Electrochemical Performances of a Solid Oxide Electrolysis Short Stack Under Multiple Steady-State and Cycling Operating Conditions

Abstract

Solid oxide electrolysis cells (SOECs) are increasingly utilized in hydrogen production from renewable energy sources, yet high degradation rates and unclear degradation mechanisms remain significant barriers to their large-scale application. Consequently, endurance testing of stacks under various operating conditions and studying the degradation mechanisms associated with these conditions is imperative. However, due to the generally poor performance consistency among stacks, multi-condition data from numerous stacks lack reliability. In this experimental study, having established a specific SOEC stack’s performance and optimal conditions, durability tests under varied conditions, including various current densities, current operation modes (cyclic or constant current), fuel utilization rates, and temperature cycles were conducted. Electrochemical analysis tools like electrochemical impedance spectroscopy and distribution of relaxation time were employed to analyze the causes of voltage fluctuations under high current densities. The results confirmed that the SOEC stack could handle current cycling at low current densities and constant-current electrolysis at high current densities and withstand at least two temperature cycles.

1. Introduction

In recent years, the promotion of renewable energy sources such as solar and wind power has been vigorous. However, given the intermittent nature of electricity generation, there is an urgent need for energy balancing and storage technologies [1]. Electrolysis presents a promising solution for large-scale energy storage by converting electricity into hydrogen [2], and solid oxide electrolysis cell (SOEC) technology has garnered lots of attention due to its highest efficiency among alkaline water electrolysis (ALK) and proton exchange membrane electrolysis cell (PEMEC) technologies [3,4]. One of the primary challenges associated with the application of SOECs is attaining a durability of the stacks sufficient to achieve a lifespan of 5 to 10 years [5].

Currently, although extensive research focuses on the steady-state electrolysis durability of single SOECs, durability test results often vary significantly due to differences in the cell structure as well as operating conditions, and the main performance degradation mechanisms remain unclear. Schefold et al. [6] conducted a long-term steady-state steam electrolysis test on a 45 cm² electrolyte-supported SOEC single cell for 20,000 h. At the current density of 0.9 A/cm², the voltage degradation rate was only 0.57%/kh. In-situ impedance testing confirmed that ohmic degradation was the dominant factor. Different from electrolyte-supported single cells, cathode (H₂ electrode)-supported cells are more often susceptible to Ni-YSZ degradation, with the Ni depletion issue at high current densities being particularly prominent [7]. Yang et al. [8] confirmed that the operating temperature affects the degradation rate of single cells. By conducting 1000 h durability tests at the thermoneutral voltage of 1.29 V at 800, 750, and 700 °C, it was found that significant initial degradation occurred within the first 300 h (~1.8, 1.5, and 1.9%/kh, respectively). The EIS data confirmed that the main degradation at all three temperature conditions could be attributed to the increase in activation impedance of the Ni/YSZ electrode. Similarly, Hauch et al. [9] found that Ni migration was not significant for the fuel electrode overpotentials below 160 mV, but significant migration of Ni occurred at fuel electrode overpotentials between 160 and 300 mV. Under high overpotentials, the microstructure of the Ni/YSZ electrode underwent irreversible changes. Park et al. [10] also suggested that SOECs are prone to failure at higher current densities and lower operating temperatures due to electrode overpotentials exceeding 200 mV. However, operating overpotentials exceeding 200 mV are expected in practical applications due to the demand for a high flow rate of hydrogen production and efficiency at a high operating current density.

Apart from innovations in materials [11,12], there are currently no effective methods to mitigate the degradation of Ni-YSZ. Some studies have attempted to improve the durability of SOECs by optimizing operating conditions. Yang et al. [6] found that an operating temperature of 750 °C with 1.29 V resulted in the lowest initial fuel electrode overpotential. At 800 °C, severe microstructural deterioration was observed in Ni/YSZ electrodes, including Ni redistribution/agglomeration and disconnection between Ni and YSZ. Chen et al. [13] suggested that the increase in activation impedance of the hydrogen electrode could be due to the growth of CGO particles or the formation of CeAlO₃ from Al³⁺ poisoning. Therefore, an open-circuit voltage (OCV) treatment was attempted using a fuel mixture of H₂O/H₂ (90%/10%) in the cathode and air in the anode for 185 h at 706 °C, because under thermodynamic calculations CeAlO₃ should decompose. However, the OCV treatment did not cause significant changes in the cell degradation behavior. Performance under a high current density was even slightly worse. Nevertheless, it is widely accepted that optimization of operational parameters (such as working voltage, temperature, gas flow rate, etc.) is crucial to mitigate irreversible long-term degradation of a SOEC [8].

