Wendy L. MaoLos Alamos National Laboratory

Elasticity of iron-rich silicate in Earth’s D” layer

U Chicago

Andrew Campbell (now at U Maryland)

Dion Heinz

Geophysical Lab

Yingwei Fei

Russell Hemley

Ho-kwang Mao

Jinfu Shu

Argonne National Lab

Yue Meng

Vitali Prakapenka

Guoyin Shen

Wolfgang Sturhahn

Jiyong Zhao

SUNY-Stony Brook

Donald Lindsley

• New phase in MgSiO3 at approximately D” discontinuity conditions– pv (Pbnm) � ppv (Cmcm)

Tsuchiya et al, EPSL 2004

pv ppv

Post-perovskite

Murakami et al, Science 2004

Post-perovskite

Tsuchiya et al, EPSL 2004

Sidorin et al, Science 1999

Mantle adiabat

Duffy, Nature 2004

What about Fe?

• D” is where the liquid Fe outer core meets the crystalline silicate mantle– Fe-rich ppv

– ULVZ

– Seismic anisotropy

Table

Culet

Gasket Sample chamber

Force

Table

Culet

Gasket Sample chamber

Force

Diamond Anvil Cell

• Samples– Orthopyroxenes: Fs10 (Mg0.9Fe0.1)SiO3,

Fs20, Fs40, Fs60, Fs80

• Thermal insulation– NaCl, SiO2

• Internal pressure standard– Pt, NaCl

• Gasket– Re, Be + graphite insert

Experimental

gasket sample

Pt

diamond anvil

laser

2θ

x-ray beam

Synchrotron XRD and laser-heated DAC

13-IDD, GSECARS

Image plate

16-IDB, HPCAT

1800 K

Fs40

130 GPa

• Gasket hole ~ 60 µm (culet =150 µm, bevel diameter = 300 µm) • X-ray beam ~ 6 x 7 µm• Double-sided Nd:YLF laser heating

W. M

ao et a

l,PNAS 2004

Fs20, M

g0.8 Fe0.2 S

iO3

Azimuth (º)

Diffra

ctio

n a

ng

le (2

θ)

0

10

0

20

0

30

0

68

10

41

21

41

6

ppv 113

ppv 020

ppv 002

ppv 022

ppv 023ppv 131

ppv 004ppv 043

ppv 152,062,200ppv 044

ppv 132

diamond spotsbeam stop holder

56

78

910

1112

1314

1516

1718

1920

2-theta

Relative intensity

NaCl

ppv 020

ppv 002

ppv 113ppv 004

ppv 152,062,200ppv 044

ppv 022

ppv 023ppv 131

ppv 132

Re

Re

NaCl

Pt

En80 138 G

Pa, quenched from

2400 K

NaCl

NaCl

Pt

56

78

910

1112

1314

1516

1718

1920

2-theta

Relative intensity

NaCl

ppv 020

ppv 002

ppv 113ppv 004

ppv 152,062,200ppv 044

ppv 022

ppv 023ppv 131

ppv 132

Re

Re

NaCl

Pt

En80 138 G

Pa, quenched from

2400 K

NaCl

NaCl

Pt

pv+ppv

ppvonly

En80, 113 G

Pa, quenched from

2000 KE

n80, 147 GP

a, quenched from 2400 K

ppv silicate can take a lot of Fe

Rel

ativ

e in

tens

ity

d-spacing (Å)

Diffraction angle (2θ)

4 3.5 3 2.5 2 1.5 1.3

6 8 10 12 14 16 18 20

Fs40, 141 GPaquenched from 1800 K

b)

Re R

e

ppv

022

ppv

131

ppv

132

ppv

113

ppv

004

*

* **** pp

v15

2,06

2,20

0pp

v04

4

**

ppv

023†

†

†

†

Re

3.5 3 2.5 2 1.5 1.3

6 8 10 12 14 16 18

c)

Fs60, 124 GPaquenched from 1800 K

Re

Re

Pt

Pt

ppv

022

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

*

** * **

ppv

152,

062,

200

ppv

044

**

Re

Au

Fs20, 147 GPaquenched from 2400 K

a)

