Abstract: To date, the benchmark Bi₂Te₃-based alloys are still the only commercial material system used for thermoelectric solid-state refrigeration. Nonetheless, the conspicuous performance imbalance between the p-type Bi_2-xSb_xTe₃ and n-type Bi₂Te_3-xSe_x legs has become a major obstacle for the improvement of cooling devices to achieve higher efficiency. In our previous study, novel n-type Bi_2-xSb_xTe₃ alloy has been proposed via manipulating donor-like effect as an alternative to mainstream n-type Bi₂Te_3-xSe_x. However, the narrow bandgap of Bi_2-xSb_xTe₃ provoked severe bipolar effect that constrained the further improvement of zT near room temperature. Herein, we have implemented band gap engineering in n-type Bi_1.5Sb_0.5Te₃ by employing isovalent Se substitution to inhibit the undesired intrinsic excitation and achieve the distinguished room-temperature zT. First, the preferential occupancy of Se at Te² site appropriately enlarges the band gap, thereby concurrently improving the Seebeck coefficient and depressing the bipolar thermal conductivity. In addition, the Se alloying mildly suppresses the compensation mechanism and essentially preserves the already optimized carrier concentration, which maintains the peak zT near room temperature. Moreover, the large strain field and mass fluctuation generated by Se alloying leads to the remarkable reduction of lattice thermal conductivity. Accordingly, the zT value of Bi_1.5Sb_0.5Te_2.8Se_0.2 reaches 1.0 at 300 K and peaks 1.1 at 360 K, which surpasses that of most well-known room-temperature n-type thermoelectric materials. These results pave the way for n-type Bi_2-xSb_xTe₃ alloys to become a new and promising top candidate for large-scale solid-state cooling applications.

Key words: Thermoelectric, Bi₂Te₃, Band gap, Bipolar effect, Point defects