1 Introduction
Solid oxide cells (SOCs) have attracted global attention due to their high energy conversion efficiency, low pollution emission, relatively flexible fuel choice, and the capability to meet the requirement for large scale energy storage and conversion [1–3]. SOCs have two reversible operating models, solid oxide fuel cell (SOFC) and solid oxide electrolysis cell (SOEC). SOFC can convert chemical fuels, such as H2, CO, and hydrocarbons, directly to electricity with high energy efficiency, high energy and power densities [4–6]. In contrast, SOEC can store electricity into chemical fuels, such as H2, CO, or syngas, through water electrolysis, CO2 electrolysis or H2O/CO2 co-electrolysis [7,8]. Due to the above characteristics, SOCs are considered as a robust and flexible energy conversion technology to cope with the wide range of operating conditions in future energy scenarios. Historically, long-term stability and thermal (cool-down/heat-up) and emergency shutdown cycle tolerance have posed critical challenges to multi-scenario applications of SOCs [1].
Interconnector is a critical component in SOC stacks, which provides electrical connection between contiguous cells as well as separation of oxygen (cathode side in SOFC model) and fuel gases and/or steam (anode side in SOFC model) [9–11]. Fe–Cr based metallic materials, such as Crofer22 APU, SUS304, and SUS430, have been widely used in the interconnector of SOCs due to their excellent electrical conductivity, high-temperature structural stability, and matched coefficient of thermal expansion (CTE) with other SOC components [12–14]. However, Cr2O3 scale, which is formed by oxidation of interconnector at a high temperature, is known to cause performance degradation of the stacks [15–17]. Furthermore, Cr2O3 can react with O2 and H2O to form CrO3 or CrO2(OH)2 to poison the cathode of SOFC (anode of SOEC) and further deteriorate SOC performance [18–20]. These two negative factors lead to poor long-term stability and thermal (cool-down/heat-up) cycle tolerance of SOCs.
Coating a conductive ceramic layer on interconnector is an effective method to inhibit the growth rate of Cr2O3 scale and prevent the diffusion of CrO3 or CrO2(OH)2 [21–23]. A lot of ceramic materials including oxides, perovskites (La1−xSrxMnO3 [24]) and spinels (MnxCo3−xO4 [25,26], CuxFe3−xO4 [27,28], NixFe3−xO4 [29,30], etc.) have been used as the coatings of interconnector. Screen printing, sol-gel, magnetron sputtering [31] and plasma spraying [32], electrophoretic deposition [33,34] and so on are used to the coating preparation. However, as an important index to evaluate the long-term stability and durability of SOC coatings, there are few studies on rising and cooling cycle test.
In this study, (La0.75Sr0.25)0.95MnO3−δ (LSM) and Mn1.5Co1.5O4 (MCO) powders are applied to the preparation of interconnector coatings. To meeting the needs of higher requirement of coating uniformity and large-scale coating fabrication, two thermal spraying technologies, atmospheric plasma spraying (APS) and low-pressure plasma spray (LPPS), are used for the coating preparation. The structure, morphology, and electrochemical performance of different coatings fabricated using different methods are investigated and compared in detail. Cr diffusion-inhibited performance of the MCO coatings is tested and evaluated on rising and cooling cycle operation.
2 Experimental
2.1 Powder characterization and granulation
The LSM and MCO powders used to fabricate coatings in this study were purchased from Ningbo SOFCMAN Energy Technology Co., Ltd. and Fuelcell Energy, Inc., respectively. X-ray diffraction (XRD, Cu Kα radiation) was used to confirm the crystal patterns and the four-electrode method was used to measure the electrical conductivity of the two powders. The median particle size (d50) of the powder in the thermal spraying process must be greater than 25 μm. However, the d50 of the initial LSM and MCO are 0.5–2 μm and 0.5–0.84 μm, respectively, which cannot meet the requirements of thermal plasma spraying process. Therefore, spray drying granulation process [35] was adopted to prepare suitable particle size powder before plasma spraying, and scanning electron microscopy (SEM) was adopted to characterize the morphology and particle size of the powders after granulation.
2.2 Interconnector coating preparation by plasma spraying process
LSM and MCO coatings were prepared on stainless steel (SUS430) interconnector. The two kinds of coatings (LSM-APS and MCO-APS) were manufactured by APS in the Oerlikon Metco UniCoat F4 system. In addition, the MCO coating was also manufactured by LPPS in the Sulzer Metco LPPS-TF system (MCO-LPPS).
