© The Author(s), published by EDP Sciences, 2022

1 Introduction

Solid Oxide Fuel Cell (SOFC) technology has great potential in the power generation field. The structures of its components are simple and because SOFC technology can be used in high temperatures (650–1000 °C), it does not require use of expensive platinum as a catalytic converter because cathode and anode materials can be used at relatively low costs instead. Furthermore, SOFC technology has a high efficiency, relatively low pollution emissions, such as NOx, SOx, and HC, and the produced CO₂ can be easily captured [1, 2]. Additionally, the materials that can be used as fuels in an SOFC are diverse, including natural gas, CO, H₂, methanol, coal gas, and even flammable exhaust gases.

To popularize the use of SOFCs, some components still require further improvement. For instance, (Y_0.08Zr₀.₉₂)O_1.96 (8 mol% yttria-stabilized zirconia, 8YSZ) is currently a commercially available electrolyte ceramic material that has a low conductivity of 0.1 S cm⁻¹ under operational temperatures of 1000 °C [3–5]. The use of a SOFC under high temperatures has the following disadvantages: (1) The composition interface is prone to reacting. (2) There can be a reduction in thermal stability. (3) Interfacial bonding is affected by the coefficient of expansion and easily results in battery cracking caused by thermal stress. (4) To achieve stable usage, electrode and bipolar plate materials with relatively high costs are required, directly affecting the manufacturing costs of the SOFC [6–10]. Electrolyte materials currently in use only possess appropriate levels of conductivity in high temperature environments. Therefore, the development of an electrolyte material with a high ionic conductivity in low temperature environments (under 600 °C) is an essential research topic in SOFC development.

In operational environments with temperatures below 600–800 °C, Rare-earth Doped CeO₂ (RDC) electrolyte materials can achieve conductivity values similar to that of 8YSZ at 1000 °C. Therefore, RDC is viewed as the electrolyte material with the greatest potential for medium to low temperature applications. However, a portion of the Ce⁴⁺ may be reduced to Ce³⁺ in doped CeO₂ oxides during high-temperature sintering, which may result in a mixed-valence state. After the resulting material is used to assemble a cell component, if the operating temperature surpasses 550 °C and the material comes into contact with an anode reduction atmosphere, the formation of a Ce⁴⁺/Ce³⁺ mixed-valence state is accelerated and the electrical conductivity is increased, which results in a decline in the open circuit voltage and operating power density [11–14].

In order to stop the Ce valence change, two electrolyte materials, a carbonate mixture and oxygen ion conductor, were combined to produce a carbonate-oxide composite; under 500 °C, it achieves a conductivity value similar to that of 8YSZ at 1000 °C and suppresses the Ce element valence change, which is an achievement that has attracted remarkable attention. This study builds upon previous research by synthesizing a carbonate-oxide composite electrolyte material through a bath method in order to ensure a high degree of density [15]. Furthermore, a series of doped CeO₂ oxides of (La_0.2Ce_0.8)O_1.9 (LaDC), (Sm_0.2Ce_0.8)O_1.9 (SDC), (Gd_0.2Ce_0.8)O_1.9 (GDC), and (Gd_0.05Y_0.15Ce_0.8)O_1.9 (GYDC) were synthesized for complete physical property analyses and discussion. The lattice parameter, microstructure and Ce⁴⁺/Ce³⁺ mixed-valence state in the RDC oxides and carbonate-oxide composite materials are shown in this research.

2 Materials and methods

The chemicals listed in Table 1 were processed with wet ball-milling to produce the LaDC, SDC, GDC, and GYDC sample powders. The manufacturing process was as follows: 3 wt.% binder (PMMA), 1 wt.% dispersant (corn oil), and 1 wt.% plasticizer (glycerin) were added to each sample powder. The resultant mixture was placed in a tungsten carbide grinding jar. A suitable amount of zirconia balls with 1 cm diameters and a volume ratio of 1:1 relative to the sample powder were added. Acetone was added, and the resulting mixture was ground with a planetary ball mill for 24 h. After filtering out the zirconia balls and placing the solution into an oven to dry the acetone, an agate mortar was used to evenly grind the dry powder. A uniaxial pressure of 840 MPa was then applied to transform the powder sample into a pellet. The pellet was placed into a furnace that increased 100 °C/h until 500 °C, after which the furnace idled for 2 h to remove any remaining organic additives. The ThermoGravimetric and Differential Thermal Analysis (TG/DTA) results of the binder, dispersant, and plasticizer are shown in Figure 1. The heat program was used by the heating rate of 5 °C/min from room temperature to 800 °C, and the N₂ gas was employed to measurement process by 20 mL/min. These results indicate that these organic additives can be completely decomposed at 500 °C. Then, the furnace continued to increase its temperature by 100 °C/h until it reached the maximal sintering temperature of 1500–1600 °C, at which point, the furnace maintained the temperature to complete sintering over 5–10 h. Finally, the furnace lowered its temperature by 100 °C/h until it reached 1000 °C, at which point, the furnace was turned off to allow for natural cooling.

