BaCo0.4Fe0.4Zr0.1Y0.1O3−σ Cathode Performance for Proton Conducting Solid Oxide Fuel Cells with BaZr0.8−xCexY0.1Yb0.1O3−δ Electrolytes

BaCo0.4Fe0.4Zr0.1Y0.1O3−σ (BCFZY) is a proton, oxygen-ion, and electron-hole conducting cathode material for intermediate temperature solid oxide fuel cells. Its electrode reaction mechanism in air with moisture is not well understood. In this study, three types of symmetrical cells with the same BCFZY cathode were fabricated over three related proton conducting electrolytes: BaZr0.8−xCexY0.1Yb0.1O3−δ (x = 0.1, 0.4, and 0.7). The cathode shows similar performance over three different electrolytes in dry air but different responses to moisture introduction. The differences are hypothesized to relate to the mutual diffusion at the cathode/electrolyte interface. Such a hypothesis is supported by different techniques such as XRD Rietveld refinement of BCFZY cathode in mixtures with different electrolytes after firing, energy-dispersive X-ray spectroscopy (EDS) line scanning for element concentration distribution at the cathode/electrolyte interface, as well as electrochemical test for a related BaCoFeO-type cathode with Zr replaced by Ce.

To study the chemical compatibility and mutual diffusion between the three BZCYYb-type electrolytes and the BCFZY cathode, the powders of each electrolyte and the BCFZY cathode were mixed at the weight ratio of 1:1 and went through the same firing process as z E-mail: zhcheng@fiu.eduECS Advances, 2023 2 044502 symmetrical cell cathode fabrication, which was 4 h at 1100 °C.The as-synthesized powders and the mixed powders after firing were analyzed by X-ray diffraction (XRD, Rigaku BD700535-01, Cu-K α radiation, λ = 1.541874Å) for phase composition.
Thermogravimetric analysis (TGA).-TGA was used to measure the weight change of each electrolyte powder in dry air.To obtain more accurate data, a two-cycle TGA was performed: in the first cycle, the sample was heated up to 800 °C at 5 °C min −1 and held for 1 h to fully remove the adsorbed water before cooling it down to room temperature.In the second cycle, the same sample from the first one was again heated to 800 °C at 5 °C min −1 and cooled down to room temperature.The final weight at the end of the first cycle would be the starting weight for the second cycle.The change of oxygen contents for the BZCYYb electrolytes with temperature can be obtained based on TGA from the second cycle, 13 by assuming the room temperature oxygen contents of the three electrolytes (Ce-rich, ZrCe-bal, Zr-rich) as 2.9.Such an assumption was supported by XRD Rietveld refinement and was also consistent with iodometric titration measurement in the literature. 20mmetrical cells fabrication and microstructure examination.-Forpellet fabrication, all electrolytes were sintered following the same process.Using BaZr 0.1 Ce 0.7 Y 0.1 Yb 0.1 O 3−δ (Ce-rich) electrolyte as an example, the Ce-rich powder was mixed with 0.5 wt% NiO as the sintering aid by hand grounding.0.2 g mixed powder was dry pressed under ∼300 MPa into a 10 mm pellet pressing die.The prepared green pellets were sintered at 1450 °C for 10 h in air in a muffle furnace to form dense pellets.In order to reduce Ba loss 21 and unexpected reaction with Al crucible, the pellets were vertically stacked, separated and surrounded by the corresponding electrolyte powder.The electrode slurries were made by mixing the cathode powder (e.g., either BCFZY or BCFCY that has all Zr replaced by Ce) and 4 wt% ethyl cellulose (EC) dissolved in alpha-terpineol with the weight ratio of 3:7 (e.g., 0.6 g of BCFZY powder and 1.4 g of 4 wt% EC in alpha-terpineol).Then the slurry was brush painted on both sides of the electrolyte pellet with an active area ∼0.16 cm 2 and calcinated at 1100 °C for 4 h in air in a muffle furnace at a lower ramp rate of 2.5 °C min −1 for both the heating and cooling process to obtain the symmetrical cell.The microstructure of the surface of three electrolytes and the cross section of the symmetrical cell were analyzed by a scanning electron microscope (SEM JEOL JSM-F100) fitted with energy dispersive X-ray spectroscopy (EDS).
