Doping Mechanisms and Electrical Properties of Sodium Niobate

Intermediate temperature solid oxide fuel cells (ITSOFCs) are one of many technologies attempting to decarbonise global energy production and utilise the hydrogen economy. Unfortunately, typical solid oxide electrolytes such as yttria-stabilised zirconia (YSZ) do not exhibit sufficient conductivity for operation in the intermediate temperature region (400-750 °C). As a result, many potential intermediate temperature solid oxide electrolytes have undergone investigation to replace YSZ. The goal of this project was to investigate the perovskite material NaNbO3 (NN) and identify its suitability as a solid oxide electrolyte.

NN ceramics were prepared via solid-state reaction of mixed oxides at 900 °C, followed by sintering at temperatures in the range 1240-1350 °C. Physical properties of NN ceramics were assessed using X-Ray diffraction (XRD), Scanning Electron Microscopy (SEM), Electron Probe Micro Analysis (EPMA) and Nuclear Magnetic Resonance (NMR). Electrical properties were investigated using impedance spectroscopy to assess bulk conductivity and pO2 dependence of the ceramics, probostat measurements have allowed for determination of oxygen ion transport numbers (tion) and the electronic band gap of NN has been determined via UV-Vis spectroscopy.

This project is divided into exploring the defect chemistry of NN and how it influences the electrical properties.

It has been shown that NN produced via solid-state synthesis varies significantly compared to hydrothermally prepared, and commercially available NN powders. The variance has been attributed to non-stoichiometry between the sodium and niobium sites caused by sodium volatility during processing. Deliberate Na:Nb non-stoichiometry in the range 0.96-1.22 showcases that sodium rich NN exhibits greater bulk conductivities (55:45 NN σ500°C ≈ 26 μS cm-1) and promotes ionic conductivity, compared to sodium deficient NN (49:51 NN σ500°C ≈ 0.13 μS cm-1).

The following doping mechanisms have been used to influence the electrical properties of NN:
1. Direct substitution of sodium for calcium to generate A-site vacancies:
2Na+ = Ca2+
2. A-site donor doping using Mg2+ and Ga3+ for the reduction of oxygen vacancies:
Na+ = Mg2+ + 0.5O2-
3. B-site donor doping using W6+ and Mo6+ for the reduction of oxygen vacancies:
Nb5+ = W6+ + 0.5O2-
4. B-site acceptor doping using Mg2+, Ga3+ and Ti4+ for the generation of oxygen vacancies:
Nb5+ + O2- = Ga3+

Generation of A-site vacancies via calcium doping leads to an increase in bulk conductivity (Na0.9Ca0.05NbO3 σ500°C ≈ 2.7 μS cm-1), with low levels of ionic conductivity attributed to sodium ion conduction.

Donor doping of the A-site using Mg2+ or Ga3+ has no significant effect on the bulk conductivity (Na0.99Mg0.01NbO3.005 σ500°C ≈ 0.24 μS cm-1), however, ionic conductivity is supressed. Similarly, donor doping the B-site with W6+ supresses ionic conductivity without influencing bulk conductivity (NaNb0.99W0.01O3.005 σ500 °C ≈ 0.69 μS cm-1). However, doping with molybdenum does not appear to be successful due to preferential formation of a molybdenum bronze.

Acceptor doping of the B-site with Ga3+ and Ti4+ promotes ionic conductivity and greatly enhances bulk conductivity (NaNb0.95Ga0.05O2.95 σ500 °C ≈ 0.08 mS cm-1/NaNb0.95Ti0.05O2.975 σ500 °C ≈ 0.05 mS cm-1). It has been shown that in gallium doped NN maximum conductivities are achieved at very low concentrations (<1at%) of dopant due to pinning of oxygen vacancies. Acceptor doping with Mg2+ appears to be unsuccessful due to a preference for A-site donor doping.

The properties of NN vary significantly according to the dopant mechanism utilised, and the Na:Nb ratio. The electronic band gap of NN appears to be unaffected by acceptor doping and remains 3.4 eV. Oxygen ion transport numbers for acceptor doped NN ceramics indicate that all variants are mixed oxide ion-electronic conductors (tion=0.2-0.7), indicating that NN based materials are unsuitable for applications as solid oxide electrolytes.