Fig. 1 TG/DTA diagrams of (a) binder (PMMA), (b) dispersant (corn oil), and (c) plasticizer (glycerin).

The theoretical density of each sample was calculated by first obtaining the theoretical molar volume: a³ × 10⁻²⁴ × 6.02 × 10²³ cm³/Z, in which a (Å) is the sample unit cell parameter. If the sample exhibited a fluorite structure, Z equals 1, whereas if the sample exhibited a pyrochlore structure, which is the superstructure of fluorite, Z equals 8. This is because the entire cells were A₁₆B₁₆O₅₆ and the typical A₂B₂O₇ is 1/8 of the unit cell. Then, the molar weight of the sample was divided by the theoretical molar volume to obtain the theoretical density of each sample. Next, a scale was used to measure the actual weight of each sample, after which the Archimedes principle was used to determine the actual volume, from which the actual density was obtained. Lastly, the actual density was divided by the theoretical density to find the relative density of the sample.

The manufacturing process of composites was as follows. First, the RDC pellet was sintered at the appropriate temperature into a porous oxide bulk material that has a relative density of approximately 60–70%. Then, Li₂CO₃ and Na₂CO₃ were mixed through wet-ball milling at a 1:1 molar ratio to produce a carbonate mixture referred to as LNCO. An appropriate amount of LNCO was then placed into an SUS304 stainless steel container that was heat resistant up to 900 °C. The container was heated to 500 °C for 24 h to produce a carbonate liquid. A porous oxide pellet was then heated with the LNCO to 500 °C for 12 h to allow the melted carbonate to penetrate and fill the pores of the pellet in order to produce a high-density composite. Zhu et al. found that the conductivity of the carbonate-oxide composite increases with increasing the amount of carbonate mixture until the carbonate volume makes up 30% of the composite [16]. Therefore, this study utilized sintering temperatures to make oxide bulk materials with porosity values between 30% and 40%, which is equal to the ratio of carbonate volumes in the samples.

Powder diffraction data used for the Rietveld refinement were collected to determine the crystal structure of the sample oxide. The General Structure Analysis System developed by Larson and von Dreele was employed to analyze the structural parameters using the Rietveld method [17]. The morphology of the sample was observed using a VEGA SBH scanning electron microscope. The X-ray Absorption Near Edge Structure spectroscopy (XANES), BL16A1 wiggler beam line was used to observe the Ce⁴⁺/Ce³⁺ mixed-valence state in both the RDC oxides and carbonate-oxide composite materials.

3 Results and discussion

Figure 2 illustrates the XRD diagram for LaDC, SDC, GDC, and GYDC pure oxide samples. Here, GDC and GYDC were sintered at 1500 °C for 5 h, while LaDC and SDC were sintered at 1600 °C for 5 h. These temperatures were selected so that the pure oxide sample bulks would exceed the relative density of 98.3(4)%, which meant that the density would be as high as an SOFC solid electrolyte. The sintered RDC series samples were all fluorite single-phase samples. As can be observed from Figures 2 and 3, the diffraction signal angles gradually shift toward the right from LaDC, SDC, GDC, to GYDC because of the respective eight coordination La³⁺, Sm³⁺, Gd³⁺ and Y³⁺ ionic radii of 1.16, 1.079, 1.053, and 1.019 Å [18]. This trend matches the Bragg diffraction formula of nλ = 2dsinθ. Figure 3 shows the XRD diagram for the LaDC, SDC, GDC, and GYDC pure oxide samples sintered at 1000 °C for 10 h. These samples also exhibited fluorite single-phase structures. Figure 4 displays the XRD diagram of the carbonate-oxide composites and LNCO. It shows the signal positions of the Li₂CO₃ (7.12° and 10.64°) and Na₂CO₃ (8.90°, 11.13°, 11.47°, and 14.09°) [15]. As the carbonate structures were amorphous, their signal range was wider and less apparent [6].

