Control of Cu-doping behavior in n-type Cu0.01Bi1.99Te2.7Se0.3 polycrystalline bulk via fabrication technique change

Abstract Bi–Te-based alloys are potential candidates for use in thermoelectric (TE) modules for low-mid-temperature energy harvesting. However, n-type Bi–Te–Se alloys are inferior to their p-type counterparts in terms of TE performance. It has been found that doping Cu atoms into n-type Bi–Te–Se alloys is effective in improving the TE properties; however, different studies have reported contrasting roles of Cu dopants in the transport properties of Bi–Te–Se alloys. This is attributed to the complex doping behaviors of Cu atoms. In this study, it is demonstrated that employing an appropriate fabrication technique can enable more Cu dopants to be directed to specific atomic sites than others. As part of the study, two n-type polycrystalline Cu0.01Bi1.99Te2.7Se0.3 bulk samples are respectively fabricated via melt-spinning followed by spark plasma sintering, and via ball-milling followed by spark plasma sintering. The majority of Cu dopants in the melt-spun sample are suspected to exist in between Te1–Te1 atomic layers (van der Waals gap), given the elongated c-axis, increased Hall carrier concentration, and improved non-degenerate mobility (intercalated Cu generates one electron and acts as an electric connector). Resultantly, a ~30% enhancement in the TE figure-of-merit (~0.92 at 360–400 K) is achieved in the melt-spun sample, relative to the ball-milled sample, wherein Cu atoms substitute Bi atoms.