskite of MgGeO3 to 123 GPa. Earth Planet. Sci. Lett. 260:166–78

Sidorin I, Gurnis M, Helmberger DV. 1999. Evidence for a ubiquitous seismic dis-continuity at the base of the mantle. Science 286:1326–31

Sinmyo R, Hirose K, O’Neill HSC, Okunishi E. 2006. Ferric iron in Al-bearingpostperovskite. Geophys. Res. Lett. 33:L12S13

596 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Speziale S, Zha C-S, Duffy TS, Hemley RJ, Mao H-K. 2001. Quasi-hydrostatic com-pression of magnesium oxide to 52 GPa: implications for the pressure-volume-temperature equations of state. J. Geophys. Res. 106:515–28

Speziale S, Milner A, Lee VE, Clark SM, Pasternak MP, Jeanloz R. 2005. Iron spintransition in Earth’s mantle. Proc. Natl. Acad. Sci. USA 102:17918–22

Stackhouse S, Brodholt JP, Price GD. 2005a. High temperature elastic anisotropyof the perovskite and postperovskite polymorphs of Al2O3. Geophys. Res. Lett.32:L13305

Stackhouse S, Brodholt JP, Dobson DP, Price GD. 2006a. Electronic spin transitionsand the seismic properties of ferrous iron-bearing MgSiO3 post-perovskite. Geo-phys. Res. Lett. 33:L12S03

Stackhouse S, Brodholt JP, Price GD. 2006b. Elastic anisotropy of FeSiO3 end-members of the perovskite and postperovskite phases. Geophys. Res. Lett.33:L01304

Stackhouse S, Brodholt JP, Wookey J, Kendall J-M, Price GD. 2005b. The effectof temperature on the seismic anisotropy of the perovskite and postperovskitepolymorphs of MgSiO3. Earth Planet. Sci. Lett. 230:1–10

Su W-J, Dziewonski AM. 1997. Simultaneous inversion for 3-D variations in shearand bulk velocity in the mantle. Phys. Earth Planet. Inter. 100:135–56

Suzuki A, Ohtani E, Morishima H, Kubo T, Kanbe Y, et al. 2000. In situ determi-nation of the phase boundary between wadsleyite and ringwoodite in Mg2SiO4.Geophys. Res. Lett. 27:803–6

Tackley PJ, Stevenson DJ, Glatzmaier GA, Schubert G. 1993. Effects of an endother-mic phase transition at 670-km depth in a spherical model of convection in theEarth’s mantle. Nature 361:699–704

Takafuji N, Hirose K, Mitome M, Bando Y. 2005. Solubilities of O and Si in liquidiron equilibrium with (Mg,Fe)SiO3 perovskite and the light elements in the core.Geophys. Res. Lett. 32:L06313

Tateno S, Hirose K, Sata N, Ohishi Y. 2005. Phase relations in Mg3Al2Si3O12 to 180GPa: effect of Al on postperovskite phase transition. Geophys. Res. Lett. 32:L15306

Tateno S, Hirose K, Sata N, Ohishi Y. 2007. Solubility of FeO in (Mg,Fe)SiO3 per-ovskite and the postperovskite phase transition. Phys. Earth Planet. Inter. 160:318–25

Trampert J, Deschamps F, Resovsky J, Yuen D. 2004. Probabilistic tomography mapschemical heterogeneities throughout the lower mantle. Science 306:853–56

Tsuchiya J, Tsuchiya T, Wentzcovitch RM. 2005a. Transition from the Rh2O3(II)-to-CaIrO3 structure and the high-pressure-temperature phase diagram of alumina.Phys. Rev. B 72:020103

Tsuchiya J, Tsuchiya T, Wentzcovitch RM. 2005b. Vibrational and thermodynamicproperties of MgSiO3 postperovskite. J. Geophys. Res. 110:B02204

Tsuchiya T. 2003. First-principles prediction of the P-V-T equation of state of goldand the 660-km discontinuity in Earth’s mantle. J. Geophys. Res. 108:2462