Fig. 2 XRD diagram of the RDC series oxides sintered at 1500–1600 °C for 5 h. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom).

Fig. 3 XRD diagram of the RDC series oxides sintered at 1000 °C for 10 h. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom).

Fig. 4 XRD diagram of the RDC series composites. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom). The symbol (●) in the diagram indicates the Li2CO3 signal, whereas (■) indicates the Na2CO3 signal.

The XRD angle of diffraction data from the RDC samples was refined to obtain the structural parameters for calculating the relative density of the samples and to evaluate their effect on sample physical properties. Figure 5 shows the Rietveld refinement results for the RDC series samples sintered at 1500–1600 °C. Table 2 lists the RDC series sample lattice parameters obtained from the Rietveld refinement. Note that the χ² values are all between 0.5 and 5 and the error values are less than 5%, indicating that the sample calculations are credible. Figure 6 shows the refined lattice parameters plotted against the doped rare earth ionic radii. The sample lattice parameters, including LaDC, SDC, GDC, and GYDC, exhibit a gradually decreasing trend that is consistent with Vegard’s Law. These results indicate that, for this series, the consistency of each step in the manufacturing process of the samples is high and human error is relatively small.

Fig. 5 The RDC series Rietveld refinement results for sample powders sintered at 1500–1600 °C, (a) LaDC, (b) SDC, (c) GDC, and (d) GYDC.

Fig. 6 LaDC, SDC, GDC, and GYDC lattice parameters plotted against R3+ ionic radius.

The heat treatment temperatures and relative densities of the RDC series of high-density oxides, porous oxides, and composite samples are listed in Table 3. As shown in the table, both LaDC and SDC sintered at 1600 °C, and GDC and GYDC sintered at 1500 °C produced high sample densities (>98.3(4)%). Furthermore, 1000 °C sintering for 10 h produced densities between 63.6(2)% and 68.4(5)%, which correspond to porosities between 31.6% and 36.4%. After the porous samples were made into composites, the sample densities approached 100%, meaning that the melted carbonate filled almost all the sample pores. A necessary condition for a solid electrolyte is high relative density close to 100%. This study verified the relative density of composites, as shown in Table 4. The composite made of a porous oxide with a known relative density, then used sandpaper to remove the excess surface carbonate of the composite bulk and measured it to determine weight W₁. Then, boiling water was used to remove the carbonate in the composite, which was then dried prior to reweighing to obtain the weight W₂ of the remaining oxide bulk. Therefore, W₁ − W₂ was the weight of the carbonate filling into the porous oxide bulk. As LNCO is composed of Li₂CO₃ and Na₂CO₃ with a molar ratio of 1:1, the respective weights, W₃ and W₄, in the composite can be obtained. The densities of Li₂CO₃ and Na₂CO₃ were used to obtain their respective volumes V₄ and V₅ in the composite filling. The volumes of Li₂CO₃ and Na₂CO₃ (V₄ + V₅) approached volume V₃ of the oxide bulk material pores, indicating that the carbonate almost completely filled the pores. Furthermore, the relative density of the oxide can be used to obtain the volume of the oxide V₂ in the composite. The sum of the two carbonate volumes V₄ + V₅ plus the volume occupied by the oxide V₂, divided by the volume of the composite V₁ is a value closely approaching 1. These results confirm that this calculation process and its results can be used to quantify the relative density of a carbonate-oxide composite bulk.

Scanning Electron Microscope (SEM) microstructure photographs help visualize the results of the high-density composite calculations. Figure 7 shows the SEM images of the RDC series high-density oxide, porous oxide, and composite samples magnified 5000 times. As shown in Figures 7a–7d, pores are barely visible in oxides that have been sintered at high temperatures (1500–1600 °C), whereas numerous pores can be seen in the oxide samples sintered at 1000 °C in Figures 7e–7h. Figures 7i–7l depict the SEM diagrams of the composite samples and clearly illustrate that the vast majority of pores in the porous oxide samples have been filled.

