Download Thermal and chemical structure at the bottom of the lower mantle

DEEP EARTH
Core–mantle boundary landscapes
The molten-iron alloy of the core meets the mantle’s silicate rock at Earth’s core–mantle boundary. Seismological
images reveal hummocks of iron-enriched material above the boundary, highlighting the heterogeneous nature of
the mantle.
Thermal structure at the bottom of the mantle
T
Upper
mantle
Thermal and chemical structure at the bottom
of the lower mantle
n
LLSVP
e
nuity
le
skit
conti
rov
Small-sca
D”-dis
t-pe Material terogeneities
Supporting Online
Pos
he
ULVZ
www.sciencemag.org/cgi/content/full/329/5998/1513/DC1
re
Co
Materials and Methods
SOM Text
Fig. S1
Tables S1 to S6
References and Notes
Post-perov
skite
Sharp
sides
LVZ
Rost, 2013 Nature Geoscience
ΔVP: -5%
ΔVS: -8%
Δρ: ~1.5-2%
H: <100 km
mantle solidus -> upper bound for geotherm
19 April 2010; accepted 28 June 2010
10.1126/science.1191056
Core
2
ΔVP: ~5-10%
ΔVS: 20-30%
Δρ: ~10%
H: <20 km
Downloaded from www.sciencemag.org on February 12, 2011
pressure high-temperature melting experiments is
Figure 1 | Schematic of the mantle. a, Cold, dense oceanic plates subduct downwards, and plumes of
a natural
KLB-1 peridotite (14). To ensure chemwarm material rise upwards in the mantle. The lowermost mantle contains large-scale heterogeneities,
ical
homogeneity
atvelocity
the smallest
and Fe
as dominated by the material postsuch
as large low shear
provincesscale
(red, LLSVP)
and regions
pervoskite
(grey
by phase
of mantle rocks at extremely high pressures.
mostly
Fe2+
, adashed
glasslines),
wascreated
prepared
bytransitions
using an
The lowermost levitation
mantle also contains
heterogeneities,
aerodynamic
devicesmaller-scale
coupled with
CO2 such as ultra-low velocity zones
(orange, ULVZ). The D"-discontinuity may represent the boundary to the post-perovskite phase. Sun et al.2
laser
heating under slightly reducing conditions
identify a previously undocumented type of low-velocity zone (dark brown, LVZ) at the core–mantle
ofboundary
oxygenbeneath
fugacity
(17).
At interpret
high temperature,
the USA
that they
as small-scale ridges of iron-enriched mantle.
b, The newly
documented
LVZsfrom
exhibitcell
distinct
characteristics
pressures
were
measured
parameters
of compared to the ULVZs (ΔVP, decrease in
P-wave seismic velocity; ΔVS, decrease in S-wave seismic velocity; Δρ, increase in density compared to
the magnesium perovskite (Mg,Fe)SiO3 by using
surrounding, typical mantle; H, height of feature).
a thermal equation
of state
reported
for
Previous
studies
byrecently
x-ray
diffraction
methods
the same KLB-1 peridotitic starting material as
Solidus temperature of the lower mantle
G
uctio
Subd
Melting of Peridotite to
140 Gigapascals
Lower
mantle
Subduction
he border between the core and
mantle is the most significant
internal boundary of our planet.
The transfer of heat across this boundary
plays an important role in controlling both
the dynamics of Earth’s outer core and the
convection of the mantle. Seismological
studies1 of the lowermost few hundred
kilometres of the mantle above the
boundary have revealed a multitude of
three-dimensional (3D) structures on
scales ranging from thousands to just a
few tens of kilometres. Writing in Earth
REPORTS
and Planetary Science Letters, Sun et al.2
use high-resolution seismic images of the
Ryuichi
Nomura
mantle
21. O. Möhler,
P. J. DeMott, G. Vali, Z. Levin, Biogeosciences
28. J. Merikanto,lowermost
D. V. Spracklen,
G. to
W.identify
Mann, S.aJ.previously
Pickering,
Tokyo Institute
Technology
undocumented
chemical
heterogeneity,
4, 1059of(2007).
