Effect of High-Temperature Exposure on Degradation of La 0.6 Sr 0.4 CoO 3 Coated Metallic Interconnects for ITSOFC and SOEC

Previous studies showed the formation of new phases affecting the electrical properties of LSC thin films deposited on stainless steel substrates, which are commonly tested for ITSOFC and SOEC interconnects. A 4.3 μm thick La 0.6 Sr 0.4 CoO 3 coating was deposited on AISI430 steel by spray pyrolysis, followed by heat treatment (800°C/2h) and an oxidation in air (800°C/96h). The La 0.6 Sr 0.4 CoO 3 phase interacted with the metallic substrate and formed SrCrO 4 , causing degradation of the perovskite into La 0.9 Sr 0.1 CoO 3 . An EDS mapping showed Sr and Cr enrichment in the coating/substrate interface. TG analysis indicated a lower mass gain for the coated substrate. The total ASR at 800°C of the interconnect before and after oxidation was 3.23 Ω.cm 2 and 3.98 Ω.cm 2 , respectively. The Ea underwent very small variation, remaining around 0.24 eV (T≤300°C) and 0.65 eV (T≥400°C). The reaction of Cr from the substrate and Sr from LSC seems to have impaired the performance of the interconnect.


Introduction
Metallic interconnects have been used in intermediate temperature solid oxide fuel cells (ITSOFC) and solid oxide electrolysis cells (SOECs) with operating temperatures between 600°C and 800°C. The maximum total area specific resistance (ASR) for the application is 0.1 Ω.cm 2 in the whole temperature range. In comparison to ceramic materials, metallic materials present a higher electrical conductivity, higher mechanical resistance, easier manufacturing, and lower cost [1][2][3][4] . Many heat resistant alloys, such as chromium, nickel or cobalt-based alloys, have been studied for this application. Ferritic stainless steels were selected due to their good oxidation resistance, thermal expansion coefficient (TEC, 10-14x10 -6 K -1 ), compatibility with the other components, and low cost compared to other alloys 5,6 . However, under operating conditions, in oxidizing and reducing atmospheres, these metallic alloys present a oxidation rate that follows the growth of a chromium oxide layer with high electrical resistance, and chromium vaporization impairing the ITSOFC and SOEC performance [7][8][9][10][11][12] .
One alternative to protect the ferritic stainless steel from high-temperature oxidation is applying a protective oxide coating to its surface [13][14][15] . These coatings must act as a barrier to avoid chromium diffusion from the metallic substrate and oxygen diffusion to the metallic alloy, maintaining thermomechanical and chemical stability during fuel cell operation 16 . The obtained coatings have the advantages of easy deposition and a cost lower than heat resistant metallic alloys that need to have specific compositions and to be free of undesirable impurities 17 .
Perovskite cobaltite coatings interact with the stainless steel and a new phase is formed, SrCrO 4 29,41 . However, the electrical behavior of these phases needs further investigation given that some authors 42 showed a beneficial effect whereas others 43 showed that this phase is non-conductive and that SrCrO 4 formation leads to lower conductivity.
As a continuation of our previous investigation 29 , the purpose of this study was to investigate the effect of oxidation on AISI 430 ferritic stainless steel coated with La 0.6 Sr 0.4 CoO 3 (LSC), obtained by spray pyrolysis. The electrical conductivity at high temperatures was evaluated in an inert atmosphere using electrochemical impedance spectroscopy (EIS). Additionally, the phase formation and stability of LSC during an air oxidation process were analyzed.

Substrate preparation
In this work, 1 mm thick with 1 cm 2 AISI 430 ferritic stainless steel was used as a substrate for the coating. The chemical composition of the AISI 430 is 16.44 Cr, 0.06 C, 0.43 Mn, 0.036 Mo, 0.40 Ni, 0.02 Co, 0.002 Al, and 0.02 Nb (wt %), with the balance Fe. The samples were cut into 10 x 10 mm pieces. SiC paper (120-600 -grid) was used to sand these pieces. . This solution was placed in a reservoir connected to an atomizer nozzle. The precursor solution was sprayed onto the ferritic stainless steel samples, which were placed on a heated metal plate at the desired temperature. The metallic substrate was placed under the spray nozzle to obtain a more homogeneous film deposition. The deposition parameters were shown in our previous paper 29 .