4 3.5 3 2.5 2 1.5 1.3

NaC

l

δ02

0 ppv

002

*

*pp

v02

2

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

***

*

**

ppv

152,

062,

200

ppv

044

* *

NaC

l

Re R

eP

t NaC

l

5 7 9 11 13 15 17 19

ppv

110

*

d)

Fs80, 142 GPaquenched from 2000 K

6 8 10 12 14 16 18

3.5 3 2.5 2 1.5 1.3

Re

Re

Pt

Pt

Pt

Au

ppv

132

ppv

113

ppv

004

ppv

131

ppv

023

ppv

022

** **

*

*

Rel

ativ

e in

tens

ity

d-spacing (Å)

Diffraction angle (2θ)

4 3.5 3 2.5 2 1.5 1.3

6 8 10 12 14 16 18 20

Fs40, 141 GPaquenched from 1800 K

b)

Re R

e

ppv

022

ppv

131

ppv

132

ppv

113

ppv

004

*

* **** pp

v15

2,06

2,20

0pp

v04

4

**

ppv

023†

†

†

†

Re

3.5 3 2.5 2 1.5 1.3

6 8 10 12 14 16 18

c)

Fs60, 124 GPaquenched from 1800 K

Re

Re

Pt

Pt

ppv

022

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

*

** * **

ppv

152,

062,

200

ppv

044

**

Re

Au

Fs20, 147 GPaquenched from 2400 K

a)

4 3.5 3 2.5 2 1.5 1.3

NaC

l

δ02

0 ppv

002

*

*pp

v02

2

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

***

*

**

ppv

152,

062,

200

ppv

044

* *

NaC

l

Re R

eP

t NaC

l

5 7 9 11 13 15 17 19

ppv

110

*

d)

Fs80, 142 GPaquenched from 2000 K

6 8 10 12 14 16 18

3.5 3 2.5 2 1.5 1.3

Re

Re

Pt

Pt

Pt

Au

ppv

132

ppv

113

ppv

004

ppv

131

ppv

023

ppv

022

** **

*

* ppv only

W. Mao et al, PNAS 2005

Fe-Mg partitioning: unsettled issue

Murakami et al, GRL 2005

Kobayashi et al, GRL 2006

Andrault et al, JGR 2001

Mao et al, Science 1997

Kessen et al, EPSL 2002 +

mw

silicate

MgFe

MgFeK

)/(

)/(=

4 6 8 10 12 14 16

ppv

020

ppv

132

ppv

002

ppv

022 pp

v02

3

ppv

131

ppv

113

ppv

004

ppv

152,

062,

200

ppv

044

mw

111

mw

200

mw

220

** **

****

ppv

110

ppv

111

ppv

114

ppv

133

*

***

**

Rel

ativ

e in

tens

ity

Diffraction angle (2θ)

Re

Fa30

4 6 8 10 12 14 16

ppv

020

ppv

132

ppv

002

ppv

022 pp

v02

3

ppv

131

ppv

113

ppv

004

ppv

152,

062,

200

ppv

044

mw

111

mw

200

mw

220

** **

****

ppv

110

ppv

111

ppv

114

ppv

133

*

***

**

Rel

ativ

e in

tens

ity

Diffraction angle (2θ)

Re

Fa30

• Starting composition: Fa30 (Fe0.3Mg0.7)2SiO4, Fa45

• After decompression, recovered mw had lattice constant, a0 = 4.2336 Åwhich corresponds to a composition: Fe0.23Mg0.77O

• For mass balance the ppv phase has a composition: Fe0.37Mg0.63SiO3

• Fe partitions strongly into ppvphase?

120 GPa

Fe-Mg partitioning: unsettled issue

How does Fe affect VS?