Electrochemical measurements.-Electrochemicalimpedance spectroscopy (EIS) measurements were performed on symmetrical cells under open circuit conditions using a potentiostat (Gamry Interface 1000) in the frequency range 10 6 to 0.01 Hz with an AC amplitude of 1 mA.Fritless silver paste was brushed onto the electrodes of the symmetrical cell to collect current.The cell was placed in a sealed tube furnace (OTF-1200), heated to 750 °C and stabilized over 12 h before EIS testing.Each symmetrical cell was first tested in dry simulated air (Ultra Zero Grade, Airgas, with <5 ppm H 2 O and CO 2 by volume).The EIS data was collected every 50 °C from 650 °C to 450 °C with an air flow of 40 cc min −1 .To test the effect of H 2 O, the cell was heated up again to 650 °C and 3% H 2 O was introduced (by passing the simulated air through a water bubbler).After holding for over 24 h, the cell hydration effect would saturate and the cell was then tested from 650 °C to 450 °C by EIS to compare the electrochemical behaviors in moist air vs simulated (dry) air.

Results and Discussion
Phase identification for as-synthesized electrolytes and cathodes.-Figure1a    TGA of BZCYYb electrolytes.-FigureS1 shows the TGA data for the three electrolytes.As explained before, the first heating cycle was for removal of absorbed water (or dehydration) while in the second heating cycle, all weight loss of the electrolyte powder could be attributed to oxygen loss.Figure 3 shows how oxygen content, calculated from the TGA data, changes with temperature.As expected, all three electrolytes have the same tendency to lose oxygen when heated up; the higher the Ce concentration, the more oxygen loss especially at high temperatures of ∼800 °C, and the data is close to literature. 20crostructure of symmetrical cells.-Themicrostructure of the three proton conducting electrolytes (Ce-rich, ZrCe-bal and Zr-rich) after sintering is shown in Figs.4a-4c.From the top view, the surface appears dense for all three electrolytes, but there are some differences: The higher the Ce concentration, the larger the grain size of the BZCYYb electrolyte.Since the electrolyte powder synthesis and pellet sintering processes were the same, the difference in morphology is attributed to the Ce/Zr ratio: it is known that the BaZrO 3 has a higher melting temperatures and is more stable than BaCeO 3 . 22igures 4d-4f shows the cross-sections of the three symmetrical cells based on the same BCFZY cathode over those three electrolytes.All three electrolytes are near fully dense, while the cathodes are porous with grain size <1 μm and thickness of ∼15-20 μm.
Electrochemical behaviors of symmetrical cells.-Figures5a-5c shows the impedance spectra for a BCFZY/Ce-rich/BCFZY symmetrical  ECS Advances, 2023 2 044502 cell in dry simulated air (<5 ppm H 2 O & CO 2 per Airgas, actual concentration might be higher) as well as moist air (3% H 2 O) at 650, 550, and 450 °C.Both ohmic resistance R Ω (from the high frequency intercept in the EIS spectra) and electrode apparent interfacial resistance R ai (from the difference between the high frequency and low frequency intercepts in the EIS spectra) decreased when introducing H 2 O. Figures 6a-6c shows the results for the BCFZY/ZrCe-bal/BCFZY symmetrical cell under the same conditions.For this system, R Ω decreased when introducing H 2 O, while R ai stayed almost the same at 550 °C and below (it increased slightly at 650 °C).Figures 7a-7c is for the BCFZY/Zr-rich/BCFZY symmetrical cell.R Ω follows the same trend as the other two, but introducing H 2 O caused R ai to increase.In Figs.7b  -7c, the dash line represents the fitting data of the EIS curve using Zview.In the fitting, the high frequency part was ignored since it presents the grain boundary effect instead of the cell resistance.