The spraying process schematic was shown in Fig.1. For the APS process, SUS430 substrates were mounted on fixture, and the granulation powders were transported to the plasma torch by carrier gas. Then, a mechanical arm moved torch along a path and the powders were sprayed by the torch to coat stationary substrates at a standoff distance. Different from APS, LPPS need to be vacuumed and maintained in an inert gas state during the spraying process. After SUS430 substrates were mounted on fixture, the LPPS system was vacuumed and filled with inert gas. Then, a spraying process similar to APS was operated. Air was not allowed into the LPPS system until the substrates were sprayed and cooled to room temperature. The SUS430 samples of 10 mm × 10 mm × 2 mm were sprayed on both sides. The operating parameters of the two plasma spraying processes are listed in Tab.1 and Tab.2, respectively. The surficial and cross-section morphologies and structures of the rectangular samples with different coatings were observed using SEM, and the distribution of characteristic elements of the samples at the interface was detected and analyzed using energy dispersive spectrometer (EDS) in linear scan mode.
Tab.1 Main parameters of APS process |
| Parameter | Value |
|---|---|
| Power/kW | 48 |
| Torch current/A | 515 |
| Flow rate of Ar/(L·min–1) | 40 |
| Flow rate of H2/(L·min–1) | 12 |
| Standoff distance/mm | 110 |
| Powder feed rate/(g·min–1) | ~10 |
| Rotational speed/% | 15 |
| Stirring rate/% | 60 |
| Cooling gas/MPa | 0.15 |
Tab.2 Main parameters of LPPS process |
| Parameter | Value |
|---|---|
| Vacuum chamber pressure/Pa | 50 |
| Power/kW | 115 |
| Plasma current/A | 2200 |
| Flow rate of Ar/(L·min–1) | 80 |
| Flow rate of He/(L·min–1) | 90 |
| Standoff distance/mm | 650 |
| Powder feed rate/(g·min–1) | 80 |
| Powder gas flow rate (Ar, L·min–1) | 10 |
| Spraying line velocity/(mm·s–1) | 300 |
| Preheating temperature/°C | 700 |
| Cooling pressure/MPa | 50 |
| Cooling time/min | 20 |
2.3 Electrochemical measurement
As shown in Fig.2, the electrical resistance of the rectangular samples was measured by using a four-point method from 650 to 800 °C in air, and interconnector samples without coating were used as comparisons. Ag paste was used as current collector in the electrochemical measurement. The resistance (R) of the samples was measured using an electrochemical workstation (IM6ex, Zanher) in the mode of electrochemical impedance spectroscopy (EIS) with an amplitude of applied voltage of 10 mV and a frequency range of 1 MHz to 0.1 Hz. The area specific resistance (ASR) was calculated using
where A is the superficial area of the rectangular samples. The stability and durability of the coatings were evaluated by several rising and cooling cycle tests from room temperature to 800 °C. After electrochemical test, the cross-section patterns of the samples were observed using SEM and the Cr diffusion phenomenon at the interface was detected and evaluated using EDS.
3 Results and discussion
3.1 Structure, morphology, and performance of LSM and MCO powder
The structure, morphology, and basic performance of the LSM and MCO powders are measured and analyzed to provide reference for the preparation and performance analysis of corresponding coatings. Crystal structures of the initial LSM and MCO powders are identified by XRD patterns in Fig.3(a) and Fig.3(b). LSM (PDF#89-4461) shows a typical perovskite crystal type while MCO (PDF#84-0482) shows a typical spinel crystal type, and no impurity phase is detected within the two powders. The diffraction pattern for the LSM powder has 8 peaks at 22.9°, 32.5°, 40.2°, 46.8°, 52.6°, 58.0°, 68.1°, and 78.0°, corresponding to (012), (110), (202), (024), (122), (300), (220), and (128), respectively, while that for the MCO powder has 9 peaks at 18.5°, 30.5°, 35.9°, 37.6°, 43.7°, 54.2°, 57.8°, 63.5°, and 76.2°, corresponding to (111), (220), (311), (222), (400), (422), (511), (440), and (622), respectively. The electronic conductivity test results are demonstrated in Fig.3(c). The electronic conductivity (σ) of LSM is obviously higher than that of MCO. At 800 °C, which is a general operating temperature of SOCs, for instance, the σ of LSM is 155.56 S/cm2, 2.6 times larger than that of MCO.
The SEM patterns of LSM and MCO granulation powders are provided in Fig.4(a) and Fig.4(b) and Fig.4(c) and Fig.4(d), respectively. The LSM and MCO powders have a typical spherical particle morphology, as shown in Fig.4(a) and Fig.4(c). The particle size and uniformity of the powders are measured and analyzed from Fig.4, which provides a preliminary judgment on whether the corresponding powder is suitable for coating preparation. The homogeneity of LSM, with a diameter ranging from 18.44 to 31.55 μm (measured in Fig.4(b)), is more uneven than that of MCO with a diameter of approximately 37.9 μm (measured in Fig.4(d)). As a result, the MCO granulation powder is more suitable for the thermal plasma spraying process due to the more uniform and larger particle morphology.