Coating deposition
The coatings obtained by this method were amorphous. To obtain a crystalline phase composed of La 0.6 Sr 0.4 CoO 3 (LSC), the films were heated to 800°C in air for 2 hours, using a 10 °C min -1 heating rate.

Coating characterization
The coating surface morphologies and cross-sections were examined by scanning electron microscopy (SEM) (JEOL -JSM 6060). The chemical composition of the films was analyzed by energy dispersive spectroscopy (EDS) (JSM 5800). The structural characteristics of the coatings before and after the oxidation test were analyzed by XRD (Philips X-Ray Analytical Equipment X'Pert-MPD System). The scans were collected in the 2Ɵ range (20-75°) with a 0.02° variation.
The oxidation behavior was observed by thermogravimetric analysis (TG) and by a furnace test on both the coated and uncoated samples. The TG analysis was performed for 24 hours at 800°C in air to measure the thermal history, whereas the furnace test was 96 hours at 800°C in air. The test was conducted with a 10 °C min -1 heating rate.
Impedance spectroscopy analyses were performed for coated samples before and after oxidation at 800°C for 96 hours. A gold sputtered cover was deposited on both sides of coated substrates to act as a contact electrode, avoiding double layer effects. The electrical measurements were carried out in an argon atmosphere, to prevent further oxidation, and the samples were stabilized for 1 hour at each temperature, in the range of 100°C to 800°C. An impedance analyzer (7260 Impedance Analyser from Materials Mates) was used in the 1 Hz to 10 MHz frequency range, at OCP with a +/-20 mV AC amplitude using a two-probe set.

