searching for earth’s lost oceans: an odyssey in mineral...

Searching for Earth’s Lost Oceans:

An Odyssey in Mineral Physics

March 16, 2004

Searching for Earth’s Lost Oceans:

An Odyssey in Mineral Physics

J. R. Smyth

with (lots of) help from C. M. Holl, H. Spetzler(University of Colorado, Boulder)

S. D. Jacobsen (Carnegie)D. J. Frost,, F. Langenhorst, and C. A.

McCammon (Bayerisches Geoinstitut, Bayreuth)M. Manghnani and G. Amulele

(University of Hawaii)

Oceans cover 71% of the planet’s surface.

71% of the surface

But only 0.025% of the mass

H-cycling in the Deep Earth

• There has been running water (implying oceans) as far back as we can see in geologic time.

• Water controls the geology and biology of the planet.

• Are oceans the surface expression of a deeper reservoir?


• Water fluxes melting at ridges.– MORBs contain 0.1 to 0.3 wt % H2O– OIBs contain 0.6 to 1.5 wt % H2O

• Plates hydrate on ocean floor.• Plates dehydrate on subduction.

– Completely?


• 0.10 percent H2O by weight in the top 10 km of the descending slab is sufficient to recycle the oceans once over 4.5 billion years at current subduction rates.

Review Mantle Mineralogy

Shear Wave Velocity Structure

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8

300 350 400 450 500 550 600 650

Depth (km)

Shea

r Wav

e Ve

loci

ty (k

m/s

)

Dry Wadsleyite and Ringwooditeare too fast to be consistent with PREM model for olivine-rich upper mantle.(Duffy & Anderson, 1988)

H-cycling: Role of Nominally Anhydrous

Phases• Synthesis Experiments

– Olivine, Wadsleyite, Wads II, Ringwoodite

• Effects of H on Transition Depths• Effects of H on P and S velocities• Isotope Effects• Natural Samples

– Eclogites and Peridotites

Laboratory Odyssey: Find the water

HydrousSynthesis

Experiments: 5000 ton

Multi-anvil Press,

BGIBayreuth

Crystals are large enough

for single crystal X-ray and neutron diffraction and GHZ

ultrasonicsin DAC.

Multi-anvil Press Sample Assembly

H-Cycling: Synthesis Experiments

• Synthesis (Bayreuth)• Characterization

– XRD single crystal (Boulder, Bayreuth)– TEM (Bayreuth)– IR / Raman (Bayreuth, Caltech, Oxford)– Mössbauer (Bayreuth)

• Property Measurements– Isothermal Compression (Boulder, APS)– Ultrasonic Vp and Vs (Bayreuth)

H-Cycling: Synthesis Experiments

• Olivine• Wadsleyite• Wadsleyite II• Ringwoodite

• Synthesis• Characterization

– XRD single crystal– TEM– IR / Raman– Mössbauer

• Property Measurements– Isothermal Compression– Ultrasonic Vp and Vs

Hydration of Olivine

• Olivine accepts up to 0.2 wt % H2O at 12GPa (Kohlstedt et al, 1996).

• Maybe up to 8000ppm in low silica runs• Hydration causes measurable volume

expansion• Hydration appears to involve divalent

cation vacancies principally at M2.• H becomes compatible at P > 10 GPa• Hydration may cause decreased bulk

modulus and seismic velocities.• H partitions strongly to wadsleyite.

Hydrous Wadsleyite

• 1.6[MgO]•0.2[Mg(OH)2] •1[SiO2]

• Up to 0.4 H pfu• H on non-silicate

oxygens• Si/(Mg+Fe) = 0.5• ρ = 3.36 g/cm3

• 3.0% H2O by weight• 10 % H2O by volume

Wadsleyite II (Spinelloid IV)

• 1.6[MgO]•0.2[Mg(OH)2]•1[SiO2]• May occur between wadsleyite and

ringwoodite (17.5 GPa @ 1400 ºC)• May require Fe3+, Cr or Al• May obscure 520km discontinuity

Wadsleyite II = Spinelloid IV

• Two new structure refinements• Calculated powder patterns• Polarized oriented FTIR + Visible• Raman• Mössbauer • TEM + Fe and Si ELNES• Isothermal bulk modulus

