Introduction

Synthesis of the composites based on doped BaCeO3 is one of the main concepts concerning improvement of the chemical stability of BaCeO3-based materials towards CO2/H2O containing atmosphere. It also results with modification of the materials microstructure and functional properties, mainly the transport properties—the mechanism and the value of conductivity. The synthesis of the composites is usually undertaken for doped barium cerate—doped ceria system, as the dopant yttrium, samarium, gadolinium or neodymium is frequently chosen [1,2,3,4,5,6,7,8,9]. However, the modification of doped BaCeO3 with carbonates [10], tungstates [11] or phase rich in iron, nickel or titanium [12,13,14] was also reported.

In the area of the composites based on the perovskites, solid-state reaction or sol–gel methods are commonly used [3, 6, 8, 9, 15]. However, spark plasma sintering, chemical solution deposition, co-precipitation, citrate–nitrate combustion or impregnation can also be applied [1, 4, 11, 16, 17]. Two different approaches in the composites synthesis are usually undertaken. In the first one, the assumption is the single-step synthesis. In the second one, composite components are separately synthesised and the composite formation is performed by direct sintering or impregnation. For both cases, usually at least one of the synthesis steps requires the use of high temperature. Thus, undesirable processes connected with the mutual reactivity between the composites components and the co-doping of the perovskite structure materials can be observed, leading to the change of the materials defects structure [3, 5]. As the result, the decrease in the total conductivity and an increase in the activation energy of the conductivity are often seen [1,2,3,4]. In this work, an introduction of an additional phase into the sintered Y-doped BaCeO3 was performed by low-temperature impregnation of the sintered perovskite with the solution of V2O5 precursor. Thus, only the solvent evaporation and the precursor decomposition at temperature below 500 °C had to be performed. Necessity of application of the relatively low temperatures should reduce cations interdiffusion and the materials co-doping, as well as the mutual reactivity between the components of the potential composite. Introduction of an additional phase often leads to the increase in the BaCeO3-based materials chemical stability in corrosive CO2/H2O-rich atmospheres [1, 4, 11], but the decrease in electrical conductivity due to high contact resistance between both composites phases was also reported [2, 4]. Thus, vanadium(V) oxide was chosen as the modifier since its positive effects on both stability and conductivity of phosphate glasses were previously observed [18, 19].

Experimental

Single-phase Y-doped BaCeO3 (BaCe0.9Y0.1O3−δ) was synthesised by solid-state reaction method. Barium carbonate, cerium(IV) oxide and 0.25 M water solution of yttrium nitrate (prepared based on Y(NO3)3 6H2O) were used as the reagents. The BaCO3 and CeO2 powders’ mixture was impregnated with the proper amount of yttrium nitrate solution and dried for 12 h at 80 °C. The received powder was calcined at 1200 °C for 24 h and then homogenized by milling in the absolute alcohol suspension using a rotation-vibration mill (operated in clockwise and anticlockwise rotation, total time of milling 1 h, speed 400 rpm) with ZrO2 grinding media and dried at 90 °C for 24 h. The final powder was formed in the pellet dies (ø = 10 mm, thickness about 4 mm) at 2.5 MPa and sintered at 1300 °C for 3 h in air. Surface of the sintered BaCe0.9Y0.1O3−δ pellets was ground using diamond paper and subsequently polished prior to all further investigations and synthesis steps in order to remove the CeO2-rich phases formed on the surface as the result of barium oxide evaporation during sintering. The total porosities of BaCe0.9Y0.1O3−δ samples determined on the basis of geometric measurements and samples’ mass were equal to 32 ± 1 vol%. The vanadyl acetylacetonate dissolved in methanol (the solutions’ concentration 0.35 M) was applied as the precursor of V2O5. Wet vacuum impregnation method was applied to introduce the V2O5 precursor solution into the sintered BaCe0.9Y0.1O3−δ.

Phase composition and unit cell parameters were determined based on XRD measurements taken on Philips X’Pert Pro diffractometer with monochromatized Cu-Kα. The Rietveld refinement was used for quantitative analysis. Scanning electron microscopy (SEM) (Nova NanoSEM 200 FEI) coupled with X-ray energy-dispersive spectroscopy (EDAX company apparatus) was used to observe the samples morphology and to determine their chemical composition. These measurements were taken on the samples polished cross sections.