Tsuchiya T, Tsuchiya J. 2006. Effect of impurity on the elasticity of perovskiteand postperovskite: velocity contrast across the postperovskite transition in(Mg,Fe,Al)(Si,Al)O3. Geophys. Res. Lett. 33:L12S04

www.annualreviews.org • The Postperovskite Transition 597

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Tsuchiya T, Tsuchiya J, Umemoto K, Wentzcovitch RM. 2004a. Elasticity of post-perovskite MgSiO3. Geophys. Res. Lett. 31:L14603

Tsuchiya T, Tsuchiya J, Umemoto K, Wentzcovitch RM. 2004b. Phase transition inMgSiO3 perovskite in the earth’s lower mantle. Earth Planet. Sci. Lett. 224:241–48

van der Hilst RD, Bass J, Matas J, Trampert J, eds. 2005. Earth’s Deep Mantle: Structure,Composition, and Evolution. Washington, DC: Am. Geophys. Union

van der Hilst RD, de Hoop MV, Wang P, Shim S-H, Ma P, Tenorio L. 2007. Seismo-stratigraphy and thermal structure of Earth’s core-mantle boundary region.Science 315:1813–17

Van der Voo R, Spakman W, Bijwaard H. 1999. Mesozoic subducted slabs underSiberia. Nature 397:246–49

Vassiliou MS, Ahrens TJ. 1981. Hugoniot equation of state of periclase to 200 GPa.Geophys. Res. Lett. 8:729–32

Walte N, Heidelbach F, Miyajima N, Frost D. 2007. Texture development and TEManalysis of deformed CaIrO3: implications for the D′′ layer at the core-mantleboundary. Geophys. Res. Lett. 34:L08306

Watanuki T, Shimomura O, Yagi T, Kondo T, Isshiki M. 2001. Contruction of laser-heated diamond anvil cell system for in situ X-ray diffraction study at SPring-8.Rev. Sci. Instrum. 72:1289–92

Wentzcovitch RM, Karki BB, Karato S, Da Silva CRS. 1998. High pressure elasticanisotropy of MgSiO3 perovskite and geophysical implications. Earth Planet. Sci.Lett. 164:371–78

Wentzcovitch RM, Stixrude L, Karki BB, Kiefer B. 2004. Akimotoite to perovskitephase transition in MgSiO3. Geophys. Res. Lett. 31:L10611

Wentzcovitch RM, Tsuchiya T, Tsuchiya J. 2006. MgSiO3 postperovskite at D′′ con-ditions. Proc. Natl. Acad. Sci. USA 103:543–46

Williams Q, Revenaugh J, Garnero E. 1998. A correlation between ultralow basalvelocities in the mantle and hot spots. Science 280:546–49

Wookey J, Stackhouse S, Kendall J-M, Brodholt J, Price GD. 2005. Efficacy of thepostperovskite phase as an explanation for lowermost-mantle seismic properties.Nature 438:1004–7

Woodward PM. 1997. Octahedral tilting in perovskite. I. Geometrical consideration.Acta Crystallogr. B53:32–43

Wysession ME, Lay T, Revenaugh J, Williams Q, Garnero EJ, et al. 1998. The D′′

discontinuity and its implications. See Gurnis et al. 1998, pp. 273–97Yamazaki D, Karato S-I. 2002. Fabric development in (Mg,Fe)O during large strain,

shear deformation: implications for seismic anisotropy in Earth’s lower mantle.Phys. Earth Planet. Inter. 131:251–67

Yamazaki D, Yoshino T, Ohfuji H, Ando J-I, Yoneda A. 2006. Origin of seismicanisotropy in the D′′ layer inferred from shear deformation experiments onpostperovskite phase. Earth Planet. Sci. Lett. 252:372–78

Young RA, ed. 1993. The Rietveld Method. New York: Oxford Univ. PressYu YG, Wentzcovitch RM, Tsuchiya T, Umemoto K, Weidner DJ. 2007. First prin-

ciples investigation of the postspinel transition in Mg2SiO4. Geophys. Res. Lett.34:L10306

598 Shim

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-EA36-18 ARI 26 March 2008 1:58

