Towards end to end technology modeling: Carbon nanotube and thermoelectric devices

Shuaib Salamat, Purdue University

Abstract

The goal of this work is to demonstrate the feasibility of end-to-end ("atoms to applications") technology modeling. Two different technologies were selected to drive this work. The first technology is carbon nanotube field-effect transistors (CNTFETs), and the goal is to model device level variability and identify the origin of variations in these devices. Recently, there has been significant progress in understanding the physics of carbon nanotube electronic devices and in identifying their potential applications. For nanotubes, the carrier mobility is high, so low bias transport across several hundred nanometers is nearly ballistic, and the deposition of high-k gate dielectrics does not degrade the carrier mobility. The conduction and valence bands are symmetric (useful for complimentary application) and the bandstructure is direct (enables optical emission). Because of these striking features, carbon nanotubes (CNTs) have received much attention. Carbon nanotubes field-effect transistors (CNTFETs) are one of the main potential candidates for large-area electronics. In this research model, systematic simulation approaches are applied to understand the intrinsic performance variability in CNTFETs. It is shown that control over diameter distribution is critically important process parameter for attaining high performance transistors and circuits with characteristics rivaling those of state-of-the-art Si technology. The second technology driver concerns the development of a multi-scale framework for thermoelectric device design. An essential step in the development of new materials and devices for thermoelectrics is to develop accurate, efficient, and realistic models. The ready availability of user friendly ab-initio codes and the ever-increasing computing power have made the band structure calculations routine. Thermoelectric device design, however, is still largely done at the effective mass level. Tools that allow device designers to make use of sophisticated electronic structure and phonon dispersion calculations are needed. We have developed a proof-of-concept, integrated, multi-scale design framework for TE technology. Beginning from full electronic and phonon dispersions, Landauer approach is used to evaluate the temperature-dependent thermoelectric transport parameters needed for device simulation. A comprehensive SPICE-based model for electro-thermal transport has also been developed to serve as a bridge between the materials and device level descriptions and the system level simulations. This prototype framework has been used to design a thermoelectric cooler for managing hot spots in the integrated circuit chips. What's more, as a byproduct of this research a suite of educational and simulation resources have been developed and deployed, on the nanoHUB.org science gateway to serve as a resource for the TE community.

Degree

Ph.D.

Advisors

Alam, Purdue University.

Subject Area

Electrical engineering|Nanotechnology

COinS