Wadsleyite II : New Data

• Wadsleyite II is Spinelloid IV• Wadsleyite II is a stable phase in TZ

– May require > 2% H2O by weight– May require > 1% trivalent cations (Fe,

Cr, Al)

• Wadsleyite II is difficult to recognize– XRPD, FTIR, Raman

• Wadsleyite II may obscure 525 km

Ringwoodite

• (γ-Mg1.63Fe0.22

H0.4 Si0.95O4)• ~10 % of Fe

present as ferric (Mössbauer)

• Dark blue color

High Pressure Single-Crystal X-ray Diffraction

Ringwoodite with Quartz @ Ringwoodite with Quartz @ 11 11 GpaGpa

Single-Crystal Compression to 11 GPa

Ringwoodite cell volume vs pressure

500

505

510

515

520

525

530

535

0 2 4 6 8 10 12

Pressure (GPa)

Unit

cell

volu

me

(Å^3

)

Ringwoodite Isothermal Compression to 11 GPa

(Single-Crystal)• Fo90 with ~8,000 ppm H2O• Third-order B-M fit

• KT = 169.0 ± 3.4 GPa• K’ = 7.9 ± 0.9

• One Percent H2O in Ringwoodite– = 600ºC on KT (at low P)

High Pressure Powder Diffraction at GSECARS

Compression to 45 GPaHydrous Ringwoodite Compression

440

450

460

470

480

490

500

510

520

530

540

0 5 10 15 20 25 30 35 40 45 50

Pressure (GPa)

Uni

t Cel

l Vol

ume

(Å^3

)

Ringwoodite Isothermal Compression to 45 GPa

• Fo90 with ~8,000 ppm H2O– Fourth-order B-M fit

– Kt = 166.6 ± 3.8 GPa– K’ = 7.6 ± 1.3– K” = -0.11 ± 0.11

• One percent H2O in ringwoodite– = 600ºC on KT (at low pressure)

Water in Ringwoodite

140145150155160165170175180185

0 1 2 3 4 5 6 7 8

Mol % H4SiO4

Bul

k M

odul

us (G

Pa)

Birch-Murnaghan EOS

• Fourth-Order• P=3K0fE(1+2fE)5/2{1+3/2(K’-4)fE +

3/2[K0K”+(K’-4)(K’-3)+35/9]fE2}

• Where finite strain, fE = [(V0/V)2/3 –1] /2• If truncated to second order, implied K’

– K’ = 4• If truncated to third order, implied K”

– K”= -1/K0 {(3-K’)(4-K’)+35/9}

Ringwoodite Compression Studies

• Bulk modulus decreases significantly with hydration.

• 1% H2O in ringwoodite = 600ºC.• Consistent with significant

hydration in TZ.

• Need better precision on K’ and K” as f(P,T,H) to constrain the amount of H that might be present.

Ringwoodite UltrasonicsAmbient (Jacobsen et al BGI)

Ringwoodite UltrasonicsAmbient (Jacobsen et al BGI)

• P and S velocity measurements:– Ghz Interferometry in Single Xtls– Vp = 9690 - 0.042 CH2O (m/s)– Vs = 5680 - 0.036 CH2O (m/s)– CH2O = ppm (wt) H2O

• One Percent H2O in Ringwoodite– = 600ºC on Vp (at Low P) (~4%)– = 1000ºC on Vs (at Low P) (~6%)

Shear Wave Velocity Structure

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8

300 350 400 450 500 550 600 650

Depth (km)

She

ar W

ave

Vel

ocity

(km

/s)

Hydration of Wadsleyite and Ringwooditeis more consistentwith model shearvelocity structure

What New Data Are Needed?• Hydrous olivine

– Crystal Chemistry– EOS to 12GPa 1400ºC– Ultrasonics / Brillouin

• Hydrous wadsleyite– Crystal Chemistry– EOS to 10 GPa HiT– Ultrasonics / Brillouin

• Hydrous ringwoodite– Crystal Chemistry– EOS to 50 GPa (Hi T?)– Ultrasonics / Brillouin

Neutron Diffraction of Wadsleyite (ISIS – RAL)

Neutron Diffraction of Wadsleyite

Lateral Velocity Variations in TZ May Reflect Hydration

• Red means Wet• Blue means Dry

Displacement of Phase Transitions

• Wadsleyite and ringwoodite accept more H than olivine or perovskite.