The impact of transient operating conditions such as cyclic current on/off on the durability of a SOEC single cell is more complex compared to the steady-state operating conditions [14]. Schefold et al. [15] carried out a one-year electrolysis durability test using an electrolyte-supported SOEC, which included 80,000 current density cycles from 0 to 0.7 A/cm² (with a cycle period of 2 min) as well as a sustained period at 0.7 A/cm² for 5800 h. The voltage degradation was consistently 0.4%/kh, demonstrating no significant influence induced by the current cycles. The dominant degradation mechanisms were ohmic degradation, minor gas conversion, and reaction losses. Petipas et al. [16] conducted a 600 h test on a cathode-supported SOEC single cell under both steady-state and transient conditions, which included steady testing at 20 A (corresponding to thermoneutral voltage), as well as 1800 current cycles ranging from 1–20 to 0–20 A. The measured voltage degradation rate was 5%/kh, confirming that the hydrogen electrode does not require a minimum current. EIS data indicate that the increase in ASR was mainly due to the rise in high-frequency ohmic resistance and electrode charge transfer impedance. Notably, the transient testing results for the durability of the same Ni-YSZ-supported SOEC single cell did not align with the analysis of the aforementioned steady-state durability experimental results.

It is well known that scaling up from a single cell to a stack can lead to a significantly increased degradation rate. Experimental results often show that the electrochemical performance and durability of a stack are inferior to those of single-cell tests due to issues such as improper flow field design, the high probability of seal failure, and the insufficient mechanical strength of the cells [17]. In a study of stack-level steady-state durability, Léon et al. [18] investigated the effects of scaling up from an electrolyte-supported single cell to a 30-cell stack and a 90-cell integrated module on degradation. Steady-state durability tests were conducted at 830 °C with a steam-to-hydrogen ratio of 90% H₂O/10%H₂ for 8756 h for the cell, 4224 h for the stack, and 2200 h for the module. The 30-cell electrolyte-supported stack achieved a low degradation rate of 0.8%/kh, approximately twice the degradation rate of the single cell. Yan et al. [19] built a short stack with four Ni-YSZ-cathode-supported cells, each with an active area of 80 cm² using the Julich F10 design [20]. Under a constant current density of 0.5 A/cm² at 800 °C, the stack was used for nearly 1000 h of steam electrolysis, with an average voltage degradation rate of 15.8%, attributed to the degradation of the air electrode. It is evident that due to manufacturing challenges, the uniformity of stacks is often more difficult to ensure compared to a single cell, and there can be significant differences even in steady-state degradation rates between different stacks. Therefore, for durability comparison under multiple operating conditions, a more reliable approach is to use single-stack testing. Although temperature fluctuations may occur during testing under multiple operating conditions, the interference with the test results is limited because SOECs have been partly proven to withstand temperature fluctuations and even cycles [21].

In transient durability studies of SOEC stacks, Léon et al. [18] conducted 16,000 cycle tests with a period of 2 min on a 30-cell electrolyte-supported stack. The degradation rate was close to that of the steady-state degradation rate they tested, without causing additional degradation. Rao et al. [22] conducted a 1000 h long-term electrolysis test on a short stack of six cathode-supported single cells (each with an active area of 80 cm²), achieving a degradation rate of less than 1%/kh. Impedance analysis confirmed a constant ohmic resistance, with only minor degradation related to processes in the fuel electrode. To induce and accelerate specific degradation phenomena, Königshofer et al. [23] used multiple variables such as high steam partial pressure, rapid load changes, high voltage, and high steam conversion rates for accelerated stress tests (ASTs). They carried out a 2000 h accelerated degradation test on a stack composed of 10 electrolyte-supported single cells, each with an active area of 125 cm². Compared to mild operating conditions, a significantly higher degradation rate was found, and the potential for degradation acceleration varied with different operating strategies.

