Yield determination and design optimization of GaAs MESFETs using a physical simulator

The authors have enhanced TEFLON, the physics based large-signal GaAs MESFET model, to predict yield of a GaAs MESFET design before fabrication. Device acceptance criteria include large-signal RF operating characteristics such as maximum power added efficiency and power out at 1 dB gain compression. SUPREM 3.5 has been incorporated into TEFLON so that optimizable parameters include processing specifications for material deposition, ion implantation, and annealing as well as physical parameters such as device geometry. The authors have incorporated Monte Carlo techniques to statistically treat disturbances of the physical and processing parameters. Finally they have integrated the yield estimation modules with an optimizer so a design can be centered for maximum yield in the presence of process disturbances.<>