UC Berkeley

There is a strong interest in thermoelectric materials for energy production and savings. The properties which are integral to thermoelectric performance are typically linked, typically changing one of these properties for the better will change another for the worse. The intertwined nature of these properties has limited bulk thermoelectrics to low efficiencies, which has curbed their use to only niche applications. There has been theoretical and experimental work which has shown that limiting these materials in one or more dimensions will result in deconvolution of properties. Nanowires of well established thermoelectrics should show impressively high performance.

Tellurium is attractive in many fields, including thermoelectrics. Nanowires of tellurium have been grown, but with limited success and with out the ability to dope the tellurium. Working on previous work with other systems, tellurium was studied in dimethyl sulfoxide

(DMSO). The electrochemical system of tellurium was found to be quite dierent from its aqueous analog, but through comprehensive cyclic voltammetric study, all events were identified and explained. The binary antimony-tellurium system was also studied, as doping

of tellurium is integral for many applications. Cyclic voltammograms of this system were studied, and the insight from these studies was used to grow nanowire arrays. Arrays of tellurium were grown and analysis showed that by using DMSO, antimony doped tellurium

nanowire arrays could be grown. Furthermore, analysis showed that the antimony doped tellurium interstitially, resulting in a n-type material.

Measurements were also performed on arrays and individual wires. Arrays of 1.15% antimony showed ZT of 0.092, with the low ZT attributed to poor contact methods. Although contacting was an obstacle towards measuring whole arrays, single wire measurements were also performed. Single wire measurements were done by a novel method which allows for easy, reproducible measurements of wire properties. So far, this method has only been used to measure the properties of bismuth telluride, but can easily be adapted to any type of nanowire. Bismuth telluride nanowires were found to have electrical resistivities matching that of bulk (1.4*10^{-5<\super>&Omega-m), and thermal conductivities lower than bulk (<1 W/m-K).}