LaxPr4−xNi3O10−δ: Mixed A-Site Cation Higher-Order Ruddlesden-Popper Phase Materials as Intermediate-Temperature Solid Oxide Fuel Cell Cathodes

Systematic studies of the air electrode and full solid oxide fuel cell performance of La3PrNi3O9.76, and La2Pr2Ni3O9.65 n = 3 Ruddlesden–Popper phases are reported. These phases were found to adopt orthorhombic symmetry with a decrease in lattice parameters on increasing Pr content, consistent with the solid solution series end members. From electrochemical impedance spectroscopy measurements of symmetrical cells, the electrodes were found to possess area specific resistances of 0.07 Ω cm2 for the La2Pr2Ni3O9.65 cathode and 0.10 Ω cm2 for the La3PrNi3O9.76 cathode at 750 ◦C, representing a significant improvement on previously reported compositions. This significant improvement in performance is attributed to the optimisation of the electrode microstructure, introduction of an electrolyte interlayer and the resulting improved adhesion of the electrode layer. Following this development, the new electrode materials were tested for their single-cell performance, with the maximum power densities obtained for La2Pr2Ni3O9.65 and La3PrNi3O9.76 being 390 mW cm−2 and 400 mW cm−2 at 800 ◦C, respectively. As these single-cell measurements were based on thick electrolytes, there is considerable scope to enhance over cell performance in future developments.

99.0%), Pr(NO 3 ) 3 ·6H 2 O (Sigma Aldrich, 99.99%) and Ni(NO 3 ) 2 ·6H 2 O (Sigma Aldrich, 99.0%) were dissolved in an aqueous solution of 10% (by weight) of citric acid (Sigma Aldrich, 99.99%). The solution was heated at 250 • C, under constant stirring for three hours, until a gel was obtained. The gels, after being decomposed in air for 12 h at 600 • C, were ground in a mortar and pestle, and the resultant powder was further annealed for 24 h in air at 1000 • C prior to X-ray diffraction (XRD) analysis. The XRD patterns of the materials were collected using a PANalytical X'Pert Pro MPD (Cu K α source, λ = 1.5406 Å) and Rietveld refinement of the XRD data, performed with the GSAS/EXPGUI software package [31,32], which was used to confirm the phase identification and extract unit cell parameters. Post-test microstructural investigations were achieved using a LEO Gemini 1525 FEG scanning electron microscope with an accelerating voltage of 5.0 kV.

