Numerical Simulation of Negative Differential Resistance Characteristics in Si/Si1¡xGex RTD at Room Temperature

Tao Li,Zhiping Yu,Yan Wang,Lei Huang,Cailan Xiang
2006-01-01
Abstract:Negative differential resistance (NDR) characteris- tics in the current-voltage curve of a p-type Si/Si1¡xGex resonant tunnelling diode (RTD) are simulated with the quantum hydrody- namic (QHD) model. An integrated difference scheme including Schafetter-Gummel (SG) method, second upwind method and second-order central difference method is used to discretize the QHD equations, which maintains both accuracy and stability. This work is the first to simulate hole transport in RTD using the QHD model. Investigations of some structure modifications have been carried out. Analysis of the results indicates that both quantum barrier thickness and hole effective mass have an impact on NDR characteristics for Si/Si1¡xGex RTD. The simulated peak-to-valley current ratio (PVCR) of 1.14 at T=293K agrees quantitatively with the experimental result when x=0.23. I. INTRODUCTION Circuits that use RTD hold promise as a technology for ul- tradense high-speed integrated digital devices. The NDR in I-V characteristic in RTD can be used to reduce the device counts per circuit function, thus increasing the circuit integration density (1). Many researches have been done on AlGaAs RTD and NDR characteristics are observed at room temperature. However, RTD based on SiGe can be made more easily than AlGaAs RTD, and it also has a good compatibility with the conventional CMOS processing technology. Simulation for a p-type double-barrier Si/Si1¡xGex RTD is performed in this paper, and an analysis on NDR is also included. For the modeling of ultra-small semiconductor devices, the QHD model was proposed by C. L. Gardner (2) from a moment expansion of the Wigner function. This model is a quantum corrected version of the classical hydrodynamic model, with O("h2) corrections to the stress tensor and energy density terms. This model has the advantages that it is com- putationally less expensive and expressed in terms of intuitive macroscopic quantities. In this case, macroscopic boundary conditions can be imposed. Moreover, a QHD model with a "smooth" potential has been derived in (3-4).
What problem does this paper attempt to address?