The exposition test: long term (700 h), low-temperature (25 °C) exposure for CO2/H2O atmosphere (10% CO2 and 100% HR) was performed on the received materials in order to evaluate the chemical stability towards carbon dioxide and water vapour. The un-modified BaCe0.9Y0.1O3−δ and materials modified by impregnation in the form of pellets were exposed to this corrosive atmosphere in the custom-made degradation chamber. To analyse the exposition test results, samples before and after the test were submitted for thermogravimetric measurements taken on SDT2960 TA Instruments apparatus. The samples with mass about 50 mg were placed in the platinum crucibles and heated in synthetic air atmosphere with 10 deg min−1 heating rate.

The DC resistance and Seebeck coefficient measurements were performed in a Fine Instruments apparatus. The measurements were taken by Keysight Technology multimeter in fully automatic system in synthetic air as a function of temperature (250–700 °C). Before each measurement, the samples were stabilized at given temperature for 30 min. The conductivity was calculated based on measured resistance and sample geometry.

Results and discussion

Sample preparation

Wet impregnation under pressure method was applied to introduce additional components into the intergranular voids of the sintered single-phase BaCe0.9Y0.1O3−δ. The selection of the proper solution for impregnation was performed experimentally. The relatively high solubility in non-aqueous solvent was the main requirement, since the chemical resistance of barium cerate towards water is significantly small. The vanadyl acetylacetonate was chosen as the V2O5 precursor since it decomposes in air at relatively low temperature (about 400 °C) with formation of vanadium(V) oxide [20, 21]. Methanol was used as a solvent since it allowed a solution with relatively high concentration (0.35 M) to be obtained. Concerning the relatively high porosity of sintered BaCe0.9Y0.1O3−δ (above 30%), the impregnation was performed under the pressure of 0.1 MPa. The higher pressure could lead to the mechanical destruction of the samples, especially that multi-step, multi-cycle impregnation was applied. Figure 1 presents the one cycle of impregnation. The BaCe0.9Y0.1O3−δ samples before impregnation were dried (120 °C for 12 h) to remove the moisture and weighted. Single-phase BaCe0.9Y0.1O3−δ samples impregnated with the solution of V2O5 precursor in methanol were drying at 80 °C for 12 h in order to evaporate the solvent. Then they were one more time impregnated (2nd step) and dried (80 °C, 12 h). This step was followed by 3rd impregnation and another drying. After three steps of impregnation, samples were annealed at 450 °C for 1 h to decompose vanadyl acetylacetonate to vanadium(V) oxide. This procedure was considered as the one cycle of impregnations (I impregnation cycle). For some samples, this procedure was repeated for the second time and for the third time. Thus, samples after I impregnation cycle (3 impregnations), II impregnation cycle (6 impregnations) and III impregnation cycle (9 impregnations) were obtained. In the following part of the paper, these samples will be labelled: BCYO_VO1, BCYO_VO2 and BCYO_VO3, respectively. For the reference single-phase BaCe0.9Y0.1O3−δ, the acronym BCYO will be used. The drying and annealing conditions were determined based on recorded TG curves.

Fig. 1
figure 1

Scheme of the one cycle of impregnation (I impregnation cycle) of BaCe0.9Y0.1O3−δ with methanol solution of vanadyl acetylacetonate

Full size image

To monitor the amount of compound introduced into the BaCe0.9Y0.1O3−δ, all samples were weighted after each drying and annealing process. The assumption was to synthesise the materials in BaCe0.9Y0.1O3 - V2O5 system; thus, the recorded masses were used to evaluate the nominal composition of the received materials (Table 1).

Table 1 Nominal composition of received BaCeO3-based materials
Full size table

Structure, phase composition and microstructure

Single-phase BaCe0.9Y0.1O3−δ with orthorhombic structure (space group Pnma, ICSD card no. 98-007-9625) was impregnated with the methanol solution of V2O5 precursor. In Fig. 2, the XRD results of the analysis of samples’ surface after I, II and III impregnation cycle are presented together with the results for BaCe0.9Y0.1O3−δ sample before impregnation, as the reference. The analysis of diffractograms showed, in the case of all synthesized materials, the presence of orthorhombic (space group Pnma) BaCeO3 as the main phase. Additional phases tetragonal CeVO4 (space group I41/amd, ICSD card no. 01-072-0282) and orthorhombic BaV2O6 (space group C222, ICSD card no. 01-086-0240) were observed for BCYO_VO2 and BCYO_VO3 materials. The presence of these phases clearly indicates the decomposition of Y-doped BaCeO3 in contact with methanol solution or during the precursor decomposition. The phase composition of the obtained materials in mole fraction, calculated based on the results of the Rietveld refinement, is presented in Table 2.

