Introduction

Solid oxide fuel cell technology has witnessed remarkable development and scale-up in the past few decades to cope with the gradual shift from fossil fuels to renewable energy. A lot of research and development is still being conducted on different aspects of this technology, and one such area is improving intermediate temperature (500–700 °C) cell kinetics by modifying the electrodes or electrode/electrolyte interfaces. Lanthanum Strontium Manganite—LaxSr1-xMnO3-δ (LSM) is a benchmark cathode material in solid oxide fuel cells (SOFC) for operation at temperatures above 800 °C due to its excellent chemical and thermal stability with Yttria stabilized Zirconia-YSZ (YxZr1-xO3-δ) electrolyte [1]. However, polarization resistance associated with the LSM cathode becomes high at intermediate temperatures due to poor catalytic activity and electronic/ionic conductivity values (electronic conductivity (σe) of 100 S cm−1 and ionic conductivity (σi) of approx. 10−6 S cm−1 at 800 °C) [2]. Therefore, significant efforts are being made to overcome such reduced efficiency of SOFC at a lower temperature range (500–700 °C) due to the sluggish kinetics of the ORR at cathode [3]. In addition to a high electrocatalytic activity for the ORR, a cathode is also required to have thermal and chemical compatibility with other cell components, especially the electrolyte within the operating temperature range. Other factors that influence the cathode performance are the cathode and the cathode/electrolyte interfacial microstructure.

The Strontium (Sr) and Iron (Fe) doped LaCoO3-δ (La1-xSrxCo1-yFeyO3-δ—LSCF) cathode is a promising mixed ionic and electronic conductor (MIEC) with electronic conductivity (σe) of 120 S cm−1 at 500 °C [Electrochemical analysis

As-prepared button cells were placed in a tubular furnace at 500 °C with a highly accurate temperature control system in air (ambient testing environment). V-I characteristics and chronoamperometry were performed at a fixed potential of 1 V to study the effect of different physical parameters (temperature and time) and cell operating conditions (OCV and under cathodic polarization). The same set of experiments was repeated at 600 °C and 700 °C for both the LSCF-B and LSCF-M. The electrochemical data were recorded using a Princeton Applied Research VersaSTAT in the frequency range of 1 MHz to 0.1 Hz at 10 mV s−1 amplitude.

Results and discussion

Electrochemical impedance (EIS) analysis

Cell performance under cathodic polarization

The electrochemical performance of LSCF-B and LSCF-M for the ORR were systematically evaluated at three different temperatures for 10 min and 24 h. Figure 1 shows the typical electrochemical impedance spectra (EIS) of LSCF-M and LSCF-B at 500 °C, 600 °C, and 700 °C under OCV. The ohmic resistance (Rohmic) was determined from the first intercept of the high-frequency arc on the x-axis, while the polarization resistance (Rpol) of the electrodes was determined from the difference between the total resistance of the cell (intercept at low frequency) and the ohmic resistance (intercept at high frequency on the real axis). The obtained value was divided by 2 to obtain the cathode polarization since the cells were symmetrical and both anode and cathode were exposed to the same environment [21]. As expected, both the cells showed a decrease in Rpol and Rohmic as an effect of increasing temperature. The ionic conductivity of the electrolyte increases with temperature resulting in a reduced Rohmic [10]. Similarly, an increase in temperature activates the electrodes and reduces resistive losses related to charge-transfer as well as mass-transfer processes of gas adsorption–desorption and diffusion [22], thus resulting in a reduced Rpol as observed here.

Fig. 1
figure 1

Electrochemical impedance spectra recorded at OCV for a LSCF-M and b LSCF-B at 500, 600, and 700 °C