K. S. Carslaw,
Atmos. Chem. Phys.
9, 8601
(2009).
interpreted
as local
iron5,enrichment,
22. B. C. Christner, C. E. Morris, C. M. Foreman, R. M. Cai,
29. R. Krejci et al.,
Atmos. Chem.
Phys.
1527 (2005).
at the
boundary
beneath
D. C. Sands, Science 319, 1214 (2008).
30. A. M. L. Ekman
et core–mantle
al., Geophys. Res.
Lett. 35,
L17810
North America.
23. R. M. Bowers et al., Appl. Environ. Microbiol. 75, 5121
(2008).
structure
of the7,lowermost
mantle
(2009).
31. U. Kuhn et al., The
Atmos.
Chem. Phys.
2855 (2007).
is complex
Seismic
waves travel
24. K. A. Pratt et al., Nat. Geosci. 2, 398 (2009).
32. J. Lelieveld et
al., Nature(Fig. 1).
452, 737
(2008).
slowly of
through
two large
regions
25. M. O. Andreae et al., Science 303, 1337 (2004).
33. Support fromunusually
a large number
colleagues,
agencies,
of the
lower mantle
located beneath
Africa
26. P. Reutter et al., Atmos. Chem. Phys. 9, 7067 (2009).
and institutions
is gratefully
acknowledged
as detailed
and
the
Pacific
Ocean.
Each
region
is
about
27. M. Kulmala et al., J. Aerosol Sci. 35, 143 (2004).
in the supporting online material.
1
15,000 km across and rises 500–1000 km
above the core-mantle boundary 1. These
regions, commonly termed large low shear
velocity provinces (LLSVPs), have sharp
peripheral boundaries and are thought to
alter the speed of passing seismic waves
because they have a different composition
and their temperature and density is higher
than that of the surrounding mantle1,3.
An undulating discontinuity (D"discontinuity,
Fig. 1) is also observed several
1
5
G. Fiquet,1* A. L. Auzende,1 J. Siebert,1 A. Corgne,2,3 H. Bureau,
H.kilometres
Ozawa,1,4
G.the
Garbarino
hundred
above
core–mantle
boundary 4. This discontinuity probably
Interrogating physical processes that occur within the lowermostmarks
mantle
is atransition
key to understanding
a phase
in the mantle rocks,
induced
by the high
pressures
at depth
that
Earth’s evolution and present-day inner composition. Among such
processes,
partial
melting
has
create a material known as post-perovskite5.
been proposed to explain mantle regions with ultralow seismic velocities
near the core-mantle
Furthermore, at the edges of the LLSVPs,
boundary, but experimental validation at the appropriate temperature
and pressure
regimesof mantle
thin intermittent
layers of 5–40 km
have been identified
that greatly
remains challenging. Using laser-heated diamond anvil cells, werocks
constructed
the solidus
curvereduce
of
6–8
CMB
the
speed
of
passing
waves
. These smaller
a natural fertile peridotite between
36 and 140 gigapascals. Melting at core-mantle boundary
Lower mantle ◀ ▶︎ Outer core
anomalies are known as ultra-low velocity
pressures occurs at 4180 T 150 kelvin, which is a value that matches
estimated
mantle
geotherms.
zones (ULVZs).
Scattered
seismic
waves
Molten regions may therefore exist at the base of the present-day
Melting phase
maymantle.
identify heterogeneities
on relations
even smaller
9
scales,
about
10 km
in
size
.
Those
and element partitioning data also show that these liquids could host many incompatible could
be created by partial melting or chemical
elements at the base of thepyrolite
mantle.
heterogeneities, possibly linked to subducted
Lower mantle
plume?