• H will move 410 to shallower depths.• H will move 660 deeper.• Effect of H on depth of 410 & 660 like

Low Temp • Effect of H on Vp & Vs like high Temp

Blue = Dry; Red = Wet

Dry Wet

Cool

Hot

Conclusions• The nominally anhydrous phases

in the mantle can incorporate many times the amount of H and O in the oceans.

• Observed seismic velocities are inconsistent with an anhydrous pyrolite transition zone.

• Velocities are consistent with 0.5 to 1.5 wt % H2O in TZ.

• Better EOS and EP data may be able to provide further constraints.

Conclusions

• Ocean volume need not be constant through geologic time.

• The oceans likely represent a dynamic equilibrium between surface and interior reservoirs.

Are the oceans just the tip of the iceberg?

Deuterium/Hydrogen ratios in the Solar System

• Estimated protosolar: ~ 2 x 10-5

• Earth: 1.5 x 10-4 (SMOW)• Mars: ~ 8 x 10-4

• Venus: ~ 2 x 10-2

• Jupiter and Saturn: ~ 2 - 2.5 x 10-5

• Neptune: ~ 6 x 10-5

• Comets, asteroids, interstellar dust: ???

D/H Fractionation of Mantle Minerals Relative to Phlogopite (after Bell and Ihinger, 2000). Fractionation increases with Pressure

Predicted D/H Fractionation

0

10

20

30

40

50

60

3000 3100 3200 3300 3400 3500 3600 3700 3800

O-H Stretching (cm-1)

D/H

Fra

ctio

natio

n R

elat

ive

to P

hlog

opite

Ringwoodite

Wadsleyite

Opx

CpxGarnetOlivine

Phlogopite

D-H Fractionation Increases With Pressure• High pressure phases prefer H• Water returned to surface is D-

enriched.• The nominally anhydrous silicates

are big, but not sufficient to hold the missing light H.

• If Earth were Proto-solar in value there would be a huge reservoir of light H in the interior.

• Lower mantle or core

How Does the Water Get There? Natural Samples: Eclogites

• Garnet: not much H– H4O4 tetrahedron large– Inhibited by pressure

• Clinopyroxene– HAlSi2O6 (hydro-jadeite)– 1000 ppm H2O by weight @ 4GPa– >5000 ppm at 10GPa

Pyroxene: 2 HAlSi2O6 ->H2O + Al2SiO5 + SiO2

H-Cycling: Natural Samples

• Nominally anhydrous minerals in eclogite can carry 2000 ppm or more

• Olivine can incorporate 2000 ppm• 1000 ppm in crustal portion of slab can

recycle the ocean volume in 4.5 Gy(at current subduction rates).

• Ocean volume may have exchanged more than once.

Interior Reservoir?

• Oceans are only 0.025% of mass.• Chondrites are ~0.10 % H2O• Nominally anhydrous mantle minerals

can incorporate many times the ocean volume.

• Is there a large missing reservoir of H in the interior?

• http://ruby.colorado.edu/~smyth/Home.html

Wadsleyite IINew Work

Wadsleyite II (Spinelloid IV)

• Wadsleyite • Imma or I2/m• a = 5.69 Å• b = 11.50 Å• c = 8.26 Å

• Wadsleyite II• Imma• a = 5.69 Å• b = 29.0 Å• c = 8.26 Å

Wadsleyite II = Spinelloid IV

• Two new structure refinements• Calculated powder patterns• Polarized oriented FTIR + Visible• Raman• Mössbauer • TEM + Fe and Si ELNES• Isothermal bulk modulus

Wadsleyite I & IICalculated XRPD

Wadsleyite II HRTEM & SAED

Wadsleyite II Mössbauer

Wadsleyite I & IIPolarized, oriented FTIR

Wadsleyite I & IIRaman

Wadsleyite II Conclusions• Wadsleyite II is Spinelloid IV• Wadsleyite II is a stable phase in TZ

– May require > 2% H2O by weight– May require > 1% trivalent cations (Fe,

Cr, Al)

• Wadsleyite II is difficult to recognize– XRPD, FTIR, Raman

• Wadsleyite II may obscure 525 km

searching for earth’s lost oceans: an odyssey in mineral...

Documents