3.2 Structure and morphology of LSM and MCO coatings
The structures and morphologies of the SUS430 samples with different coatings are shown in Fig.5. As demonstrated in Fig.5(a), Fig.5(d), and Fig.5(g), the surface of SUS430 with different coatings is not obviously spherical particles of powders, but in a flat and dense structure. The locally enlarged images of Fig.5(a), Fig.5(d), and Fig.5(g) shown in Fig.5(b), Fig.5(e), and Fig.5(h) reveal that the three coatings are composed of micro/nano-particles and the particles size is smaller than the granulation powder before spraying, suggesting that the sprayed powder undergoes a process of melting and recrystallization, which results in a strong bonding between powders caused by melting, improving the gas barrier capacity of the coatings and benefiting for avoiding the oxidation of the interconnector [32].
In addition, Fig.5(b), Fig.5(e), and Fig.5(h) also show that the micro/nano-particle distribution of MCO-LPPS is the densest and most uniform, followed by MCO-APS and LSM-APS. The cross-section SEM patterns in Fig.5(c), Fig.5(f), and Fig.5(i) exhibit continuous coating structures of the three samples. This can be explained by the fact that under the function of high temperature plasma, the powders are attached to the surface of SUS430 in a molten state, and fuse with each other, and finally form a dense and consecutive coating structure. Such melting-induced strong bonding among the powders would result in a high gas tightness of the coating, which is conducive to inhibiting the formation of oxide layer and preventing the diffusion of toxic chromium compounds.
Furthermore, it is also known from Fig.5(c), Fig.5(f), and Fig.5(i) that the coating thickness of LSM-APS, MCO-APS, and MCO-LPPS is approximately 43, 35, and 35 μm, respectively. The porosity is analyzed by using ImageJ software. Compared to MCO-APS and MCO-LPPS coatings (the porosity being 5.6% and 3.4%, respectively), more pores can be observed locally in the LSM-APS coating (the porosity being 10.1%), suggesting that MCO powders can form a denser coating and the LPPS process can fabricate the lowest porosity coating. As the porosity decreases, the gas barrier performance increases, which can reduce the decay rate of the electrochemical performance of SOCs in long-term operation.
3.3 Conductivity and stability of different coatings
The R of different rectangular samples was tested by EIS. As Fig.6(a) shows, the EIS results of SUS430 with LSM-APS, the point the curve intersect with the horizontal coordinate is the R of the sample at different temperatures, from which, the corresponding ASR can be calculated using Eq. (1). The ASR of other samples was obtained by the same method. Fig.6(b) compares the ASR of bare SUS430 and SUS430 with LSM-APS and MCO-APS at different temperatures. The MCO-APS exhibits the lowest ASR values of 0.0086, 0.0063, 0.0047, and 0.0033 Ω·cm2 at 650, 700, 750, and 800 °C, respectively.
The ASR of the LSM-APS is higher than that of MCO-APS but lower than that of the bare sample. At 800 °C, for instance, the ASR of the LSM-APS is 0.0100 Ω·cm2, much higher than that of MCO-APS (0.0033 Ω·cm2) and lower than that of the bare sample (0.0340 Ω·cm2). The results illustrate that the oxidation of the bare sample at high temperatures is considerably rapid so that the impedance is significantly higher than that of the coated samples. Although the electronic conductivity of the oxide coatings is lower than that of the interconnector itself, the ASR of the sample with coatings is lower than that of the bare sample because the coatings can inhibit the oxidation of the interconnector, which can significantly reduce the electronic conductivity of the interconnector and eventually lead to a rapid performance degradation of the SOC stack. Moreover, Chen et al. [24] fabricated LSM-APS on SUS430. The initial ASR of the sample is similar (0.0150 Ω·cm2) to the result in the present study, and the ASR of the former after oxidation for 300 h in air at 800 °C has barely increased, indicating an excellent long-term oxidation resistance [24]. Furthermore, Fig.6(b) also indicates that the initial inoxidizability of the MCO coating is better than that of the LSM coating, as the electronic conductivity of the MCO powder is smaller but the ASR of MCO-APS is higher.
The inoxidizability and stability of the rising and cooling cycle test of MCO-APS and MCO-LPPS are analyzed and compared to select a more suitable thermal spraying method for coating preparation. Fig.6(c) shows that the ASR of MCO-LPPS (0.0027 Ω·cm2 at 800 °C) is lower than that of MCO-APS (0.0033 Ω·cm2 at 800 °C) and the bare sample, which means that the initial inoxidizability of MCO-LPPS is stronger than that of MCO-APS. Fig.6(d) and Fig.6(f) show the ASR of MCO-APS and MCO-LPPS for the 4 rising and cooling cycle tests, respectively. The result indicates that at 800 °C, the ASRs of the MCO-APS and MCO-LPPS are still smaller than the initial ASR of the bare sample even after 4 cycle tests, showing the good oxidation resistance and cycle stability of MCO coating.