Sebastian Rost
oceanic plates. However, little is known about scale structure at depth by analysing details of
the fine-scale structure of the deep Earth.
the waveforms6 of seismic waves generated by
Sun et al.2 use seismic data from
deep earthquakes in the Philippines recorded
10
Earthscope’s
USArray
—
a
dense
grid
of
in the mid-western USA.
100
3750 KThe authors identify several ridges of
more thanXRD
400 seismometers112
deployed across
the United States — to analyse the detailed
mantle rocks at the core–mantle boundary
210
structure
that cause passing seismic waves to slow
80 of the core–mantle boundary
beneath North America. They image the fine- considerably. The ridges are present in some
gasket
004
020
200
NATURE GEOSCIENCE | VOL 6 | FEBRUARY 2013 | www.nature.com/naturegeoscience
89
60
zones
may be partially molten (5).
X-ray
Intensity (counts)
is that these
© 2013 Macmillan Publishers Limited. All rights reserved
Recent high-resolution waveform studies also
002
200
find evidence
the ULVZ material is denser
110
40
Multi-anvilthat
press!
than the
+ surrounding
above solidus! mantle (11). These partially
Melting criteria!
103
below solidus!
molten+ regions
have not been detected to be lat! -CaPv/Fp melting!
20
!
111
erally continuous and have a thickness ranging
!
Diamond anvil cell!
from a■ few
kilometers
up to about 50 km.
Fiquet+,
2010 Science!
! -diffuse scattering
0
It is
linkEPSL
these observations with
■ attractive
Andrault+,to
2011
250
an episode of extensive melting that probably
Fiquet+, 2010 Science
2715 K
affected the primitive Earth, leading to the formation of a deep magma ocean. If the evolution
200
of a terrestrial magma ocean resulted in the
Difficulty for detecting the initiation of melting
formation of a layer of melt at the base of the
4
150
mantle early in Earth history, its survival depends
on whether it was (and maybe still is) gravitationally and chemically stable (12). If this is the
100
case, such a layer would be an ideal candidate for
an unsampled geochemical reservoir hosting a
50
variety of incompatible species, notably the planet’s missing budget of heat-producing elements
0
(13). The presence of high-pressure melts would
6
10
12
14
8
also have consequences for chemical reactions
Diffraction angle (2-θ)
between the mantle and core, the dynamics of
REPORTS
the lowermost mantle, and the heat flow across
The lower abundance of ~10- to 100-mm paranalyzed here.
mass GPa
of each mineral (table
Fig. 1. Diffraction patterns collected
atThe61
1
S1) was obtained from the modal abundances, ticles in the smooth terrain can potentially be
the CMB.
Institut de Minéralogie et de Physique des Milieux Condensés,
and the density calculated
from its mean chem- explained by (i) smaller grains having higher
after normalized reference background
subtracical compositions (3). The total mass of our ex- ejection velocity and therefore higher loss rates
To
constrain
the
existence
of
melt
at
the
base
Institut
de
Physique
du
Globe
de
Paris,
Université
Pierre
et
X-ray
micro-CT methods
Results
amined particles
mg. From the mineral from Itokawa after impacts (13), (ii) selective
tion: subsolidus at 2715 K (bottom)
andis 14.5
above
Marie Curie, UMR CNRS 7590, Université Paris Diderot, 140
mass and the porosity, we obtained an average electrostatic levitation of smaller grains (13), and/or
of the mantle, we performed melting experiments solidus at 3750 K (top). The diffuse
scattering
density
of 3.4 g/cm . liquid
This corresponds to grain (iii) size-dependent segregation by vibration [the
rue de Lourmel, 75015 Paris, France. 2Institut de Physique du
density and is comparable to the measured grain Brazil-nut effect (14)].
on a fertile peridotite composition over a range of contribution is outlined by the shaded
area (3.54
as Ta0.13 g/cm ) (10).
