A Highly Efficient Statistical Compact Model Parameter Extraction Scheme

A new method of determining statistical compact model parameters is proposed. The variations of measured device characteristics can be efficiently translated into the variations of a set of transistor model parameters. Since the target of fitting is not the I-V curves of individual samples, but the statistically analyzed results of the I-V data, the extraction is fulfilled in only one optimization step. The method is applicable to any compact model platforms. Therefore, high accuracy and efficiency can be achieved at the same time.