Full size image

As seen from Fig. 1, at all test temperatures, Rohmic is lower for LSCF-B (without GDC interlayer) compared to LSCF-M (with GDC interlayer). It is to be noted here that the 3–5 µm thick GDC layer on either side of the electrolyte would impart an additional Rohmic of 0.11 Ωcm2 at 700 °C [23]. Similar observations have been reported by Hsieh et al. [24] who introduced an electron-blocking layer of ScSZ between a LSCF cathode on GDC electrolyte in a tubular cell while studying the effect of the interface on the performance of microtubular electrolyte-supported SOFC between 650 and 800 °C. At 500 °C and 600 °C, Rpol at OCV increased for LSCF-M. However, at 700 °C, this trend reversed and Rpol dropped from 0.42 Ωcm2 for LSCF-B to 0.18 Ωcm2 for LSCF-M. This could possibly be due to changes in interfacial cation chemistry at the interface at higher temperature. Overall, the resistance decreased with an increase in temperature due to thermal activation, improving the materials’ electrochemical properties such as electronic structure and charge transfer processes. Furthermore, at high-temperature, GDC generates more oxygen vacancies in situ that aids the “oxygen hop**” process across the GDC lattice and at the interface [25]. The OCV testing of both cells after polarization treatment of 10 min and 24 h at 1 V showed improved Rohmic and Rpol when compared to initial OCV values, which indicates that the polarization of the cells activated the cathode and the interfacial layer.

Figure 2 shows the effect of cathodic polarization at 1 V on LSCF-B and LSCF-M at the three testing temperatures (500 °C, 600 °C, and 700 °C). Figure 2a shows the EIS recorded after 10 min of cathodic polarization for LSCF-B and LSCF-M, while 2b shows the EIS recorded after 24 h of chronoamperometry at a constant potential of 1 V for LSCF-B and LSCF-M, respectively. Here we observe an improved performance for LSCF-M compared to LSCF-B at 600 °C and 700 °C under polarization (CV — constant voltage) conditions. For both the samples, EIS at 500 °C did not intercept the x-axis, possibly due to the upper-frequency limits (100 kHz) of the VersaStat and possibly due to enlarged grain boundary/higher resistance of ScSZ at low temperatures of 500 °C. To rule out the possibility of this behavior being related to our samples, we repeated the testing of LSCF-B starting from higher (700 °C) to lower temperature (500 °C) at the same operating conditions [10 min and 24 h under polarization (CV) treatment]. This has been referred to here as testing of LSCF-B in a reverse trend, and the nature of the EIS at 500 °C remained unchanged. Interestingly, we noticed a decrease in Rpol at all temperatures in a reverse trend of testing temperature (700 °C to 500 °C) as compared to the normal trend (500 °C to 700 °C) of LSCF-B (Table 1). Cathode activation at high operating temperatures has already been reported [26].

Fig. 2
figure 2

Electrochemical impedance spectra (EIS) recorded under polarization condition at 1 V for a 10 min, b 24 h with LSCF-B (hollow symbols) and LSCF-M (solid symbols) at 500 °C (red), 600 °C (green), and 700 °C (blue). EIS at 500 °C shows the effect of instrument limitation at a lower testing temperature in both samples under cathodic polarization conditions

Full size image
Table 1 Comparison of Rpol between normal and reverse trend of testing temperature for LSCF-B
Full size table

Operating the cell at a high temperature (700 °C) may have activated the cathode thus improving cell performance as discussed by Alder et al. [26]. The reason can be alteration in cation compositional changes in the LSCF bulk or at the interface between the LSCF and electrolyte triggered by high operating temperature (700 °C) under polarization.

Next, we studied the effect of cell cathodic polarization for LSCF-B and LSCF-M at 600 °C and 700 ℃. An increased cathode performance has been seen in the past by using GDC as an interlayer between different cathodes and electrolytes [27]. In this study, with the introduction of the GDC interlayer between the LSCF and ScSZ electrolyte, the polarization resistance of the cathode (Rpol) drastically dropped from 3.04 Ωcm−2 to 0.48 Ωcm−2 at 600 °C and from 2.11 to 0.17 Ωcm−2 at 700 °C following 24 h chronoamperometry at 1 V (Table 2).

Table 2 Comparison of Rpol and Rohmic under different testing conditions in LSCF-B and LSCF-M at 600 °C and 700 °C
Full size table

The enhanced performance can be ascribed to the good adhesion between the layers leading to a better contact, and hence enlarged active area at the cathode/interlayer and interlayer/electrolyte interfaces, as discussed later in the “Scanning electron microscopy and energy dispersive spectroscopy” section [SEM images of LSCF-M in Fig. 9a, b and LSCF – B in Fig. 9c, d]. Wang et al. [28] also reported the structural damage of the interface between LSM cathode and YSZ electrolyte after polarization of the cell. Another explanation for the improved performance of LSCF-M can be the increased ion conduction in the cathode bulk due to the stress developed by the higher oxide ion conductive GDC interfacial layer. GDC electrocatalytic activity is also well known in solid oxide cell operation. Addition of a mixed ionic and electronic conductor promotes the electrochemical activity as has been reported frequently [29]. Coddet et al. [30] discussed the positive effect of barrier layers and stated that the GDC barrier layer can enhance oxygen transport at the cathode side. The stress developed by the underlaying GDC barrier layer is due to structural defects/voids of the GDC and can be beneficial to the transport of oxygen ions compared to the bare cathode.

In the AC Impedance spectroscopy studies, the EIS arcs of the LSCF-B and LSCF-M clearly showed a distinct high frequency (HF) arc and a low frequency (LF) arc (Fig. 2). In LSCF based cathodes, the HF processes are charge transfer processes [31], while the LF processes are mass transfer processes like adsorption/desorption and bulk/surface diffusion. It can be seen in Fig. 2 that there is a significant change in both arcs by introducing the GDC interfacial layer in LSCF-B.

All the processes contribute to the total polarization of the cathode. For LSCF-B, at 600 °C, the polarization resistance increased after 24 h of cell operation. It is evident from Fig. 2 that both the HF and LF arc increased significantly, possibly due to microstructural instability, which increased the charge transfer resistance and associated processes. A similar trend has also been observed for LSCF under polarization for 48 h at 750 °C [27].

Contrarily, in LSCF-M at 600 °C, the contribution of the LF processes to the total Rpol of the cathode is low under both testing conditions of 10 min and 24 h. In this case, as evidenced by HF frequency arc, charge transfer processes are dominating the total polarization resistance of the cathode within 10 min and 24 h as compared to LSCF-B. The performance of the LSCF-M is stable for 24 h under polarization. No significant increase in charge transfer resistance was observed from the HF arc. The reduction in the resistance offered by HF processes in 24 h represents improved charge transfer and related processes and is consistent with increasing stability in current density vs time analysis as shown in Fig. 3.

Fig. 3
figure 3

Current density as a function of time for a LSCF-M and b LSCF-B, at a cathodic polarization of 1 V for 24 h

Full size image

Rohmic depicted opposite trends for LSCF-B and LSCF-M. For LSCF-B, Rohmic at 700 °C increased post 24 h of operation at constant potential of 1 V, possibly due to Zr interdiffusion and formation of trace amounts of non-conducting phases at the electrode-electrolyte interface as seen from SEM-EDS maps discussed in the “Scanning electron microscopy and energy dispersive spectroscopy” section. In contrast, for LSCF-M, Rohmic at both 600 and 700 °C dropped by more than 10% post 24 h of operation at constant potential of 1 V. The reasons are (1) activation of GDC interlayer resulting in better electronic conductivity, (2) improved thermal stability at the electrode-electrolyte interface resulting in better adhesion and reduced contact losses, and (3) mitigation of any Zr interdiffusion that can otherwise form non-conducting phases at the electrode-electrolyte interface as seen from SEM-EDS maps discussed in the “Scanning electron microscopy and energy dispersive spectroscopy” section.

In conclusion, with the introduction of GDC between the LSCF cathode and the ScSZ electrolyte in LSCF-M, the reaction contributing to the polarization of the cathode is changed from LF processes (mass transfer) to HF (charge transfer) processes, when compared to LSCF-B at both the testing temperatures, i.e., 600 °C and 700 °C.

Short-term stability testing

There is a clear decrease in current density with time for LSCF-B at all testing temperatures (Fig. 3a), especially at high temperatures (700 °C) where it started initially at 256 mA cm−2 and decreased to 219 mA cm−2 in 24 h. The same behavior is observed for 600 °C and 500 °C; however, the decline is not as drastic. At high temperatures, the decrease in current density might be due to a collapse in porosity from catalyst agglomeration and nonconductive phases formation which are common at high temperatures. Wilde et al. [32] documented the non-conductive phase formation at 1100 °C as the sintering temperature, the samples in this study are sintered at 1050 °C with adequate densification [33] which is close to 1100 °C, and the possibility of non-conductive phases cannot be ruled out and has been confirmed from our SEM images as well (discussed in the “Scanning electron microscopy and energy dispersive spectroscopy” section). Another possibility is the voltage-assisted nonconductive phase formation as we have sintered our samples at 1050 °C and no secondary phase formation was detected in XRD. At a lower testing temperature (500 °C), this effect (decrease in stability over 24 h of testing under polarization) is not very prominent, showing the stability of the cathode microstructure under polarization. The reduction in current density at high temperatures could be due to the onset of phase formation of impurities at the interface between the LSCF cathode and the ScSZ electrolyte since high temperature triggers voltage-assisted cross-reactions at interfaces. On the other hand, LSCF-M showed a constant increase in current density with time rather than a decline over 24 h at all testing temperatures (Fig. 3b). This demonstrates the improvement at the interface between the LSCF cathode and ScSZ electrolyte in terms of adhesion, with increased active area or/and enhanced oxygen ion conductivity. A similar effect is shown by Kulkarni et al. [19] who reported the increased active area and better adhesion of the cathode to the electrolyte by introducing the GDC ultra-thin layer at the interface that reduced the Rpol of the cathode. It is obvious from Fig. 3 that the difference between the current density of LSCF-B and LSCF-M is even more significant at higher temperatures (700 °C). This shows the beneficial effect of the GDC interfacial layer on cathode performance at high temperature. The current-voltage characteristics of LSCF-M (Fig. 4a) and LSCF-B (Fig. 4b) at all temperatures also show a clear difference in the behavior of the two electrodes, especially at 600 °C and 700 °C after 24 h chronoamperometry at a constant voltage of 1 V. The LSCF-B shows a mixed reaction mechanism indicated by the two slopes in the V-I curves at the current density of 30 mA cm−2 and 60 mA cm−2 (at 600 °C), and 100 mA cm−2 and 160 mA cm−2 (at 700 °C). This mixed reaction mechanism indicates the contribution of both ohmic and polarization resistances to the performance of the LSCF-B. The behavior (contribution of both ohmic and polarization resistances) becomes dominating at a high testing temperature which is consistent with EIS analysis and stability study. The V-I curve for the LSCF- M shows that the introduction of GDC has reduced the polarization resistance contribution, as indicated by the straight lines at all temperatures, particularly at higher temperatures. Figure 5 shows a comparison of the current-voltage characteristics of LSCF-M and LSCF-B at 600 °C.

Fig. 4
figure 4

Voltage-current curves at 500 °C, 600 °C, and 700 °C of LSCF-M (a) and LSCF-B (b) after 24 h of polarization at 1 V

Full size image
Fig. 5
figure 5

Voltage-current curves of LSCF-B and LSCF-M at 600 °C

Full size image

Physical characterization

X-rays diffraction

X-ray diffractograms were obtained for the powder precursors of LSCF, ScSz, and GDC mixed in 50–50 wt% (two precursors at a time) and sintered at 800 °C at different time intervals to investigate the formation of secondary phases. Figure 6 shows the XRD spectra of these precursors sintered at 800 °C for 100 h.

Fig. 6
figure 6

X-ray diffractograms obtained from powder precursors sintered at 800 °C for 100 h a LaSrCoFeO3-δ (LSCF) mixed with Scandium stabilized zirconia (ScSZ), b LaSrCoFeO3-δ (LSCF) mixed with Gd-doped Ceria (GDC), and c Scandium stabilized zirconia (ScSZ) mixed with Gd-doped Ceria (GDC). The inset in the figure shows the spectrum of the secondary phase formed in a

