Résumé:
The improvement of the depth resolution of Secondary Ions Mass Spectrometry (SIMS) analysis
for the actual needs of microelectronics is necessary. In this work, we chose to improve the depth
resolution by signal processing techniques, in particular the deconvolution. A deconvolution
technique is proposed. It is able to produce regular solutions similar to those obtained by Van
Cittert with Tikhonov-Miller regularization and hard constraints algorithm (algorithm already
proved its effectiveness in SIMS) with a faster convergence. Moreover, by using the proposed
algorithm the gains in depth resolution and in maximum are better than those achieved using the
algorithm of Van Cittert regularized and constrained. Techniques to overcome the limits of the
partial deconvolution procedure are also proposed. It was shown that the amplification of the
noise can be controlled by filtering the measured profile before the application of the
deconvolution or by applying the partial deconvolution procedure through Tikhonov-Miller
regularization. The two problems amplification of the noise and appearance of negative values
in the deconvolved profiles can be treated by using iterative methods.
