The search for new thermoelectric materials: Intermediate valent cerium intermetallics and multivalley semiconductors

Two classes of materials were evaluated for their potential in thermoelectric cooling applications: intermediate valent cerium intermetallics and “multivalley” semiconductors. In evaluating these materials, we use the thermoelectric figure of merit, Z · T = S 2T/ρκ where S is the thermopower, T is the temperature, ρ is the electrical resistivity, and κ is the thermal conductivity. As Z · T increases, the thermoelectric efficiency of the material in a cooling device also increases.; The physical properties of intermediate valent materials display numerous anomalies, including large S. The observed combination of large S and low ρ makes them a potentially attractive alternative to the small band gap semiconductors currently used for thermoelectric cooling applications. The systems investigated were Ce_1–xR_x Pd₃ where R = Nd and Th; Ce₂Ni_3–xM _vSi₅ where M = Co, Cu, or Pd; Ce₂Ni₂In _1–xSn_x; and polycrystalline and single crystalline Ce₂CoSi₃ and Ce₅Cu₁₉P₁₂. The room temperature _S was found to be between 20 and 80 μV/K for these materials. We attempted to improve S in some systems by substitutions, but the thermopowers were usually reduced. Even in the cases where S was increased, the magnitude remained below 80 μV/K. All measured S were too low to produce significant Z · T so we began researching semiconductors.; We have coined the term “multivalley” to describe semiconductors that have a large number, N_V, of extrema (valleys) in the electronic band structure at the band gap. The maximum possible N_V increases with increasing crystal symmetry and thermoelectric theory predicts that Z · T will increase with increasing N_V. We are trying to find multivalley materials by incorporating tetrahedral, polyatomic anions into semiconducting networks. The synthesis of A₄[SnTe₄] where A = Na and K was indicated by the color of methanolic; solutions but could not be verified by diffraction techniques. The previously reported phases M₂[SiS₄] where M = Fe and Mn were successfully prepared, but attempts to make the M = Cd and Zn versions, as well as M₄[SiS₄] where M = Cu and Au, were unsuccessful. Our attempts to produce new multivalley materials by reacting M₂[SiS₄] with Bi₂S₃ at 800–850°C were also unsuccessful. Our search for multivalley semiconductors is continuing with other tetrahedral anions.