Grain Rotation in Thin Films of Gold
- PDF / 2,201,311 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 106 Downloads / 255 Views
EXPERIMENTAL RESULTS Figure 1 shows a sequence of images taken between 250*C anneals. The images are from a region of the film in front of a growing void, which is seen at the bottom of the images in Figures Ic and d. Changes in grain orientation were apparent because of the changing orientations of the twins within the grains. The grain rotation observed in our specimens was recognized only after all anneals were completed, so diffraction patterns from individual grains are not available to determine changes in grain boundary misorientation. The changes in angles between twins in adjacent grains provides evidence of the rotation, but they are insufficient to fully determine misorientation because of a 1800 ambiguity relating to the sign of the (111) surface normal. We used the image analysis package NIH Image to measure the angle between twins in neighboring grains, as illustrated in Figure 1. Lines are drawn parallel to twin boundaries in grains 1 and 2. The angle between these twins decreases from 30* to 190 from Figure Ia to Figure Id. Eight grain boundaries are labeled in Figure 2. The angle between twins in the grains on either side
250nm Figure 1. Grain rotation during annealing of a gold thin film at 250 C. TEM images talen before annealing (a) and after annealing 1hour (b), 2 hours (c), 3 hours (d).
16
150nm Figure 2. Grain boundary labels for the plot inFigure 3. This is an enlargement of Fig. Id. of each boundary was measured after each anneal. All of these inter-twin angles are plotted in Figure 3. If only grain 2 were rotating, the angular changes for boundaries A-F would all be the same. The angles measured across boundaries A and C decrease monotonically, but angles for the other boundaries both increase and decrease during the anneals. The variations among these angles indicate that several grain boundaries experience changes in misorientations during the annealing process and that all of the grains were rotating relative to each other. All of the twins are identifiable from image to image and we surmise that the observed changes of misorientation occur by a rigid body rotation mechanism of some kind.
801 70,
60
"50,I
4140 E
A
IT 30 j
F
20 10 0 0
1
2
3
Anneal (hours)
Figure 3. Angles between twins in neighboring grains plotted versus annealing time in hours for eight grain boundaries. Letters refer to labels in Figure 2.
17
Other changes in the microstructure accompany this grain rotation. The most obvious is the change in the geometry of the triple junction between grains 1, 2, and 3. The triple junction angles are plotted in Figure 4 where an angle is named with the same letter as the boundary opposite it. The variation in angle A is within the measurement error of 2', but angles B and G change drastically in a complimentary fashion. Angle B decreases from Figure la to Figure lc, but increases from Figure Ic to Figure Id. The angles formed by the boundaries at the triple point at the other end of boundary A maintain more constant values. Their maximum angular variation is 170
Data Loading...
