Comparison of the Pitting Corrosion Resistance of Bainite and Martensite in Fe-0.4C-1.5Si-2Mn Steel

Abstract

Specimens with different microstructures (bainite, as-quenched martensite, and tempered martensite) were fabricated using a Fe-0.4C-1.5Si-2Mn steel sheet, and the pitting corrosion resistances of these microstructures were compared. Retained austenite was barely detected in the X-ray diffraction analysis. The Vickers hardness values of the microstructures were ordered as (high) as-quenched martensite > tempered martensite ≈ bainite in the 325°C-austempered specimen > bainite in the 425°C-austempered specimen (low). The pitting corrosion resistance of each microstructure was evaluated by potentiodynamic polarization in boric-borate buffer solutions containing NaCl (pH 8.0) under naturally aerated conditions. The pitting corrosion resistances of the microstructures were ordered as (high) as-quenched martensite > bainite in the 325°C-austempered specimen > tempered martensite > bainite in the 425°C-austempered specimen (low). The lower active dissolution rates of the microstructures were determined to provide superior pitting corrosion resistance.

1. Introduction

Martensitic and bainitic transformations have been used in the development of high-strength and/or high-ductility steels.^1,2,3,4) Improving the properties of steel by controlling these transformations requires an in-depth understanding of the relationship between the properties and microstructures. In particular, the relationship between the pitting corrosion resistance and microstructure has yet to be well investigated compared to that of mechanical properties and microstructure. Steels containing bainite have recently been intensively developed; hence, there is an urgent need to investigate the pitting corrosion resistance of bainitic microstructures. In this study, the pitting corrosion resistances of bainite and martensite in Fe-0.4C-1.5Si-2Mn steel were compared.

Although carbon steels can be successfully protected against pitting corrosion in chloride environments by coating and/or painting, cut edges and coating defects are vulnerable to corrosion damage. Therefore, comprehending the relationship between the microstructure and pitting corrosion resistance of carbon steels becomes imperative for extending their service life and minimizing the maintenance cost associated with steel structures.

The pitting corrosion resistance of carbon steels is often evaluated using NaCl solutions, such as 3.5 mass% NaCl.^5,6,7,8) The use of solutions with excessive chloride ion concentrations and pH values below the depassivation pH of carbon steels (approximately pH 7) has hindered the understanding of the relationship between the microstructure and the pitting corrosion resistance. Solutions with low chloride ion concentrations and approximately pH 8, in which carbon steels are spontaneously passivated, are suitable for understanding the pitting corrosion resistance of microstructures.^9,10,11) For example, Kadowaki et al. discovered that the pitting corrosion resistance of typical microstructures of AISI 1045 carbon steel was ordered as (high) as-quenched martensite > tempered martensite ≈ primary ferrite > pearlite (low).^10,11) However, the pitting corrosion resistance of bainite in carbon steels is yet to be determined. Nishimoto et al. reported that a microstructure comprising bainite and degenerate pearlite exhibited higher pitting corrosion resistance than tempered martensite and ferrite-pearlite microstructures.¹²⁾ However, to the best of our knowledge, no study has compared the pitting corrosion resistance of bainite and martensite.

In this study, specimens composed of bainite, as-quenched martensite, and tempered martensite were prepared using Fe-0.4C-1.5Si-2Mn steel. The pitting corrosion resistance of each microstructure was compared using potentiodynamic polarization in near-neutral pH solutions containing chloride ions.

2. Experimental

2.1. Specimens

An Fe-0.4C-1.5Si-2Mn steel sheet was used in this study; its chemical composition is listed in Table 1. The steel sheet was cut into approximately 15 × 12 × 1.5 mm coupons.

Table 1. Chemical composition of the Fe-0.4C-1.5Si-2Mn steel sheet (mass%).

C	Si	Mn	P	S	Al	N	O
0.39	1.51	2.01	0.010	0.002	0.04	0.004	<0.001

To obtain specimens containing bainitic structures, one set of coupons was heat-treated using a Thermecmastor-Z thermomechanical processing simulator (Fuji Electronic Industrial Co., Ltd.). The coupons were austenitized at 1050°C for 0.5 h under vacuum and subsequently cooled to either 425°C or 325°C at a cooling rate of 25°C s^–1 using N₂ gas, followed by isothermal holding for 0.5 h. Finally, the specimens were cooled to a room temperature using N₂ gas. The coupons were not compressed during the heat treatment. The specimens isothermally held at 425°C and 325°C are referred to as “425°C-austempered” and “325°C-austempered” specimens, respectively, in this paper.