In summary, it is clear that the degradation of stacks is significantly greater than that of a single cell. However, results in the literature suggest that both electrolyte-supported and Ni-YSZ-cathode-supported cells and stacks generally exhibit low degradation rates under steady-state and transient currents. Nevertheless, due to differences in operating conditions and stack structures, the reported degradation rates vary significantly, and the degradation mechanisms remain unclear. Therefore, the correlation between the steady-state and dynamic degradation mechanisms of SOECs and operating conditions needs further clarification. Due to the poor consistency among stacks, comparing degradation under multiple operating conditions and states (steady-state and dynamic) across multiple stacks lacks reliability. Moreover, there are few studies reporting simultaneous steady-state, dynamic, and thermal cycle tests on the same stack. Additionally, aside from improvements at the single-cell level, there are no effective solutions yet to address stack degradation.

This study firstly assesses the optimal performance of a specific stack under steady-state conditions, as well as the influence of hydrogen partial pressure, fuel utilization rate, and current step size on steady-state and dynamic performance. Then, leading factors in SOEC steady-state and dynamic degradation and their correlation with operating conditions are revealed by conducting steady-state and dynamic durability tests under multiple conditions. Finally, the performance variation of the stack at different temperatures is clarified and the lowest stable working temperature is determined by performing thermal cycle tests.

2. Results

2.1. Parametric Study of Operating Conditions on SOEC Stack Performances

It is crucial to understand the stable performances of the SOEC short stack with various operating parameters in order to determine reasonable operating conditions. This section investigates the influences of key operating parameters on the fuel side of the stack, including the fuel utilization (FU) factor, the hydrogen partial pressure in the fuel gas mixture, and the current step size in dynamic operations. A parametric study of these conditions is important for determining reasonable operating conditions in the subsequent long-term operation tests.

2.1.1. Parametric Study in the Steady-State Tests

To maintain the required reducing environment in the Ni-YSZ cathode, the hydrogen partial pressure should be kept above 10%. Meanwhile, to ensure a certain level of hydrogen production efficiency and prevent fuel depletion, the FU is typically controlled between 50 and 90%. As shown in Figure 1, a lower hydrogen partial pressure (higher steam partial pressure) can lower the OCV of the SOEC stack and reduce the overall energy consumption for hydrogen production. The performance differences at steady state between 50% and 75% FU are found to be negligible. Therefore, from the perspective of steady-state performance, hydrogen partial pressure has a more significant impact than FU and should be maintained at 10% in regard to energy consumption and efficiency.

The impedance test results in Figure 2 show that at any combinations of FU and hydrogen partial pressure, the total impedance is the highest at the current density of 0.1 A/cm², mainly due to the higher activation overpotential at low current densities. As the current density increases, the impedance decreases, transitioning from activation impedance domination to ohmic impedance domination. Since a SOEC typically operates in the ohmic polarization-dominated region, the working current density of 0.2–0.4 A/cm² is considered a reasonable operating range for the studied SOEC short stack.

2.1.2. Parametric Study in the Dynamic Operation

Figure 3 illustrates the effects of FU and hydrogen partial pressures on the dynamic performances of the SOEC short stack. The use of a dual piston pump in the experiments can lead to intermittent fuel shortages, resulting in regular voltage pulses in the voltage response. The magnitude of these pulses is closely related to the demand for the fuel gas. For instance, an increase in the current density leads to higher fuel demand, thereby intensifying the voltage pulse amplitude. A higher FU reduces the fuel feeding into the stack and results in similar increases in the amplitude of voltage pulses. Thus, the voltage pulses generated by the intermittent effect of the piston pump are more pronounced under conditions of high FU and high current density.

