Changes in the Burgers Vector of Perfect Dislocation Loops without Contact with the External Dislocations

We report the observations of a new type of changing process in the Burgers vector of dislocations by in situ transmission electron microscopy. Small interstitial-type perfect dislocation loops in bcc iron with diameters less than approximately 50 nm are transformed from a $1/2⟨111⟩$ loop to another $1/2⟨111⟩$ one or an energetically unfavorable $⟨100⟩$ one; furthermore, a $⟨100⟩$ loop is transformed to a $1/2⟨111⟩$ one. These transformations occurred on high-energy electron irradiation or simple heating without contact with external dislocations. The origin of these phenomena is discussed.

Figure 1

Spontaneous change in the projected direction of the motion of an interstitial-type dislocation loop along the observation axis of [1 0 5] upon 1000-keV electron irradiation with a beam flux of $1\times {10}^{24}{e}^{-}/{\mathrm{m}}^2\mathrm{s}$ at 293 K. The reflection $\mathbf{g}=020$ was adopted. The moving direction of the marked loop—indicated by arrows and additional lines—spontaneously changes from $[110]$ to $[1\overline{1}0]$ at 1.50 s. This spontaneous change corresponds to that in the Burgers vector of the loop from $1/2[11\pm 1]$ to $1/2[\overline{1}1\pm 1]$.

Figure 2

Spontaneous change in the TEM image of an interstitial-type dislocation loop upon heating at 570 K. The beam incident direction was $[\overline{1}05]$. The reflection $\mathbf{g}=0\overline{2}0$ and deviation parameter $s=0.1{\mathrm{nm}}^{-1}$ were adopted. The foil thickness was approximately 100 nm.

Figure 3

Comparison among the TEM images of the dislocation loop before and after the change shown in Fig. 2 and generated images of the circular loops of various types. Panels (a) and (b) are the experimental images of the loop: (a) image before the change; (b) image after the change. Panels (c)–(l) are generated images. The Burgers vector, habit plane, and diameter of the loops for the calculations are as follows: (c) $1/2[\overline{1}11](1\overline{1}0)$, 12 nm; (d) $1/2[\overline{1}11](3\overline{3}\overline{1})$, 11 nm; (e) $1/2[\overline{1}11](2\overline{2}\overline{1})$, 11 nm; (f) $1/2[\overline{1}11](1\overline{1}\overline{1})$, 11 nm; (g) $1/2[\overline{1}1\overline{1}](1\overline{1}0)$, 12 nm; (h) $1/2[\overline{1}11](3\overline{3}\overline{1})$, 8 nm; (i) $[010](2\overline{2}\overline{1})$, 12 nm; (j) $1/2[\overline{1}1\overline{1}](1\overline{1}1)$, 11 nm; (k) $[010](0\overline{1}0)$, 10 nm; (l) $1/2[\overline{1}11](0\overline{1}0)$, 10 nm. In the calculation, the deviation of the observation axis from the $[001]$ pole and the absorption of electrons were taken into account. The central position of the loop was located at the center of depth of the specimen. The diameters of the loops in (k) and (l) satisfy a condition that the major length of their coffee-bean images is equal to that of the experimental image of the loop after the change shown in (b). The diameters of the loops shown in (c)–(j) except (h) were set so that the number of SIAs contained within these loops was equal to that in the loop shown in (k). The diameter of the loop shown in (h) was set so that the number of SIAs in the loop was the half of that in the loop shown in (k).