378 9 Intermetallic Alloys
Stacking Faults
In MoSi
2
, stacking faults on (001) habit planes were described in a number
of papers, e.g., [623–625]. Taking into account the results on faults observed
after deformation at high temperatures, specific climb dissociation reactions
were considered, e.g., in [616,626], leading to APBs and CSFs on (001) planes.
As a second process, the formation of grown-in condensation faults on (001)
planes is considered due to a loss of Si during the growth of MoSi
2
single
crystals at temperatures up to 1,600
◦
C [624,625]. These faults are bordered by
Frank type dislocations having a Burgers vector parallel to [001]. During the
in situ straining experiments described above, faults on (001) planes formed
instantaneously. These faults were studied in detail in [627].
The faults were created principally during in situ straining or heating in the
HVEM at temperatures between 440 and 1,000
◦
C and only if dislocations with
1/2111 Burgers vectors were present. These dislocations may originate either
from in situ loading or from a pre-deformation in macroscopic experiments.
Figure 9.47 presents an example of the formation of such faults during in situ
annealing at 630
◦
C. The faults are generated so quickly that the process itself
has never been recorded.
The displacement vector R
F
of the faults and the Burgers vectors of the
bordering dislocations have been analyzed by the methods of TEM diffraction
contrast, e.g., the contrast extinction rule (2.9) in Sect. 2.3. An example of a
fault bounded by dislocations A and B is presented in Fig. 9.48. The details
of the analysis are described in [627] including further figures. The analy-
sis clearly showed that the displacement vector R
F
equals either 1/6[001] or
1/2[001], which both characterize intrinsic stacking faults, and that the Burg-
ers vector of one dislocation bordering each fault is parallel to [001] but that
the other dislocation does not have the same Burgers vector, see in particu-
lar Fig. 9.48g where dislocation B is extinguished, but A is not. The Burgers
vector of the second dislocation should be parallel to [110]. Although the
whole analysis in [627] is not unambiguous, it is argued that the Burgers
vectors of the bordering dislocations are 1/2[001] and 1/2[110]. Thus, these
defects cannot be stacking faults, and produced by the condensation of vacan-
cies due to a loss of Si, which on all edges should be surrounded by Frank
partial dislocations. Besides, the instantaneous formation of the defects con-
tradicts the creation by a diffusion process. Since the generation of the defects
is bound to the presence of dislocations with 1/2111 Burgers vectors, the
best explanation of their formation is the dissociation of these dislocations
according to
1/2[111](1
¯
10) → 1/2[001](1
¯
10) + 1/2[110](001). (9.8)
The reaction leads to faults with a fault vector equal to the absolute value
of the Burgers vector of the Frank-type partial with R
F
= −1/2[001]. The
faults can be created either by climb of the Frank-type dislocation with a
1/2[001] Burgers vector, or by glide of the partial dislocation with the 1/2[110]