Fig. 2
figure 2

XRD results of the initial BaCe0.9Y0.1O3−δ sample and BaCe0.9Y0.1O3−δ samples after impregnation

Full size image
Table 2 Phase composition of modified BaCeO3-based materials
Full size table

The increase in the content of the additional phases with the impregnation cycles suggests progressive decomposition of the perovskite. The relatively high content of CeVO4 and BaV2O6 in comparison with assumed materials’ composition indicates their segregation on the samples surface.

The BaCeO3 lattice parameters for initial single-phase BaCe0.9Y0.1O3−δ sample and impregnated BaCe0.9Y0.1O3−δ samples (Table 3) are significantly comparable. Thus, eventual co-doping of Y-doped BaCeO3 with vanadium or other changes in the chemical composition of the perovskite is excluded.

Table 3 Lattice parameters of the initial and modified BaCe0.9Y0.1O3−δ
Full size table

SEM microphotographs of pellets polished cross sections recorded for modified BaCe0.9Y0.1O3−δ samples show porous materials with irregular pore structure (Fig. 3). EDS spectra indicate the presences of vanadium at the surface and in the interior parts of the materials. For the surface part of the samples, the increase in the number of the impregnation cycles leads to higher vanadium content. It stays in agreement with the results of XRD measurements indicating segregation of vanadium rich phases on the samples surface. For all modified BaCe0.9Y0.1O3−δ materials, the small vanadium amount was also observed in the interior part of the materials. It suggests that applied impregnation method allows locating the additional vanadium-based phase not only on the materials surface. However, it must be noticed that due to the limitations of EDS method presented values should be treated for comparison only but not as absolute values.

Fig. 3
figure 3

SEM microphotographs of the cross section of the modified BaCe0.9Y0.1O3−δ samples together with results of the EDS analysis

Full size image

Chemical stability

It is well known that the main disadvantage of Y-doped BaCeO3 is the poor chemical stability in the CO2 and water vapour-rich atmosphere [22,23,24]. Thus, the good electrical properties of these materials, including the proton conductivity, cannot be fully exploited since the materials relatively easily undergo chemical degradation resulting in the decrease in both mechanical and transport properties. To evaluate the chemical stability of synthesised materials, samples were exposed to carbon dioxide and water vapour-rich atmosphere (10% CO2, 100% HR). During this long time (700 h), low-temperature (25 °C) exposition physical processes, e.g., adsorption of H2O and/or CO2 and chemical reaction between the material and a surrounding atmosphere can process. As it was discussed in the literature, Y-doped BaCeO3 in CO2/H2O atmosphere can undergo degradation leading to the formation of BaCO3 and Ba(OH)2 [25, 26]. Thus, the phase composition of the material before and after the exposition can be different. TG curves for single-phase BaCe0.9Y0.1O3−δ and BaCe0.9Y0.1O3−δ modified by impregnation before and after the exposition test are presented in Fig. 4.

Fig. 4
figure 4

TG curves for initial BaCe0.9Y0.1O3−δ sample and modified BaCe0.9Y0.1O3−δ samples before and after the exposition test

