A Quantitative Error Analysis of the Mobility Extraction According to the Matthiessen Rule in Advanced MOS Transistors

This paper presents a quantitative analysis of the errors produced by the Matthiessen rule in the extraction of the inversion-layer mobility in metal-oxide-semiconductor transistors. We show that the Matthiessen rule results in large inaccuracies in the mobility extraction, and most of all, it can lead to wrong trends, namely, to an incorrect dependence of the mobility on the temperature or the strain level. Consequently, when the Matthiessen rule is used to infer a given mobility component from the experiments, the inaccuracy of the extraction procedure can yield apparent discrepancies between experiments and simulations. Our results demonstrate that the mobility components extracted from the measurements by using the Matthiessen rule should not be regarded as experimental data because the extraction procedure relies on assumptions that are not fulfilled in most practical cases.