Effects of A- and B-site (co-)acceptor doping on the structure and proton conductivity of LaNbO₄

Abstract

B-site and A- and B-site co-doped lanthanum niobates, La_(1-x)D_A(x)Nb_(1-y)D_B(y)O_(4-δ), where D_A stands for Ca or Ba and D_B for Ga, Ge or In (x = 0 or 0.01; y = 0.01), have been synthesized via the solid-state route. Essentially single-phase materials were obtained after 10 h at 1500 °C with monoclinic Fergusonite structure at room temperature. The unit cell volume increases slightly with the presence of the acceptor. The electrical conductivity of the materials, characterized as a function of the temperature from 300 °C to 900 °C under wet reducing conditions with AC impedance spectroscopy, was dominated by protons. The maximum conductivity of 4.1 · 10⁻⁴ S · cm⁻¹ at 900 °C was obtained for La_0.99Ca_0.01Nb_0.99Ga_0.01O_4-δ, which is a factor of ~ 2.5 lower than the highest conductivity reported for the A-site Ca-doped LaNbO₄. Based on a semi-quantitative evaluation it was shown that the hydration of the B-site and A- and B-site co-doped LaNbO₄ is slightly more exothermic and that the proton mobility has higher activation enthalpies than Ca-doped LaNbO₄. This is accounted for by a stronger association between the acceptor and the proton for B-site than for A-site doping.

Introduction

Acceptor-doped lanthanum ortho-niobates (LaNbO₄) exhibit proton conduction at elevated temperatures accompanied by high chemical and thermal stability [1], [2]. Additionally, good mechanical properties, high red-ox stability and compatibility with NiO and Ni, all together make these compounds candidates for electrolyte membranes in proton-conducting solid oxide fuel cells (PC-SOFC) [3], [4], [5], [6], [7], [8].

Rare earth ortho-niobates (LnNbO₄) crystallize in a low temperature Fergusonite structure with monoclinic symmetry, but change to a tetragonal Scheelite structure at higher temperatures [9], [10], [11], [12]. The phase transition between the two polymorphs is reversible and takes place in the range of 495–520 °C [10], [13]. The Scheelite structure (ABO₄) can be described as two intercalated diamond lattices, one for the A-site cation and the other for the B-site cation [14]. The A-site cation is coordinated by eight oxygen ions and the B-site cation is tetrahedrally coordinated by the oxygen ions. In the Fergusonite polymorph, A- and B-sites are coordinated similarly to the Scheelite form, with two additional near-oxygen sites at each B-site cation. As reported previously [11], [13], [14], [15], the Scheelite-to-Fergusonite transition can be described as a shear transformation in which the Scheelite structure is partially conserved while certain sheets are slightly shifted.

Undoped LaNbO₄ shows conductivity at high temperatures in the order of 10⁻⁵ S · cm⁻¹ under wet H₂ (p(H₂O) ~ 0.025 atm) [16] which is far too low for any high-drain application and acceptor doping is required to improve the functional characteristics. The electrical properties of nominally stoichiometric and A-site acceptor-substituted (Ca, Sr and Ba) LaNbO₄ have been characterized to a detailed level [1], [2], [16], [17], [18]. The materials are mixed conductors and the relative influence of the different charge carriers varies with the composition and conditions. Ionic conductivity predominates up to temperatures of 800 °C, above which p-type conduction starts to contribute under oxidizing conditions according to the defect chemical reaction:

$$\frac{1}{2}{O}_2\left(g\right)+v_O^{••}=O_O^x+2h^{•}$$

n-type conductivity is virtually negligible up to ~ 1100 °C and contributes only under very reducing conditions:

$$O_O^x=v_O^{••}+2e^{\prime }+\frac{1}{2}{O}_2\left(g\right)\text{.}$$

Protons are the prevailing ionic charge carriers under wet conditions; e.g. 1% Ca substituted LaNbO₄ is an essentially pure proton conductor at temperatures up to 950 °C [1], [2]. The conductivity of LaNbO₄ increases significantly (1 to 2 orders of magnitude) with acceptor substitution, reaching proton conductivities in the order of ~ 1 · 10⁻³ S · cm⁻¹ at 950 °C in wet H₂ (p(H₂O) ~ 0.025 atm) [1]. A further increase in the A-site doping concentration or changing to another alkaline-earth element (Sr, Ba) did not increase significantly the conductivity in the case of LaNbO₄ [16].

Substitution with aliovalent cations is indeed a frequently used strategy to modify and control the electrical properties of ceramic materials. Therefore, when selecting the type of doping element and its concentration, the targeted defect formation and the structural features of the host crystal lattice have to be taken into account. Acceptors in LaNbO₄ are charge compensated by formation of oxygen vacancies [19]. The vacancy formation is accompanied by a local structural relaxation, associated with rearrangement of [NbO₄]^3- tetrahedra forming [Nb₂O₇]^4- and [Nb₃O₁₁]^7- units [20]. Moreover, as shown in Ref. [19] for the tetragonal symmetry of LaNbO₄, the solution energies of foreign elements into the La- and Nb-sublattices depend on their ionic radii, pointing out that geometrical similarity is energetically favorable.

Dissolution of protons in LaNbO₄ follows the well accepted scheme where oxygen vacancies are hydrated by water vapor, forming hydroxide defects [21], [22], [23]:

$$H_{2}O\left(g\right)+v_O^{••}+O_O^x=2O{H}_O^{•}$$

Accordingly, the concentration of protons increases with increasing the doping level which is generally also beneficial for the proton conductivity. Looking at the literature of proton conductors there have been a few attempts to empirically correlate hydration thermodynamics and proton transport to physical parameters and structure [23], [24], [25], [26], [27], [28], [29]. The hydration enthalpy was correlated to the ionic radius along the series of Ca-doped rare-earth ortho-niobates [1], whereas for perovskite structured oxides other parameters like basicity and electronegativity have been tested [22], [23]. Within one class of materials, variations in the proton mobility generally depend on changes within the structure, of which the ionic radius obviously plays an important role. In order to select the elements to (co-)acceptor dope on the A- and/or B-site in LaNbO₄, we have therefore chosen to evaluate the ionic radius and the electronegativity. Fig. 1 shows the Pauling electronegativity as a function of the ionic radius (according to Shannon [30]) for the selected doping elements, as well as for the parent cations. The materials in the present study have a chemical composition corresponding to the general formula La_(1-x)D_A(x)Nb_(1-y)D_B(y)O_(4-δ), where D_A is either Ca or Ba and D_B is one of Ga, Ge or In (x = 0 or 0.01; y = 0.01) as summarized in Table 1. Different doping schemes of LaNbO₄, A- and/or B-site, were implemented aiming to correlate the conductivity of doped materials with the chemical and structural features of the host–guest matrix. Ultimately, this strategy may lead to an improved understanding from which the conductivity behavior of acceptor-doped LaNbO₄ materials may be optimized.