Another set of coupons was heat-treated in an electric furnace to obtain martensitic structures. The coupons were austenitized at 1050°C for 0.5 h under vacuum and quenched in water. The specimens prepared using this method are referred to as “water-quenched” specimens in this paper. To obtain a tempered martensitic structure, the water-quenched specimens were heat-treated at 425°C for 0.5 h and quenched in water. The specimens prepared using this method are referred to as “WQ-tempered” specimens in this paper.

After heat treatment, the specimen surfaces were ground using a milling machine to eliminate the decarburized layers. Subsequently, the surfaces were ground using a series of SiC papers up to 1500 grit, followed by polishing using 6 and 1 μm diamond pastes. Finally, the specimens were rinsed with ethanol.

2.2. Potentiodynamic Anodic Polarization

To evaluate the pitting corrosion resistance in chloride environments, potentiodynamic anodic polarization was conducted in boric-borate buffer solutions with NaCl additions (pH 8.0) at 25°C under naturally aerated conditions. The solutions were prepared by mixing NaCl-added 0.35 M H₃BO₃ and NaCl-added 75 mM Na₂B₄O₇. The NaCl concentrations were 5 and 10 mM. Polarization curves were recorded once for each condition. Measurements were performed on an electrode area of approximately 64 mm² (8 mm × 8 mm) using a three-electrode cell. A platinum plate was used as the counter electrode, and the reference electrode was an Ag/AgCl (3.33 M KCl) electrode (0.206 V vs. standard hydrogen electrode (SHE) at 25°C). All potentials reported herein refer to those of the SHE. The scan rate of the electrode potential was 20 mV min^–1.

To analyze the intrinsic dissolution behavior without the influence of passivation, potentiodynamic polarization curves were measured in a 0.35 M H₃BO₃-10 mM NaCl-0.1 M Na₂SO₄ solution (pH 4.5) at 25°C under deaerated conditions. In this solution, Na₂SO₄ was added to ensure electroconductivity. Prior to the potentiodynamic polarization, a cathodic treatment (applying a constant potential at approximately −1.0 V for 300 s) was conducted to remove any air-formed oxide films on the specimens.

2.3. Microstructure Characterization

X-ray diffraction (XRD) analysis using Cu-Kα radiation (Kα1: 1.5405 Å and Kα2: 1.5443 Å) with a Ni filter was conducted. The XRD data were collected with a step width of 0.02°. The scan speed was 2° min⁻¹. The Kα2 peaks were stripped (stripping ratio Kα2/Kα1 = 0.5) from the scans using computer software.

In order to observe the specimen surfaces, field-emission scanning electron microscopy (FE-SEM) equipped with energy-dispersive X-ray spectroscopy (EDS) was used, capturing secondary electron images under an accelerating voltage of 20 kV. The specimens were etched with 3 vol% nital to reveal their microstructures.

The Vickers hardness values of the specimens were measured under an applied load of 0.2 N. The measurements were conducted five times for each specimen.

3. Results and Discussion

3.1. Microstructure of the Specimens

Bainite was produced by austempering between the B_s and M_s points. The bainite transformation temperature (B_s) is determined by the concentration of the alloying elements:^13,14)

  

$$\begin{array}{l}{B}_{\text{s}}\text{[}°\text{C]}=656-57.7\times (\text{\%C})-75\times (\text{\%Si})-35\times (\text{\%Mn})\\ -15.3\times (\text{\%Ni})-34\times (\text{\%Cr})-41.2\times (\text{\%Mo})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }(\mathrm{i}\mathrm{n}\mathrm{\hspace{0.17em} }\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\%)\end{array}$$(1)

The martensite transformation temperature (M_s) is determined using the following formula:¹⁵⁾

  

$$\begin{array}{l}{M}_{\text{s}}\text{[}°\text{C]}=499-308\times (\text{\%C})-10.8\times (\text{\%Si})-32.4\times (\text{\%Mn})\\ -16.2\times (\text{\%Ni})-27\times (\text{\%Cr})-10.8\times (\text{\%Mo})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }(\mathrm{i}\mathrm{n}\mathrm{\hspace{0.17em} }\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\%)\end{array}$$(2)

According to Eqs. (1) and (2), the B_s and M_s points of the Fe-0.4C-1.5Si-2Mn steel were determined to be 450°C and 297°C, respectively. The specimens in this study were austempered at either 425°C or 325°C.

