Study of Filled Tungsten Bronze Strontium Barium Niobate for Thermoelectric Applications

Open Access
Chan, Jason Hl
Graduate Program:
Materials Science and Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
November 09, 2015
Committee Members:
  • Clive A Randall, Thesis Advisor
  • Susan E Trolier Mckinstry, Thesis Advisor
  • thermoelectric
  • SBN
  • strontium barium niobate
  • defect
  • nonstoichiometry
  • filled bronze
The criteria for developing efficient thermoelectric materials have been well-established: a high electrical conductivity, high Seebeck coefficient, and a low thermal conductivity are all required. These three material properties are encompassed in a dimensionless material figure of merit, zT. Commercial thermoelectrics which use materials with zTs ~ 1 are usually either expensive or non-sustainable. The formulation of an oxide thermoelectric would facilitate achieving sustainability and cost-effectiveness. Currently, there are only a handful of high performance thermoelectric oxides, especially of the n-type variety. The n-type oxide, strontium barium niobate (SBN), had previously been reported to have a figure of merit of 0.5 to 1.0 at 550 K. The higher thermopower is correlated with the precipitation of an NbO2-x secondary phase within the parent material after reaching a particular degree of oxygen-deficiency. The process was found to be reversible, leading to the hypothesis that A-site occupancy within the structure is directly linked to the thermoelectric properties. In this work, A-site filled strontium barium niobate (SrxBa6-xNb10O30-δ) ceramics prepared by conventional solid-state sintering were explored. The barium end member, Ba6Nb10O30-δ, served as a model system of study with respect to synthesis and electrical properties. Ceramic samples were prepared by solid-state sintering of powders for the filled SBN compositions. Sr-Ba ratios of 0-100, 10-90, 20-40, 40-60, 60-40, 80-20, and 100-0 were fabricated under 1300-1350°C at ~10-16 atm pO2. Densities of ~95% were reached. Phase purity was assessed by X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and energy dispersive spectroscopy (EDS). Electrical conductivities, Seebeck coefficients and Hall measurements were also used to provide insight into potential thermoelectric properties. XRD phase purity, and hence solid-solution, was achieved up to the 80-20 composition. Backscatter imaging from SEM and TEM micrographs showed no secondary phases on the microscopic level. Quantitative EDS showed that the 60-40 SBN sample had an A-site cation content of 6.21 +/- 0.33 and 5.07 +/- 0.25 for the filled and unfilled compositions, respectively. These numbers were within error of the ideal values of 6.00 and 5.00, proving the difference in site occupancy of the filled and unfilled bronze, respectively. The electrical conductivities of phase pure, non-oxidized filled SBN reached values of at least 120 S/cm. Kinetics were demonstrated to affect the final conductivities such that 30 hour anneals at 1300°C were required for 2.75 mm thick samples. For thin samples (1.38 mm) a 5 hour sinter was enough to achieve high conductivities. The largest conductivities and room temperature conductivities were demonstrated by the 40-60 samples; the values were 700 S/cm at 600 K and 400 S/cm, respectively. The high conductivity was consistent with A-site filling. However, measuring electrical conductivities on samples annealed at 10-14 atm pO2 at 1300°C showed one to two orders of magnitude decrease in the conductivity, despite being phase pure. The carrier concentrations in the filled and unfilled 60-40 composition were on the order of 1019 carriers/cm3 to 1021 carriers/cm3, respectively, implying a different defect compensation between the filled and unfilled bronzes. These two results suggest that the A-site filling plays a part in the high thermoelectric performance, but may not be solely responsible. Seebeck coefficients were taken to compare the power factors of the filled SBN to those of the unfilled SBNs. The largest power factor was demonstrated in the 60-40 composition with a value of 2.7 μW/(m-K^2 ) for samples processed under 2.5% H2 (10-16 atm pO2). Hall measurements were conducted to assess the carrier concentration of the filled and unfilled SBNs. It was found that the carrier concentration was on the order of 1021 carriers/cm3 for these heavily reduced samples, which was consistent with the A-site model hypothesis. This suggests that for the filled SBNs heat-treated at 10-16 atm pO2 are over-reduced and optimization of the power factor would require a higher partial pressure of oxygen during processing.