222

3

4

)(

Φ=−=

∂∂−=

υυυρ SP

KV

PVK

7 8 9 10 11 12 13 14 15 16 17 182-theta

Rel

ativ

e co

un

ts

d-spacing (Å)

2.7 2.3 2 1.8 1.6 1.5 1.4 1.3 1.2 1.1

Diffraction angle (2θ)

Rel

ativ

e in

tens

ity

ppv

022

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

ppv

152,

062

ppv

044

ppv

110

ppv

200

ppv

025

ppv

222

ppv

111

ppv

040

ppv

042

ppv

133

ppv

151

ppv

114

Pt

Pt

PtPt

ppv

041

7 8 9 10 11 12 13 14 15 16 17 182-theta

Rel

ativ

e co

un

ts

d-spacing (Å)

2.7 2.3 2 1.8 1.6 1.5 1.4 1.3 1.2 1.1

Diffraction angle (2θ)

Rel

ativ

e in

tens

ity

ppv

022

ppv

023

ppv

131

ppv

132

ppv

113

ppv

004

ppv

152,

062

ppv

044

ppv

110

ppv

200

ppv

025

ppv

222

ppv

111

ppv

040

ppv

042

ppv

133

ppv

151

ppv

114

Pt

Pt

PtPt

ppv

041

W. Mao et al, Science 2006

Fs40 ppv, 140 GPa

116

118

120

122

124

126

120 130 140 150 160 170Pressure (GPa)

Vo

lum

e (A

3)V

olum

e (Å

3 )

Pressure (GPa)

Fs40, ppv

116

118

120

122

124

126

120 130 140 150 160 170Pressure (GPa)

Vo

lum

e (A

3)V

olum

e (Å

3 )

Pressure (GPa)

Fs40, ppv

Nuclear resonant x-ray spectroscopy (NRXS)

APD

APD HRM M

KBmirrors

x-rays

DAC 150nsec

NRIXS signal

SMS signal

nonmagnetic

magnetic

time

Energy

Inte

nsity

Anti-stokes Stokes

Elastic line

Synchrotron 57Fe Mössbauer spectroscopy

XOR, Sector 3, APS, ANLXOR, Sector 3, APS, ANL

Nuclear resonant inelastic x-ray spectroscopy

Phonon density of states

ω

g(ω

) de

nsity

of s

tate

s Debye Model

“Real” Crystal

2322

)( ωυπ

ω Vg =

333

213

SpD υυυ+=

0 20 40 60 80 100 120

Fs40, ppv

Energy (meV)

Pho

non

DO

S, g

(E)

0 20 40 60 80 100 120

Fs40, ppv

Energy (meV)

Pho

non

DO

S, g

(E)

D

VP, km/sec VS, km/sec ν PREM, mantle side of CMB 13.72 7.26 0.30 ULVZ (Thorne, JGR 2004) 12.35 5.08 0.40 Fs40 ppv at 130 GPa-300 K 12.72 4.86 0.41 Fs40 ppv at 130 GPa-3000 K 11.91 4.05 0.43

W. Mao et al, Science 2006

Fe in ULVZ

• Fe-rich ppv has low enough Vs to explain the depression of velocity in ULVZ

• When Fe-poor mantle contacts core, a thin layer of Fe-rich ppvforms

• Mantle convection sweeps the thin layer into a thickened pile to form ULVZ

Outer Core

Fe

D’’Fe-poor ppv

Mantle Upwelling

ULVZFe-rich ppv

Radial x-ray diffraction

Magnitude of waviness-- deviatoric strain & shear strength

Magnitude dependence on hkl-- elasticity tensor

Intensity vs azimuthal angle -- lattice preferred orientation

Deformation of a germanate analog

In contrast with phenomenological considerations suggesting (010) as a slip plane, lattice planes near (100) became aligned perpendicular to the compression direction, suggesting that slip on (100) or (110) dominated plastic deformation.

Elastic anisotropy of silicate ppv at 140 GPa

2θ (deg.)

Azi

mut

hala

ngle

, η(d

eg.)

