Copper chalcogenides-based thermoelectric material for energy conversion application

Abstract

The superionic thermoelectric material, copper selenide (Cu₂-xSe) has gained great attention due to its liquid-like behavior at high temperature. The liquid-like property could highly suppress the lattice thermal conductivity owing to the high mobility of Cu while maintain an excellent electrical conductivity, resulting in the large enhancement of thermoelectric efficiency. However, the high mobility of Cu also has some issues relating to thermal stability and data repeatability during thermal cycling tests, which is under great concern in recent years. Up to now, seldom reported articles systematically studied the data repeatability and the factors affecting the data repeatability and thermal stability. In this study, the Cu vacancy and selenium (Se) vacancy were studied using a conventional sintering method. Besides, the isothermal condition was also explored to suppress the evaporation of Se during the sintering progress. Finally, the parameters for the fabrication of Cu₂-xSe were determined, and the enhancement of thermoelectric efficiency was further studied by alloying with indium, In. Cu₂-xSe (x = 0.01, 0.03, 0.06), superionic thermoelectric material (denoted as Cuvac-1, Cuvac-3, Cuvac-6 ) was studied systematically by employing the conventional sintering method for the first time with regards to the phase structure, thermal stability, data repeatability and doping effect as well. The data repeatability of the Cu/Se system at heating and cooling cycling have been studied accordingly. It was found that the Cu vacancy strongly affects the data repeatability of the thermoelectric performance for the conventional sintered samples. Finally, the Cu1.97Se sample can have the excellent zT performance of 0.8 at 800 K with good repeatable data and cycling performance. To explore the effect of sintering temperature on the thermoelectric performance and data repeatability, the Cuvac-3 sample was sintered at 873 K, 973 K, and 1073 K for 2 hours using the conventional sintering method. It was found that the sintering temperature could reach as low as 873 K to obtain a densified sample with good data repeatability in thermal cycling. However, the thermoelectric performance is relatively poor when compared to the ones sintered at 973 K and 1073 K. In the thermogravimetric analysis (TGA) analysis for the Cuvac-3 powder, the evaporation of Se occurs at around 750 K, which is even slightly lower than the lowest sintering temperature adopted in the study. This means that the sample sintered at 873 K might also have the composition deviation with regard to the nominal composition. To minimize the evaporation of selenium (Se) during the high-temperature sintering process, the vapor control and isothermal control were conducted during the conventional sintering process. The control experiment was conducted between the Cuvac-3 sample sintered in a conventional tube furnace with the temperature gradient along the sintering chamber, and the sample sintered in a sealed small quartz tube located in the isothermal area of the tube furnace (Cuvac-3A). By comparing the thermoelectric performance between these two samples, the conclusion can be made that the Cu-Se materials experience the severe evaporation of Se during the sintering process. The evaporation process is strongly related to the sintering temperature and the isothermal environment. The results provide the valuable information for the selection of spark plasma sintering (SPS) or hot press sintering (HPS) method to fabricate the Cu₂Se material. To further enhance the thermoelectric performance, indium (In) is chosen as the dopant of Cuvac-3A. With the introduction of In, the electrical conductivity, σ, of the doped Cuvac-3A is reduced, which is probably due to the reduced carrier concentration, n. It is found that the In-Cu-Se system is uncommon compared to the other dopants due to the partial solubility of In into the Cu-Se system. The In dopant acts as a softening agent, which reduces the specific heat capacity, cp, and the total thermal conductivity, Ktot, of the sample due to the impurity phase of CuInSe₂. After applying the single parabolic model (SPB) model, the lattice thermal conductivity, Kl, was found to be greatly reduced. The Kl values of the Cu₁.₉₇In₀.₀₃Se sample (Cuvac-3A-In) with the micron grain size are even lower than those of the samples with the nano grains. With the optimized doping amount of In, the zT of Cuvac-3A-In approached 1 at 800 K. The SPB model predicts the optimized zT for Cu2-xInxSe that could reach to 1.8 at 800 K if could be further reduced.