Cell Preparation
A commercially available La 0.8 Sr 0.2 Ga 0.8 Mg 0.2 O 3−δ (LSGM) electrolyte powder, purchased from Praxair Surface Technologies (Lot # 03-P6948DM), was pressed (0.04 MPa) into 13 mm diameter disks using a uniaxial press, followed by the cold isostatic press (300 MPa), and then sintered at 1450 • C for 8 h. The thickness of the LSGM pellets after sintering and polishing with silicon carbide paper (1200 grit) was in the 400-500 µm range. L3P1N3 and L2P2N3 electrode inks were prepared by mixing the as-prepared powders with an organic ink vehicle supplied by fuel cell materials (Lot # R1835, Item # 311006, USA) in a 2:1 weight ratio, and were used to prepare symmetrical cells using the screen-printing technique. After ball milling the electrode powders for 24 h, the electrode slurries were prepared by mixing the powders with the ink vehicle and triple roll milling the prepared ink to produce a slurry with a uniform particle size (~3 µm). A further ink consisting of only the electrolyte LSGM powder was prepared in a similar way to be used as an interlayer between the electrode and the LSGM pellet.
Symmetrical cells of the configuration L2P2N3|LSGMInk|LSGM|LSGMInk|L2P2N3 and L3P1N3|LSGMInk|LSGM|LSGMInk|L3P1N3 were prepared. Firstly, an LSGM interlayer ink was screen-printed on one face of the LSGM pellet, which was then dried for 10 min at 200 • C. An identical procedure was repeated for screen-printing the other face of the pellet, and this was later followed by annealing the cells for 1 h at 1000 • C. This was followed by screen-printing of the appropriate electrode layer ink on both sides of the samples, again drying the inks for 10 min at 200 • C at each step. All cells prepared had an active area of 0.5 cm 2 . Two cells of each composition were prepared and sintered at two different temperatures; one sample of each composition, L2P2N3 (sample 1) and L3P1N3 (sample 2), were sintered at 1100 • C for 1 h. Since both the L2P2N3 and L3P1N3 compositions decompose at temperatures above 1050 • C, slow cooling (0.5 • C min −1 ) to 950 • C from 1100 • C under an oxygen flow was required, which leads to the re-formation of the L2P2N3 and L3P1N3 higher-order phases. To make sure that the L2P2N3 and L3P1N3 compositions were fully recovered, the cells were maintained at 950 • C for 6 h. Another set of samples, L2P2N3 (sample 3) and L3P1N3 (sample 4), were sintered for 2.5 h at 1050 • C under an ambient atmosphere. The sample details are summarised in Table 1.
Pt mesh was used as a current collector for the electrochemical measurements. Impedance measurements of the symmetrical cells were performed using a Solartron 1260 impedance gain/phase analyser in combination with a Solartron 1287 electrochemical interface in the 0.1 to 10 6 Hz frequency range with a perturbation amplitude of 10 mV. The cell configurations used for the single-cell performance measurements were NiO-GDC|LDC|LSGMPellet|LSGM Ink|Cathode Ink (where the cathode was L2P2N3 or L3P1N3). A LDC (La 0.4 Ce 0.6 O 1.8 ) buffer layer was used for interfacial stability between the anode and LSGM electrolyte [33]. The LSGM ink and LDC buffer layer were screen-printed onto the electrolyte pellets, and the cells were sintered for 1 h at 1000 • C. This was followed by screen-printing the cathode and anode and then sintering the complete cell at 1050 • C for 2.5 h. The current-power-voltage (IPV) curves were recorded in the range of 600-800 • C using a Solartron 1287 electrochemical interface controlled by CorrWare software, where the cell voltage was varied from OCV to 0.3 V. Humidified H 2 (~3% H 2 O) was fed as fuel to the anode with a flow rate of 40 mL min −1 , with pure oxygen used as the oxidant at the cathode with a flow rate of 100 mL min −1 .

X-ray Diffraction
Powder X-ray diffraction was used to initially characterise the synthesised cathode materials. It is well known that the Ruddlesden-Popper phases can adopt a variety of space group symmetries depending on the stoichiometry of the material. To fully evaluate the crystal chemistry of these new materials, three space groups, orthorhombic Bmab, orthorhombic Fmmm and monoclinic P2 1 /a, as previously proposed, were evaluated [27,34,35] for both of the prepared compositions. The XRD characterisation, followed by Rietveld refinement of L3P1N3, concluded that this material crystallised with an orthorhombic structure adopting the Fmmm space group (Figure 1). It should be noted that the oxygen content was fixed to full stoichiometry for the refinement, i.e., all oxygen positions had an occupancy of 1 corresponding to the La 3 PrNi 3 O 10 composition. This was necessary as refining oxygen content from XRD data is inherently unreliable. Similarly, XRD data collected previously for the L2P2N3 and L1P3N3 compositions [29,36] were refined in the orthorhombic Fmmm and monoclinic P2 1 /a space groups, respectively.
The lattice parameters for L3P1N3 were determined to be a = 5.406(1) Å, b = 5.466(1) Å and c = 27.873(9) Å, and are in close agreement with the previously reported results for the end member materials [27,37]. The lattice parameters of the L2P2N3 and L1P3N3 compositions, obtained through the application of Rietveld refinement previously reported by the authors, are shown in Table 2. The lattice parameters, when compared to the two end members, La 4 Ni 3 O 9.78 and Pr 4 Ni 3 O 10.10 , show a smooth decrease with increasing Pr content (Table 2), which is consistent with the variation of the ionic radius [38], leading to an overall decrease in the cell volume. This is not surprising and is attributed to the ionic radius of the Pr 3+ ion (1.179 Å) being smaller than La 3+ (1.216Å) in 9-fold coordination [38]. Table 2. Unit-cell parameter comparison of the prepared compositions of L2P2N3 and L3P1N3. The two end members, L4N3 and P4N3, are also included to show the systematic decrease of lattice parameters with increasing Pr content, which is consistent with the variation of the ionic radii, leading to an overall decrease in the cell volume with increasing Pr content. Note the data for L2P2N3 are from reference 35.

