GOALI: Integrated Circuit Silicon Nanowire Thermoelectric Generators for On-chip Micropower Generation

University Of Texas At Dallas, Richardson TX

Investigators

Abstract

Thermoelectric generators are of great scientific and technological interest to society because they can recycle some of the unwanted waste heat generated from the operation of machinery, electronics, living organisms, etc. back into useful electrical power. One of the major challenges limiting widespread use of thermoelectric devices is the difficulty in adapting materials having excellent thermoelectric properties into industrially and economically scalable production. The ideal would be to make thermoelectric generators in an industrial production manner similar to what is done with silicon in commercial integrated circuit technology. Recently it was discovered that silicon in the form of a very small wire (called a nanowire) has good thermoelectric characteristics and can be potentially integrated with existing silicon electronics. The main goal of this project is to determine how to fulfill that potential and pioneer a route towards integrating silicon nanowire based thermoelectric generators having useful performance characteristics into existing industrial silicon technology. The development of efficient silicon integrated circuit thermoelectric devices scalable to industrial production will open up numerous applications of immediate and future benefit to industry, including improving energy efficiency in large scale electronics by harvesting waste heat, regulating temperature build up in high power electronics, and establishing an environmentally ¡°green¡± way to power microelectronic circuits. This research also provides a uniquely valuable cross-disciplinary training opportunity for undergraduate and graduate students. This project will expose students to an industrial perspective and approach to research, which will be particularly valuable to both students and society when they enter the workforce. Integrated circuit thermoelectric (TE) generators (IC TEG) have compelling importance as an on-chip, small form factor way to energize or actuate microcircuits and sensors or regulate on-chip temperature in a localized manner. Most current research in TEGs focuses on developing materials having a high TE figure-of-merit ZT. Widely studied high ZT materials such as Bi2Te3 have ZT slightly larger than 1 but are incompatible with industrial Si processing and thus are difficult to incorporate as microelectronic TEGs in a commercially scalable manner. Recently it was found that Si nanowires (SiNWs) can have ZT up to 0.6. While this opens a route towards silicon TEGs, the ZT of SiNWs is highly sensitive to fabrication, being typically ¡Ü 0.3 in practical circumstances. However, how well a TEG converts temperature difference into electrical power tends to be limited not only by ZT but by parasitic device impedances that degrade heat input and electrical power output. These impedances can be designed and controlled in Si processing, so the performance of SiNW TEGs could be made competitive with higher ZT materials by carefully optimizing the electrical and thermal impedance matches. This project seeks to maximize SiNW based IC TEG voltage/power generation not simply through a materials approach but by understanding, controlling, and optimizing TEG design, parasitic losses, and impedance matching aspects. The SiNW TEG circuits to be studied were fabricated with baseline industrial silicon technology and thus can be integrated into larger circuits in a highly compact, monolithic, and controlled manner. The main goal is to demonstrate SiNW IC TEGs useful to the microelectronics industry and fabricated in a commercially scalable manner.