464 10 Transmission Electron Microscopy
high Tc superconducting oxides [51 – 54] . To make sure that recrystallization only
takes place at the interface of the crystal and the amorphous layer, the beam bright-
ness should be carefully controlled to allow a lengthy process. Otherwise, individ-
ual nanocrystallites of metal oxides, for example copper oxide and barium oxide,
may be created in the amorphous layer. On the other hand, if the composition of
the decomposed surface coating layer changes, the amorphous layer cannot be
repaired. For example, the surface of HgBa
2
CuO
4+ δ
crystals is often decomposed
by loss of Hg as detected by a narrow electron beam EDX analysis (Figure 10.12 b)
[55] . This amorphous layer cannot recrystallize into the parent crystal, although
formation of barium copper oxide microcrystallites is still possible. The surface
crystallinity of these high Tc superconducting materials certainly affects the physi-
cal properties of the materials [56] .
It is worth mentioning that a very small electron beam, for example about 2 nm
in the case in Figure 10.12 b for HgBa
2
CuO
4+ δ
, should be used in EDX. The collec-
tion time for the EDX spectrum shown in the left inset was several hours, while
the EDX spectrum in the right inset was obtained in only a few minutes. During
such a long collection time, the operator must make sure that the beam location
is always at the surface amorphous region. Narrow beam diffraction patterns may
help in this respect. When the beam is very small, different diffraction patterns
from different areas can be observed. From the crystalline region, a single crystal
pattern with individual diffraction disks is shown. The pattern from the disordered
region shows one or more diffraction rings. When the beam shifts away from the
particle, no diffraction signals can be observed. Exposing the specimen area to an
electron beam for a long time may result in heavy carbon contamination. The
beam location at the amorphous layer should therefore be moved many times
along the surface.
When a clean surface of a crystallite is observed without any coating layer, it is
possible to determine the terminal atomic planes on various crystal surfaces. For
example, three distinct surface terminal planes are marked in Figure 10.12 a.
HRTEM studies of the [001] surface of YBa
2
Cu
4
O
8
by examination of profi le
images indicated that the surface usually terminated with the CuO
2
plane corre-
sponding to a sequence of bulk – Y – CuO
2 − x
, while the [001] surface of YBa
2
Cu
3
O
7
prefers to terminate with the BaO atomic plane [57] . Zandbergen investigated a
CuO terminal plane on the [001] surface of GdBa
2
Cu
3
O
7
and found that the ter-
minal atomic plane often had a different image contrast indicating lattice distor-
tion including surface relaxation [58] , especially when oxygen in the termination
plane was partially lost.
By exposing a clean crystal surface to the electron beam, a series of HRTEM
surface profi le images can be recorded and a movement of surface atoms from a
high - energy site to a low - energy site may be observed [47] .
Observation of various surface defects is an important topic in HRTEM surface
profi le imaging. In Figure 10.11 b, it can be seen that over a large area the coating
layer of C - La
2
O
3
is not continuous. Instead, the coating layer is broken into some
small islands. This is because the lattices of C - La
2
O
3
and La
2
CuO
4
as shown in
Figure 10.11 c do not match each other perfectly. Very interesting surface profi le