Fig. 7 RDC sintered at 1500–1600 °C for 5 h; sintered at 1000 °C for 10 h; SEM diagram of carbonate composite magnified 5000 times. (Here, (a), (e), (i) have R = La; (b), (f), (j) have R = Sm; (c), (g), (k) have R = Gd; (d), (h), (l) have R = Gd + Y.)

Numerous studies have indicated that after CeO₂-doped oxide electrolyte materials are assembled into SOFC unit cell components, the reduction of Ce⁴⁺ into Ce³⁺ results in a mixed-valence state. This study proposes that materials of this type are likely to begin producing mixed-valence states during high-temperature sintering densification process. Operational temperatures higher than 550 °C and the effects of the reductive environment result in an even more apparent mixed-valence phenomenon. The resulting electronic conductivity is increased, causing lower open circuit values, which adversely affect the operation power density of the SOFC single cell. Therefore, this study utilized typical high-temperature sintering densification on the RDC oxides, low-temperature heat-treated samples, and sample powders resulting from composite production, with Ce(NO₃)₃ and CeO₂ as reference standards to analyze the average Ce valence.

The RDC series oxide measured by XANES spectrum approached that of the CeO₂ reference standard, exhibiting single maximum differential energy shoulders that were almost identical. Thus, this study used the strength “1” position as the RDC series reference line, as shown by the horizontal lines in Figure 8. Each energy shoulder at the intersection between each energy reference line and spectrum was used to infer the Ce element valence of each sample, as shown in Figure 9. The analysis results indicated that the RDC samples that underwent high temperature treatments exhibited lower Ce valences with an average of 3.93(2), whereas the RDC samples treated at 1000 °C temperatures exhibited higher Ce valences with an average of 3.96(1). Thus, a multi-valence element, like Ce, might tend to reduce its state during higher heat-treated temperatures. The RDC oxides were first sintered at 1000 °C to get a porous sample bulk, followed by 500 °C treatments with the LNCO carbonate mixture. The resultant average valence of Ce in the composite was 3.96(1), which is similar to the RDC oxides treated at 1000 °C. The Ce valence in the RDC oxides with different heat-treated temperature and in the carbonate-oxide composites measured by XANES are listed in Tables 5–7. As seen in the results, the Ce valence in the RDC oxide would not be affected by combining with carbonate. In other words, using the LNCO carbonate and RDC oxide to form a composite creates a stable electrolyte material that can be applied in low temperature SOFCs.

Fig. 8 (- – -) Ce(NO3)3, (‒) CeO2, (□) LaDC, (○) SDC, (∆) GDC, and (◊) GYDC samples with 1500–1600 °C (a), 1000 °C (b), and composite samples with 500 °C (c) treatment, Ce L-edge XANES spectrum after normalization.

Fig. 9 Ce(NO3)3 and CeO2 reference standard, (□) LaDC, (○) SDC, (∆) GDC, and (◊) GYDC samples treated with 1500–1600 °C (a), 1000 °C (b), and composite samples with 500 °C (c); valence plot of maximal first order differential values of Ce element.

4 Conclusion

The XRD structure appraisal and Rietveld refinement results indicated that the RDC oxide samples produced through the solid-state synthesis method possess a single-phase fluorite phase. When the samples were assembled into bulk materials, sintering them at 1500–1600 °C achieved relative densities exceeding 98.3%, whereas sintering at 1000 °C for 10 h maintained densities between 63.6(2)% and 68.4(5)%, indicating porosities between 31.6% and 36.4%. The bath method was used to heat porous oxide bulk materials with a carbonate mixture, LNCO, to 500 °C for 12 h, then allowing the melted carbonate to fill the oxide pores and produce a dense carbonate-oxide composite electrolyte. Boiling water was used to remove the carbonate. Then, structural refinement was used to obtain the lattice parameters for the pore and oxide volume calculations. The volume of LNCO used to fill the pores approached the volume of the pores of the RDC oxide sintered at 1000 °C, indicating that the melted carbonate permeated and filled the oxide pores. This result was supported by the SEM observations. The XANES spectrum analysis results indicated that the RDC samples sintered at 1500–1600 °C exhibited a lower average Ce valence of 3.93(2), while samples treated at 1000 °C exhibited a higher average Ce valence of 3.96(1), which is identical to that of the composite samples. These results illustrate that a carbonate-oxide composite is a stable electrolyte material for manufacturing under lower temperatures to further inhibit the development of Ce⁴⁺/Ce³⁺ mixed-valence situations. The composite electrolyte has potential for low temperature SOFC operation under 500 °C. This can increase the working performance and decrease the degradation rate of the peripheral components. Thus, the fabrication cost can be further decreased to improve market acceptance.