• Changes in ohmic resistance R Ω Figure 8a summarizes ohmic resistance R Ω vs inverse temperature for all three types of symmetrical cells in both simulated air and 3% moist air.Note that, first, as Ce concentration in the BZCYYb electrolyte increases, R Ω decreases, meaning increase in bulk ionic conductivity.In fact, from the plot, 2 4 Ce rich ZrCe bal Zr rich in the temperature range of ∼450-650 °C.Such an observation is consistent with literature 8 showing Ce-rich electrolytes for Ba-based electrolytes tend to give higher ionic conductivity than Zr-rich electrolytes.It is also consistent with the oxygen contents determined from TGA (see Fig. 3), showing the Ce-rich electrolyte has lower oxygen content, i.e., higher oxygen vacancy concentration, than the ZrCe-bal and Zr-rich electrolytes at elevated temperature (e.g., 550 °C).
A second observation is that, upon introducing moisture, R Ω dropped for all three electrolytes.The drop in R Ω is attributed to the hydration reaction of water combining with oxygen vacancy (from Y and Yb acceptor doping) to form protons (OH)o • : leading to higher bulk conductivity of the electrolyte, and it is well known for proton conducting oxide electrolytes.
• Changes in apparent interfacial resistance R ai Figures 8b-8d summarizes cathode apparent interfacial resistance R ai vs inverse temperature for each of the three types of symmetrical cells.It should be noted that for proton conducting electrolytes such as BZCYYb that display mixed ionic and electronic conduction in air, R ai -the difference between the high frequency and low frequency intercepts in an EIS spectrum is only an approximation for the actual cathode interfacial resistance R i .This is due to electronic leakage through the proton conducting electrolyte, especially in dry air, that tends to make the measured apparent interfacial resistance appear smaller than the actual value, i.e., R ai < R i .Researchers have developed equations that help calculate the actual interfacial resistance R i . 23,24However, these calculations are all based on assumptions about certain parameters such as electronic and ionic transport numbers (t e and t i ) for the electrolyte in the interested temperature range.Without reliable experimental or theoretical t e or t i values for current materials, any estimation or correction based on those equations would be highly arbitrary and does not improve the accuracy of the analysis.As a result, in this  study, R ai was used directly as first approximation of the electrode activity.
Nevertheless, in dry simulated air, the BCFZY cathode shows comparable R ai among all three types of symmetrical cells, despite the difference in BZCYYb electrolytes with varying Ce/Zr ratios.This is understandable due to the same cathode material and processing condition.This also indicates that the oxygen reduction reaction (ORR) is unlikely to be limited by the bulk ionic conductivity of the electrolytes.
However, upon H 2 O introduction, as mentioned before, the cathode symmetrical cells behave very differently: apparent interfacial resistance R ai increased over the Zr-rich electrolyte, but decreased over the Ce-rich electrolyte, while staying the same over the ZrCe-bal electrolyte.
As to the behavior of increasing cathode R ai upon moisture introduction for the BCFZY/Zr-rich/BCFZY symmetrical cell (Figs.7 and 8d), similar observation has been made by other researcher 1 as well as for other cathodes such as LSCF and BSCF, 15 though most are not triple conducting cathodes and the electrolyte used were mostly Ce-rich BZCYYb electrolyte. 19,25One earlier explanation is that moisture addition to air might lead to strong adsorption of H 2 O on the cathode (and electrolyte) surface, which blocks ORR reaction site.However, the triple conducting nature of the BCFZY cathode means the entire cathode surface would be active for ORR and a low percentage of moisture is not expected to impact much.
Therefore, a more likely explanation is related to the depletion of electron hole by the defect reaction in the electrolyte between electron holes, water vapor, and lattice oxygen (O O ×) to form protons (or protonated oxygen) while releasing oxygen gas, as below: Introducing moisture would shift the above reaction to the right side, leading to a drop in hole concentration. 26The result will be higher electronic resistance through the electrolyte and less electronic leakage.This would increase the ionic transport number t i of the electrolyte toward unity, which, in turn, would lead to higher electrode apparent interfacial resistance R ai , as observed in Figs.7  and 8d.On the other hand, the behavior of R ai with moisture introduction for the BCFZY/Ce-rich/BCFZY symmetrical cell is also interesting.There have been similar observations on alternative cathodes such as Ba 0.95 La 0.02 Fe 0.8 Zn 0.2 O 3−δ 27 and Pr 2 NiO 4+δ • 28 The exact reason for reducing R ai upon moisture introduction in these symmetrical cells is not clear at this moment.As mentioned before, the hydration of the proton conducting electrolyte in air would suppress electronic leakage, which should lead to increase in R ai .