In the transient current stepwise increase processes, voltage overshoot exhibits patterns similar to the pulsing phenomenon, with noticeable concentration response hysteresis occurring under conditions of low steam supply and high current density. However, as shown in Figure 3b, when fuel supply is reduced at FU = 75%, increasing the steam partial pressure (reducing hydrogen partial pressure) to boost the overall steam supply can effectively mitigate the voltage pulses and overshoot. Nevertheless, compared to steam partial pressure, the FU has a more significant effect on the dynamic response characteristics of the SOEC stack.

The operating conditions of a SOEC often depend on the input from power supplies such as renewable energy sources, whose variability significantly increases the complexity of the SOEC operation. Consequently, the current step size is one essential parameter in transient processes, and its impact on the dynamic performances of the SOEC short stack is analyzed in Figure 4.

Figure 4a,b present the dynamic voltage responses and the corresponding V-I curve and EIS data measured at steady state in different current step sizes. It is worth mentioning that momentary voltage fluctuations caused by the pump have a small probability of creating a bad data point in the low-frequency region of the EIS. This is very easy to identify in the Nyquist plot and has been removed from the results. It is observed that larger current step size results in higher steady-state voltages for the same current density in the range of 0–0.2 A/cm², which is not dominated by concentration polarization. The impedance analysis in Figure 4c indicates that the ohmic resistance remains consistent across the three current step size tests, with the difference attributed to variations in activation overpotentials. Additionally, at high current densities, the hysteresis phenomenon caused by concentration polarization is severe, with larger current steps leading to greater voltage overshoot. Therefore, to maintain coherence in the I-V curves while the power input varies and to minimize the impact of voltage overshoot, a small current step size of 0.1 A/cm² is more representative of the real electrochemical performances in I-V characteristics.

2.2. Performance Under Multiple Operating Conditions

2.2.1. Test Conditions and the Overall Performance

After appropriate operating conditions for the SOEC stack were determined, a long-term SOEC test was conducted by maintaining a 10% hydrogen partial pressure in the fuel gas mixture. Based on the parameters with the greatest impact on dynamic performance, four sets of SOEC tests were carried out for a total operating time of 350 h.

Table 1. Steps and corresponding conditions for the 350 h long-term SOEC test.

Set	Segment	Current Density (A/cm2)	Operation	Fuel Utilization	Duration (h)
01	1–8	0–0.2	Cycling (a)	50%	40
	10–12	0.2	Galvanostatic	50%	20
02	15–17	0–0.4	Cycling (a)	50%	60
	18	0–0.5	Cycling (a)	50%	20
	22–23	0.4	Galvanostatic	50%	30
03	26–30	0–0.2	Cycling (a)	75%	40
	32–34	0.2	Galvanostatic	75%	25
04	37–42	0–0.4	Cycling (a)	75%	60
	45–47	0.4	Galvanostatic	75%	50

^(a) Each on/off cycle is two minutes with a 50% duty cycle.

lists details of the test procedures for the four sets. Each test set was divided into several segments as shown in the upper bar of Figure 5a for the convenience of results discussion.

The overall test results are presented in Figure 5a, with the enlarged view of the dynamic current cycling test results labeled as b shown in Figure 5b, and the enlarged view of the steady-state test results labeled as c and d displayed in Figure 5c,d. The spikes in current appear at the time points to measure the V-I curves and EIS. The green line in Figure 5a connects the voltage values V[i] at the last second of each on/off cycle during the cycling test, which reflects the quasi-steady-state performance changes of the stack during the dynamic processes.

V[i] = V|_T=120×i

(1)

where i is the number of on/off cycles.

The purple line represents the averaged voltage values during constant-current electrolysis, after filtering out voltage spikes caused by intermittent operation of pumps. The voltage data were filtered to include only values within ±0.1 V of the average voltage, and then a rolling average of every 30 data points was calculated to resample effective voltage data. This method allows for much better visualization of the changes in electrochemical performance during the electrolysis process than simply showing the raw experimental data.

