Direct measurement of thin-film thermoelectric figure of merit

Abstract

Thermoelectric materials are playing an increasingly important role in a technologically advanced and resourcelimited world. An important application of thermoelectric materials is in direct thermal-to-electrical energy conversion. Because a substantial amount of energy in our world is in thermal form, there are tremendous applications for thermoelectric devices that efficiently convert thermal energy into electricity. The energy conversion efficiency of a thermoelectric material is a function of its dimensionless figure of merit that is defined as ZTS 2 T /, where S, , , and T are the material’s Seebeck coefficient, electrical conductivity, thermal conductivity, and ambient temperature, respectively. The numerator of ZT is known as the thermoelectric power factor P. Recent work to improve the ZT of materials designed for high-temperature power generation has focused on thin films due to the freedom to engineer the materials at the nanoscale. In bulk thermoelectric materials, there is a dramatic tradeoff between S and free carrier concentration n. Therefore, n is chosen to optimize P at a particular T. Efforts have been made to decouple S and n to increase the P of thin films beyond bulk values using low dimensional structures or with electronic energy filtering for large n in tall barrier heterostructures. 1 However, the majority of improvement in ZT has been achieved through the reduction in due to interfacial phonon scattering. An extension of this concept was recently demonstrated in isotropic nanoparticle thin films that exhibit a that is a factor of 2 below the bulk alloy limit due to enhanced scattering of longer wavelength phonons. 2 This reduction in is achieved while enhancing due to an increase in n by the semimetallic nanoparticles. Nanostructured thin-film research offers great potential to increase the ZT of thermoelectric materials. Accurate characterization of thin film thermoelectrics is an important step in the development of these materials. In order to evaluate the ZT of thin films, we utilize direct crossplane ZT measurement using the Harman method. 3 The Harman method utilizes measurements of transient voltages in a thermoelectric device to extract the resistive voltage drop and the Seebeck-induced voltage drop slower thermal response. Further study of the Seebeck voltage under opposite polarities permits the separation of the Joule effect-induced and the Peltier effect-induced Seebeck voltages. However, the Harman method is quite difficult to apply to thin films. Due to the fast thermal response of thin films, high-speed detection of the small transient Seebeck voltage is necessary. 4 In addition, because this method relies on closedcircuit electrical measurements, electrical parasitics such as resistances in series with the material under investigation will result in a reduction in the measured ZT relative to the intrinsic material value. Additionally, the side of the measurement device that is not heat-sinked must be free of any parasitic heat load, making electrical probing difficult for the requirement of uniform device current injection. Specific device geometries are used to obtain accurate measurement of the cross-plane ZT of InGaAlAs thin films with embedded ErAs nanoparticles Fig. 1. The devices are