Multi-Spectral imaging in materials microanalysis

Multi-spectral imaging in materials microanalysis is reviewed, with illustrative results from scanning Auger microscopy and energy-dispersive x-ray imaging. A new approach to defining the signal-to-noise ratio for multi-spectral images suggests that the number of resolvable signal levels increases multiplicatively with the number of orthogonal dimensions containing the signal. As the collected signal increases only as the sum of the number of dimensions, this can lead to considerable advantages of multi-spectral over single-channel imaging. One disadvantage of multi-spectral imaging is the large amount of data that are collected. However, in practice, most of the signal in an image can be represented in a few dimensions using principal component analysis. An example of the use of principal component analysis shows that the information in a data set from an electron microscope image of seven energy-dispersive x-ray images can be viewed in two or three dimensions and can thus be efficiently presented on a windowed graphics screen.