Equivalence and Accuracy of MOSFET Channel Length Measurement Techniques

It is shown that the MOSFET channel length measurement techniques of Terada and Muta, Peng et al., Whitfield, Suciu and Johnston, and De La Moneda et al. are actually equivalent, i.e. merely different expressions of the same formula for channel length in terms of measured resistance, and that some of the transresistance methods of Jain, although not equivalent, are also related to the same formula. The accuracy of this formula is evaluated for the general case and related to the error components due to source and drain resistance asymmetry, short channel geometry effect, and variation of series resistance with bias. No independent error component due to field-induced mobility degradation is found. Finally the errors in the methods of Terada and Muta, Chen et al., Sheu et al., Wordeman et al. and Jain, are determined and compared. The gate transresistance technique is found to be the most accurate method.