Speaker
Description
Scandium nitride (ScN) is a material with high electrical conductivity and a high thermoelectric power factor, making it interesting for thermoelectric conversion applications. Unfortunately, from this perspective, the relatively high thermal conductivity in its bulk form limits its global thermoelectric efficiency. To achieve a configuration that reduces the thermal transport, thin films of ScN are currently being intensely studied [1,2].
Recent experimental results show that the alternating interruption of the cathode potential during the reactive magnetron sputtering procedure, on magnesium oxide substrates, that is normally used to prepare ScN thin films [2], gave rise to twin domains in the structure of the deposited ScN thin films, having important effects on the measured thermoelectric performance [3].
Here, we propose a theoretical interpretation for these results by assuming that the alteration of the thermoelectric properties is mainly due to the presence of defected layers (i.e. with impurities) in the ScN structure, as a consequence of interruptions in the cathode potential, with vacancies in the crystal structure or the addition of oxygen atoms, replacing nitrogen.
We interpret these results by preparing simple atomistic models of these impurities (oxygen and vacancies) and studying their effects on the thermoelectric performance of the material, compared to similar models of pure ScN. Within the framework of the Landauer approach to quantum transport, the nonequilibrium Green's function formalism is used to model electronic transport as the probability of carriers (electrons or phonons) being transmitted from a semi-infinite source to a semi-infinite drain through a central conductor containing the impurity, which is treated as a scattering element [4].
The employed methodology is suitable for the calculations of all the structural properties affecting the thermoelectric performance, namely the Seebeck coefficient, the electronic conductivity and the thermal conductivity [5]. Here we limit our analysis to the thermoelectric power factor, resulting from the electronic structure of the systems.
[1]. P. Eklund et al., J. Mater. Chem. C, 4 (2016), 3905-3914.
[2]. J. More-Chevalier et al., AIP Adv. 9 (2019), 015317.
[3]. J. More-Chevalier et al., not yet published (2024).
[4]. M. B. Nardelli, Phys. Rev. B 60 (1999), 7828–7833.
[5]. L. Cigarini et al., J. Phys. D: Appl. Phys. 50 (2017), 395502.