Globe de Paris, Equipe de Minéralogie
à l’Institut de Minéralogie
density of LL chondrites
BL47XU@SPring8
the collected sample is representative of Itokawa
lower-mantle pressures between 36 and 140 GPa guide; it does not correspond to a Ifphysical
et de Physique des Milieux Condensés, 140 rue de Lourmel,
structural
has the average
porosity
of LL chondrites,
its
(B)
subsolidus
(A) above
solidus
GPa/3690
K and
151
GPa/3680 K
bulk density would be 3.1 T 0.2 g/cm . The
using a laser-heated diamond-anvil cell (DAC) model of142
75015 Paris, France. 3Observatoire Midi-Pyrénées, UMR CNRS
the liquid. HKL indexes
are ofgiven
forthen be 39 T 6%
macroporosity
Itokawa would
keV
7
keV
5562, 14 rue Edouard Belin, 31400 Toulouse, France. 4Department
on the basis of the bulk density
of Itokawa (1.9 T
coupled with in situ synchrotron measurements 7
remaining diffraction peaks that0.13can
assigned
g/cm be
) (1). This
is consistent with a rubbleof Earth and Planetary Sciences, Tokyo Institute of Technology
asteroid model of Itokawa (1).
(14). Our study thus extends the pressure range of to magnesium silicate perovskite,pileobserved
above
The sphere-equivalent diameters of tapping
2-12-1 Ookayama, Meguro-ku, Tokyo 152-8551, Japan. 5Eurosample particles
calculated
from their volumes
previous measurements (15, 16) of the solidus the solidus temperature at this pressure
(top).
Stars
pean Synchrotron Radiation Facility, BP220, 38043 Grenoble
range from 14 to 114 mm (median 36.8 mm),
whereas the diameters
of spatula
and liquidus temperatures of a mantle-like com- denote diffraction peaks of Ca-perovskite
and
fer-sample particles
cedex, France.
range from 0.5 to 32 mm (median 3.5 mm). These
are smaller
the size-sorted, mm- to
position to depths exceeding those of the CMB at ropericlase affected by partial particles
melting
atthanthese
*To whom correspondence should be addressed. E-mail:
cm-sized particles observed in close-up images of
MUSES-C Regio (2). The mm- to cm-sized parti2900 km. The starting material used for the high- conditions
[email protected]
8bykeV
8 keV (bottom).
cles were not comminuted
the pressure of the
eophysical and geochemical observations favor the presence of chemical
heterogeneities in the lowermost
chondritemantle.
These are thought to be either primitive mantle
residues (1), dense subducted slab components
(2), products of chemical interactions between
the core and mantle (3, 4), or dense melts perhaps
as old as the Earth itself (5). The core-mantle
boundary is a complex region that has been the
focus of numerous geophysical studies. Seismologic studies suggest the presence of two large
low-shear velocity provinces (LLSVPs) under the
African continent and in the pacific basin (6, 7).
3
The consensus view is that these slow regions
(which are possibly up to 1000 km thick) exhibit
an anomalously low shear velocity and increased
bulk modulus but are not usually thought to be
partially molten (8). Additionally, extensively documented ultralow-velocity zones (ULVZs) correspond to localized features at the core mantle
boundary (CMB), with strong reductions in
seismic velocities (in the range of 10 to 30%)
for both P and S waves (9, 10); the interpretation
*
*
*
*
3
3
Figure 3A shows the shape distri
tapping samples. The mean b/a an
0.71 T 0.13 and 0.43 T 0.14, resp
and c are longest, middle, and
diameters, respectively, of a best-fi
The distribution among polyminer
mineralic particles does not have
difference based on the Kolmog
3
3
1516
17 SEPTEMBER 2010
Spatial resolution!
~70nm (for <100um sample)!
~200nm(for >100um sample)!
X-ray energy!