Full size image

For all materials before the test, only one mass loss in the temperature range 500–800 °C is observed. This effect can be attributed to degradation of the protonic defects created in the perovskite as the result of yttrium doping [27, 28]. The shape of TG curves after the exposition is significantly different, since a few additional effects are clearly seen, where desorption of H2O/CO2 and decomposition of BaCO3 formed during the explosion are the main one. The first one takes place below 400 °C, the second one at temperatures above 800 °C. This temperature is reported in the literature as the temperature of barium carbonate decomposition; however, it strongly depends on the surrounding atmosphere [29]. The total mass loss for single-phase BaCe0.9Y0.1O3−δ is 3.8%, while for modified materials the total mass loss is 2.4%, 2.2% and 1.8% for BCYO_VO1, BCYO_VO2 and BCYO_VO3, respectively. Based on the mass loss above 800 °C, the amount of BaCO3 in the samples after the exposition was calculated. The barium carbonate content was 5.9% for single-phase BaCe0.9Y0.1O3−δ and about 1.9–2.0% for modified materials. Thus, it can be postulated that BaCe0.9Y0.1O3−δ materials after modification with the solution of V2O5 precursor are not completely stable in CO2/H2O-rich atmosphere; however, their stability is higher than stability of the single-phase BaCe0.9Y0.1O3−δ. As the reason of the perovskites’ stability improvement, two factors are usually discussed, structure stabilization due to co-doping of BaCe0.9Y0.1O3−δ and the presence of an additional phase. It was reported in the literature that incorporation of even a small amount of pentavalent ion (Nb5+) into the barium cerate structure leads to significant improvement of the perovskite chemical stability due to the structure stabilization [30]. This co-doping should be followed by the change of the BaCeO3 lattice parameters, as an ion with smaller ionic radii (Nb5+, 0.78 Å CN = 6 [31]) is incorporated into Ce4+ position (1.01 Å for CN = 6 [31]). For BaCe0.9Y0.1O3−δ modified by impregnation, the change of the lattice parameters compared to single-phase BaCe0.9Y0.1O3−δ was not observed. Thus, co-doping of the perovskite structure cannot be considered as the reason of chemical stability improvement. The presence of the additional phases CeVO4 and BaV2O6 is the most probable reason of the enhancement of stability towards CO2/H2O. As was discussed based on the XRD measurements results, materials received after impregnation were multi-phase except of BCYO_VO1 sample. Concerning the comparable chemical stability of BCYO_VO1 with the stability of BCYO_VO2 and BCYO_VO3 (similar content of barium carbonate), it can be postulated that even after only one impregnation cycle partial decomposition of the perovskite occurs and the additional phases are formed; however, their content is below the XRD detection point. The small differences between the mass losses for modified materials are mainly the reason of different amount of absorbed water, as the main difference is seen for temperatures below 100 °C. It cannot be excluded that the amount of secondary phases influence the possibility of water absorption, as the smaller mass loss is observed for the material with the highest content of the additional phases (BCYO_VO3).

Electrical properties

Figure 5 presents the total electrical conductivities of modified BaCe0.9Y0.1O3−δ materials in Arrhenius coordinates, together with the results for single-phase BaCe0.9Y0.1O3−δ as the reference. Measurements were taken in synthetic air in the temperature range 250–700 °C. At these conditions ionic-hole transport occurs in Y-doped BaCeO3 with the activation energy Ea of this total conductivity about 0.7–0.8 eV [32, 33]. Thus, the conductivity results for single-phase BaCe0.9Y0.1O3−δ stay in general agreement with the literature data and the total conductivity observed is the result of oxide ion conductivity via oxygen vacancies and hole conductivity. According to the literature, CeVO4 is the mixed ionic (oxygen ion) and p-type conductor, with the activation energy of conductivity about 0.4 eV [34, 35]. For barium vanadium oxide, the polaron hopping mechanism of conductivity was reported with the activation energy 0.89 eV [36]. The electrical conductivity of BCYO_VO3 sample is one–two orders of magnitude lower than that for the reference BaCe0.9Y0.1O3−δ. However, the activation energy for BCYO_VO3 sample (0.73 eV ± 0.02 eV) is comparable with the value observed for the reference (0.82 eV ± 0.03 eV); thus, the mechanism of the conductivity is rather not affected by the presence of the additional phases, especially that in both BaCe0.9Y0.1O3−δ and CeVO4 the similar mechanism of conductivity is observed. An observed slight decrease in the activation energy can be the result of lower activation energy for CeVO4 conductivity. As was mentioned based on the XRD measurements results, the CeVO4 and BaV2O6 phases are probably located at the samples surface. Thus, this decrease in the value of conductivity can be the result of the blocking effect of these vanadium rich phases on the total materials conductivity. This effect is also seen for the second multi-phase material (BCYO_VO2); however, the observed changes are non-monotonic. It suggests that after two cycles of impregnation with the methanol solution the surface of the sample is not fully covered by an uniform layer of the additional phases or the thickness of the layer is diverse; thus, the resultant effect is observed. According to the XRD measurements results, BCYO_VO1 sample is single-phase; however, exposition test results clearly imply the presence of an additional phase/phases. The electrical conductivity results also suggest the non-single-phase material, since the conductivity is lower than that for reference un-modified BaCe0.9Y0.1O3−δ.

