Atomic resolution electron microscopy was used to obtain images of moving dislocation kinks on partial dislocations at 600C. Video difference images were used to obtain direct estimates of the kink velocity. Observations of kink delay at obstacles (possibly O atoms at the dislocation core) yielded de-pinning energies and parameters for the obstacle theory of kink motion. The kink formation energy was deduced from the distribution of kink pair separations in low-dose images, and was compared with the kink migration energy. It was found that, contrary to the case of metals, kink migration rather than kink formation controlled the velocity of unobstructed dislocations. Measurements of the 30 partial furnished kink densities that were about one third of those for the 90 dislocation. The free energy of kink formation was 0.797eV. It was found that, contrary to the case of metals, kink mobility rather than formation was the rate-limiting step which controlled the motion of free dislocations.

H.R.Kolar, J.C.H.Spence, H.Alexander: Physical Review Letters, 1996, 77[19], 4031-4