Small-signal modeling of RF CMOS

Continuous scaling of CMOS technology has now reached a state of evolution, in terms of both frequency and noise, where it is becoming a serious contender for radio frequency (RF) applications in the GHz range. While complete broadband characterization and accurate modeling of the MOSFETs are critical requirements for circuit designs, the RF behavior and physics are not well understood. This dissertation explores the physical origins of small-signal characteristics of RF MOSFETs and discusses many of the challenges facing the CMOS RF circuit designer in terms of device modeling and characterization.; A distributed NQS model is discussed and an estimation of the theoretical limit up to which quasi-static MOSFET models are reasonable is presented. The NQS frequency (fNQS) is defined as the frequency when NQS effects start to occur. Bias dependency of the NQS frequency is examined for critical modeling parameters (i.e., Cgg, C gs, and Cgd); the strong bias dependency of the parameters is due to the charge sharing between source and drain, and mobility. At high frequencies, the existence of substrate network cannot be ignored. The substrate network is modeled as parallel Rsub and Csub, and the product (RsubC sub) shows conservative behavior (i.e., dielectric relaxation time constant of Si). The existence of substrate resistance impacts terminal admittances, even at low frequencies, and this fact suggests that the QS model needs to include substrate components in order to achieve accurate terminal characteristics. In addition, the negative capacitance (NC) effect of a diode is discussed and the corresponding compact model is developed. Instead of analyzing a small-signal model from measurement data alone, two-dimensional simulation is employed in order to develop a physically acceptable small-signal model of RF MOSFETs.; When dealing with small-signal equivalent circuits, the extraction of parameters is of great importance. Accurate extraction of parasitics is critical in RF circuit simulation and accurate modeling of intrinsic device. A compact model of MOSFETs for the common source configuration is developed to identify two-port characteristics and the model is used to effectively de-embed extrinsic parasitics with good accuracy.