SIMS PROFILE SIMULATION USING DELTA FUNCTION DISTRIBUTIONS

SIMS profiles of abrupt dopant distributions in semiconductor materials contain distrotions introduced by the measurement process. In quantitative depth profile studies, it is important to determine and remove these errors. We have used the idea that a given input profile may be represented by a series of adjacent delta-doped planes. Given a knowledge of the SIMS profile obtained from one single plane (the resolution function), it is possible to simulate the expected SIMS profile from a given input profile by the convolution with the resolution function. Experimental data have been obtained from special samples of molecular beam epitaxial GaAs containing both delta-doped planes and uniformly doped regions (1-20 nm thick). The dopants investigated were Si, Be and Al. It is shown that the simulated and experimental profiles of the uniformly doped regions show good agreement, thus demonstrating the validity of the method. We also present results of the convolution of an input function having exponential decays on the leading and trailing edges, which show the corrections that should be made to experimentally determined profiles. Additional simulations have been made with Gaussian inputs, which are relevant in dopant diffusion studies. Silicon diffusion lengths as small as 4 atomic planes (1.1 nm) have been deduced.