2018 Volume 59 Issue 1 Pages 19-22
The influences of In- and Y-doping on the electrical conduction properties of barium zirconate were investigated. The electrical conductivity measured on of BaZr1−x−yInxYyO3−δ (x = 0, 0.1, 0.2 and y = 0, 0.1, 0.2) could be understood that yttrium doping causes high bulk conductivity and indium doping leads to lowering activation energy of the grain boundary. Co-doping of yttrium and indium promotes the enhancement effect of improving the bulk conductivity and lowering of activation energy, and it is thus expected that the co-doping with yttrium and indium can work for controlling the bulk and grain boundary conduction specifically in the materials.
Trivalent-cation-doped barium zirconate shows proton conduction and is a promising candidate electrolyte material for intermediate temperature solid oxide fuel cells and steam electrolysis1–7). The material has the cubic perovskite-type structure. Partial substitution of trivalent cation for the tetravalent zirconium ion causes the formation of oxide ion vacancy and the hydration of the oxide ion vacancy with ambient moisture results in the formation of protonic charge carrier1). Yttrium has been reported to be the best dopant as long as the bulk proton conductivity is concerned2), whereas the material is characterized by high grain-boundary resistance. One reason for the high grain-boundary resistance is due to the poor sinterability, causing limited grain growth even by sintering at high temperature (～1800℃)8–10). It is supposed that such high resistance at the grain boundaries is originated from high potential barriers at the grain boundaries11,12). The addition of yttrium assists the grain growth of barium zirconate, and this effect become stronger with the increase of the doping level7). Indium is another possible dopant for barium zirconate and is reported to enhance the sinterbility, whereas the conductivity is less than that of Y-doped case, due to low bulk conductivity of In-doped BaZrO313–15).
In this study, the effect of yttrium and indium doping to barium zirconate, i.e., BaZr1−xYxO3−δ, BaZr1−xInxO3−δ, (x = 0.1, 0.2), on the electrical conductivity is compared. It was found that the yttrium doping results in higher conductivity than indium doping, while the indium doping leads to lower activation energy of the grain boundary conduction. In response to this result, yttrium-indium co-doping, i.e. BaZr1−x−yInxYyO3−δ (x, y = 0.1, 0.2) has been investigated to see how the conductivity and activation energy changes. These investigated compounds are referred to hereafter as BZY, BZI and BZIY, or composition-specifically as BZY91, BZY82, BZI91, BZI82, BZIY811, BZIY721 and BZIY712, respectively.
BaZr1−x−yInxYyO3−δ (x = 0, 0.1, 0.2 and y = 0, 0.1, 0.2) electrolytes were prepared by a chemical solution method using aqueous solution. Ba(NO3)2 (Wako, 99%), ZrO(NO3)2∙xH2O (Zirconyl nitrate solution, Aldrich, 35 mass%, 99%), Y(NO3)3∙6H2O (Wako, 99.9%) and In(NO3)3∙3H2O (Wako, 98%) were used. Citric acid (Wako, 99.5%) and ethylene diamine tetraacetic acid (EDTA, Dojindo, 99%) were used as chelating and complexing agents. The molar ratio between total metal cations, EDTA, and citric acid was set at 1:1.5:1.5 and the appropriate amounts of the materials are dissolved in deionized water under vigorous stirring. NH3 water (Chameleon reagent, 28%) was added to the solution to adjust the pH to approximately 9 to 10. The aqueous solution was dehydrated on a hot plate at 260℃ to generate a viscous liquid. This material was then dried at 240℃ for overnight in vacuum and calcined at 900℃ for 10 h to obtain precursor powder. The powder was then ball milled in ethanol at 300 rpm for 5 days with 2 mm zirconia balls and dried at 120℃ for overnight in vacuum, sieved (150 µm), pressed into pellets at 250 MPa for 10 min and finally sintered at 1600℃ for 10 h in air. The crystal structure of the sintered samples was determined by X-ray diffraction (XRD, CuKɑ, 40 kV–40 mA, Rigaku Ultima IV). The microstructures of the sintered pellets were observed via a scanning electron microscopy (SEM, Topcon, SM-350) conducted on the fracture surface. The electrical conductivity was measured by AC impedance method (Versa STAT 3). The sintered and polished samples were painted with platinum electrode (Tanaka Kikinzoku Kogyo, TR-7907) on the bar sample surface, then heat-treatment at 950℃ for 1 h in air (4-probe impedance cell).