Moreover, Fig.6(d) and Fig.6(e) also shows that the ASR of the coated samples increase with the increase of the cycle number, illustrating that the coatings can only delay but cannot completely prevent the oxidation of the interconnector. Furthermore, Fig.6(f) proves that the ASR of MCO-APS is smaller than that of MCO-LPPS at 800 °C after 2 and 3 cycle tests but the difference is tiny (0.0020 Ω·cm2 after 2 cycle test and 0.0017 Ω·cm2 after 3 cycle test). However, opposite result is observed after 4 cycle and the ASR of MCO-LPPS is 0.0032 Ω·cm2 lower than that of MCO-APS. As a result, it can be concluded that MCO-LPPS has a better oxidation resistance and stability in long-term and frequent rising and cooling cycle operation. The LPPS technology is more suitable for the preparation of the MCO coating for the actual interconnector.
3.4 Cr diffusion-inhibited performance
Element distribution on the interface of the two-type sprayed MCO coatings before and after cycle test is analyzed and compared to evaluate the Cr diffusion inhibition performance of the coatings. Fe and Cr are the characteristic elements of SUS430, while Mn and Co are the characteristic elements of MCO. As the EDS patterns demonstrates in Fig.7(a) and Fig.7(c), the initial MCO-APS and MCO-LPPS samples exhibit a consistent distribution of the characteristic elements near the interface (yellow box area in Fig.7). Along the direction of the white arrow in Fig.7(a) and Fig.7(c), Fe and Cr decrease rapidly, while Mn and Co increase significantly at the SUS430/coating interfaces.
After 4 cycle tests, Mn, Co, and Fe exhibit the same distribution pattern as the initial sample, while the element distribution of Cr of the two samples shows a marked difference compared to the initial samples. Cr shows an obvious aggregation and increase at the interface in Fig.7(b) and Fig.7(d), and the relative content is higher than that of the body phase. The results demonstrate that even if coatings are applied to the surface of SUS430, there still exists the Cr migrating from bulk to interface and forming oxides at the interface during the cycle operations. Additionally, the relative intensity of Cr at the interface in MCO-APS is much higher than that in MCO-LPPS, which means that the MCO-LPPS coating has a better performance than the MCO-APS coating in inhibiting Cr migration and oxide layer growth.
4 Conclusions
Interconnector is a critical component for SOC stack construction. Coating on the interconnector surface is an important method to inhibit the oxidation and Cr migration of the interconnector. In this study, APS and LPPS technology was applied to manufacture LSM and MCO coatings on SOC interconnector surface. In addition, the electrochemical performance and Cr diffusion-inhibited performance of the coatings were analyzed and evaluated. The major conclusions are as follows:
1) The ASRs of MCO-APS and LSM-APS are 0.0010 and 0.0033 Ω·cm2 at 800 °C, respectively. Although the electrical conductivity of the LSM powder is much higher than that of the MCO powder, MCO-APS has a better electrochemical performance, possibly because MCO-APS can form a denser coating, thereby improving the antioxidant capacity of MCO-APS.
2) The initial ASR of MCO-LLPS is 0.0027 Ω·cm2 at 800 °C, lower than that of MCO-APS. Moreover, the ASR of MCO-LPPS is 0.0032 Ω·cm2 lower than that of MCO-APS after 4 cycle operation. Compared to APS, the higher flame temperature and cleaner operating environment of LPPS enable the preparation of denser coatings that exhibit a better oxidation resistance and cycle stability.
3) MCO-LPPS coating has a better performance than MCO-APS coating in inhibiting Cr migration of SUS430. The relative intense of Cr at the interface of SUS430 with MCO-LPPS is lower than that of MCO-APS after 4 rising and cooling cycle operations.
This work provides data support and reference for the performance changes of MCO coatings in frequent start-stop applications in SOC systems. The MCO-LPPS method will be used for the coating preparation of actual interconnectors for the construction of SOC stacks. The electrochemical performance, long-term stability and thermal (cool-down/heat-up) cycles tolerance of the SOC stacks with MCO-LPPS coatings will be tested and evaluated in the SOFC mode and the SOEC mode in the following study. In addition, this work also provides an advanced coating preparation method, which can not only be used for the fabrication of interconnector coatings, but also offer guidance for the coating preparation of key components of other electrochemical devices such as proton exchange membrane fuel cell.