Figure 5a allows for a visual comparison of the overall performance across different test sets. First, it is observed that the voltages are relatively stable at the current density of 0.2 A/cm² in both the steady-state and on/off cycling tests. In contrast, significant voltage fluctuation occurs at a higher current density of 0.4 A/cm² during the cycling tests. This becomes even more severe when the current density increases to 0.5 A/cm² in on/off cycling. However, the voltage can remain stable at 0.4 A/cm² in steady-state electrolysis. Second, the voltages are almost the same at 0.2 A/cm² in steady-state and cycling operations. However, it has higher values in cycling than the steady-state tests at 0.4 A/cm². Lastly, analyzing from the perspective of FU, it is evident that increasing fuel (lower FU) has a limited effect at the current density of 0.2 A/cm². The effect of fuel increase is more pronounced at the current density of 0.4 A/cm². Clearly, the stack cannot operate stably under on/off cycling at 0.4 A/cm², but no significant performance degradation is observed.

2.2.2. Performance Analysis in Each Test Set

In test set 01, a current density of 0.2 A/cm² and FU of 25%, which represent the mildest operating conditions in this study, are used. The I-V-t curves in Figure 6a and the electrochemical impedance spectroscopy (EIS) results indicate that the performances are relatively stable during both the current cycling (Figure 7a,b) and the constant-current electrolysis processes (Figure 7c,d).

In test set 02, more stringent operating conditions are implemented by setting a current density of 0.4 A/cm² and 0.5 A/cm² in on/off cycling tests while maintaining a FU of 50%. Then, a galvanostatic test was conducted at 0.4 A/cm². Figure 6b illustrates that cycling between 0 and 0.4 A/cm² causes periodic performance fluctuations. The green line represents near steady-state performance within each cycle, thus eliminating the impact of plunger-pump-induced voltage fluctuations. The reasons for these fluctuations will be analyzed in detail later. If the current density is further increased to 0.5 A/cm², the voltage response is likely to show even more drastic fluctuations. Nevertheless, the performance of constant-current electrolysis at the current density of 0.4 A/cm² remains stable.

The I-V-t curves displayed in Figure 6c reflect the testing set 03 results at 0.2 A/cm² with 75% FU, which is higher than the FU in testing set 01. The results indicate that both during the current cycling and the constant-current electrolysis processes, the performances are very stable. Comparing these results with those at 0.2 A/cm² and 50% FU, it is confirmed that the impact of FU at low current densities is quite limited. Therefore, a higher FU can be adopted to enhance system efficiency. Similarly, the I-V and EIS data shown in Figure 7e–h suggest that the electrochemical performance remains stable throughout the current cycling and galvanostatic electrolysis processes.

In test set 04, the most rigorous conditions were applied with a high current density of 0.4 A/cm² and high FU at 75%. Figure 7d shows that the on/off cycling between 0 and 0.4 A/cm² results in periodic performance fluctuations, with the fluctuation period and pattern being similar to those observed in Figure 7b. The voltage fluctuation range was also similar (~1 V). This implies that the fluctuations are more likely due to activation processes rather than concentration processes, which will be the focus of the following analysis. The constant-current electrolysis at 0.4 A/cm² remains stable.

2.2.3. Discussion on Voltage Fluctuations Under High-Current Cycling

To investigate the cause of voltage fluctuations under high-current-density cycling and to seek possible solutions, another 130 h on/off cycling test at 0–0.4 A/cm² was conducted as shown in Figure 8. The main difference between this test and test set 02 in the former 350 h test was the inclusion of multiple I-V and EIS test points during the on/off cycling. In addition, four OCV treatments using pure H₂ at segment numbers 4, 10, 17, and 22 were conducted in the first 60 h trying to alleviate the voltage fluctuations, followed by continued electrolysis in the subsequent 70 h of operation.

Figure 9 displays I-V and EIS results at different voltage states, specifically at the valley, plain, and peak points as shown in Figure 8. It is observed that the total impedance at peak (#21) is significantly higher. To further investigate the origins of this increased impedance, a Distribution of Relaxation Times (DRT) analysis was conducted using the Tikhonov regularization method. The regularization parameter was set at 0.001, which provides a balance between resolution and uncertainty in the DRT results for typical noise levels in input data. The results from the DRT analysis suggest that the main difference originates from the activation overpotential of the Ni-YSZ cathode. The instability of the cathode catalyst under high-overpotential conditions leads to continuous changes in activation impedance.

