Advanced III-V / Si nano-scale transistors and contacts: Modeling and analysis

Seung Hyun Park, Purdue University

Abstract

The exponential miniaturization of Si CMOS technology has been a key to the electronics revolution. However, the continuous downscaling of the gate length becomes the biggest challenge to maintain higher speed, lower power, and better electrostatic integrity for each following generation. Hence, novel devices and better channel materials than Si are considered to improve the metal-oxide-semiconductor field-effect transistors (MOSFETs) device performance. III-V compound semiconductors and multi-gate structures are being considered as promising candidates in the next CMOS technology. III-V and Si nano-scale transistors in different architectures are investigated (1) to compare the performance between InGaAs of III-V compound semiconductors and strained-Si in planar FETs and triple-gate non-planar FinFETs. (2) to demonstrate whether or not these technologies are viable alternatives to Si and conventional planar FETs. The simulation results indicate that III-V FETs do not outperform Si FETs in the ballistic transport regime, and triple-gate FinFETs surely represent the best architecture for sub-15nm gate contacts, independently from the choice of channel material. ^ This work also proves that the contact resistance becomes a limiting factor of device performance as it takes larger fraction of the total on-state resistance. Hence, contact resistance must be reduced to meet the next ITRS requirements. However, from a modeling point of view, the understanding of the contacts still remains limited due to its size and multiple associated scattering effects, while the intrinsic device performance can be projected. Therefore, a precise theoretical modeling is required to advance optimized contact design to improve overall device performance. In this work, various factors of the contact resistances are investigated within realistic contact-to-channel structure of III-V quantum well field-effect transistors (QWFET). The key finding is that the contact-to-channel resistance is mainly caused by structural reasons: 1) barriers between multiple layers in the contact region 2) Schottky barrier between metal and contact pad. These two barriers work as bottleneck of the system conductance. The extracted contact resistance matches with the experimental value. The approximation of contact resistance from quantum transport simulation can be very useful to guide better contact designs of the future technology nodes. ^ The theoretical modeling of these nano-scale devices demands a proper treatment of quantum effects such as the energy-level quantization caused by strong quantum confinement of electrons and band structure non-parabolicity. 2-D and 3-D quantum transport simulator that solves non-equilibrium Green's functions (NEGF) transport and Poisson equations self-consistently within a real-space effective mass approximation. The sp3d5s* empirical tight-binding method is employed to include non-parabolicity to obtain more accurate effective masses in confined nano-structures. The accomplishment of this work would aid in designing, engineering and manufacturing nano-scale devices, as well as next-generation microchips and other electronics with nano-scale features.^

Degree

Ph.D.

Advisors

Gerhard Klimeck, Purdue University.

Subject Area

Engineering, Electronics and Electrical|Physics, Quantum|Nanotechnology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS