2023 年 64 巻 10 号 p. 2376-2385
Alkaline water electrolysis (AWE) is a hydrogen manufacturing process that generates “green hydrogen” using electricity derived from renewable energies. Stainless steel (SS), specifically austenitic SS, has recently attracted attention as an anode material for the oxygen evolution reaction (OER) of the AWE. SS anode surfaces are generally activated by generating surface catalyst layers (SCL) for the OER through specific chemical pre-treatment, although the precise chemical compositions and microstructures of the SCL remain under debate. Furthermore, because fluctuations in the electrode potential derived from renewable energies cause remarkable elution of the constituent elements into the electrolyte, corrosion behaviors of the SS anodes should be clarified. This review introduces the recent progress of the SS anodes, particularly in the context of surface treatments to generate surface catalyst layers with high OER performances under simulated AWE conditions. In general, recent reports have clearly shown that surface-treated SS anodes are superior to the commonly employed Ni-based anodes for AWE applications.
Hydrogen is a next-generation energy medium to resolve global climate issues reducing CO2 emissions. Hydrogen manufactured by water electrolysis using electricity generated by renewable energies, such as photovoltaics and wind power is known as “green hydrogen”. Such green hydrogen is considered to be a clean energy source because CO2 emissions are negligible during its production. Therefore, the application of green hydrogen in various fields, such as the chemical, iron and steel making, and transportation industries, has received growing interest.1–3) However, further research and development is required for the widespread application of green hydrogen. More specifically, energy efficiency of the water electrolysis process used to generate hydrogen should be improved significantly.
The water electrolysis processes operated at temperatures close to room temperature can be classified into two main types,4) namely polymer electrolyte membrane water electrolysis (PEMWE) and alkaline water electrolysis (AWE). PEMWE is characterized by high current densities that allow smaller electrolysis stack; however, noble metal and noble metal oxides are required as the electrocatalyst for both cathode and anode, respectively. In contrast, AWE (Fig. 1) allows the use of more cost-effective materials (e.g., nickel, and iron) for the cathode and anode. The two-type water electrolysis can therefore be differentiated in terms of their electrolyzer sizes and material costs.5,6)
Schematic representation of the alkaline water electrolysis and the key features of the SS anodes.
In both PEMWE and AWE, the overpotential of the oxygen evolution reaction (OER) at the anode is significantly larger than that of the hydrogen evolution reaction (HER) at the cathode. In other words, the OER is the rate-limiting step and so the development of a highly OER-active and durable anode is key to enhancing the energy efficiency of hydrogen manufacturing. From the viewpoint of the OER electrocatalytic activity and corrosion resistance against the strongly alkaline environments of AWE, a mesh and a foam Ni-based current collector electrode with the OER-active surface catalyst layer of the Ni-containing metal oxide, are currently adopted for the anode of the commercial AWE electrolyzer.6–8) To improve the current efficiency and to reduce the AWE electrolyzer costs, the development of more cost-effective anode materials with low OER overpotentials and high corrosion resistance properties is essential. Thus, to date, various metals and alloys,9,10) metal oxides,11–17) hydroxides,18–22) and metal sulfides,23–26) have been investigated for AWE anode materials.
Stainless steel (SS) has recently attracted attention as an AWE anode material for replacing the currently used Ni-based anodes.9–12) The SS anode is not only suitable as an effective current collector,9–13,27) but it also reduces the OER overpotential by generating an OER-active surface catalyst layer (SCL) following specific (electro)chemical pre-treatments.28–40) Previous studies on the SS anode (OER) and cathode (HER) have been summarized in a number of review articles,41–43) in which various surface treatments are listed for activation of the OER/HER. In recent years, remarkable progress has been achieved in the area of SS anode materials, including in the development of effective surface treatment methods for generating OER-active SCLs, and in the investigation of their detailed chemical and morphological structures on the nanometer and atomic scales. Furthermore, in terms of the practical application of SS anodes for AWE, dissolution of the constituent elements has been also reported under practical operating conditions, such as a high current density44,45) and power fluctuation.7,46)
Thus, the aforementioned recent progress in SS anodes is summarized in this review. Particular focus is placed on OER-active SCL generation by specific chemical pre-treatment, and the corrosion behaviors of SS anodes under potential cycling simulating power fluctuations of renewable energies. Finally, the prospects of the SS anode for green hydrogen manufacturing by AWE are discussed.