-150

-100

-50

0

50

100

150

9 10 11 12 13 14 15 16

diamond spot

022

113

004

132

131

023,

130

Pt

040

110

Be

100

Be

002

Be

101

042

022

113

004

132

131

023,

130

Pt

040

110

Be

100

Be

002

Be

101

042

-150

-100

-50

0

50

100

150

9 10 11 12 13 14 15 16

W. Mao et al, in prep.

Fs40

Determination of single-crystal elasticity tensor

1-3cos2ψ

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

Dev

iato

ricst

rain

(ε ψ

)

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

0.004052.441

0.003222.355

0.004272.032

022 110 040

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

0.003411.821

0.003251.744

0.004291.691

023 131 042

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

1-3cosy2

0.003941.540

0.003391.565

0.003371.527

132 113 004

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

-2 -1 0 1

G(hkl)R-1 = (3H4 - H 2) s11 + (3K4 - K 2) s22 + (3L4 - L 2) s33 + (6H2K2 + L2 - 1) s12 +

(6H2L2 + K2 - 1) s13 + (6K2L2 + H2 – 1) s23 + 3K2L2 s44+ 3H2L2 s55+ 3H2K2 s66

H = hd/a, K = kd/b, L = ld/c

)()cos31()(

)()()( 2 hklQ

hkld

hkldhkldhkl

P

P ⋅−=−

= ψε ψψ

[Q(hkl)/<Q>] G-1 = α G(hkl)R-1 + (1-α) G-1

αψ

2θ

η

A

q

X

0°

cos ψ = cos η ⋅ cos θ ⋅ sin α – sin θ ⋅ cos α

Formulism from Singh et al J. Appl. Phys., 1998

Linear compressibilities of ppv at 140 GPa

χa = s11 + s12 + s13

χb = s22 + s12 + s23

χc = s33 + s13 + s23

6.1

6.12

6.14

6.16

6.18

6.2

6.22

6.24

95 105 115 125 135 145

Pressure (GPa)

Latti

ce p

aram

eter

(Å

)

8.12

8.16

8.2

8.24

8.28

8.32

95 105 115 125 135 145

2.45

2.46

2.47

2.48

2.49

2.5

2.51

95 105 115 125 135 145

a

b

c6.1

6.12

6.14

6.16

6.18

6.2

6.22

6.24

95 105 115 125 135 145

Pressure (GPa)

Latti

ce p

aram

eter

(Å

)

8.12

8.16

8.2

8.24

8.28

8.32

95 105 115 125 135 145

2.45

2.46

2.47

2.48

2.49

2.5

2.51

95 105 115 125 135 145

a

b

c

c11 c22 c33 c12 c13 c23 c44 c55 c66

1129 1011 1119 814 722 805 97 124 229

In GPa

W. Mao et al, in prep.

Iitaka et al, Nature 2004

Tsuchiya et al, GRL 2004

MgSiO3 ppv

Theoretically determined velocity anisotropy

Stackhouse et al, EPSL 2005

2

4

6

8

10

12

14

0 90 180 270

VS1

VS2

VP

Vel

ocity

(km

/sec

)

Propagation direction

[100] [100][010] [001]

a b c a

2

4

6

8

10

12

14

0 90 180 270

VS1

VS2

VP

Vel

ocity

(km

/sec

)

Propagation direction

[100] [100][010] [001]

a b c a

Fs40ppv

W. Mao et al, in prep.

• Very large azimuthal VS anisotropy, (VS,max - VS,min)/VS,aggregate = 44%

• Small azimuthal VP anisotropy, (VP,max - VP,min)/VP,aggregate = 8%

This implies:

• Seismically observable VS splitting even with the 100 slip plane and low degree of LPO

• No seismically observable VPanisotropy

Possible implications

• Clapeyron slope of pv-ppv– Topography of the top of D”

– Double crossing

• Melting of ppv– CMB Temperature

• FTIR and visible spectroscopy– Radiative heat transfer in D”

• Brillouin spectroscopy – Elasticity of low-Fe ppv

• Element partitioning among pv, ppv, mw, and Fe– Geochemistry of D”

Current/Future work in D”

Fs40ppv

W. Mao et al, in prep.

Experimentally determined velocity anisotropy

c11 c22 c33 c12 c13 c23 c44 c55 c66

1129 1011 1119 814 722 805 97 124 229

Stackhouse et al, GRL 2006

Stackhouse et al, EPSL 2005