One possible explanation for the observed reduction in R ai is about the greater availability of H 2 O molecules for the ORR over proton conducting electrolytes: especially for the reverse (half-cell) reaction from water and lattice oxygen O O × to protons.Since EIS for a cathode symmetrical cell is obtained under zero bias condition, when there are very few H 2 O molecules, as in dry simulated air, the reverse of reaction 3 from H 2 O and lattice oxygen O O × to protons, despite the low AC bias, might be limited due to the near absence of H 2 O. Introducing moisture would naturally alleviate the problem of "starvation" or inadequacy of H 2 O as a reactant, accelerate the reverse of reaction 3, and, possibly, decrease R ai in EIS measurements.
An alternative explanation is that ORR in dry simulated air over the Ce-rich electrolyte (or any proton conducting electrolytes) would follow a different route involving only oxygen vacancy and no proton, as below: After introducing moisture, ORR might change, maybe partially, to the route involving protons (i.e., reaction 3).The kinetics for ORR by the proton route (reaction 3) might be faster than the oxygen vacancy route (reaction 4), leading to a large drop in R ai in EIS measurements.
Either way, for Ce-rich electrolyte, the effect of moisture on electrode kinetics is hypothesized to be large enough to overcome the effect from reduced electronic leakage, and overall R ai drop upon introducing moisture.
Interactions between BCFZY cathode and BZCYYb electrolytes.-However,despite the above hypotheses, there are questions still unanswered, especially about why the same BCFZY cathode behaved differently (i.e., R ai to moisture) over different BZCYYb electrolytes, all of which supposed to be proton conducting.To understand this, as stated, the interactions between the BCFZY cathode and the three electrolytes had been studied, and the results are presented below.
• XRD for cathode/electrolyte mixed powders Figures 9a-9c show XRD Rietveld refinement of three mixtures, each contained BCFZY cathode and one of the three BZCYYb electrolytes at a weight ratio of 1:1 and had gone through heat treatment at 1100 °C for 4 h.The unit cell volume of the BCFZY cathode after the heat treatment was 70.70, 70.21, 69.90 Å 3 when mixed with Ce-rich, ZrCe-bal, Zr-rich electrolyte, respectively.As shown in Fig. 9d, the unit cell volume of BCFZY cathode in all three mixtures were all larger than the as-synthesized sample (69.57Å 3 ).This is because under the cathode firing process at 1100 °C, there would be mutual diffusion between the cathode and the electrolyte.The larger high-concentration B-site elements such as Ce and Zr in the electrolyte would diffuse into the cathode and expand its lattice.In addition, it is observed that the higher the electrolyte Ce concentration, the greater the cathode volume expansion after firing (see Fig. 9d).Such a phenomenon can be understood in two aspects.Firstly, Ce ion is larger than Zr ion.Secondly, Ce should have greater tendency to diffuse than Zr given BaCeO 3 is known to have Figures 10a-10c shows EDS line scans along the cross-sections for three representative symmetrical cells: BCFZY/Ce-rich/BCFZY, BCFZY/ZrCe-bal/BCFZY, and BCFZY/Zr-rich/BCFZY, respectively.The origin or "0" in the horizontal "distance" axis corresponds to the interface between the BCFZY cathode and BZCYYb electrolyte.(Only data from the dense electrolyte side are presented.This is due to the large difference in porosity between the porous cathode and the dense electrolyte leading to a large gap of signal intensity at the interface).Figure 10a is for the cell with Ce-rich electrolyte.Within ∼1 μm from the interface into the dense electrolyte, the intensity of Ce gradually increases, while the intensity of Co and Fe decreases.It means that Ce in the bulk of electrolyte was lost due to diffusion into the cathode, while Co and Fe in the BCFZY cathode diffused into the electrolyte.By contrast, in Figs.10b-10c, the intensity change for either Zr or Co/Fe around the cathode/electrolyte interface is much less obvious, illustrating the ZrCe-bal and Zr-rich electrolytes with a cubic structure are more stable than the Ce-rich electrolyte with an orthorhombic structure and there was much less interdiffusion involved.It should be mentioned that the trends in EDS line scan about Ce/Co/Fe intensity were repeatable over multiple regions across multiple samples, indicating the validity of the observations.