During the development of a suitable SS anode, austenitic SS has been examined to a greater extent than ferritic and martensitic SS owing to its high OER activity and corrosion resistance.47) The origin of the superior OER properties of austenitic SS stems from the specific chemical and morphological microstructures generated on the electrode surface by (electro)chemical pre-treatment. Indeed, in most cases, Ni- and Fe-containing oxides and (oxy)hydroxides are generated by such treatment.28,41,48) In particular, electrochemical oxidation in an alkaline solution generates surface Ni-based (oxy)hydroxides on the SS surface, which act as an OER-active SCL. More specifically, a portion of the Fe present in the SS electrode is incorporated into the OER-active Ni-based (hydro)oxide SCL, thereby contributing to its enhanced OER properties.13,28,32,49) Based on the results presented in the aforementioned studies, it became apparent that the OER activities of NiFe hydroxides are strongly affected by the Ni/Fe atomic ratio.18,22,28,49–52) However, the other constituent elements of austenitic SS, including Cr, Mn and Mo, are easily eluted into the electrolyte under AWE conditions; therefore, such elements do not contribute to enhancing the OER activity of the SCL.
Hydroxides containing both Ni and Fe, and in particular NiFe layered double hydroxides (NiFe-LDH), are known as OER-active electrocatalysts in alkaline solutions.19,20,22,51,53–62) To date, two OER mechanisms have been proposed for NiFe-LDH, namely the adsorption evolution mechanism (AEM) and the lattice oxygen participation mechanism (LOM).19,20,53,63) In the AEM, the OER proceeds via the adsorption and desorption of OER-related species (OH, O, etc.) on the metal element sites (left-hand cycle, Fig. 2(a)). In contrast, in the LOM, the lattice oxygen of the NiFe-LDH is involved in the OER (right-hand cycle, Fig. 2(a)). For example, density functional theory (DFT) calculations for the OER of NiFe-LDH suggest that the on-top site of the Fe atom adjacent to Ni, or the bridge site between Ni and Fe, correlates with reduction of the OER overpotential (i.e., higher OER activity) via the AEM (Fig. 2(b)).19,64) Furthermore, Zhang et al. predicted by DFT calculations that the (010), (100), and $(\bar{2}10)$ planes of the γ-phase NiFe-LDH, which correspond to the edges of the (001) basal plane (Fig. 2(c)), show lower OER overpotentials through the AEM (Fig. 2(d)). Indeed, they experimentally confirmed that the OER overpotential of the NiFe-LDH decreases with increasing the edge ratio of the LDH.19) On the other hand, Dionigi et al. deduced theoretically that the OER proceeds on these LDH edges via the LOM.20) They proposed that the potential limiting step in the LOM is the oxidation of adsorbed OH to adsorbed O, and the bridge site between Ni and Fe is responsible for the higher OER activity.
(a) Illustration of the competition between the adsorbate evolution mechanism (AEM) and the lattice-oxygen participation mechanism (LOM). Reprinted with permission.63) Copyright 2018 American Chemical Society. (b) Theoretical calculations based on density functional theory were implemented to explore the active sites of γ-phase NiFe-LDH for OER. Reprinted with permission.64) Copyright 2015 American Chemical Society. (c) The Wulff shape of the NiFe-LDH. (d) 2D volcano plot between the calculated overpotential based on the AEM and the descriptor of ΔGO–ΔGOH. Reprinted with permission.19) Copyright 2021 Wiley-VCH.
At present, although no definitive conclusion has been reached for the OER mechanism on NiFe LDH, the presence of neighboring Ni and Fe sites is expected to be key to ensuring a high OER activity. Therefore, regardless of the proposed mechanism, the OER activity of the NiFe hydroxide SCL generated on the SS anode is dominated by the chemical compositions and microstructures. Accordingly, the chemical and microstructural fine-tuning of these species is essential for achieving enhanced OER activities through the appropriate surface pre-treatment of SS.
Various surface pretreatments have been applied to SS electrodes to generate OER-active SCLs, as summarized in Table 1 for austenitic SS.10,11,28–41,49,65–67) For example, Zhong et al. used 316 SS plates as the starting anode material (Figs. 3(a) and 3(d)), and the OER-active SCL was synthesized by combining hydrothermal and electrochemical anodization treatments. Initially, the SS surface was corroded by hydrothermal treatment in an aqueous ammonia solution to increase the surface area of the pristine SS (Figs. 3(b) and 3(e)). Subsequently, electrochemical oxidation was conducted to generate an OER-active electrocatalyst (Figs. 3(c) and 3(f)).68) As shown in Fig. 3(g), the OER overpotentials of the surface-treated SS were lower in the following order: hydrothermally-treated and electrochemically oxidized SS (SPN50) < hydrothermally-treated SS (SPN; aimed at increasing electrochemical surface area) < electrochemically oxidized SS (SP50; for OER-active SCL synthesis) < non-treated SS (SP). The above trend suggests that surface treatments aimed at generating high-surface-area OER-active SCL are key to enhancing the OER performances of SS anodes.