Symmetrical Cell Performance Testing
LSGM was selected as the electrolyte for cell testing because of its good low-temperature electrolyte properties [40] and its reported superior performance and chemical compatibility with Ruddlesden-Popper phases [14,41,42]. The cathode performance of the L2P2N3 (samples 1 and 3) and L3P1N3 (samples 2 and 4) compositions was initially investigated using impedance spectroscopy to determine the area specific resistance (ASR) of the materials. Typical Nyquist plots and fitting results with equivalent circuit LR(CPE1R1), where L is the inductance, R represents the ohmic resistance and (CPE1R1) represents the constant phase element and resistance of the electrode process, are shown in Figure 2.
In all samples, the total resistance decreased with increasing temperature, as expected, indicating the thermal behaviour of the electrode reaction processes. The ASR values obtained for the L2P2N3 electrode were 0.15 Ω cm 2 and 0.07 Ω cm 2 at 750 °C for samples 1 and 3, respectively. This contrasts dramatically with our earlier results, where a higher ASR-0.56 Ω cm 2 at 750 °C -was reported [36]. Our optimisation procedure, consisting of ball milling the initial electrode and using a LSGM interlayer, evidently improved the overall electrode performance. The corresponding ASR values for the L3P1N3 electrode were 0.17 Ω cm 2 and 0.10 Ω cm 2 at 750 °C for samples 2 and 4, respectively. One clear observation is that sintering of the electrode for a shortened period of time (2.5 h (samples 3, 4)) under an ambient atmosphere presents a lower ASR than the corresponding ASR obtained for the same electrode sintered under an oxygen atmosphere followed by slow cooling (samples 1, 2). One possible reason for this observation could be the elongated periods of time (11-14 h) required for the cell sintering procedure under an oxygen atmosphere because the overall process involves sintering at a lower temperature (950 °C for 6 h) to allow for the reformation of the

Symmetrical Cell Performance Testing
LSGM was selected as the electrolyte for cell testing because of its good low-temperature electrolyte properties [40] and its reported superior performance and chemical compatibility with Ruddlesden-Popper phases [14,41,42]. The cathode performance of the L2P2N3 (samples 1 and 3) and L3P1N3 (samples 2 and 4) compositions was initially investigated using impedance spectroscopy to determine the area specific resistance (ASR) of the materials. Typical Nyquist plots and fitting results with equivalent circuit LR(CPE 1 R 1 ), where L is the inductance, R represents the ohmic resistance and (CPE 1 R 1 ) represents the constant phase element and resistance of the electrode process, are shown in Figure 2.
In all samples, the total resistance decreased with increasing temperature, as expected, indicating the thermal behaviour of the electrode reaction processes. The ASR values obtained for the L2P2N3 electrode were 0.15 Ω cm 2 and 0.07 Ω cm 2 at 750 • C for samples 1 and 3, respectively. This contrasts dramatically with our earlier results, where a higher ASR-0.56 Ω cm 2 at 750 • C-was reported [36]. Our optimisation procedure, consisting of ball milling the initial electrode and using a LSGM interlayer, evidently improved the overall electrode performance. The corresponding ASR values for the L3P1N3 electrode were 0.17 Ω cm 2 and 0.10 Ω cm 2 at 750 • C for samples 2 and 4, respectively. One clear observation is that sintering of the electrode for a shortened period of time (2.5 h (samples 3, 4)) under an ambient atmosphere presents a lower ASR than the corresponding ASR obtained for the same electrode sintered under an oxygen atmosphere followed by slow cooling (samples 1, 2). One possible reason for this observation could be the elongated periods of time (11-14 h) required for the cell sintering procedure under an oxygen atmosphere because the overall process involves sintering at a lower temperature (950 • C for 6 h) to allow for the reformation of the n = 3 Ruddlesden-Popper phase. This was not the case with ambient atmosphere sintering of the electrode, used for preparing samples 2 and 4. To put the results obtained in perspective, Table 3 shows the resistances obtained at different temperatures for the L3P1N3 electrode (samples 2 and 4) and L2P2N3 electrode (samples 1 and 3). It is evident from the data in Table 3, and from previous studies [29], that the calculated ASR values decrease with the increase of Pr content (25% in the L3P1N3 composition to 75% in the L1P3N3 composition [29]). A similar finding was reported for the n = 1 RP phase (La2-xPrxNiO4+ δ) by Vibhu et al. [25], wherein they reported that electrode performance increased with an increase in the Pr content of a composition. This was ascribed to improved catalytic properties of Pr as compared to La.  To put the results obtained in perspective, Table 3 shows the resistances obtained at different temperatures for the L3P1N3 electrode (samples 2 and 4) and L2P2N3 electrode (samples 1 and 3). It is evident from the data in Table 3, and from previous studies [29], that the calculated ASR values decrease with the increase of Pr content (25% in the L3P1N3 composition to 75% in the L1P3N3 composition [29]). A similar finding was reported for the n = 1 RP phase (La 2-x Pr x NiO 4+δ) by Vibhu et al. [25], wherein they reported that electrode performance increased with an increase in the Pr content of a composition. This was ascribed to improved catalytic properties of Pr as compared to La.

Single Button Cell Testing
Single-cells of each composition were tested under open circuit conditions (OCV) with a measured OCV of 1.10 V, which agrees well with the theoretical value calculated from the Nernst equation, thereby indicating minimum gas leakage. The maximum power densities (MPD) obtained over a range of temperatures for both compositions are outlined in Figure 3 and summarised below in Table 4. The highest MPD was obtained for the L3P1N3 composition, 400 mW cm −2 at 800 • C. It should be noted that in each case, the cells were based on relatively thick (~400 µm) electrolytes which could be thinned to minimise the Ohmic losses and thus achieve higher power densities. We also noted that there was a small difference between the minimum ASRs obtained and the maximum power densities. This is ascribed to small differences in the electrolyte and electrode layer thicknesses, as shown in the subsequent section. It was noted that the post-test XRD (not shown) found all of the material compositions retained their structural integrity and that there was no discernible phase evolution of the electrode.

Single Button Cell Testing
Single-cells of each composition were tested under open circuit conditions (OCV) with a measured OCV of 1.10 V, which agrees well with the theoretical value calculated from the Nernst equation, thereby indicating minimum gas leakage. The maximum power densities (MPD) obtained over a range of temperatures for both compositions are outlined in Figure 3 and summarised below in Table 4. The highest MPD was obtained for the L3P1N3 composition, 400 mW cm −2 at 800 °C. It should be noted that in each case, the cells were based on relatively thick (~400 μm) electrolytes which could be thinned to minimise the Ohmic losses and thus achieve higher power densities. We also noted that there was a small difference between the minimum ASRs obtained and the maximum power densities. This is ascribed to small differences in the electrolyte and electrode layer thicknesses, as shown in the subsequent section. It was noted that the post-test XRD (not shown) found all of the material compositions retained their structural integrity and that there was no discernible phase evolution of the electrode.

Symmetrical Cells
The secondary electron SEM images-in both the top view and cross-section view of the electrode layer and cell, respectively-obtained for the L2P2N3 electrode (samples 1 and 3) and