7, 8 keV (Fe K edge: 7.11 keV)
VOL 329
SCIENCE
www.sciencemag.org
5um
melt
spacecraft during touchdown [~0.02 MPa (7)] if
they are coherent (11), and the collected small
particles should be original regolith particles from
the smooth terrain. Three sampling mechanisms are
possible (7): (i) impact by the sampler horn, (ii)
electrostatic interaction between charged particles
and possibly charged sampler horn, and (iii) levitation by thruster jets from the ascending spacecraft. Some mechanisms may have caused some
size sorting. However, because details concerning
the touchdown conditions are not known, we
cannot specify the mechanism(s) and such effect.
The cumulative size distribution of the
tapping samples has a log slope of about –2 in
the range of 30 to 100 mm (Fig. 2). Large particles might have been selectively picked up
from the tapping samples. The spatula samples
have a log slope of –2.8 in the range of 5 to 20 mm
(Fig. 2). However, sweeping by the spatula
would have pulverized some of the particles,
and the slope is an upper limit to the original
slope. Thus, the slope
for the fine particles (~5
6
to 100 mm) in the smooth terrain should be
shallower than –2.8 and probably around –2. This
slope is shallower than that of Itokawa boulders
of 5 to 30 m (–3.1 T 0.1) (12). If transition of the
slope from about –3 to –2 occurs at the mm- to
cm-sized region, then we can explain the observation of abundant mm- to cm-sized regolith
(2). The lack of mm-sized particle in the Hayabusa
samples might be explained by a small probability
of collecting these small particles. However, the
possibility of size selection biases during sampling
from Itokawa or agglomerations of small particles
(11) cannot be excluded. In contrast, abundant
sub-mm regolith powder was observed on the
Moon, and the size distribution from lunar sam-
heated region
Nomura+, 2014 Science
0
642
LAC (cm-1)
Fig. 1. Slice images of Itokawa particles obtained by microtomography with a gray sca
linear attenuation coefficient (LAC) of objects (from 0 to X cm−1), where X is the maximum L
CT images. (A) Sample RA-QD02-0063 (7 keV, X = 431 cm−1). (B) RA-QD02-0014 (7 keV,
Some voids define a 3D plane (arrows). (C) RA-QD02-0042 (7 keV, X = 575 cm−1). (D) R
(7 keV, X = 431 cm−1). Concentric structure is a ring artifact. Bright edges of particles
artifacts resulting from refraction contrast. Ol indicates olivine; LPx, low-Ca pyroxene
pyroxene; Pl, plagioclase; CP, Ca phosphate; Tr, troilite; and Meso, mesostasis.
Iron-rich region in the hottest part
5
Fig. 2. Cumulative size distribution of Itokawa particles. Sphereequivalent diameters of the tapping samples and diameters of the
spatula samples are shown.
Solidus temperature of the pyrolitic mantle
Thermal structure at the bottom of the mantle
CMB
Lower mantle ◀ ▶︎ Outer core
CMB
Lower mantle ◀ ▶︎ Outer core
Solidus
Nomura+, 2014 Science
Nomura+, 2014 Science
Solidus temperature: 3570(200) K @CMB
Mantle solidus: upper bound for geotherm
7
8
news & views
–mantle boundary landscapes
lloy of the core meets the mantle’s silicate rock at Earth’s core–mantle boundary. Seismological
mmocks of iron-enriched material above the boundary, highlighting the heterogeneous nature of
Chemical structure at the bottom of the mantle
letters to nature
le
Small-sca es
neiti
heteroge
ULVZ
re
Co
Sharp
sides
600
700
800
4,500
4.2
4,000
4.0
MORB
3.8
Post-perov PYROLITE
3.6
skite
3.4
14
16
LVZ 18
660 km depth
20 22 24 26
Pressure (GPa)
28
S
Core
adiabat
120
140
3,000
Mantle adiabat
2,000
0
20
40
60
80
100
solid circles represent melting temperatures of MORB and MgSiO3, respectively,
Δρ: ~1.5-2% to perovskitite lithology, but once MORB transbecause of the transformation
H: <100
forms to perovskitite
at 720km
km depth, it is no longer buoyant in the deep mantle.
