poster-icon
Description
Evan Witkoske1 Yining Feng1 Xufeng Wang1 Mark Lundstrom1 Na Lu1

1, Purdue University, West Lafayette, Indiana, United States

The potential benefits of new thermoelectric materials can be drastically reduced by a failure to optimally dope the material to its zT maximum point. When analyzing experimental data of an arbitrary material, the important question is: Can we determine the optimal doping concentration to reach the maximum zT in the given material? A procedure to do so under the assumption of parabolic energy bands has recently been reported [1]. In this paper, we use a technique that makes use of the full, complex band structure obtained by DFT simulation and apply it to experimental results for CuAlO2.

An integrated computational and experimental study using the 2H phase of CuAlO2 as a model material is presented. The 2H phase of CuAlO2 has gained interest as a promising metal oxide candidate for high temperature p-type thermoelectric applications [2-4]. Using a method similar to one that has recently been reported but without assuming parabolic energy bands, we will show how a complex TE material can be computationally assessed and how experimental data can be analyzed using first-principles informed calculations to answer the question posed above.

Experimentally, 2H phase CuAl1-xFexO2 (0≤x≤0.5) nanobulk [5] were synthesized using solid-state methods, and their thermoelectric properties will be measured using temperature dependent characterization tools. Specifically, electrical properties will be obtained by Van der Pauw method, thermal conductivity will be measured by Laser Flash Method using Netzsch LFA 567, an the Seebeck will be obtained by a home-built high temperature system. From there we can extract the Fermi level and subsequent theoretical doping concentration that maximizes zT for this material. Optimized doping concentrations will be achieved using Al and Fe as dopants through solid-state methods to achieve the optimized zT as predicated using our computational approach. The experimental data will be compared to the modeling work to refine our model for better experimental design guidance. Full band, first principles informed calculations will be compared to calculations that use parabolic energy bands [1], and the differences will be discussed.

[1] S. D. Kang and G.J. Snyder, arXiv:1710.06896v1, 2017.
[2] X. G. Zheng, K. Taniguchi, A. Takahashi, Y. Liu, C. N. Xu: Appl. Phys. Lett. Vol. 85 (2004),
p.1728
[3] W. T. Liu, Y. Y. Luo, Z. T. Liu, Z. M. Wei, "Density Functional Theory Study of P-Type
Transparent Conducting 2H-CuAlO2 Oxide", Applied Mechanics and Materials, Vol. 252,
pp. 263-266, 2013
[4] Y. Feng, X. Jiang, E. Ghafari, B. Kucukgok, C. Zhang, I. Ferguson, and N. Lu, “Metal Ox- ides for Thermoelectric Power Generation and Beyond,” Adv. Comp. Sci., 2017.
[5] Q. Hao, D. Xu, N. Lu, and H. Zhao, “High-throughput predictions of nanoporous bulk
materials as next-generation thermoelectric materials: A material genome approach,”
Phys. Rev. B, 93, 205206, 2016.

Tags