Conflicts of interest

The authors declare that they have no conflict of interest.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Authors’ contributions

Conceptualization, Yong-Xin Liang, Xin Wu, L.C. Wen, and Gengyu Cao; methodology, Yong-Xin Liang, L.C. Wen, and Ze-Rong Ma; software, Yong-Xin Liang, Rui-Sen Huang, and Si-Ting Yu; validation, Xin-Yue He, L.C. Wen, and H.M. Liu; formal analysis, Yong-Xin Liang, L.C. Wen, and Xu-Yang Ke; investigation, Yong-Xin Liang, Ze-Rong Ma, L.C. Wen, and Ri-Feng Yan; resources, Yong-Xin Liang, L.C. Wen, and Xiao-Xian Liang; data curation, L.C. Wen and Gengyu Cao; writing – original draft preparation, Yong-Xin Liang, L.C. Wen; writing – review and editing, Yong-Xin Liang, Ze-Rong Ma, L.C. Wen, and Gengyu Cao; visualization, Yong-Xin Liang, L.C. Wen; supervision, L.C. Wen; project administration, L.C. Wen and Gengyu Cao; funding acquisition, Gengyu Cao. All authors have read and agreed to the published version of the manuscript.

Acknowledgments

This research was funded by the Development and Testing of ITSOFC Single Cell, which belongs to the Skilled Personnel Recommend Programme of Guangdong University of Petrochemical Technology, grant number 702-519150.

Funding

This research was funded by the Development and Testing of ITSOFC Single Cell, which belongs to the Skilled Personnel Recommend Programme of Guangdong University of Petrochemical Technology, grant number 702-519150.

References

All Tables

All Figures

	Fig. 1 TG/DTA diagrams of (a) binder (PMMA), (b) dispersant (corn oil), and (c) plasticizer (glycerin).
In the text

	Fig. 2 XRD diagram of the RDC series oxides sintered at 1500–1600 °C for 5 h. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom).
In the text

	Fig. 3 XRD diagram of the RDC series oxides sintered at 1000 °C for 10 h. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom).
In the text

	Fig. 4 XRD diagram of the RDC series composites. The angle of diffraction 2θ values are for X-wavelength 0.15418 nm (top) and 0.0495941 nm (bottom). The symbol (●) in the diagram indicates the Li2CO3 signal, whereas (■) indicates the Na2CO3 signal.
In the text

	Fig. 5 The RDC series Rietveld refinement results for sample powders sintered at 1500–1600 °C, (a) LaDC, (b) SDC, (c) GDC, and (d) GYDC.
In the text

	Fig. 6 LaDC, SDC, GDC, and GYDC lattice parameters plotted against R3+ ionic radius.
In the text

	Fig. 7 RDC sintered at 1500–1600 °C for 5 h; sintered at 1000 °C for 10 h; SEM diagram of carbonate composite magnified 5000 times. (Here, (a), (e), (i) have R = La; (b), (f), (j) have R = Sm; (c), (g), (k) have R = Gd; (d), (h), (l) have R = Gd + Y.)
In the text

	Fig. 8 (- – -) Ce(NO3)3, (‒) CeO2, (□) LaDC, (○) SDC, (∆) GDC, and (◊) GYDC samples with 1500–1600 °C (a), 1000 °C (b), and composite samples with 500 °C (c) treatment, Ce L-edge XANES spectrum after normalization.
In the text

	Fig. 9 Ce(NO3)3 and CeO2 reference standard, (□) LaDC, (○) SDC, (∆) GDC, and (◊) GYDC samples treated with 1500–1600 °C (a), 1000 °C (b), and composite samples with 500 °C (c); valence plot of maximal first order differential values of Ce element.
In the text