NEURAL-NET COMPUTING FOR INTERPRETATION OF SEMICONDUCTOR FILM OPTICALELLIPSOMETRY PARAMETERS

Optical ellipsometry has been found to be a promising technique for mo nitoring process parameters, such as film composition and film thickne ss, of semiconductor wafers grown with molecular beam epitaxy. Whereas it is a straightforward task to calculate ellipsometry angles given t he thickness of the film and the refractive indexes of the film and su bstrate, it is a difficult task to invert that mathematical relationsh ip. However, the process must be inverted if the measured parameters a re to be interpreted meaningfully in terms of film composition and fil m thickness. This paper reports on the use of neural-net computing for the inverse mapping of measured ellipsometry parameters, We used a fu nctional-link net which is very efficient in function approximation. T he advantage of using the net, however, is not only its speed, but als o because some other net architecture characteristics allow us to perf orm the task in a holistic manner.