• Test of Ce-containing BCFCY cathode over Zr-rich electrolyte
To explain why introducing moisture caused the same BCFZY cathode to display decreasing apparent interfacial resistance R ai over the Zr-rich electrolyte but increasing R ai over the Ce-rich electrolyte, we further hypothesize that Ce doping into the Ze-containing BCFZY cathode would enhance proton conduction in the cathode, while neat BCFZY, though active in dry air, has limited proton conduction capability.0][31] This means the Zr-rich BZCYYb electrolyte would have greater electronic leakage, while the Ce-rich electrolyte would be less influenced.
Therefore, on the Zr-rich electrolyte with slightly lower t i , the limited proton conduction in the BCFZY cathode means ORR would largely be limited to the triple phase boundary region.Despite the greater availability of H 2 O molecules and potentially faster ORR for the proton route (reaction 3), the effect of introducing moisture might be dominated by (i) water suppressing electronic leakage, and (ii) limited ORR proceeding via the proton route.These two effects lead to the observations of higher R ai with introducing moisture, as in Figs.7 and 8d.
In comparison, on the Ce-rich electrolyte, it is less impacted by electronic leakage due to higher t i .Meanwhile, due to extensive Ce diffusion into the BCFZY cathode, when moisture is introduced, the cathode, at least near the interface, conducts proton better and the sites for faster ORR via the proton route might greatly expand.As a result, R ai decreases upon moisture introduction.
For the ZrCe-bal electrolyte, the two effects (water suppressing electronic leakage causing R ai to increase vs promoting ORR via the proton route causing R ai to drop) might roughly balance each other.Therefore, no obvious change in R ai was observed.
To test the hypothesis that Ce doping into BCFZY cathode might enhance proton conduction and cathode activity, especially in humidified air, a related cathode BaCo 0.4 Fe 0.4 Ce 0.1 Y 0.1 O 3−δ (BCFCY) is synthesized by replacing all Zr (in BCFZY) with Ce. Figure 11 shows the impedance spectra for such a BCFCY/Zr-rich/ BCFCY symmetrical cell in both simulated air and moist air (3% H 2 O) at 650 °C.Ohmic resistance R Ω of the Ce-containing BCFCY symmetrical cell in both atmospheres were close to the Zr-containing symmetrical cell, which is expected from the same Zr-rich electrolyte composition and geometry.After introducing 3% H 2 O, the new Ce-containing BCFCY cathode does show reduction in R ai at 650 °C, as opposed to increasing R ai for the Zr-containing BCFZY cathode (see Figs. 7a and 8d).This seems to confirm that Ce doping to this BaCoFeO-type cathode might promote proton conduction in the cathode and improve the activity upon introducing moisture, even over the Zr-rich electrolyte.
Finally, it should be mentioned that all the above analyses only constitute simplified frameworks that serve as a starting point for further understanding.There are many questions that remain unanswered for such BaCoFe-type cathodes over BZCYYb proton conducting electrolytes and there might be alternative explanations.Future studies including detailed materials characterizations, ab initial calculation (e.g., by the density functional theory or DFT method), as well as electrochemical modeling are needed to fully understand conductivity of BCFZY/BCFCY cathodes in different atmospheres, their water permeation rates, ionic transport numbers for different BZCYYb electrolytes, as well as many of the observed complex electrochemical behaviors.

