Defect engineering and nano-structuring are the core stratagems for improving thermoelectric properties. In bismuth telluride alloys nanosizing individual crystallites has been extensively studied in efforts to reduce the thermal conductivity, but nanostructuring with second phases has been more challenging. In this study, we demonstrate a thermoelectric figure of merit ZT of 1.4 at 400 K, realized in Zn-containing BiSbTe alloys (specifically Bi_0.46Sb_1.54Te₃) by integrating defect complexity with nanostructuring. We have succeeded in creating nanostructured BiSbTe alloys containing ZnTe nanoprecipitates. We present a melt-spinning-based synthesis that forms in situ ZnTe nanoprecipitates to produce an extremely low lattice thermal conductivity of ∼0.35 W m⁻¹ K⁻¹ at 400 K, approaching the amorphous limit in the Bi_2−xSb_xTe₃ system, while preserving the high power factor of Bi_0.46Sb_1.54Te₃. These samples show excellent repeatability and thermal stability at temperatures up to 523 K. DFT calculations and experimental results show that Zn is inclined to form dual site defects, including two substitutional defects Zn_Bi/Sb′ and a Te vacancy, to achieve full charge compensation, which was further explicitly corroborated by Positron annihilation measurement. The strong enhancement of thermoelectric properties was validated in a thermoelectric module fabricated with the melt-spun p-legs (ZnTe-nanostructured BiSbTe) and zone-melt n-legs (conventional BiTeSe) which achieved a thermoelectric conversion efficiency of 5.0% when subjected to a temperature gradient of 250 K, representing about 40% improvement compared with a commercial zone-melt-based module. The results presented here represent a significant step forward for applications in thermoelectric power generation.

While zone-melted (ZM) Bi₂Te₃ is a standard commercially available thermoelectric (TE) material, it suffers from not being sufficiently mechanically robust due to the presence of the van der Waals bonded Te–Te layers that reduce the product yield and compromise operational reliability. Polycrystalline materials prepared by powder metallurgy techniques exhibit improved mechanical properties but usually lose the desired texture exhibited by ZM ingots, and this affects their TE performance. It is highly desirable to be able to fabricate Bi₂Te₃-based bulk materials with anisotropies similar to a single crystal, yet being mechanically strong as the polycrystalline specimens. Herein, we combine for the first time the thermal explosion technique with selective laser melting (SLM) to synthesize the highly textured p-type Bi_0.4Sb_1.6Te₃ bulk material. Structural analysis (FESEM and XRD) indicates that the slender columnar grains grew along the building direction (BD) of the structure and the orientation factor reached up to 0.9, close to that representing a single crystal. TEM images revealed a high density of dislocations inside the grains. Since the printed compound has a high degree of texture, the TE and mechanical properties exhibit a highly anisotropic behavior. The maximum ZT of annealed samples parallel to the BD was 1.1, similar to that of the single crystal. However, the compressive strength of the structure reached up to 91 MPa, some 2.5 times the strength of a typical single crystal (37 MPa), and even higher than that of Spark Plasma Sintered (SPS) polycrystalline samples (80 MPa). Meanwhile, the mechanical cutting performance was much superior compared to that of the ZM ingot, and TE legs could be cut to sizes as small as 0.2 mm. A micro-TE module assembled using SLM-printed high performance p-type BiSbTe and SPS-compacted n-type BiTeSe materials showed the maximum cooling temperature difference of 62 °C. The work provides a facile and effective solution for preparation of Bi₂Te₃-based materials with high texture, robust mechanical properties, and excellent TE performance. As such, it lays a solid foundation for rapid in situ 3D printing of Bi₂Te₃-based micro-TE devices.

Reduction of lattice thermal conductivity in Bi _0.5 Sb _1.5 Te _{3+ x} is absent despite dense dislocation arrays at grain boundaries.