314 p. B. HIRSCH, R. W. HORNE AND M. J. WHELAN 



Table 1. Siininiary of results of boundary misorientation measurements in three cases. 



Plate 

 numbers 



Normal 

 to foil 



Type of 



boundary 



assumed 



Rotation 

 axis 



Component 

 angle of 

 rotation 



Calculated 



dislocation 



spacing 



Observed 

 approximate 



spacing of 

 dislocations 



Resolution 



94, 95 



[001] 



293, 294, 296 [001] 

 299, 300, 302 [001] 



(110) 

 Simple tilt 



(211) 

 Mixed tilt 



Hexagonal 

 network 

 near (111). 

 Twist 

 boundary 



[112] 



[Oil] 

 [111] 



2.1° 



1.36 

 0.35 = 



80 A 



94 A 



270 A 



130 A Poor 



100 A Good in places 

 300 A Fair 



Dark field experiments, in which the 30 // objective 

 aperture is moved to accept several Bragg reflections 

 in turn, have shown that the contrast is reversed, 

 and that the dislocations and boundaries are illu- 

 minated usually in several Bragg reflections, whereas 

 strain free areas show up in only one or two reflec- 

 tions. The contours visible at A and B in fig. 1 are 

 extinction contours first studied by Heidenreich (6). 

 It is thought that these may be due to bending in the 

 foil produced by large range strains in otherwise 

 perfect grains. 



Arrangement of dislocations. — The direct observa- 

 tions of dislocations in the subgrain boundaries en- 

 able dislocation theory to be checked quantitatively. 

 Foracomplete discussion of networks and boundaries 

 in face-centred cubic crystals reference should be made 

 to Frank (3), Ball and Hirsch (2), and Amelinckx ( I ). 

 Dislocation theory predicts that simple boundaries 

 may be considered as surface arrays of dislocations. 

 A probable example of a simple (1 10) tilt boundary 

 has already been given in fig. 2. This type of boun- 

 dary contains dislocations of one of the twelve cubic 

 slip systems only. More complicated boundaries 

 may contain dislocations from two or more slip 

 systems. Fig. 3 (a) shows an example of a simple 

 twist boundary on (100), consisting of a crossed 

 grid of screw dislocations. The crystallographic 

 orientation of this field of view was determined by 

 diffraction. The normal to the foil was close to [100] 

 and the lines run parallel to [110] directions. The 

 network therefore makes a shallow angle with the 

 surface. It can be seen at A that dislocation lines 

 end on the surface. 



Another network, predicted theoretically (3) and 

 observed extensively in AgBr and NaCI crystals 

 (1, 5), is a hexagonal network of screw dislocations 

 lying in a (111) plane, forming a twist boundary. 

 Fig. 3 (b) shows a small piece of hexagonal network 

 in aluminium. Large hexagonal networks are not 

 observed, presumably because the foil is thin and 

 areas of [111] orientation are rare because of the 

 preferred orientation. 



Experiments have been performed to check the 

 theory quantitatively by microdiffraction by selecting 

 a small area across a subgrain boundary. From the 

 splitting of the diffraction spots it is possible to 

 obtain the component misorientation angle about 

 the axis of the instrument. A high resolution micro- 

 graph often enables the type of boundary to be de- 

 termined. This is possible when the resolution is 

 particularly good and when interference effects pro- 

 duced by the overlap of crystals at the boundary are 

 absent. These interference effects produce fringes at 

 the boundary, which tend to mask partially or com- 

 pletely the dislocation network. Table 1 contains a 

 summary of the results obtained in three cases. The 

 dislocation spacing, calculated from the observed 

 misorientation angle and the assumed network, is 

 compared with the approximate observed spacing. 

 In these three cases the type of boundary present is 

 only tentatively proposed. In the first two cases the 

 boundaries were inferred from the trace of the boun- 

 dary plane. In the third poorly resolved hexagons 

 could be seen. It is seen that there is agreement to a 

 factor of 1.5. This is evidence in favour of disloca- 

 tion theory, and shows that the boundaries do con- 

 tain dislocations of unit Burgers vector. 



Movement of dislocations. — When working under 

 fine focus illumination conditions with the double 

 condenser lens, dislocations are observed to move. 

 The reason for the movement is not at present clear, 

 but it is thought that it may be due to a combined 

 effect of heating and straining due to thermal gra- 

 dients in the specimen. The temperature rise in the 

 object is not known.Thefirst indication of movement 

 is that certain lines bow out. This is presumably 

 direct confirmation of the mechanism suggested for 

 the decrease in elastic modulus due to dislocations 

 (11). Two types of movement can be observed. The 

 first is slow and jerky, probably requiring consider- 

 able cross-slip and climb, which possibly corresponds 

 to creep. 



The second type of movement is very fast and 

 usually it is only possible to observe the appearance 