Conclusions
This study compares the behaviors of BaCo 0.4 Fe 0.4 Zr 0.1 Y 0.1 O 3−δ •(BCFZY) cathode over BaZr 0.8−x Ce x Y 0.1 Yb 0.1 O 3−δ (BZCYYb) electrolytes with different Ce/Zr ratios (i.e., x = 0.1, 0.4, and 0.7).Consistent with expectation, for the BZCYYb electrolytes, the higher the Ce concentration, the bigger the grain size in sintered pellet and more oxygen vacancy, leading to higher conductivity especially at high temperature (e.g., 550 °C).Furthermore, the comparable BCFZY cathode apparent interfacial resistance R ai over the three different electrolytes in dry simulated air suggests that the cathode oxygen reduction reaction is limited by oxygen adsorption/dissociation process instead of the ionic conductivity of the electrolyte.After introducing ∼3% moisture to air, cathode R ai drops for the symmetrical cell with Ce-rich electrolyte, while it increases for the Zr-rich electrolyte.Analyses such as EDS line scanning at the cathode/electrolyte interfaces and XRD Rietveld refinement comparing the cathode's unit cell volume change when fired together with different BZCYYb electrolytes suggest that Ce has a stronger tendency than Zr to diffuse from the electrolyte to the cathode.And the diffused Ce likely increases the BCFZY cathode activity.Such a hypothesis is supported by the observation that replacing all Zr by Ce in the BaCoFeO-type cathode over the Zr-rich electrolyte also increases the cathode activity upon moisture introduction at 650 °C.Additional experiments and theoretical calculations will be carried out in the future to fully understand such phenomenon.
shows the XRD patterns of the three electrolyte powders BaZr 0.1 Ce 0.7 Y 0.1 Yb 0.1 O 3−δ (Ce-rich), BaZr 0.4 Ce 0.4 Y 0.1 Yb 0.1 O 3−δ (ZrCe-bal), and BaZr 0.7 Ce 0.1 Y 0.1 Yb 0.1 O 3−δ (Zr-rich) after heat treatment at 1300 °C for 5 h.The result indicates that all electrolytes are pure phase perovskites and the peaks shift to the right or higher 2θ angle with higher Zr concentrations.This is because the Zr ion is smaller than the Ce ion.Figure1b shows the XRD Rietveld refinement of the Ce-rich electrolyte, indicating that it has an orthorhombic Pnma space group with lattice parameters of a = 6.195Å, b = 8.695 Å, c = 6.201Å (4 formula per unit cell?Equivalent pseudo-cubic lattice parameter = 4.3709 Å).With lower Ce concentration, the crystal structure of the electrolyte becomes cubic: Fig.1cshows the refinement for the ZrCe-bal electrolyte, illustrating that it has a cubic Pm3m ̅ space group with a lattice parameter of a = 4.311 Å.The refinement of the Zr-rich electrolyte is shown in Fig.1d.It also has the cubic Pm3m ̅ space group with an even smaller lattice parameter of 4.216 Å.The refinement R wp values of these electrolytes are 9.43%, 8.80% and 9.68%, respectively, indicating the calculated data match well with the XRD patterns.The trend in lattice volume (or (equivalent) lattice parameters) also matches with the shift in obtained XRD patterns.

Figure 3 .
Figure 3. Oxygen contents of the three BZCYYb electrolytes.As stated, all data were from the second heating/cooling cycle, while the adsorbed water was removed in the first heating/cooling cycle.

Figure 5 .
Figure 5. Impedance spectra for a BCFZY/Ce-rich/BCFZY symmetrical cell in dry simulated air (<5 ppm H 2 O and CO 2 by volume) vs moist air (3% H 2 O by volume) at (a) 650 °C, (b) 550 °C, (c) 450 °C.Note the apparent interfacial resistance R ai decreases upon H 2 O introduction, especially at lower temperature.

Figure 6 .
Figure 6.Impedance spectra for a BCFZY/ZrCe-bal/BCFZY symmetrical cell in dry simulated air (<5 ppm H 2 O and CO 2 by volume) vs moist air (3% H 2 O by volume) at (a) 650 °C, (b) 550 °C, (c) 450 °C.Note the apparent interfacial resistance R ai stays roughly the same upon H 2 O introduction.

Figure 7 .
Figure 7. Impedance spectra for a BCFZY/Zr-rich/BCFZY symmetrical cell in dry simulated air (<5 ppm H 2 O and CO 2 by volume) vs moist air (3% H 2 O by volume) at (a) 650 °C, (b) 550 °C, (c) 450 °C.Note the apparent interfacial resistance R ai increases upon H 2 O introduction.