Is SEM Noise Gaussian?

Noise in the scanning electron microscope (SEM) is defined as being the random fluctuations which occur in the signal observed from a particular pixel in the image even under conditions where the incident beam, the sample, and the recording conditions are kept constant. This noise is then the result of the fact that electron production from the gun, and electron interactions with the specimen are statistical in nature and consequently are different for every individual electron. It is usual to assume that the noise satisfies Gaussian statistics. Thus if the mean number of electrons emitted from a given pixel is N, then the standard deviation of the signal, and hence the noise component, is N . Since noise significantly affects the predicted performance of an electron-beam tool, especially in applications such as defect detection during semiconductor fabrication, it is necessary to test this assumption. In addition there is increasing interest in simulating SEM images for various purposes, and the need to incorporate realistic noise in such cases also raises the question as to its exact nature.