Figure 1 shows SEM images of the specimens after nital etching. In the 425°C-austempered specimen (Fig. 1(a)), a microstructure with elongated lath structures and cementite phases along the lath boundaries was observed. It has been proposed that bainite is classified by the morphology of ferrite: lath-like and plate-like morphologies are upper bainite and lower bainite, respectively.^16,17,18) According to this morphology-based classification, the microstructure of the 425°C-austempered specimen is supposed to be upper bainite. In the 325°C-austempered specimen (Fig. 1(b)), the microstructure was also composed of lath structures. The austempering at 325°C likely resulted in the formation of finer blocks and packets of upper bainite than the 425°C-austempered specimen. The microstructure of the water-quenched specimen (Fig. 1(c)) was entirely martensitic. In addition to the above three specimens, WQ-tempered specimens were also prepared because martensitic steels are generally tempered to balance their strength and toughness. Tempering was conducted at 425°C for 0.5 h. Figure 1(d) shows that the WQ-tempered specimen exhibited a tempered martensitic structure.

Fig. 1. Secondary electron images of the (a) 425°C-austempered, (b) 325°C-austempered, (c) water-quenched, and (d) WQ-tempered specimens after nital etching.

The additions of Si and Mn to carbon steel tend to promote retained austenite at room temperatures. The dissolution resistance of retained austenite in Fe-0.4C-1.5Si.2Mn steel in boric borate buffer solutions containing NaCl (pH 7.0) has been reported to be superior to that of a bainite matrix, and the retained austenite acts as a barrier against the propagation of pitting corrosion.¹⁹⁾ To confirm the presence of retained austenite, the phases in the specimens were identified using XRD, and the results are presented in Fig. 2. All specimens exhibited several peaks at 44.3, 64.6, and 81.9°, corresponding to the bcc (or bct) structure. However, few or no peaks of the fcc structure were detected. This implies that the amount of retained austenite was very small; therefore, its contribution to the pitting corrosion resistance was negligible.

Fig. 2. XRD patterns (Cu-Kα radiation) of the 425°C-austempered, 325°C-austempered, and water-quenched specimens. (Online version in color.)

Figure 3 shows the Vickers hardness values of the specimens. The hardnesses of the 425°C- and 325°C-austempered specimens were approximately 360 and 500 HV, respectively. The water-quenched specimen exhibited the highest hardness (800 HV); however, after tempering at 425°C for 0.5 h, the hardness decreased to approximately 500 HV, comparable to that of the 325°C-austempered specimen.

Fig. 3. Vickers hardness of the specimens. The measurements were conducted five times for each specimen, and the average values are plotted with error bars indicating the maximum and minimum values.

3.2. Evaluation of the Pitting Corrosion Resistance

The pitting corrosion resistance of each specimen was investigated by potentiodynamic polarization in boric-borate buffer solutions containing NaCl (pH 8.0). Figure 4 shows the polarization curves of the specimens. As shown in Fig. 4(a), the experiments were conducted with an NaCl concentration of 10 mM. Potentiodynamic polarization began at 0.02 V, and cathodic currents flowed first. The anodic currents were measured after the cathodic currents reached zero. The anodic current density gradually increased to the level of the passive state of carbon steels in near-neutral pH solutions.^12,20) Current oscillations in the passive state indicate metastable pitting events, and the subsequent continuous increase in the current density was attributed to the initiation of stable pitting. The pitting potential in this study was defined as the critical potential at which the current density exceeds 0.1 A m⁻². The pitting potential of the water-quenched specimen (black curve) was higher than those of the other specimens. The pitting potential of the WQ-tempered specimen (gray curve) was 0.103 V. The pitting corrosion resistance of the tempered martensite was lower than that of the as-quenched martensite. The pitting potential of the 325°C-austempered specimen (red curve) was 0.109 V, which suggests that the pitting corrosion resistance of bainite in the 325°C-austempered specimen was lower than that of as-quenched martensite and higher than that of tempered martensite. The pitting potential of the 425°C-austempered specimen (blue curve) was 0.094 V, indicating that the pitting corrosion resistance of bainite in the 425°C-austempered specimen was lower than that of the other microstructures.