Zha CS, Mao H-K, Hemley RJ. 2000. Elasticity of MgO and a primary pressure scaleto 55 GPa. Proc. Natl. Acad. Sci. USA 97:13494–99

Zhang F, Oganov AR. 2006a. Mechanisms of Al3+ incorporation in MgSiO3 post-perovskite at high pressures. Earth Planet. Sci. Lett. 248:54–61

Zhang F, Oganov AR. 2006b. Valence state and spin transitions of iron in Earth’smantle silicates. Earth Planet. Sci. Lett. 249:436–43

Zhang J, Weidner DJ. 1999. Thermal equation of state of aluminum-enriched silicateperovskite. Science 284:782–84

www.annualreviews.org • The Postperovskite Transition 599

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-FM ARI 29 March 2008 2:32

Annual Reviewof Earth andPlanetary SciencesVolume 36, 2008Contents

FrontispieceMargaret Galland Kivelson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �xii

The Rest of the Solar SystemMargaret Galland Kivelson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �1

Abrupt Climate Changes: How Freshening of the Northern AtlanticAffects the Thermohaline and Wind-Driven Oceanic CirculationsMarcelo Barreiro, Alexey Fedorov, Ronald Pacanowski, and S. George Philander � � � � 33

Geodynamic Significance of Seismic Anisotropy of the Upper Mantle:New Insights from Laboratory StudiesShun-ichiro Karato, Haemyeong Jung, Ikuo Katayama, and Philip Skemer � � � � � � � � � � 59

The History and Nature of Wind Erosion in DesertsAndrew S. Goudie � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 97

Groundwater Age and Groundwater Age DatingCraig M. Bethke and Thomas M. Johnson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �121

Diffusion in Solid Silicates: A Tool to Track Timescales of ProcessesComes of AgeSumit Chakraborty � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �153

Spacecraft Observations of the Martian AtmosphereMichael D. Smith � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �191

Crinoid Ecological MorphologyTomasz K. Baumiller � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �221

Oceanic Euxinia in Earth History: Causes and ConsequencesKatja M. Meyer and Lee R. Kump � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �251

The Basement of the Central Andes: The Arequipaand Related TerranesVictor A. Ramos � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �289

Modeling the Dynamics of Subducting SlabsMagali I. Billen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �325

vii

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.

ANRV341-FM ARI 29 March 2008 2:32

Geology and Evolution of the Southern Dead Sea Fault with Emphasison Subsurface StructureZvi Ben-Avraham, Zvi Garfunkel, and Michael Lazar � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �357

The Redox State of Earth’s MantleDaniel J. Frost and Catherine A. McCammon � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �389

The Seismic Structure and Dynamics of the Mantle WedgeDouglas A. Wiens, James A. Conder, and Ulrich H. Faul � � � � � � � � � � � � � � � � � � � � � � � � � � � � �421

The Iron Isotope Fingerprints of Redox and Biogeochemical Cyclingin the Modern and Ancient EarthClark M. Johnson, Brian L. Beard, and Eric E. Roden � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �457

The Cordilleran Ribbon Continent of North AmericaStephen T. Johnston � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �495

Rheology of the Lower Crust and Upper Mantle: Evidencefrom Rock Mechanics, Geodesy, and Field ObservationsRoland Burgmann and Georg Dresen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �531

The Postperovskite TransitionSang-Heon Shim � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �569

Coastal Impacts Due to Sea-Level RiseDuncan M. FitzGerald, Michael S. Fenster, Britt A. Argow,and Ilya V. Buynevich � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �601

Indexes

Cumulative Index of Contributing Authors, Volumes 26–36 � � � � � � � � � � � � � � � � � � � � � � � �649

Cumulative Index of Chapter Titles, Volumes 26–36 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �653

Errata

An online log of corrections to Annual Review of Earth and Planetary Sciences articlesmay be found at http://earth.annualreviews.org

viii Contents

Ann

u. R

ev. E

arth

Pla

net.

Sci.

2008

.36:

569-

599.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

MA

SSA

CH

USE

TT

S IN

STIT

UT

E O

F T

EC

HN

OL

OG

Y o

n 05

/01/

08. F

or p

erso

nal u

se o

nly.