(a)–(c) Schematic representation of the fabrication process used to obtain the rusty stainless steel plate. 3D AFM images of (a) the pristine SS plate, (b) the corroded SS plate, and (c) the EORC-activated rusty SS plate. SEM images of (d) the SS plate, (e) the corroded SS plate, and (f) the EORC-activated rusty SS plate (inset: the corresponding photographic images). (g) Polarization curves for OER of the rusty SS plates. Reprinted with permission.68) Copyright 2016 Wiley-VCH. (h) Schematic illustration of the fabrication process used to obtain the SS electrodes subjected to chemical etching, thermal oxidization, and plasma treatment. (i) Cross sectional TEM images of OESSC sample. (j) Polarization curves of the SS, ESS, OESS and OESSC samples for the OER. Reprinted with permission.33) Copyright 2019 Elsevier.
In addition, as shown in Fig. 3(h), Lyu et al. attempted to form a high-surface-area OER-active SCL on 316 L SS by combining chemical and plasma treatments.33) They etched the SS in an HCl solution to generate oxides of the constituent elements (mainly Fe oxides) with a high surface area. Subsequently, CH4 plasma treatment was applied to the etched SS. The resulting OER-active SCL was a mixture of graphene-encapsulated iron carbide particles (C@Fe3C) and an amorphous carbon layer (a-C layer) (Fig. 3(i)). Figure 3(j) shows the polarization curves of the treated samples, indicating that the lowest OER overpotential was achieved for the combined surface-treated sample (oxidized and etched stainless steel with CH4 treatment; OESSC), which therefore represents the most OER-active SS among the tested specimens. Furthermore, the OESSC exhibited a higher corrosion resistance than the non-treated pristine SS. Indeed, the surface composition of Fe, evaluated by X-ray photoelectron spectroscopy (XPS), remained essentially unchanged after constant potential electrolysis at 1.56 V vs. RHE. Given these results, the authors concluded that the high OER performance of OESSC can be attributed to coating of the Fe3C catalyst particles with graphene.
Using SS316L mesh as the starting anode material, Gomaa et al. reported morphological changes following various surface pretreatments, and subsequently, they evaluated the corresponding OER properties of the obtained anodes. Their method for generating an OER-active SCL involves anodization of the SS in a solution of ammonium fluoride in ethylene glycol, followed by heat treatment under various gas streams (Fig. 4(a)).65) As shown by the morphological changes observed by scanning electron microscopy (SEM, Fig. 4(b)), anodization and subsequent heat treatment (ASS-x; x = Air, H2, O2) brought about a remarkable increase in the surface roughness of the starting SS mesh. Furthermore, the surface microstructures were found to depend on the stream gas employed during heat treatments, wherein the oxygen stream (ASS-O2) led to the largest surface area. Moreover, the polarization curves of the prepared samples (Fig. 4(c)) clearly revealed that ASS-O2 treatment gave the lowest OER onset potential, i.e., the highest OER activity.
(a) Schematic representation of the fabrication process used to form the OER-active catalysts on the SS 316 L mesh: (i) electrochemical anodization, (ii) annealing under a gas stream, and (iii) electrochemical characterization. (b) FESEM images of the SS-AR, ASS-Air, ASS-H2, and ASS-O2 samples. (c) OER polarization curves of the various SS mesh anodes recorded in a 1 M KOH solution. Reprinted with permission.65) Copyright 2022 American Chemical Society.
As mentioned above, OER-active SCLs can be generated by various electrochemical pretreatment approaches. Our group previously investigated the OER performance of the SCL generated on 316 SS by electrochemical oxidation (i.e., constant current density electrolysis (CCE), 30 mA cm−2, 1 M KOH, 75°C).49) Thus, Fig. 5(a) shows a cross-sectional scanning transmission electron microscopy (STEM) image of 316 SS subjected to the aforementioned CCE process for 5 h. A nanofiber-like structure with a width of 3–5 nm and length of ∼50 nm can be seen on a dense buffer layer with a thickness of ∼30 nm. The high-resolution STEM image shown in Fig. 5(b) reveals that these nanofibers were composed of a NiFe-LDH, and energy dispersive X-ray spectroscopic (EDS) analyses indicated that the nanofibers contained both Ni and Fe as constituent elements, while no signals due to Cr and Mn, both of which are constituent elements of 316 SS, were detected. The OER overpotential of the CCE-treated 316 SS (CCE-10h; red) was significantly lower than those of iridium oxide (IrOx)69) and the non-treated pristine 316 SS (316 L-SS),36,68) and was comparable to that of the aforementioned NiFe-LDH powder catalysts.70,71) Figure 5(d) provides a schematic summary of the time evolution of the NiFe nanostructured SCL during CCE treatment. It has also been reported that CCE treatment can be applied to generate nanostructured catalyst layers on 303 and 310S SS,28) and so overall, these reports clarify the correlation between the OER overpotential and the Fe/Ni atomic ratio in the SCL generated by CCE treatment.