Fig. 4. Potentiodynamic polarization curves of the specimens measured in a boric-borate buffer solution with NaCl (pH 8.0, naturally aerated). The NaCl concentrations were (a) 10 mM and (b) 5 mM. (c) Pitting potentials of the specimens. (Online version in color.)

To confirm the reproducibility of the order of the pitting corrosion resistance of the microstructures, polarization was also conducted in a boric-borate buffer solution containing 5 mM NaCl, as shown in Fig 4(b). In addition to verifying reproducibility, the NaCl concentration was reduced to 5 mM in order to clarify the difference in the pitting potential values. The pitting potentials of the specimens increased when the NaCl concentration was reduced to 5 mM; however, the difference in the pitting potentials remained small. Nevertheless, the order was the same as that for 10 mM NaCl. Figure 5 shows SEM images of the pits after polarization. The pits in all specimens were round; however, the corrosion morphologies inside the pits differed. Figures 5(a) and 5(b) show that the inner surfaces of the pits in bainitic microstructures were rough. This was attributed to the superior corrosion resistance of cementite, as compared to that of ferrite.²¹⁾ The EDS analysis was conducted at points 1 and 2 in Fig. 5(b). The relative atomic ratio at point 1 was Fe:C:Si:Mn:O = 93:<1:3:3:<1. The relative atomic ratio at point 2 was Fe:C:Si:Mn:O = 27:48:1:<1:23. Although it is difficult to measure the exact C concentration by EDS analysis, it implies that cementite tended to remain inside the pit. The difference in the dissolution rates of the ferrite and cementite phases roughened the inner surface of the pits. As shown in Fig. 5(c), the surface inside the pit of the as-quenched martensite was relatively smooth; however, the precipitation of cementite in tempered martensite (Fig. 5(d)) resulted in a rough surface inside the pit. The different corrosion morphologies suggest that the pitting potentials are affected by the microstructure. The pitting potentials of the specimens are presented in Fig. 4(c). The pitting corrosion resistance of the microstructures followed the order of: (high) as-quenched martensite > bainite in the 325°C-austempered specimen > tempered martensite > bainite in the 425°C-austempered specimen (low).

Fig. 5. Secondary electron images of the (a) 425°C-austempered, (b) 325°C-austempered, (c) water-quenched, and (d) WQ-tempered specimens after polarization in a boric-borate buffer solution containing 5 mM NaCl.

During the early stages of the pitting corrosion process, the active dissolution rate of steel would determine whether a transition from metastable pitting to stable pit growth occurs. To analyze the intrinsic dissolution rates of the specimens, potentiodynamic polarization curves were measured in a 0.35 M H₃BO₃-10 mM NaCl-0.1 M Na₂SO₄ solution (pH 4.5). Figure 6 shows the polarization curves of the specimens. Since cathodic treatment was conducted at −1.0 V for 300 s prior to the scanning of the electrode potential, surface oxide films were removed before the polarization curves were measured. Figure 6(b) shows the magnified polarization behavior near the corrosion potential, which shows that the corrosion potential varied for each specimen. Little or no difference was observed on the cathodic side of the polarization curves, indicating that the higher corrosion potential was mainly attributed to a decrease in the active dissolution rates. The corrosion potentials of the microstructures in the pH 4.5 solution were ordered as (high) as-quenched martensite > bainite in the 325°C-austempered specimen > tempered martensite > bainite in the 425°C-austempered specimen (low), which is consistent with the order of the pitting potentials.

Fig. 6. (a) Potentiodynamic polarization curves of the specimens measured in a 0.35 M H₃BO₃-10 mM NaCl-0.1 M Na₂SO₄ solution (pH 4.5, deaerated). (b) Enlarged view in the potential range of −0.55 V to −0.4 V. (Online version in color.)

