down the [111] zone axis of the KIT - 6 templated mesopore structure (Figure
10.19 a), seeing a hexagonal pattern of the pores, a hexagonal pattern of SAED
corresponding to the [001] zone axis of the rhombohedral Cr
2
O
3
can be obtained
without specimen tilting (Figure 10.19 b). Consequently, the [111] axis of the KIT -
6 - like cubic mesoporous unit cell and the [001] axis of Cr
2
O
3
are perfectly parallel.
When a particle is imaged on the [100] projection of the cubic mesoporous system
(Figure 10.19 c), a perfect square pattern of the mesopores can be seen. Figure
10.19 d is the corresponding HRTEM image recorded from a corner of the particle
in Figure 10.19 c. The image viewed down the
221
[]
axis of Cr
2
O
3
shows a contrast
pattern which is the closest to a square pattern from the Cr
2
O
3
crystal structure.
From this sample, many particles have these peculiar orientations, which seem to
have a close relation with the structural symmetry of KIT - 6. This unique phenom-
enon has not been observed with other samples of porous oxides. At present, the
reason for the formation of such coherence between crystal structure and the
symmetry of the nanopore system is not understood.
If we compare the image contrast pattern in Figure 10.19 c with the correspond-
ing one from KIT - 6, we can conclude that Cr
2
O
3
only replicated one of two inter-
laced pores in KIT - 6. Figure 10.19 a is a rare example showing replication of both
pores in the central area. On the other hand, TEM images from KIT - 6 - templated
Co
3
O
4
showed that two interlaced pores were replicated in all the particles [100] .
Pore size and pore volume measurements by N
2
adsorption/desorption of these
two specimens confi rmed the above conclusion. This implies that there are some
small channels connecting the two interlaced pores in KIT - 6.
Crystal growth of metal oxides has also been performed in the pore systems of
some cage - containing mesoporous silicas, such as FDU - 12 [79, 80] and SBA - 16
[77] . SBA - 16 has a body centered cubic structure (space group
Im
3
) and FDU - 12
is face centered cubic (space group
Fm m3 ). Accordingly, each spherical nanocage
is 8 coordinated by the neighboring nanocages in SBA - 16, but 12 coordinated in
FDU - 12. When crystals of oxides form in these two silica phases, it can be expected
that the structures of the products are 3D arrangements of solid nanospheres
connected to each other by some very short nanorods [107, 109] . The HRTEM
technique has been used to examine the yields and the single crystal properties of
these porous oxides. The small bridges are barely revealed in the TEM images.
However, it is possible to see a bridge at an edge of a particle where two adjacent
nanospheres can be identifi ed. Another diffi culty in imaging the porous crystals
templated by these mesoporous silicas is that, since the crystal orientation has no
relation to the mesostructure, it is diffi cult to fi nd a view direction parallel with
principal axes of both the crystal structure and the mesopore structure.
It seems to be possible that most transition metal oxides can be made in porous
crystals with different morphologies using various mesoporous silicas as tem-
plates. It is expected that these materials have potential in applications such as
catalysis, fuel cell, gas sensors and Li - batteries. Their physical properties would
fall in between nanoparticles and bulk specimens, although our knowledge about
these properties is still very limited.
10.7 Porous Materials 475