determined in a laser-heated diamond cell. Open squares represent melting
temperatures of MORB, determined in the multi-anvil apparatus. Melting curves of
St14, Ca-pv14,15 and Mg-pv13,14 are substantially higher than that of
10 MORB. The
melting curves of Mg-pv from refs 13 and 14 (labelled as Mg-pv13 and Mg-pv14,
respectively) are plotted for comparison. The mantle adiabats are from Boehler23.
shallower than that expected1,2,8,9. The transition boundary has a
positive
pressure±temperature
Figure 1 | Schematic of the mantle. a, Cold, dense oceanic plates subduct
downwards,
and plumes of slope, whereas the transition boundary inlarge-scale
the underlying
harzburgite has a negative slope10. Within a
warm material rise upwards in the mantle. The lowermost mantle contains
heterogeneities,
cool slabby
(for
such as large low shear velocity provinces (red, LLSVP) and regions dominated
theexample
material1,000
post-8C at the 660 km depth), the transforperovskite
lithology
in the basaltic crust and the underpervoskite (grey dashed lines), created by phase transitions of mantle mation
rocks at to
extremely
high
pressures.
lying harzburgite layer would occur at a similar depth, implying that
The lowermost mantle also contains smaller-scale heterogeneities, such as ultra-low velocity zones
the delamination of the basaltic crust
from the slab is unlikely at the
2
(orange, ULVZ). The D"-discontinuity may represent the boundary to the
post-perovskite
phase.
Sun et al.
660 km discontinuity.
identify a previously undocumented type of low-velocity zone (dark brown,
LVZ) atexperiments
the core–mantle
Melting
were carried out using a multi-anvil
3
boundary beneath the USA that they interpret as small-scale ridges of apparatus
iron-enriched
mantle.
and
laser-heated diamond cell11 in a pressure range
between
and(ΔV
64 PGPa
(Fig. in
3). The solidus temperatures are
b, The newly documented LVZs exhibit distinct characteristics compared
to the 16
ULVZs
, decrease
2,400
K
and
2,700
at 22 GPa
P-wave seismic velocity; ΔVS, decrease in S-wave seismic velocity; Δρ, increase in densityKcompared
to and 27 GPa, respectively, based on
the multi-anvil experiments. The solidus phase assemblage at
surrounding, typical mantle; H, height of feature).
27 GPa is Al-bearing Mg±perovskite and Ca±perovskite, stishovite,
Al-phase and trace majorite (,3%). No important phase transformations have been reported at higher pressures up to 100 GPa
oceanic plates. However, little is known about scale structure(ref.
at depth
by analysing details of
9), suggesting that this assemblage remains in the deep mantle,
6
the fine-scale structure of the deep Earth.
the waveformsexcept
of seismic
waves would
generated
by
that majorite
be completely
transformed to perovskite
2
Sun et al. use seismic data from
deep earthquakes
in the Philippines
recorded
at pressures
slightly above
27 GPa. Analysis of run products above
Earthscope’s USArray 10 — a dense grid of
in the mid-western
USA.temperature at 27 GPa showed that Ca±perovskite is the
the solidus
phase,
followed
more than 400 seismometers deployed across
The authorsliquidus
identify
several
ridgesbyofstishovite, Al-Ca-phase and Al-bearing
Mg±perovskite.
