In-situ ellipsometry solutions using a radial basis function network

The drive to improve semiconductor device performance has led to a need for advanced semiconductor materials and complex multilayer structures. Fabrication of such devices requires precise control of semiconductor layer parameters such as thickness, composition and interface sharpness. Calculation of thickness and composition from nondestructive ellipsometric measurements is an ill-posed problem whose solution, until recently, has only been capable ex-situ using cpu intensive nonlinear model fitting techniques, making real time control of semiconductor growth difficult. This paper describes the development of a radial basis function neural network to monitor the composition of the top 10 /spl Aring/ of the growing semiconductor. The network is assessed on theoretically modelled structures and a limited set of real structures. Promising initial results are obtained which highlight several technical aspects to be addressed before an operational system can be developed.