Enhancing eCHORD Orientation Mapping by Means of a Derivative Approach.

Abstract

The core indexing engine of the electron Channelling ORientation Determination (eCHORD) technique has been modified to significantly minimize indexation errors in the resulting orientation maps, thus enhancing accurate microstructure characterization. Orientation mapping using eCHORD involves acquiring a sequence of backscattered electron images by rotating the region of interest (ROI) within the scanning electron microscope, the sample being tilted in the range ∼10°-15°. Such an image series forms a data-cube from which intensity profiles can be extracted at each position in the ROI, reflecting the variation of the backscattered electron signal as a function of sample rotation. These profiles are compared to a database of theoretical profiles obtained by dynamical diffraction simulations. Here, the methodology consists of emphasizing the importance of diffraction peak positions in these profiles, by differentiating the data extracted from the backscattered electron images rather than using them in their raw form. The implementation of the profile derivative significantly improves the correspondence between experimental data and its theoretical counterpart. The characterization of sub-micron twin boundaries in a copper interconnect on a microelectronic chip provides insights into the effects of this indexing refinement. A comparison with Electron Back Scattered Diffraction data on a twinned nickel-based sample confirms the results.