Energy technology development is guided by international goals for sustainable energy supply to overcome the environmental crisis. Thermoelectric (TE) energy conversion is one among various approaches to feed global energy needs. It is a versatile option for harvesting and recovering waste heat by direct conversion of thermal into electrical energy, having important advantages such as absence of any harmful emission and moving parts. However, there are still challenges in the TE device technology. A crucial one is the contact between the TE material and the metallic bridge to build up a functional module for energy conversion, namely a Thermoelectric Generator (TEG) [1].

With the aim of bringing together the gained experimental knowledge and a robust computational approach, we apply Integrated Computational Materials Engineering (ICME) [2] to accelerate and transform the development of contact solutions as a critical step of the module making. As presented in the Figure, ICME enables integration of material knowledge encoded in databases and materials processing to overcome the current status based on mainly empirical trial.

The challenge lies in the intrinsic interdisciplinarity necessary for designing the contacts, a field that develops multiple intermetallic layers as well as new interfaces and has to satisfy mechanical performance without hindering the functional stability of the TE material. This is addressed in this work by establishing an integrated multiscale computational methodology for the particular case of Mg-Si-based compounds, which are suitable TE materials for high temperature applications with non-toxic, abundant and inexpensive elements. Specifically, the compositions Mg₂(Si_1−xSn_x) x = 0.6, 0.7 and a Cu electrode were selected to present our approach based on thermodynamic [3], kinetic and mechanical modelling using the CALPHAD, first-principles thermodynamics and a finite element framework as implemented in the Thermo-Calc and Ansys packages. Diffusion profiles, interfacial energies estimation, prediction of growing layers, thermomechanical stresses among other relevant properties were calculated and confronted with experiments. These results represent an important step towards achieving a customized full integration strategy for efficient TEG design with data transfer/storage enabling software interoperability compatible with nowadays autonomous laboratory workflows. 

References

[1] Ying, P., He, R., Mao, J. et al. Nat Commun, 2021, 12, 1121.

[2] Xiong, W., Olson, G. npj Comput Mater, 2016, 2, 15009.

[3] S. Tumminello, S. Ayachi, S. G. Fries, E. Müller, and J. de Boor J. Mater. Chem. A, 2021, 9, 20436-20452.