Geoscience Reference
In-Depth Information
T
=
2000 K
10
29
10
14
s
=
0.1 MPa
10
27
V*
=
20 cm
3
/mol
10
12
10
25
(1)
10
10
(2)
(3)
10
23
10
8
10
(4)
10
21
10
6
(5)
3
(6)
10
19
10
4
(7)
(8)
1
10
17
10
2
10
15
10
0
0
50
100
150 200
depth, km
250
300
350
400
0
20
40
60
80
100
pressure, GPa
Fig. 4.9
Viscosity-depth relationships calculated using
various experimental results on high-pressure
rheological properties of olivine (Karato, 2010b). Solid
curves correspond to results by Kawazoe
et al
. (2009)
(1) (dry) and Karato and Jung (2003) (7,8) ((7): 1,000 ppm
H/Si, (8): 10,000 ppm H/Si) where water content was
controlled, deformation mechanisms are identified and
appropriate thermodynamic formula was used. Other
results ((2), (3), (4), (5), (6)) are from studies where
unreasonable extrapolations are made or the influence
of water content was not examined. For the detailed
discussions on these results, see Karato (2010b).
Fig. 4.10
Influence of pressure on effective viscosity
for various values of activation volume
η/η
0
=
exp
PV
∗
RT
is used. T
=
2000 K.
Also critical is the influence of water. In many
previous studies, substantial amounts of water
were dissolved in the samples particularly at high
pressures, but water contents were not well char-
acterized. Because the water reduces the effective
viscosity (the creep strength) dramatically, dis-
solved water can lead tomisleading interpretation
of the results. For more details on these issues,
see Karato (2010b).
Using a newly developed deformation appa-
ratus, RDA (rotational Drickamer apparatus) at
the synchrotron X-ray facility, Kawazoe
et al
.
(2009) performed deformation experiments on
olivine aggregates to
one changes the pressure by 0.3GPa, one will see
the change in strain-rate by
25% (at 1600K) (if
strain-rate is fixed then a change in stress is
∼
7%).
To determine the activation volume with 10% er-
ror, it will be necessary to measure the difference
in strain-rate by
∼
0.7%
error). Measuring the change in strain-rate (or
stress) with this resolution is difficult particularly
when grain-size affects the strength. In contrast,
if pressure difference of 10GPa is used, then the
change in strain-rate will be a factor of
∼
2-3% error (or stress by
∼
∼
∼
1900K) for
nearly water-free (dry) samples. The water con-
tent and microstructures (grain-size, dislocation
structures) were carefully examined and the evi-
dence for power-law creep involving dislocation
recoverywas found at high temperatures. Samples
show low water contents (below the detection
limit) and the data represent the creep strength of
dry olivine. The data from Kawazoe
et al
. (2009)
corresponding to the power-law creep regime
are summarized in Figure 4.11. These data were
10GPa (to
∼
1800
(a factor of
9 change in stress). Even with less
precise mechanical measurements under high-
pressure environment, changes in properties by
high pressure are so large that much more precise
determination of pressure effects can be made if
one uses a high-pressure deformation apparatus.
∼