Fig. 5
figure 5

Electrical conductivity of initial BaCe0.9Y0.1O3−δ sample and modified BaCe0.9Y0.1O3−δ samples in Arrhenius coordinates

Full size image

Yttrium-doped BaCeO3 and CeVO4 are considered as mixed type conductors with the presence of p-type conductivity. The type of dominant carriers was confirmed by Seebeck coefficient measurements (Fig. 6). For all modified materials, similar as for reference BaCe0.9Y0.1O3−δ, the positive value of the Seebeck coefficient was observed. However, only for BCYO sample the Seebeck coefficient was measured with the high correlation (> 0.99) for the wide temperature range (450–700 °C). For BCYO_VO1, BCYO_VO2, BCYO_VO3 samples, the results with high correlation were obtained only for higher temperatures. Thus, the blocking effect related to the presence of the additional phases can also be considered.

Fig. 6
figure 6

Seebeck coefficient as a function of temperatures for initial BaCe0.9Y0.1O3−δ sample and modified BaCe0.9Y0.1O3−δ samples

Full size image

Impregnation efficiency

The sample with the highest number of the impregnation procedures (BCVO_VO3) was chosen to evaluate the impregnation efficiency. Sample dimensions after III impregnation cycle were 9.90 mm in diameter and 3.80 mm in thickness. The surface of this pellet was polished on both sides using diamond paper to reduce the thickness for about 20%. As-prepared sample was submitted for XRD and electrical conductivity measurements. The main identified phase was orthorhombic BaCeO3 (space group Pnma); however, small traces of additional phases tetragonal CeVO4 (space group I41/amd) and rhombohedral Ba3V2O8 (space group R-3 m, ICSD card no. 01-082-2057) were also observed (Fig. 7—only the position of peaks with the 100% intensity in the reference card for CeVO4 and Ba3V2O8 were indicated). The results clearly indicate that the applied impregnation method allows to introduce the precursor solution into the inner part of the sintered samples; nevertheless, it is still associated with the partial decomposition of the initial Y-doped BaCeO3. The formation of barium-rich Ba3V2O8 phase instead of BaV2O6 observed on the materials surface is probably the result of the smaller vanadium content.

Fig. 7
figure 7

Phase composition and electrical conductivity in Arrhenius coordinates of BCYO_VO3 samples after polishing

Full size image

The electrical conductivity of polished BCYO_VO3 sample was compared with the electrical conductivity of BCYO and BCYO_VO3 (Fig. 7). The general decrease in the conductivity was observed in comparison with single-phase BaCe0.9Y0.1O3−δ material; however, the materials’ conductivity was higher than for BCYO_VO3, where the relatively high content of additional phases was detected. Thus, the blocking effect of the vanadium containing phases on the total materials conductivity is clearly seen. Moreover, the values of the conductivity and the shape of ln(σT) = f(1000/T) dependency for BCYO_VO3 sample after polishing were similar to those observed for BCYO_VO1 sample. This is a further factor indicating the presence of the additional phases on the surface of the sample after only one impregnation cycle.

Conclusions

An attempt of introduction of an additional phase into the intergranular voids of sintered porous single-phase BaCe0.9Y0.1O3−δ material by wet vacuum impregnation method was undertaken. The methanol solution of vanadyl acetylacetonate as the V2O5 precursor was applied. The procedure was performed in three cycles; each cycle consisted of three steps and the proper thermal treatment was applied to decompose the vanadyl acetylacetonate to V2O5. Unfortunately, during this synthesis procedure the BaCe0.9Y0.1O3−δ undergoes partial decomposition; thus, CeVO4 and barium vanadium oxides were observed as the additional phases, located mainly on the samples surface. However, vanadium containing phases were also seen in the inner parts of the materials. The BaCeO3 lattice parameters calculated based on the XRD measurements results for initial single-phase BaCe0.9Y0.1O3−δ and for modified materials proved that co-doping of Y-doped BaCeO3 structure by vanadium does not occur. Also, the defects structure of the initial material was not changed since the presence of the proton defect was still observed for materials after impregnations.

Chemical stability of the obtained multi-phase materials towards CO2 and water vapour was higher than for reference single-phase BaCe0.9Y0.1O3−δ since the amount of BaCO3 formed during the exposition test was lower for materials received by impregnation. Moreover, the smaller ability to absorption of gaseous was observed for modified material, probably as the result of segregation of the additional phases mainly on the materials surface. It was also the probable reason of the chemical stability improvement. The slight decrease in transport properties as the result of performed impregnations was observed, since the decrease in the total conductivity for modified materials was noticed in comparison with un-modified BaCe0.9Y0.1O3−δ. Still, the mixed oxide ion and hole conductivity were observed; however, the presence of the additional phases had a blocking effect on the total materials conductivity.