Symmetrical Cells
The secondary electron SEM images-in both the top view and cross-section view of the electrode layer and cell, respectively-obtained for the L2P2N3 electrode (samples 1 and 3) and L3P1N3 electrode (samples 2 and 4) after symmetrical-cell testing are presented in Figure 4; Figure 5. Figure 4 illustrates the post-test microstructural SEM images of samples 1 and 2, i.e., those sintered at 1150 • C with subsequent annealing at 950 • C for 6 h in oxygen. It is clear from these images that the electrode/interlayer and interlayer/electrolyte interfaces are well adhered, in contrast to the data previously reported for the L2P2N3 composition [36]. This highlights the significant improvement in the cell preparation methods developed. In addition, these microstructural data correlate well with the cathode performance determined by impedance spectroscopy measurements. As mentioned previously, the ASR of the L2P2N3 electrodes decreased from 0.56 Ω cm 2 (reported previously [36]) to 0.15 Ω cm 2 at a measurement temperature of 750 • C, which is ascribed to the ball milling of the electrode material and introduction of an LSGM buffer layer. A further reduction in the ASR from 0.15 Ω cm 2 to 0.07 Ω cm 2 at 750 • C is attributed to the reduction of the effective cell sintering time, of which the direct impact is negligible particle coarsening ( Figure 5). Similar behaviour is observed in the L3P1N3 composition; sample 2 in Figure 4 shows mild particle coarsening and sintering (~5 µm) compared to sample 4, which shows almost no particle coarsening ( Figure 5). This is also attributed to the reduction of the electrode sintering time, and this is also reflected in the half-cell performance of L3P1N3, with samples 2 and 4 presenting ASR values of 0.17 Ω cm 2 and 0.10 Ω cm 2 at 750 • C, respectively. correlate well with the cathode performance determined by impedance spectroscopy measurements. As mentioned previously, the ASR of the L2P2N3 electrodes decreased from 0.56 Ω cm 2 (reported previously [36]) to 0.15 Ω cm 2 at a measurement temperature of 750 °C , which is ascribed to the ball milling of the electrode material and introduction of an LSGM buffer layer. A further reduction in the ASR from 0.15 Ω cm 2 to 0.07 Ω cm 2 at 750 °C is attributed to the reduction of the effective cell sintering time, of which the direct impact is negligible particle coarsening ( Figure 5). Similar behaviour is observed in the L3P1N3 composition; sample 2 in Figure 4 shows mild particle coarsening and sintering (~5 μm) compared to sample 4, which shows almost no particle coarsening ( Figure 5). This is also attributed to the reduction of the electrode sintering time, and this is also reflected in the half-cell performance of L3P1N3, with samples 2 and 4 presenting ASR values of 0.17 Ω cm 2 and 0.10 Ω cm 2 at 750 °C, respectively. The introduction of the buffer layer shows a clear impact by increasing the number of contact points at the electrode/electrolyte interface, which ultimately decreases the interface polarisation resistance contribution. Furthermore, as illustrated in Figure 5, SEM images of samples 3 and 4 taken at high magnification show excellent particle interconnectivity, and the cathodes present porous The introduction of the buffer layer shows a clear impact by increasing the number of contact points at the electrode/electrolyte interface, which ultimately decreases the interface polarisation resistance contribution. Furthermore, as illustrated in Figure 5, SEM images of samples 3 and 4 taken at high magnification show excellent particle interconnectivity, and the cathodes present porous structures. This interconnectivity of the electrode particles and the porous structure of the cathode are expected to facilitate oxygen transport by providing an extended network of active pathways for oxygen diffusion and charge transfer.

Single Fuel Cell Microstructure Analysis
Post-test single-cell microstructural studies showed similar results, as discussed in the previous section, for the symmetrical half-cells. The SEM images illustrated in Figure 6 highlight the microstructure and interface adherence in post-test single-cells based on L2P2N3 and L3P1N3 cathodes. The images demonstrate that a porous cathode structure with considerable particle interconnectivity is maintained through the cell test, although minor particle coarsening/sintering was observed, in contrast to the half-cells measured at OCV. The electrode screen-printed over the electrolyte interface layer, as observed in the previous section, shows strong adhesion.

