Geoscience Reference
In-Depth Information
microtomography has led to the production of
high-resolution 3D images of melt geometry (Zhu
et al
., 2011). Quantitative analysis of such 3D
images will greatly help in determining the devia-
tion from the standard model due to faceting. The
images will also contribute greatly to observing a
deviation from the equilibrium geometry due to
a nonhydrostatic stress, which will be discussed
in Section 3.2.2.
In summary, texturally equilibrated partially
molten rocks, with a melt fraction of less than
a few percent, are expected to have a homo-
geneous distribution of interconnected melt
tubules. Allowing for the continuing debate and
uncertainty regarding the effects of faceting,
the model of equilibrium melt geometry is the
best available for dealing with partially melted
rocks in the Earth's interior, given the lack of
a reliable model for nonequilibrium geometry.
This being the case, the mechanical properties
of texturally equilibrated partially molten rocks
can be treated as ''standard'' properties, and they
will be discussed in detail in Section 3.5.
factors that affect the microstructure under
stress are poorly understood, a systematic control
of experimental conditions is still difficult to
achieve. However, several studies have obtained
highly reproducible results on stress-induced
melt redistributions. The distributions can be
divided into two types: melt redistribution at
the grain scale, and melt redistribution over
distances much greater than the grain scale. In
both types, anisotropy is commonly observed.
Grain scale melt redistribution has been ob-
served in partially molten rocks deformed under
uniaxial compression and in simple shear. Let
σ
1
and
σ
3
be the maximum and minimum principal
(compressive) stresses, respectively. In uniaxial
compression, as schematically illustrated in
Figure 3.2a, melt is redistributed preferentially
along grain boundaries oriented at 15
◦
to 20
◦
from the
σ
1
direction (Daines & Kohlstedt, 1997).
In simple shear deformation, as schematically
illustrated in Figure 3.2b, the long axes of melt
pockets are oriented predominantly at an angle
of 20
◦
from the
σ
1
direction (Zimmerman
et al
.,
1999). In analogues of partially molten rock
(borneol
melt), deformed under pure shear,
the anisotropy was observed through a contin-
uous and nondestructive monitoring of sample
microstructures with ultrasonic shear waves
(Takei, 2010). Takei reported that grain-to-grain
contact faces with their normals oriented nearly
parallel to the
+
3.2.2 Disequilibrium geometry under stress
Any possible deviation of microstructure from
the equilibrium geometry, caused by non-
hydrostatic stress, is important, because the
mechanical and transport properties that are
predicted on the basis of equilibrium geometry
sometimes fail to explain geophysical and
geochemical data. As shown in Section 3.2.1,
the equilibrium geometry is characterized by the
melt geometry on the grain scale, as well as a
homogeneous distribution of melt over distances
much greater than the grain scale. Therefore,
I use the term ''disequilibrium'' when at least
one of these two conditions is not satisfied.
The results of several experimental studies
show that the microstructures in deformed
partially molten rocks deviate substantially from
the equilibrium geometry (e.g., Zimmerman
et al
., 1999; Holtzman
et al
., 2003a,b). On the
other hand, the results from several other studies
have shown that the deviation is small and
invisible (e.g., Hirth & Kohlstedt, 1995). Because
σ
3
direction show a decrease in
area, whereas other contact faces remain almost
unchanged (Figure 3.2c). Although the types of
deformation in these studies were different, and
the microstructural anisotropies were described
in different ways, the results show the same
tendency for enhanced wetting to occur on
grain boundaries that have their normals nearly
parallel to the
σ
3
direction.
Jin
et al
. (1994) reported that deformation
enhanced the wetting of all grain boundaries
isotropically. This phenomenon, called dynamic
wetting, may be described by Equation (3.1) using
A
3. However, dynamic wetting has not been
reproduced in other experiments performed under
similar conditions (e.g., Hirth & Kohlstedt, 1995),
and hence it will not be discussed here. The grain
>
2
.
