Bismuth telluride (Bi2Te3)-alloy thermoelectric generator modules with high average figure of merit still convert heat to electricity at efficiencies well below the theoretical estimate for their own leg materials, and their reported efficiencies cannot be compared across laboratories because heat-flow metrology for large modules is unstandardized. The source demonstration (Nozariasbmarz et al., 2020) achieved a record 8% single-couple efficiency at a 230 °C temperature difference but measured ~15% below ideal theory, and its authors identified parasitic losses, contact resistances at leg/metal and metal/header junctions, internal leg temperature uncertainty, and the absence of accurate standardized heat-flow measurement for large multi-couple modules as unresolved device- and metrology-level problems. Later literature (2021-2025) provides partial progress on contact barrier layers and measurement procedures but leaves all of these problems open. The problem is to quantify and suppress the device-level loss channels so that measured module efficiency approaches the theoretical limit for the same leg materials, to establish an accurate, uncertainty-quantified, transferable metrology for module efficiency, and to demonstrate that these gains survive scale-up to field-deployable multi-couple modules.

Background

A thermoelectric generator (TEG) is a solid-state heat engine: it converts a heat flux directly into electrical power through the Seebeck effect, in which a temperature difference across a semiconductor leg produces a voltage. A module is built from many 'couples' — electrically connected pairs of p-type and n-type legs — sandwiched between ceramic headers (here aluminum nitride) that conduct heat to and from the legs, with copper interconnects and joining (solder/braze/metallization) layers completing the electrical and thermal paths. The conversion efficiency of an ideal module is set by the leg materials' dimensionless figure of merit zT=S2σT/κzT = S^2\sigma T/\kappa (where SS is the Seebeck coefficient, σ\sigma the electrical conductivity, and κ\kappa the thermal conductivity) averaged over the operating window, and by the temperature difference ΔT\Delta T between the hot and cold headers; for example, a constant-property model gives maximum efficiencies of 7.9%, 9.3%, 10.4%, and 11.5% at ΔT=230\Delta T = 230 °C for average zTzT values of 0.8, 1.0, 1.2, and 1.4, respectively. Bi2Te3-based alloys are the workhorse commercial thermoelectric materials for near-room-temperature applications, and waste-heat recovery over roughly 25–250 °C is their principal power-generation market. Nozariasbmarz, Poudel, Li, Kang, Zhu, and Priya (iScience 23, 101340, 2020) used compositional and defect engineering to obtain average zTzT of 1.08 (p-type Bi0.5Sb1.5Te3-based) and 0.84 (n-type Bi2Te2.7Se0.3-based) between 25 and 250 °C, and built single-couple modules reaching 8% efficiency and 2.1 W/cm² at ΔT=230\Delta T = 230 °C — about 40% above contemporary commercial modules — with an 18-couple full module whose output power scaled as expected (~18×). Nevertheless, the measured efficiencies were about 15% below the theoretical estimates for the same leg materials. The authors attribute this gap to parasitic heat losses through the thermoelectric legs and junctions and to the actual internal leg temperature being lower than the external header temperature that is measured and imposed in the model; they further note that electrical and thermal contact resistances at the leg/metal-interconnect and interconnect/substrate junctions — governed by contact metallization, soldering/brazing materials, and fabrication quality — have made reliable low-resistance joining a persistent difficulty in thermoelectric device fabrication. In their stated limitations, they add that comparing literature and commercial modules is unreliable because modules and legs differ in size, and that measuring the efficiency of large-size modules is challenging mainly because the input heat flow QinQ_{in} (the denominator of the efficiency) cannot be measured accurately. These device- and metrology-level obstacles sit between materials advances (higher zTzT) and the delivered cost-effective waste-heat conversion that motivates the field; as the paper puts it, such materials techniques 'do not assure high conversion efficiency' of TEGs, mainly due to low average zTzT or high device contact resistance.

Problem Statement

For thermoelectric generator modules based on bismuth telluride alloys operating over roughly 25–250 °C, where the leg materials already reach average figures of merit (zT)avg≈0.84(zT)_{avg}\approx 0.84–1.08 and single-couple devices have demonstrated 8% conversion efficiency at ΔT=230\Delta T=230 °C (Nozariasbmarz et al. 2020), resolve the device- and metrology-level obstacles that separate measured module performance from the materials' thermodynamic potential, as one coherent objective with three jointly supporting parts. (1) Loss quantification and suppression: identify and quantify the parasitic heat-flow channels, the electrical and thermal contact resistances at the leg/metal-interconnect and interconnect/header junctions, and the mismatch between the internal leg temperature profile and the externally measured header temperatures, and reduce them so that the measured efficiency of a multi-couple module approaches the theoretical estimate computed from the same leg materials' (zT)avg(zT)_{avg} — closing the roughly 15% shortfall observed in the source demonstration — including reliable low-resistance joining/metallization schemes that can be reproduced at manufacturing scale. (2) Metrology: establish an accurate, standardized method, with a quantified uncertainty budget, for measuring the input heat flow QinQ_{in} and hence the conversion efficiency of large-size, multi-couple modules, so that efficiencies reported by different laboratories and manufacturers can be meaningfully compared. (3) Scale-up verification: demonstrate that (1) and (2) hold for scaled, large-area multi-couple modules — preserved (zT)avg(zT)_{avg}, low parasitics, and efficiency consistent with the validated measurement protocol. Part (2) is the enabling measurement science without which the closure claims of (1) and the preserved-performance claims of (3) cannot be independently verified. These three parts share that module technology but not one device: parts (1) and (2) may be satisfied on the same multi-couple module, whereas part (3) requires a separate scale-up demonstration on a module of larger active area and larger couple count than the module used for (1)–(2), so that the loss-closure of (1) and the metrology of (2) are shown to survive the increase in scale. Part (1): the measured efficiency of a multi-couple module, supported by a quantified loss budget and a reproducible joining/metallization scheme, should come within 5% (relative) of the constant-property theoretical estimate for the same (zT)avg(zT)_{avg} at ΔT=230\Delta T=230 °C — closing at least two-thirds of the ~15% shortfall reported in the source demonstration. Part (2): the measurement method should have a relative QinQ_{in} uncertainty budget of at most 5% and should agree with an independent reference heat-flow technique within that budget. Part (3): the same measured, uncertainty-bounded target of part (1) should hold on the scaled-up module, whose scale-up must be nondegenerate — the active area and couple count must each exceed those of the module used for (1)–(2) by at least an order of magnitude. The closure and preservation claims of parts (1) and (3) must be demonstrated by measurement on real multi-couple modules rather than by analysis or prediction alone.

