Neural network modeling of GaAs IC material and MESFET device characteristics

This paper pro,ides an overview of research focused on the utilization of n eurocomputing technology to model critical in-process integrated circuit ma terial and device characteristics. Artificial neural networks are employed to develop models of complex relationships between material and device char acteristics at critical stages of the semiconductor fabrication process. Me asurements taken and subsequently used in modeling include doping concentra tions, layer thicknesses, planar geometries, resistivities, device voltages , and currents. The neural network architecture utilized in this research i s the multilayer perceptron neural network (MLPNN). The MLPNN is trained in the supervised mode using the generalized delta learning rule. The MLPNN h as demonstrated with good results the ability to model these characteristic s, and provide an effective tool for parametric yield prediction and whole wafer characterization in semiconductor manufacturing. (C) 1999 John Wiley & Sons, Inc.