The XRD patterns of BZY and BZI electrolytes as sintered at 1600℃ for 10 h in air are shown in Fig. 1. The patterns of all the specimen show the cubic structure without secondary phase. The lattice parameter and relative density are indicated in Table 1. With respect to BZY and BZI systems, the lattice parameters are in the order of BZY82 > BZY91 > BZI82 > BZI91. The ionic radius of the six-coordinated In3+, Y3+ and Zr4+ are 80 pm, 90 pm and 72 pm, respectively16), and the order of the lattice parameters are consistent with the ionic radii. The relative density obtained for all the synthesized specimens were above 90% and is dependent on the doping level. 10% Y- and In-doped barium zirconate showed relative densities lower than 92%. As mentioned above, low sinterability is a characteristic feature of barium zirconate based electrolytes. As doping level increases, the relative density increased 91.8% to 97.1% in BZY system. The same tendency is found in the BZI system as increased 91.5% to 96.9%.
XRD patterns of BZY and BZI electrolytes as sintered at 1600℃ for 10 h in air.
SEM images of the fractured cross-sections of BZY and BZI samples are shown in Fig. 2. BZY91 and BZI91 have small grains with porosity. Dense microstructure with uniform grain size was obtained with BZY82 and BZI82 samples. The increase in the doping level of either Y or In cause well developed grains with uniform grain size and high density. This tendency is consistent with those reported on BZY7). Doping of yttrium and indium showed similar effect on the enhancement of sintering and the enhancement is more pronounced in the case of In doping as recognized in Fig. 2, although the difference is not significant.
SEM images of fractured cross-sections of (a) BZY91, (b) BZY82, (c) BZI91 and (d) BZI82 as sintered disk.
The Arrhenius plots of the electrical conductivity of BZY and BZI in moist 1% H2 atmosphere are shown in Fig. 3. The plots are not straight but bending particularly in the case of BZY82. This bending feature is due to the change in the process dominating the electrical conductivity17,18): temperature dependence of the grain boundary resistance is larger than that of the bulk resistance. Since the two resistances are in series, the larger component dominates the total resistivity, and as a result the grain boundary dominates the total resistivity (conductivity) at low temperature with a higher slope in the Arrhenius plots. As temperature increases, the contribution of the bulk resistivity increases, and hence the slope of the plots would decrease. These temperature dependent changes between the bulk and grain boundary contributions would result in the bending of the Arrhenius plots17,18). The activation energy calculated from the slope of the plots in the temperature range of 300–400℃ which would thus be grain boundary dominant region is listed in Table 1. It is notable that the activation energy of BZI, which is around 0.46–0.52 eV, is markedly smaller than that of BZY, 0.75–0.82 eV. Thus it can be concluded that the activation energy of grain boundary conductivity is smaller than that of BZY. It is generally supposed that the origin of the grain boundary resistance is the electric potential barrier developed by the electric double layer (space charge layer)11,12). The present result suggests that the formation of such space charge is more significant in the case of BZY than in the case of BZI.
The temperature dependence of the electrical conductivity of BZY and BZI in moist 1% H2 atmosphere (pH2O = 1.9 kPa).
On the other hand, at the higher temperature region where the bulk conductivity would become dominant, the conductivity of BZY is higher than that of BZI. This suggests that Y doping is advantageous in increasing the bulk conductivity compared to In doping in consistent with the idea shown by Kreuer that Y is the best dopant for barium zirconate2,15).