Full size image

No secondary phase formation was noticed in the powder precursors annealed at 800 °C for 8 h in the air. It can be clearly seen that SrZrO3 formation was observed along with some traces of cobalt iron oxide sintering under similar conditions in 100 h (Fig. 6). Sr segregation at higher firing temperature (~ 1200 °C) of the cell is possible and is reported frequently [34]. We report the formation of SrZrO3 even at low firing temperature of powders at 800 °C, over a period of 100 h as compared to 1200 °C. This suggests that non-conducting phase formation can take place at 800 °C when heated for a longer time. Railsback et al. [34] also reported Co interdiffusion from LSCF in GDC layers suggesting the tendency of Co to leach-out from LSCF into other layers. We detected Co oxide in LSCF powder mixed ScSZ powder confirming pass out of Cobalt from LSCF at 800 °C in 100 h (Fig. 6a). We also obtained the XRD diffractograms of the as-prepared and samples tested for 10 min and 24 h. First, we obtained the Bragg‐Brentano diffractograms from the surfaces of the samples, and afterwards, we obtained the Grazing Incident diffractograms after scra** the top layer of the LSCF-B and LSCF-M. The only phases reflected by all diffractograms are pure LSCF, Ag, GDC, and multiple zirconia‐based phases both in as prepared and tested samples. The zirconia-based phases were all detected in the off-the-shelf disks purchased from fuel cell materials while the Ag peak comes from the silver current collector layer on the top. The absence of additional phases in the Grazing incidence obtained from the scraped samples, as well as the Bragg-Brentano diffractograms obtained from top layers of as-prepared and tested cells, confirmed that no inter-phases had formed because of an unwanted reaction between LSCF and ScSZ or LSCF and GDC, suggesting chemical compatibility at cell operating temperatures. The diffractograms collected from chosen regions of as-prepared and tested LSCF-B and LSCF-M samples are given in Fig. 7. The best match phases from the ICDD‐JCPDS powder diffraction database found in these samples are listed in Table 3.

Fig. 7
figure 7

The X-ray diffractograms a and b are from as-prepared and tested samples of LSCF-B, respectively, while c and d are from as-prepared and tested samples of LSCF-M, respectively

Full size image
Table 3 X-ray powder diffraction patterns of the annealed powder precursors at 800 °C for 8 h and 100 h of both cells
Full size table

Scanning electron microscopy and energy dispersive spectroscopy

Figure 8 shows the scanning electron microscopy images of the LSCF-M while Fig. 9 shows the electrode–electrolyte and electrode-interlayer-electrolyte interfaces of the LSCF-B and LSCF-M, respectively. These images show cells consisting of a LSCF cathode, ScSZ electrolyte, and the GDC interfacial layer in LSCF-M. It is clear from the SEM images that the LSCF-B has poor contact at the interface between the cathode and electrolyte. In this study, the cathode was sintered on electrolyte at 775 °C to control the high temperature-triggered secondary phase formation. In Fig. 9c and d, the separation of layers/delamination at the interface between LSCF and ScSZ electrolyte can be easily noticed in LSCF-B. This might be due to the lack of enough densification, because of the low firing temperature required for better contact between the layers. This poor contact is causing the loss of active area and structural damage after testing. In contrast, at the same firing temperature (775 °C) of LSCF with GDC layer, the LSCF-M has better contact between the GDC interlayer and LSCF cathode. After testing, there is not much transformation at the cathode/GDC interfacial layer/electrolyte interfaces demonstrating the stability of the interfaces. It is well-known that GDC and ScSZ have similar thermal expansion coefficients (TEC). The TEC values of LSCF, GDC, and ScSZ are 14.6 × 10−6 K−1, 12 × 10−6 K−1 at 700 °C [35] 10.7 × 10−6 K−1 [36], respectively). Since GDC has a TEC value intermediate between LSCF and ScSZ, a GDC interfacial layer improves adhesion between LSCF cathode and ScSZ electrolyte. Such improved adhesion enlarges the active area along the extended triple phase boundary (TPB) which helps to improve the overall performance of the cell. Similar results were reported by Chen et al. [37] and Kulkarni et al. [19]. On the other hand, a poor contact between the LSCF cathode and SSZ electrolyte in tested LSCF-B can be attributed to dissimilar thermal expansion coefficients of LSCF and ScSZ, resulting in non-uniform expansion and delamination at higher operating temperatures and hence poor performance.

