It is known that high-resolution electron microscopy (HREM) can provide quantitative information on the properties of crystalline materials. The HREM patterns of layered structures of III-V semiconductors vary with the chemical composition of the latter within the sublattices, which is also influenced by interdiffusion. Local variations of the crystal cell similarity are recorded for image analysis and compared with templates of known material composition.
Of the advanced theories of data interpretation, the now well-established fuzzy logic is highly suited for corresponding image processing techniques. Combining neighbouring image cell similarities, the underlying chemical composition is evaluated by applying fuzzy logic criteria of inference to masks of about 1 nm×1 nm in size.
The new approach can be used to localize regions of significant changes in composition, i.e. edge detection, and to determine the composition across the interface region. The methods introduced prove successfully applicable to simulated as well as to experimental images of AlAs/AlxGal-xAs. Similarity/composition relations of nonlinear as well as nonmonotonic characteristics are studied to establish an alternative fuzzy logic approach.