Cross-sectional STEM images of the NiFe-hydroxide/oxide nanostructures generated on the 316SS substrate after constant current density electrolysis (CCE). The STEM images observed at (a) low and (b) high resolutions. (c) Comparison of the OER overpotential of the CCE-treated 316SS with those of previously reported catalysts. (d) Schematic representation of the time evolution of the hetero-layered nanostructures generated on the 316SS substrate under CCE conditions. Reprinted with permission.49) Copyright 2019 American Chemical Society.
In the practical hydrogen production by AWE, a current density of at least 400 mA cm−2 is desirable.6) Therefore, in recent years, OER performances have been investigated under high current densities of 400–1000 mA cm−2, which is close to the practical AWE condition.7,23,44,45,72–75) In contrast, the OER performances of surface-treated SSs have been mainly evaluated for electrolysis conducted under relatively low current densities (i.e., ∼10 mA cm−2),31,48,65,66) and so further investigations are required for the SCLs generated on SS to verify the feasibility of SS anodes.
Figure 6(a) shows the OER overpotential changes under constant current density electrolysis at 400 mA cm−2 in a 1 M KOH solution at both 20 and 75°C for the NiFe hydroxide/oxide (NiFe-HyOx) SCL generated on 316 SS. During continuous electrolysis for up to 20 h, the OER overpotential remained nearly constant, except at the beginning of the electrolysis process, thereby indicating that the SCL is tolerant to the high current density conditions. However, after electrolysis for 20 h, 30 and 80 mV increase in the overpotential were observed at temperatures of 20 and 75°C, respectively (Fig. 6(b)). STEM observations revealed that the nanofiber-like structured SCL remained after electrolysis, although the SCL decreased in thickness, probably due to (electro)chemical dissolution and/or physical detachment upon aggressive bubble formation under the electrolysis conditions (Fig. 6(c)).
Durability of the NiFe-HyOx/SS under CCE conditions at 400 mA cm−2: (a) Chronopotentiometry curves recorded at a constant current density of 400 mA cm−2 in a 1 M KOH solution at both 20 and 75°C; (b) polarization curves for the OER recorded at 20°C before and after the 20 and 75°C tests; the inset shows the OER overpotentials at 400 mA cm−2-geo; (c) Cross-sectional STEM images before and after electrolysis for 20 h at 20°C. Reprinted with permission.49) Copyright 2019 American Chemical Society.
Independently, Xiao et al. prepared an Se-doped NiFe oxyhydroxide SCL generated by the electrochemical oxidation treatment of 304 SS (Se-SS, Fig. 7(a)) and evaluated the OER properties.67) As shown in Fig. 7(b), Se-SS exhibited a superior OER overpotential compared to the non-treated (SS) and the treated but non-Se-doped SS (ESS) samples. Furthermore, Se-SS maintained a nearly constant potential during electrolysis at 500 mA cm−2 for 95 h, thereby demonstrating its practical feasibility for use as an AWE anode (Fig. 7(c)).
(a) Schematic illustration of the structural evolution of the Se-doped NixFe1−x–OOH catalyst layers on the SS substrate (Se-SS). (b) LSV curves recorded during the OER, and (c) the corresponding chronopotentiometric curves. The inset shows the LSV results for the Se-SS specimen before and after the stability test. Reprinted with permission.67) Copyright 2020 Elsevier.