As shown in Fig. 4(c), the pitting corrosion resistance of the as-quenched martensite was higher than those of the other microstructures. The higher pitting corrosion resistance of the as-quenched martensite may be attributed to interstitial carbon. The order of the interstitial carbon concentration was assumed to be (high) as-quenched martensite > tempered martensite ≥ bainitic ferrite (low). Kadowaki et al. reported that interstitial carbon reduces the active dissolution rate of martensite in AISI 1045 carbon steel and improves the pitting corrosion resistance.^11,22) Li et al. reported that the bonds between interstitial carbon and metal atoms in stainless steels are 1.4–2.0 times stronger than the metal-metal bonds.²³⁾ It has also been reported that the pitting potential of tempered martensite decreases as the interstitial carbon concentration decreases with tempering.²⁴⁾ Therefore, the superior pitting corrosion resistance of the as-quenched martensite, as compared to the other microstructures, was attributed to a higher concentration of interstitial carbon. The lower pitting corrosion potential of tempered martensite was attributed to the lower interstitial carbon content resulting from tempering.

The solubility limit of carbon in ferrite is considerably low; therefore, the difference in the concentrations of interstitial carbon in the bainitic ferrite of the 325°C- and 425°C-austempered specimens seems insignificant. However, bainite in the 325°C-austempered specimen exhibited superior pitting corrosion resistance than bainite in the 425°C-austempered specimen, as shown in Fig. 4(c). The difference between the pitting corrosion resistances of bainite of the 325°C- and 425°C-austempered specimens may be related to the precipitation morphology of the cementite. Because the dissolution resistance of cementite is reportedly higher than that of ferrite, cementite likely acts as a barrier against the propagation of pitting corrosion.¹⁰⁾ As shown in Fig. 1, cementite appears to be more finely precipitated in bainite in the 325°C-austempered specimen than in bainite in the 425°C-austempered specimen. The fine precipitation morphology of cementite may contribute to the improvement in pitting corrosion resistance; however, the corresponding details remain unclear and should be explored in future research. In summary, this study determined the order of the pitting corrosion resistance of bainite, as-quenched martensite, and tempered martensite in Fe-0.4C-1.5Si-2Mn steel. The pitting corrosion resistance was determined to vary with the interstitial carbon content and precipitation morphology of the carbides.

4. Conclusions

Specimens with different microstructures (bainite, as-quenched martensite, and tempered martensite) were fabricated using a Fe-0.4C-1.5Si-2Mn steel sheet to investigate the pitting corrosion resistances of the microstructures. Retained austenite was barely detected in the XRD analysis. The Vickers hardness values of the microstructures were ordered as (high) as-quenched martensite > tempered martensite ≈ bainite in the 325°C-austempered specimen > bainite in the 425°C-austempered specimen (low).

The pitting corrosion resistance of each microstructure was evaluated by potentiodynamic polarization in boric-borate buffer solutions containing NaCl (pH 8.0) under naturally aerated conditions. The pitting corrosion resistances of the microstructures were ordered as (high) as-quenched martensite > bainite in the 325°C-austempered specimen > tempered martensite > bainite in the 425°C-austempered specimen (low).

Acknowledgments

This study was supported by the 31st ISIJ Research Promotion Grant from the Iron and Steel Institute of Japan. This work was also supported by JSPS KAKENHI Grant Number JP21K14430.