References

  1. Amin Nozariasbmarz, Bed Poudel, Wenjie Li, Han Byul Kang, Hang Tian Zhu, Shashank Priya (2020). Bismuth Telluride Thermoelectrics with 8% Module Efficiency for Waste Heat Recovery Application. iScience 23, 101340. DOI: 10.1016/j.isci.2020.101340; https://doi.org/10.1016/j.isci.2020.101340; full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC7369584/
  2. Pawel Ziolkowski, Przemyslaw Blaschkewitz, Eckhard Müller (2021). Heat flow measurement as a key to standardization of thermoelectric generator module metrology: A comparison of reference and absolute techniques. Measurement 167, 108273. DOI: 10.1016/j.measurement.2020.108273; https://doi.org/10.1016/j.measurement.2020.108273
  3. Wu Zhu, Ping Wei, Jianqiang Zhang, Longzhou Li, Wanting Zhu, Xiaolei Nie, Xiahan Sang, Qingjie Zhang, Wenyu Zhao (2022). Fabrication and Excellent Performances of Bismuth Telluride-Based Thermoelectric Devices. ACS Applied Materials & Interfaces 14, 12276–12283. DOI: 10.1021/acsami.1c24627; https://doi.org/10.1021/acsami.1c24627
  4. Zhonglin Bu, Xinyue Zhang, Yixin Hu, Zhiwei Chen, Siqi Lin, Wen Li, Chong Xiao, Yanzhong Pei (2022). A record thermoelectric efficiency in tellurium-free modules for low-grade waste heat recovery. Nature Communications 13, 237. DOI: 10.1038/s41467-021-27916-y; https://doi.org/10.1038/s41467-021-27916-y
  5. S. Harish, D. Sivaprahasam, B. Jayachandran, R. Gopalan, G. Sundararajan (2021). Performance of bismuth telluride modules under thermal cycling in an automotive exhaust thermoelectric generator. Energy Conversion and Management 232, 113900. DOI: 10.1016/j.enconman.2021.113900; https://doi.org/10.1016/j.enconman.2021.113900
  6. Gang Wu, Qiang Zhang, Yuntian Fu, Xiaojian Tan, Jacques G. Noudem, Zongwei Zhang, Chen Cui, Peng Sun, Haoyang Hu, Jiehua Wu, Guo-Qiang Liu, Jun Jiang (2023). High-Efficiency Thermoelectric Module Based on High-Performance Bi0.42Sb1.58Te3 Materials. Advanced Functional Materials 33, 2305686. DOI: 10.1002/adfm.202305686; https://doi.org/10.1002/adfm.202305686
  7. Vaskuri C. S. Theja, Vaithinathan Karthikeyan, Dani S. Assi, Hongli Huang, Venkatramanan Kannan, Yue Chen, Chan-Hung Shek, Vellaisamy A. L. Roy (2024). 2D MXene Interface Engineered Bismuth Telluride Thermoelectric Module with Improved Efficiency for Waste Heat Recovery. Advanced Materials Technologies 9, 2301722. DOI: 10.1002/admt.202301722; https://doi.org/10.1002/admt.202301722
  8. Jhonatan Samuel Ferrer Caro, Eckhard Müller, Pawel Ziolkowski (2024). Characterization of thermoelectric generator modules: Numerical and analytical study on heat flow determination under transient temperature conditions. AIP Advances 14, 085019. DOI: 10.1063/5.0147567; https://doi.org/10.1063/5.0147567
  9. Weng Pin Wong, George Elsa, Muhammad Norhaffis Mustafa, Rashmi Walvekar, Arshid Numan, Mohammad Khalid, Phei Li Lau (2025). Bismuth telluride-based thermoelectric generators: Advances in synthesis, performance enhancement, and device applications. Journal of Power Sources 653, 237722. DOI: 10.1016/j.jpowsour.2025.237722; https://doi.org/10.1016/j.jpowsour.2025.237722