During the generation of green hydrogen using electricity obtained from renewable energies, the inevitable power fluctuations need to be taken into account. Thus, the SS must be tolerant both to a high current density and to power fluctuations.76,77) For example, the reverse current78,79) which flows under the startup–shutdown conditions78,80) leads to degradation of the AWE anode and cathode. To prevent such degradation, protective current is applied to retain the electrode potentials close to the onset potentials of the anode (OER) and the cathode (HER) during shutdown.81) However, the use of protective current increases the electrolyzer costs. Thus, the development of durable electrode materials is underway,82–85) although few studies have been reported for the SS anode.13,29,30)
Previously, our group investigated the correlation between the microstructure of the NiFe-HyOx SCL generated on 316 SS (Fig. 4(a)) and the corresponding OER properties under potential cycling that simulated power fluctuations of renewable energies. As shown in Fig. 8(a), the OER overpotential remained relatively unchanged over 20,000 potential cycles of 0.5 and 1.8 V vs. RHE, which resulted in an increase in the OER overpotential for the pure Ni anode. The cross-sectional STEM images collected after 20,000 potential cycles (Figs. 8(b) and 8(c)) show that the OER-active, nanofiber-like SCL with a thickness of ∼50 nm remained essentially unchanged in its topmost surface region. However, potential cycling generated an amorphous Ni–Fe (hydro)oxide interlayer with a thickness of ∼850 nm between the SCL and the 316 SS substrate. These results suggest that corrosion of the SS substrate proceeds under potential cycle loading. More specifically, the repeated oxidation and reduction of the SS generated an amorphous interlayer of Ni–Fe (hydro)oxide. In addition, due to the depletion of Cr in the interlayer, the Cr present in the 316 SS should leach into the electrolyte. Cr exists as toxic hexavalent chromium (CrO42−) in a strong alkaline solution,86,87) and so reduction of Cr dissolution is desirable for practical green hydrogen manufacturing by AWE using SS-based anodes. As shown in Fig. 8(d), the dissolution of Fe and Cr was suppressed with lowering the Fe/Ni ratio of the starting SS, with the lowest degree of dissolution being observed for 310S SS. A cross-sectional structural observation of 310S (Fig. 8(e)) revealed that the 5 µm-thick oxide film consisting of Fe-rich hydroxides and oxides (Fe/Ni ratio, ∼10) exhibited numerous cracks and voids, wherein the 310S SS (Fig. 8(f)) showed a ∼1 µm-thick Ni-rich hydroxide film with nanopores. These results suggest that the different corrosion behaviors of SS can be attributed to differences in the microstructures of the oxide films generated as corrosion products (Fig. 8(g)).
(a)–(c) Electrochemical stability of the 316 SS electrodes under simulated potential fluctuation conditions: (a) Changes in the OER overpotentials of the NiFe-HyOx/SS during the potential cycles; the inset shows the applied potential cycle protocol. (b) Cross-sectional HAADF-STEM images of the NiFe-HyOx/316SS electrode after 20,000 potential cycles, and (c) a magnified image of the topmost surface region in (b). Reprinted with permission.30) Copyright 2021 Elsevier. (d) Amounts of dissolved Fe and Cr vs. the Fe/Ni atomic ratios of the various austenitic SS electrodes. Cross-sectional SEM images of (e) the 301 SS and (f) 310S SS electrodes collected after 20,000 potential cycles. (g) Schematic representation of the corrosion behaviors of the 301 and 310S SS electrodes during the potential cycle loadings. Reprinted with permission.29) Copyright 2022 Elsevier.
This review introduces recent progress in stainless steel anodes for use in alkaline water hydrolysis (AWE), focusing on the surface pre-treatment methods available to generate surface catalyst layers (SCLs) that are active for the oxygen evolution reaction (OER). The OER performances of the prepared anodes are also discussed under a constant current density and under power fluctuation conditions of renewable energies. Recent studies have clearly shown that the OER performances of SS anodes with SCL strongly depend on the chemical compositions and microstructures of the SCL at both the nanometer and atomic scales; in particular, the Ni–Fe hydroxides generated under specific treatment conditions are essential for enhancing the OER performance. As summarized in this review, the surface-treated SS anodes, whose surfaces are comprised of NiFe-hydroxide-based SCLs, exhibited superior OER performances compared to those of practical Ni-based anodes, and SS anodes with the SCL enabled AWE at current densities of 400–500 mA cm−2. In particular, Se doping of the NiFe-hydroxide SCL was found to be beneficial for high current density operations. Furthermore, the NiFe-hydroxide/oxide SCL was stable against simulated power fluctuation conditions. Although 310S SS possessed the highest corrosion resistance among the various austenitic SS specimens examined, leaching of its constituent elements (e.g., Fe and Cr) was detrimental to the electrolyzer performance; the dissolution of these major elements depends significantly on the ratios of the constituent elements. Such corrosion-induced degradations should therefore be mitigated by the careful selection of suitable SS grades and by fine-tuning of the minor elements. Currently, research and developments remain ongoing into the practical application of SS as an AWE anode. Ultimately, for the development of novel high-performance SS anodes, surface treatments should be optimized, and the corrosion mechanisms of the various SS grades should be clarified.
This study was partly supported by the New Energy and Industrial Technology Development Organization (NEDO), JSPS KAKENHI Grant Number 21H01661, and the Toyota Mobility Foundation Hydrogen Initiative.