Results and Discussion
The XRD patterns obtained in this study are exhibited in Figure 1. Comparing the uncoated substrate before and after oxidation for 96 hours at 800°C (Figure 1a and Figure 1b), significant amounts of two main oxides were formed during the oxidation process. Intense peaks of Cr 2 O 3 (PDF nº 00-038-1479) were detected with equally intense peaks of MnCr 2 O 4 (PDF nº 01-075-1614), in agreement with the oxidation products of AISI 430 found by Rufner et al. 44 .
After spray pyrolysis at 550°C for 30 min, the deposited coating has no distinct XRD reflections and the coating was amorphous at the XRD analysis scale (not shown). However, after heat treatment at 800°C for 2 hours in air (Figure 1c), it was possible to observe the formation of the desired phase, La 0.6 Sr 0.4 CoO 3 (LSC) perovskite (PDF nº 01-089-5719), as expected from the precursor solution preparation fractions. Moreover, the substrate phase could be distinguished (PDF nº 01-087-0722) among the perovskite peaks. Also, very low-intensity peaks were observed and can be attributed to a very low concentration of SrCrO 4 (PDF nº 00-015-0356), which appears to start forming even with a time as low as 2 hours at 800°C (heat treatment). The formation of SrCrO 4 may occur due to the hexavalent chromium ion, which diffuses through the coating 45 . Residual Co 3 O 4 (PDF nº 03-065-3103) that was not fully incorporated into the perovskite lattice was observed.
Considering the result of the LSC coated sample, it is noted that the oxidation at 800°C for 96 hours (Figure 1d) caused some modifications in the phases present in the film. First, the dissolution and incorporation of Co 3 O 4 , which was no longer detected in the pattern. Also, the more intense peaks of SrCrO 4 may indicate the formation of higher amounts of this compound, followed by a loss of strontium in the LSC perovskite, resulting in the La 0.9 Sr 0.1 CoO 3 composition (PDF nº 00-028-1229). The presence of Sr in the LSC coating was determinant in the evolution of phases in the surface of the AISI 430 steel due to the reactivity of strontium oxide (SrO). The SrO rapidly reacted with Cr 2 O 3 , forming SrCrO 4 , which seemed to have suppressed the formation of the MnCr 2 O 4 phase. According to Chen et al. 46 , the formation of SrCrO 4 at the interface is thermodynamically favorable. However, as described above, it cost the perovskite a severe Sr impoverishment. Figure 2 shows a cross-section of the AISI 430 coated with LSC after heat treatment for 2 hours at 800°C. The film thickness is heterogeneous, the average thickness was 4.3 ± 0.9 μm. The irregular shapes of formed films are related to the spray pyrolysis process. The small droplets land on the substrate surface to undergo a pyrolytic reaction. Therefore, isolated agglomerates with irregular shapes are formed 47 . Figure 3 shows images obtained by elemental distribution using EDS for cross-sections of the LSC coated substrate after heat treatment for 2 hours at 800°C. The elemental distribution demonstrates the presence of the coating on top of the substrate surface (Figure 3b, 3d, 3e, and 3f). In Figure 3e and 3f, it is possible to observe Sr and Cr enrichment in the substrate/coating interface. This possibly indicates the location of the SrCrO 4 , detected in the XRD analysis (Figure 1c), which starts with the reaction between the Sr from perovskite and the Cr from the metallic substrate 48 . Liu and Konysheva 49 indicated through impedance analysis, that SrCrO 4 precipitation occurs by growing a layer between the Sr rich and Cr rich interfaces. In previous work 29 , a computational simulation was performed, that showed the stability of the SrCrO 4 oxide when a system containing Sr and variable amounts of Cr was exposed to an environment rich in O 2 at 800°C, causing perovskite degradation from the La 0.6 Sr 0.4 CoO 3 phase to the La 0.9 Sr 0.1 CoO 3 phase, in agreement with the observed degradation of the perovskite in the XRD results. Furthermore, Yokokawa et al. 43 showed that SrCrO 4 formation is thermodynamically favored for perovskites containing less stable tetravalent ions (Fe 4+ or Co 4+ ). Additionally, the presence of Cr was verified in the coating region (Figure 3e), suggesting that some amount of Cr ions diffuse from the substrate through the coating.
The TG analysis ( Figure 4) shows that the coated substrate presented a lower mass gain than the uncoated substrate after 24 hours at 800°C, which may be associated with an increase in the oxidation resistance of the surface coated with La 0.6 Sr 0.4 CoO 3 . The uncoated substrate had a mass gain of 1.745 mg cm -2 , while the coated substrate had a mass gain of 0.215 mg cm -2 . The initial oxidation (from t=0 to t= 88 minutes) is larger for the uncoated sample, which may be due to free oxidation of the substrate, probably associated with the formation of phases that can be observed in Figure 1b. The initial oxidation (from t=0 to t= 147 minutes) for the coated substrate may be due to oxidation of the uncoated side because the coating was deposited on one side of the metallic substrate and the SrCrO 4 formation, that can be observed in Figure 1c, which according to Chen et al. 46 , occurs through the reaction of oxides (Cr and Sr) with O 2 (g). Both substrates, uncoated from t=197 to t=855 minutes and coated from t=147 to t=951 minutes, showed a mass loss, which is associated with chromium vaporization from the metallic substrate 50-52 . Quadakkers et al. 48 observed that chromium transport happens due to volatile species, such as CrO 3 , and Cr 2 (OH) 2 in chromium-based alloys at 950°C. According to Kurokawa et al. 53 , the density of ceramic coatings deposited on stainless steel substrates dominates the suppression of chromium release rate. In this sense, it may be inferred that the coating obtained in this work was not effective to promote a barrier to chromium vaporization, because the chromium release rate observed in the TGA (between t~150 min and t~900 min) is similar to the coated and uncoated substrates.
