Defect Analysis and Optimization of the Indium Oxide (Zinc Oxide) System for Energy Applications.

摘要

The defect chemistry of compounds in the In2O3(ZnO) k system (k = 3, 5, 7, and 9) was investigated via analysis of the dependence of conductivity on oxygen partial pressure at high temperature (750°C). Defect mechanisms were proposed based on the resulting Brouwer slopes, and the presence of two defect regimes was attributed to competition between In 2O3--like (k = 3) and ZnO--like behavior (k = 5, 7, 9). The donor in both cases is proposed to be the In antisite defect ( In•Zn ). To support the validity of the proposed models, density functional theory was used to calculate the formation energies of the proposed defects.;X-ray and ultraviolet photoelectron spectroscopy studies were performed on the surfaces of thin film (k = 2) and bulk specimens (k = 3, 5, 7, and 9) to investigate their work functions, Fermi levels, and ionization potentials. The work functions and Fermi levels for all samples fell on a line of constant ionization potential (∼7.7--7.8 eV), which was independent of composition and unchanged by oxidation and reduction treatments. This ionization potential was very close to those of oxidized In2O3 and ZnO surfaces. The work functions of each sample were similar when measured under vacuum or ambient conditions (4.7--4.9 eV).;The high-temperature (750°C) electrical conductivity and thermopower of bulk samples of In2O3(ZnO)k were analyzed to assess their potential for use in thermoelectric power applications. The density of states-mobility products were determined by Jonker analysis (thermopower vs. ln conductivity) of the measured electrical properties. Ioffe analysis (log-log plots of maximum power factor vs. density of states-mobility product) was employed to predict the maximum thermoelectric power factor that could be achieved for each compound with an optimized carrier concentration, which agreed well with the best performances reported in the literature. The maximum predicted power factors were similar for the k = 1--7 phases, between 2 and 15x10-4 W/mK2, and were comparable to those of other n-type thermoelectric oxides. The predicted power factors were much lower for the k = 9 phase.