Experimental Characterization of the MOSFET Fano Factor at Cryogenic Temperatures for Accurate Cryo-CMOS RF Modeling

Multiple independent researchers have hypothesized that MOSFET channel noise can be described down to deep cryogenic temperatures using a temperature-independent suppressed shot noise model containing just one fitting parameter, the Fano factor $\digamma $ . However, experimental data to back-up these claims have not been available to date. In this article, we describe a carefully constructed experiment through which we are able to provide such data. Using a combination of room temperature and cryogenic MOSFET measurements in conjunction with cryogenic amplifier measurements, we extract $\digamma $ at both room temperature and 8K for transistors from 22 nm FDSOI, 45 nm PDSOI and 65nm bulk CMOS processes over a range of current densities. We find that, as predicted, $\digamma $ does indeed appear to be temperature-independent. The implications of this finding are described, both in terms of the transistor noise parameters as well as the parameters used in conventional MOSFET channel noise models.