Fig. 8
figure 8

SEM images of tested LSCF-M showing good contact to LSCF cathode and ScSZ electrolyte

Full size image
Fig. 9
figure 9

SEM images of tested LSCF-M at a 5KX and b 20 KX and tested LSCF-B at c 5KX, d 20 KX, respectively

Full size image

The formation of CeZrO2 layer at the GDC/ScSZ interface was investigated through EDS by measuring the percentage quantities of Ce and Zr at various positions at the interface of GDC interfacial layer and ScSZ electrolyte. The EDS results of LSCF-M are depicted in Fig. 10a, b for tested samples and c, d for fresh samples while Fig. 11a, b shows the EDS spectrum of LSCF-B, respectively. Ce/Sr + Zr atomic ratio across the different layers of the samples including LSCF cathode, GDC/ScSZ interface, and ScSZ electrolyte as a function of position was studied. Jung et al. [18] studied the optimization of microstructure of GDC interfacial layer as a function of sintering temperature for the elimination of secondary phase of CeZrO2. However, here we report no CeZrO2 layer formation for GDC/ScSZ interface in both fresh and tested samples of the LSCF-M (Fig. 10a, b and c, d) sintered at 775 ℃ in air. Similarly, no cross-diffusion of any cation from GDC interfacial layer and ScSZ electrolyte in LSCF cathode was detected. A small amount of Sr from LSCF cathode is seen in GDC interfacial layer in LSCF-M in spectrum 2 (Fig. 10a, b). However, in Fig. 11a, b, the EDX analysis of the LSCF-B tested samples shows that diffusion of Zr has taken place in to the LSCF cathode layer [refer to EDS spectrum 2 in Fig. 11a, b). But it can be said that this cross-diffusion of Zr is very small as it was not detected by the detailed analysis through XRD both at the interface and at the surface of the fresh or tested samples. It is proposed that the actual quantity of the Zr diffusion will be much lesser than the width of the cross-diffusion zone in cathode of LSCF-B cell as it was not detected at position 1 and 5 shown in Fig. 11a, b in spectrum 1 and 5, respectively. Clearly, the GDC interlayer is effective in mitigating any cross-diffusion of Zr even in minute quantity as has been envisaged for LSCF-B.

Fig. 10
figure 10

EDS analysis of LSCF-M tested (a, b) and of fresh (c, d) samples at various positions across LSCF cathode, GDC interfacial layer, and ScSZ electrolyte

Full size image
Fig. 11
figure 11

EDS analysis for tested LSCF-B (a, b) at various positions across LSCF cathode, GDC interfacial layer, and ScSZ electrolyte

Full size image

Conclusion

The Electrochemical Impendence Spectroscopic analysis of symmetrical button cell with cell configuration of LSCF | GDC | ScSZ | GDC | LSCF reveals a significant improvement in polarization resistance for the cell with an interlayer between the electrolyte and cathode at all temperatures under cathodic polarization conditions when compared to LSCF | ScSZ | LSCF cells with no interfacial layer modification. Polarization resistance was measured to be 0.48 Ωcm2 at 600 °C and 0.17 Ωcm2 at 700 °C after 24 h of polarization condition at 1 V. The current density improvement at higher temperatures (600 °C and 700 °C) shows the positive impact of the gadolinium-doped ceria as an interfacial layer on the overall performance of cell, possibly due to the improved adherence of cathode to the electrolyte, and possible retardation of cross-reaction between cathode and electrolyte, resulting in an improvement in the oxygen reduction reaction process. The improved performance at 600 °C and 700 °C suggests an induced increase in ionic conductivity of the Lanthanum Strontium Cobalt Ferrite cathode. The potentiodynamic study also revealed a possible change in the oxygen reduction reaction mechanism due to interfacial layer deposition as compared to the bare sample. This suggests that the deposition of the gadolinium-doped ceria interface layer between the Lanthanum Strontium Cobalt Ferrite cathode and Scandium stabilized zirconia electrolyte is not only beneficial for interfacial modification but has an impact on the oxygen reduction reaction mechanism as well.