Analysis And Modeling Of Electrical Parameters For Nanoscale MOSFETs

As CMOS technology dimensions are being aggressively scaled to reach a limit where device performance must be assessed against fundamental limits, nanoscale device modeling is needed to provide innovative new MOS devices as well as to understand the limits of the scaling process. The objectives of this thesis are to examine and assess new features of carrier transport in futuristic nanoscale transistors and to implement the appropriate physics and methodology for nanoscale device modeling.To examine and assess new features of carrier transport in futuristic nanoscale transistors: As the essential physics of carrier transport departs that of conventional approaches because the effects of quasi-ballistic transport and quantum phenomena on device and circuit performance are becoming more important. Hence, it is of great interest to develop a new macroscopic approach for the simulation of nanoscale devices operating near the ballistic limit. In this study, we try to understand essential physics of quasi-ballistic transport and its implications to nanoscale device simulation based on macroscopic transport models.To implement the appropriate physics and methodology for nanoscale device modeling, which consists of four parts- threshold voltage, gate capacitance, mobility, and sub-threshold characteristic modeling: 1) For advance MOSFETs, it is very necessary to account for the velocity and the eletrical field distribution in two dimensions, especially in the condition that the transverse field (the gate field) and the longitudinal field (the drain field) are simultaneously high. In additions, the carriers do not distribute according the Feimi or Boltzmann function due to the high electrical field in channel. In this work, a methodology is presented for calculating mobility of nano-scaled MOSFET's from the Boltzmann transport equation(BTE). Approximate solution of the BTE for electrons in nano-scaled MOSFET's is given, and the improved distribution function of the carriers is used to model the mobility of carriers. A new model is presented for two-dimensional characteristic field-dependent mobility.2) As critical transistor dimensions scale below the 100 ran (nanoscale) regime, quantum mechanical effects begin to manifest themselves and affect important device performance metrics such as threshold voltage. Based on the modified triangular potential well and the first subband approximation, a physical-based model of MOSFETs threshold voltage is presented. Then, on the basis of WKB theory, the concept of "localization" was proposed to redefine the boundary of the first sunband, and converted the QM effects in the horizontal open boundary into the revision of vertical threshold voltage. The new model accounts for 2-D quantum mechanical effects for future generation MOS devices and integration circuits.3) With the scaling of MOSFET dimensions, the gate oxides become thinner. Due to the quantum mechanical effects (QME's), the carrier distributions in the silicon substrate and polysilicon electrodes play more important role for the gate capacitance. Based on improved triangular potential well approximation and least-squares curve fit, a simplified analytical model combined the impact of quantum effects in inversion and polysilicon gate regions is proposed. And we quantitatively explore the impact of quantum effect on gate capacitance.4) Sub-threshold behaviour of MOSFETs is of special interest at the present time because the operating voltages are lowering towards 1.5V and gate lengths are scaled down to nanometer. In this work, we present a new method for studying the subthreshold characteristic of nano-scaled MOSFETs called the regular perturbation method. Poisson Equation is solved using regular perturbation method for the first time. Especially, the depletion approximation and charge-sheet model in Poisson equation are avoided due to their invalidity in nano-scaled MOSFETs. These analytical models predict the subthreshold behaviors, similar to Medici simulation results. Therefore, our models are useful to predict the subthreshold behavior of nanoscale MOSFETs, and to give insights into device design and their scaling limits.