Theboundary
partial melt composition is enriched in MgO
the United States — to analyse the detailed
mantle rocks at
the core–mantle
and FeO
andwaves
depleted
in SiO2 and Al2O3, reØecting that Mg±
structure of the core–mantle boundary
that cause passing
seismic
to slow
perovskite
is
eliminated
near the solidus temperature. The
beneath North America. They image the fine- considerably. The ridges are present inÆrst
some
partial melt is probably denser than the solid residue in MORB
composition at lower mantle pressures because of its higher iron
6 | FEBRUARY 2013 | www.nature.com/naturegeoscience
89
content. The melting curve of MORB determined in the laserheated diamond cell is consistent with that determined by multi© 2013 Macmillan Publishers Limited. All rights reserved
anvil experiments for the same starting material between 22 and
27 GPa (Fig. 3), and also with previous results at low pressures12. It is
substantially lower than that of each constituent mineral: Mg±
perovskite13,14, Ca±perovskite14,15 and stishovite14. The present
measurements of the melting temperature of MgSiO3 are consistent
with previous studies13,14,16 at 12±28 GPa.
The melting temperature of basalt is about 250 K lower than that
of mantle peridotite17 at a depth of 1,500 km (corresponding to a
pressure of 64 GPa). Extrapolation to 135 GPa yields a melting
temperature of MORB of about 4,000 K at the core±mantle bound-
Future work: melting temperature measurement
ancient crust, BIF, MORB, and BMO residues…
Multi-anvil press
NATURE | VOL 397 | 7 JANUARY 1999 | www.nature.com
3600(100) K
Temperature uncertainty
Melting criteria!
! -Temperature jump with increasing laser power
the melting relationships of Simon22 (S) and Kraut and Kennedy23 (KK). Open and
27 GPa from X-rayΔV
diffraction
P: -5% and microprobe data. The density proÆle of pyrolite
ancient crust, BIF, MORB, BMO residues…
Laser-heated diamond anvil cell!
(LH-DAC)
Figure 3 Melting curve of MORB extrapolated to the core±mantle boundary using
pyrolite (dashed line). Solid circles represent the calculated densities at 24, 26 and
is from a previousΔV
study
. Pyrolite becomes denser than MORB at 660 km depth
S: -8%
9
ary, which is substantially lower than that of single-phase MgSiO3 ±
perovskite (Fig. 3). If the temperature locally reaches 4,000 K in the
D0 region, which may be a graveyard for subducted lithosphere, the
former basaltic crust could partially melt. This may provide an
explanation for the recent seismic observations of the seismic
anisotropy18,19 and anomalously slow P-wave velocities20,21 at the
base of the mantle. Under such a scenario, the temperature of the
outer core must be higher than the 4,000 K required for melting of
MORB perovskitite (Fig. 3). The temperature difference over the
thermal boundary between core and mantle may reach 1,500 K, and
hot mantle plumes, including partially molten slab materials, are
likely to arise from this depth.
M
.........................................................................................................................
Conclusions and future works
Methods
We have determined the phase relations and the melting temperatures of
MORB over a wide pressure range by using two complementary high-pressure
techniques, the multi-anvil apparatus and the diamond-anvil cell. The multianvil experiments provide detailed chemical composition information of the
individual minerals in the high-pressure assemblages through the electron
!
microprobe and structure data through X-ray diffraction. The phase identiÆcation was also conÆrmed by micro-Raman spectroscopy. The temperatures
in the multi-anvil experiments were measured with a W5Re-W26Re ther!
mocouple,
whereas pressures were determined based on Æxed pressure
calibration points, including the ilmenite±perovskite transition in MgSiO3
CMB
(ref. 3). Pressure calibration at high temperature was based on the Al2O3
solubility in the MgSiO3 ±perovskite structure at pressures above 23 GPa (ref. 6).
Combining accurate temperature measurements with a self-consistent pressure
calibration, the present multi-anvil experiments provide a reliable determination of the dP/dT slope of the majorite±perovskite transformation in MORB
and relatively accurate measurements of melting temperature up to 27 GPa, the
maximum attainable pressure in the multi-anvil apparatus.
