Analytical current noise model based on nanoscale MOSFET.docVIP

Analytical current noise model based on nanoscale MOSFET 　　Abstract. In this paper， a physical understanding of excess noise component associated with transport mechanism in nanoscale MOSFET is developed. Based on the current flow image， current noise models valid for both uncorrelated and correlated carrier injection are derived. The variation of suppression-factors with source-drain voltage， gate voltage， temperature and source-drain doping are investigated. The results we obtained with considering the combination of the two effects are consistent with those from experiments and the theoretically explain is given. 　　Key words： Nanoscale MOSFETs， Current Noise， Fano. 　　While the drain thermal noise of long channel MOSFET agrees with the van der Ziel model[1]， considerably larger excess noise has been observed in MOSFET with channel lengths of deep sub-micron which severely hinder the applications of RF CMOS. Since the channel length of MOSFET continues to downscaling， it becomes an urgent need to address the issue of noise reduction. A clear understanding and accurate prediction of this excess noise play an integral role in realization of optimal noise behavior. 　　However， it is still an argument about the component of excess noise. Many researchers have pointed out that short channel effects[2-6] such as the hot electron effect， channel length modulation （CLM）， carrier heating and velocity saturation （mobility degradation） would lead to excess noise. Together with experimental measurements， they developed all kinds of analytical thermal noise correctional models based on the above short channel effects. Although the thermal noise correctional models can well describe the noise characteristic for deep submicron MOSFET， they are not valid for nanoscale MOSFET[7]. Some researchers attribute the failure of the thermal noise correctional models to the presence of shot noise[8，9]. Recent experiments， numerical simulations and theoretical results have proved the