Image processing techniques have been used for decades in many branches of science. with the advent of low cost, highresolution CCD cameras and the advances in personal computing, techniques previously used in other disciplines are increasingly being applied by transmission electron microscopists. The present paper gives an example of using image processing techniques for characterizing the number and size of second phase precipitates in an oxide matrix.

Si inclusions in the form of Si precipitates can occur in silicon dioxide films. The inclusions are contained within the films and effectively reduce the local thickness of the oxide. This thinning results in a reduction in the voltage necessary to cause oxide breakdown; the larger is the precipitate, the lower the breakdown voltage. Knowledge of the precipitate size and density is therefore important when assessing the dielectric integrity of these films. The Si precipitates are crystalline and more or less randomly oriented within the matrix.