Oxygen ion conductivity in ceria-based electrolytes co-doped with samarium and gadolinium

Highlights

• Thorough, systematic study of eleven Gd- and Sm-doped ceria electrolytes

• Co-doping Gd and Sm improved conductivity over singly-doped parent compositions.

• Total dopant concentration of 0.175 and Sm/Gd ratio of 50/50 was best.

• Conductivity values favourably comparable with the state-of-the-art

• Kinetic parameters and capacitance values give insights into conduction process.

Abstract

In a systematic study, two compositional series of ceria-based oxides, both co-doped with Sm and Gd, were synthesised using a low temperature method and evaluated as oxygen ion-conducting electrolytes for Intermediate Temperature Solid Oxide Fuel Cells (IT-SOFCs). Series one, Ce_1-2xSm_xGd_xO_2-x, had equal concentrations of Sm and Gd but varying total dopant concentration. Series two, Ce_0.825Sm_xGd_0.175-xO_1.9125, had a fixed total dopant concentration but the Sm:Gd concentration ratio was varied. The materials were characterised using scanning and transmission electron microscopy, inductively coupled plasma mass spectrometry and X-ray diffraction. Impedance spectra were recorded on dense pellets of these materials. From these, total, bulk and grain boundary conductivities and capacitances along with activation energies, pre-exponential constants and enthalpies of ion migration and defect association were obtained. These gave a detailed insight into the fundamental conduction processes in the materials. Ce_0.825Sm_0.0875Gd_0.0875O_1.9125 had the highest total ionic conductivity at temperatures of 550 °C and above and also demonstrated an enhanced conductivity with respect to its singly-doped parent compounds, Ce_0.825Sm_0.175O_1.9125 and Ce_0.825Gd_0.175O_1.9125, at 400 °C and above. This compares favourably with previously-reported values and has promising implications for the development of IT-SOFCs.

Introduction

A promising technology to help address our global energy and environmental challenges are fuel cells, which convert chemical potential energy from a fuel and oxidant directly into electrical energy. Solid oxide fuel cells (SOFCs) in particular have high efficiencies, low emissions and fuel flexibility [1]. SOFCs are likely to have a large impact on energy conversion in the future, over a range of commercial applications from small domestic power units to large industrial facilities [[2], [3], [4]]. Currently, SOFCs typically operate at temperatures of 800 to 1000 °C [5]. By reducing this to intermediate temperatures (ITs) of between 500 and 750 °C, the auxiliary plant could be manufactured from low cost materials such as standard steels rather than expensive technical ceramics (e.g. precision-made alumina parts, lanthanum chromite-based interconnects). As well as allowing wider material choices, IT-SOFCs would reduce loss of performance and component degradation caused by electrode sintering, interfacial diffusion between electrolyte and electrodes and thermal stress [6,7].

For successful implementation of IT-SOFCs, electrolytes with higher ionic conductivity are necessary. A partially-occupied oxygen ion sub-lattice containing a large number of interconnected and equivalent sites is required for high conductivity. This is obtained by doping the material with acceptor cations and so creating oxygen vacancies through which oxide ions are transferred by a hopping mechanism [8]. Yttria stabilised zirconia (YSZ) is the electrolyte material used in first generation SOFCs, but at intermediate temperatures, ceria-based electrolytes demonstrate higher oxide ion conductivity [9]. Reports have shown that ceria doped with trivalent rare earth ions – such as Gd as in Eq. (1) – gives higher ionic conductivity than those doped with other elements, and in certain cases using multiple rare earth dopants gives rise to higher conductivity than the use of a single dopant [4,10].

$$G{d}_2{O}_3\to 2G{d_{\mathit{Ce}}}^{\prime }+3O_O^x+V_O^{··}$$

One interesting recent approach for optimising ionic conductivity is to search for an average dopant ionic radius which causes least distortion to the host lattice. This is reported to minimise both strain and activation energy for oxygen vacancy diffusion, E_a, and occurs when the repulsive elastic (related to dopant radius) and attractive electronic parts of the interactions between vacancies and dopant ions balance [[11], [12], [13], [14]]. According to research by Andersson and co-workers using ab initio methods, this is the case for a hypothetical atomic number between 61 (Pm) and 62 (Sm) [15]. Following this computational work, Omar et al. experimentally studied the energies of interaction between the oxygen vacancy and the dopant cations as a function of dopant ionic size. They concluded that the ionic conductivity was not a function solely of elastic strain, and therefore, a structure–ionic conductivity relationship based on critical radius was not sufficient to explain ionic conductivity behaviour in doped ceria [16]. Indeed, the local structure around dopants in CeO₂, such as the formation by electrostatic attraction of clusters of vacancies and dopant ions, has been reported to influence ionic mobility [17]. Connecting these concepts with experimental observations is valuable for electrolyte design.

Previous work by the authors on co-doped ceria electrolytes found that of the Ce_0.8Sm_xGd_yNd_zO_1.9 materials studied (where x, y and z = 0.2, 0.1, 0.0667 or 0 and x + y + z = 0.2), Ce_0.8Sm_0.1Gd_0.1O_1.9 exhibited the highest conductivity between 300 °C and 700 °C [4]. Wang and co-workers observed a higher conductivity for Ce_0.85Gd_0.15-ySm_yO_1.925 (0.05 ≤ y ≤ 0.1) than for Ce_0.85Gd_0.15O_1.925 or Ce_0.85Sm_0.15O_1.925 between 500 and 700 °C [18]. This was attributed to the suppression of the ordering of oxygen vacancies leading to a lower E_a for the co-doped than for singly doped ceria. Dikmen and co-workers also found a four-fold increase in conductivity for Ce_0.8Gd_0.2-xSm_xO_2-δ at 700 °C for x = 0.1 over x = 0. Greater increases were observed at the lower temperatures of 500 and 600 °C [19]. However, using La or Nd as co-dopants with Gd failed to surpass the conductivity of the material co-doped with Sm and Gd, at 700 °C. Zając and Molenda demonstrated that Ce_0.85Sm_0.075Gd_0.075O_2-x/2 had a higher bulk conductivity at 700 °C than the singly-doped parent compounds or the co-doped sample, Ce_0.85Nd_0.075Gd_0.075O_2-x/2 [20].

It is apparent from the prior work of the authors and the relevant literature that Sm and Gd as either single or co-dopants in ceria electrolytes enhance conductivity to a greater extent than other rare earth dopants. Therefore, it is worth conducting a detailed study of how the total dopant concentration and the ratio of these individual co-dopants affect ionic conductivity in ceria co-doped with Sm and Gd. In this work, two compositional series, Ce_1-2xSm_xGd_xO_2-x, where x = 0.125, 0.1, 0.0875, 0.075 or 0.05, and Ce_0.825Sm_xGd_0.175-xO_1.9125, where x = 0, 0.035, 0.070, 0.0875, 0.105, 0.140 or 0.175, were synthesised using a low temperature citrate method known to give high purity nanopowders [3,4]. The composition, powder nanostructure and crystal phase of these products were studied as were the microstructure and ionic conductivity of the dense electrolyte bodies prepared from them. The results gave insight into how subtle changes in composition affect the parameters determining favourable performance for IT-SOFC electrolytes. It is hoped that they will also aid future computational work on these materials.