In the above discussion, grain boundary and bulk dominant temperature regions have been discussed only from the slope of the Arrhenius plots. Contribution of these may usually be more clearly discussed by impedance analysis. In order to measure more intrinsic total conductivity, the 4-probe AC impedance method was applied in this study. Given the sample shape and configuration in this study, however, it is difficult to separate between the bulk and grain boundary with the 4-probe method.
As discussed above, BZY has high bulk conductivity and BZI is characterized by low activation energy of grain boundary conductivity. Thus co-doping of Y and In, i.e. BZIY, will bring us interests on how these properties change on the co-existence of the two dopant species. Figure 4 shows the XRD patterns of BZIY811, BZIY712 and BZIY721 pellets after sintered at 1600℃ for 10 h. The XRD results show the cubic structure without secondary phase. The lattice parameter and relative density are listed in Table 2. With respect to the BZIY system, the lattice parameter size are in the order of BZIY712 > BZIY721 > BZIY811. Figure 5 shows SEM observation of the fractured cross section of the specimens. Good grain growth was observed in BZIY712 and BZIY721. It appears that the high doping concentration promotes grain growth of BZIY.
XRD patterns of BZIY electrolytes as sintered at 1600℃ for 10 h in air.
SEM images of fractured cross-sections of (a) BZIY811, (b) BZIY721 and (c) BZIY712 as sintered disk.
The Arrhenius plots of the electrical conductivity of BZIY in moist 1% H2 are shown in Fig. 6. As mentioned above, the bending feature of the electrical conductivity is due to the change in bulk (above 500℃) and grain boundary (below 400℃) dominating of the electrical conductivities, and here how these conductivities are affected by the co-substitution is discussed. BZIY811 is the sample in which the half of In in BZI82 is substituted by Y. In response to this substitution, the electrical conductivity in the region dominated by bulk conductivity became higher than that of BZI82: the conductivity is higher than the logarithmic average of the conductivities of BZI82 and BZY82 in the bulk-dominant temperature region (above 500℃). On the other hand, the activation energy of the grain boundary conductivity of BZIY811, 0.58 eV, is smaller than the average of the activation energies of BZI82 and BZY82. It is apparent that the co-substitution of Y and In synergistically works, i.e. it results in enhancing the bulk conductivity and lowering the activation energy of grain boundary conductivity from these averages.
The temperature dependence of the electrical conductivity of BZIY in moist 1% H2 atmosphere (pH2O = 1.9 kPa).
Then either Y or In doping is increased by another 10% from BZIY811, resulting in the compositions of BZIY712 and BZIY721, respectively. In BZIY712, increasing amount of dopant leads to the increase in the bulk conductivities, however, the activation energy in grain boundary conduction also increased. In comparison with the conductivity of BZY82, the conductivity of BZIY712 is almost the same. A small decrease at high temperature suggests that In doping cause a small degradation of the conductivity of BZY82. On the other hand, BZY721 showed inferior properties in both the bulk conductivity and activation energy of the grain boundary conductivity to those of BZY811. As a result, BZIY721 electrolyte show slightly decreased conductivities than that of BZIY811 in all examined temperature region. From these results, the co-doping may result in degradation of the absolute values of conductivity when the original conductivity is already high. However, the present results suggest that the co-doping of yttrium and indium can work for specific purposes of controlling bulk and/or grain boundary conductivities by choosing each doping level appropriately.
The electrical properties of trivalent-cation-doped BaZr1−x−yInxYyO3−δ (x = 0, 0.1, 0.2 and y = 0, 0.1, 0.2) have been investigated. Comparison between the properties of BZY and BZI suggests that BZY is characterized by higher bulk conductivity, but BZI has lower activation energy of the grain boundary. Co-doping of Y and In promotes the enhancement of the bulk conductivity and lowering of activation energy of the grain boundary conductivity in the inferior materials. Although, no example could be found in increasing the conductivity which is already high, such co-doping possibly control the specific bulk-grain-boundary conductivity feature of the materials.
This work was supported by International Institute for Carbon Neutral Research (I2CNER) and The World Premium International Research Center Initiative (WPI), Japan.