Stuart Wright, EDAX and Ryan Larsen, Brigham Young University
Stuart Wright of EDAX and Ryan Larsen of Brigham Young University have developed an automated method for discriminating between coherent and non-coherent twins in orientation imaging microscopy (OIM) scan data, extending some earlier semiautomatic work of Randle (Randle, V. (2001). Scripta Materialia, in press.)
For a boundary to be considered a twin boundary, the misorientation across the grain boundary must be very near the twin misorientation relationship. For example, the primary recrystallization twin in FCC materials can be described as a 60 degree rotation about a <111> crystal axis. Thus, a boundary segment which has a misorientation of 60.7 degrees about a <10 10 11> axis could be considered a twin. However, for a boundary segment to be considered a coherent twin, it must satisfy an additional requirement. The boundary plan must coincide with the twinning plane.
For the example already given, this means the 111 planes of the crystals on either side of the grain boundary must be aligned (within a given tolerance) with the grain boundary plane. Since OIM scans are inherently two dimensional; it is not possible to determine whether a given boundary satisfies this criterion without performing serial sectioning or some other three-dimensional sampling technique. However, a partial check can be made: whether the trace of the boundary plane is aligned with the trace of the twinning plane. Randle has shown that such a check is valid in three dimensions 90% of the time. Since the boundary segments in an OIM map follow the scan grid, reconstructed boundaries are used for making the alignment check between the boundary and twin plane traces.
Using this methodology, the impact of twin boundaries on grain size distributions can be more accurately determined. The following figure shows results for copper interconnect lines.