References

• 1)   X.  Zhang,  G.  Miyamoto,  Y.  Toji,  Y.  Zhang and  T.  Furuhara: Acta Mater., 209 (2021), 116772. https://doi.org/10.1016/j.actamat.2021.116772
• 2)   N.  Sakaguchi,  H.  Ohno and  N.  Nakada: ISIJ Int., 62 (2022), 586. https://doi.org/10.2355/isijinternational.ISIJINT-2021-478
• 3)   T.  Ogawa,  N.  Maruyama,  N.  Sugiura and  N.  Yoshinaga: ISIJ Int., 50 (2010), 469. https://doi.org/10.2355/isijinternational.50.469
• 4)   S.  Takagi,  Y.  Toji,  M.  Yoshino and  K.  Hasegawa: ISIJ Int., 52 (2012), 316. https://doi.org/10.2355/isijinternational.52.316
• 5)   J.  Yang,  Y.  Lu,  Z.  Guo,  J.  Gu and  C.  Gu: Corros. Sci., 130 (2018), 64. https://doi.org/10.1016/j.corsci.2017.10.027
• 6)   Y.-B.  Guo,  C.  Li,  Y.-C.  Liu,  L.-M.  Yu,  Z.-Q.  Ma,  C.-X.  Liu and  H.-J.  Li: Int. J. Miner. Metall. Mater., 22 (2015), 604. https://doi.org/10.1007/s12613-015-1113-z
• 7)   Y.  Wang,  G.  Cheng and  Y.  Li: Corros. Sci., 111 (2016), 508. https://doi.org/10.1016/j.corsci.2016.05.037
• 8)   P. K.  Katiyar,  S.  Misra and  K.  Mondal: Metall. Mater. Trans. A, 50A (2019), 1489. https://doi.org/10.1007/s11661-018-5086-1
• 9)   K.  Fushimi,  R.  Nakagawa,  Y.  Kitagawa and  Y.  Hasegawa: J. Electrochem. Soc., 166 (2019), C3409. https://doi.org/10.1149/2.0461911jes
• 10)   M.  Kadowaki,  I.  Muto,  Y.  Sugawara,  T.  Doi,  K.  Kawano and  N.  Hara: J. Electrochem. Soc., 164 (2017), C261. https://doi.org/10.1149/2.0801706jes
• 11)   M.  Kadowaki,  I.  Muto,  Y.  Sugawara,  T.  Doi,  K.  Kawano and  N.  Hara: J. Electrochem. Soc., 164 (2017), C962. https://doi.org/10.1149/2.0541714jes
• 12)   M.  Nishimoto,  I.  Muto,  T.  Doi,  K.  Kawano and  Y.  Sugawara: Mater. Corros., 74 (2023), 724. https://doi.org/10.1002/maco.202213658
• 13)   S.  Kang,  S.  Yoon and  S.-J.  Lee: ISIJ Int., 54 (2014), 997. https://doi.org/10.2355/isijinternational.54.997
• 14)   J. S.  Kirkaldy and  D.  Venugopalan: Phase Transformations in Ferrous Alloys, ed. by A. R. Marder and J. I. Goldstein, TMS-AIME, Warrendale, PA, (1984), 125.
• 15)   C. Y.  Kung and  J. J.  Rayment: Metall. Trans. A, 13A (1982), 328. https://doi.org/10.1007/BF02643327
• 16)   Y.  Ohmori and  T.  Maki: Mater. Trans., JIM, 32 (1991), 631. https://doi.org/10.2320/matertrans1989.32.631
• 17)   Y.  Ohmori: ISIJ Int., 41 (2001), 554. https://doi.org/10.2355/isijinternational.41.554
• 18)   K.  Tsuzaki,  K.  Fujiwara and  T.  Maki: Mater. Trans., JIM, 32 (1991), 667. https://doi.org/10.2320/matertrans1989.32.667
• 19)   M.  Kadowaki,  I.  Muto,  Y.  Sugawara and  N.  Hara: Corros. Sci., 200 (2022), 110251. https://doi.org/10.1016/j.corsci.2022.110251
• 20)   M.  Nishimoto,  I.  Muto and  Y.  Sugawara: ISIJ Int., 62 (2022), 750. https://doi.org/10.2355/isijinternational.ISIJINT-2021-567
• 21)   M.  Kadowaki,  I.  Muto,  K.  Takahashi,  T.  Doi,  H.  Masuda,  H.  Katayama,  K.  Kawano,  Y.  Sugawara and  N.  Hara: J. Electrochem. Soc., 166 (2019), C345. https://doi.org/10.1149/2.1311912jes
• 22)   M.  Kadowaki,  A.  Saengdeejing,  I.  Muto,  Y.  Chen,  T.  Doi,  K.  Kawano,  Y.  Sugawara,  T.  Doi,  K.  Kawano and  N.  Hara: J. Electrochem. Soc., 168 (2021), 111503. https://doi.org/10.1149/1945-7111/ac3276
• 23)   T.  Li,  S.-C.  Chien,  Z.  Ren,  W.  Windl,  F.  Ernst and  G. S.  Frankel: Acta Mater., 221 (2021), 117433. https://doi.org/10.1016/j.actamat.2021.117433
• 24)   M.  Kadowaki,  I.  Muto,  Y.  Sugawara,  T.  Doi,  K.  Kawano and  N.  Hara: J. Electrochem. Soc., 165 (2018), C711. https://doi.org/10.1149/2.0071811jes