If the barrier was effective, it would be expected a smaller chromium release rate in the coated substrate, even with only one side protected. Starting at 865 minutes, the uncoated substrate showed a mass gain associated with the growth of Cr 2 O 3 and/or Fe 2 O 3 54 . The distinction, by XRD, between Fe 2 O 3 and Cr 2 O 3 is not easy because of the similarities of the cell parameters 51,55 . The coated substrate, starting at 1082 minutes, showed a mass gain, which is associated with continued SrCrO 4 growth.
The impedance spectroscopy results of coated samples before and after oxidation at 800°C for 96 hours presented a resistance, as illustrated in the Nyquist diagram obtained at 800°C, shown in Figure 5a. From the resistance measured at different temperatures the Arrhenius diagram was obtained and can be seen in Figure 5b. For the coated heat-treated substrate (800 °C/2h) before oxidation, a decrease in the area specific resistance (ASR) was observed, with a rise in the temperature, from 11531 Ωcm 2 at 100°C to 3.23 Ωcm 2 at 800 °C. After oxidation at 800°C for 96 hours in air, the electrical resistance values at 100°C and 800°C were 14324 Ωcm 2 and 3.98 Ωcm 2 , respectively. Both results indicate that the LSC coated AISI 430 steels present semiconductor  behavior after the heat treatment (800 °C/2h) and after oxidation (800 °C/96h). Wu et al. 56 determined, using a DC four-wire method, that strontium-doped lanthanum cobaltites (La 1-x Sr x CoO 3 ) deposited on a samarium-doped ceria (SDC) electrolyte present a metallic-like conduction behavior, decreasing conductivity with rising temperature. De Souza and Kilner 57 tested the oxygen diffusion of La 1-x Sr x CoO 3 pellets, considering a band model where electrons occupy a wide and partially filled band. This seems to agree with the metallic-like behavior because large number of carriers may be available. Instead, the additional SrCrO 4 phase detected in the XRD and other secondary phases usually has typical semiconductor behavior and is less conductive than LSC, causing deleterious effects on the coating conductivity when present 43,49 . Liu and Konisheva 49 analyzed the influence of SrCrO 4 in the deterioration of lanthanum cobaltite cathodes in half cells and reported the semiconductor like behavior of SrCrO 4 . They also measured its total conductivity as 0.182 mS cm -1 at 800°C, which is 6 orders of magnitude lower than the lanthanum cobaltite cathode. In this study, considering the average thickness of 4.3 μm, the total conductivities of the samples at 800°C were 0.124 mS cm -1 after heat treatment (800 °C/2h) and 0.100 mS cm -1 after oxidation (800 °C/96h), which are values close to those cited above. This suggests that the small amount of SrCrO 4 , detected by very low intensity peaks in XRD analysis before the oxidation (Figure 1c), is already enough to cause severe deterioration on the total ASR. In turn, it can be related to the formation of a SrCrO 4 continuous layer in the perovskite/steel interface, as observed in the EDS mapping before oxidation (Figure 3e and 3f), instead of a random phase distribution. After the oxidation at 800°C for 96h, this SrCrO 4 layer is not greatly increased, due to the depletion of Sr in the perovskite, but still resulted in a slight increase in the total ASR.
The presence of SrCrO 4 seems to drastically deteriorate the performance of the LSC coating with respect to the electrical conductivity. The obtained values of ASR are too high. For a demanding role such as an ITSOFC interconnect, the total interconnect ASR must remain below 0.1 Ωcm 2 for proper performance 31 . However, compared to insulating phases such as Cr 2 O 3 and MnCr 2 O 4 , some authors may consider the formation of SrCrO 4 somewhat favorable to the overall performance of the interconnect 42,58 because the service life may be considered. Figure 5b, are virtually the same for samples before and after oxidation. In both samples, a variation in the E a values occurred between 300°C and 400°C.At low temperatures (T≤300°C) the E a before oxidation was 0.22 eV and varied very slightly to 0.27 eV after oxidation. At higher temperatures (T≥400°C) the E a increased to 0.65 eV and 0.66 eV, respectively. Wu et al. 56 observed a change in the slope, provoked by the shift in E a , and attributed it to a crystal transformation of LSC from a rhombohedral to a cubic structure. The E a shift, however, occurred at higher temperatures than what was determined in this work (300°C to 400°C) and a much lower Ea was also attributed to LSC, on the order of 10 -2 eV. An Ea of 0.35 eV, obtained by Song et al. 59 at lower temperatures, was attributed to the formation of Cr 2 O 3 . Higher Ea values (>1.0 eV) attributed to LSC were also reported when the oxygen reduction is considered. Yet, this work assumed that only electronic conduction occurred in the sample since the measurements were carried out in an argon flow.

Conclusions
LSC coatings were successfully deposited on an AISI 430 steel surface by spray pyrolysis, with an average thickness of 4.3 ± 0.9 μm. Heat treatment at 800°C for 2 hours in air led to the formation of the desired La 0.6 Sr 0.4 CoO 3 phase. However, it also initiated a reaction between the Sr in the coating and the Cr in the substrate, forming insulating SrCrO 4 , which grows larger during the oxidation process (800°C for 96 hours). This resulted in a Sr deficient LSC and decreased electrical properties in the coating. The ASR at 800°C after heat treatment was 3.23 Ω.cm 2 and showed only a small rise to 3.98 Ω.cm 2 after the oxidation process at 800°C for 96h. The high-temperature reaction of Cr from the substrate and Sr from the LSC, forming SrCrO4, seriously impaired the interconnect performance, preventing the joint use of these materials in ITSOFC and SOEC devices.