Modeling of Unified Channel Mobility for Quantum Hydrodynamic Simulation of Nanoscale MOSFETs

University Of Massachusetts Amherst, Amherst MA

Investigators

Abstract

The CMOS technology is now being ushered from l8Onm node to 13Onm node. For lOOnm-l3Onm generation technology, many key issues of transistor design must be reexamined. One of them is non-equilibrium transport which takes place in nonoscale MOS devices. Because of continuous down-scaling of the device, the gate length now becomes comparable to or smaller than the inelastic mean-free-path. In this regime, the non-equilibrium /non-local transport of carriers in MOS devices requires substantially different formulation from that of the conventional one. In this research project, it is proposed to construct a unified mobility model for carriers in the channel of bulk, SOI, thin-body SOI, and DG MOSFETs applicable to gate length below lOOnm. Based on sound physics of non-equilibrium transport including quantum effects, the conventional low-field inversion layer mobility will be extended to the high-field regime by seeking an effective average carrier energy which characterizes the non-local transport occurring in the device. A series of self-consistent Monte Carlo (MC) simulations of carriers in inversion layers with a quantum mechanically corrected potential will be carried out in the region encompassing retarding-, low-, and high-fields. A unified channel mobility will be constructed in such a way that it can accurately predict velocity overshoot at the drain end of the channel, quantization effects in the inversion layer including tunneling, as well as thermionic emission-limited current density across the source-channel barrier. This unified channel mobility will then be incorporated in the quantum hydrodynamic (QHD) transport equations for numerical simulation of nanometer-scale bulk, SOI, thin-body SOI, and DG MOSFETs. It is expected that this research will provide a consistent and easy means of moving the simulation hierachy from drift-diffusion to MC Boltzmann via a hydrodynamic (HD) formulation. The project will enable simulation of MOS devices with the channel length of 100nm and below. Such capability is required of simulation of next generation of integrated circuits, which are expected in a few years. The proposed research will also benefit both graduate and undergraduate students on campus and engineers in industry who take P.I.'s device simulation course through the distance learning. The P.I. has constantly updated course materials based on research results of the past NSF projects and other research projects supported by the industry.