2.3. Temperature Cycling of the SOEC Stack

2.3.1. Impact of Temperature Cycles on Stack Performances

The ability to endure temperature cycling is one of the key indicators for the performance evaluation of a SOEC stack. Figure 10a shows the I-V-t curve of the stack undergoing three thermal cycles. In the first two cycles, 10% H₂/90% H₂O with the same composition as the fuel gas mixture was fed into the stack during the cooling process. In comparison, pure H₂ was used in the third thermal cycle in the cooling stage. The temperature cooling rate was set as a constant of 0.5 °C/min for all three thermal cycles to ensure that the EIS data tested at each temperature point would not exhibit thermal drift. In the first cycle, the stack operating temperature gradually decreased from 750 °C at segment #1, shown in the upper bar in Figure 10a, to 550 °C and then increased to 750 °C at segment #3 for performance measurement. In the second cycle, the lowest temperature was further reduced to 450 °C. The stack performance was measured at segment #6 when the operating temperature reached 750 °C again. EIS data were recorded at open-circuit voltage throughout the process. As can be seen in Figure 10a, the voltage curves are symmetric around the center at the lowest temperature point in the first two thermal cycles, indicating that the short stack has acceptable thermal cycling performance. However, the OCV increased after the thermal cycles, as shown by the red curve at segments #1, #3, and #6. The DRT presented in Figure 10c indicates that differences are observed not only in the activation impedance but also in the concentration impedance following temperature cycling.

2.3.2. Performance Change by Cooling with the Fuel Gas Mixture

As illustrated in Figure 11, the impedance changes under open-circuit conditions were tested at every 50 °C interval during cooling and reheating with the 10% H₂/90% H₂O mixture in the second thermal cycle. Good repeatability of the EIS at the same temperature throughout the cycles was observed. The increase in ohmic resistance was the primary change after the decrease in stack temperature due to increased ionic transfer resistance in the electrolyte at low temperatures. There was a marked increase in concentration impedance between 650 and 600 °C, which was also accompanied by increases in activation impedance, indicating a slower reaction rate, and concentration impedance, indicating slower mass transfer.

2.3.3. Performance Change by Cooling with Pure Hydrogen

Similarly, the impedance changes under open-circuit conditions were tested at every 50 °C interval during the cooling process in the third thermal cycle with pure hydrogen. As observed in Figure 12a, the low-frequency EIS is more stable, displaying an almost 45° line at 650 °C and 750 °C without the interference of water vapor, and it is necessary to use the more complex ECM method for EIS interpretation.

Figure 12b–d separately display the ECM decomposition results, the ECM circuit, the trends in impedances, and the DRT analysis results. Based on a previous study on a similar type of SOEC [16], the four peaks (five impedance values) are attributed to the following electrochemical processes: R0 represents the ohmic resistance (R_s) as shown in Figure 12c, where the increase in ohmic resistance is one of the main mechanisms affected by cooling. The ohmic resistance linearly increases between 750 °C and 650 °C, with a more significant rise at 600 °C. Compared with Figure 12a, R_s seems to be minimally affected by the gas composition introduced during cooling. R_ct (corresponding to the high-frequency process P1) is attributed to the charge transfer resistance at the electrode/electrolyte interface and the electrochemical reaction kinetics, which increases with the decrease in temperature. R_{Ni/YSZ_TPB} represents the charge transfer resistance within the Ni/YSZ electrode (corresponding to P2). R_ct and R_{Ni/YSZ_TPB} together make up the polarization resistance R_Ni/YSZ of the Ni/YSZ electrode, contributing to the partial performance deduction during temperature drop. P3 is likely due to the oxygen evolution reaction at the oxygen electrode active sites, which becomes significantly larger below 700 °C. P4 is usually associated with gas diffusion mass transfer and gas conversion polarization processes, which are difficult to distinguish from P3 above 600 °C due to overlap but shift to lower frequencies and become significantly larger below 600 °C. The frequency ranges corresponding to the processes of peaks P1–P4 match well with those reported in existing studies [19,20].

