Geoscience Reference
In-Depth Information
discontinuity, which is characterized by a 2.5
−
4.0% increase of seismic shear wave velocity,
occurs about 250 km above the CMB (Loper &
Lay, 1995; Wysession
et al
., 1998). Although
a shear velocity jump of
V
S
=
phase, in-situ soundwave velocity measurements
under thermodynamic stability field of MgSiO
3
post-perovskite phase is therefore prerequisite.
Murakami
et al
. (2007b) first performed the direct
shear wave velocity measurements of MgSiO
3
post-perovsktie phase by Brillouin spectroscopy
in a DAC up to a pressure of 172 GPa, in com-
bination with infrared laser annealing for sam-
ple synthesis and synchrotron X-ray diffraction
measurements for phase identification and pres-
sure determination. The experimental procedure
of this study is almost the same as that in
Murakami
et al
. (2007a). To ensure that single-
phase post-perovskite was obtained, sample syn-
thesis was performed under the conditions well
within the stability field of the post-perovskite
phase. After each increase of pressure, the sam-
ple was annealed for
2
.
75 is typical
(Wysession
et al
., 1998), recent seismic observa-
tions have reported relatively small discontinu-
ities of
V
S
∼
1.1% beneath the Central
Pacific region (Lay, 2006). This rapid velocity
change is among the most enigmatic of seismic
features and its origin is not fully understood.
This discontinuity is seen in many regions around
the world with a variation in the observed depth
of
0.6
−
100 km. This discontinuity may not, how-
ever, be a global and ubiquitous feature above the
CMB (Vidale & Benz, 1993).
A number of explanations for the D
disconti-
nuity has been offered. These include (Wysession
et al
., 1992) a thermo-chemical boundary (Davies
& Gurnis, 1986), a phase transition (Sidorin
et al
.,
1998; Murakami
et al
., 2003), or a breakdown
of the constituent mantle minerals (Stixrude &
Cohen, 1993). However, there has been no con-
sensus on how the phenomena of the D
layer
can be explained. The recent experimental iden-
tification of the post-perovskite phase transition
in MgSiO
3
(Murakami
et al
., 2004) may offer new
insights into the properties of D
region. Theo-
retical calculations (Iitaka
et al
., 2004; Oganov
& Ono, 2004; Tsuchiya
et al
., 2004; Stackhouse
et al
., 2005; Wentzcovitch
et al
., 2006) and inter-
pretations of seismic data (Hernlund
et al
., 2005)
suggest that the perovskite to post-perovskite
phase transition and the properties of this new
phase could reasonably explain most of the seis-
mic characteristics of the D
layer. However, the
magnitude of the velocity jump caused by the per-
ovskite to post-perovskite phase transition was
not been determined through direct experimen-
tal measurements, and it is important to have
those experimental data in order to interpret D
phenomena.
Unlike the MgSiO
3
perovskite, the MgSiO
3
post-perovskite phase is unquenchable, which
converts to the amorphous state at ambient condi-
tion after releasing the pressure. In order to obtain
the elasticity data from MgSiO
3
post-perovsktie
∼
∼
1hourbyCO
2
laser at
∼
2000 K, and the laser was rastered over the
entire sample area several times to ensure pres-
sure homogeneity and complete transformation
of the sample to post-perovskite phase. Again,
this procedure substantially improved the qual-
ity of the Brillouin spectra at high pressure and
the reliability of the Brillouin measurements,
allowing us to dramatically extend the upper limit
of pressure for Brillouin measurements. Repre-
sentative high-pressure Brillouin spectrum and
the pressure dependence of the aggregate shear
velocities for the post-perovskite phase are shown
in Figure 6.6a. Fitted line of those data by Eule-
rian finite strain equations gives the
G
0
,and
G
0
.
For this regression, the zero pressure volume,
isothermal bulk modulus and pressure derivative
of the bulk modulus were adopted from recent
computational results at ambient temperature
(300 K). The best-fit values are
G
0
=
136 GPa and
G
0
=
2.0. These best-fit values are in excellent
agreement with the computational calculations
(Tsuchiya
et al
., 2004, 2005; Caracas & Cohen,
2005), although their uncertainties (15 and 13%,
respectively) are rather high owing to the small
number of data points.
From the results of Murakami
et al
. (2007b) and
those for MgSiO
3
perovskite (Murakami
et al
.,
2007a), the shear wave velocity contrast across
perovskite to post-perovsktie phase transition