! To extend the melting-temperature measurements for MORB to pressures
higher than 27 GPa, we conducted experiments in a laser-heated diamondanvil cell. In the present diamond cell experiments, a double-sided heating
system with a multimode Nd:YAG laser was used to minimize both axial and
radical temperature gradients in the heated sample11. The MORB glass sample
was sandwiched between two Re foils with Al2O3 layers, as thermal insulators,
on the top and bottom of the sandwich assemblage. The interface between the
Thermal structure!
-Solidus temperature of pyrolite 3570(200) K @CMB!
!
-> upper bound for T
Nomura+, 2014 Science
!
Chemical heterogeneities!
!
!
<- melting temperature of ancient crust, BIF, MORB and BMO !
! residues
© 1999 Macmillan Magazines Ltd
55
Zhang+Fei, 2008 GRL
11
160
Hirose+, 1999 Nature
Pressure (GPa)
Figure 2 Comparison of zero-pressure density changes in MORB (solid line) and
ΔVP: ~5-10%
ΔVS: 20-30%
Δρ: ~10%
H: <20 km
Upper bound for TCMB
3,500
1,500
10
Core
KK
MORB
2,500
30
Rost, 2013 Nature Geoscience
LLSVP: iron-rich?!
ULVZ: partially molten?
Mg
-pv 1
3
Mg
-pv 1
4
500
4.4
Temperature (K)
Zero-Pressure Sub
Density
(g cm–3)
duction
Lower mantle
plume?
n
LLSVP
e
nuity
skit
conti
rov
D”-dis
t-pe
Pos
5,000
Depth (km)
Lower
mantle
uctio
e lowermost mantle
ismic waves travel
ugh two large regions
cated beneath Africa
Each region is about
rises 500–1000 km
boundary 1. These
rmed large low shear
SVPs), have sharp
and are thought to
ing seismic waves
fferent composition
and density is higher
unding mantle1,3.
ontinuity (D"also observed several
ove the core–mantle
ntinuity probably
on in the mantle rocks,
essures at depth that
n as post-perovskite5.
ges of the LLSVPs,
s of 5–40 km of mantle
fied that greatly reduce
aves6–8. These smaller
as ultra-low velocity
red seismic waves
neities on even smaller
size9. Those could
elting or chemical
ly linked to subducted
Melting temperature: MORB
Upper
mantle
Subd
en the core and
ost significant
ry of our planet.
ross this boundary
e in controlling both
’s outer core and the
tle. Seismological
ost few hundred
tle above the
ed a multitude of
D) structures on
ousands to just a
. Writing in Earth
Letters, Sun et al.2
ismic images of the
dentify a previously
cal heterogeneity,
on enrichment,
undary beneath
12
CMB
Lower mantle ◀ ▶︎ Outer core
pure iron
Anzellini+, 2013 Science
Fe-O-S liquidus
Terasaki+, 2012 EPSL
FeH
Sakamaki+, 2009 PEPI
Nomura+, 2014 Science
0.6wt% hydrogen to make the “liquid” outer core
13
14
Temperature gradient in LH-DAC
Conclusions and future works
Thermal structure!
3800
Radia
3300
!
l (mea
sured
-Solidus temperature of pyrolite 3570(200) K @CMB!
)
Temperature (K)
Nomura+, 2014 Science
!
!
2800
Thermal conductivity (κ∝)!
1/T (lattice)!
const.!
T3 (radiation)
2300
1800
1300
-lower bound for core geotherm
-> FeHx Tm only up to 20 GPa
Sakamaki+, 2009 PEPI
Calculation after!
Manga+Jeanloz, 1996 GRL
800
-> upper bound for TCMB!
-> Tm in Fe-H-X system over 135 GPa is necessary
Chemical heterogeneities!
300
0
1
2
3
4
Axial distance from hotspot (µm)
5
!
!
!
<- melting temperature of ancient crust, BIF, MORB and BMO !
! residues
15
16
17
18