3. Experimental Setup

In the experimental setup, a cross-flow configuration was employed using a commercial short stack from Ningbo SOFCMAN Energy Technology Co., Ltd., Ningbo, Zhejiang, China. The stack consisted of six cells, each measuring 10 × 10 cm with an active area of 69 cm², as shown in Figure 13. The cell construction included NiO/YSZ‖YSZ‖CGO‖LSCF/CGO layers, with thicknesses of 400 μm for NiO/YSZ, 10~15 μm for YSZ, 2~3 μm for GDC, and 20~25 μm for LSCF-GDC.

Figure 14 is the schematic of the test setup. The stack was mounted in a bolt-loaded test fixture and placed inside a furnace for testing. Gas flow to and from the stack was facilitated through gas pipes connected to the fixture. The fuel side utilized a high-precision plunger pump and a steam vaporizer with mixing capabilities (produced by Suzhou Friends Experimental Equipment Co., Ltd., Suzhou, China, model FD-HG20), along with a hydrogen generator to supply a mixture of steam and hydrogen gas at 180 °C. The flow rate of hydrogen was precisely controlled by a mass flow controller (Sevenstar Co., Ltd., Beijing, China, model CS200). An air compressor and the same model of mass flow controller were used on the air side to supply purge air. Air was supplied at a constant rate of 5 SLM. Fuel flow varied based on the operating conditions, including fuel utilization and steam fraction. The setup operated under atmospheric pressure conditions, as no backpressure devices were used. Both the fuel gas mixture and air were preheated sufficiently in coils placed in a muffle furnace before feeding into the stack. The stack was placed within a muffle furnace equipped with a programmable controller (manufactured by Hefei MTI, Hefei, Anhui, China, model VBF-1200) to maintain the working temperature of 750 °C. The current was applied to the stack using current collectors on both sides. An 800W/80V/60A DC power supply (provided by ITECH Electronic Co., Ltd., Farmington, NM, USA, model IT6502D) was used to deliver the required current for water electrolysis. Impedance testing was conducted using a Solartron eight-channel electrochemical workstation (Solartron Co., Ltd., Bognor Regis, UK, including a potentiostat 1470E, impedance analyzer 1260, and frequency response analyzer 1455A). Although the configuration of the electrochemical workstation paralleled with an external power supply may have introduced some interference to the EIS measurements, it is common in EIS testing of large-area cells [24]. At OCV, a disturbance of 10 mA was applied. In other cases, the disturbance current was set at 3% of the current electrolysis current.

4. Conclusions

This study uses a short stack composed of six SOEC single cells to investigate the impact of multiple operating conditions on SOEC durability. Steady-state performance tests were conducted, and appropriate conditions were selected to assess the stability and durability of the SOECs under different FU and current states. I-V curves, EIS, and DRT analysis tools were used to examine voltage fluctuations caused by current cycling at high current densities. In addition, the stability of the SOECs under temperature cycling was tested to verify the stack’s temperature cyclability. The following specific conclusions were drawn from these studies:

(1)

For the specific stack tested, a fuel composition of 10% H₂-90% H₂O yields the best performance. Due to the overshoot of voltage response, it is suggested that the steady-state I-V test current step should be 0.1 A/cm². FU does not substantially enhance steady-state performance, while a lower FU can mitigate overshooting and fluctuations caused by mass transfer lag in dynamic processes.

(2)

The stack is capable of withstanding constant-current electrolysis within the range of 0–0.4 A/cm², as well as cyclic currents of 0–0.2 A/cm². A cyclic current of 0.4 A/cm² results in larger fluctuations, which, according to impedance analysis, could be due to instability in the cathode’s activation impedance.

(3)

The SOEC short stack can withstand at least two temperature cycles. All impedances increase as the temperature decreases, with a dramatic performance drop between 650 and 600 °C, suggesting that 650 °C is the minimum operating temperature of the stack.