Single Fuel Cell Microstructure Analysis
Post-test single-cell microstructural studies showed similar results, as discussed in the previous section, for the symmetrical half-cells. The SEM images illustrated in Figure 6 highlight the microstructure and interface adherence in post-test single-cells based on L2P2N3 and L3P1N3 cathodes. The images demonstrate that a porous cathode structure with considerable particle interconnectivity is maintained through the cell test, although minor particle coarsening/sintering was observed, in contrast to the half-cells measured at OCV. The electrode screen-printed over the electrolyte interface layer, as observed in the previous section, shows strong adhesion.
It is pertinent here to point out that there are some discernible differences between the post-test microstructure obtained after half-and single-cell testing that could be the reason behind the anomalous half-and single-cell behaviour described in the previous section. The interlayer appears denser in the prepared single-cells and the electrode is significantly more porous when compared to the half-cells. The thickness of the interlayer varies slightly, with L2P2N3 and L3P1N3 single-cells showing a thickness of~8 µm, while all the half-cells are slightly thicker at around 10 µm. Irrespective of these slight variations, it is clear that the introduction of an electrolyte interlayer has a significant positive effect on the adhesion, and thus performance of La 4−x Pr x Ni 3 O 10+/−δ electrodes for solid oxide fuel cells.
Crystals 2020, 10, x 10 of 13 anomalous half-and single-cell behaviour described in the previous section. The interlayer appears denser in the prepared single-cells and the electrode is significantly more porous when compared to the half-cells. The thickness of the interlayer varies slightly, with L2P2N3 and L3P1N3 single-cells showing a thickness of ~8 μm, while all the half-cells are slightly thicker at around 10 μm. Irrespective of these slight variations, it is clear that the introduction of an electrolyte interlayer has a significant positive effect on the adhesion, and thus performance of La4−xPrxNi3O10+/−δ electrodes for solid oxide fuel cells.

Conclusions
Two higher-order Ruddlesden-Popper phase materials were investigated for IT-SOFC cathode applications. The materials, L2P2N3 and L3P1N3, were synthesised by a sol-gel method and were later characterised by X-ray powder diffraction. Lattice parameters showed a systematic decrease with increasing Pr content, which is in agreement with the concept of effective ionic radii being the main driver of this observation; the ionic radius of Pr is lower than La [27,38] in equivalent coordination environments.
The best half-cell performance of 0.07 Ω cm 2 and 0.10 Ω cm 2 , respectively, for a L2P2N3 and L3P1N3 electrode operating at 750 °C was obtained, which fits in well with our previous report of 0.05 Ω cm 2 at 750 °C for a L1P3N3 electrode. This is the first report showing such low ASR values for n = 3 Ruddlesden-Popper compositions with LSGM electrolytes, and offers opportunities to develop a new series of high-performance devices. All of these electrodes in half-cell configurations exceeded the requirement of <0.15 Ω cm 2 at 700 °C, and with further optimisation, will have applications in the low operating temperature regime (<600 °C).

Conclusions
Two higher-order Ruddlesden-Popper phase materials were investigated for IT-SOFC cathode applications. The materials, L2P2N3 and L3P1N3, were synthesised by a sol-gel method and were later characterised by X-ray powder diffraction. Lattice parameters showed a systematic decrease with increasing Pr content, which is in agreement with the concept of effective ionic radii being the main driver of this observation; the ionic radius of Pr is lower than La [27,38] in equivalent coordination environments.
The best half-cell performance of 0.07 Ω cm 2 and 0.10 Ω cm 2 , respectively, for a L2P2N3 and L3P1N3 electrode operating at 750 • C was obtained, which fits in well with our previous report of 0.05 Ω cm 2 at 750 • C for a L1P3N3 electrode. This is the first report showing such low ASR values for n = 3 Ruddlesden-Popper compositions with LSGM electrolytes, and offers opportunities to develop a new series of high-performance devices. All of these electrodes in half-cell configurations exceeded the requirement of <0.15 Ω cm 2 at 700 • C, and with further optimisation, will have applications in the low operating temperature regime (<600 • C).
Testing of single-cells revealed maximum power densities for the L2P2N3 and L3P1N3 compositions of 390 mW cm −2 and 400 mW cm −2 at 800 • C, respectively, highlighting a marginally higher performance at a lower Pr content. Post-test microstructural analysis after both half-and single-cell testing revealed that the cathode microstructure was porous with considerable particle interconnectivity. It was observed that the use of a buffer layer of LSGM between the electrode and LSGM electrolyte significantly improved the contact and is suggested to be the driver for the performance improvement in these materials.