A Constant Properties Model for the Performance Estimation in Segmented Thermoelectric Generator Elements and Its Experimental Validation Using an n-Type Mg2Si0.3Sn0.7–Bi2Te2.7Se0.3 Segmented Leg

Abstract

Segmenting multiple compatible thermoelectric (TE) materials can improve the conversion efficiency in thermoelectric generators (TEGs) compared to single-material TEGs. A model to accurately estimate the performance in segmented TEGs is essential for optimizing the geometric parameters. Constant properties models (CPMs) help in the quick estimation of the performance of TEGs working under large temperature differences using averaged material properties. This work discusses a CPM to predict the performance of single and segmented TEG elements using a combination of temperature-averaged and spatially-averaged TE properties. Spatial averaging requires the knowledge of the temperature profile across the TEG element length, which was obtained by solving the heat balance equation with a numerical integration approach as suggested by Ponnusamy et al. In this work, we develop the CPM for the segmented TEG elements (CPMS). The model predictions were validated experimentally from the performance evaluation of single Mg2(Si0.3Sn0.7)0.98Bi0.02 and segmented Mg2(Si0.3Sn0.7)0.98Bi0.02–Bi2(Te2.7Se0.3)0.98(CuI)0.02 TEG elements. Single and segmented TEG elements with Cu contacts were prepared by a hot-pressing technique. Both Mg2(Si0.3Sn0.7)0.98Bi0.02 and Bi2(Te2.7Se0.3)0.98(CuI)0.02 TE materials show peak zT > 1 in their measurement range. The experimental results show a maximum efficiency (ηmax) of ∼8.25% with 18% improvement for the segmented structure at ΔT ≈ 302 K. The improvement mainly originated from lower heat input (Qin) even though the maximum power output (Pmax) became lower by ∼30%. The performance prediction using the CPM matches well with <10% error for Pmax and ηmax for the single TEG element, and using the CPMS, the predictions were <10% error for Pmax and <25% error for ηmax for the segmented TEG element experimental results. This work demonstrates how CPMS can be used to predict device performance characteristics in segmented TEGs.