Efficient methodology for the accurate generation of customized compact model parameters from electrical test data

Abstract

Disclosed herein are embodiments of an automated, fast and efficient method of generating a customized compact model that represents a semiconductor device at the chip, wafer or multi-wafer level in a specific manufacturing environment. Specifically, measurement data is collected from a specific manufacturing environment and sorted by channel lengths. Then, an optimizer is used to generate customized modeling parameters based on the measurement data. The optimization processes is a multi-step process. First, a first set of modeling parameters is generated based on measurement data associated with a long channel length. Second, a second set of modeling parameters is generated based on the first set and on measurement data associated with a short channel length. Finally, the customized modeling parameters are generated based on both the first set and the second set. The customized modeling parameters are used to generate a customized compact device model representative of the specific manufacturing environment.