Neural network modeling of GaAs IC material and MESFET device characteristics

This paper provides an overview of research focused on the utilization of neurocomputing technology to model critical in-process integrated circuit material and device characteristics. Artificial neural networks are employed to develop models of complex relationships between material and device characteristics at critical stages of the semiconductor fabrication process. Measurements taken and subsequently used in modeling include doping concentrations, layer thicknesses, planar geometries, resistivities, device voltages, and currents. The neural network architecture utilized in this research is the multilayer perceptron neural network (MLPNN). The MLPNN is trained in the supervised mode using the generalized delta learning rule. The MLPNN has demonstrated with good results the ability to model these characteristics, and provide an effective tool for parametric yield prediction and whole wafer characterization in semiconductor manufacturing.